dm-raid.c 83.2 KB
Newer Older
NeilBrown's avatar
NeilBrown committed
1
2
/*
 * Copyright (C) 2010-2011 Neil Brown
3
 * Copyright (C) 2010-2016 Red Hat, Inc. All rights reserved.
NeilBrown's avatar
NeilBrown committed
4
5
6
7
8
 *
 * This file is released under the GPL.
 */

#include <linux/slab.h>
9
#include <linux/module.h>
NeilBrown's avatar
NeilBrown committed
10
11

#include "md.h"
12
#include "raid1.h"
NeilBrown's avatar
NeilBrown committed
13
#include "raid5.h"
14
#include "raid10.h"
NeilBrown's avatar
NeilBrown committed
15
16
#include "bitmap.h"

Alasdair G Kergon's avatar
Alasdair G Kergon committed
17
18
#include <linux/device-mapper.h>

NeilBrown's avatar
NeilBrown committed
19
#define DM_MSG_PREFIX "raid"
20
#define	MAX_RAID_DEVICES	253 /* md-raid kernel limit */
NeilBrown's avatar
NeilBrown committed
21

22
23
static bool devices_handle_discard_safely = false;

NeilBrown's avatar
NeilBrown committed
24
/*
25
26
 * The following flags are used by dm-raid.c to set up the array state.
 * They must be cleared before md_run is called.
NeilBrown's avatar
NeilBrown committed
27
 */
28
#define FirstUse 10		/* rdev flag */
NeilBrown's avatar
NeilBrown committed
29
30
31
32

struct raid_dev {
	/*
	 * Two DM devices, one to hold metadata and one to hold the
33
	 * actual data/parity.	The reason for this is to not confuse
NeilBrown's avatar
NeilBrown committed
34
35
36
37
38
39
40
41
42
43
44
	 * ti->len and give more flexibility in altering size and
	 * characteristics.
	 *
	 * While it is possible for this device to be associated
	 * with a different physical device than the data_dev, it
	 * is intended for it to be the same.
	 *    |--------- Physical Device ---------|
	 *    |- meta_dev -|------ data_dev ------|
	 */
	struct dm_dev *meta_dev;
	struct dm_dev *data_dev;
45
	struct md_rdev rdev;
NeilBrown's avatar
NeilBrown committed
46
47
48
};

/*
Heinz Mauelshagen's avatar
Heinz Mauelshagen committed
49
 * Flags for rs->ctr_flags field.
50
51
52
 *
 * 1 = no flag value
 * 2 = flag with value
NeilBrown's avatar
NeilBrown committed
53
 */
54
55
56
57
58
59
60
61
62
63
64
65
#define CTR_FLAG_SYNC		   0x1	 /* 1 */ /* Not with raid0! */
#define CTR_FLAG_NOSYNC		   0x2	 /* 1 */ /* Not with raid0! */
#define CTR_FLAG_REBUILD	   0x4	 /* 2 */ /* Not with raid0! */
#define CTR_FLAG_DAEMON_SLEEP	   0x8	 /* 2 */ /* Not with raid0! */
#define CTR_FLAG_MIN_RECOVERY_RATE 0x10	 /* 2 */ /* Not with raid0! */
#define CTR_FLAG_MAX_RECOVERY_RATE 0x20	 /* 2 */ /* Not with raid0! */
#define CTR_FLAG_MAX_WRITE_BEHIND  0x40	 /* 2 */ /* Only with raid1! */
#define CTR_FLAG_WRITE_MOSTLY	   0x80	 /* 2 */ /* Only with raid1! */
#define CTR_FLAG_STRIPE_CACHE	   0x100 /* 2 */ /* Only with raid4/5/6! */
#define CTR_FLAG_REGION_SIZE	   0x200 /* 2 */ /* Not with raid0! */
#define CTR_FLAG_RAID10_COPIES	   0x400 /* 2 */ /* Only with raid10 */
#define CTR_FLAG_RAID10_FORMAT	   0x800 /* 2 */ /* Only with raid10 */
66
/* New for v1.8.0 */
67
68
#define CTR_FLAG_DELTA_DISKS	      0x1000 /* 2 */ /* Only with reshapable raid4/5/6/10! */
#define CTR_FLAG_DATA_OFFSET	      0x2000 /* 2 */ /* Only with reshapable raid4/5/6/10! */
69
#define CTR_FLAG_RAID10_USE_NEAR_SETS 0x4000 /* 2 */ /* Only with raid10! */
70

71
72
73
74
75
76
77
78
79
/*
 * Definitions of various constructor flags to
 * be used in checks of valid / invalid flags
 * per raid level.
 */
/* Define all any sync flags */
#define	CTR_FLAGS_ANY_SYNC		(CTR_FLAG_SYNC | CTR_FLAG_NOSYNC)

/* Define flags for options without argument (e.g. 'nosync') */
80
81
#define	CTR_FLAG_OPTIONS_NO_ARGS	(CTR_FLAGS_ANY_SYNC | \
					 CTR_FLAG_RAID10_USE_NEAR_SETS)
82
83
84
85
86
87
88
89
90
91
92

/* Define flags for options with one argument (e.g. 'delta_disks +2') */
#define CTR_FLAG_OPTIONS_ONE_ARG (CTR_FLAG_REBUILD | \
				  CTR_FLAG_WRITE_MOSTLY | \
				  CTR_FLAG_DAEMON_SLEEP | \
				  CTR_FLAG_MIN_RECOVERY_RATE | \
				  CTR_FLAG_MAX_RECOVERY_RATE | \
				  CTR_FLAG_MAX_WRITE_BEHIND | \
				  CTR_FLAG_STRIPE_CACHE | \
				  CTR_FLAG_REGION_SIZE | \
				  CTR_FLAG_RAID10_COPIES | \
93
94
95
				  CTR_FLAG_RAID10_FORMAT | \
				  CTR_FLAG_DELTA_DISKS | \
				  CTR_FLAG_DATA_OFFSET)
96
97
98
99
100
101
102
103
104
105
106
107
108

/* All ctr optional arguments */
#define ALL_CTR_FLAGS		(CTR_FLAG_OPTIONS_NO_ARGS | \
				 CTR_FLAG_OPTIONS_ONE_ARG)

/* Invalid options definitions per raid level... */

/* "raid0" does not accept any options */
#define RAID0_INVALID_FLAGS ALL_CTR_FLAGS

/* "raid1" does not accept stripe cache or any raid10 options */
#define RAID1_INVALID_FLAGS	(CTR_FLAG_STRIPE_CACHE | \
				 CTR_FLAG_RAID10_COPIES | \
109
110
111
				 CTR_FLAG_RAID10_FORMAT | \
				 CTR_FLAG_DELTA_DISKS | \
				 CTR_FLAG_DATA_OFFSET)
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126

/* "raid10" does not accept any raid1 or stripe cache options */
#define RAID10_INVALID_FLAGS	(CTR_FLAG_WRITE_MOSTLY | \
				 CTR_FLAG_MAX_WRITE_BEHIND | \
				 CTR_FLAG_STRIPE_CACHE)
/*
 * "raid4/5/6" do not accept any raid1 or raid10 specific options
 *
 * "raid6" does not accept "nosync", because it is not guaranteed
 * that both parity and q-syndrome are being written properly with
 * any writes
 */
#define RAID45_INVALID_FLAGS	(CTR_FLAG_WRITE_MOSTLY | \
				 CTR_FLAG_MAX_WRITE_BEHIND | \
				 CTR_FLAG_RAID10_FORMAT | \
127
128
				 CTR_FLAG_RAID10_COPIES | \
				 CTR_FLAG_RAID10_USE_NEAR_SETS)
129
130
131
#define RAID6_INVALID_FLAGS	(CTR_FLAG_NOSYNC | RAID45_INVALID_FLAGS)
/* ...invalid options definitions per raid level */

132
133
134
135
136
137
138
139
140
141
142
143
144
/*
 * Flags for rs->runtime_flags field
 * (RT_FLAG prefix meaning "runtime flag")
 *
 * These are all internal and used to define runtime state,
 * e.g. to prevent another resume from preresume processing
 * the raid set all over again.
 */
#define RT_FLAG_RS_PRERESUMED		0x1
#define RT_FLAG_RS_RESUMED		0x2
#define RT_FLAG_RS_BITMAP_LOADED	0x4
#define RT_FLAG_UPDATE_SBS		0x8

145
146
147
/* Array elements of 64 bit needed for rebuild/write_mostly bits */
#define DISKS_ARRAY_ELEMS ((MAX_RAID_DEVICES + (sizeof(uint64_t) * 8 - 1)) / sizeof(uint64_t) / 8)

148
149
150
151
152
153
154
155
156
/*
 * raid set level, layout and chunk sectors backup/restore
 */
struct rs_layout {
	int new_level;
	int new_layout;
	int new_chunk_sectors;
};

NeilBrown's avatar
NeilBrown committed
157
158
159
struct raid_set {
	struct dm_target *ti;

160
	uint32_t bitmap_loaded;
Heinz Mauelshagen's avatar
Heinz Mauelshagen committed
161
	uint32_t ctr_flags;
162
163
164
	uint32_t runtime_flags;

	uint64_t rebuild_disks[DISKS_ARRAY_ELEMS];
NeilBrown's avatar
NeilBrown committed
165

166
167
	int raid_disks;
	int delta_disks;
168
	int data_offset;
169
170
	int raid10_copies;

171
	struct mddev md;
NeilBrown's avatar
NeilBrown committed
172
173
	struct raid_type *raid_type;
	struct dm_target_callbacks callbacks;
174
	struct rs_layout rs_layout;
NeilBrown's avatar
NeilBrown committed
175
176
177
178

	struct raid_dev dev[0];
};

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
/* Backup/restore raid set configuration helpers */
static void _rs_config_backup(struct raid_set *rs, struct rs_layout *l)
{
	struct mddev *mddev = &rs->md;

	l->new_level = mddev->new_level;
	l->new_layout = mddev->new_layout;
	l->new_chunk_sectors = mddev->new_chunk_sectors;
}

static void rs_config_backup(struct raid_set *rs)
{
	return _rs_config_backup(rs, &rs->rs_layout);
}

static void _rs_config_restore(struct raid_set *rs, struct rs_layout *l)
{
	struct mddev *mddev = &rs->md;

	mddev->new_level = l->new_level;
	mddev->new_layout = l->new_layout;
	mddev->new_chunk_sectors = l->new_chunk_sectors;
}

static void rs_config_restore(struct raid_set *rs)
{
	return _rs_config_restore(rs, &rs->rs_layout);
}
/* END: backup/restore raid set configuration helpers */

209
210
211
212
213
214
/* raid10 algorithms (i.e. formats) */
#define	ALGORITHM_RAID10_DEFAULT	0
#define	ALGORITHM_RAID10_NEAR		1
#define	ALGORITHM_RAID10_OFFSET		2
#define	ALGORITHM_RAID10_FAR		3

NeilBrown's avatar
NeilBrown committed
215
216
217
218
219
220
221
222
223
/* Supported raid types and properties. */
static struct raid_type {
	const char *name;		/* RAID algorithm. */
	const char *descr;		/* Descriptor text for logging. */
	const unsigned parity_devs;	/* # of parity devices. */
	const unsigned minimal_devs;	/* minimal # of devices in set. */
	const unsigned level;		/* RAID level. */
	const unsigned algorithm;	/* RAID algorithm. */
} raid_types[] = {
224
225
226
	{"raid0",	  "raid0 (striping)",			    0, 2, 0,  0 /* NONE */},
	{"raid1",	  "raid1 (mirroring)",			    0, 2, 1,  0 /* NONE */},
	{"raid10_far",	  "raid10 far (striped mirrors)",	    0, 2, 10, ALGORITHM_RAID10_FAR},
227
	{"raid10_offset", "raid10 offset (striped mirrors)",	    0, 2, 10, ALGORITHM_RAID10_OFFSET},
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
	{"raid10_near",	  "raid10 near (striped mirrors)",	    0, 2, 10, ALGORITHM_RAID10_NEAR},
	{"raid10",	  "raid10 (striped mirrors)",		    0, 2, 10, ALGORITHM_RAID10_DEFAULT},
	{"raid4",	  "raid4 (dedicated last parity disk)",	    1, 2, 4,  ALGORITHM_PARITY_N}, /* raid4 layout = raid5_n */
	{"raid5_n",	  "raid5 (dedicated last parity disk)",	    1, 2, 5,  ALGORITHM_PARITY_N},
	{"raid5_ls",	  "raid5 (left symmetric)",		    1, 2, 5,  ALGORITHM_LEFT_SYMMETRIC},
	{"raid5_rs",	  "raid5 (right symmetric)",		    1, 2, 5,  ALGORITHM_RIGHT_SYMMETRIC},
	{"raid5_la",	  "raid5 (left asymmetric)",		    1, 2, 5,  ALGORITHM_LEFT_ASYMMETRIC},
	{"raid5_ra",	  "raid5 (right asymmetric)",		    1, 2, 5,  ALGORITHM_RIGHT_ASYMMETRIC},
	{"raid6_zr",	  "raid6 (zero restart)",		    2, 4, 6,  ALGORITHM_ROTATING_ZERO_RESTART},
	{"raid6_nr",	  "raid6 (N restart)",			    2, 4, 6,  ALGORITHM_ROTATING_N_RESTART},
	{"raid6_nc",	  "raid6 (N continue)",			    2, 4, 6,  ALGORITHM_ROTATING_N_CONTINUE},
	{"raid6_n_6",	  "raid6 (dedicated parity/Q n/6)",	    2, 4, 6,  ALGORITHM_PARITY_N_6},
	{"raid6_ls_6",	  "raid6 (left symmetric dedicated Q 6)",   2, 4, 6,  ALGORITHM_LEFT_SYMMETRIC_6},
	{"raid6_rs_6",	  "raid6 (right symmetric dedicated Q 6)",  2, 4, 6,  ALGORITHM_RIGHT_SYMMETRIC_6},
	{"raid6_la_6",	  "raid6 (left asymmetric dedicated Q 6)",  2, 4, 6,  ALGORITHM_LEFT_ASYMMETRIC_6},
	{"raid6_ra_6",	  "raid6 (right asymmetric dedicated Q 6)", 2, 4, 6,  ALGORITHM_RIGHT_ASYMMETRIC_6}
NeilBrown's avatar
NeilBrown committed
244
245
};

246
247
248
249
250
251
/* True, if @v is in inclusive range [@min, @max] */
static bool _in_range(long v, long min, long max)
{
	return v >= min && v <= max;
}

252
253
254
255
256
257
258
259
/* ctr flag bit manipulation... */
/* Set single @flag in @flags */
static void _set_flag(uint32_t flag, uint32_t *flags)
{
	WARN_ON_ONCE(hweight32(flag) != 1);
	*flags |= flag;
}

260
261
262
263
264
265
266
/* Clear single @flag in @flags */
static void _clear_flag(uint32_t flag, uint32_t *flags)
{
	WARN_ON_ONCE(hweight32(flag) != 1);
	*flags &= ~flag;
}

267
268
269
270
271
272
273
/* Test single @flag in @flags */
static bool _test_flag(uint32_t flag, uint32_t flags)
{
	WARN_ON_ONCE(hweight32(flag) != 1);
	return (flag & flags) ? true : false;
}

274
275
276
277
278
279
/* Test multiple @flags in @all_flags */
static bool _test_flags(uint32_t flags, uint32_t all_flags)
{
	return (flags & all_flags) ? true : false;
}

280
281
282
283
284
285
/* Clear (multiple) @flags in @all_flags */
static void _clear_flags(uint32_t flags, uint32_t *all_flags)
{
	*all_flags &= ~flags;
}

286
287
288
289
290
291
292
293
294
/* Return true if single @flag is set in @*flags, else set it and return false */
static bool _test_and_set_flag(uint32_t flag, uint32_t *flags)
{
	if (_test_flag(flag, *flags))
		return true;

	_set_flag(flag, flags);
	return false;
}
295
296
297
298
299
300
301
302
303
304
305

/* Return true if single @flag is set in @*flags and clear it, else return false */
static bool _test_and_clear_flag(uint32_t flag, uint32_t *flags)
{
	if (_test_flag(flag, *flags)) {
		_clear_flag(flag, flags);
		return true;
	}

	return false;
}
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
/* ...ctr and runtime flag bit manipulation */

/* All table line arguments are defined here */
static struct arg_name_flag {
	const uint32_t flag;
	const char *name;
} _arg_name_flags[] = {
	{ CTR_FLAG_SYNC, "sync"},
	{ CTR_FLAG_NOSYNC, "nosync"},
	{ CTR_FLAG_REBUILD, "rebuild"},
	{ CTR_FLAG_DAEMON_SLEEP, "daemon_sleep"},
	{ CTR_FLAG_MIN_RECOVERY_RATE, "min_recovery_rate"},
	{ CTR_FLAG_MAX_RECOVERY_RATE, "max_recovery_rate"},
	{ CTR_FLAG_MAX_WRITE_BEHIND, "max_write_behind"},
	{ CTR_FLAG_WRITE_MOSTLY, "writemostly"},
	{ CTR_FLAG_STRIPE_CACHE, "stripe_cache"},
	{ CTR_FLAG_REGION_SIZE, "region_size"},
	{ CTR_FLAG_RAID10_COPIES, "raid10_copies"},
	{ CTR_FLAG_RAID10_FORMAT, "raid10_format"},
325
326
327
	{ CTR_FLAG_DATA_OFFSET, "data_offset"},
	{ CTR_FLAG_DELTA_DISKS, "delta_disks"},
	{ CTR_FLAG_RAID10_USE_NEAR_SETS, "raid10_use_near_sets"},
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
};

/* Return argument name string for given @flag */
static const char *_argname_by_flag(const uint32_t flag)
{
	if (hweight32(flag) == 1) {
		struct arg_name_flag *anf = _arg_name_flags + ARRAY_SIZE(_arg_name_flags);

		while (anf-- > _arg_name_flags)
			if (_test_flag(flag, anf->flag))
				return anf->name;

	} else
		DMERR("%s called with more than one flag!", __func__);

	return NULL;
}

346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
/*
 * bool helpers to test for various raid levels of a raid set,
 * is. it's level as reported by the superblock rather than
 * the requested raid_type passed to the constructor.
 */
/* Return true, if raid set in @rs is raid0 */
static bool rs_is_raid0(struct raid_set *rs)
{
	return !rs->md.level;
}

/* Return true, if raid set in @rs is raid10 */
static bool rs_is_raid10(struct raid_set *rs)
{
	return rs->md.level == 10;
}

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
/*
 * bool helpers to test for various raid levels of a raid type
 */

/* Return true, if raid type in @rt is raid0 */
static bool rt_is_raid0(struct raid_type *rt)
{
	return !rt->level;
}

/* Return true, if raid type in @rt is raid1 */
static bool rt_is_raid1(struct raid_type *rt)
{
	return rt->level == 1;
}

/* Return true, if raid type in @rt is raid10 */
static bool rt_is_raid10(struct raid_type *rt)
{
	return rt->level == 10;
}

/* Return true, if raid type in @rt is raid4/5 */
static bool rt_is_raid45(struct raid_type *rt)
{
	return _in_range(rt->level, 4, 5);
}

/* Return true, if raid type in @rt is raid6 */
static bool rt_is_raid6(struct raid_type *rt)
{
	return rt->level == 6;
}
396
397
398
399
400
401

/* Return true, if raid type in @rt is raid4/5/6 */
static bool rt_is_raid456(struct raid_type *rt)
{
	return _in_range(rt->level, 4, 6);
}
402
403
/* END: raid level bools */

404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
/*
 * Convenience functions to set ti->error to @errmsg and
 * return @r in order to shorten code in a lot of places
 */
static int ti_error_ret(struct dm_target *ti, const char *errmsg, int r)
{
	ti->error = (char *) errmsg;
	return r;
}

static int ti_error_einval(struct dm_target *ti, const char *errmsg)
{
	return ti_error_ret(ti, errmsg, -EINVAL);
}
/* END: convenience functions to set ti->error to @errmsg... */

420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
/* Return invalid ctr flags for the raid level of @rs */
static uint32_t _invalid_flags(struct raid_set *rs)
{
	if (rt_is_raid0(rs->raid_type))
		return RAID0_INVALID_FLAGS;
	else if (rt_is_raid1(rs->raid_type))
		return RAID1_INVALID_FLAGS;
	else if (rt_is_raid10(rs->raid_type))
		return RAID10_INVALID_FLAGS;
	else if (rt_is_raid45(rs->raid_type))
		return RAID45_INVALID_FLAGS;
	else if (rt_is_raid6(rs->raid_type))
		return RAID6_INVALID_FLAGS;

	return ~0;
}

/*
 * Check for any invalid flags set on @rs defined by bitset @invalid_flags
 *
 * Has to be called after parsing of the ctr flags!
 */
static int rs_check_for_invalid_flags(struct raid_set *rs)
{
444
445
	if (_test_flags(rs->ctr_flags, _invalid_flags(rs)))
		return ti_error_einval(rs->ti, "Invalid flag combined");
446
447
448
449

	return 0;
}

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

/* MD raid10 bit definitions and helpers */
#define RAID10_OFFSET			(1 << 16) /* stripes with data copies area adjacent on devices */
#define RAID10_BROCKEN_USE_FAR_SETS	(1 << 17) /* Broken in raid10.c: use sets instead of whole stripe rotation */
#define RAID10_USE_FAR_SETS		(1 << 18) /* Use sets instead of whole stripe rotation */
#define RAID10_FAR_COPIES_SHIFT		8	  /* raid10 # far copies shift (2nd byte of layout) */

/* Return md raid10 near copies for @layout */
static unsigned int _raid10_near_copies(int layout)
{
	return layout & 0xFF;
}

/* Return md raid10 far copies for @layout */
static unsigned int _raid10_far_copies(int layout)
{
	return _raid10_near_copies(layout >> RAID10_FAR_COPIES_SHIFT);
}

/* Return true if md raid10 offset for @layout */
static unsigned int _is_raid10_offset(int layout)
{
	return layout & RAID10_OFFSET;
}

/* Return true if md raid10 near for @layout */
static unsigned int _is_raid10_near(int layout)
{
	return !_is_raid10_offset(layout) && _raid10_near_copies(layout) > 1;
}

/* Return true if md raid10 far for @layout */
static unsigned int _is_raid10_far(int layout)
{
	return !_is_raid10_offset(layout) && _raid10_far_copies(layout) > 1;
}

/* Return md raid10 layout string for @layout */
static const char *raid10_md_layout_to_format(int layout)
489
490
{
	/*
491
492
493
	 * Bit 16 stands for "offset"
	 * (i.e. adjacent stripes hold copies)
	 *
494
495
	 * Refer to MD's raid10.c for details
	 */
496
	if (_is_raid10_offset(layout))
497
498
		return "offset";

499
	if (_raid10_near_copies(layout) > 1)
500
501
		return "near";

502
503
	WARN_ON(_raid10_far_copies(layout) < 2);

504
505
506
	return "far";
}

507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
/* Return md raid10 algorithm for @name */
static const int raid10_name_to_format(const char *name)
{
	if (!strcasecmp(name, "near"))
		return ALGORITHM_RAID10_NEAR;
	else if (!strcasecmp(name, "offset"))
		return ALGORITHM_RAID10_OFFSET;
	else if (!strcasecmp(name, "far"))
		return ALGORITHM_RAID10_FAR;

	return -EINVAL;
}


/* Return md raid10 copies for @layout */
static unsigned int raid10_md_layout_to_copies(int layout)
523
{
524
525
	return _raid10_near_copies(layout) > 1 ?
	       _raid10_near_copies(layout) : _raid10_far_copies(layout);
526
527
}

528
529
530
531
/* Return md raid10 format id for @format string */
static int raid10_format_to_md_layout(struct raid_set *rs,
				      unsigned int algorithm,
				      unsigned int copies)
532
{
533
	unsigned int n = 1, f = 1, r = 0;
534

535
536
537
538
539
540
541
542
543
544
	/*
	 * MD resilienece flaw:
	 *
	 * enabling use_far_sets for far/offset formats causes copies
	 * to be colocated on the same devs together with their origins!
	 *
	 * -> disable it for now in the definition above
	 */
	if (algorithm == ALGORITHM_RAID10_DEFAULT ||
	    algorithm == ALGORITHM_RAID10_NEAR)
545
		n = copies;
546
547
548
549
550
551
552
553

	else if (algorithm == ALGORITHM_RAID10_OFFSET) {
		f = copies;
		r = RAID10_OFFSET;
		if (!_test_flag(CTR_FLAG_RAID10_USE_NEAR_SETS, rs->ctr_flags))
			r |= RAID10_USE_FAR_SETS;

	} else if (algorithm == ALGORITHM_RAID10_FAR) {
554
		f = copies;
555
556
557
		r = !RAID10_OFFSET;
		if (!_test_flag(CTR_FLAG_RAID10_USE_NEAR_SETS, rs->ctr_flags))
			r |= RAID10_USE_FAR_SETS;
558

559
560
561
562
563
564
	} else
		return -EINVAL;

	return r | (f << RAID10_FAR_COPIES_SHIFT) | n;
}
/* END: MD raid10 bit definitions and helpers */
565

566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
/* Check for any of the raid10 algorithms */
static int _got_raid10(struct raid_type *rtp, const int layout)
{
	if (rtp->level == 10) {
		switch (rtp->algorithm) {
		case ALGORITHM_RAID10_DEFAULT:
		case ALGORITHM_RAID10_NEAR:
			return _is_raid10_near(layout);
		case ALGORITHM_RAID10_OFFSET:
			return _is_raid10_offset(layout);
		case ALGORITHM_RAID10_FAR:
			return _is_raid10_far(layout);
		default:
			break;
		}
	}
582

583
	return 0;
584
585
}

586
/* Return raid_type for @name */
587
static struct raid_type *get_raid_type(const char *name)
NeilBrown's avatar
NeilBrown committed
588
{
589
	struct raid_type *rtp = raid_types + ARRAY_SIZE(raid_types);
NeilBrown's avatar
NeilBrown committed
590

591
592
593
	while (rtp-- > raid_types)
		if (!strcasecmp(rtp->name, name))
			return rtp;
NeilBrown's avatar
NeilBrown committed
594
595
596
597

	return NULL;
}

598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
/* Return raid_type for @name based derived from @level and @layout */
static struct raid_type *get_raid_type_by_ll(const int level, const int layout)
{
	struct raid_type *rtp = raid_types + ARRAY_SIZE(raid_types);

	while (rtp-- > raid_types) {
		/* RAID10 special checks based on @layout flags/properties */
		if (rtp->level == level &&
		    (_got_raid10(rtp, layout) || rtp->algorithm == layout))
			return rtp;
	}

	return NULL;
}

613
614
615
616
617
618
619
620
621
622
623
624
625
/*
 * Set the mddev properties in @rs to the current
 * ones retrieved from the freshest superblock
 */
static void rs_set_cur(struct raid_set *rs)
{
	struct mddev *mddev = &rs->md;

	mddev->new_level = mddev->level;
	mddev->new_layout = mddev->layout;
	mddev->new_chunk_sectors = mddev->chunk_sectors;
}

626
627
628
629
630
631
632
633
634
635
636
/*
 * Set the mddev properties in @rs to the new
 * ones requested by the ctr
 */
static void rs_set_new(struct raid_set *rs)
{
	struct mddev *mddev = &rs->md;

	mddev->level = mddev->new_level;
	mddev->layout = mddev->new_layout;
	mddev->chunk_sectors = mddev->new_chunk_sectors;
637
	mddev->raid_disks = rs->raid_disks;
638
639
640
641
	mddev->delta_disks = 0;
}


NeilBrown's avatar
NeilBrown committed
642
643
644
645
646
static struct raid_set *context_alloc(struct dm_target *ti, struct raid_type *raid_type, unsigned raid_devs)
{
	unsigned i;
	struct raid_set *rs;

647
648
	if (raid_devs <= raid_type->parity_devs)
		return ERR_PTR(ti_error_einval(ti, "Insufficient number of devices"));
NeilBrown's avatar
NeilBrown committed
649
650

	rs = kzalloc(sizeof(*rs) + raid_devs * sizeof(rs->dev[0]), GFP_KERNEL);
651
652
	if (!rs)
		return ERR_PTR(ti_error_ret(ti, "Cannot allocate raid context", -ENOMEM));
NeilBrown's avatar
NeilBrown committed
653
654
655

	mddev_init(&rs->md);

656
657
658
	rs->raid_disks = raid_devs;
	rs->delta_disks = 0;

NeilBrown's avatar
NeilBrown committed
659
660
661
662
663
664
665
666
	rs->ti = ti;
	rs->raid_type = raid_type;
	rs->md.raid_disks = raid_devs;
	rs->md.level = raid_type->level;
	rs->md.new_level = rs->md.level;
	rs->md.layout = raid_type->algorithm;
	rs->md.new_layout = rs->md.layout;
	rs->md.delta_disks = 0;
667
	rs->md.recovery_cp = rs_is_raid0(rs) ? MaxSector : 0;
NeilBrown's avatar
NeilBrown committed
668
669
670
671
672
673
674
675
676
677

	for (i = 0; i < raid_devs; i++)
		md_rdev_init(&rs->dev[i].rdev);

	/*
	 * Remaining items to be initialized by further RAID params:
	 *  rs->md.persistent
	 *  rs->md.external
	 *  rs->md.chunk_sectors
	 *  rs->md.new_chunk_sectors
678
	 *  rs->md.dev_sectors
NeilBrown's avatar
NeilBrown committed
679
680
681
682
683
684
685
686
687
	 */

	return rs;
}

static void context_free(struct raid_set *rs)
{
	int i;

688
689
690
	for (i = 0; i < rs->md.raid_disks; i++) {
		if (rs->dev[i].meta_dev)
			dm_put_device(rs->ti, rs->dev[i].meta_dev);
691
		md_rdev_clear(&rs->dev[i].rdev);
NeilBrown's avatar
NeilBrown committed
692
693
		if (rs->dev[i].data_dev)
			dm_put_device(rs->ti, rs->dev[i].data_dev);
694
	}
NeilBrown's avatar
NeilBrown committed
695
696
697
698
699
700
701
702
703

	kfree(rs);
}

/*
 * For every device we have two words
 *  <meta_dev>: meta device name or '-' if missing
 *  <data_dev>: data device name or '-' if missing
 *
704
705
706
707
708
709
710
711
712
713
 * The following are permitted:
 *    - -
 *    - <data_dev>
 *    <meta_dev> <data_dev>
 *
 * The following is not allowed:
 *    <meta_dev> -
 *
 * This code parses those words.  If there is a failure,
 * the caller must use context_free to unwind the operations.
NeilBrown's avatar
NeilBrown committed
714
 */
715
static int parse_dev_params(struct raid_set *rs, struct dm_arg_set *as)
NeilBrown's avatar
NeilBrown committed
716
717
718
719
{
	int i;
	int rebuild = 0;
	int metadata_available = 0;
720
	int r = 0;
721
	const char *arg;
NeilBrown's avatar
NeilBrown committed
722

723
724
725
726
727
728
	/* Put off the number of raid devices argument to get to dev pairs */
	arg = dm_shift_arg(as);
	if (!arg)
		return -EINVAL;

	for (i = 0; i < rs->md.raid_disks; i++) {
NeilBrown's avatar
NeilBrown committed
729
730
731
732
733
734
735
736
737
738
739
740
		rs->dev[i].rdev.raid_disk = i;

		rs->dev[i].meta_dev = NULL;
		rs->dev[i].data_dev = NULL;

		/*
		 * There are no offsets, since there is a separate device
		 * for data and metadata.
		 */
		rs->dev[i].rdev.data_offset = 0;
		rs->dev[i].rdev.mddev = &rs->md;

741
742
743
744
745
746
		arg = dm_shift_arg(as);
		if (!arg)
			return -EINVAL;

		if (strcmp(arg, "-")) {
			r = dm_get_device(rs->ti, arg,
747
748
					    dm_table_get_mode(rs->ti->table),
					    &rs->dev[i].meta_dev);
749
			if (r)
750
				return ti_error_ret(rs->ti, "RAID metadata device lookup failure", r);
751
752
753

			rs->dev[i].rdev.sb_page = alloc_page(GFP_KERNEL);
			if (!rs->dev[i].rdev.sb_page)
754
				return ti_error_ret(rs->ti, "Failed to allocate superblock page", -ENOMEM);
NeilBrown's avatar
NeilBrown committed
755
756
		}

757
758
759
760
761
		arg = dm_shift_arg(as);
		if (!arg)
			return -EINVAL;

		if (!strcmp(arg, "-")) {
NeilBrown's avatar
NeilBrown committed
762
			if (!test_bit(In_sync, &rs->dev[i].rdev.flags) &&
763
764
			    (!rs->dev[i].rdev.recovery_offset))
				return ti_error_einval(rs->ti, "Drive designated for rebuild not specified");
NeilBrown's avatar
NeilBrown committed
765

766
			if (rs->dev[i].meta_dev)
767
				return ti_error_einval(rs->ti, "No data device supplied with metadata device");
768

NeilBrown's avatar
NeilBrown committed
769
770
771
			continue;
		}

772
		r = dm_get_device(rs->ti, arg,
NeilBrown's avatar
NeilBrown committed
773
774
				    dm_table_get_mode(rs->ti->table),
				    &rs->dev[i].data_dev);
775
776
		if (r)
			return ti_error_ret(rs->ti, "RAID device lookup failure", r);
NeilBrown's avatar
NeilBrown committed
777

778
779
780
781
		if (rs->dev[i].meta_dev) {
			metadata_available = 1;
			rs->dev[i].rdev.meta_bdev = rs->dev[i].meta_dev->bdev;
		}
NeilBrown's avatar
NeilBrown committed
782
		rs->dev[i].rdev.bdev = rs->dev[i].data_dev->bdev;
783
		list_add_tail(&rs->dev[i].rdev.same_set, &rs->md.disks);
NeilBrown's avatar
NeilBrown committed
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
		if (!test_bit(In_sync, &rs->dev[i].rdev.flags))
			rebuild++;
	}

	if (metadata_available) {
		rs->md.external = 0;
		rs->md.persistent = 1;
		rs->md.major_version = 2;
	} else if (rebuild && !rs->md.recovery_cp) {
		/*
		 * Without metadata, we will not be able to tell if the array
		 * is in-sync or not - we must assume it is not.  Therefore,
		 * it is impossible to rebuild a drive.
		 *
		 * Even if there is metadata, the on-disk information may
		 * indicate that the array is not in-sync and it will then
		 * fail at that time.
		 *
		 * User could specify 'nosync' option if desperate.
		 */
		DMERR("Unable to rebuild drive while array is not in-sync");
805
		return ti_error_einval(rs->ti, "Unable to rebuild drive while array is not in-sync");
NeilBrown's avatar
NeilBrown committed
806
807
808
809
810
	}

	return 0;
}

811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
/*
 * validate_region_size
 * @rs
 * @region_size:  region size in sectors.  If 0, pick a size (4MiB default).
 *
 * Set rs->md.bitmap_info.chunksize (which really refers to 'region size').
 * Ensure that (ti->len/region_size < 2^21) - required by MD bitmap.
 *
 * Returns: 0 on success, -EINVAL on failure.
 */
static int validate_region_size(struct raid_set *rs, unsigned long region_size)
{
	unsigned long min_region_size = rs->ti->len / (1 << 21);

	if (!region_size) {
		/*
827
		 * Choose a reasonable default.	 All figures in sectors.
828
829
		 */
		if (min_region_size > (1 << 13)) {
830
			/* If not a power of 2, make it the next power of 2 */
831
			region_size = roundup_pow_of_two(min_region_size);
832
833
834
835
836
837
838
839
840
841
			DMINFO("Choosing default region size of %lu sectors",
			       region_size);
		} else {
			DMINFO("Choosing default region size of 4MiB");
			region_size = 1 << 13; /* sectors */
		}
	} else {
		/*
		 * Validate user-supplied value.
		 */
842
843
		if (region_size > rs->ti->len)
			return ti_error_einval(rs->ti, "Supplied region size is too large");
844
845
846
847

		if (region_size < min_region_size) {
			DMERR("Supplied region_size (%lu sectors) below minimum (%lu)",
			      region_size, min_region_size);
848
			return ti_error_einval(rs->ti, "Supplied region size is too small");
849
850
		}

851
852
		if (!is_power_of_2(region_size))
			return ti_error_einval(rs->ti, "Region size is not a power of 2");
853

854
855
		if (region_size < rs->md.chunk_sectors)
			return ti_error_einval(rs->ti, "Region size is smaller than the chunk size");
856
857
858
859
860
861
862
863
864
865
	}

	/*
	 * Convert sectors to bytes.
	 */
	rs->md.bitmap_info.chunksize = (region_size << 9);

	return 0;
}

866
/*
867
 * validate_raid_redundancy
868
869
 * @rs
 *
870
871
 * Determine if there are enough devices in the array that haven't
 * failed (or are being rebuilt) to form a usable array.
872
873
874
 *
 * Returns: 0 on success, -EINVAL on failure.
 */
875
static int validate_raid_redundancy(struct raid_set *rs)
876
877
{
	unsigned i, rebuild_cnt = 0;
878
	unsigned rebuilds_per_group = 0, copies, d;
879
	unsigned group_size, last_group_start;
880
881

	for (i = 0; i < rs->md.raid_disks; i++)
882
883
		if (!test_bit(In_sync, &rs->dev[i].rdev.flags) ||
		    !rs->dev[i].rdev.sb_page)
884
885
886
887
888
889
890
891
892
893
894
895
896
897
			rebuild_cnt++;

	switch (rs->raid_type->level) {
	case 1:
		if (rebuild_cnt >= rs->md.raid_disks)
			goto too_many;
		break;
	case 4:
	case 5:
	case 6:
		if (rebuild_cnt > rs->raid_type->parity_devs)
			goto too_many;
		break;
	case 10:
898
899
900
901
902
903
904
905
906
907
908
909
910
911
		copies = raid10_md_layout_to_copies(rs->md.layout);
		if (rebuild_cnt < copies)
			break;

		/*
		 * It is possible to have a higher rebuild count for RAID10,
		 * as long as the failed devices occur in different mirror
		 * groups (i.e. different stripes).
		 *
		 * When checking "near" format, make sure no adjacent devices
		 * have failed beyond what can be handled.  In addition to the
		 * simple case where the number of devices is a multiple of the
		 * number of copies, we must also handle cases where the number
		 * of devices is not a multiple of the number of copies.
912
913
914
		 * E.g.	   dev1 dev2 dev3 dev4 dev5
		 *	    A	 A    B	   B	C
		 *	    C	 D    D	   E	E
915
		 */
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
		if (!strcmp("near", raid10_md_layout_to_format(rs->md.layout))) {
			for (i = 0; i < rs->md.raid_disks * copies; i++) {
				if (!(i % copies))
					rebuilds_per_group = 0;
				d = i % rs->md.raid_disks;
				if ((!rs->dev[d].rdev.sb_page ||
				     !test_bit(In_sync, &rs->dev[d].rdev.flags)) &&
				    (++rebuilds_per_group >= copies))
					goto too_many;
			}
			break;
		}

		/*
		 * When checking "far" and "offset" formats, we need to ensure
		 * that the device that holds its copy is not also dead or
		 * being rebuilt.  (Note that "far" and "offset" formats only
		 * support two copies right now.  These formats also only ever
		 * use the 'use_far_sets' variant.)
		 *
		 * This check is somewhat complicated by the need to account
937
		 * for arrays that are not a multiple of (far) copies.	This
938
939
940
941
942
943
944
945
		 * results in the need to treat the last (potentially larger)
		 * set differently.
		 */
		group_size = (rs->md.raid_disks / copies);
		last_group_start = (rs->md.raid_disks / group_size) - 1;
		last_group_start *= group_size;
		for (i = 0; i < rs->md.raid_disks; i++) {
			if (!(i % copies) && !(i > last_group_start))
946
				rebuilds_per_group = 0;
947
948
			if ((!rs->dev[i].rdev.sb_page ||
			     !test_bit(In_sync, &rs->dev[i].rdev.flags)) &&
949
			    (++rebuilds_per_group >= copies))
950
					goto too_many;
951
952
		}
		break;
953
	default:
954
955
		if (rebuild_cnt)
			return -EINVAL;
956
957
958
959
960
961
962
963
	}

	return 0;

too_many:
	return -EINVAL;
}

NeilBrown's avatar
NeilBrown committed
964
965
966
967
/*
 * Possible arguments are...
 *	<chunk_size> [optional_args]
 *
968
969
 * Argument definitions
 *    <chunk_size>			The number of sectors per disk that
970
 *					will form the "stripe"
971
 *    [[no]sync]			Force or prevent recovery of the
972
 *					entire array
NeilBrown's avatar
NeilBrown committed
973
 *    [rebuild <idx>]			Rebuild the drive indicated by the index
974
 *    [daemon_sleep <ms>]		Time between bitmap daemon work to
975
 *					clear bits
NeilBrown's avatar
NeilBrown committed
976
977
 *    [min_recovery_rate <kB/sec/disk>]	Throttle RAID initialization
 *    [max_recovery_rate <kB/sec/disk>]	Throttle RAID initialization
978
 *    [write_mostly <idx>]		Indicate a write mostly drive via index
NeilBrown's avatar
NeilBrown committed
979
980
 *    [max_write_behind <sectors>]	See '-write-behind=' (man mdadm)
 *    [stripe_cache <sectors>]		Stripe cache size for higher RAIDs
981
 *    [region_size <sectors>]		Defines granularity of bitmap
982
983
 *
 * RAID10-only options:
984
 *    [raid10_copies <# copies>]	Number of copies.  (Default: 2)
985
 *    [raid10_format <near|far|offset>] Layout algorithm.  (Default: near)
NeilBrown's avatar
NeilBrown committed
986
 */
987
static int parse_raid_params(struct raid_set *rs, struct dm_arg_set *as,
NeilBrown's avatar
NeilBrown committed
988
989
			     unsigned num_raid_params)
{
990
	int raid10_format = ALGORITHM_RAID10_DEFAULT;
991
	unsigned raid10_copies = 2;
992
	unsigned i;
993
	unsigned value, region_size = 0;
994
	sector_t sectors_per_dev = rs->ti->len;
995
	sector_t max_io_len;
996
	const char *arg, *key;
997
	struct raid_dev *rd;
998
	struct raid_type *rt = rs->raid_type;
999
1000

	arg = dm_shift_arg(as);