xfs_ialloc_btree.c 18.5 KB
Newer Older
D
Dave Chinner 已提交
1
// SPDX-License-Identifier: GPL-2.0
L
Linus Torvalds 已提交
2
/*
3 4
 * Copyright (c) 2000-2001,2005 Silicon Graphics, Inc.
 * All Rights Reserved.
L
Linus Torvalds 已提交
5 6
 */
#include "xfs.h"
7
#include "xfs_fs.h"
8
#include "xfs_shared.h"
9
#include "xfs_format.h"
10 11
#include "xfs_log_format.h"
#include "xfs_trans_resv.h"
12
#include "xfs_bit.h"
L
Linus Torvalds 已提交
13 14
#include "xfs_mount.h"
#include "xfs_btree.h"
15
#include "xfs_btree_staging.h"
L
Linus Torvalds 已提交
16
#include "xfs_ialloc.h"
17
#include "xfs_ialloc_btree.h"
L
Linus Torvalds 已提交
18 19
#include "xfs_alloc.h"
#include "xfs_error.h"
20
#include "xfs_trace.h"
21
#include "xfs_trans.h"
22
#include "xfs_rmap.h"
23
#include "xfs_ag.h"
L
Linus Torvalds 已提交
24

25 26 27 28 29
STATIC int
xfs_inobt_get_minrecs(
	struct xfs_btree_cur	*cur,
	int			level)
{
D
Darrick J. Wong 已提交
30
	return M_IGEO(cur->bc_mp)->inobt_mnr[level != 0];
31
}
L
Linus Torvalds 已提交
32

33 34 35 36 37
STATIC struct xfs_btree_cur *
xfs_inobt_dup_cursor(
	struct xfs_btree_cur	*cur)
{
	return xfs_inobt_init_cursor(cur->bc_mp, cur->bc_tp,
38
			cur->bc_ag.agbp, cur->bc_ag.pag, cur->bc_btnum);
39 40
}

41 42 43 44 45 46
STATIC void
xfs_inobt_set_root(
	struct xfs_btree_cur	*cur,
	union xfs_btree_ptr	*nptr,
	int			inc)	/* level change */
{
47
	struct xfs_buf		*agbp = cur->bc_ag.agbp;
C
Christoph Hellwig 已提交
48
	struct xfs_agi		*agi = agbp->b_addr;
49 50 51 52 53 54

	agi->agi_root = nptr->s;
	be32_add_cpu(&agi->agi_level, inc);
	xfs_ialloc_log_agi(cur->bc_tp, agbp, XFS_AGI_ROOT | XFS_AGI_LEVEL);
}

55 56 57 58 59 60
STATIC void
xfs_finobt_set_root(
	struct xfs_btree_cur	*cur,
	union xfs_btree_ptr	*nptr,
	int			inc)	/* level change */
{
61
	struct xfs_buf		*agbp = cur->bc_ag.agbp;
C
Christoph Hellwig 已提交
62
	struct xfs_agi		*agi = agbp->b_addr;
63 64 65 66 67 68 69

	agi->agi_free_root = nptr->s;
	be32_add_cpu(&agi->agi_free_level, inc);
	xfs_ialloc_log_agi(cur->bc_tp, agbp,
			   XFS_AGI_FREE_ROOT | XFS_AGI_FREE_LEVEL);
}

70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88
/* Update the inode btree block counter for this btree. */
static inline void
xfs_inobt_mod_blockcount(
	struct xfs_btree_cur	*cur,
	int			howmuch)
{
	struct xfs_buf		*agbp = cur->bc_ag.agbp;
	struct xfs_agi		*agi = agbp->b_addr;

	if (!xfs_sb_version_hasinobtcounts(&cur->bc_mp->m_sb))
		return;

	if (cur->bc_btnum == XFS_BTNUM_FINO)
		be32_add_cpu(&agi->agi_fblocks, howmuch);
	else if (cur->bc_btnum == XFS_BTNUM_INO)
		be32_add_cpu(&agi->agi_iblocks, howmuch);
	xfs_ialloc_log_agi(cur->bc_tp, agbp, XFS_AGI_IBLOCKS);
}

89
STATIC int
90
__xfs_inobt_alloc_block(
91 92 93
	struct xfs_btree_cur	*cur,
	union xfs_btree_ptr	*start,
	union xfs_btree_ptr	*new,
94 95
	int			*stat,
	enum xfs_ag_resv_type	resv)
96 97 98 99 100 101 102 103
{
	xfs_alloc_arg_t		args;		/* block allocation args */
	int			error;		/* error return value */
	xfs_agblock_t		sbno = be32_to_cpu(start->s);

	memset(&args, 0, sizeof(args));
	args.tp = cur->bc_tp;
	args.mp = cur->bc_mp;
104
	args.oinfo = XFS_RMAP_OINFO_INOBT;
105
	args.fsbno = XFS_AGB_TO_FSB(args.mp, cur->bc_ag.pag->pag_agno, sbno);
106 107 108 109
	args.minlen = 1;
	args.maxlen = 1;
	args.prod = 1;
	args.type = XFS_ALLOCTYPE_NEAR_BNO;
110
	args.resv = resv;
111 112

	error = xfs_alloc_vextent(&args);
C
Carlos Maiolino 已提交
113
	if (error)
114
		return error;
C
Carlos Maiolino 已提交
115

116 117 118 119 120 121 122 123
	if (args.fsbno == NULLFSBLOCK) {
		*stat = 0;
		return 0;
	}
	ASSERT(args.len == 1);

	new->s = cpu_to_be32(XFS_FSB_TO_AGBNO(args.mp, args.fsbno));
	*stat = 1;
124
	xfs_inobt_mod_blockcount(cur, 1);
125 126 127
	return 0;
}

128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144
STATIC int
xfs_inobt_alloc_block(
	struct xfs_btree_cur	*cur,
	union xfs_btree_ptr	*start,
	union xfs_btree_ptr	*new,
	int			*stat)
{
	return __xfs_inobt_alloc_block(cur, start, new, stat, XFS_AG_RESV_NONE);
}

STATIC int
xfs_finobt_alloc_block(
	struct xfs_btree_cur	*cur,
	union xfs_btree_ptr	*start,
	union xfs_btree_ptr	*new,
	int			*stat)
{
145
	if (cur->bc_mp->m_finobt_nores)
146
		return xfs_inobt_alloc_block(cur, start, new, stat);
147 148 149 150
	return __xfs_inobt_alloc_block(cur, start, new, stat,
			XFS_AG_RESV_METADATA);
}

151
STATIC int
152
__xfs_inobt_free_block(
153
	struct xfs_btree_cur	*cur,
154 155
	struct xfs_buf		*bp,
	enum xfs_ag_resv_type	resv)
156
{
157
	xfs_inobt_mod_blockcount(cur, -1);
158
	return xfs_free_extent(cur->bc_tp,
159
			XFS_DADDR_TO_FSB(cur->bc_mp, XFS_BUF_ADDR(bp)), 1,
160
			&XFS_RMAP_OINFO_INOBT, resv);
161 162 163 164 165 166 167 168 169 170 171 172 173 174 175
}

STATIC int
xfs_inobt_free_block(
	struct xfs_btree_cur	*cur,
	struct xfs_buf		*bp)
{
	return __xfs_inobt_free_block(cur, bp, XFS_AG_RESV_NONE);
}

STATIC int
xfs_finobt_free_block(
	struct xfs_btree_cur	*cur,
	struct xfs_buf		*bp)
{
176
	if (cur->bc_mp->m_finobt_nores)
177 178
		return xfs_inobt_free_block(cur, bp);
	return __xfs_inobt_free_block(cur, bp, XFS_AG_RESV_METADATA);
179
}
180

181 182 183 184 185
STATIC int
xfs_inobt_get_maxrecs(
	struct xfs_btree_cur	*cur,
	int			level)
{
D
Darrick J. Wong 已提交
186
	return M_IGEO(cur->bc_mp)->inobt_mxr[level != 0];
187 188
}

189 190
STATIC void
xfs_inobt_init_key_from_rec(
191 192
	union xfs_btree_key		*key,
	const union xfs_btree_rec	*rec)
193 194 195 196
{
	key->inobt.ir_startino = rec->inobt.ir_startino;
}

197 198
STATIC void
xfs_inobt_init_high_key_from_rec(
199 200
	union xfs_btree_key		*key,
	const union xfs_btree_rec	*rec)
201
{
202
	__u32				x;
203 204 205 206 207 208

	x = be32_to_cpu(rec->inobt.ir_startino);
	x += XFS_INODES_PER_CHUNK - 1;
	key->inobt.ir_startino = cpu_to_be32(x);
}

209 210 211 212 213 214
STATIC void
xfs_inobt_init_rec_from_cur(
	struct xfs_btree_cur	*cur,
	union xfs_btree_rec	*rec)
{
	rec->inobt.ir_startino = cpu_to_be32(cur->bc_rec.i.ir_startino);
215 216 217 218 219 220 221 222 223 224
	if (xfs_sb_version_hassparseinodes(&cur->bc_mp->m_sb)) {
		rec->inobt.ir_u.sp.ir_holemask =
					cpu_to_be16(cur->bc_rec.i.ir_holemask);
		rec->inobt.ir_u.sp.ir_count = cur->bc_rec.i.ir_count;
		rec->inobt.ir_u.sp.ir_freecount = cur->bc_rec.i.ir_freecount;
	} else {
		/* ir_holemask/ir_count not supported on-disk */
		rec->inobt.ir_u.f.ir_freecount =
					cpu_to_be32(cur->bc_rec.i.ir_freecount);
	}
225 226 227
	rec->inobt.ir_free = cpu_to_be64(cur->bc_rec.i.ir_free);
}

228
/*
M
Malcolm Parsons 已提交
229
 * initial value of ptr for lookup
230 231 232 233 234 235
 */
STATIC void
xfs_inobt_init_ptr_from_cur(
	struct xfs_btree_cur	*cur,
	union xfs_btree_ptr	*ptr)
{
236
	struct xfs_agi		*agi = cur->bc_ag.agbp->b_addr;
237

238
	ASSERT(cur->bc_ag.pag->pag_agno == be32_to_cpu(agi->agi_seqno));
239 240 241 242

	ptr->s = agi->agi_root;
}

243 244 245 246 247
STATIC void
xfs_finobt_init_ptr_from_cur(
	struct xfs_btree_cur	*cur,
	union xfs_btree_ptr	*ptr)
{
248
	struct xfs_agi		*agi = cur->bc_ag.agbp->b_addr;
249

250
	ASSERT(cur->bc_ag.pag->pag_agno == be32_to_cpu(agi->agi_seqno));
251 252 253
	ptr->s = agi->agi_free_root;
}

254
STATIC int64_t
255
xfs_inobt_key_diff(
256 257
	struct xfs_btree_cur		*cur,
	const union xfs_btree_key	*key)
258
{
259
	return (int64_t)be32_to_cpu(key->inobt.ir_startino) -
260 261 262
			  cur->bc_rec.i.ir_startino;
}

263 264
STATIC int64_t
xfs_inobt_diff_two_keys(
265 266 267
	struct xfs_btree_cur		*cur,
	const union xfs_btree_key	*k1,
	const union xfs_btree_key	*k2)
268 269 270 271 272
{
	return (int64_t)be32_to_cpu(k1->inobt.ir_startino) -
			  be32_to_cpu(k2->inobt.ir_startino);
}

273
static xfs_failaddr_t
274
xfs_inobt_verify(
275 276
	struct xfs_buf		*bp)
{
277
	struct xfs_mount	*mp = bp->b_mount;
278
	struct xfs_btree_block	*block = XFS_BUF_TO_BLOCK(bp);
279
	xfs_failaddr_t		fa;
280 281
	unsigned int		level;

282 283 284
	if (!xfs_verify_magic(bp, block->bb_magic))
		return __this_address;

285 286 287 288 289 290 291 292 293 294
	/*
	 * During growfs operations, we can't verify the exact owner as the
	 * perag is not fully initialised and hence not attached to the buffer.
	 *
	 * Similarly, during log recovery we will have a perag structure
	 * attached, but the agi information will not yet have been initialised
	 * from the on disk AGI. We don't currently use any of this information,
	 * but beware of the landmine (i.e. need to check pag->pagi_init) if we
	 * ever do.
	 */
295
	if (xfs_sb_version_hascrc(&mp->m_sb)) {
296 297 298
		fa = xfs_btree_sblock_v5hdr_verify(bp);
		if (fa)
			return fa;
299
	}
300

301
	/* level verification */
302
	level = be16_to_cpu(block->bb_level);
D
Darrick J. Wong 已提交
303
	if (level >= M_IGEO(mp)->inobt_maxlevels)
304
		return __this_address;
305

D
Darrick J. Wong 已提交
306 307
	return xfs_btree_sblock_verify(bp,
			M_IGEO(mp)->inobt_mxr[level != 0]);
308 309 310
}

static void
311
xfs_inobt_read_verify(
312 313
	struct xfs_buf	*bp)
{
314 315
	xfs_failaddr_t	fa;

316
	if (!xfs_btree_sblock_verify_crc(bp))
317 318 319 320 321 322
		xfs_verifier_error(bp, -EFSBADCRC, __this_address);
	else {
		fa = xfs_inobt_verify(bp);
		if (fa)
			xfs_verifier_error(bp, -EFSCORRUPTED, fa);
	}
323

324
	if (bp->b_error)
325
		trace_xfs_btree_corrupt(bp, _RET_IP_);
326
}
327

328 329
static void
xfs_inobt_write_verify(
330 331
	struct xfs_buf	*bp)
{
332 333 334 335
	xfs_failaddr_t	fa;

	fa = xfs_inobt_verify(bp);
	if (fa) {
336
		trace_xfs_btree_corrupt(bp, _RET_IP_);
337
		xfs_verifier_error(bp, -EFSCORRUPTED, fa);
338
		return;
339 340 341
	}
	xfs_btree_sblock_calc_crc(bp);

342 343
}

344
const struct xfs_buf_ops xfs_inobt_buf_ops = {
345
	.name = "xfs_inobt",
346
	.magic = { cpu_to_be32(XFS_IBT_MAGIC), cpu_to_be32(XFS_IBT_CRC_MAGIC) },
347 348
	.verify_read = xfs_inobt_read_verify,
	.verify_write = xfs_inobt_write_verify,
349
	.verify_struct = xfs_inobt_verify,
350 351
};

352 353
const struct xfs_buf_ops xfs_finobt_buf_ops = {
	.name = "xfs_finobt",
354 355
	.magic = { cpu_to_be32(XFS_FIBT_MAGIC),
		   cpu_to_be32(XFS_FIBT_CRC_MAGIC) },
356 357 358 359 360
	.verify_read = xfs_inobt_read_verify,
	.verify_write = xfs_inobt_write_verify,
	.verify_struct = xfs_inobt_verify,
};

361 362
STATIC int
xfs_inobt_keys_inorder(
363 364 365
	struct xfs_btree_cur		*cur,
	const union xfs_btree_key	*k1,
	const union xfs_btree_key	*k2)
366 367 368 369 370 371 372
{
	return be32_to_cpu(k1->inobt.ir_startino) <
		be32_to_cpu(k2->inobt.ir_startino);
}

STATIC int
xfs_inobt_recs_inorder(
373 374 375
	struct xfs_btree_cur		*cur,
	const union xfs_btree_rec	*r1,
	const union xfs_btree_rec	*r2)
376 377 378 379 380
{
	return be32_to_cpu(r1->inobt.ir_startino) + XFS_INODES_PER_CHUNK <=
		be32_to_cpu(r2->inobt.ir_startino);
}

381
static const struct xfs_btree_ops xfs_inobt_ops = {
382 383 384
	.rec_len		= sizeof(xfs_inobt_rec_t),
	.key_len		= sizeof(xfs_inobt_key_t),

385
	.dup_cursor		= xfs_inobt_dup_cursor,
386
	.set_root		= xfs_inobt_set_root,
387
	.alloc_block		= xfs_inobt_alloc_block,
388
	.free_block		= xfs_inobt_free_block,
389
	.get_minrecs		= xfs_inobt_get_minrecs,
390
	.get_maxrecs		= xfs_inobt_get_maxrecs,
391
	.init_key_from_rec	= xfs_inobt_init_key_from_rec,
392
	.init_high_key_from_rec	= xfs_inobt_init_high_key_from_rec,
393
	.init_rec_from_cur	= xfs_inobt_init_rec_from_cur,
394 395
	.init_ptr_from_cur	= xfs_inobt_init_ptr_from_cur,
	.key_diff		= xfs_inobt_key_diff,
396
	.buf_ops		= &xfs_inobt_buf_ops,
397
	.diff_two_keys		= xfs_inobt_diff_two_keys,
398 399
	.keys_inorder		= xfs_inobt_keys_inorder,
	.recs_inorder		= xfs_inobt_recs_inorder,
400 401
};

402 403 404 405 406 407
static const struct xfs_btree_ops xfs_finobt_ops = {
	.rec_len		= sizeof(xfs_inobt_rec_t),
	.key_len		= sizeof(xfs_inobt_key_t),

	.dup_cursor		= xfs_inobt_dup_cursor,
	.set_root		= xfs_finobt_set_root,
408
	.alloc_block		= xfs_finobt_alloc_block,
409
	.free_block		= xfs_finobt_free_block,
410 411 412
	.get_minrecs		= xfs_inobt_get_minrecs,
	.get_maxrecs		= xfs_inobt_get_maxrecs,
	.init_key_from_rec	= xfs_inobt_init_key_from_rec,
413
	.init_high_key_from_rec	= xfs_inobt_init_high_key_from_rec,
414 415 416
	.init_rec_from_cur	= xfs_inobt_init_rec_from_cur,
	.init_ptr_from_cur	= xfs_finobt_init_ptr_from_cur,
	.key_diff		= xfs_inobt_key_diff,
417
	.buf_ops		= &xfs_finobt_buf_ops,
418
	.diff_two_keys		= xfs_inobt_diff_two_keys,
419 420 421 422
	.keys_inorder		= xfs_inobt_keys_inorder,
	.recs_inorder		= xfs_inobt_recs_inorder,
};

423
/*
424
 * Initialize a new inode btree cursor.
425
 */
426 427
static struct xfs_btree_cur *
xfs_inobt_init_common(
428 429
	struct xfs_mount	*mp,		/* file system mount point */
	struct xfs_trans	*tp,		/* transaction pointer */
430
	struct xfs_perag	*pag,
431
	xfs_btnum_t		btnum)		/* ialloc or free ino btree */
432 433 434
{
	struct xfs_btree_cur	*cur;

435
	cur = kmem_cache_zalloc(xfs_btree_cur_zone, GFP_NOFS | __GFP_NOFAIL);
436 437
	cur->bc_tp = tp;
	cur->bc_mp = mp;
438
	cur->bc_btnum = btnum;
439
	if (btnum == XFS_BTNUM_INO) {
D
Dave Chinner 已提交
440
		cur->bc_statoff = XFS_STATS_CALC_INDEX(xs_ibt_2);
441
		cur->bc_ops = &xfs_inobt_ops;
442
	} else {
D
Dave Chinner 已提交
443
		cur->bc_statoff = XFS_STATS_CALC_INDEX(xs_fibt_2);
444
		cur->bc_ops = &xfs_finobt_ops;
445 446
	}

447 448
	cur->bc_blocklog = mp->m_sb.sb_blocklog;

449 450
	if (xfs_sb_version_hascrc(&mp->m_sb))
		cur->bc_flags |= XFS_BTREE_CRC_BLOCKS;
451

452 453
	/* take a reference for the cursor */
	atomic_inc(&pag->pag_ref);
454
	cur->bc_ag.pag = pag;
455 456 457 458 459 460 461 462 463
	return cur;
}

/* Create an inode btree cursor. */
struct xfs_btree_cur *
xfs_inobt_init_cursor(
	struct xfs_mount	*mp,
	struct xfs_trans	*tp,
	struct xfs_buf		*agbp,
464
	struct xfs_perag	*pag,
465 466 467 468 469
	xfs_btnum_t		btnum)
{
	struct xfs_btree_cur	*cur;
	struct xfs_agi		*agi = agbp->b_addr;

470
	cur = xfs_inobt_init_common(mp, tp, pag, btnum);
471 472 473 474 475 476 477
	if (btnum == XFS_BTNUM_INO)
		cur->bc_nlevels = be32_to_cpu(agi->agi_level);
	else
		cur->bc_nlevels = be32_to_cpu(agi->agi_free_level);
	cur->bc_ag.agbp = agbp;
	return cur;
}
478

479 480 481 482 483
/* Create an inode btree cursor with a fake root for staging. */
struct xfs_btree_cur *
xfs_inobt_stage_cursor(
	struct xfs_mount	*mp,
	struct xbtree_afakeroot	*afake,
484
	struct xfs_perag	*pag,
485 486 487 488
	xfs_btnum_t		btnum)
{
	struct xfs_btree_cur	*cur;

489
	cur = xfs_inobt_init_common(mp, NULL, pag, btnum);
490
	xfs_btree_stage_afakeroot(cur, afake);
491 492
	return cur;
}
493

494 495 496 497 498 499 500 501 502 503 504 505
/*
 * Install a new inobt btree root.  Caller is responsible for invalidating
 * and freeing the old btree blocks.
 */
void
xfs_inobt_commit_staged_btree(
	struct xfs_btree_cur	*cur,
	struct xfs_trans	*tp,
	struct xfs_buf		*agbp)
{
	struct xfs_agi		*agi = agbp->b_addr;
	struct xbtree_afakeroot	*afake = cur->bc_ag.afake;
506
	int			fields;
507 508 509 510

	ASSERT(cur->bc_flags & XFS_BTREE_STAGING);

	if (cur->bc_btnum == XFS_BTNUM_INO) {
511
		fields = XFS_AGI_ROOT | XFS_AGI_LEVEL;
512 513
		agi->agi_root = cpu_to_be32(afake->af_root);
		agi->agi_level = cpu_to_be32(afake->af_levels);
514 515 516 517 518
		if (xfs_sb_version_hasinobtcounts(&cur->bc_mp->m_sb)) {
			agi->agi_iblocks = cpu_to_be32(afake->af_blocks);
			fields |= XFS_AGI_IBLOCKS;
		}
		xfs_ialloc_log_agi(tp, agbp, fields);
519 520
		xfs_btree_commit_afakeroot(cur, tp, agbp, &xfs_inobt_ops);
	} else {
521
		fields = XFS_AGI_FREE_ROOT | XFS_AGI_FREE_LEVEL;
522 523
		agi->agi_free_root = cpu_to_be32(afake->af_root);
		agi->agi_free_level = cpu_to_be32(afake->af_levels);
524 525 526 527 528
		if (xfs_sb_version_hasinobtcounts(&cur->bc_mp->m_sb)) {
			agi->agi_fblocks = cpu_to_be32(afake->af_blocks);
			fields |= XFS_AGI_IBLOCKS;
		}
		xfs_ialloc_log_agi(tp, agbp, fields);
529 530 531 532
		xfs_btree_commit_afakeroot(cur, tp, agbp, &xfs_finobt_ops);
	}
}

533 534 535 536 537 538 539 540 541
/*
 * Calculate number of records in an inobt btree block.
 */
int
xfs_inobt_maxrecs(
	struct xfs_mount	*mp,
	int			blocklen,
	int			leaf)
{
542
	blocklen -= XFS_INOBT_BLOCK_LEN(mp);
543 544 545 546 547

	if (leaf)
		return blocklen / sizeof(xfs_inobt_rec_t);
	return blocklen / (sizeof(xfs_inobt_key_t) + sizeof(xfs_inobt_ptr_t));
}
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

/*
 * Convert the inode record holemask to an inode allocation bitmap. The inode
 * allocation bitmap is inode granularity and specifies whether an inode is
 * physically allocated on disk (not whether the inode is considered allocated
 * or free by the fs).
 *
 * A bit value of 1 means the inode is allocated, a value of 0 means it is free.
 */
uint64_t
xfs_inobt_irec_to_allocmask(
	struct xfs_inobt_rec_incore	*rec)
{
	uint64_t			bitmap = 0;
	uint64_t			inodespbit;
	int				nextbit;
	uint				allocbitmap;

	/*
	 * The holemask has 16-bits for a 64 inode record. Therefore each
	 * holemask bit represents multiple inodes. Create a mask of bits to set
	 * in the allocmask for each holemask bit.
	 */
	inodespbit = (1 << XFS_INODES_PER_HOLEMASK_BIT) - 1;

	/*
	 * Allocated inodes are represented by 0 bits in holemask. Invert the 0
	 * bits to 1 and convert to a uint so we can use xfs_next_bit(). Mask
	 * anything beyond the 16 holemask bits since this casts to a larger
	 * type.
	 */
	allocbitmap = ~rec->ir_holemask & ((1 << XFS_INOBT_HOLEMASK_BITS) - 1);

	/*
	 * allocbitmap is the inverted holemask so every set bit represents
	 * allocated inodes. To expand from 16-bit holemask granularity to
	 * 64-bit (e.g., bit-per-inode), set inodespbit bits in the target
	 * bitmap for every holemask bit.
	 */
	nextbit = xfs_next_bit(&allocbitmap, 1, 0);
	while (nextbit != -1) {
		ASSERT(nextbit < (sizeof(rec->ir_holemask) * NBBY));

		bitmap |= (inodespbit <<
			   (nextbit * XFS_INODES_PER_HOLEMASK_BIT));

		nextbit = xfs_next_bit(&allocbitmap, 1, nextbit + 1);
	}

	return bitmap;
}
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

#if defined(DEBUG) || defined(XFS_WARN)
/*
 * Verify that an in-core inode record has a valid inode count.
 */
int
xfs_inobt_rec_check_count(
	struct xfs_mount		*mp,
	struct xfs_inobt_rec_incore	*rec)
{
	int				inocount = 0;
	int				nextbit = 0;
	uint64_t			allocbmap;
	int				wordsz;

	wordsz = sizeof(allocbmap) / sizeof(unsigned int);
	allocbmap = xfs_inobt_irec_to_allocmask(rec);

	nextbit = xfs_next_bit((uint *) &allocbmap, wordsz, nextbit);
	while (nextbit != -1) {
		inocount++;
		nextbit = xfs_next_bit((uint *) &allocbmap, wordsz,
				       nextbit + 1);
	}

	if (inocount != rec->ir_count)
		return -EFSCORRUPTED;

	return 0;
}
#endif	/* DEBUG */
630 631 632

static xfs_extlen_t
xfs_inobt_max_size(
633 634
	struct xfs_mount	*mp,
	xfs_agnumber_t		agno)
635
{
636 637
	xfs_agblock_t		agblocks = xfs_ag_block_count(mp, agno);

638
	/* Bail out if we're uninitialized, which can happen in mkfs. */
D
Darrick J. Wong 已提交
639
	if (M_IGEO(mp)->inobt_mxr[0] == 0)
640 641
		return 0;

642 643 644 645 646 647 648 649 650
	/*
	 * The log is permanently allocated, so the space it occupies will
	 * never be available for the kinds of things that would require btree
	 * expansion.  We therefore can pretend the space isn't there.
	 */
	if (mp->m_sb.sb_logstart &&
	    XFS_FSB_TO_AGNO(mp, mp->m_sb.sb_logstart) == agno)
		agblocks -= mp->m_sb.sb_logblocks;

D
Darrick J. Wong 已提交
651
	return xfs_btree_calc_size(M_IGEO(mp)->inobt_mnr,
652 653
				(uint64_t)agblocks * mp->m_sb.sb_inopblock /
					XFS_INODES_PER_CHUNK);
654 655
}

656 657 658 659 660
/* Read AGI and create inobt cursor. */
int
xfs_inobt_cur(
	struct xfs_mount	*mp,
	struct xfs_trans	*tp,
661
	struct xfs_perag	*pag,
662 663 664 665 666 667 668 669 670 671
	xfs_btnum_t		which,
	struct xfs_btree_cur	**curpp,
	struct xfs_buf		**agi_bpp)
{
	struct xfs_btree_cur	*cur;
	int			error;

	ASSERT(*agi_bpp == NULL);
	ASSERT(*curpp == NULL);

672
	error = xfs_ialloc_read_agi(mp, tp, pag->pag_agno, agi_bpp);
673 674 675
	if (error)
		return error;

676
	cur = xfs_inobt_init_cursor(mp, tp, *agi_bpp, pag, which);
677 678 679 680
	*curpp = cur;
	return 0;
}

681 682 683
static int
xfs_inobt_count_blocks(
	struct xfs_mount	*mp,
684
	struct xfs_trans	*tp,
685
	struct xfs_perag	*pag,
686 687 688
	xfs_btnum_t		btnum,
	xfs_extlen_t		*tree_blocks)
{
689 690
	struct xfs_buf		*agbp = NULL;
	struct xfs_btree_cur	*cur = NULL;
691 692
	int			error;

693
	error = xfs_inobt_cur(mp, tp, pag, btnum, &cur, &agbp);
694 695 696 697
	if (error)
		return error;

	error = xfs_btree_count_blocks(cur, tree_blocks);
698
	xfs_btree_del_cursor(cur, error);
699
	xfs_trans_brelse(tp, agbp);
700 701 702 703

	return error;
}

704 705 706 707 708
/* Read finobt block count from AGI header. */
static int
xfs_finobt_read_blocks(
	struct xfs_mount	*mp,
	struct xfs_trans	*tp,
709
	struct xfs_perag	*pag,
710 711 712 713 714 715
	xfs_extlen_t		*tree_blocks)
{
	struct xfs_buf		*agbp;
	struct xfs_agi		*agi;
	int			error;

716
	error = xfs_ialloc_read_agi(mp, tp, pag->pag_agno, &agbp);
717 718 719 720 721 722 723 724 725
	if (error)
		return error;

	agi = agbp->b_addr;
	*tree_blocks = be32_to_cpu(agi->agi_fblocks);
	xfs_trans_brelse(tp, agbp);
	return 0;
}

726 727 728 729 730 731
/*
 * Figure out how many blocks to reserve and how many are used by this btree.
 */
int
xfs_finobt_calc_reserves(
	struct xfs_mount	*mp,
732
	struct xfs_trans	*tp,
733
	struct xfs_perag	*pag,
734 735 736 737 738 739 740 741 742
	xfs_extlen_t		*ask,
	xfs_extlen_t		*used)
{
	xfs_extlen_t		tree_len = 0;
	int			error;

	if (!xfs_sb_version_hasfinobt(&mp->m_sb))
		return 0;

743
	if (xfs_sb_version_hasinobtcounts(&mp->m_sb))
744
		error = xfs_finobt_read_blocks(mp, tp, pag, &tree_len);
745
	else
746
		error = xfs_inobt_count_blocks(mp, tp, pag, XFS_BTNUM_FINO,
747
				&tree_len);
748 749 750
	if (error)
		return error;

751
	*ask += xfs_inobt_max_size(mp, pag->pag_agno);
752 753 754
	*used += tree_len;
	return 0;
}
755 756 757 758 759 760 761

/* Calculate the inobt btree size for some records. */
xfs_extlen_t
xfs_iallocbt_calc_size(
	struct xfs_mount	*mp,
	unsigned long long	len)
{
D
Darrick J. Wong 已提交
762
	return xfs_btree_calc_size(M_IGEO(mp)->inobt_mnr, len);
763
}