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
|
/* $OpenBSD: softraid_raidp.c,v 1.38 2013/03/30 14:41:37 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/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 P functions. */
int sr_raidp_create(struct sr_discipline *, struct bioc_createraid *,
int, int64_t);
int sr_raidp_assemble(struct sr_discipline *, struct bioc_createraid *,
int, void *);
int sr_raidp_alloc_resources(struct sr_discipline *);
int sr_raidp_free_resources(struct sr_discipline *);
int sr_raidp_rw(struct sr_workunit *);
int sr_raidp_openings(struct sr_discipline *);
void sr_raidp_intr(struct buf *);
void sr_raidp_set_chunk_state(struct sr_discipline *, int, int);
void sr_raidp_set_vol_state(struct sr_discipline *);
void sr_raidp_xor(void *, void *, int);
int sr_raidp_addio(struct sr_workunit *wu, int, daddr64_t, daddr64_t,
void *, int, int, void *);
void sr_dump(void *, int);
void sr_raidp_scrub(struct sr_discipline *);
void *sr_get_block(struct sr_discipline *, int);
void sr_put_block(struct sr_discipline *, void *, int);
/* discipline initialisation. */
void
sr_raidp_discipline_init(struct sr_discipline *sd, u_int8_t type)
{
/* Fill out discipline members. */
sd->sd_type = type;
if (sd->sd_type == SR_MD_RAID4)
strlcpy(sd->sd_name, "RAID 4", sizeof(sd->sd_name));
else
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_RAIDP_NOWU;
/* Setup discipline specific function pointers. */
sd->sd_alloc_resources = sr_raidp_alloc_resources;
sd->sd_assemble = sr_raidp_assemble;
sd->sd_create = sr_raidp_create;
sd->sd_free_resources = sr_raidp_free_resources;
sd->sd_openings = sr_raidp_openings;
sd->sd_scsi_rw = sr_raidp_rw;
sd->sd_scsi_intr = sr_raidp_intr;
sd->sd_set_chunk_state = sr_raidp_set_chunk_state;
sd->sd_set_vol_state = sr_raidp_set_vol_state;
}
int
sr_raidp_create(struct sr_discipline *sd, struct bioc_createraid *bc,
int no_chunk, int64_t coerced_size)
{
if (no_chunk < 3)
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 &
~((sd->sd_meta->ssdi.ssd_strip_size >> DEV_BSHIFT) - 1)) *
(no_chunk - 1);
return 0;
}
int
sr_raidp_assemble(struct sr_discipline *sd, struct bioc_createraid *bc,
int no_chunk, void *data)
{
return 0;
}
int
sr_raidp_openings(struct sr_discipline *sd)
{
return (sd->sd_max_wu >> 1); /* 2 wu's per IO */
}
int
sr_raidp_alloc_resources(struct sr_discipline *sd)
{
int rv = EINVAL;
DNPRINTF(SR_D_DIS, "%s: sr_raidp_alloc_resources\n",
DEVNAME(sd->sd_sc));
if (sr_wu_alloc(sd))
goto bad;
if (sr_ccb_alloc(sd))
goto bad;
/* setup runtime values */
sd->mds.mdd_raidp.srp_strip_bits =
sr_validate_stripsize(sd->sd_meta->ssdi.ssd_strip_size);
if (sd->mds.mdd_raidp.srp_strip_bits == -1)
goto bad;
rv = 0;
bad:
return (rv);
}
int
sr_raidp_free_resources(struct sr_discipline *sd)
{
int rv = EINVAL;
DNPRINTF(SR_D_DIS, "%s: sr_raidp_free_resources\n",
DEVNAME(sd->sd_sc));
sr_wu_free(sd);
sr_ccb_free(sd);
rv = 0;
return (rv);
}
void
sr_raidp_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_raidp_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_raidp_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;
}
int
sr_raidp_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;
int s, i;
daddr64_t blk, lbaoffs, strip_no, chunk, row_size;
daddr64_t strip_size, no_chunk, lba, chunk_offs, phys_offs;
daddr64_t strip_bits, length, parity, strip_offs, datalen;
void *xorbuf, *data;
/* blk and scsi error will be handled by sr_validate_io */
if (sr_validate_io(wu, &blk, "sr_raidp_rw"))
goto bad;
strip_size = sd->sd_meta->ssdi.ssd_strip_size;
strip_bits = sd->mds.mdd_raidp.srp_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)
/* create write workunit */
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->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;
if (sd->sd_type == SR_MD_RAID4)
parity = no_chunk; /* RAID4: Parity is always drive N */
else {
/* 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_raidp_addio(wu, chunk, lba, length,
data, xs->flags, 0, NULL))
goto bad;
break;
case BIOC_SDOFFLINE:
case BIOC_SDREBUILD:
case BIOC_SDHOTSPARE:
/*
* XXX only works if this LBA has already
* been scrubbed
*/
printf("Disk %llx offline, "
"regenerating buffer\n", chunk);
memset(data, 0, length);
for (i = 0; i <= no_chunk; i++) {
/*
* read all other drives: xor result
* into databuffer.
*/
if (i != chunk) {
if (sr_raidp_addio(wu, i, lba,
length, NULL, SCSI_DATA_IN,
SR_CCBF_FREEBUF, 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_get_block(sd, length);
if (xorbuf == NULL)
goto bad;
memcpy(xorbuf, data, length);
/* xor old data */
if (sr_raidp_addio(wu_r, chunk, lba, length, NULL,
SCSI_DATA_IN, SR_CCBF_FREEBUF, xorbuf))
goto bad;
/* xor old parity */
if (sr_raidp_addio(wu_r, parity, lba, length, NULL,
SCSI_DATA_IN, SR_CCBF_FREEBUF, xorbuf))
goto bad;
/* write new data */
if (sr_raidp_addio(wu, chunk, lba, length, data,
xs->flags, 0, NULL))
goto bad;
/* write new parity */
if (sr_raidp_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) {
/* 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;
}
/* rebuild io, let rebuild routine deal with it */
if (wu->swu_flags & SR_WUF_REBUILD)
goto queued;
/* current io failed, restart */
if (wu->swu_state == SR_WU_RESTART)
goto start;
/* deferred io failed, don't restart */
if (wu->swu_state == SR_WU_REQUEUE)
goto queued;
if (sr_check_io_collision(wu))
goto queued;
start:
sr_raid_startwu(wu);
queued:
splx(s);
return (0);
bad:
/* wu is unwound by sr_wu_put */
if (wu_r)
sr_scsi_wu_put(sd, wu_r);
return (1);
}
void
sr_raidp_intr(struct buf *bp)
{
struct sr_ccb *ccb = (struct sr_ccb *)bp;
struct sr_workunit *wu = ccb->ccb_wu, *wup;
struct sr_discipline *sd = wu->swu_dis;
struct scsi_xfer *xs = wu->swu_xs;
struct sr_softc *sc = sd->sd_sc;
int s;
DNPRINTF(SR_D_INTR, "%s: sr_raidp_intr bp %p xs %p\n",
DEVNAME(sc), bp, xs);
s = splbio();
sr_ccb_done(ccb);
/* XOR data to result. */
if (ccb->ccb_state == SR_CCB_OK && ccb->ccb_opaque)
sr_raidp_xor(ccb->ccb_opaque, ccb->ccb_buf.b_data,
ccb->ccb_buf.b_bcount);
/* Free allocated data buffer. */
if (ccb->ccb_flag & SR_CCBF_FREEBUF) {
sr_put_block(sd, ccb->ccb_buf.b_data, ccb->ccb_buf.b_bcount);
ccb->ccb_buf.b_data = NULL;
}
DNPRINTF(SR_D_INTR, "%s: sr_intr: comp: %d count: %d failed: %d\n",
DEVNAME(sc), wu->swu_ios_complete, wu->swu_io_count,
wu->swu_ios_failed);
if (wu->swu_ios_complete < wu->swu_io_count)
goto done;
if (xs != NULL)
xs->error = XS_NOERROR;
/* if all ios failed, retry reads and give up on writes */
if (wu->swu_ios_failed == wu->swu_ios_complete) {
/* XXX xs could be NULL here. */
if (xs->flags & SCSI_DATA_IN) {
printf("%s: retrying read on block %lld\n",
DEVNAME(sc), ccb->ccb_buf.b_blkno);
sr_wu_release_ccbs(wu);
wu->swu_state = SR_WU_RESTART;
if (sd->sd_scsi_rw(wu) == 0)
goto done;
xs->error = XS_DRIVER_STUFFUP;
} else {
printf("%s: permanently fail write on block %lld\n",
DEVNAME(sc), ccb->ccb_buf.b_blkno);
xs->error = XS_DRIVER_STUFFUP;
}
}
TAILQ_FOREACH(wup, &sd->sd_wu_pendq, swu_link)
if (wu == wup)
break;
if (wup == NULL)
panic("%s: wu %p not on pending queue",
DEVNAME(sd->sd_sc), wu);
TAILQ_REMOVE(&sd->sd_wu_pendq, wu, swu_link);
if (wu->swu_collider) {
if (wu->swu_ios_failed)
sr_raid_recreate_wu(wu->swu_collider);
/* XXX Should the collider be failed if this xs failed? */
/* restart deferred wu */
wu->swu_collider->swu_state = SR_WU_INPROGRESS;
TAILQ_REMOVE(&sd->sd_wu_defq, wu->swu_collider, swu_link);
sr_raid_startwu(wu->swu_collider);
}
if (wu->swu_flags & SR_WUF_REBUILD)
wu->swu_flags |= SR_WUF_REBUILDIOCOMP;
if (wu->swu_flags & SR_WUF_WAKEUP)
wakeup(wu);
if (!(wu->swu_flags & SR_WUF_REBUILD)) {
if (xs == NULL) {
sr_scsi_wu_put(sd, wu);
} else {
sr_scsi_done(sd, xs);
}
}
done:
splx(s);
}
int
sr_raidp_addio(struct sr_workunit *wu, int dsk, daddr64_t blk, daddr64_t len,
void *data, int flag, int ccbflag, void *xorbuf)
{
struct sr_discipline *sd = wu->swu_dis;
struct sr_ccb *ccb;
ccb = sr_ccb_get(sd);
if (!ccb)
return (-1);
/* allocate temporary buffer */
if (data == NULL) {
data = sr_get_block(sd, len);
if (data == NULL)
return (-1);
}
DNPRINTF(0, "%sio: %d.%llx %llx %s\n",
flag & SCSI_DATA_IN ? "read" : "write",
dsk, blk, len,
xorbuf ? "X0R" : "-");
ccb->ccb_flag = ccbflag;
if (flag & SCSI_POLL) {
ccb->ccb_buf.b_flags = 0;
ccb->ccb_buf.b_iodone = NULL;
} else {
ccb->ccb_buf.b_flags = B_CALL;
ccb->ccb_buf.b_iodone = sr_raidp_intr;
}
if (flag & SCSI_DATA_IN)
ccb->ccb_buf.b_flags |= B_READ;
else
ccb->ccb_buf.b_flags |= B_WRITE;
/* add offset for metadata */
ccb->ccb_buf.b_flags |= B_PHYS;
ccb->ccb_buf.b_blkno = blk;
ccb->ccb_buf.b_bcount = len;
ccb->ccb_buf.b_bufsize = len;
ccb->ccb_buf.b_resid = len;
ccb->ccb_buf.b_data = data;
ccb->ccb_buf.b_error = 0;
ccb->ccb_buf.b_proc = curproc;
ccb->ccb_buf.b_dev = sd->sd_vol.sv_chunks[dsk]->src_dev_mm;
ccb->ccb_buf.b_vp = sd->sd_vol.sv_chunks[dsk]->src_vn;
ccb->ccb_buf.b_bq = NULL;
if ((ccb->ccb_buf.b_flags & B_READ) == 0)
ccb->ccb_buf.b_vp->v_numoutput++;
ccb->ccb_wu = wu;
ccb->ccb_target = dsk;
ccb->ccb_opaque = xorbuf;
LIST_INIT(&ccb->ccb_buf.b_dep);
TAILQ_INSERT_TAIL(&wu->swu_ccb, ccb, ccb_link);
DNPRINTF(SR_D_DIS, "%s: %s: sr_raidp: b_bcount: %d "
"b_blkno: %x b_flags 0x%0x b_data %p\n",
DEVNAME(sd->sd_sc), sd->sd_meta->ssd_devname,
ccb->ccb_buf.b_bcount, ccb->ccb_buf.b_blkno,
ccb->ccb_buf.b_flags, ccb->ccb_buf.b_data);
wu->swu_io_count++;
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_raidp_xor(void *a, void *b, int len)
{
uint32_t *xa = a, *xb = b;
len >>= 2;
while (len--)
*xa++ ^= *xb++;
}
#if 0
void
sr_raidp_scrub(struct sr_discipline *sd)
{
daddr64_t strip_no, strip_size, no_chunk, parity, max_strip, strip_bits;
daddr64_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_raidp.srp_strip_bits;
max_strip = sd->sd_meta->ssdi.ssd_size >> strip_bits;
for (strip_no = 0; strip_no < max_strip; strip_no++) {
if (sd->sd_type == SR_MD_RAID4)
parity = no_chunk;
else
parity = no_chunk - ((strip_no / no_chunk) %
(no_chunk + 1));
xorbuf = sr_get_block(sd, strip_size);
for (i = 0; i <= no_chunk; i++) {
if (i != parity)
sr_raidp_addio(wu_r, i, 0xBADCAFE, strip_size,
NULL, SCSI_DATA_IN, SR_CCBF_FREEBUF,
xorbuf);
}
sr_raidp_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);
if (sr_check_io_collision(wu_r))
goto queued;
sr_raid_startwu(wu_r);
queued:
splx(s);
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
void *
sr_get_block(struct sr_discipline *sd, int length)
{
return dma_alloc(length, PR_NOWAIT | PR_ZERO);
}
void
sr_put_block(struct sr_discipline *sd, void *ptr, int length)
{
dma_free(ptr, length);
}
|