xfs_bmap.c 173.9 KB
Newer Older
L
Linus Torvalds 已提交
1
/*
2
 * Copyright (c) 2000-2006 Silicon Graphics, Inc.
3
 * All Rights Reserved.
L
Linus Torvalds 已提交
4
 *
5 6
 * This program is free software; you can redistribute it and/or
 * modify it under the terms of the GNU General Public License as
L
Linus Torvalds 已提交
7 8
 * published by the Free Software Foundation.
 *
9 10 11 12
 * This program is distributed in the hope that it would be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
L
Linus Torvalds 已提交
13
 *
14 15 16
 * You should have received a copy of the GNU General Public License
 * along with this program; if not, write the Free Software Foundation,
 * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
L
Linus Torvalds 已提交
17 18
 */
#include "xfs.h"
19
#include "xfs_fs.h"
20
#include "xfs_shared.h"
21 22 23
#include "xfs_format.h"
#include "xfs_log_format.h"
#include "xfs_trans_resv.h"
24
#include "xfs_bit.h"
L
Linus Torvalds 已提交
25
#include "xfs_sb.h"
26
#include "xfs_mount.h"
27
#include "xfs_defer.h"
28
#include "xfs_da_format.h"
29
#include "xfs_da_btree.h"
30
#include "xfs_dir2.h"
L
Linus Torvalds 已提交
31
#include "xfs_inode.h"
32
#include "xfs_btree.h"
33
#include "xfs_trans.h"
34
#include "xfs_inode_item.h"
L
Linus Torvalds 已提交
35 36 37
#include "xfs_extfree_item.h"
#include "xfs_alloc.h"
#include "xfs_bmap.h"
D
Dave Chinner 已提交
38
#include "xfs_bmap_util.h"
39
#include "xfs_bmap_btree.h"
L
Linus Torvalds 已提交
40 41 42 43 44
#include "xfs_rtalloc.h"
#include "xfs_error.h"
#include "xfs_quota.h"
#include "xfs_trans_space.h"
#include "xfs_buf_item.h"
C
Christoph Hellwig 已提交
45
#include "xfs_trace.h"
46
#include "xfs_symlink.h"
47 48
#include "xfs_attr_leaf.h"
#include "xfs_filestream.h"
49
#include "xfs_rmap.h"
50
#include "xfs_ag_resv.h"
51
#include "xfs_refcount.h"
L
Linus Torvalds 已提交
52 53 54 55 56


kmem_zone_t		*xfs_bmap_free_item_zone;

/*
57
 * Miscellaneous helper functions
L
Linus Torvalds 已提交
58 59 60
 */

/*
61 62
 * Compute and fill in the value of the maximum depth of a bmap btree
 * in this filesystem.  Done once, during mount.
L
Linus Torvalds 已提交
63
 */
64 65 66 67 68 69 70 71 72 73 74 75
void
xfs_bmap_compute_maxlevels(
	xfs_mount_t	*mp,		/* file system mount structure */
	int		whichfork)	/* data or attr fork */
{
	int		level;		/* btree level */
	uint		maxblocks;	/* max blocks at this level */
	uint		maxleafents;	/* max leaf entries possible */
	int		maxrootrecs;	/* max records in root block */
	int		minleafrecs;	/* min records in leaf block */
	int		minnoderecs;	/* min records in node block */
	int		sz;		/* root block size */
L
Linus Torvalds 已提交
76

77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97
	/*
	 * The maximum number of extents in a file, hence the maximum
	 * number of leaf entries, is controlled by the type of di_nextents
	 * (a signed 32-bit number, xfs_extnum_t), or by di_anextents
	 * (a signed 16-bit number, xfs_aextnum_t).
	 *
	 * Note that we can no longer assume that if we are in ATTR1 that
	 * the fork offset of all the inodes will be
	 * (xfs_default_attroffset(ip) >> 3) because we could have mounted
	 * with ATTR2 and then mounted back with ATTR1, keeping the
	 * di_forkoff's fixed but probably at various positions. Therefore,
	 * for both ATTR1 and ATTR2 we have to assume the worst case scenario
	 * of a minimum size available.
	 */
	if (whichfork == XFS_DATA_FORK) {
		maxleafents = MAXEXTNUM;
		sz = XFS_BMDR_SPACE_CALC(MINDBTPTRS);
	} else {
		maxleafents = MAXAEXTNUM;
		sz = XFS_BMDR_SPACE_CALC(MINABTPTRS);
	}
98
	maxrootrecs = xfs_bmdr_maxrecs(sz, 0);
99 100 101 102 103 104 105 106 107 108 109
	minleafrecs = mp->m_bmap_dmnr[0];
	minnoderecs = mp->m_bmap_dmnr[1];
	maxblocks = (maxleafents + minleafrecs - 1) / minleafrecs;
	for (level = 1; maxblocks > 1; level++) {
		if (maxblocks <= maxrootrecs)
			maxblocks = 1;
		else
			maxblocks = (maxblocks + minnoderecs - 1) / minnoderecs;
	}
	mp->m_bm_maxlevels[whichfork] = level;
}
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
STATIC int				/* error */
xfs_bmbt_lookup_eq(
	struct xfs_btree_cur	*cur,
	xfs_fileoff_t		off,
	xfs_fsblock_t		bno,
	xfs_filblks_t		len,
	int			*stat)	/* success/failure */
{
	cur->bc_rec.b.br_startoff = off;
	cur->bc_rec.b.br_startblock = bno;
	cur->bc_rec.b.br_blockcount = len;
	return xfs_btree_lookup(cur, XFS_LOOKUP_EQ, stat);
}

STATIC int				/* error */
xfs_bmbt_lookup_ge(
	struct xfs_btree_cur	*cur,
	xfs_fileoff_t		off,
	xfs_fsblock_t		bno,
	xfs_filblks_t		len,
	int			*stat)	/* success/failure */
{
	cur->bc_rec.b.br_startoff = off;
	cur->bc_rec.b.br_startblock = bno;
	cur->bc_rec.b.br_blockcount = len;
	return xfs_btree_lookup(cur, XFS_LOOKUP_GE, stat);
}

139
/*
140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160
 * Check if the inode needs to be converted to btree format.
 */
static inline bool xfs_bmap_needs_btree(struct xfs_inode *ip, int whichfork)
{
	return XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_EXTENTS &&
		XFS_IFORK_NEXTENTS(ip, whichfork) >
			XFS_IFORK_MAXEXT(ip, whichfork);
}

/*
 * Check if the inode should be converted to extent format.
 */
static inline bool xfs_bmap_wants_extents(struct xfs_inode *ip, int whichfork)
{
	return XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_BTREE &&
		XFS_IFORK_NEXTENTS(ip, whichfork) <=
			XFS_IFORK_MAXEXT(ip, whichfork);
}

/*
 * Update the record referred to by cur to the value given
161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176
 * by [off, bno, len, state].
 * This either works (return 0) or gets an EFSCORRUPTED error.
 */
STATIC int
xfs_bmbt_update(
	struct xfs_btree_cur	*cur,
	xfs_fileoff_t		off,
	xfs_fsblock_t		bno,
	xfs_filblks_t		len,
	xfs_exntst_t		state)
{
	union xfs_btree_rec	rec;

	xfs_bmbt_disk_set_allf(&rec.bmbt, off, bno, len, state);
	return xfs_btree_update(cur, &rec);
}
177

L
Linus Torvalds 已提交
178
/*
179 180
 * Compute the worst-case number of indirect blocks that will be used
 * for ip's delayed extent of length "len".
L
Linus Torvalds 已提交
181
 */
182 183 184 185
STATIC xfs_filblks_t
xfs_bmap_worst_indlen(
	xfs_inode_t	*ip,		/* incore inode pointer */
	xfs_filblks_t	len)		/* delayed extent length */
L
Linus Torvalds 已提交
186
{
187 188 189 190
	int		level;		/* btree level number */
	int		maxrecs;	/* maximum record count at this level */
	xfs_mount_t	*mp;		/* mount structure */
	xfs_filblks_t	rval;		/* return value */
L
Linus Torvalds 已提交
191 192

	mp = ip->i_mount;
193 194 195 196 197 198 199 200 201 202 203 204
	maxrecs = mp->m_bmap_dmxr[0];
	for (level = 0, rval = 0;
	     level < XFS_BM_MAXLEVELS(mp, XFS_DATA_FORK);
	     level++) {
		len += maxrecs - 1;
		do_div(len, maxrecs);
		rval += len;
		if (len == 1)
			return rval + XFS_BM_MAXLEVELS(mp, XFS_DATA_FORK) -
				level - 1;
		if (level == 0)
			maxrecs = mp->m_bmap_dmxr[1];
L
Linus Torvalds 已提交
205
	}
206
	return rval;
L
Linus Torvalds 已提交
207 208 209
}

/*
210
 * Calculate the default attribute fork offset for newly created inodes.
L
Linus Torvalds 已提交
211
 */
212 213 214
uint
xfs_default_attroffset(
	struct xfs_inode	*ip)
L
Linus Torvalds 已提交
215
{
216 217
	struct xfs_mount	*mp = ip->i_mount;
	uint			offset;
L
Linus Torvalds 已提交
218

219
	if (mp->m_sb.sb_inodesize == 256) {
220
		offset = XFS_LITINO(mp, ip->i_d.di_version) -
221 222 223
				XFS_BMDR_SPACE_CALC(MINABTPTRS);
	} else {
		offset = XFS_BMDR_SPACE_CALC(6 * MINABTPTRS);
L
Linus Torvalds 已提交
224
	}
225

226
	ASSERT(offset < XFS_LITINO(mp, ip->i_d.di_version));
227
	return offset;
L
Linus Torvalds 已提交
228 229 230
}

/*
231 232 233
 * Helper routine to reset inode di_forkoff field when switching
 * attribute fork from local to extent format - we reset it where
 * possible to make space available for inline data fork extents.
234 235
 */
STATIC void
236 237 238
xfs_bmap_forkoff_reset(
	xfs_inode_t	*ip,
	int		whichfork)
239
{
240 241 242 243 244 245 246 247 248
	if (whichfork == XFS_ATTR_FORK &&
	    ip->i_d.di_format != XFS_DINODE_FMT_DEV &&
	    ip->i_d.di_format != XFS_DINODE_FMT_UUID &&
	    ip->i_d.di_format != XFS_DINODE_FMT_BTREE) {
		uint	dfl_forkoff = xfs_default_attroffset(ip) >> 3;

		if (dfl_forkoff > ip->i_d.di_forkoff)
			ip->i_d.di_forkoff = dfl_forkoff;
	}
249 250
}

251 252 253 254 255 256 257 258
#ifdef DEBUG
STATIC struct xfs_buf *
xfs_bmap_get_bp(
	struct xfs_btree_cur	*cur,
	xfs_fsblock_t		bno)
{
	struct xfs_log_item_desc *lidp;
	int			i;
259

260 261 262 263 264 265 266 267
	if (!cur)
		return NULL;

	for (i = 0; i < XFS_BTREE_MAXLEVELS; i++) {
		if (!cur->bc_bufs[i])
			break;
		if (XFS_BUF_ADDR(cur->bc_bufs[i]) == bno)
			return cur->bc_bufs[i];
L
Linus Torvalds 已提交
268
	}
269

270 271 272 273 274 275 276 277
	/* Chase down all the log items to see if the bp is there */
	list_for_each_entry(lidp, &cur->bc_tp->t_items, lid_trans) {
		struct xfs_buf_log_item	*bip;
		bip = (struct xfs_buf_log_item *)lidp->lid_item;
		if (bip->bli_item.li_type == XFS_LI_BUF &&
		    XFS_BUF_ADDR(bip->bli_buf) == bno)
			return bip->bli_buf;
	}
278

279 280
	return NULL;
}
C
Christoph Hellwig 已提交
281

282 283 284 285 286 287 288 289 290 291
STATIC void
xfs_check_block(
	struct xfs_btree_block	*block,
	xfs_mount_t		*mp,
	int			root,
	short			sz)
{
	int			i, j, dmxr;
	__be64			*pp, *thispa;	/* pointer to block address */
	xfs_bmbt_key_t		*prevp, *keyp;
L
Linus Torvalds 已提交
292

293
	ASSERT(be16_to_cpu(block->bb_level) > 0);
C
Christoph Hellwig 已提交
294

295 296 297 298
	prevp = NULL;
	for( i = 1; i <= xfs_btree_get_numrecs(block); i++) {
		dmxr = mp->m_bmap_dmxr[0];
		keyp = XFS_BMBT_KEY_ADDR(mp, block, i);
C
Christoph Hellwig 已提交
299

300 301 302
		if (prevp) {
			ASSERT(be64_to_cpu(prevp->br_startoff) <
			       be64_to_cpu(keyp->br_startoff));
L
Linus Torvalds 已提交
303
		}
304
		prevp = keyp;
L
Linus Torvalds 已提交
305 306

		/*
307
		 * Compare the block numbers to see if there are dups.
L
Linus Torvalds 已提交
308
		 */
309 310 311 312
		if (root)
			pp = XFS_BMAP_BROOT_PTR_ADDR(mp, block, i, sz);
		else
			pp = XFS_BMBT_PTR_ADDR(mp, block, i, dmxr);
C
Christoph Hellwig 已提交
313

314 315 316 317 318 319 320 321 322 323 324 325
		for (j = i+1; j <= be16_to_cpu(block->bb_numrecs); j++) {
			if (root)
				thispa = XFS_BMAP_BROOT_PTR_ADDR(mp, block, j, sz);
			else
				thispa = XFS_BMBT_PTR_ADDR(mp, block, j, dmxr);
			if (*thispa == *pp) {
				xfs_warn(mp, "%s: thispa(%d) == pp(%d) %Ld",
					__func__, j, i,
					(unsigned long long)be64_to_cpu(*thispa));
				panic("%s: ptrs are equal in node\n",
					__func__);
			}
L
Linus Torvalds 已提交
326
		}
327 328
	}
}
L
Linus Torvalds 已提交
329

330 331
/*
 * Check that the extents for the inode ip are in the right order in all
332 333 334 335
 * btree leaves. THis becomes prohibitively expensive for large extent count
 * files, so don't bother with inodes that have more than 10,000 extents in
 * them. The btree record ordering checks will still be done, so for such large
 * bmapbt constructs that is going to catch most corruptions.
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
 */
STATIC void
xfs_bmap_check_leaf_extents(
	xfs_btree_cur_t		*cur,	/* btree cursor or null */
	xfs_inode_t		*ip,		/* incore inode pointer */
	int			whichfork)	/* data or attr fork */
{
	struct xfs_btree_block	*block;	/* current btree block */
	xfs_fsblock_t		bno;	/* block # of "block" */
	xfs_buf_t		*bp;	/* buffer for "block" */
	int			error;	/* error return value */
	xfs_extnum_t		i=0, j;	/* index into the extents list */
	xfs_ifork_t		*ifp;	/* fork structure */
	int			level;	/* btree level, for checking */
	xfs_mount_t		*mp;	/* file system mount structure */
	__be64			*pp;	/* pointer to block address */
	xfs_bmbt_rec_t		*ep;	/* pointer to current extent */
	xfs_bmbt_rec_t		last = {0, 0}; /* last extent in prev block */
	xfs_bmbt_rec_t		*nextp;	/* pointer to next extent */
	int			bp_release = 0;

	if (XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_BTREE) {
		return;
	}

361 362 363 364
	/* skip large extent count inodes */
	if (ip->i_d.di_nextents > 10000)
		return;

365 366 367 368 369 370 371 372 373 374 375 376 377
	bno = NULLFSBLOCK;
	mp = ip->i_mount;
	ifp = XFS_IFORK_PTR(ip, whichfork);
	block = ifp->if_broot;
	/*
	 * Root level must use BMAP_BROOT_PTR_ADDR macro to get ptr out.
	 */
	level = be16_to_cpu(block->bb_level);
	ASSERT(level > 0);
	xfs_check_block(block, mp, 1, ifp->if_broot_bytes);
	pp = XFS_BMAP_BROOT_PTR_ADDR(mp, block, 1, ifp->if_broot_bytes);
	bno = be64_to_cpu(*pp);

C
Christoph Hellwig 已提交
378
	ASSERT(bno != NULLFSBLOCK);
379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394
	ASSERT(XFS_FSB_TO_AGNO(mp, bno) < mp->m_sb.sb_agcount);
	ASSERT(XFS_FSB_TO_AGBNO(mp, bno) < mp->m_sb.sb_agblocks);

	/*
	 * Go down the tree until leaf level is reached, following the first
	 * pointer (leftmost) at each level.
	 */
	while (level-- > 0) {
		/* See if buf is in cur first */
		bp_release = 0;
		bp = xfs_bmap_get_bp(cur, XFS_FSB_TO_DADDR(mp, bno));
		if (!bp) {
			bp_release = 1;
			error = xfs_btree_read_bufl(mp, NULL, bno, 0, &bp,
						XFS_BMAP_BTREE_REF,
						&xfs_bmbt_buf_ops);
395
			if (error)
396
				goto error_norelse;
L
Linus Torvalds 已提交
397
		}
398 399 400
		block = XFS_BUF_TO_BLOCK(bp);
		if (level == 0)
			break;
L
Linus Torvalds 已提交
401 402

		/*
403 404
		 * Check this block for basic sanity (increasing keys and
		 * no duplicate blocks).
L
Linus Torvalds 已提交
405
		 */
C
Christoph Hellwig 已提交
406

407 408 409
		xfs_check_block(block, mp, 0, 0);
		pp = XFS_BMBT_PTR_ADDR(mp, block, 1, mp->m_bmap_dmxr[1]);
		bno = be64_to_cpu(*pp);
410 411
		XFS_WANT_CORRUPTED_GOTO(mp,
					XFS_FSB_SANITY_CHECK(mp, bno), error0);
412 413 414
		if (bp_release) {
			bp_release = 0;
			xfs_trans_brelse(NULL, bp);
L
Linus Torvalds 已提交
415
		}
416
	}
C
Christoph Hellwig 已提交
417

418 419 420 421 422 423 424 425 426 427 428 429 430 431
	/*
	 * Here with bp and block set to the leftmost leaf node in the tree.
	 */
	i = 0;

	/*
	 * Loop over all leaf nodes checking that all extents are in the right order.
	 */
	for (;;) {
		xfs_fsblock_t	nextbno;
		xfs_extnum_t	num_recs;


		num_recs = xfs_btree_get_numrecs(block);
L
Linus Torvalds 已提交
432 433

		/*
434
		 * Read-ahead the next leaf block, if any.
L
Linus Torvalds 已提交
435
		 */
436

437
		nextbno = be64_to_cpu(block->bb_u.l.bb_rightsib);
L
Linus Torvalds 已提交
438 439

		/*
440 441 442
		 * Check all the extents to make sure they are OK.
		 * If we had a previous block, the last entry should
		 * conform with the first entry in this one.
L
Linus Torvalds 已提交
443
		 */
C
Christoph Hellwig 已提交
444

445 446 447 448 449 450 451 452 453 454 455 456 457
		ep = XFS_BMBT_REC_ADDR(mp, block, 1);
		if (i) {
			ASSERT(xfs_bmbt_disk_get_startoff(&last) +
			       xfs_bmbt_disk_get_blockcount(&last) <=
			       xfs_bmbt_disk_get_startoff(ep));
		}
		for (j = 1; j < num_recs; j++) {
			nextp = XFS_BMBT_REC_ADDR(mp, block, j + 1);
			ASSERT(xfs_bmbt_disk_get_startoff(ep) +
			       xfs_bmbt_disk_get_blockcount(ep) <=
			       xfs_bmbt_disk_get_startoff(nextp));
			ep = nextp;
		}
L
Linus Torvalds 已提交
458

459 460 461 462 463 464 465
		last = *ep;
		i += num_recs;
		if (bp_release) {
			bp_release = 0;
			xfs_trans_brelse(NULL, bp);
		}
		bno = nextbno;
L
Linus Torvalds 已提交
466
		/*
467
		 * If we've reached the end, stop.
L
Linus Torvalds 已提交
468
		 */
469 470
		if (bno == NULLFSBLOCK)
			break;
471

472 473 474 475 476 477 478
		bp_release = 0;
		bp = xfs_bmap_get_bp(cur, XFS_FSB_TO_DADDR(mp, bno));
		if (!bp) {
			bp_release = 1;
			error = xfs_btree_read_bufl(mp, NULL, bno, 0, &bp,
						XFS_BMAP_BTREE_REF,
						&xfs_bmbt_buf_ops);
479
			if (error)
480
				goto error_norelse;
L
Linus Torvalds 已提交
481
		}
482
		block = XFS_BUF_TO_BLOCK(bp);
483
	}
484

485
	return;
486

487 488 489 490 491 492 493 494 495
error0:
	xfs_warn(mp, "%s: at error0", __func__);
	if (bp_release)
		xfs_trans_brelse(NULL, bp);
error_norelse:
	xfs_warn(mp, "%s: BAD after btree leaves for %d extents",
		__func__, i);
	panic("%s: CORRUPTED BTREE OR SOMETHING", __func__);
	return;
L
Linus Torvalds 已提交
496 497 498
}

/*
499
 * Add bmap trace insert entries for all the contents of the extent records.
L
Linus Torvalds 已提交
500
 */
501 502 503 504 505 506
void
xfs_bmap_trace_exlist(
	xfs_inode_t	*ip,		/* incore inode pointer */
	xfs_extnum_t	cnt,		/* count of entries in the list */
	int		whichfork,	/* data or attr fork */
	unsigned long	caller_ip)
L
Linus Torvalds 已提交
507
{
508 509 510
	xfs_extnum_t	idx;		/* extent record index */
	xfs_ifork_t	*ifp;		/* inode fork pointer */
	int		state = 0;
511

512 513
	if (whichfork == XFS_ATTR_FORK)
		state |= BMAP_ATTRFORK;
514

515 516 517 518 519
	ifp = XFS_IFORK_PTR(ip, whichfork);
	ASSERT(cnt == (ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t)));
	for (idx = 0; idx < cnt; idx++)
		trace_xfs_extlist(ip, idx, whichfork, caller_ip);
}
520

521 522
/*
 * Validate that the bmbt_irecs being returned from bmapi are valid
523 524
 * given the caller's original parameters.  Specifically check the
 * ranges of the returned irecs to ensure that they only extend beyond
525 526 527 528 529 530 531 532 533 534 535 536
 * the given parameters if the XFS_BMAPI_ENTIRE flag was set.
 */
STATIC void
xfs_bmap_validate_ret(
	xfs_fileoff_t		bno,
	xfs_filblks_t		len,
	int			flags,
	xfs_bmbt_irec_t		*mval,
	int			nmap,
	int			ret_nmap)
{
	int			i;		/* index to map values */
537

538
	ASSERT(ret_nmap <= nmap);
539

540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560
	for (i = 0; i < ret_nmap; i++) {
		ASSERT(mval[i].br_blockcount > 0);
		if (!(flags & XFS_BMAPI_ENTIRE)) {
			ASSERT(mval[i].br_startoff >= bno);
			ASSERT(mval[i].br_blockcount <= len);
			ASSERT(mval[i].br_startoff + mval[i].br_blockcount <=
			       bno + len);
		} else {
			ASSERT(mval[i].br_startoff < bno + len);
			ASSERT(mval[i].br_startoff + mval[i].br_blockcount >
			       bno);
		}
		ASSERT(i == 0 ||
		       mval[i - 1].br_startoff + mval[i - 1].br_blockcount ==
		       mval[i].br_startoff);
		ASSERT(mval[i].br_startblock != DELAYSTARTBLOCK &&
		       mval[i].br_startblock != HOLESTARTBLOCK);
		ASSERT(mval[i].br_state == XFS_EXT_NORM ||
		       mval[i].br_state == XFS_EXT_UNWRITTEN);
	}
}
561

562 563 564 565
#else
#define xfs_bmap_check_leaf_extents(cur, ip, whichfork)		do { } while (0)
#define	xfs_bmap_validate_ret(bno,len,flags,mval,onmap,nmap)
#endif /* DEBUG */
566

567 568 569
/*
 * bmap free list manipulation functions
 */
570

571 572 573 574 575 576
/*
 * Add the extent to the list of extents to be free at transaction end.
 * The list is maintained sorted (by block number).
 */
void
xfs_bmap_add_free(
577 578 579 580 581
	struct xfs_mount		*mp,
	struct xfs_defer_ops		*dfops,
	xfs_fsblock_t			bno,
	xfs_filblks_t			len,
	struct xfs_owner_info		*oinfo)
582
{
583
	struct xfs_extent_free_item	*new;		/* new element */
584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599
#ifdef DEBUG
	xfs_agnumber_t		agno;
	xfs_agblock_t		agbno;

	ASSERT(bno != NULLFSBLOCK);
	ASSERT(len > 0);
	ASSERT(len <= MAXEXTLEN);
	ASSERT(!isnullstartblock(bno));
	agno = XFS_FSB_TO_AGNO(mp, bno);
	agbno = XFS_FSB_TO_AGBNO(mp, bno);
	ASSERT(agno < mp->m_sb.sb_agcount);
	ASSERT(agbno < mp->m_sb.sb_agblocks);
	ASSERT(len < mp->m_sb.sb_agblocks);
	ASSERT(agbno + len <= mp->m_sb.sb_agblocks);
#endif
	ASSERT(xfs_bmap_free_item_zone != NULL);
600

601
	new = kmem_zone_alloc(xfs_bmap_free_item_zone, KM_SLEEP);
602 603
	new->xefi_startblock = bno;
	new->xefi_blockcount = (xfs_extlen_t)len;
604 605 606 607
	if (oinfo)
		new->xefi_oinfo = *oinfo;
	else
		xfs_rmap_skip_owner_update(&new->xefi_oinfo);
608 609
	trace_xfs_bmap_free_defer(mp, XFS_FSB_TO_AGNO(mp, bno), 0,
			XFS_FSB_TO_AGBNO(mp, bno), len);
610
	xfs_defer_add(dfops, XFS_DEFER_OPS_TYPE_FREE, &new->xefi_list);
611
}
C
Christoph Hellwig 已提交
612

613 614 615
/*
 * Inode fork format manipulation functions
 */
L
Linus Torvalds 已提交
616

617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639
/*
 * Transform a btree format file with only one leaf node, where the
 * extents list will fit in the inode, into an extents format file.
 * Since the file extents are already in-core, all we have to do is
 * give up the space for the btree root and pitch the leaf block.
 */
STATIC int				/* error */
xfs_bmap_btree_to_extents(
	xfs_trans_t		*tp,	/* transaction pointer */
	xfs_inode_t		*ip,	/* incore inode pointer */
	xfs_btree_cur_t		*cur,	/* btree cursor */
	int			*logflagsp, /* inode logging flags */
	int			whichfork)  /* data or attr fork */
{
	/* REFERENCED */
	struct xfs_btree_block	*cblock;/* child btree block */
	xfs_fsblock_t		cbno;	/* child block number */
	xfs_buf_t		*cbp;	/* child block's buffer */
	int			error;	/* error return value */
	xfs_ifork_t		*ifp;	/* inode fork data */
	xfs_mount_t		*mp;	/* mount point structure */
	__be64			*pp;	/* ptr to block address */
	struct xfs_btree_block	*rblock;/* root btree block */
640
	struct xfs_owner_info	oinfo;
L
Linus Torvalds 已提交
641

642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663
	mp = ip->i_mount;
	ifp = XFS_IFORK_PTR(ip, whichfork);
	ASSERT(ifp->if_flags & XFS_IFEXTENTS);
	ASSERT(XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_BTREE);
	rblock = ifp->if_broot;
	ASSERT(be16_to_cpu(rblock->bb_level) == 1);
	ASSERT(be16_to_cpu(rblock->bb_numrecs) == 1);
	ASSERT(xfs_bmbt_maxrecs(mp, ifp->if_broot_bytes, 0) == 1);
	pp = XFS_BMAP_BROOT_PTR_ADDR(mp, rblock, 1, ifp->if_broot_bytes);
	cbno = be64_to_cpu(*pp);
	*logflagsp = 0;
#ifdef DEBUG
	if ((error = xfs_btree_check_lptr(cur, cbno, 1)))
		return error;
#endif
	error = xfs_btree_read_bufl(mp, tp, cbno, 0, &cbp, XFS_BMAP_BTREE_REF,
				&xfs_bmbt_buf_ops);
	if (error)
		return error;
	cblock = XFS_BUF_TO_BLOCK(cbp);
	if ((error = xfs_btree_check_block(cur, cblock, 0, cbp)))
		return error;
664 665
	xfs_rmap_ino_bmbt_owner(&oinfo, ip->i_ino, whichfork);
	xfs_bmap_add_free(mp, cur->bc_private.b.dfops, cbno, 1, &oinfo);
666 667 668 669 670 671 672 673 674 675 676 677
	ip->i_d.di_nblocks--;
	xfs_trans_mod_dquot_byino(tp, ip, XFS_TRANS_DQ_BCOUNT, -1L);
	xfs_trans_binval(tp, cbp);
	if (cur->bc_bufs[0] == cbp)
		cur->bc_bufs[0] = NULL;
	xfs_iroot_realloc(ip, -1, whichfork);
	ASSERT(ifp->if_broot == NULL);
	ASSERT((ifp->if_flags & XFS_IFBROOT) == 0);
	XFS_IFORK_FMT_SET(ip, whichfork, XFS_DINODE_FMT_EXTENTS);
	*logflagsp = XFS_ILOG_CORE | xfs_ilog_fext(whichfork);
	return 0;
}
C
Christoph Hellwig 已提交
678

679 680 681 682 683 684 685 686 687
/*
 * Convert an extents-format file into a btree-format file.
 * The new file will have a root block (in the inode) and a single child block.
 */
STATIC int					/* error */
xfs_bmap_extents_to_btree(
	xfs_trans_t		*tp,		/* transaction pointer */
	xfs_inode_t		*ip,		/* incore inode pointer */
	xfs_fsblock_t		*firstblock,	/* first-block-allocated */
688
	struct xfs_defer_ops	*dfops,		/* blocks freed in xaction */
689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707
	xfs_btree_cur_t		**curp,		/* cursor returned to caller */
	int			wasdel,		/* converting a delayed alloc */
	int			*logflagsp,	/* inode logging flags */
	int			whichfork)	/* data or attr fork */
{
	struct xfs_btree_block	*ablock;	/* allocated (child) bt block */
	xfs_buf_t		*abp;		/* buffer for ablock */
	xfs_alloc_arg_t		args;		/* allocation arguments */
	xfs_bmbt_rec_t		*arp;		/* child record pointer */
	struct xfs_btree_block	*block;		/* btree root block */
	xfs_btree_cur_t		*cur;		/* bmap btree cursor */
	xfs_bmbt_rec_host_t	*ep;		/* extent record pointer */
	int			error;		/* error return value */
	xfs_extnum_t		i, cnt;		/* extent record index */
	xfs_ifork_t		*ifp;		/* inode fork pointer */
	xfs_bmbt_key_t		*kp;		/* root block key pointer */
	xfs_mount_t		*mp;		/* mount structure */
	xfs_extnum_t		nextents;	/* number of file extents */
	xfs_bmbt_ptr_t		*pp;		/* root block address pointer */
L
Linus Torvalds 已提交
708

709
	mp = ip->i_mount;
710 711
	ifp = XFS_IFORK_PTR(ip, whichfork);
	ASSERT(XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_EXTENTS);
C
Christoph Hellwig 已提交
712

713 714 715 716 717
	/*
	 * Make space in the inode incore.
	 */
	xfs_iroot_realloc(ip, 1, whichfork);
	ifp->if_flags |= XFS_IFBROOT;
C
Christoph Hellwig 已提交
718

719 720 721 722
	/*
	 * Fill in the root.
	 */
	block = ifp->if_broot;
723 724 725 726 727 728 729 730
	if (xfs_sb_version_hascrc(&mp->m_sb))
		xfs_btree_init_block_int(mp, block, XFS_BUF_DADDR_NULL,
				 XFS_BMAP_CRC_MAGIC, 1, 1, ip->i_ino,
				 XFS_BTREE_LONG_PTRS | XFS_BTREE_CRC_BLOCKS);
	else
		xfs_btree_init_block_int(mp, block, XFS_BUF_DADDR_NULL,
				 XFS_BMAP_MAGIC, 1, 1, ip->i_ino,
				 XFS_BTREE_LONG_PTRS);
C
Christoph Hellwig 已提交
731

732 733 734 735 736
	/*
	 * Need a cursor.  Can't allocate until bb_level is filled in.
	 */
	cur = xfs_bmbt_init_cursor(mp, tp, ip, whichfork);
	cur->bc_private.b.firstblock = *firstblock;
737
	cur->bc_private.b.dfops = dfops;
738 739 740 741 742 743 744 745
	cur->bc_private.b.flags = wasdel ? XFS_BTCUR_BPRV_WASDEL : 0;
	/*
	 * Convert to a btree with two levels, one record in root.
	 */
	XFS_IFORK_FMT_SET(ip, whichfork, XFS_DINODE_FMT_BTREE);
	memset(&args, 0, sizeof(args));
	args.tp = tp;
	args.mp = mp;
746
	xfs_rmap_ino_bmbt_owner(&args.oinfo, ip->i_ino, whichfork);
747 748 749 750
	args.firstblock = *firstblock;
	if (*firstblock == NULLFSBLOCK) {
		args.type = XFS_ALLOCTYPE_START_BNO;
		args.fsbno = XFS_INO_TO_FSB(mp, ip->i_ino);
751
	} else if (dfops->dop_low) {
752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771
		args.type = XFS_ALLOCTYPE_START_BNO;
		args.fsbno = *firstblock;
	} else {
		args.type = XFS_ALLOCTYPE_NEAR_BNO;
		args.fsbno = *firstblock;
	}
	args.minlen = args.maxlen = args.prod = 1;
	args.wasdel = wasdel;
	*logflagsp = 0;
	if ((error = xfs_alloc_vextent(&args))) {
		xfs_iroot_realloc(ip, -1, whichfork);
		xfs_btree_del_cursor(cur, XFS_BTREE_ERROR);
		return error;
	}
	/*
	 * Allocation can't fail, the space was reserved.
	 */
	ASSERT(args.fsbno != NULLFSBLOCK);
	ASSERT(*firstblock == NULLFSBLOCK ||
	       args.agno == XFS_FSB_TO_AGNO(mp, *firstblock) ||
772
	       (dfops->dop_low &&
773 774 775 776 777 778 779 780 781 782 783
		args.agno > XFS_FSB_TO_AGNO(mp, *firstblock)));
	*firstblock = cur->bc_private.b.firstblock = args.fsbno;
	cur->bc_private.b.allocated++;
	ip->i_d.di_nblocks++;
	xfs_trans_mod_dquot_byino(tp, ip, XFS_TRANS_DQ_BCOUNT, 1L);
	abp = xfs_btree_get_bufl(mp, tp, args.fsbno, 0);
	/*
	 * Fill in the child block.
	 */
	abp->b_ops = &xfs_bmbt_buf_ops;
	ablock = XFS_BUF_TO_BLOCK(abp);
784 785 786 787 788 789 790 791 792
	if (xfs_sb_version_hascrc(&mp->m_sb))
		xfs_btree_init_block_int(mp, ablock, abp->b_bn,
				XFS_BMAP_CRC_MAGIC, 0, 0, ip->i_ino,
				XFS_BTREE_LONG_PTRS | XFS_BTREE_CRC_BLOCKS);
	else
		xfs_btree_init_block_int(mp, ablock, abp->b_bn,
				XFS_BMAP_MAGIC, 0, 0, ip->i_ino,
				XFS_BTREE_LONG_PTRS);

793 794 795 796 797 798 799 800
	arp = XFS_BMBT_REC_ADDR(mp, ablock, 1);
	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
	for (cnt = i = 0; i < nextents; i++) {
		ep = xfs_iext_get_ext(ifp, i);
		if (!isnullstartblock(xfs_bmbt_get_startblock(ep))) {
			arp->l0 = cpu_to_be64(ep->l0);
			arp->l1 = cpu_to_be64(ep->l1);
			arp++; cnt++;
L
Linus Torvalds 已提交
801
		}
802 803 804
	}
	ASSERT(cnt == XFS_IFORK_NEXTENTS(ip, whichfork));
	xfs_btree_set_numrecs(ablock, cnt);
L
Linus Torvalds 已提交
805

806 807 808 809 810 811 812 813 814
	/*
	 * Fill in the root key and pointer.
	 */
	kp = XFS_BMBT_KEY_ADDR(mp, block, 1);
	arp = XFS_BMBT_REC_ADDR(mp, ablock, 1);
	kp->br_startoff = cpu_to_be64(xfs_bmbt_disk_get_startoff(arp));
	pp = XFS_BMBT_PTR_ADDR(mp, block, 1, xfs_bmbt_get_maxrecs(cur,
						be16_to_cpu(block->bb_level)));
	*pp = cpu_to_be64(args.fsbno);
C
Christoph Hellwig 已提交
815

816 817 818 819 820 821 822 823 824 825 826
	/*
	 * Do all this logging at the end so that
	 * the root is at the right level.
	 */
	xfs_btree_log_block(cur, abp, XFS_BB_ALL_BITS);
	xfs_btree_log_recs(cur, abp, 1, be16_to_cpu(ablock->bb_numrecs));
	ASSERT(*curp == NULL);
	*curp = cur;
	*logflagsp = XFS_ILOG_CORE | xfs_ilog_fbroot(whichfork);
	return 0;
}
C
Christoph Hellwig 已提交
827

828 829 830 831 832 833
/*
 * Convert a local file to an extents file.
 * This code is out of bounds for data forks of regular files,
 * since the file data needs to get logged so things will stay consistent.
 * (The bmap-level manipulations are ok, though).
 */
834 835 836 837 838 839 840 841 842 843 844
void
xfs_bmap_local_to_extents_empty(
	struct xfs_inode	*ip,
	int			whichfork)
{
	struct xfs_ifork	*ifp = XFS_IFORK_PTR(ip, whichfork);

	ASSERT(XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_LOCAL);
	ASSERT(ifp->if_bytes == 0);
	ASSERT(XFS_IFORK_NEXTENTS(ip, whichfork) == 0);

845
	xfs_bmap_forkoff_reset(ip, whichfork);
846 847 848 849 850 851
	ifp->if_flags &= ~XFS_IFINLINE;
	ifp->if_flags |= XFS_IFEXTENTS;
	XFS_IFORK_FMT_SET(ip, whichfork, XFS_DINODE_FMT_EXTENTS);
}


852 853 854 855 856 857 858 859
STATIC int				/* error */
xfs_bmap_local_to_extents(
	xfs_trans_t	*tp,		/* transaction pointer */
	xfs_inode_t	*ip,		/* incore inode pointer */
	xfs_fsblock_t	*firstblock,	/* first block allocated in xaction */
	xfs_extlen_t	total,		/* total blocks needed by transaction */
	int		*logflagsp,	/* inode logging flags */
	int		whichfork,
860 861
	void		(*init_fn)(struct xfs_trans *tp,
				   struct xfs_buf *bp,
862 863 864
				   struct xfs_inode *ip,
				   struct xfs_ifork *ifp))
{
865
	int		error = 0;
866 867
	int		flags;		/* logging flags returned */
	xfs_ifork_t	*ifp;		/* inode fork pointer */
868 869 870
	xfs_alloc_arg_t	args;		/* allocation arguments */
	xfs_buf_t	*bp;		/* buffer for extent block */
	xfs_bmbt_rec_host_t *ep;	/* extent record pointer */
C
Christoph Hellwig 已提交
871

872 873 874 875
	/*
	 * We don't want to deal with the case of keeping inode data inline yet.
	 * So sending the data fork of a regular inode is invalid.
	 */
D
Dave Chinner 已提交
876
	ASSERT(!(S_ISREG(VFS_I(ip)->i_mode) && whichfork == XFS_DATA_FORK));
877 878
	ifp = XFS_IFORK_PTR(ip, whichfork);
	ASSERT(XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_LOCAL);
879 880 881 882 883 884 885

	if (!ifp->if_bytes) {
		xfs_bmap_local_to_extents_empty(ip, whichfork);
		flags = XFS_ILOG_CORE;
		goto done;
	}

886 887
	flags = 0;
	error = 0;
888 889 890 891 892
	ASSERT((ifp->if_flags & (XFS_IFINLINE|XFS_IFEXTENTS|XFS_IFEXTIREC)) ==
								XFS_IFINLINE);
	memset(&args, 0, sizeof(args));
	args.tp = tp;
	args.mp = ip->i_mount;
893
	xfs_rmap_ino_owner(&args.oinfo, ip->i_ino, whichfork, 0);
894 895 896 897 898 899 900 901
	args.firstblock = *firstblock;
	/*
	 * Allocate a block.  We know we need only one, since the
	 * file currently fits in an inode.
	 */
	if (*firstblock == NULLFSBLOCK) {
		args.fsbno = XFS_INO_TO_FSB(args.mp, ip->i_ino);
		args.type = XFS_ALLOCTYPE_START_BNO;
902
	} else {
903 904
		args.fsbno = *firstblock;
		args.type = XFS_ALLOCTYPE_NEAR_BNO;
905
	}
906 907 908 909 910 911 912 913 914 915 916 917
	args.total = total;
	args.minlen = args.maxlen = args.prod = 1;
	error = xfs_alloc_vextent(&args);
	if (error)
		goto done;

	/* Can't fail, the space was reserved. */
	ASSERT(args.fsbno != NULLFSBLOCK);
	ASSERT(args.len == 1);
	*firstblock = args.fsbno;
	bp = xfs_btree_get_bufl(args.mp, tp, args.fsbno, 0);

918
	/*
919
	 * Initialize the block, copy the data and log the remote buffer.
920
	 *
921 922 923 924
	 * The callout is responsible for logging because the remote format
	 * might differ from the local format and thus we don't know how much to
	 * log here. Note that init_fn must also set the buffer log item type
	 * correctly.
925
	 */
926 927
	init_fn(tp, bp, ip, ifp);

928
	/* account for the change in fork size */
929 930
	xfs_idata_realloc(ip, -ifp->if_bytes, whichfork);
	xfs_bmap_local_to_extents_empty(ip, whichfork);
931
	flags |= XFS_ILOG_CORE;
932 933 934 935 936 937 938 939 940 941 942 943 944

	xfs_iext_add(ifp, 0, 1);
	ep = xfs_iext_get_ext(ifp, 0);
	xfs_bmbt_set_allf(ep, 0, args.fsbno, 1, XFS_EXT_NORM);
	trace_xfs_bmap_post_update(ip, 0,
			whichfork == XFS_ATTR_FORK ? BMAP_ATTRFORK : 0,
			_THIS_IP_);
	XFS_IFORK_NEXT_SET(ip, whichfork, 1);
	ip->i_d.di_nblocks = 1;
	xfs_trans_mod_dquot_byino(tp, ip,
		XFS_TRANS_DQ_BCOUNT, 1L);
	flags |= xfs_ilog_fext(whichfork);

945 946 947 948
done:
	*logflagsp = flags;
	return error;
}
C
Christoph Hellwig 已提交
949

950 951 952 953 954 955 956 957
/*
 * Called from xfs_bmap_add_attrfork to handle btree format files.
 */
STATIC int					/* error */
xfs_bmap_add_attrfork_btree(
	xfs_trans_t		*tp,		/* transaction pointer */
	xfs_inode_t		*ip,		/* incore inode pointer */
	xfs_fsblock_t		*firstblock,	/* first block allocated */
958
	struct xfs_defer_ops	*dfops,		/* blocks to free at commit */
959 960 961 962 963 964
	int			*flags)		/* inode logging flags */
{
	xfs_btree_cur_t		*cur;		/* btree cursor */
	int			error;		/* error return value */
	xfs_mount_t		*mp;		/* file system mount struct */
	int			stat;		/* newroot status */
C
Christoph Hellwig 已提交
965

966 967 968 969 970
	mp = ip->i_mount;
	if (ip->i_df.if_broot_bytes <= XFS_IFORK_DSIZE(ip))
		*flags |= XFS_ILOG_DBROOT;
	else {
		cur = xfs_bmbt_init_cursor(mp, tp, ip, XFS_DATA_FORK);
971
		cur->bc_private.b.dfops = dfops;
972 973 974 975
		cur->bc_private.b.firstblock = *firstblock;
		if ((error = xfs_bmbt_lookup_ge(cur, 0, 0, 0, &stat)))
			goto error0;
		/* must be at least one entry */
976
		XFS_WANT_CORRUPTED_GOTO(mp, stat == 1, error0);
977 978 979 980
		if ((error = xfs_btree_new_iroot(cur, flags, &stat)))
			goto error0;
		if (stat == 0) {
			xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR);
D
Dave Chinner 已提交
981
			return -ENOSPC;
L
Linus Torvalds 已提交
982
		}
983 984 985
		*firstblock = cur->bc_private.b.firstblock;
		cur->bc_private.b.allocated = 0;
		xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR);
L
Linus Torvalds 已提交
986
	}
987 988 989 990 991
	return 0;
error0:
	xfs_btree_del_cursor(cur, XFS_BTREE_ERROR);
	return error;
}
992

993 994 995 996 997 998 999 1000
/*
 * Called from xfs_bmap_add_attrfork to handle extents format files.
 */
STATIC int					/* error */
xfs_bmap_add_attrfork_extents(
	xfs_trans_t		*tp,		/* transaction pointer */
	xfs_inode_t		*ip,		/* incore inode pointer */
	xfs_fsblock_t		*firstblock,	/* first block allocated */
1001
	struct xfs_defer_ops	*dfops,		/* blocks to free at commit */
1002 1003 1004 1005
	int			*flags)		/* inode logging flags */
{
	xfs_btree_cur_t		*cur;		/* bmap btree cursor */
	int			error;		/* error return value */
1006

1007 1008 1009
	if (ip->i_d.di_nextents * sizeof(xfs_bmbt_rec_t) <= XFS_IFORK_DSIZE(ip))
		return 0;
	cur = NULL;
1010
	error = xfs_bmap_extents_to_btree(tp, ip, firstblock, dfops, &cur, 0,
1011
		flags, XFS_DATA_FORK);
1012 1013
	if (cur) {
		cur->bc_private.b.allocated = 0;
1014 1015
		xfs_btree_del_cursor(cur,
			error ? XFS_BTREE_ERROR : XFS_BTREE_NOERROR);
1016
	}
L
Linus Torvalds 已提交
1017 1018 1019
	return error;
}

1020 1021 1022 1023 1024 1025 1026 1027 1028
/*
 * Called from xfs_bmap_add_attrfork to handle local format files. Each
 * different data fork content type needs a different callout to do the
 * conversion. Some are basic and only require special block initialisation
 * callouts for the data formating, others (directories) are so specialised they
 * handle everything themselves.
 *
 * XXX (dgc): investigate whether directory conversion can use the generic
 * formatting callout. It should be possible - it's just a very complex
1029
 * formatter.
1030 1031 1032 1033 1034 1035
 */
STATIC int					/* error */
xfs_bmap_add_attrfork_local(
	xfs_trans_t		*tp,		/* transaction pointer */
	xfs_inode_t		*ip,		/* incore inode pointer */
	xfs_fsblock_t		*firstblock,	/* first block allocated */
1036
	struct xfs_defer_ops	*dfops,		/* blocks to free at commit */
1037 1038 1039
	int			*flags)		/* inode logging flags */
{
	xfs_da_args_t		dargs;		/* args for dir/attr code */
1040

1041 1042
	if (ip->i_df.if_bytes <= XFS_IFORK_DSIZE(ip))
		return 0;
1043

D
Dave Chinner 已提交
1044
	if (S_ISDIR(VFS_I(ip)->i_mode)) {
1045
		memset(&dargs, 0, sizeof(dargs));
1046
		dargs.geo = ip->i_mount->m_dir_geo;
1047 1048
		dargs.dp = ip;
		dargs.firstblock = firstblock;
1049
		dargs.dfops = dfops;
1050
		dargs.total = dargs.geo->fsbcount;
1051 1052 1053
		dargs.whichfork = XFS_DATA_FORK;
		dargs.trans = tp;
		return xfs_dir2_sf_to_block(&dargs);
L
Linus Torvalds 已提交
1054
	}
1055

D
Dave Chinner 已提交
1056
	if (S_ISLNK(VFS_I(ip)->i_mode))
1057 1058 1059
		return xfs_bmap_local_to_extents(tp, ip, firstblock, 1,
						 flags, XFS_DATA_FORK,
						 xfs_symlink_local_to_remote);
1060

1061 1062
	/* should only be called for types that support local format data */
	ASSERT(0);
D
Dave Chinner 已提交
1063
	return -EFSCORRUPTED;
1064
}
C
Christoph Hellwig 已提交
1065

1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076
/*
 * Convert inode from non-attributed to attributed.
 * Must not be in a transaction, ip must not be locked.
 */
int						/* error code */
xfs_bmap_add_attrfork(
	xfs_inode_t		*ip,		/* incore inode pointer */
	int			size,		/* space new attribute needs */
	int			rsvd)		/* xact may use reserved blks */
{
	xfs_fsblock_t		firstblock;	/* 1st block/ag allocated */
1077
	struct xfs_defer_ops	dfops;		/* freed extent records */
1078 1079 1080 1081 1082 1083
	xfs_mount_t		*mp;		/* mount structure */
	xfs_trans_t		*tp;		/* transaction pointer */
	int			blks;		/* space reservation */
	int			version = 1;	/* superblock attr version */
	int			logflags;	/* logging flags */
	int			error;		/* error return value */
C
Christoph Hellwig 已提交
1084

1085
	ASSERT(XFS_IFORK_Q(ip) == 0);
L
Linus Torvalds 已提交
1086

1087 1088
	mp = ip->i_mount;
	ASSERT(!XFS_NOT_DQATTACHED(mp, ip));
1089

1090
	blks = XFS_ADDAFORK_SPACE_RES(mp);
1091 1092 1093 1094

	error = xfs_trans_alloc(mp, &M_RES(mp)->tr_addafork, blks, 0,
			rsvd ? XFS_TRANS_RESERVE : 0, &tp);
	if (error)
1095
		return error;
1096

1097 1098 1099 1100
	xfs_ilock(ip, XFS_ILOCK_EXCL);
	error = xfs_trans_reserve_quota_nblks(tp, ip, blks, 0, rsvd ?
			XFS_QMOPT_RES_REGBLKS | XFS_QMOPT_FORCE_RES :
			XFS_QMOPT_RES_REGBLKS);
1101 1102
	if (error)
		goto trans_cancel;
1103
	if (XFS_IFORK_Q(ip))
1104
		goto trans_cancel;
1105
	if (ip->i_d.di_aformat != XFS_DINODE_FMT_EXTENTS) {
L
Linus Torvalds 已提交
1106
		/*
1107
		 * For inodes coming from pre-6.2 filesystems.
L
Linus Torvalds 已提交
1108
		 */
1109 1110 1111 1112
		ASSERT(ip->i_d.di_aformat == 0);
		ip->i_d.di_aformat = XFS_DINODE_FMT_EXTENTS;
	}
	ASSERT(ip->i_d.di_anextents == 0);
C
Christoph Hellwig 已提交
1113

1114
	xfs_trans_ijoin(tp, ip, 0);
1115
	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
L
Linus Torvalds 已提交
1116

1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131
	switch (ip->i_d.di_format) {
	case XFS_DINODE_FMT_DEV:
		ip->i_d.di_forkoff = roundup(sizeof(xfs_dev_t), 8) >> 3;
		break;
	case XFS_DINODE_FMT_UUID:
		ip->i_d.di_forkoff = roundup(sizeof(uuid_t), 8) >> 3;
		break;
	case XFS_DINODE_FMT_LOCAL:
	case XFS_DINODE_FMT_EXTENTS:
	case XFS_DINODE_FMT_BTREE:
		ip->i_d.di_forkoff = xfs_attr_shortform_bytesfit(ip, size);
		if (!ip->i_d.di_forkoff)
			ip->i_d.di_forkoff = xfs_default_attroffset(ip) >> 3;
		else if (mp->m_flags & XFS_MOUNT_ATTR2)
			version = 2;
L
Linus Torvalds 已提交
1132
		break;
1133 1134
	default:
		ASSERT(0);
D
Dave Chinner 已提交
1135
		error = -EINVAL;
1136
		goto trans_cancel;
1137
	}
L
Linus Torvalds 已提交
1138

1139 1140 1141 1142
	ASSERT(ip->i_afp == NULL);
	ip->i_afp = kmem_zone_zalloc(xfs_ifork_zone, KM_SLEEP);
	ip->i_afp->if_flags = XFS_IFEXTENTS;
	logflags = 0;
1143
	xfs_defer_init(&dfops, &firstblock);
1144 1145
	switch (ip->i_d.di_format) {
	case XFS_DINODE_FMT_LOCAL:
1146
		error = xfs_bmap_add_attrfork_local(tp, ip, &firstblock, &dfops,
1147 1148 1149 1150
			&logflags);
		break;
	case XFS_DINODE_FMT_EXTENTS:
		error = xfs_bmap_add_attrfork_extents(tp, ip, &firstblock,
1151
			&dfops, &logflags);
1152 1153
		break;
	case XFS_DINODE_FMT_BTREE:
1154
		error = xfs_bmap_add_attrfork_btree(tp, ip, &firstblock, &dfops,
1155 1156 1157 1158
			&logflags);
		break;
	default:
		error = 0;
L
Linus Torvalds 已提交
1159 1160
		break;
	}
1161 1162 1163
	if (logflags)
		xfs_trans_log_inode(tp, ip, logflags);
	if (error)
1164
		goto bmap_cancel;
1165 1166
	if (!xfs_sb_version_hasattr(&mp->m_sb) ||
	   (!xfs_sb_version_hasattr2(&mp->m_sb) && version == 2)) {
1167
		bool log_sb = false;
1168 1169 1170 1171

		spin_lock(&mp->m_sb_lock);
		if (!xfs_sb_version_hasattr(&mp->m_sb)) {
			xfs_sb_version_addattr(&mp->m_sb);
1172
			log_sb = true;
1173 1174 1175
		}
		if (!xfs_sb_version_hasattr2(&mp->m_sb) && version == 2) {
			xfs_sb_version_addattr2(&mp->m_sb);
1176
			log_sb = true;
1177
		}
1178
		spin_unlock(&mp->m_sb_lock);
1179 1180
		if (log_sb)
			xfs_log_sb(tp);
L
Linus Torvalds 已提交
1181
	}
1182

1183
	error = xfs_defer_finish(&tp, &dfops, NULL);
1184
	if (error)
1185
		goto bmap_cancel;
1186
	error = xfs_trans_commit(tp);
1187 1188 1189 1190
	xfs_iunlock(ip, XFS_ILOCK_EXCL);
	return error;

bmap_cancel:
1191
	xfs_defer_cancel(&dfops);
1192
trans_cancel:
1193
	xfs_trans_cancel(tp);
1194 1195
	xfs_iunlock(ip, XFS_ILOCK_EXCL);
	return error;
L
Linus Torvalds 已提交
1196 1197 1198
}

/*
1199
 * Internal and external extent tree search functions.
L
Linus Torvalds 已提交
1200
 */
1201

1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225
/*
 * Read in the extents to if_extents.
 * All inode fields are set up by caller, we just traverse the btree
 * and copy the records in. If the file system cannot contain unwritten
 * extents, the records are checked for no "state" flags.
 */
int					/* error */
xfs_bmap_read_extents(
	xfs_trans_t		*tp,	/* transaction pointer */
	xfs_inode_t		*ip,	/* incore inode */
	int			whichfork) /* data or attr fork */
{
	struct xfs_btree_block	*block;	/* current btree block */
	xfs_fsblock_t		bno;	/* block # of "block" */
	xfs_buf_t		*bp;	/* buffer for "block" */
	int			error;	/* error return value */
	xfs_exntfmt_t		exntf;	/* XFS_EXTFMT_NOSTATE, if checking */
	xfs_extnum_t		i, j;	/* index into the extents list */
	xfs_ifork_t		*ifp;	/* fork structure */
	int			level;	/* btree level, for checking */
	xfs_mount_t		*mp;	/* file system mount structure */
	__be64			*pp;	/* pointer to block address */
	/* REFERENCED */
	xfs_extnum_t		room;	/* number of entries there's room for */
1226

1227 1228 1229 1230 1231 1232
	bno = NULLFSBLOCK;
	mp = ip->i_mount;
	ifp = XFS_IFORK_PTR(ip, whichfork);
	exntf = (whichfork != XFS_DATA_FORK) ? XFS_EXTFMT_NOSTATE :
					XFS_EXTFMT_INODE(ip);
	block = ifp->if_broot;
L
Linus Torvalds 已提交
1233
	/*
1234
	 * Root level must use BMAP_BROOT_PTR_ADDR macro to get ptr out.
L
Linus Torvalds 已提交
1235
	 */
1236 1237 1238 1239
	level = be16_to_cpu(block->bb_level);
	ASSERT(level > 0);
	pp = XFS_BMAP_BROOT_PTR_ADDR(mp, block, 1, ifp->if_broot_bytes);
	bno = be64_to_cpu(*pp);
C
Christoph Hellwig 已提交
1240
	ASSERT(bno != NULLFSBLOCK);
1241 1242
	ASSERT(XFS_FSB_TO_AGNO(mp, bno) < mp->m_sb.sb_agcount);
	ASSERT(XFS_FSB_TO_AGBNO(mp, bno) < mp->m_sb.sb_agblocks);
L
Linus Torvalds 已提交
1243
	/*
1244 1245
	 * Go down the tree until leaf level is reached, following the first
	 * pointer (leftmost) at each level.
L
Linus Torvalds 已提交
1246
	 */
1247 1248 1249 1250 1251 1252 1253 1254 1255 1256
	while (level-- > 0) {
		error = xfs_btree_read_bufl(mp, tp, bno, 0, &bp,
				XFS_BMAP_BTREE_REF, &xfs_bmbt_buf_ops);
		if (error)
			return error;
		block = XFS_BUF_TO_BLOCK(bp);
		if (level == 0)
			break;
		pp = XFS_BMBT_PTR_ADDR(mp, block, 1, mp->m_bmap_dmxr[1]);
		bno = be64_to_cpu(*pp);
1257 1258
		XFS_WANT_CORRUPTED_GOTO(mp,
			XFS_FSB_SANITY_CHECK(mp, bno), error0);
1259
		xfs_trans_brelse(tp, bp);
L
Linus Torvalds 已提交
1260 1261
	}
	/*
1262
	 * Here with bp and block set to the leftmost leaf node in the tree.
L
Linus Torvalds 已提交
1263
	 */
1264 1265
	room = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
	i = 0;
L
Linus Torvalds 已提交
1266
	/*
1267
	 * Loop over all leaf nodes.  Copy information to the extent records.
L
Linus Torvalds 已提交
1268
	 */
1269 1270 1271 1272 1273
	for (;;) {
		xfs_bmbt_rec_t	*frp;
		xfs_fsblock_t	nextbno;
		xfs_extnum_t	num_recs;
		xfs_extnum_t	start;
C
Christoph Hellwig 已提交
1274

1275 1276 1277 1278 1279 1280 1281 1282 1283
		num_recs = xfs_btree_get_numrecs(block);
		if (unlikely(i + num_recs > room)) {
			ASSERT(i + num_recs <= room);
			xfs_warn(ip->i_mount,
				"corrupt dinode %Lu, (btree extents).",
				(unsigned long long) ip->i_ino);
			XFS_CORRUPTION_ERROR("xfs_bmap_read_extents(1)",
				XFS_ERRLEVEL_LOW, ip->i_mount, block);
			goto error0;
L
Linus Torvalds 已提交
1284 1285
		}
		/*
1286
		 * Read-ahead the next leaf block, if any.
L
Linus Torvalds 已提交
1287
		 */
1288 1289 1290 1291
		nextbno = be64_to_cpu(block->bb_u.l.bb_rightsib);
		if (nextbno != NULLFSBLOCK)
			xfs_btree_reada_bufl(mp, nextbno, 1,
					     &xfs_bmbt_buf_ops);
L
Linus Torvalds 已提交
1292
		/*
1293
		 * Copy records into the extent records.
L
Linus Torvalds 已提交
1294
		 */
1295 1296 1297 1298 1299 1300
		frp = XFS_BMBT_REC_ADDR(mp, block, 1);
		start = i;
		for (j = 0; j < num_recs; j++, i++, frp++) {
			xfs_bmbt_rec_host_t *trp = xfs_iext_get_ext(ifp, i);
			trp->l0 = be64_to_cpu(frp->l0);
			trp->l1 = be64_to_cpu(frp->l1);
L
Linus Torvalds 已提交
1301
		}
1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317
		if (exntf == XFS_EXTFMT_NOSTATE) {
			/*
			 * Check all attribute bmap btree records and
			 * any "older" data bmap btree records for a
			 * set bit in the "extent flag" position.
			 */
			if (unlikely(xfs_check_nostate_extents(ifp,
					start, num_recs))) {
				XFS_ERROR_REPORT("xfs_bmap_read_extents(2)",
						 XFS_ERRLEVEL_LOW,
						 ip->i_mount);
				goto error0;
			}
		}
		xfs_trans_brelse(tp, bp);
		bno = nextbno;
L
Linus Torvalds 已提交
1318
		/*
1319
		 * If we've reached the end, stop.
L
Linus Torvalds 已提交
1320
		 */
1321 1322 1323 1324 1325 1326 1327
		if (bno == NULLFSBLOCK)
			break;
		error = xfs_btree_read_bufl(mp, tp, bno, 0, &bp,
				XFS_BMAP_BTREE_REF, &xfs_bmbt_buf_ops);
		if (error)
			return error;
		block = XFS_BUF_TO_BLOCK(bp);
L
Linus Torvalds 已提交
1328
	}
1329 1330 1331 1332 1333 1334
	ASSERT(i == (ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t)));
	ASSERT(i == XFS_IFORK_NEXTENTS(ip, whichfork));
	XFS_BMAP_TRACE_EXLIST(ip, i, whichfork);
	return 0;
error0:
	xfs_trans_brelse(tp, bp);
D
Dave Chinner 已提交
1335
	return -EFSCORRUPTED;
1336
}
1337 1338


1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356
/*
 * Search the extent records for the entry containing block bno.
 * If bno lies in a hole, point to the next entry.  If bno lies
 * past eof, *eofp will be set, and *prevp will contain the last
 * entry (null if none).  Else, *lastxp will be set to the index
 * of the found entry; *gotp will contain the entry.
 */
STATIC xfs_bmbt_rec_host_t *		/* pointer to found extent entry */
xfs_bmap_search_multi_extents(
	xfs_ifork_t	*ifp,		/* inode fork pointer */
	xfs_fileoff_t	bno,		/* block number searched for */
	int		*eofp,		/* out: end of file found */
	xfs_extnum_t	*lastxp,	/* out: last extent index */
	xfs_bmbt_irec_t	*gotp,		/* out: extent entry found */
	xfs_bmbt_irec_t	*prevp)		/* out: previous extent entry found */
{
	xfs_bmbt_rec_host_t *ep;		/* extent record pointer */
	xfs_extnum_t	lastx;		/* last extent index */
1357

1358 1359 1360 1361 1362 1363 1364 1365 1366
	/*
	 * Initialize the extent entry structure to catch access to
	 * uninitialized br_startblock field.
	 */
	gotp->br_startoff = 0xffa5a5a5a5a5a5a5LL;
	gotp->br_blockcount = 0xa55a5a5a5a5a5a5aLL;
	gotp->br_state = XFS_EXT_INVALID;
	gotp->br_startblock = 0xffffa5a5a5a5a5a5LL;
	prevp->br_startoff = NULLFILEOFF;
1367

1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383
	ep = xfs_iext_bno_to_ext(ifp, bno, &lastx);
	if (lastx > 0) {
		xfs_bmbt_get_all(xfs_iext_get_ext(ifp, lastx - 1), prevp);
	}
	if (lastx < (ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t))) {
		xfs_bmbt_get_all(ep, gotp);
		*eofp = 0;
	} else {
		if (lastx > 0) {
			*gotp = *prevp;
		}
		*eofp = 1;
		ep = NULL;
	}
	*lastxp = lastx;
	return ep;
L
Linus Torvalds 已提交
1384 1385
}

1386
/*
1387 1388 1389 1390 1391
 * Search the extents list for the inode, for the extent containing bno.
 * If bno lies in a hole, point to the next entry.  If bno lies past eof,
 * *eofp will be set, and *prevp will contain the last entry (null if none).
 * Else, *lastxp will be set to the index of the found
 * entry; *gotp will contain the entry.
1392
 */
1393
xfs_bmbt_rec_host_t *                 /* pointer to found extent entry */
1394 1395 1396 1397 1398 1399 1400 1401
xfs_bmap_search_extents(
	xfs_inode_t     *ip,            /* incore inode pointer */
	xfs_fileoff_t   bno,            /* block number searched for */
	int             fork,      	/* data or attr fork */
	int             *eofp,          /* out: end of file found */
	xfs_extnum_t    *lastxp,        /* out: last extent index */
	xfs_bmbt_irec_t *gotp,          /* out: extent entry found */
	xfs_bmbt_irec_t *prevp)         /* out: previous extent entry found */
1402
{
1403 1404
	xfs_ifork_t	*ifp;		/* inode fork pointer */
	xfs_bmbt_rec_host_t  *ep;            /* extent record pointer */
1405

1406
	XFS_STATS_INC(ip->i_mount, xs_look_exlist);
1407
	ifp = XFS_IFORK_PTR(ip, fork);
1408

1409
	ep = xfs_bmap_search_multi_extents(ifp, bno, eofp, lastxp, gotp, prevp);
1410

1411 1412 1413 1414 1415
	if (unlikely(!(gotp->br_startblock) && (*lastxp != NULLEXTNUM) &&
		     !(XFS_IS_REALTIME_INODE(ip) && fork == XFS_DATA_FORK))) {
		xfs_alert_tag(ip->i_mount, XFS_PTAG_FSBLOCK_ZERO,
				"Access to block zero in inode %llu "
				"start_block: %llx start_off: %llx "
1416
				"blkcnt: %llx extent-state: %x lastx: %x",
1417 1418 1419 1420 1421 1422 1423 1424
			(unsigned long long)ip->i_ino,
			(unsigned long long)gotp->br_startblock,
			(unsigned long long)gotp->br_startoff,
			(unsigned long long)gotp->br_blockcount,
			gotp->br_state, *lastxp);
		*lastxp = NULLEXTNUM;
		*eofp = 1;
		return NULL;
1425
	}
1426 1427
	return ep;
}
1428

1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458
/*
 * Returns the file-relative block number of the first unused block(s)
 * in the file with at least "len" logically contiguous blocks free.
 * This is the lowest-address hole if the file has holes, else the first block
 * past the end of file.
 * Return 0 if the file is currently local (in-inode).
 */
int						/* error */
xfs_bmap_first_unused(
	xfs_trans_t	*tp,			/* transaction pointer */
	xfs_inode_t	*ip,			/* incore inode */
	xfs_extlen_t	len,			/* size of hole to find */
	xfs_fileoff_t	*first_unused,		/* unused block */
	int		whichfork)		/* data or attr fork */
{
	int		error;			/* error return value */
	int		idx;			/* extent record index */
	xfs_ifork_t	*ifp;			/* inode fork pointer */
	xfs_fileoff_t	lastaddr;		/* last block number seen */
	xfs_fileoff_t	lowest;			/* lowest useful block */
	xfs_fileoff_t	max;			/* starting useful block */
	xfs_fileoff_t	off;			/* offset for this block */
	xfs_extnum_t	nextents;		/* number of extent entries */

	ASSERT(XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_BTREE ||
	       XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_EXTENTS ||
	       XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_LOCAL);
	if (XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_LOCAL) {
		*first_unused = 0;
		return 0;
1459
	}
1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477
	ifp = XFS_IFORK_PTR(ip, whichfork);
	if (!(ifp->if_flags & XFS_IFEXTENTS) &&
	    (error = xfs_iread_extents(tp, ip, whichfork)))
		return error;
	lowest = *first_unused;
	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
	for (idx = 0, lastaddr = 0, max = lowest; idx < nextents; idx++) {
		xfs_bmbt_rec_host_t *ep = xfs_iext_get_ext(ifp, idx);
		off = xfs_bmbt_get_startoff(ep);
		/*
		 * See if the hole before this extent will work.
		 */
		if (off >= lowest + len && off - max >= len) {
			*first_unused = max;
			return 0;
		}
		lastaddr = off + xfs_bmbt_get_blockcount(ep);
		max = XFS_FILEOFF_MAX(lastaddr, lowest);
1478
	}
1479 1480 1481 1482 1483
	*first_unused = max;
	return 0;
}

/*
1484
 * Returns the file-relative block number of the last block - 1 before
1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507
 * last_block (input value) in the file.
 * This is not based on i_size, it is based on the extent records.
 * Returns 0 for local files, as they do not have extent records.
 */
int						/* error */
xfs_bmap_last_before(
	xfs_trans_t	*tp,			/* transaction pointer */
	xfs_inode_t	*ip,			/* incore inode */
	xfs_fileoff_t	*last_block,		/* last block */
	int		whichfork)		/* data or attr fork */
{
	xfs_fileoff_t	bno;			/* input file offset */
	int		eof;			/* hit end of file */
	xfs_bmbt_rec_host_t *ep;		/* pointer to last extent */
	int		error;			/* error return value */
	xfs_bmbt_irec_t	got;			/* current extent value */
	xfs_ifork_t	*ifp;			/* inode fork pointer */
	xfs_extnum_t	lastx;			/* last extent used */
	xfs_bmbt_irec_t	prev;			/* previous extent value */

	if (XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_BTREE &&
	    XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_EXTENTS &&
	    XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_LOCAL)
D
Dave Chinner 已提交
1508
	       return -EIO;
1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522
	if (XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_LOCAL) {
		*last_block = 0;
		return 0;
	}
	ifp = XFS_IFORK_PTR(ip, whichfork);
	if (!(ifp->if_flags & XFS_IFEXTENTS) &&
	    (error = xfs_iread_extents(tp, ip, whichfork)))
		return error;
	bno = *last_block - 1;
	ep = xfs_bmap_search_extents(ip, bno, whichfork, &eof, &lastx, &got,
		&prev);
	if (eof || xfs_bmbt_get_startoff(ep) > bno) {
		if (prev.br_startoff == NULLFILEOFF)
			*last_block = 0;
1523
		else
1524
			*last_block = prev.br_startoff + prev.br_blockcount;
1525 1526
	}
	/*
1527
	 * Otherwise *last_block is already the right answer.
1528
	 */
1529 1530 1531
	return 0;
}

D
Dave Chinner 已提交
1532
int
1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547
xfs_bmap_last_extent(
	struct xfs_trans	*tp,
	struct xfs_inode	*ip,
	int			whichfork,
	struct xfs_bmbt_irec	*rec,
	int			*is_empty)
{
	struct xfs_ifork	*ifp = XFS_IFORK_PTR(ip, whichfork);
	int			error;
	int			nextents;

	if (!(ifp->if_flags & XFS_IFEXTENTS)) {
		error = xfs_iread_extents(tp, ip, whichfork);
		if (error)
			return error;
1548 1549
	}

1550 1551 1552 1553 1554
	nextents = ifp->if_bytes / sizeof(xfs_bmbt_rec_t);
	if (nextents == 0) {
		*is_empty = 1;
		return 0;
	}
1555

1556 1557
	xfs_bmbt_get_all(xfs_iext_get_ext(ifp, nextents - 1), rec);
	*is_empty = 0;
1558 1559 1560
	return 0;
}

1561 1562 1563 1564 1565 1566
/*
 * Check the last inode extent to determine whether this allocation will result
 * in blocks being allocated at the end of the file. When we allocate new data
 * blocks at the end of the file which do not start at the previous data block,
 * we will try to align the new blocks at stripe unit boundaries.
 *
1567
 * Returns 1 in bma->aeof if the file (fork) is empty as any new write will be
1568 1569 1570 1571 1572 1573
 * at, or past the EOF.
 */
STATIC int
xfs_bmap_isaeof(
	struct xfs_bmalloca	*bma,
	int			whichfork)
L
Linus Torvalds 已提交
1574
{
1575 1576 1577
	struct xfs_bmbt_irec	rec;
	int			is_empty;
	int			error;
L
Linus Torvalds 已提交
1578

1579 1580 1581
	bma->aeof = 0;
	error = xfs_bmap_last_extent(NULL, bma->ip, whichfork, &rec,
				     &is_empty);
1582
	if (error)
1583
		return error;
L
Linus Torvalds 已提交
1584

1585 1586 1587 1588 1589
	if (is_empty) {
		bma->aeof = 1;
		return 0;
	}

L
Linus Torvalds 已提交
1590
	/*
1591 1592
	 * Check if we are allocation or past the last extent, or at least into
	 * the last delayed allocated extent.
L
Linus Torvalds 已提交
1593
	 */
1594 1595 1596 1597 1598
	bma->aeof = bma->offset >= rec.br_startoff + rec.br_blockcount ||
		(bma->offset >= rec.br_startoff &&
		 isnullstartblock(rec.br_startblock));
	return 0;
}
L
Linus Torvalds 已提交
1599

1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613
/*
 * Returns the file-relative block number of the first block past eof in
 * the file.  This is not based on i_size, it is based on the extent records.
 * Returns 0 for local files, as they do not have extent records.
 */
int
xfs_bmap_last_offset(
	struct xfs_inode	*ip,
	xfs_fileoff_t		*last_block,
	int			whichfork)
{
	struct xfs_bmbt_irec	rec;
	int			is_empty;
	int			error;
1614

1615
	*last_block = 0;
1616

1617 1618
	if (XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_LOCAL)
		return 0;
1619

1620 1621
	if (XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_BTREE &&
	    XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_EXTENTS)
D
Dave Chinner 已提交
1622
	       return -EIO;
1623

1624 1625
	error = xfs_bmap_last_extent(NULL, ip, whichfork, &rec, &is_empty);
	if (error || is_empty)
1626
		return error;
1627 1628

	*last_block = rec.br_startoff + rec.br_blockcount;
1629 1630 1631
	return 0;
}

1632 1633 1634 1635 1636 1637 1638 1639 1640
/*
 * Returns whether the selected fork of the inode has exactly one
 * block or not.  For the data fork we check this matches di_size,
 * implying the file's range is 0..bsize-1.
 */
int					/* 1=>1 block, 0=>otherwise */
xfs_bmap_one_block(
	xfs_inode_t	*ip,		/* incore inode */
	int		whichfork)	/* data or attr fork */
C
Christoph Hellwig 已提交
1641
{
1642 1643 1644 1645
	xfs_bmbt_rec_host_t *ep;	/* ptr to fork's extent */
	xfs_ifork_t	*ifp;		/* inode fork pointer */
	int		rval;		/* return value */
	xfs_bmbt_irec_t	s;		/* internal version of extent */
C
Christoph Hellwig 已提交
1646

1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663
#ifndef DEBUG
	if (whichfork == XFS_DATA_FORK)
		return XFS_ISIZE(ip) == ip->i_mount->m_sb.sb_blocksize;
#endif	/* !DEBUG */
	if (XFS_IFORK_NEXTENTS(ip, whichfork) != 1)
		return 0;
	if (XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_EXTENTS)
		return 0;
	ifp = XFS_IFORK_PTR(ip, whichfork);
	ASSERT(ifp->if_flags & XFS_IFEXTENTS);
	ep = xfs_iext_get_ext(ifp, 0);
	xfs_bmbt_get_all(ep, &s);
	rval = s.br_startoff == 0 && s.br_blockcount == 1;
	if (rval && whichfork == XFS_DATA_FORK)
		ASSERT(XFS_ISIZE(ip) == ip->i_mount->m_sb.sb_blocksize);
	return rval;
}
C
Christoph Hellwig 已提交
1664

1665 1666 1667
/*
 * Extent tree manipulation functions used during allocation.
 */
C
Christoph Hellwig 已提交
1668

1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691
/*
 * Convert a delayed allocation to a real allocation.
 */
STATIC int				/* error */
xfs_bmap_add_extent_delay_real(
	struct xfs_bmalloca	*bma)
{
	struct xfs_bmbt_irec	*new = &bma->got;
	int			diff;	/* temp value */
	xfs_bmbt_rec_host_t	*ep;	/* extent entry for idx */
	int			error;	/* error return value */
	int			i;	/* temp state */
	xfs_ifork_t		*ifp;	/* inode fork pointer */
	xfs_fileoff_t		new_endoff;	/* end offset of new entry */
	xfs_bmbt_irec_t		r[3];	/* neighbor extent entries */
					/* left is 0, right is 1, prev is 2 */
	int			rval=0;	/* return value (logging flags) */
	int			state = 0;/* state bits, accessed thru macros */
	xfs_filblks_t		da_new; /* new count del alloc blocks used */
	xfs_filblks_t		da_old; /* old count del alloc blocks used */
	xfs_filblks_t		temp=0;	/* value for da_new calculations */
	xfs_filblks_t		temp2=0;/* value for da_new calculations */
	int			tmp_rval;	/* partial logging flags */
1692
	int			whichfork = XFS_DATA_FORK;
1693
	struct xfs_mount	*mp;
C
Christoph Hellwig 已提交
1694

1695
	mp = bma->ip->i_mount;
1696
	ifp = XFS_IFORK_PTR(bma->ip, whichfork);
C
Christoph Hellwig 已提交
1697

1698 1699 1700 1701 1702
	ASSERT(bma->idx >= 0);
	ASSERT(bma->idx <= ifp->if_bytes / sizeof(struct xfs_bmbt_rec));
	ASSERT(!isnullstartblock(new->br_startblock));
	ASSERT(!bma->cur ||
	       (bma->cur->bc_private.b.flags & XFS_BTCUR_BPRV_WASDEL));
C
Christoph Hellwig 已提交
1703

1704
	XFS_STATS_INC(mp, xs_add_exlist);
C
Christoph Hellwig 已提交
1705

1706 1707 1708
#define	LEFT		r[0]
#define	RIGHT		r[1]
#define	PREV		r[2]
C
Christoph Hellwig 已提交
1709 1710

	/*
1711
	 * Set up a bunch of variables to make the tests simpler.
C
Christoph Hellwig 已提交
1712
	 */
1713 1714 1715 1716 1717 1718 1719 1720 1721
	ep = xfs_iext_get_ext(ifp, bma->idx);
	xfs_bmbt_get_all(ep, &PREV);
	new_endoff = new->br_startoff + new->br_blockcount;
	ASSERT(PREV.br_startoff <= new->br_startoff);
	ASSERT(PREV.br_startoff + PREV.br_blockcount >= new_endoff);

	da_old = startblockval(PREV.br_startblock);
	da_new = 0;

C
Christoph Hellwig 已提交
1722
	/*
1723 1724
	 * Set flags determining what part of the previous delayed allocation
	 * extent is being replaced by a real allocation.
C
Christoph Hellwig 已提交
1725
	 */
1726 1727 1728 1729
	if (PREV.br_startoff == new->br_startoff)
		state |= BMAP_LEFT_FILLING;
	if (PREV.br_startoff + PREV.br_blockcount == new_endoff)
		state |= BMAP_RIGHT_FILLING;
C
Christoph Hellwig 已提交
1730 1731

	/*
1732 1733
	 * Check and set flags if this segment has a left neighbor.
	 * Don't set contiguous if the combined extent would be too large.
C
Christoph Hellwig 已提交
1734
	 */
1735 1736 1737
	if (bma->idx > 0) {
		state |= BMAP_LEFT_VALID;
		xfs_bmbt_get_all(xfs_iext_get_ext(ifp, bma->idx - 1), &LEFT);
1738

1739 1740
		if (isnullstartblock(LEFT.br_startblock))
			state |= BMAP_LEFT_DELAY;
1741 1742
	}

1743 1744 1745 1746 1747 1748
	if ((state & BMAP_LEFT_VALID) && !(state & BMAP_LEFT_DELAY) &&
	    LEFT.br_startoff + LEFT.br_blockcount == new->br_startoff &&
	    LEFT.br_startblock + LEFT.br_blockcount == new->br_startblock &&
	    LEFT.br_state == new->br_state &&
	    LEFT.br_blockcount + new->br_blockcount <= MAXEXTLEN)
		state |= BMAP_LEFT_CONTIG;
1749

L
Linus Torvalds 已提交
1750
	/*
1751 1752 1753
	 * Check and set flags if this segment has a right neighbor.
	 * Don't set contiguous if the combined extent would be too large.
	 * Also check for all-three-contiguous being too large.
L
Linus Torvalds 已提交
1754
	 */
1755
	if (bma->idx < ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t) - 1) {
1756 1757
		state |= BMAP_RIGHT_VALID;
		xfs_bmbt_get_all(xfs_iext_get_ext(ifp, bma->idx + 1), &RIGHT);
1758

1759 1760
		if (isnullstartblock(RIGHT.br_startblock))
			state |= BMAP_RIGHT_DELAY;
L
Linus Torvalds 已提交
1761
	}
1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776

	if ((state & BMAP_RIGHT_VALID) && !(state & BMAP_RIGHT_DELAY) &&
	    new_endoff == RIGHT.br_startoff &&
	    new->br_startblock + new->br_blockcount == RIGHT.br_startblock &&
	    new->br_state == RIGHT.br_state &&
	    new->br_blockcount + RIGHT.br_blockcount <= MAXEXTLEN &&
	    ((state & (BMAP_LEFT_CONTIG | BMAP_LEFT_FILLING |
		       BMAP_RIGHT_FILLING)) !=
		      (BMAP_LEFT_CONTIG | BMAP_LEFT_FILLING |
		       BMAP_RIGHT_FILLING) ||
	     LEFT.br_blockcount + new->br_blockcount + RIGHT.br_blockcount
			<= MAXEXTLEN))
		state |= BMAP_RIGHT_CONTIG;

	error = 0;
L
Linus Torvalds 已提交
1777
	/*
1778
	 * Switch out based on the FILLING and CONTIG state bits.
L
Linus Torvalds 已提交
1779
	 */
1780 1781 1782 1783
	switch (state & (BMAP_LEFT_FILLING | BMAP_LEFT_CONTIG |
			 BMAP_RIGHT_FILLING | BMAP_RIGHT_CONTIG)) {
	case BMAP_LEFT_FILLING | BMAP_LEFT_CONTIG |
	     BMAP_RIGHT_FILLING | BMAP_RIGHT_CONTIG:
L
Linus Torvalds 已提交
1784
		/*
1785 1786
		 * Filling in all of a previously delayed allocation extent.
		 * The left and right neighbors are both contiguous with new.
L
Linus Torvalds 已提交
1787
		 */
1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805
		bma->idx--;
		trace_xfs_bmap_pre_update(bma->ip, bma->idx, state, _THIS_IP_);
		xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, bma->idx),
			LEFT.br_blockcount + PREV.br_blockcount +
			RIGHT.br_blockcount);
		trace_xfs_bmap_post_update(bma->ip, bma->idx, state, _THIS_IP_);

		xfs_iext_remove(bma->ip, bma->idx + 1, 2, state);
		bma->ip->i_d.di_nextents--;
		if (bma->cur == NULL)
			rval = XFS_ILOG_CORE | XFS_ILOG_DEXT;
		else {
			rval = XFS_ILOG_CORE;
			error = xfs_bmbt_lookup_eq(bma->cur, RIGHT.br_startoff,
					RIGHT.br_startblock,
					RIGHT.br_blockcount, &i);
			if (error)
				goto done;
1806
			XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
1807 1808 1809
			error = xfs_btree_delete(bma->cur, &i);
			if (error)
				goto done;
1810
			XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
1811 1812 1813
			error = xfs_btree_decrement(bma->cur, 0, &i);
			if (error)
				goto done;
1814
			XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825
			error = xfs_bmbt_update(bma->cur, LEFT.br_startoff,
					LEFT.br_startblock,
					LEFT.br_blockcount +
					PREV.br_blockcount +
					RIGHT.br_blockcount, LEFT.br_state);
			if (error)
				goto done;
		}
		break;

	case BMAP_LEFT_FILLING | BMAP_RIGHT_FILLING | BMAP_LEFT_CONTIG:
1826
		/*
1827 1828
		 * Filling in all of a previously delayed allocation extent.
		 * The left neighbor is contiguous, the right is not.
1829
		 */
1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846
		bma->idx--;

		trace_xfs_bmap_pre_update(bma->ip, bma->idx, state, _THIS_IP_);
		xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, bma->idx),
			LEFT.br_blockcount + PREV.br_blockcount);
		trace_xfs_bmap_post_update(bma->ip, bma->idx, state, _THIS_IP_);

		xfs_iext_remove(bma->ip, bma->idx + 1, 1, state);
		if (bma->cur == NULL)
			rval = XFS_ILOG_DEXT;
		else {
			rval = 0;
			error = xfs_bmbt_lookup_eq(bma->cur, LEFT.br_startoff,
					LEFT.br_startblock, LEFT.br_blockcount,
					&i);
			if (error)
				goto done;
1847
			XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
1848 1849 1850 1851 1852 1853 1854 1855 1856 1857
			error = xfs_bmbt_update(bma->cur, LEFT.br_startoff,
					LEFT.br_startblock,
					LEFT.br_blockcount +
					PREV.br_blockcount, LEFT.br_state);
			if (error)
				goto done;
		}
		break;

	case BMAP_LEFT_FILLING | BMAP_RIGHT_FILLING | BMAP_RIGHT_CONTIG:
1858
		/*
1859 1860
		 * Filling in all of a previously delayed allocation extent.
		 * The right neighbor is contiguous, the left is not.
1861
		 */
1862 1863 1864 1865 1866
		trace_xfs_bmap_pre_update(bma->ip, bma->idx, state, _THIS_IP_);
		xfs_bmbt_set_startblock(ep, new->br_startblock);
		xfs_bmbt_set_blockcount(ep,
			PREV.br_blockcount + RIGHT.br_blockcount);
		trace_xfs_bmap_post_update(bma->ip, bma->idx, state, _THIS_IP_);
1867

1868 1869 1870 1871 1872 1873 1874 1875 1876 1877
		xfs_iext_remove(bma->ip, bma->idx + 1, 1, state);
		if (bma->cur == NULL)
			rval = XFS_ILOG_DEXT;
		else {
			rval = 0;
			error = xfs_bmbt_lookup_eq(bma->cur, RIGHT.br_startoff,
					RIGHT.br_startblock,
					RIGHT.br_blockcount, &i);
			if (error)
				goto done;
1878
			XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
1879 1880 1881 1882 1883 1884 1885 1886 1887 1888
			error = xfs_bmbt_update(bma->cur, PREV.br_startoff,
					new->br_startblock,
					PREV.br_blockcount +
					RIGHT.br_blockcount, PREV.br_state);
			if (error)
				goto done;
		}
		break;

	case BMAP_LEFT_FILLING | BMAP_RIGHT_FILLING:
1889
		/*
1890 1891 1892
		 * Filling in all of a previously delayed allocation extent.
		 * Neither the left nor right neighbors are contiguous with
		 * the new one.
1893
		 */
1894 1895 1896
		trace_xfs_bmap_pre_update(bma->ip, bma->idx, state, _THIS_IP_);
		xfs_bmbt_set_startblock(ep, new->br_startblock);
		trace_xfs_bmap_post_update(bma->ip, bma->idx, state, _THIS_IP_);
1897

1898 1899 1900 1901 1902 1903 1904 1905 1906 1907
		bma->ip->i_d.di_nextents++;
		if (bma->cur == NULL)
			rval = XFS_ILOG_CORE | XFS_ILOG_DEXT;
		else {
			rval = XFS_ILOG_CORE;
			error = xfs_bmbt_lookup_eq(bma->cur, new->br_startoff,
					new->br_startblock, new->br_blockcount,
					&i);
			if (error)
				goto done;
1908
			XFS_WANT_CORRUPTED_GOTO(mp, i == 0, done);
1909 1910 1911 1912
			bma->cur->bc_rec.b.br_state = XFS_EXT_NORM;
			error = xfs_btree_insert(bma->cur, &i);
			if (error)
				goto done;
1913
			XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
1914 1915
		}
		break;
L
Linus Torvalds 已提交
1916

1917
	case BMAP_LEFT_FILLING | BMAP_LEFT_CONTIG:
L
Linus Torvalds 已提交
1918
		/*
1919 1920
		 * Filling in the first part of a previous delayed allocation.
		 * The left neighbor is contiguous.
L
Linus Torvalds 已提交
1921
		 */
1922 1923 1924 1925 1926 1927
		trace_xfs_bmap_pre_update(bma->ip, bma->idx - 1, state, _THIS_IP_);
		xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, bma->idx - 1),
			LEFT.br_blockcount + new->br_blockcount);
		xfs_bmbt_set_startoff(ep,
			PREV.br_startoff + new->br_blockcount);
		trace_xfs_bmap_post_update(bma->ip, bma->idx - 1, state, _THIS_IP_);
L
Linus Torvalds 已提交
1928

1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940
		temp = PREV.br_blockcount - new->br_blockcount;
		trace_xfs_bmap_pre_update(bma->ip, bma->idx, state, _THIS_IP_);
		xfs_bmbt_set_blockcount(ep, temp);
		if (bma->cur == NULL)
			rval = XFS_ILOG_DEXT;
		else {
			rval = 0;
			error = xfs_bmbt_lookup_eq(bma->cur, LEFT.br_startoff,
					LEFT.br_startblock, LEFT.br_blockcount,
					&i);
			if (error)
				goto done;
1941
			XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
1942 1943 1944 1945 1946
			error = xfs_bmbt_update(bma->cur, LEFT.br_startoff,
					LEFT.br_startblock,
					LEFT.br_blockcount +
					new->br_blockcount,
					LEFT.br_state);
C
Christoph Hellwig 已提交
1947
			if (error)
L
Linus Torvalds 已提交
1948 1949
				goto done;
		}
1950 1951 1952 1953 1954 1955 1956 1957 1958
		da_new = XFS_FILBLKS_MIN(xfs_bmap_worst_indlen(bma->ip, temp),
			startblockval(PREV.br_startblock));
		xfs_bmbt_set_startblock(ep, nullstartblock(da_new));
		trace_xfs_bmap_post_update(bma->ip, bma->idx, state, _THIS_IP_);

		bma->idx--;
		break;

	case BMAP_LEFT_FILLING:
L
Linus Torvalds 已提交
1959
		/*
1960 1961
		 * Filling in the first part of a previous delayed allocation.
		 * The left neighbor is not contiguous.
L
Linus Torvalds 已提交
1962
		 */
1963 1964 1965 1966 1967 1968 1969 1970
		trace_xfs_bmap_pre_update(bma->ip, bma->idx, state, _THIS_IP_);
		xfs_bmbt_set_startoff(ep, new_endoff);
		temp = PREV.br_blockcount - new->br_blockcount;
		xfs_bmbt_set_blockcount(ep, temp);
		xfs_iext_insert(bma->ip, bma->idx, 1, new, state);
		bma->ip->i_d.di_nextents++;
		if (bma->cur == NULL)
			rval = XFS_ILOG_CORE | XFS_ILOG_DEXT;
L
Linus Torvalds 已提交
1971
		else {
1972 1973 1974 1975 1976 1977
			rval = XFS_ILOG_CORE;
			error = xfs_bmbt_lookup_eq(bma->cur, new->br_startoff,
					new->br_startblock, new->br_blockcount,
					&i);
			if (error)
				goto done;
1978
			XFS_WANT_CORRUPTED_GOTO(mp, i == 0, done);
1979 1980 1981
			bma->cur->bc_rec.b.br_state = XFS_EXT_NORM;
			error = xfs_btree_insert(bma->cur, &i);
			if (error)
L
Linus Torvalds 已提交
1982
				goto done;
1983
			XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
L
Linus Torvalds 已提交
1984
		}
1985

1986
		if (xfs_bmap_needs_btree(bma->ip, whichfork)) {
1987
			error = xfs_bmap_extents_to_btree(bma->tp, bma->ip,
1988
					bma->firstblock, bma->dfops,
1989
					&bma->cur, 1, &tmp_rval, whichfork);
1990 1991 1992
			rval |= tmp_rval;
			if (error)
				goto done;
L
Linus Torvalds 已提交
1993
		}
1994 1995 1996 1997 1998 1999
		da_new = XFS_FILBLKS_MIN(xfs_bmap_worst_indlen(bma->ip, temp),
			startblockval(PREV.br_startblock) -
			(bma->cur ? bma->cur->bc_private.b.allocated : 0));
		ep = xfs_iext_get_ext(ifp, bma->idx + 1);
		xfs_bmbt_set_startblock(ep, nullstartblock(da_new));
		trace_xfs_bmap_post_update(bma->ip, bma->idx + 1, state, _THIS_IP_);
L
Linus Torvalds 已提交
2000 2001
		break;

2002
	case BMAP_RIGHT_FILLING | BMAP_RIGHT_CONTIG:
L
Linus Torvalds 已提交
2003
		/*
2004 2005
		 * Filling in the last part of a previous delayed allocation.
		 * The right neighbor is contiguous with the new allocation.
L
Linus Torvalds 已提交
2006
		 */
2007 2008
		temp = PREV.br_blockcount - new->br_blockcount;
		trace_xfs_bmap_pre_update(bma->ip, bma->idx + 1, state, _THIS_IP_);
L
Linus Torvalds 已提交
2009
		xfs_bmbt_set_blockcount(ep, temp);
2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023
		xfs_bmbt_set_allf(xfs_iext_get_ext(ifp, bma->idx + 1),
			new->br_startoff, new->br_startblock,
			new->br_blockcount + RIGHT.br_blockcount,
			RIGHT.br_state);
		trace_xfs_bmap_post_update(bma->ip, bma->idx + 1, state, _THIS_IP_);
		if (bma->cur == NULL)
			rval = XFS_ILOG_DEXT;
		else {
			rval = 0;
			error = xfs_bmbt_lookup_eq(bma->cur, RIGHT.br_startoff,
					RIGHT.br_startblock,
					RIGHT.br_blockcount, &i);
			if (error)
				goto done;
2024
			XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
2025 2026 2027 2028 2029 2030 2031
			error = xfs_bmbt_update(bma->cur, new->br_startoff,
					new->br_startblock,
					new->br_blockcount +
					RIGHT.br_blockcount,
					RIGHT.br_state);
			if (error)
				goto done;
L
Linus Torvalds 已提交
2032
		}
2033 2034 2035 2036 2037 2038 2039 2040

		da_new = XFS_FILBLKS_MIN(xfs_bmap_worst_indlen(bma->ip, temp),
			startblockval(PREV.br_startblock));
		trace_xfs_bmap_pre_update(bma->ip, bma->idx, state, _THIS_IP_);
		xfs_bmbt_set_startblock(ep, nullstartblock(da_new));
		trace_xfs_bmap_post_update(bma->ip, bma->idx, state, _THIS_IP_);

		bma->idx++;
L
Linus Torvalds 已提交
2041 2042
		break;

2043
	case BMAP_RIGHT_FILLING:
L
Linus Torvalds 已提交
2044
		/*
2045 2046
		 * Filling in the last part of a previous delayed allocation.
		 * The right neighbor is not contiguous.
L
Linus Torvalds 已提交
2047
		 */
2048 2049
		temp = PREV.br_blockcount - new->br_blockcount;
		trace_xfs_bmap_pre_update(bma->ip, bma->idx, state, _THIS_IP_);
L
Linus Torvalds 已提交
2050
		xfs_bmbt_set_blockcount(ep, temp);
2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061
		xfs_iext_insert(bma->ip, bma->idx + 1, 1, new, state);
		bma->ip->i_d.di_nextents++;
		if (bma->cur == NULL)
			rval = XFS_ILOG_CORE | XFS_ILOG_DEXT;
		else {
			rval = XFS_ILOG_CORE;
			error = xfs_bmbt_lookup_eq(bma->cur, new->br_startoff,
					new->br_startblock, new->br_blockcount,
					&i);
			if (error)
				goto done;
2062
			XFS_WANT_CORRUPTED_GOTO(mp, i == 0, done);
2063 2064 2065 2066
			bma->cur->bc_rec.b.br_state = XFS_EXT_NORM;
			error = xfs_btree_insert(bma->cur, &i);
			if (error)
				goto done;
2067
			XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
L
Linus Torvalds 已提交
2068
		}
2069

2070
		if (xfs_bmap_needs_btree(bma->ip, whichfork)) {
2071
			error = xfs_bmap_extents_to_btree(bma->tp, bma->ip,
2072
				bma->firstblock, bma->dfops, &bma->cur, 1,
2073
				&tmp_rval, whichfork);
2074 2075 2076
			rval |= tmp_rval;
			if (error)
				goto done;
L
Linus Torvalds 已提交
2077
		}
2078 2079 2080 2081 2082 2083 2084 2085
		da_new = XFS_FILBLKS_MIN(xfs_bmap_worst_indlen(bma->ip, temp),
			startblockval(PREV.br_startblock) -
			(bma->cur ? bma->cur->bc_private.b.allocated : 0));
		ep = xfs_iext_get_ext(ifp, bma->idx);
		xfs_bmbt_set_startblock(ep, nullstartblock(da_new));
		trace_xfs_bmap_post_update(bma->ip, bma->idx, state, _THIS_IP_);

		bma->idx++;
L
Linus Torvalds 已提交
2086 2087 2088 2089
		break;

	case 0:
		/*
2090 2091 2092 2093 2094 2095 2096 2097 2098 2099 2100 2101 2102 2103 2104 2105 2106 2107
		 * Filling in the middle part of a previous delayed allocation.
		 * Contiguity is impossible here.
		 * This case is avoided almost all the time.
		 *
		 * We start with a delayed allocation:
		 *
		 * +ddddddddddddddddddddddddddddddddddddddddddddddddddddddd+
		 *  PREV @ idx
		 *
	         * and we are allocating:
		 *                     +rrrrrrrrrrrrrrrrr+
		 *			      new
		 *
		 * and we set it up for insertion as:
		 * +ddddddddddddddddddd+rrrrrrrrrrrrrrrrr+ddddddddddddddddd+
		 *                            new
		 *  PREV @ idx          LEFT              RIGHT
		 *                      inserted at idx + 1
L
Linus Torvalds 已提交
2108
		 */
2109 2110 2111 2112 2113 2114 2115 2116 2117 2118 2119 2120 2121 2122 2123 2124 2125 2126 2127 2128 2129 2130
		temp = new->br_startoff - PREV.br_startoff;
		temp2 = PREV.br_startoff + PREV.br_blockcount - new_endoff;
		trace_xfs_bmap_pre_update(bma->ip, bma->idx, 0, _THIS_IP_);
		xfs_bmbt_set_blockcount(ep, temp);	/* truncate PREV */
		LEFT = *new;
		RIGHT.br_state = PREV.br_state;
		RIGHT.br_startblock = nullstartblock(
				(int)xfs_bmap_worst_indlen(bma->ip, temp2));
		RIGHT.br_startoff = new_endoff;
		RIGHT.br_blockcount = temp2;
		/* insert LEFT (r[0]) and RIGHT (r[1]) at the same time */
		xfs_iext_insert(bma->ip, bma->idx + 1, 2, &LEFT, state);
		bma->ip->i_d.di_nextents++;
		if (bma->cur == NULL)
			rval = XFS_ILOG_CORE | XFS_ILOG_DEXT;
		else {
			rval = XFS_ILOG_CORE;
			error = xfs_bmbt_lookup_eq(bma->cur, new->br_startoff,
					new->br_startblock, new->br_blockcount,
					&i);
			if (error)
				goto done;
2131
			XFS_WANT_CORRUPTED_GOTO(mp, i == 0, done);
2132 2133 2134 2135
			bma->cur->bc_rec.b.br_state = XFS_EXT_NORM;
			error = xfs_btree_insert(bma->cur, &i);
			if (error)
				goto done;
2136
			XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
L
Linus Torvalds 已提交
2137 2138
		}

2139
		if (xfs_bmap_needs_btree(bma->ip, whichfork)) {
2140
			error = xfs_bmap_extents_to_btree(bma->tp, bma->ip,
2141
					bma->firstblock, bma->dfops, &bma->cur,
2142
					1, &tmp_rval, whichfork);
2143 2144 2145 2146 2147 2148 2149 2150 2151
			rval |= tmp_rval;
			if (error)
				goto done;
		}
		temp = xfs_bmap_worst_indlen(bma->ip, temp);
		temp2 = xfs_bmap_worst_indlen(bma->ip, temp2);
		diff = (int)(temp + temp2 - startblockval(PREV.br_startblock) -
			(bma->cur ? bma->cur->bc_private.b.allocated : 0));
		if (diff > 0) {
2152 2153
			error = xfs_mod_fdblocks(bma->ip->i_mount,
						 -((int64_t)diff), false);
2154 2155 2156 2157 2158 2159 2160 2161 2162 2163 2164 2165 2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182 2183
			ASSERT(!error);
			if (error)
				goto done;
		}

		ep = xfs_iext_get_ext(ifp, bma->idx);
		xfs_bmbt_set_startblock(ep, nullstartblock((int)temp));
		trace_xfs_bmap_post_update(bma->ip, bma->idx, state, _THIS_IP_);
		trace_xfs_bmap_pre_update(bma->ip, bma->idx + 2, state, _THIS_IP_);
		xfs_bmbt_set_startblock(xfs_iext_get_ext(ifp, bma->idx + 2),
			nullstartblock((int)temp2));
		trace_xfs_bmap_post_update(bma->ip, bma->idx + 2, state, _THIS_IP_);

		bma->idx++;
		da_new = temp + temp2;
		break;

	case BMAP_LEFT_FILLING | BMAP_LEFT_CONTIG | BMAP_RIGHT_CONTIG:
	case BMAP_RIGHT_FILLING | BMAP_LEFT_CONTIG | BMAP_RIGHT_CONTIG:
	case BMAP_LEFT_FILLING | BMAP_RIGHT_CONTIG:
	case BMAP_RIGHT_FILLING | BMAP_LEFT_CONTIG:
	case BMAP_LEFT_CONTIG | BMAP_RIGHT_CONTIG:
	case BMAP_LEFT_CONTIG:
	case BMAP_RIGHT_CONTIG:
		/*
		 * These cases are all impossible.
		 */
		ASSERT(0);
	}

2184 2185 2186 2187 2188
	/* add reverse mapping */
	error = xfs_rmap_map_extent(mp, bma->dfops, bma->ip, whichfork, new);
	if (error)
		goto done;

2189
	/* convert to a btree if necessary */
2190
	if (xfs_bmap_needs_btree(bma->ip, whichfork)) {
2191 2192 2193 2194
		int	tmp_logflags;	/* partial log flag return val */

		ASSERT(bma->cur == NULL);
		error = xfs_bmap_extents_to_btree(bma->tp, bma->ip,
2195
				bma->firstblock, bma->dfops, &bma->cur,
2196
				da_old > 0, &tmp_logflags, whichfork);
2197 2198 2199 2200 2201 2202 2203 2204 2205 2206 2207 2208
		bma->logflags |= tmp_logflags;
		if (error)
			goto done;
	}

	/* adjust for changes in reserved delayed indirect blocks */
	if (da_old || da_new) {
		temp = da_new;
		if (bma->cur)
			temp += bma->cur->bc_private.b.allocated;
		ASSERT(temp <= da_old);
		if (temp < da_old)
2209 2210
			xfs_mod_fdblocks(bma->ip->i_mount,
					(int64_t)(da_old - temp), false);
2211
	}
2212 2213 2214 2215 2216

	/* clear out the allocated field, done with it now in any case. */
	if (bma->cur)
		bma->cur->bc_private.b.allocated = 0;

2217
	xfs_bmap_check_leaf_extents(bma->cur, bma->ip, whichfork);
L
Linus Torvalds 已提交
2218
done:
2219
	bma->logflags |= rval;
L
Linus Torvalds 已提交
2220
	return error;
2221 2222 2223
#undef	LEFT
#undef	RIGHT
#undef	PREV
L
Linus Torvalds 已提交
2224 2225 2226
}

/*
2227
 * Convert an unwritten allocation to a real allocation or vice versa.
L
Linus Torvalds 已提交
2228
 */
2229 2230 2231 2232 2233 2234 2235 2236
STATIC int				/* error */
xfs_bmap_add_extent_unwritten_real(
	struct xfs_trans	*tp,
	xfs_inode_t		*ip,	/* incore inode pointer */
	xfs_extnum_t		*idx,	/* extent number to update/insert */
	xfs_btree_cur_t		**curp,	/* if *curp is null, not a btree */
	xfs_bmbt_irec_t		*new,	/* new data to add to file extents */
	xfs_fsblock_t		*first,	/* pointer to firstblock variable */
2237
	struct xfs_defer_ops	*dfops,	/* list of extents to be freed */
2238
	int			*logflagsp) /* inode logging flags */
L
Linus Torvalds 已提交
2239
{
2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250 2251
	xfs_btree_cur_t		*cur;	/* btree cursor */
	xfs_bmbt_rec_host_t	*ep;	/* extent entry for idx */
	int			error;	/* error return value */
	int			i;	/* temp state */
	xfs_ifork_t		*ifp;	/* inode fork pointer */
	xfs_fileoff_t		new_endoff;	/* end offset of new entry */
	xfs_exntst_t		newext;	/* new extent state */
	xfs_exntst_t		oldext;	/* old extent state */
	xfs_bmbt_irec_t		r[3];	/* neighbor extent entries */
					/* left is 0, right is 1, prev is 2 */
	int			rval=0;	/* return value (logging flags) */
	int			state = 0;/* state bits, accessed thru macros */
2252
	struct xfs_mount	*mp = tp->t_mountp;
L
Linus Torvalds 已提交
2253

2254
	*logflagsp = 0;
L
Linus Torvalds 已提交
2255

2256 2257
	cur = *curp;
	ifp = XFS_IFORK_PTR(ip, XFS_DATA_FORK);
2258

2259 2260 2261 2262
	ASSERT(*idx >= 0);
	ASSERT(*idx <= ifp->if_bytes / sizeof(struct xfs_bmbt_rec));
	ASSERT(!isnullstartblock(new->br_startblock));

2263
	XFS_STATS_INC(mp, xs_add_exlist);
2264 2265 2266 2267

#define	LEFT		r[0]
#define	RIGHT		r[1]
#define	PREV		r[2]
2268

L
Linus Torvalds 已提交
2269
	/*
2270
	 * Set up a bunch of variables to make the tests simpler.
L
Linus Torvalds 已提交
2271
	 */
2272 2273 2274 2275 2276 2277 2278 2279 2280 2281
	error = 0;
	ep = xfs_iext_get_ext(ifp, *idx);
	xfs_bmbt_get_all(ep, &PREV);
	newext = new->br_state;
	oldext = (newext == XFS_EXT_UNWRITTEN) ?
		XFS_EXT_NORM : XFS_EXT_UNWRITTEN;
	ASSERT(PREV.br_state == oldext);
	new_endoff = new->br_startoff + new->br_blockcount;
	ASSERT(PREV.br_startoff <= new->br_startoff);
	ASSERT(PREV.br_startoff + PREV.br_blockcount >= new_endoff);
2282

L
Linus Torvalds 已提交
2283
	/*
2284 2285
	 * Set flags determining what part of the previous oldext allocation
	 * extent is being replaced by a newext allocation.
L
Linus Torvalds 已提交
2286
	 */
2287 2288 2289 2290 2291
	if (PREV.br_startoff == new->br_startoff)
		state |= BMAP_LEFT_FILLING;
	if (PREV.br_startoff + PREV.br_blockcount == new_endoff)
		state |= BMAP_RIGHT_FILLING;

L
Linus Torvalds 已提交
2292
	/*
2293 2294
	 * Check and set flags if this segment has a left neighbor.
	 * Don't set contiguous if the combined extent would be too large.
L
Linus Torvalds 已提交
2295
	 */
2296 2297 2298 2299 2300 2301
	if (*idx > 0) {
		state |= BMAP_LEFT_VALID;
		xfs_bmbt_get_all(xfs_iext_get_ext(ifp, *idx - 1), &LEFT);

		if (isnullstartblock(LEFT.br_startblock))
			state |= BMAP_LEFT_DELAY;
L
Linus Torvalds 已提交
2302
	}
2303 2304 2305 2306 2307 2308 2309 2310

	if ((state & BMAP_LEFT_VALID) && !(state & BMAP_LEFT_DELAY) &&
	    LEFT.br_startoff + LEFT.br_blockcount == new->br_startoff &&
	    LEFT.br_startblock + LEFT.br_blockcount == new->br_startblock &&
	    LEFT.br_state == newext &&
	    LEFT.br_blockcount + new->br_blockcount <= MAXEXTLEN)
		state |= BMAP_LEFT_CONTIG;

L
Linus Torvalds 已提交
2311
	/*
2312 2313 2314
	 * Check and set flags if this segment has a right neighbor.
	 * Don't set contiguous if the combined extent would be too large.
	 * Also check for all-three-contiguous being too large.
L
Linus Torvalds 已提交
2315
	 */
2316 2317 2318 2319 2320
	if (*idx < ip->i_df.if_bytes / (uint)sizeof(xfs_bmbt_rec_t) - 1) {
		state |= BMAP_RIGHT_VALID;
		xfs_bmbt_get_all(xfs_iext_get_ext(ifp, *idx + 1), &RIGHT);
		if (isnullstartblock(RIGHT.br_startblock))
			state |= BMAP_RIGHT_DELAY;
L
Linus Torvalds 已提交
2321
	}
2322

2323 2324 2325 2326 2327 2328 2329 2330 2331 2332 2333 2334
	if ((state & BMAP_RIGHT_VALID) && !(state & BMAP_RIGHT_DELAY) &&
	    new_endoff == RIGHT.br_startoff &&
	    new->br_startblock + new->br_blockcount == RIGHT.br_startblock &&
	    newext == RIGHT.br_state &&
	    new->br_blockcount + RIGHT.br_blockcount <= MAXEXTLEN &&
	    ((state & (BMAP_LEFT_CONTIG | BMAP_LEFT_FILLING |
		       BMAP_RIGHT_FILLING)) !=
		      (BMAP_LEFT_CONTIG | BMAP_LEFT_FILLING |
		       BMAP_RIGHT_FILLING) ||
	     LEFT.br_blockcount + new->br_blockcount + RIGHT.br_blockcount
			<= MAXEXTLEN))
		state |= BMAP_RIGHT_CONTIG;
2335

L
Linus Torvalds 已提交
2336
	/*
2337
	 * Switch out based on the FILLING and CONTIG state bits.
L
Linus Torvalds 已提交
2338
	 */
2339 2340 2341 2342 2343 2344 2345 2346 2347
	switch (state & (BMAP_LEFT_FILLING | BMAP_LEFT_CONTIG |
			 BMAP_RIGHT_FILLING | BMAP_RIGHT_CONTIG)) {
	case BMAP_LEFT_FILLING | BMAP_LEFT_CONTIG |
	     BMAP_RIGHT_FILLING | BMAP_RIGHT_CONTIG:
		/*
		 * Setting all of a previous oldext extent to newext.
		 * The left and right neighbors are both contiguous with new.
		 */
		--*idx;
C
Christoph Hellwig 已提交
2348

2349 2350 2351 2352 2353
		trace_xfs_bmap_pre_update(ip, *idx, state, _THIS_IP_);
		xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, *idx),
			LEFT.br_blockcount + PREV.br_blockcount +
			RIGHT.br_blockcount);
		trace_xfs_bmap_post_update(ip, *idx, state, _THIS_IP_);
C
Christoph Hellwig 已提交
2354

2355 2356 2357 2358 2359 2360 2361 2362 2363 2364
		xfs_iext_remove(ip, *idx + 1, 2, state);
		ip->i_d.di_nextents -= 2;
		if (cur == NULL)
			rval = XFS_ILOG_CORE | XFS_ILOG_DEXT;
		else {
			rval = XFS_ILOG_CORE;
			if ((error = xfs_bmbt_lookup_eq(cur, RIGHT.br_startoff,
					RIGHT.br_startblock,
					RIGHT.br_blockcount, &i)))
				goto done;
2365
			XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
2366 2367
			if ((error = xfs_btree_delete(cur, &i)))
				goto done;
2368
			XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
2369 2370
			if ((error = xfs_btree_decrement(cur, 0, &i)))
				goto done;
2371
			XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
2372 2373
			if ((error = xfs_btree_delete(cur, &i)))
				goto done;
2374
			XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
2375 2376
			if ((error = xfs_btree_decrement(cur, 0, &i)))
				goto done;
2377
			XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
2378 2379 2380 2381 2382 2383 2384
			if ((error = xfs_bmbt_update(cur, LEFT.br_startoff,
				LEFT.br_startblock,
				LEFT.br_blockcount + PREV.br_blockcount +
				RIGHT.br_blockcount, LEFT.br_state)))
				goto done;
		}
		break;
C
Christoph Hellwig 已提交
2385

2386 2387 2388 2389 2390 2391
	case BMAP_LEFT_FILLING | BMAP_RIGHT_FILLING | BMAP_LEFT_CONTIG:
		/*
		 * Setting all of a previous oldext extent to newext.
		 * The left neighbor is contiguous, the right is not.
		 */
		--*idx;
2392

2393 2394 2395 2396
		trace_xfs_bmap_pre_update(ip, *idx, state, _THIS_IP_);
		xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, *idx),
			LEFT.br_blockcount + PREV.br_blockcount);
		trace_xfs_bmap_post_update(ip, *idx, state, _THIS_IP_);
L
Linus Torvalds 已提交
2397

2398 2399 2400 2401 2402 2403 2404 2405 2406 2407
		xfs_iext_remove(ip, *idx + 1, 1, state);
		ip->i_d.di_nextents--;
		if (cur == NULL)
			rval = XFS_ILOG_CORE | XFS_ILOG_DEXT;
		else {
			rval = XFS_ILOG_CORE;
			if ((error = xfs_bmbt_lookup_eq(cur, PREV.br_startoff,
					PREV.br_startblock, PREV.br_blockcount,
					&i)))
				goto done;
2408
			XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
2409 2410
			if ((error = xfs_btree_delete(cur, &i)))
				goto done;
2411
			XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
2412 2413
			if ((error = xfs_btree_decrement(cur, 0, &i)))
				goto done;
2414
			XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
2415 2416 2417 2418 2419 2420 2421
			if ((error = xfs_bmbt_update(cur, LEFT.br_startoff,
				LEFT.br_startblock,
				LEFT.br_blockcount + PREV.br_blockcount,
				LEFT.br_state)))
				goto done;
		}
		break;
L
Linus Torvalds 已提交
2422

2423
	case BMAP_LEFT_FILLING | BMAP_RIGHT_FILLING | BMAP_RIGHT_CONTIG:
L
Linus Torvalds 已提交
2424
		/*
2425 2426
		 * Setting all of a previous oldext extent to newext.
		 * The right neighbor is contiguous, the left is not.
L
Linus Torvalds 已提交
2427
		 */
2428 2429 2430 2431 2432 2433 2434 2435 2436 2437 2438 2439 2440 2441 2442
		trace_xfs_bmap_pre_update(ip, *idx, state, _THIS_IP_);
		xfs_bmbt_set_blockcount(ep,
			PREV.br_blockcount + RIGHT.br_blockcount);
		xfs_bmbt_set_state(ep, newext);
		trace_xfs_bmap_post_update(ip, *idx, state, _THIS_IP_);
		xfs_iext_remove(ip, *idx + 1, 1, state);
		ip->i_d.di_nextents--;
		if (cur == NULL)
			rval = XFS_ILOG_CORE | XFS_ILOG_DEXT;
		else {
			rval = XFS_ILOG_CORE;
			if ((error = xfs_bmbt_lookup_eq(cur, RIGHT.br_startoff,
					RIGHT.br_startblock,
					RIGHT.br_blockcount, &i)))
				goto done;
2443
			XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
2444 2445
			if ((error = xfs_btree_delete(cur, &i)))
				goto done;
2446
			XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
2447 2448
			if ((error = xfs_btree_decrement(cur, 0, &i)))
				goto done;
2449
			XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
2450 2451 2452 2453 2454
			if ((error = xfs_bmbt_update(cur, new->br_startoff,
				new->br_startblock,
				new->br_blockcount + RIGHT.br_blockcount,
				newext)))
				goto done;
L
Linus Torvalds 已提交
2455
		}
2456
		break;
2457

2458 2459 2460 2461 2462 2463 2464 2465 2466
	case BMAP_LEFT_FILLING | BMAP_RIGHT_FILLING:
		/*
		 * Setting all of a previous oldext extent to newext.
		 * Neither the left nor right neighbors are contiguous with
		 * the new one.
		 */
		trace_xfs_bmap_pre_update(ip, *idx, state, _THIS_IP_);
		xfs_bmbt_set_state(ep, newext);
		trace_xfs_bmap_post_update(ip, *idx, state, _THIS_IP_);
2467

2468 2469 2470 2471 2472 2473 2474 2475
		if (cur == NULL)
			rval = XFS_ILOG_DEXT;
		else {
			rval = 0;
			if ((error = xfs_bmbt_lookup_eq(cur, new->br_startoff,
					new->br_startblock, new->br_blockcount,
					&i)))
				goto done;
2476
			XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
2477 2478 2479 2480 2481 2482
			if ((error = xfs_bmbt_update(cur, new->br_startoff,
				new->br_startblock, new->br_blockcount,
				newext)))
				goto done;
		}
		break;
2483

2484 2485 2486 2487 2488 2489 2490 2491 2492 2493 2494
	case BMAP_LEFT_FILLING | BMAP_LEFT_CONTIG:
		/*
		 * Setting the first part of a previous oldext extent to newext.
		 * The left neighbor is contiguous.
		 */
		trace_xfs_bmap_pre_update(ip, *idx - 1, state, _THIS_IP_);
		xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, *idx - 1),
			LEFT.br_blockcount + new->br_blockcount);
		xfs_bmbt_set_startoff(ep,
			PREV.br_startoff + new->br_blockcount);
		trace_xfs_bmap_post_update(ip, *idx - 1, state, _THIS_IP_);
L
Linus Torvalds 已提交
2495

2496 2497 2498 2499 2500 2501
		trace_xfs_bmap_pre_update(ip, *idx, state, _THIS_IP_);
		xfs_bmbt_set_startblock(ep,
			new->br_startblock + new->br_blockcount);
		xfs_bmbt_set_blockcount(ep,
			PREV.br_blockcount - new->br_blockcount);
		trace_xfs_bmap_post_update(ip, *idx, state, _THIS_IP_);
2502

2503
		--*idx;
2504

2505 2506 2507 2508 2509 2510 2511 2512
		if (cur == NULL)
			rval = XFS_ILOG_DEXT;
		else {
			rval = 0;
			if ((error = xfs_bmbt_lookup_eq(cur, PREV.br_startoff,
					PREV.br_startblock, PREV.br_blockcount,
					&i)))
				goto done;
2513
			XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
2514 2515 2516 2517 2518 2519 2520 2521 2522 2523 2524 2525 2526 2527
			if ((error = xfs_bmbt_update(cur,
				PREV.br_startoff + new->br_blockcount,
				PREV.br_startblock + new->br_blockcount,
				PREV.br_blockcount - new->br_blockcount,
				oldext)))
				goto done;
			if ((error = xfs_btree_decrement(cur, 0, &i)))
				goto done;
			error = xfs_bmbt_update(cur, LEFT.br_startoff,
				LEFT.br_startblock,
				LEFT.br_blockcount + new->br_blockcount,
				LEFT.br_state);
			if (error)
				goto done;
2528
		}
2529
		break;
2530

2531 2532 2533 2534 2535 2536 2537 2538 2539 2540 2541 2542 2543
	case BMAP_LEFT_FILLING:
		/*
		 * Setting the first part of a previous oldext extent to newext.
		 * The left neighbor is not contiguous.
		 */
		trace_xfs_bmap_pre_update(ip, *idx, state, _THIS_IP_);
		ASSERT(ep && xfs_bmbt_get_state(ep) == oldext);
		xfs_bmbt_set_startoff(ep, new_endoff);
		xfs_bmbt_set_blockcount(ep,
			PREV.br_blockcount - new->br_blockcount);
		xfs_bmbt_set_startblock(ep,
			new->br_startblock + new->br_blockcount);
		trace_xfs_bmap_post_update(ip, *idx, state, _THIS_IP_);
L
Linus Torvalds 已提交
2544

2545 2546 2547 2548 2549 2550 2551 2552 2553 2554
		xfs_iext_insert(ip, *idx, 1, new, state);
		ip->i_d.di_nextents++;
		if (cur == NULL)
			rval = XFS_ILOG_CORE | XFS_ILOG_DEXT;
		else {
			rval = XFS_ILOG_CORE;
			if ((error = xfs_bmbt_lookup_eq(cur, PREV.br_startoff,
					PREV.br_startblock, PREV.br_blockcount,
					&i)))
				goto done;
2555
			XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
2556 2557 2558 2559 2560 2561 2562 2563 2564
			if ((error = xfs_bmbt_update(cur,
				PREV.br_startoff + new->br_blockcount,
				PREV.br_startblock + new->br_blockcount,
				PREV.br_blockcount - new->br_blockcount,
				oldext)))
				goto done;
			cur->bc_rec.b = *new;
			if ((error = xfs_btree_insert(cur, &i)))
				goto done;
2565
			XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
2566 2567
		}
		break;
L
Linus Torvalds 已提交
2568

2569 2570 2571 2572 2573 2574 2575 2576 2577
	case BMAP_RIGHT_FILLING | BMAP_RIGHT_CONTIG:
		/*
		 * Setting the last part of a previous oldext extent to newext.
		 * The right neighbor is contiguous with the new allocation.
		 */
		trace_xfs_bmap_pre_update(ip, *idx, state, _THIS_IP_);
		xfs_bmbt_set_blockcount(ep,
			PREV.br_blockcount - new->br_blockcount);
		trace_xfs_bmap_post_update(ip, *idx, state, _THIS_IP_);
2578

2579
		++*idx;
L
Linus Torvalds 已提交
2580

2581 2582 2583 2584 2585
		trace_xfs_bmap_pre_update(ip, *idx, state, _THIS_IP_);
		xfs_bmbt_set_allf(xfs_iext_get_ext(ifp, *idx),
			new->br_startoff, new->br_startblock,
			new->br_blockcount + RIGHT.br_blockcount, newext);
		trace_xfs_bmap_post_update(ip, *idx, state, _THIS_IP_);
L
Linus Torvalds 已提交
2586

2587 2588 2589 2590 2591 2592 2593 2594
		if (cur == NULL)
			rval = XFS_ILOG_DEXT;
		else {
			rval = 0;
			if ((error = xfs_bmbt_lookup_eq(cur, PREV.br_startoff,
					PREV.br_startblock,
					PREV.br_blockcount, &i)))
				goto done;
2595
			XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
2596 2597 2598 2599 2600 2601 2602 2603 2604 2605 2606 2607 2608 2609
			if ((error = xfs_bmbt_update(cur, PREV.br_startoff,
				PREV.br_startblock,
				PREV.br_blockcount - new->br_blockcount,
				oldext)))
				goto done;
			if ((error = xfs_btree_increment(cur, 0, &i)))
				goto done;
			if ((error = xfs_bmbt_update(cur, new->br_startoff,
				new->br_startblock,
				new->br_blockcount + RIGHT.br_blockcount,
				newext)))
				goto done;
		}
		break;
L
Linus Torvalds 已提交
2610

2611 2612 2613 2614 2615 2616 2617 2618 2619
	case BMAP_RIGHT_FILLING:
		/*
		 * Setting the last part of a previous oldext extent to newext.
		 * The right neighbor is not contiguous.
		 */
		trace_xfs_bmap_pre_update(ip, *idx, state, _THIS_IP_);
		xfs_bmbt_set_blockcount(ep,
			PREV.br_blockcount - new->br_blockcount);
		trace_xfs_bmap_post_update(ip, *idx, state, _THIS_IP_);
L
Linus Torvalds 已提交
2620

2621 2622
		++*idx;
		xfs_iext_insert(ip, *idx, 1, new, state);
2623

2624 2625 2626 2627 2628 2629 2630 2631 2632
		ip->i_d.di_nextents++;
		if (cur == NULL)
			rval = XFS_ILOG_CORE | XFS_ILOG_DEXT;
		else {
			rval = XFS_ILOG_CORE;
			if ((error = xfs_bmbt_lookup_eq(cur, PREV.br_startoff,
					PREV.br_startblock, PREV.br_blockcount,
					&i)))
				goto done;
2633
			XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
2634 2635 2636 2637 2638 2639 2640 2641 2642
			if ((error = xfs_bmbt_update(cur, PREV.br_startoff,
				PREV.br_startblock,
				PREV.br_blockcount - new->br_blockcount,
				oldext)))
				goto done;
			if ((error = xfs_bmbt_lookup_eq(cur, new->br_startoff,
					new->br_startblock, new->br_blockcount,
					&i)))
				goto done;
2643
			XFS_WANT_CORRUPTED_GOTO(mp, i == 0, done);
2644 2645 2646
			cur->bc_rec.b.br_state = XFS_EXT_NORM;
			if ((error = xfs_btree_insert(cur, &i)))
				goto done;
2647
			XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
2648 2649 2650 2651
		}
		break;

	case 0:
L
Linus Torvalds 已提交
2652
		/*
2653 2654 2655
		 * Setting the middle part of a previous oldext extent to
		 * newext.  Contiguity is impossible here.
		 * One extent becomes three extents.
L
Linus Torvalds 已提交
2656
		 */
2657 2658 2659 2660
		trace_xfs_bmap_pre_update(ip, *idx, state, _THIS_IP_);
		xfs_bmbt_set_blockcount(ep,
			new->br_startoff - PREV.br_startoff);
		trace_xfs_bmap_post_update(ip, *idx, state, _THIS_IP_);
2661

2662 2663 2664 2665 2666 2667
		r[0] = *new;
		r[1].br_startoff = new_endoff;
		r[1].br_blockcount =
			PREV.br_startoff + PREV.br_blockcount - new_endoff;
		r[1].br_startblock = new->br_startblock + new->br_blockcount;
		r[1].br_state = oldext;
2668

2669 2670 2671 2672 2673 2674 2675 2676 2677 2678 2679 2680
		++*idx;
		xfs_iext_insert(ip, *idx, 2, &r[0], state);

		ip->i_d.di_nextents += 2;
		if (cur == NULL)
			rval = XFS_ILOG_CORE | XFS_ILOG_DEXT;
		else {
			rval = XFS_ILOG_CORE;
			if ((error = xfs_bmbt_lookup_eq(cur, PREV.br_startoff,
					PREV.br_startblock, PREV.br_blockcount,
					&i)))
				goto done;
2681
			XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
2682 2683 2684 2685 2686 2687 2688 2689 2690 2691 2692
			/* new right extent - oldext */
			if ((error = xfs_bmbt_update(cur, r[1].br_startoff,
				r[1].br_startblock, r[1].br_blockcount,
				r[1].br_state)))
				goto done;
			/* new left extent - oldext */
			cur->bc_rec.b = PREV;
			cur->bc_rec.b.br_blockcount =
				new->br_startoff - PREV.br_startoff;
			if ((error = xfs_btree_insert(cur, &i)))
				goto done;
2693
			XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
2694 2695 2696 2697 2698 2699 2700 2701 2702
			/*
			 * Reset the cursor to the position of the new extent
			 * we are about to insert as we can't trust it after
			 * the previous insert.
			 */
			if ((error = xfs_bmbt_lookup_eq(cur, new->br_startoff,
					new->br_startblock, new->br_blockcount,
					&i)))
				goto done;
2703
			XFS_WANT_CORRUPTED_GOTO(mp, i == 0, done);
2704 2705 2706 2707
			/* new middle extent - newext */
			cur->bc_rec.b.br_state = new->br_state;
			if ((error = xfs_btree_insert(cur, &i)))
				goto done;
2708
			XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
2709
		}
L
Linus Torvalds 已提交
2710
		break;
2711 2712 2713 2714 2715 2716 2717 2718 2719 2720 2721

	case BMAP_LEFT_FILLING | BMAP_LEFT_CONTIG | BMAP_RIGHT_CONTIG:
	case BMAP_RIGHT_FILLING | BMAP_LEFT_CONTIG | BMAP_RIGHT_CONTIG:
	case BMAP_LEFT_FILLING | BMAP_RIGHT_CONTIG:
	case BMAP_RIGHT_FILLING | BMAP_LEFT_CONTIG:
	case BMAP_LEFT_CONTIG | BMAP_RIGHT_CONTIG:
	case BMAP_LEFT_CONTIG:
	case BMAP_RIGHT_CONTIG:
		/*
		 * These cases are all impossible.
		 */
L
Linus Torvalds 已提交
2722 2723
		ASSERT(0);
	}
2724

2725 2726 2727 2728 2729
	/* update reverse mappings */
	error = xfs_rmap_convert_extent(mp, dfops, ip, XFS_DATA_FORK, new);
	if (error)
		goto done;

2730 2731 2732 2733 2734
	/* convert to a btree if necessary */
	if (xfs_bmap_needs_btree(ip, XFS_DATA_FORK)) {
		int	tmp_logflags;	/* partial log flag return val */

		ASSERT(cur == NULL);
2735
		error = xfs_bmap_extents_to_btree(tp, ip, first, dfops, &cur,
2736 2737 2738 2739
				0, &tmp_logflags, XFS_DATA_FORK);
		*logflagsp |= tmp_logflags;
		if (error)
			goto done;
L
Linus Torvalds 已提交
2740
	}
2741

2742 2743 2744 2745
	/* clear out the allocated field, done with it now in any case. */
	if (cur) {
		cur->bc_private.b.allocated = 0;
		*curp = cur;
L
Linus Torvalds 已提交
2746
	}
2747

2748 2749 2750
	xfs_bmap_check_leaf_extents(*curp, ip, XFS_DATA_FORK);
done:
	*logflagsp |= rval;
L
Linus Torvalds 已提交
2751
	return error;
2752 2753 2754
#undef	LEFT
#undef	RIGHT
#undef	PREV
L
Linus Torvalds 已提交
2755 2756 2757
}

/*
2758
 * Convert a hole to a delayed allocation.
L
Linus Torvalds 已提交
2759
 */
2760 2761 2762 2763 2764
STATIC void
xfs_bmap_add_extent_hole_delay(
	xfs_inode_t		*ip,	/* incore inode pointer */
	xfs_extnum_t		*idx,	/* extent number to update/insert */
	xfs_bmbt_irec_t		*new)	/* new data to add to file extents */
L
Linus Torvalds 已提交
2765
{
2766 2767 2768 2769 2770 2771 2772
	xfs_ifork_t		*ifp;	/* inode fork pointer */
	xfs_bmbt_irec_t		left;	/* left neighbor extent entry */
	xfs_filblks_t		newlen=0;	/* new indirect size */
	xfs_filblks_t		oldlen=0;	/* old indirect size */
	xfs_bmbt_irec_t		right;	/* right neighbor extent entry */
	int			state;  /* state bits, accessed thru macros */
	xfs_filblks_t		temp=0;	/* temp for indirect calculations */
L
Linus Torvalds 已提交
2773

2774 2775 2776
	ifp = XFS_IFORK_PTR(ip, XFS_DATA_FORK);
	state = 0;
	ASSERT(isnullstartblock(new->br_startblock));
L
Linus Torvalds 已提交
2777 2778

	/*
2779
	 * Check and set flags if this segment has a left neighbor
L
Linus Torvalds 已提交
2780
	 */
2781 2782 2783 2784 2785 2786
	if (*idx > 0) {
		state |= BMAP_LEFT_VALID;
		xfs_bmbt_get_all(xfs_iext_get_ext(ifp, *idx - 1), &left);

		if (isnullstartblock(left.br_startblock))
			state |= BMAP_LEFT_DELAY;
L
Linus Torvalds 已提交
2787 2788
	}

2789 2790 2791 2792 2793 2794 2795
	/*
	 * Check and set flags if the current (right) segment exists.
	 * If it doesn't exist, we're converting the hole at end-of-file.
	 */
	if (*idx < ip->i_df.if_bytes / (uint)sizeof(xfs_bmbt_rec_t)) {
		state |= BMAP_RIGHT_VALID;
		xfs_bmbt_get_all(xfs_iext_get_ext(ifp, *idx), &right);
L
Linus Torvalds 已提交
2796

2797 2798
		if (isnullstartblock(right.br_startblock))
			state |= BMAP_RIGHT_DELAY;
L
Linus Torvalds 已提交
2799
	}
2800

L
Linus Torvalds 已提交
2801
	/*
2802 2803
	 * Set contiguity flags on the left and right neighbors.
	 * Don't let extents get too large, even if the pieces are contiguous.
L
Linus Torvalds 已提交
2804
	 */
2805 2806 2807 2808 2809 2810 2811 2812 2813 2814 2815 2816
	if ((state & BMAP_LEFT_VALID) && (state & BMAP_LEFT_DELAY) &&
	    left.br_startoff + left.br_blockcount == new->br_startoff &&
	    left.br_blockcount + new->br_blockcount <= MAXEXTLEN)
		state |= BMAP_LEFT_CONTIG;

	if ((state & BMAP_RIGHT_VALID) && (state & BMAP_RIGHT_DELAY) &&
	    new->br_startoff + new->br_blockcount == right.br_startoff &&
	    new->br_blockcount + right.br_blockcount <= MAXEXTLEN &&
	    (!(state & BMAP_LEFT_CONTIG) ||
	     (left.br_blockcount + new->br_blockcount +
	      right.br_blockcount <= MAXEXTLEN)))
		state |= BMAP_RIGHT_CONTIG;
2817 2818

	/*
2819
	 * Switch out based on the contiguity flags.
2820
	 */
2821 2822 2823 2824 2825 2826 2827 2828 2829 2830
	switch (state & (BMAP_LEFT_CONTIG | BMAP_RIGHT_CONTIG)) {
	case BMAP_LEFT_CONTIG | BMAP_RIGHT_CONTIG:
		/*
		 * New allocation is contiguous with delayed allocations
		 * on the left and on the right.
		 * Merge all three into a single extent record.
		 */
		--*idx;
		temp = left.br_blockcount + new->br_blockcount +
			right.br_blockcount;
2831

2832 2833 2834 2835 2836 2837 2838 2839 2840
		trace_xfs_bmap_pre_update(ip, *idx, state, _THIS_IP_);
		xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, *idx), temp);
		oldlen = startblockval(left.br_startblock) +
			startblockval(new->br_startblock) +
			startblockval(right.br_startblock);
		newlen = xfs_bmap_worst_indlen(ip, temp);
		xfs_bmbt_set_startblock(xfs_iext_get_ext(ifp, *idx),
			nullstartblock((int)newlen));
		trace_xfs_bmap_post_update(ip, *idx, state, _THIS_IP_);
L
Linus Torvalds 已提交
2841

2842 2843
		xfs_iext_remove(ip, *idx + 1, 1, state);
		break;
L
Linus Torvalds 已提交
2844

2845 2846 2847 2848 2849 2850 2851 2852
	case BMAP_LEFT_CONTIG:
		/*
		 * New allocation is contiguous with a delayed allocation
		 * on the left.
		 * Merge the new allocation with the left neighbor.
		 */
		--*idx;
		temp = left.br_blockcount + new->br_blockcount;
L
Linus Torvalds 已提交
2853

2854 2855 2856 2857 2858 2859 2860 2861 2862
		trace_xfs_bmap_pre_update(ip, *idx, state, _THIS_IP_);
		xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, *idx), temp);
		oldlen = startblockval(left.br_startblock) +
			startblockval(new->br_startblock);
		newlen = xfs_bmap_worst_indlen(ip, temp);
		xfs_bmbt_set_startblock(xfs_iext_get_ext(ifp, *idx),
			nullstartblock((int)newlen));
		trace_xfs_bmap_post_update(ip, *idx, state, _THIS_IP_);
		break;
L
Linus Torvalds 已提交
2863

2864 2865 2866 2867 2868 2869 2870 2871 2872 2873 2874 2875 2876 2877 2878 2879 2880 2881 2882 2883 2884 2885 2886 2887 2888 2889
	case BMAP_RIGHT_CONTIG:
		/*
		 * New allocation is contiguous with a delayed allocation
		 * on the right.
		 * Merge the new allocation with the right neighbor.
		 */
		trace_xfs_bmap_pre_update(ip, *idx, state, _THIS_IP_);
		temp = new->br_blockcount + right.br_blockcount;
		oldlen = startblockval(new->br_startblock) +
			startblockval(right.br_startblock);
		newlen = xfs_bmap_worst_indlen(ip, temp);
		xfs_bmbt_set_allf(xfs_iext_get_ext(ifp, *idx),
			new->br_startoff,
			nullstartblock((int)newlen), temp, right.br_state);
		trace_xfs_bmap_post_update(ip, *idx, state, _THIS_IP_);
		break;

	case 0:
		/*
		 * New allocation is not contiguous with another
		 * delayed allocation.
		 * Insert a new entry.
		 */
		oldlen = newlen = 0;
		xfs_iext_insert(ip, *idx, 1, new, state);
		break;
L
Linus Torvalds 已提交
2890
	}
2891 2892
	if (oldlen != newlen) {
		ASSERT(oldlen > newlen);
2893 2894
		xfs_mod_fdblocks(ip->i_mount, (int64_t)(oldlen - newlen),
				 false);
L
Linus Torvalds 已提交
2895
		/*
2896
		 * Nothing to do for disk quota accounting here.
L
Linus Torvalds 已提交
2897 2898 2899 2900 2901
		 */
	}
}

/*
2902
 * Convert a hole to a real allocation.
L
Linus Torvalds 已提交
2903
 */
2904 2905 2906 2907
STATIC int				/* error */
xfs_bmap_add_extent_hole_real(
	struct xfs_bmalloca	*bma,
	int			whichfork)
2908
{
2909 2910 2911 2912 2913 2914 2915 2916
	struct xfs_bmbt_irec	*new = &bma->got;
	int			error;	/* error return value */
	int			i;	/* temp state */
	xfs_ifork_t		*ifp;	/* inode fork pointer */
	xfs_bmbt_irec_t		left;	/* left neighbor extent entry */
	xfs_bmbt_irec_t		right;	/* right neighbor extent entry */
	int			rval=0;	/* return value (logging flags) */
	int			state;	/* state bits, accessed thru macros */
2917
	struct xfs_mount	*mp;
2918

2919
	mp = bma->ip->i_mount;
2920
	ifp = XFS_IFORK_PTR(bma->ip, whichfork);
2921

2922 2923 2924 2925 2926
	ASSERT(bma->idx >= 0);
	ASSERT(bma->idx <= ifp->if_bytes / sizeof(struct xfs_bmbt_rec));
	ASSERT(!isnullstartblock(new->br_startblock));
	ASSERT(!bma->cur ||
	       !(bma->cur->bc_private.b.flags & XFS_BTCUR_BPRV_WASDEL));
2927

2928
	XFS_STATS_INC(mp, xs_add_exlist);
2929

2930 2931 2932
	state = 0;
	if (whichfork == XFS_ATTR_FORK)
		state |= BMAP_ATTRFORK;
2933

2934 2935 2936 2937 2938 2939 2940 2941 2942
	/*
	 * Check and set flags if this segment has a left neighbor.
	 */
	if (bma->idx > 0) {
		state |= BMAP_LEFT_VALID;
		xfs_bmbt_get_all(xfs_iext_get_ext(ifp, bma->idx - 1), &left);
		if (isnullstartblock(left.br_startblock))
			state |= BMAP_LEFT_DELAY;
	}
2943 2944

	/*
2945 2946
	 * Check and set flags if this segment has a current value.
	 * Not true if we're inserting into the "hole" at eof.
2947
	 */
2948 2949 2950 2951 2952 2953
	if (bma->idx < ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t)) {
		state |= BMAP_RIGHT_VALID;
		xfs_bmbt_get_all(xfs_iext_get_ext(ifp, bma->idx), &right);
		if (isnullstartblock(right.br_startblock))
			state |= BMAP_RIGHT_DELAY;
	}
2954

2955 2956 2957 2958 2959 2960 2961 2962 2963 2964
	/*
	 * We're inserting a real allocation between "left" and "right".
	 * Set the contiguity flags.  Don't let extents get too large.
	 */
	if ((state & BMAP_LEFT_VALID) && !(state & BMAP_LEFT_DELAY) &&
	    left.br_startoff + left.br_blockcount == new->br_startoff &&
	    left.br_startblock + left.br_blockcount == new->br_startblock &&
	    left.br_state == new->br_state &&
	    left.br_blockcount + new->br_blockcount <= MAXEXTLEN)
		state |= BMAP_LEFT_CONTIG;
2965

2966 2967 2968 2969 2970 2971 2972 2973 2974
	if ((state & BMAP_RIGHT_VALID) && !(state & BMAP_RIGHT_DELAY) &&
	    new->br_startoff + new->br_blockcount == right.br_startoff &&
	    new->br_startblock + new->br_blockcount == right.br_startblock &&
	    new->br_state == right.br_state &&
	    new->br_blockcount + right.br_blockcount <= MAXEXTLEN &&
	    (!(state & BMAP_LEFT_CONTIG) ||
	     left.br_blockcount + new->br_blockcount +
	     right.br_blockcount <= MAXEXTLEN))
		state |= BMAP_RIGHT_CONTIG;
2975

2976 2977 2978 2979 2980 2981 2982 2983 2984 2985 2986 2987 2988 2989 2990 2991 2992
	error = 0;
	/*
	 * Select which case we're in here, and implement it.
	 */
	switch (state & (BMAP_LEFT_CONTIG | BMAP_RIGHT_CONTIG)) {
	case BMAP_LEFT_CONTIG | BMAP_RIGHT_CONTIG:
		/*
		 * New allocation is contiguous with real allocations on the
		 * left and on the right.
		 * Merge all three into a single extent record.
		 */
		--bma->idx;
		trace_xfs_bmap_pre_update(bma->ip, bma->idx, state, _THIS_IP_);
		xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, bma->idx),
			left.br_blockcount + new->br_blockcount +
			right.br_blockcount);
		trace_xfs_bmap_post_update(bma->ip, bma->idx, state, _THIS_IP_);
2993

2994
		xfs_iext_remove(bma->ip, bma->idx + 1, 1, state);
2995

2996 2997 2998 2999 3000 3001 3002 3003 3004 3005 3006
		XFS_IFORK_NEXT_SET(bma->ip, whichfork,
			XFS_IFORK_NEXTENTS(bma->ip, whichfork) - 1);
		if (bma->cur == NULL) {
			rval = XFS_ILOG_CORE | xfs_ilog_fext(whichfork);
		} else {
			rval = XFS_ILOG_CORE;
			error = xfs_bmbt_lookup_eq(bma->cur, right.br_startoff,
					right.br_startblock, right.br_blockcount,
					&i);
			if (error)
				goto done;
3007
			XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
3008 3009 3010
			error = xfs_btree_delete(bma->cur, &i);
			if (error)
				goto done;
3011
			XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
3012 3013 3014
			error = xfs_btree_decrement(bma->cur, 0, &i);
			if (error)
				goto done;
3015
			XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
3016 3017 3018 3019 3020 3021 3022 3023 3024 3025
			error = xfs_bmbt_update(bma->cur, left.br_startoff,
					left.br_startblock,
					left.br_blockcount +
						new->br_blockcount +
						right.br_blockcount,
					left.br_state);
			if (error)
				goto done;
		}
		break;
3026

3027 3028 3029 3030 3031 3032 3033 3034 3035 3036 3037
	case BMAP_LEFT_CONTIG:
		/*
		 * New allocation is contiguous with a real allocation
		 * on the left.
		 * Merge the new allocation with the left neighbor.
		 */
		--bma->idx;
		trace_xfs_bmap_pre_update(bma->ip, bma->idx, state, _THIS_IP_);
		xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, bma->idx),
			left.br_blockcount + new->br_blockcount);
		trace_xfs_bmap_post_update(bma->ip, bma->idx, state, _THIS_IP_);
L
Linus Torvalds 已提交
3038

3039 3040 3041 3042 3043 3044 3045 3046 3047
		if (bma->cur == NULL) {
			rval = xfs_ilog_fext(whichfork);
		} else {
			rval = 0;
			error = xfs_bmbt_lookup_eq(bma->cur, left.br_startoff,
					left.br_startblock, left.br_blockcount,
					&i);
			if (error)
				goto done;
3048
			XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
3049 3050 3051 3052 3053 3054 3055 3056 3057
			error = xfs_bmbt_update(bma->cur, left.br_startoff,
					left.br_startblock,
					left.br_blockcount +
						new->br_blockcount,
					left.br_state);
			if (error)
				goto done;
		}
		break;
3058

3059 3060 3061 3062 3063 3064 3065 3066 3067 3068 3069 3070
	case BMAP_RIGHT_CONTIG:
		/*
		 * New allocation is contiguous with a real allocation
		 * on the right.
		 * Merge the new allocation with the right neighbor.
		 */
		trace_xfs_bmap_pre_update(bma->ip, bma->idx, state, _THIS_IP_);
		xfs_bmbt_set_allf(xfs_iext_get_ext(ifp, bma->idx),
			new->br_startoff, new->br_startblock,
			new->br_blockcount + right.br_blockcount,
			right.br_state);
		trace_xfs_bmap_post_update(bma->ip, bma->idx, state, _THIS_IP_);
3071

3072 3073 3074 3075 3076 3077 3078 3079 3080 3081
		if (bma->cur == NULL) {
			rval = xfs_ilog_fext(whichfork);
		} else {
			rval = 0;
			error = xfs_bmbt_lookup_eq(bma->cur,
					right.br_startoff,
					right.br_startblock,
					right.br_blockcount, &i);
			if (error)
				goto done;
3082
			XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
3083 3084 3085 3086 3087 3088 3089 3090 3091 3092 3093 3094 3095 3096 3097 3098 3099 3100 3101 3102 3103 3104 3105 3106 3107 3108 3109 3110 3111
			error = xfs_bmbt_update(bma->cur, new->br_startoff,
					new->br_startblock,
					new->br_blockcount +
						right.br_blockcount,
					right.br_state);
			if (error)
				goto done;
		}
		break;

	case 0:
		/*
		 * New allocation is not contiguous with another
		 * real allocation.
		 * Insert a new entry.
		 */
		xfs_iext_insert(bma->ip, bma->idx, 1, new, state);
		XFS_IFORK_NEXT_SET(bma->ip, whichfork,
			XFS_IFORK_NEXTENTS(bma->ip, whichfork) + 1);
		if (bma->cur == NULL) {
			rval = XFS_ILOG_CORE | xfs_ilog_fext(whichfork);
		} else {
			rval = XFS_ILOG_CORE;
			error = xfs_bmbt_lookup_eq(bma->cur,
					new->br_startoff,
					new->br_startblock,
					new->br_blockcount, &i);
			if (error)
				goto done;
3112
			XFS_WANT_CORRUPTED_GOTO(mp, i == 0, done);
3113 3114 3115 3116
			bma->cur->bc_rec.b.br_state = new->br_state;
			error = xfs_btree_insert(bma->cur, &i);
			if (error)
				goto done;
3117
			XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
3118 3119 3120 3121
		}
		break;
	}

3122 3123 3124 3125 3126
	/* add reverse mapping */
	error = xfs_rmap_map_extent(mp, bma->dfops, bma->ip, whichfork, new);
	if (error)
		goto done;

3127 3128 3129 3130 3131 3132
	/* convert to a btree if necessary */
	if (xfs_bmap_needs_btree(bma->ip, whichfork)) {
		int	tmp_logflags;	/* partial log flag return val */

		ASSERT(bma->cur == NULL);
		error = xfs_bmap_extents_to_btree(bma->tp, bma->ip,
3133
				bma->firstblock, bma->dfops, &bma->cur,
3134 3135 3136 3137 3138 3139 3140 3141 3142 3143 3144 3145 3146 3147
				0, &tmp_logflags, whichfork);
		bma->logflags |= tmp_logflags;
		if (error)
			goto done;
	}

	/* clear out the allocated field, done with it now in any case. */
	if (bma->cur)
		bma->cur->bc_private.b.allocated = 0;

	xfs_bmap_check_leaf_extents(bma->cur, bma->ip, whichfork);
done:
	bma->logflags |= rval;
	return error;
L
Linus Torvalds 已提交
3148 3149 3150
}

/*
3151
 * Functions used in the extent read, allocate and remove paths
L
Linus Torvalds 已提交
3152 3153
 */

3154 3155 3156
/*
 * Adjust the size of the new extent based on di_extsize and rt extsize.
 */
D
Dave Chinner 已提交
3157
int
3158 3159 3160 3161 3162 3163 3164 3165 3166 3167 3168
xfs_bmap_extsize_align(
	xfs_mount_t	*mp,
	xfs_bmbt_irec_t	*gotp,		/* next extent pointer */
	xfs_bmbt_irec_t	*prevp,		/* previous extent pointer */
	xfs_extlen_t	extsz,		/* align to this extent size */
	int		rt,		/* is this a realtime inode? */
	int		eof,		/* is extent at end-of-file? */
	int		delay,		/* creating delalloc extent? */
	int		convert,	/* overwriting unwritten extent? */
	xfs_fileoff_t	*offp,		/* in/out: aligned offset */
	xfs_extlen_t	*lenp)		/* in/out: aligned length */
3169
{
3170 3171 3172 3173 3174 3175 3176 3177
	xfs_fileoff_t	orig_off;	/* original offset */
	xfs_extlen_t	orig_alen;	/* original length */
	xfs_fileoff_t	orig_end;	/* original off+len */
	xfs_fileoff_t	nexto;		/* next file offset */
	xfs_fileoff_t	prevo;		/* previous file offset */
	xfs_fileoff_t	align_off;	/* temp for offset */
	xfs_extlen_t	align_alen;	/* temp for length */
	xfs_extlen_t	temp;		/* temp for calculations */
3178

3179
	if (convert)
3180 3181
		return 0;

3182 3183 3184
	orig_off = align_off = *offp;
	orig_alen = align_alen = *lenp;
	orig_end = orig_off + orig_alen;
L
Linus Torvalds 已提交
3185 3186

	/*
3187 3188
	 * If this request overlaps an existing extent, then don't
	 * attempt to perform any additional alignment.
L
Linus Torvalds 已提交
3189
	 */
3190 3191 3192 3193 3194 3195
	if (!delay && !eof &&
	    (orig_off >= gotp->br_startoff) &&
	    (orig_end <= gotp->br_startoff + gotp->br_blockcount)) {
		return 0;
	}

L
Linus Torvalds 已提交
3196
	/*
3197 3198 3199 3200 3201
	 * If the file offset is unaligned vs. the extent size
	 * we need to align it.  This will be possible unless
	 * the file was previously written with a kernel that didn't
	 * perform this alignment, or if a truncate shot us in the
	 * foot.
L
Linus Torvalds 已提交
3202
	 */
3203 3204 3205 3206
	temp = do_mod(orig_off, extsz);
	if (temp) {
		align_alen += temp;
		align_off -= temp;
L
Linus Torvalds 已提交
3207
	}
3208 3209 3210 3211 3212 3213

	/* Same adjustment for the end of the requested area. */
	temp = (align_alen % extsz);
	if (temp)
		align_alen += extsz - temp;

L
Linus Torvalds 已提交
3214
	/*
3215 3216 3217 3218 3219 3220
	 * For large extent hint sizes, the aligned extent might be larger than
	 * MAXEXTLEN. In that case, reduce the size by an extsz so that it pulls
	 * the length back under MAXEXTLEN. The outer allocation loops handle
	 * short allocation just fine, so it is safe to do this. We only want to
	 * do it when we are forced to, though, because it means more allocation
	 * operations are required.
L
Linus Torvalds 已提交
3221
	 */
3222 3223 3224 3225
	while (align_alen > MAXEXTLEN)
		align_alen -= extsz;
	ASSERT(align_alen <= MAXEXTLEN);

L
Linus Torvalds 已提交
3226
	/*
3227 3228
	 * If the previous block overlaps with this proposed allocation
	 * then move the start forward without adjusting the length.
L
Linus Torvalds 已提交
3229
	 */
3230 3231 3232 3233 3234 3235 3236 3237 3238 3239 3240 3241 3242 3243 3244 3245 3246 3247 3248 3249 3250 3251 3252 3253 3254 3255 3256 3257 3258 3259 3260 3261 3262 3263 3264 3265 3266 3267 3268 3269 3270 3271 3272
	if (prevp->br_startoff != NULLFILEOFF) {
		if (prevp->br_startblock == HOLESTARTBLOCK)
			prevo = prevp->br_startoff;
		else
			prevo = prevp->br_startoff + prevp->br_blockcount;
	} else
		prevo = 0;
	if (align_off != orig_off && align_off < prevo)
		align_off = prevo;
	/*
	 * If the next block overlaps with this proposed allocation
	 * then move the start back without adjusting the length,
	 * but not before offset 0.
	 * This may of course make the start overlap previous block,
	 * and if we hit the offset 0 limit then the next block
	 * can still overlap too.
	 */
	if (!eof && gotp->br_startoff != NULLFILEOFF) {
		if ((delay && gotp->br_startblock == HOLESTARTBLOCK) ||
		    (!delay && gotp->br_startblock == DELAYSTARTBLOCK))
			nexto = gotp->br_startoff + gotp->br_blockcount;
		else
			nexto = gotp->br_startoff;
	} else
		nexto = NULLFILEOFF;
	if (!eof &&
	    align_off + align_alen != orig_end &&
	    align_off + align_alen > nexto)
		align_off = nexto > align_alen ? nexto - align_alen : 0;
	/*
	 * If we're now overlapping the next or previous extent that
	 * means we can't fit an extsz piece in this hole.  Just move
	 * the start forward to the first valid spot and set
	 * the length so we hit the end.
	 */
	if (align_off != orig_off && align_off < prevo)
		align_off = prevo;
	if (align_off + align_alen != orig_end &&
	    align_off + align_alen > nexto &&
	    nexto != NULLFILEOFF) {
		ASSERT(nexto > prevo);
		align_alen = nexto - align_off;
	}
L
Linus Torvalds 已提交
3273

3274 3275 3276 3277 3278
	/*
	 * If realtime, and the result isn't a multiple of the realtime
	 * extent size we need to remove blocks until it is.
	 */
	if (rt && (temp = (align_alen % mp->m_sb.sb_rextsize))) {
L
Linus Torvalds 已提交
3279
		/*
3280 3281
		 * We're not covering the original request, or
		 * we won't be able to once we fix the length.
L
Linus Torvalds 已提交
3282
		 */
3283 3284 3285
		if (orig_off < align_off ||
		    orig_end > align_off + align_alen ||
		    align_alen - temp < orig_alen)
D
Dave Chinner 已提交
3286
			return -EINVAL;
L
Linus Torvalds 已提交
3287
		/*
3288
		 * Try to fix it by moving the start up.
L
Linus Torvalds 已提交
3289
		 */
3290 3291 3292
		if (align_off + temp <= orig_off) {
			align_alen -= temp;
			align_off += temp;
L
Linus Torvalds 已提交
3293
		}
3294 3295 3296 3297 3298 3299 3300 3301 3302 3303 3304 3305
		/*
		 * Try to fix it by moving the end in.
		 */
		else if (align_off + align_alen - temp >= orig_end)
			align_alen -= temp;
		/*
		 * Set the start to the minimum then trim the length.
		 */
		else {
			align_alen -= orig_off - align_off;
			align_off = orig_off;
			align_alen -= align_alen % mp->m_sb.sb_rextsize;
L
Linus Torvalds 已提交
3306 3307
		}
		/*
3308
		 * Result doesn't cover the request, fail it.
L
Linus Torvalds 已提交
3309
		 */
3310
		if (orig_off < align_off || orig_end > align_off + align_alen)
D
Dave Chinner 已提交
3311
			return -EINVAL;
3312 3313
	} else {
		ASSERT(orig_off >= align_off);
3314 3315 3316
		/* see MAXEXTLEN handling above */
		ASSERT(orig_end <= align_off + align_alen ||
		       align_alen + extsz > MAXEXTLEN);
L
Linus Torvalds 已提交
3317 3318
	}

C
Christoph Hellwig 已提交
3319
#ifdef DEBUG
3320 3321 3322 3323 3324
	if (!eof && gotp->br_startoff != NULLFILEOFF)
		ASSERT(align_off + align_alen <= gotp->br_startoff);
	if (prevp->br_startoff != NULLFILEOFF)
		ASSERT(align_off >= prevp->br_startoff + prevp->br_blockcount);
#endif
L
Linus Torvalds 已提交
3325

3326 3327 3328
	*lenp = align_alen;
	*offp = align_off;
	return 0;
L
Linus Torvalds 已提交
3329 3330
}

3331 3332
#define XFS_ALLOC_GAP_UNITS	4

D
Dave Chinner 已提交
3333
void
3334
xfs_bmap_adjacent(
D
Dave Chinner 已提交
3335
	struct xfs_bmalloca	*ap)	/* bmap alloc argument struct */
L
Linus Torvalds 已提交
3336
{
3337 3338 3339 3340 3341
	xfs_fsblock_t	adjust;		/* adjustment to block numbers */
	xfs_agnumber_t	fb_agno;	/* ag number of ap->firstblock */
	xfs_mount_t	*mp;		/* mount point structure */
	int		nullfb;		/* true if ap->firstblock isn't set */
	int		rt;		/* true if inode is realtime */
L
Linus Torvalds 已提交
3342

3343 3344 3345 3346 3347 3348
#define	ISVALID(x,y)	\
	(rt ? \
		(x) < mp->m_sb.sb_rblocks : \
		XFS_FSB_TO_AGNO(mp, x) == XFS_FSB_TO_AGNO(mp, y) && \
		XFS_FSB_TO_AGNO(mp, x) < mp->m_sb.sb_agcount && \
		XFS_FSB_TO_AGBNO(mp, x) < mp->m_sb.sb_agblocks)
L
Linus Torvalds 已提交
3349

3350 3351
	mp = ap->ip->i_mount;
	nullfb = *ap->firstblock == NULLFSBLOCK;
3352 3353
	rt = XFS_IS_REALTIME_INODE(ap->ip) &&
		xfs_alloc_is_userdata(ap->datatype);
3354 3355 3356 3357 3358 3359 3360 3361 3362 3363 3364 3365 3366 3367 3368 3369 3370 3371 3372 3373 3374 3375 3376 3377 3378 3379 3380 3381 3382 3383 3384 3385 3386 3387 3388 3389 3390 3391 3392 3393 3394 3395 3396 3397 3398 3399 3400 3401 3402 3403 3404 3405 3406 3407 3408 3409 3410 3411 3412 3413 3414 3415 3416 3417 3418 3419
	fb_agno = nullfb ? NULLAGNUMBER : XFS_FSB_TO_AGNO(mp, *ap->firstblock);
	/*
	 * If allocating at eof, and there's a previous real block,
	 * try to use its last block as our starting point.
	 */
	if (ap->eof && ap->prev.br_startoff != NULLFILEOFF &&
	    !isnullstartblock(ap->prev.br_startblock) &&
	    ISVALID(ap->prev.br_startblock + ap->prev.br_blockcount,
		    ap->prev.br_startblock)) {
		ap->blkno = ap->prev.br_startblock + ap->prev.br_blockcount;
		/*
		 * Adjust for the gap between prevp and us.
		 */
		adjust = ap->offset -
			(ap->prev.br_startoff + ap->prev.br_blockcount);
		if (adjust &&
		    ISVALID(ap->blkno + adjust, ap->prev.br_startblock))
			ap->blkno += adjust;
	}
	/*
	 * If not at eof, then compare the two neighbor blocks.
	 * Figure out whether either one gives us a good starting point,
	 * and pick the better one.
	 */
	else if (!ap->eof) {
		xfs_fsblock_t	gotbno;		/* right side block number */
		xfs_fsblock_t	gotdiff=0;	/* right side difference */
		xfs_fsblock_t	prevbno;	/* left side block number */
		xfs_fsblock_t	prevdiff=0;	/* left side difference */

		/*
		 * If there's a previous (left) block, select a requested
		 * start block based on it.
		 */
		if (ap->prev.br_startoff != NULLFILEOFF &&
		    !isnullstartblock(ap->prev.br_startblock) &&
		    (prevbno = ap->prev.br_startblock +
			       ap->prev.br_blockcount) &&
		    ISVALID(prevbno, ap->prev.br_startblock)) {
			/*
			 * Calculate gap to end of previous block.
			 */
			adjust = prevdiff = ap->offset -
				(ap->prev.br_startoff +
				 ap->prev.br_blockcount);
			/*
			 * Figure the startblock based on the previous block's
			 * end and the gap size.
			 * Heuristic!
			 * If the gap is large relative to the piece we're
			 * allocating, or using it gives us an invalid block
			 * number, then just use the end of the previous block.
			 */
			if (prevdiff <= XFS_ALLOC_GAP_UNITS * ap->length &&
			    ISVALID(prevbno + prevdiff,
				    ap->prev.br_startblock))
				prevbno += adjust;
			else
				prevdiff += adjust;
			/*
			 * If the firstblock forbids it, can't use it,
			 * must use default.
			 */
			if (!rt && !nullfb &&
			    XFS_FSB_TO_AGNO(mp, prevbno) != fb_agno)
				prevbno = NULLFSBLOCK;
L
Linus Torvalds 已提交
3420
		}
3421 3422 3423 3424 3425 3426 3427 3428 3429 3430 3431 3432 3433 3434 3435 3436 3437 3438 3439 3440 3441 3442 3443 3444 3445 3446 3447 3448 3449 3450 3451 3452 3453 3454 3455 3456 3457 3458 3459 3460 3461 3462 3463 3464 3465 3466 3467 3468 3469 3470 3471 3472 3473 3474 3475 3476 3477
		/*
		 * No previous block or can't follow it, just default.
		 */
		else
			prevbno = NULLFSBLOCK;
		/*
		 * If there's a following (right) block, select a requested
		 * start block based on it.
		 */
		if (!isnullstartblock(ap->got.br_startblock)) {
			/*
			 * Calculate gap to start of next block.
			 */
			adjust = gotdiff = ap->got.br_startoff - ap->offset;
			/*
			 * Figure the startblock based on the next block's
			 * start and the gap size.
			 */
			gotbno = ap->got.br_startblock;
			/*
			 * Heuristic!
			 * If the gap is large relative to the piece we're
			 * allocating, or using it gives us an invalid block
			 * number, then just use the start of the next block
			 * offset by our length.
			 */
			if (gotdiff <= XFS_ALLOC_GAP_UNITS * ap->length &&
			    ISVALID(gotbno - gotdiff, gotbno))
				gotbno -= adjust;
			else if (ISVALID(gotbno - ap->length, gotbno)) {
				gotbno -= ap->length;
				gotdiff += adjust - ap->length;
			} else
				gotdiff += adjust;
			/*
			 * If the firstblock forbids it, can't use it,
			 * must use default.
			 */
			if (!rt && !nullfb &&
			    XFS_FSB_TO_AGNO(mp, gotbno) != fb_agno)
				gotbno = NULLFSBLOCK;
		}
		/*
		 * No next block, just default.
		 */
		else
			gotbno = NULLFSBLOCK;
		/*
		 * If both valid, pick the better one, else the only good
		 * one, else ap->blkno is already set (to 0 or the inode block).
		 */
		if (prevbno != NULLFSBLOCK && gotbno != NULLFSBLOCK)
			ap->blkno = prevdiff <= gotdiff ? prevbno : gotbno;
		else if (prevbno != NULLFSBLOCK)
			ap->blkno = prevbno;
		else if (gotbno != NULLFSBLOCK)
			ap->blkno = gotbno;
L
Linus Torvalds 已提交
3478
	}
3479
#undef ISVALID
L
Linus Torvalds 已提交
3480 3481
}

3482 3483 3484 3485 3486 3487 3488 3489 3490 3491 3492 3493 3494 3495 3496 3497 3498 3499 3500 3501 3502 3503 3504 3505
static int
xfs_bmap_longest_free_extent(
	struct xfs_trans	*tp,
	xfs_agnumber_t		ag,
	xfs_extlen_t		*blen,
	int			*notinit)
{
	struct xfs_mount	*mp = tp->t_mountp;
	struct xfs_perag	*pag;
	xfs_extlen_t		longest;
	int			error = 0;

	pag = xfs_perag_get(mp, ag);
	if (!pag->pagf_init) {
		error = xfs_alloc_pagf_init(mp, tp, ag, XFS_ALLOC_FLAG_TRYLOCK);
		if (error)
			goto out;

		if (!pag->pagf_init) {
			*notinit = 1;
			goto out;
		}
	}

3506
	longest = xfs_alloc_longest_free_extent(mp, pag,
3507 3508
				xfs_alloc_min_freelist(mp, pag),
				xfs_ag_resv_needed(pag, XFS_AG_RESV_NONE));
3509 3510 3511 3512 3513 3514 3515 3516 3517 3518 3519 3520 3521 3522 3523 3524 3525 3526 3527 3528 3529 3530 3531 3532 3533 3534 3535 3536 3537 3538 3539 3540 3541 3542 3543 3544
	if (*blen < longest)
		*blen = longest;

out:
	xfs_perag_put(pag);
	return error;
}

static void
xfs_bmap_select_minlen(
	struct xfs_bmalloca	*ap,
	struct xfs_alloc_arg	*args,
	xfs_extlen_t		*blen,
	int			notinit)
{
	if (notinit || *blen < ap->minlen) {
		/*
		 * Since we did a BUF_TRYLOCK above, it is possible that
		 * there is space for this request.
		 */
		args->minlen = ap->minlen;
	} else if (*blen < args->maxlen) {
		/*
		 * If the best seen length is less than the request length,
		 * use the best as the minimum.
		 */
		args->minlen = *blen;
	} else {
		/*
		 * Otherwise we've seen an extent as big as maxlen, use that
		 * as the minimum.
		 */
		args->minlen = args->maxlen;
	}
}

3545
STATIC int
3546 3547 3548 3549
xfs_bmap_btalloc_nullfb(
	struct xfs_bmalloca	*ap,
	struct xfs_alloc_arg	*args,
	xfs_extlen_t		*blen)
3550
{
3551 3552 3553
	struct xfs_mount	*mp = ap->ip->i_mount;
	xfs_agnumber_t		ag, startag;
	int			notinit = 0;
3554 3555
	int			error;

3556
	args->type = XFS_ALLOCTYPE_START_BNO;
3557
	args->total = ap->total;
3558

3559 3560 3561
	startag = ag = XFS_FSB_TO_AGNO(mp, args->fsbno);
	if (startag == NULLAGNUMBER)
		startag = ag = 0;
3562

3563
	while (*blen < args->maxlen) {
3564 3565 3566 3567
		error = xfs_bmap_longest_free_extent(args->tp, ag, blen,
						     &notinit);
		if (error)
			return error;
3568

3569 3570 3571 3572 3573
		if (++ag == mp->m_sb.sb_agcount)
			ag = 0;
		if (ag == startag)
			break;
	}
3574

3575 3576 3577 3578 3579 3580 3581 3582 3583 3584 3585 3586 3587 3588 3589 3590 3591 3592 3593 3594 3595 3596 3597 3598 3599 3600 3601 3602 3603 3604 3605 3606 3607 3608 3609 3610 3611 3612 3613
	xfs_bmap_select_minlen(ap, args, blen, notinit);
	return 0;
}

STATIC int
xfs_bmap_btalloc_filestreams(
	struct xfs_bmalloca	*ap,
	struct xfs_alloc_arg	*args,
	xfs_extlen_t		*blen)
{
	struct xfs_mount	*mp = ap->ip->i_mount;
	xfs_agnumber_t		ag;
	int			notinit = 0;
	int			error;

	args->type = XFS_ALLOCTYPE_NEAR_BNO;
	args->total = ap->total;

	ag = XFS_FSB_TO_AGNO(mp, args->fsbno);
	if (ag == NULLAGNUMBER)
		ag = 0;

	error = xfs_bmap_longest_free_extent(args->tp, ag, blen, &notinit);
	if (error)
		return error;

	if (*blen < args->maxlen) {
		error = xfs_filestream_new_ag(ap, &ag);
		if (error)
			return error;

		error = xfs_bmap_longest_free_extent(args->tp, ag, blen,
						     &notinit);
		if (error)
			return error;

	}

	xfs_bmap_select_minlen(ap, args, blen, notinit);
3614 3615

	/*
3616 3617
	 * Set the failure fallback case to look in the selected AG as stream
	 * may have moved.
3618
	 */
3619
	ap->blkno = args->fsbno = XFS_AGB_TO_FSB(mp, ag, 0);
3620 3621 3622
	return 0;
}

3623 3624
STATIC int
xfs_bmap_btalloc(
D
Dave Chinner 已提交
3625
	struct xfs_bmalloca	*ap)	/* bmap alloc argument struct */
3626
{
3627 3628
	xfs_mount_t	*mp;		/* mount point structure */
	xfs_alloctype_t	atype = 0;	/* type for allocation routines */
3629
	xfs_extlen_t	align = 0;	/* minimum allocation alignment */
3630 3631 3632 3633 3634 3635 3636 3637 3638
	xfs_agnumber_t	fb_agno;	/* ag number of ap->firstblock */
	xfs_agnumber_t	ag;
	xfs_alloc_arg_t	args;
	xfs_extlen_t	blen;
	xfs_extlen_t	nextminlen = 0;
	int		nullfb;		/* true if ap->firstblock isn't set */
	int		isaligned;
	int		tryagain;
	int		error;
3639
	int		stripe_align;
3640

3641
	ASSERT(ap->length);
3642

3643
	mp = ap->ip->i_mount;
3644 3645 3646 3647 3648 3649 3650 3651

	/* stripe alignment for allocation is determined by mount parameters */
	stripe_align = 0;
	if (mp->m_swidth && (mp->m_flags & XFS_MOUNT_SWALLOC))
		stripe_align = mp->m_swidth;
	else if (mp->m_dalign)
		stripe_align = mp->m_dalign;

3652 3653
	if (xfs_alloc_is_userdata(ap->datatype))
		align = xfs_get_extsz_hint(ap->ip);
3654 3655 3656 3657 3658 3659
	if (unlikely(align)) {
		error = xfs_bmap_extsize_align(mp, &ap->got, &ap->prev,
						align, 0, ap->eof, 0, ap->conv,
						&ap->offset, &ap->length);
		ASSERT(!error);
		ASSERT(ap->length);
3660
	}
3661 3662


3663 3664 3665
	nullfb = *ap->firstblock == NULLFSBLOCK;
	fb_agno = nullfb ? NULLAGNUMBER : XFS_FSB_TO_AGNO(mp, *ap->firstblock);
	if (nullfb) {
3666 3667
		if (xfs_alloc_is_userdata(ap->datatype) &&
		    xfs_inode_is_filestream(ap->ip)) {
3668 3669 3670 3671 3672 3673 3674 3675
			ag = xfs_filestream_lookup_ag(ap->ip);
			ag = (ag != NULLAGNUMBER) ? ag : 0;
			ap->blkno = XFS_AGB_TO_FSB(mp, ag, 0);
		} else {
			ap->blkno = XFS_INO_TO_FSB(mp, ap->ip->i_ino);
		}
	} else
		ap->blkno = *ap->firstblock;
3676

3677
	xfs_bmap_adjacent(ap);
3678

3679 3680 3681 3682 3683 3684 3685 3686 3687 3688 3689 3690 3691 3692 3693 3694
	/*
	 * If allowed, use ap->blkno; otherwise must use firstblock since
	 * it's in the right allocation group.
	 */
	if (nullfb || XFS_FSB_TO_AGNO(mp, ap->blkno) == fb_agno)
		;
	else
		ap->blkno = *ap->firstblock;
	/*
	 * Normal allocation, done through xfs_alloc_vextent.
	 */
	tryagain = isaligned = 0;
	memset(&args, 0, sizeof(args));
	args.tp = ap->tp;
	args.mp = mp;
	args.fsbno = ap->blkno;
3695
	xfs_rmap_skip_owner_update(&args.oinfo);
3696

3697
	/* Trim the allocation back to the maximum an AG can fit. */
3698
	args.maxlen = MIN(ap->length, mp->m_ag_max_usable);
3699 3700 3701
	args.firstblock = *ap->firstblock;
	blen = 0;
	if (nullfb) {
3702 3703 3704 3705 3706
		/*
		 * Search for an allocation group with a single extent large
		 * enough for the request.  If one isn't found, then adjust
		 * the minimum allocation size to the largest space found.
		 */
3707 3708
		if (xfs_alloc_is_userdata(ap->datatype) &&
		    xfs_inode_is_filestream(ap->ip))
3709 3710 3711
			error = xfs_bmap_btalloc_filestreams(ap, &args, &blen);
		else
			error = xfs_bmap_btalloc_nullfb(ap, &args, &blen);
3712 3713
		if (error)
			return error;
3714
	} else if (ap->dfops->dop_low) {
3715 3716 3717 3718 3719 3720 3721 3722 3723
		if (xfs_inode_is_filestream(ap->ip))
			args.type = XFS_ALLOCTYPE_FIRST_AG;
		else
			args.type = XFS_ALLOCTYPE_START_BNO;
		args.total = args.minlen = ap->minlen;
	} else {
		args.type = XFS_ALLOCTYPE_NEAR_BNO;
		args.total = ap->total;
		args.minlen = ap->minlen;
3724
	}
3725 3726 3727 3728 3729
	/* apply extent size hints if obtained earlier */
	if (unlikely(align)) {
		args.prod = align;
		if ((args.mod = (xfs_extlen_t)do_mod(ap->offset, args.prod)))
			args.mod = (xfs_extlen_t)(args.prod - args.mod);
3730
	} else if (mp->m_sb.sb_blocksize >= PAGE_SIZE) {
3731 3732 3733
		args.prod = 1;
		args.mod = 0;
	} else {
3734
		args.prod = PAGE_SIZE >> mp->m_sb.sb_blocklog;
3735 3736
		if ((args.mod = (xfs_extlen_t)(do_mod(ap->offset, args.prod))))
			args.mod = (xfs_extlen_t)(args.prod - args.mod);
3737
	}
3738
	/*
3739 3740 3741 3742 3743 3744 3745
	 * If we are not low on available data blocks, and the
	 * underlying logical volume manager is a stripe, and
	 * the file offset is zero then try to allocate data
	 * blocks on stripe unit boundary.
	 * NOTE: ap->aeof is only set if the allocation length
	 * is >= the stripe unit and the allocation offset is
	 * at the end of file.
3746
	 */
3747
	if (!ap->dfops->dop_low && ap->aeof) {
3748
		if (!ap->offset) {
3749
			args.alignment = stripe_align;
3750 3751 3752 3753 3754 3755 3756 3757 3758 3759 3760 3761 3762 3763 3764 3765 3766 3767 3768 3769 3770 3771 3772 3773
			atype = args.type;
			isaligned = 1;
			/*
			 * Adjust for alignment
			 */
			if (blen > args.alignment && blen <= args.maxlen)
				args.minlen = blen - args.alignment;
			args.minalignslop = 0;
		} else {
			/*
			 * First try an exact bno allocation.
			 * If it fails then do a near or start bno
			 * allocation with alignment turned on.
			 */
			atype = args.type;
			tryagain = 1;
			args.type = XFS_ALLOCTYPE_THIS_BNO;
			args.alignment = 1;
			/*
			 * Compute the minlen+alignment for the
			 * next case.  Set slop so that the value
			 * of minlen+alignment+slop doesn't go up
			 * between the calls.
			 */
3774 3775
			if (blen > stripe_align && blen <= args.maxlen)
				nextminlen = blen - stripe_align;
3776 3777
			else
				nextminlen = args.minlen;
3778
			if (nextminlen + stripe_align > args.minlen + 1)
3779
				args.minalignslop =
3780
					nextminlen + stripe_align -
3781 3782 3783
					args.minlen - 1;
			else
				args.minalignslop = 0;
3784 3785
		}
	} else {
3786 3787
		args.alignment = 1;
		args.minalignslop = 0;
3788
	}
3789 3790
	args.minleft = ap->minleft;
	args.wasdel = ap->wasdel;
3791
	args.resv = XFS_AG_RESV_NONE;
3792 3793
	args.datatype = ap->datatype;
	if (ap->datatype & XFS_ALLOC_USERDATA_ZERO)
3794 3795 3796 3797
		args.ip = ap->ip;

	error = xfs_alloc_vextent(&args);
	if (error)
3798
		return error;
3799

3800 3801 3802 3803 3804 3805 3806
	if (tryagain && args.fsbno == NULLFSBLOCK) {
		/*
		 * Exact allocation failed. Now try with alignment
		 * turned on.
		 */
		args.type = atype;
		args.fsbno = ap->blkno;
3807
		args.alignment = stripe_align;
3808 3809 3810 3811 3812
		args.minlen = nextminlen;
		args.minalignslop = 0;
		isaligned = 1;
		if ((error = xfs_alloc_vextent(&args)))
			return error;
3813
	}
3814 3815 3816 3817 3818 3819 3820 3821 3822
	if (isaligned && args.fsbno == NULLFSBLOCK) {
		/*
		 * allocation failed, so turn off alignment and
		 * try again.
		 */
		args.type = atype;
		args.fsbno = ap->blkno;
		args.alignment = 0;
		if ((error = xfs_alloc_vextent(&args)))
3823 3824
			return error;
	}
3825 3826 3827 3828 3829 3830 3831
	if (args.fsbno == NULLFSBLOCK && nullfb &&
	    args.minlen > ap->minlen) {
		args.minlen = ap->minlen;
		args.type = XFS_ALLOCTYPE_START_BNO;
		args.fsbno = ap->blkno;
		if ((error = xfs_alloc_vextent(&args)))
			return error;
3832
	}
3833 3834 3835 3836 3837 3838 3839
	if (args.fsbno == NULLFSBLOCK && nullfb) {
		args.fsbno = 0;
		args.type = XFS_ALLOCTYPE_FIRST_AG;
		args.total = ap->minlen;
		args.minleft = 0;
		if ((error = xfs_alloc_vextent(&args)))
			return error;
3840
		ap->dfops->dop_low = true;
3841 3842 3843 3844 3845 3846 3847 3848 3849
	}
	if (args.fsbno != NULLFSBLOCK) {
		/*
		 * check the allocation happened at the same or higher AG than
		 * the first block that was allocated.
		 */
		ASSERT(*ap->firstblock == NULLFSBLOCK ||
		       XFS_FSB_TO_AGNO(mp, *ap->firstblock) ==
		       XFS_FSB_TO_AGNO(mp, args.fsbno) ||
3850
		       (ap->dfops->dop_low &&
3851 3852
			XFS_FSB_TO_AGNO(mp, *ap->firstblock) <
			XFS_FSB_TO_AGNO(mp, args.fsbno)));
3853

3854 3855 3856 3857
		ap->blkno = args.fsbno;
		if (*ap->firstblock == NULLFSBLOCK)
			*ap->firstblock = args.fsbno;
		ASSERT(nullfb || fb_agno == args.agno ||
3858
		       (ap->dfops->dop_low && fb_agno < args.agno));
3859 3860 3861 3862 3863 3864 3865 3866 3867 3868 3869 3870 3871 3872 3873 3874 3875 3876 3877
		ap->length = args.len;
		ap->ip->i_d.di_nblocks += args.len;
		xfs_trans_log_inode(ap->tp, ap->ip, XFS_ILOG_CORE);
		if (ap->wasdel)
			ap->ip->i_delayed_blks -= args.len;
		/*
		 * Adjust the disk quota also. This was reserved
		 * earlier.
		 */
		xfs_trans_mod_dquot_byino(ap->tp, ap->ip,
			ap->wasdel ? XFS_TRANS_DQ_DELBCOUNT :
					XFS_TRANS_DQ_BCOUNT,
			(long) args.len);
	} else {
		ap->blkno = NULLFSBLOCK;
		ap->length = 0;
	}
	return 0;
}
3878

3879 3880 3881 3882 3883 3884 3885 3886 3887 3888 3889 3890 3891 3892 3893 3894 3895 3896 3897 3898 3899 3900 3901 3902 3903 3904 3905 3906 3907 3908 3909 3910 3911 3912 3913 3914 3915 3916 3917 3918 3919 3920 3921 3922 3923 3924 3925 3926 3927 3928 3929 3930 3931 3932 3933 3934 3935
/*
 * For a remap operation, just "allocate" an extent at the address that the
 * caller passed in, and ensure that the AGFL is the right size.  The caller
 * will then map the "allocated" extent into the file somewhere.
 */
STATIC int
xfs_bmap_remap_alloc(
	struct xfs_bmalloca	*ap)
{
	struct xfs_trans	*tp = ap->tp;
	struct xfs_mount	*mp = tp->t_mountp;
	xfs_agblock_t		bno;
	struct xfs_alloc_arg	args;
	int			error;

	/*
	 * validate that the block number is legal - the enables us to detect
	 * and handle a silent filesystem corruption rather than crashing.
	 */
	memset(&args, 0, sizeof(struct xfs_alloc_arg));
	args.tp = ap->tp;
	args.mp = ap->tp->t_mountp;
	bno = *ap->firstblock;
	args.agno = XFS_FSB_TO_AGNO(mp, bno);
	args.agbno = XFS_FSB_TO_AGBNO(mp, bno);
	if (args.agno >= mp->m_sb.sb_agcount ||
	    args.agbno >= mp->m_sb.sb_agblocks)
		return -EFSCORRUPTED;

	/* "Allocate" the extent from the range we passed in. */
	trace_xfs_bmap_remap_alloc(ap->ip, *ap->firstblock, ap->length);
	ap->blkno = bno;
	ap->ip->i_d.di_nblocks += ap->length;
	xfs_trans_log_inode(ap->tp, ap->ip, XFS_ILOG_CORE);

	/* Fix the freelist, like a real allocator does. */
	args.datatype = ap->datatype;
	args.pag = xfs_perag_get(args.mp, args.agno);
	ASSERT(args.pag);

	/*
	 * The freelist fixing code will decline the allocation if
	 * the size and shape of the free space doesn't allow for
	 * allocating the extent and updating all the metadata that
	 * happens during an allocation.  We're remapping, not
	 * allocating, so skip that check by pretending to be freeing.
	 */
	error = xfs_alloc_fix_freelist(&args, XFS_ALLOC_FLAG_FREEING);
	if (error)
		goto error0;
error0:
	xfs_perag_put(args.pag);
	if (error)
		trace_xfs_bmap_remap_alloc_error(ap->ip, error, _RET_IP_);
	return error;
}

3936 3937 3938 3939 3940 3941
/*
 * xfs_bmap_alloc is called by xfs_bmapi to allocate an extent for a file.
 * It figures out where to ask the underlying allocator to put the new extent.
 */
STATIC int
xfs_bmap_alloc(
D
Dave Chinner 已提交
3942
	struct xfs_bmalloca	*ap)	/* bmap alloc argument struct */
3943
{
3944 3945
	if (ap->flags & XFS_BMAPI_REMAP)
		return xfs_bmap_remap_alloc(ap);
3946 3947
	if (XFS_IS_REALTIME_INODE(ap->ip) &&
	    xfs_alloc_is_userdata(ap->datatype))
3948 3949 3950
		return xfs_bmap_rtalloc(ap);
	return xfs_bmap_btalloc(ap);
}
3951

3952 3953 3954 3955 3956 3957 3958 3959 3960 3961 3962 3963 3964 3965 3966 3967 3968 3969 3970 3971 3972
/*
 * Trim the returned map to the required bounds
 */
STATIC void
xfs_bmapi_trim_map(
	struct xfs_bmbt_irec	*mval,
	struct xfs_bmbt_irec	*got,
	xfs_fileoff_t		*bno,
	xfs_filblks_t		len,
	xfs_fileoff_t		obno,
	xfs_fileoff_t		end,
	int			n,
	int			flags)
{
	if ((flags & XFS_BMAPI_ENTIRE) ||
	    got->br_startoff + got->br_blockcount <= obno) {
		*mval = *got;
		if (isnullstartblock(got->br_startblock))
			mval->br_startblock = DELAYSTARTBLOCK;
		return;
	}
3973

3974 3975 3976 3977 3978 3979 3980 3981 3982 3983
	if (obno > *bno)
		*bno = obno;
	ASSERT((*bno >= obno) || (n == 0));
	ASSERT(*bno < end);
	mval->br_startoff = *bno;
	if (isnullstartblock(got->br_startblock))
		mval->br_startblock = DELAYSTARTBLOCK;
	else
		mval->br_startblock = got->br_startblock +
					(*bno - got->br_startoff);
3984
	/*
3985 3986 3987 3988 3989
	 * Return the minimum of what we got and what we asked for for
	 * the length.  We can use the len variable here because it is
	 * modified below and we could have been there before coming
	 * here if the first part of the allocation didn't overlap what
	 * was asked for.
3990
	 */
3991 3992 3993 3994 3995
	mval->br_blockcount = XFS_FILBLKS_MIN(end - *bno,
			got->br_blockcount - (*bno - got->br_startoff));
	mval->br_state = got->br_state;
	ASSERT(mval->br_blockcount <= len);
	return;
3996 3997
}

3998 3999 4000 4001 4002 4003 4004 4005 4006 4007 4008 4009
/*
 * Update and validate the extent map to return
 */
STATIC void
xfs_bmapi_update_map(
	struct xfs_bmbt_irec	**map,
	xfs_fileoff_t		*bno,
	xfs_filblks_t		*len,
	xfs_fileoff_t		obno,
	xfs_fileoff_t		end,
	int			*n,
	int			flags)
4010
{
4011
	xfs_bmbt_irec_t	*mval = *map;
4012

4013 4014 4015 4016
	ASSERT((flags & XFS_BMAPI_ENTIRE) ||
	       ((mval->br_startoff + mval->br_blockcount) <= end));
	ASSERT((flags & XFS_BMAPI_ENTIRE) || (mval->br_blockcount <= *len) ||
	       (mval->br_startoff < obno));
4017

4018 4019 4020 4021 4022 4023 4024 4025 4026 4027 4028 4029 4030 4031 4032 4033 4034 4035 4036 4037 4038 4039 4040 4041 4042 4043 4044 4045 4046 4047 4048 4049 4050
	*bno = mval->br_startoff + mval->br_blockcount;
	*len = end - *bno;
	if (*n > 0 && mval->br_startoff == mval[-1].br_startoff) {
		/* update previous map with new information */
		ASSERT(mval->br_startblock == mval[-1].br_startblock);
		ASSERT(mval->br_blockcount > mval[-1].br_blockcount);
		ASSERT(mval->br_state == mval[-1].br_state);
		mval[-1].br_blockcount = mval->br_blockcount;
		mval[-1].br_state = mval->br_state;
	} else if (*n > 0 && mval->br_startblock != DELAYSTARTBLOCK &&
		   mval[-1].br_startblock != DELAYSTARTBLOCK &&
		   mval[-1].br_startblock != HOLESTARTBLOCK &&
		   mval->br_startblock == mval[-1].br_startblock +
					  mval[-1].br_blockcount &&
		   ((flags & XFS_BMAPI_IGSTATE) ||
			mval[-1].br_state == mval->br_state)) {
		ASSERT(mval->br_startoff ==
		       mval[-1].br_startoff + mval[-1].br_blockcount);
		mval[-1].br_blockcount += mval->br_blockcount;
	} else if (*n > 0 &&
		   mval->br_startblock == DELAYSTARTBLOCK &&
		   mval[-1].br_startblock == DELAYSTARTBLOCK &&
		   mval->br_startoff ==
		   mval[-1].br_startoff + mval[-1].br_blockcount) {
		mval[-1].br_blockcount += mval->br_blockcount;
		mval[-1].br_state = mval->br_state;
	} else if (!((*n == 0) &&
		     ((mval->br_startoff + mval->br_blockcount) <=
		      obno))) {
		mval++;
		(*n)++;
	}
	*map = mval;
4051 4052 4053
}

/*
4054
 * Map file blocks to filesystem blocks without allocation.
4055 4056
 */
int
4057 4058 4059
xfs_bmapi_read(
	struct xfs_inode	*ip,
	xfs_fileoff_t		bno,
4060
	xfs_filblks_t		len,
4061 4062
	struct xfs_bmbt_irec	*mval,
	int			*nmap,
4063
	int			flags)
4064
{
4065 4066 4067 4068 4069 4070 4071 4072 4073 4074
	struct xfs_mount	*mp = ip->i_mount;
	struct xfs_ifork	*ifp;
	struct xfs_bmbt_irec	got;
	struct xfs_bmbt_irec	prev;
	xfs_fileoff_t		obno;
	xfs_fileoff_t		end;
	xfs_extnum_t		lastx;
	int			error;
	int			eof;
	int			n = 0;
4075 4076 4077
	int			whichfork = (flags & XFS_BMAPI_ATTRFORK) ?
						XFS_ATTR_FORK : XFS_DATA_FORK;

4078 4079 4080
	ASSERT(*nmap >= 1);
	ASSERT(!(flags & ~(XFS_BMAPI_ATTRFORK|XFS_BMAPI_ENTIRE|
			   XFS_BMAPI_IGSTATE)));
4081
	ASSERT(xfs_isilocked(ip, XFS_ILOCK_SHARED|XFS_ILOCK_EXCL));
4082

4083 4084 4085 4086 4087
	if (unlikely(XFS_TEST_ERROR(
	    (XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_EXTENTS &&
	     XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_BTREE),
	     mp, XFS_ERRTAG_BMAPIFORMAT, XFS_RANDOM_BMAPIFORMAT))) {
		XFS_ERROR_REPORT("xfs_bmapi_read", XFS_ERRLEVEL_LOW, mp);
D
Dave Chinner 已提交
4088
		return -EFSCORRUPTED;
4089
	}
4090

4091
	if (XFS_FORCED_SHUTDOWN(mp))
D
Dave Chinner 已提交
4092
		return -EIO;
4093

4094
	XFS_STATS_INC(mp, xs_blk_mapr);
4095 4096 4097 4098 4099 4100 4101

	ifp = XFS_IFORK_PTR(ip, whichfork);

	if (!(ifp->if_flags & XFS_IFEXTENTS)) {
		error = xfs_iread_extents(NULL, ip, whichfork);
		if (error)
			return error;
4102 4103
	}

4104 4105 4106
	xfs_bmap_search_extents(ip, bno, whichfork, &eof, &lastx, &got, &prev);
	end = bno + len;
	obno = bno;
4107

4108 4109 4110 4111 4112 4113 4114 4115 4116 4117 4118 4119 4120 4121 4122 4123 4124
	while (bno < end && n < *nmap) {
		/* Reading past eof, act as though there's a hole up to end. */
		if (eof)
			got.br_startoff = end;
		if (got.br_startoff > bno) {
			/* Reading in a hole.  */
			mval->br_startoff = bno;
			mval->br_startblock = HOLESTARTBLOCK;
			mval->br_blockcount =
				XFS_FILBLKS_MIN(len, got.br_startoff - bno);
			mval->br_state = XFS_EXT_NORM;
			bno += mval->br_blockcount;
			len -= mval->br_blockcount;
			mval++;
			n++;
			continue;
		}
4125

4126 4127 4128 4129 4130 4131 4132 4133 4134 4135 4136 4137 4138 4139 4140
		/* set up the extent map to return. */
		xfs_bmapi_trim_map(mval, &got, &bno, len, obno, end, n, flags);
		xfs_bmapi_update_map(&mval, &bno, &len, obno, end, &n, flags);

		/* If we're done, stop now. */
		if (bno >= end || n >= *nmap)
			break;

		/* Else go on to the next record. */
		if (++lastx < ifp->if_bytes / sizeof(xfs_bmbt_rec_t))
			xfs_bmbt_get_all(xfs_iext_get_ext(ifp, lastx), &got);
		else
			eof = 1;
	}
	*nmap = n;
4141 4142 4143
	return 0;
}

4144
int
4145 4146 4147 4148 4149 4150 4151 4152
xfs_bmapi_reserve_delalloc(
	struct xfs_inode	*ip,
	xfs_fileoff_t		aoff,
	xfs_filblks_t		len,
	struct xfs_bmbt_irec	*got,
	struct xfs_bmbt_irec	*prev,
	xfs_extnum_t		*lastx,
	int			eof)
L
Linus Torvalds 已提交
4153
{
4154
	struct xfs_mount	*mp = ip->i_mount;
4155 4156 4157 4158 4159 4160
	struct xfs_ifork	*ifp = XFS_IFORK_PTR(ip, XFS_DATA_FORK);
	xfs_extlen_t		alen;
	xfs_extlen_t		indlen;
	char			rt = XFS_IS_REALTIME_INODE(ip);
	xfs_extlen_t		extsz;
	int			error;
4161

4162 4163 4164
	alen = XFS_FILBLKS_MIN(len, MAXEXTLEN);
	if (!eof)
		alen = XFS_FILBLKS_MIN(alen, got->br_startoff - aoff);
L
Linus Torvalds 已提交
4165

4166 4167 4168 4169 4170 4171 4172 4173 4174 4175 4176 4177 4178 4179 4180 4181 4182 4183 4184 4185 4186 4187 4188 4189 4190 4191 4192 4193 4194
	/* Figure out the extent size, adjust alen */
	extsz = xfs_get_extsz_hint(ip);
	if (extsz) {
		error = xfs_bmap_extsize_align(mp, got, prev, extsz, rt, eof,
					       1, 0, &aoff, &alen);
		ASSERT(!error);
	}

	if (rt)
		extsz = alen / mp->m_sb.sb_rextsize;

	/*
	 * Make a transaction-less quota reservation for delayed allocation
	 * blocks.  This number gets adjusted later.  We return if we haven't
	 * allocated blocks already inside this loop.
	 */
	error = xfs_trans_reserve_quota_nblks(NULL, ip, (long)alen, 0,
			rt ? XFS_QMOPT_RES_RTBLKS : XFS_QMOPT_RES_REGBLKS);
	if (error)
		return error;

	/*
	 * Split changing sb for alen and indlen since they could be coming
	 * from different places.
	 */
	indlen = (xfs_extlen_t)xfs_bmap_worst_indlen(ip, alen);
	ASSERT(indlen > 0);

	if (rt) {
D
Dave Chinner 已提交
4195
		error = xfs_mod_frextents(mp, -((int64_t)extsz));
4196
	} else {
4197
		error = xfs_mod_fdblocks(mp, -((int64_t)alen), false);
4198 4199 4200 4201 4202
	}

	if (error)
		goto out_unreserve_quota;

4203
	error = xfs_mod_fdblocks(mp, -((int64_t)indlen), false);
4204 4205 4206 4207 4208 4209 4210 4211 4212 4213 4214 4215 4216 4217 4218 4219 4220 4221 4222 4223 4224 4225 4226 4227 4228 4229
	if (error)
		goto out_unreserve_blocks;


	ip->i_delayed_blks += alen;

	got->br_startoff = aoff;
	got->br_startblock = nullstartblock(indlen);
	got->br_blockcount = alen;
	got->br_state = XFS_EXT_NORM;
	xfs_bmap_add_extent_hole_delay(ip, lastx, got);

	/*
	 * Update our extent pointer, given that xfs_bmap_add_extent_hole_delay
	 * might have merged it into one of the neighbouring ones.
	 */
	xfs_bmbt_get_all(xfs_iext_get_ext(ifp, *lastx), got);

	ASSERT(got->br_startoff <= aoff);
	ASSERT(got->br_startoff + got->br_blockcount >= aoff + alen);
	ASSERT(isnullstartblock(got->br_startblock));
	ASSERT(got->br_state == XFS_EXT_NORM);
	return 0;

out_unreserve_blocks:
	if (rt)
D
Dave Chinner 已提交
4230
		xfs_mod_frextents(mp, extsz);
4231
	else
4232
		xfs_mod_fdblocks(mp, alen, false);
4233 4234 4235 4236 4237 4238 4239
out_unreserve_quota:
	if (XFS_IS_QUOTA_ON(mp))
		xfs_trans_unreserve_quota_nblks(NULL, ip, (long)alen, 0, rt ?
				XFS_QMOPT_RES_RTBLKS : XFS_QMOPT_RES_REGBLKS);
	return error;
}

4240 4241
static int
xfs_bmapi_allocate(
4242 4243 4244 4245 4246 4247 4248 4249 4250 4251 4252
	struct xfs_bmalloca	*bma)
{
	struct xfs_mount	*mp = bma->ip->i_mount;
	int			whichfork = (bma->flags & XFS_BMAPI_ATTRFORK) ?
						XFS_ATTR_FORK : XFS_DATA_FORK;
	struct xfs_ifork	*ifp = XFS_IFORK_PTR(bma->ip, whichfork);
	int			tmp_logflags = 0;
	int			error;

	ASSERT(bma->length > 0);

L
Linus Torvalds 已提交
4253
	/*
4254 4255
	 * For the wasdelay case, we could also just allocate the stuff asked
	 * for in this bmap call but that wouldn't be as good.
L
Linus Torvalds 已提交
4256
	 */
4257 4258 4259 4260 4261 4262
	if (bma->wasdel) {
		bma->length = (xfs_extlen_t)bma->got.br_blockcount;
		bma->offset = bma->got.br_startoff;
		if (bma->idx != NULLEXTNUM && bma->idx) {
			xfs_bmbt_get_all(xfs_iext_get_ext(ifp, bma->idx - 1),
					 &bma->prev);
L
Linus Torvalds 已提交
4263
		}
4264 4265 4266 4267 4268
	} else {
		bma->length = XFS_FILBLKS_MIN(bma->length, MAXEXTLEN);
		if (!bma->eof)
			bma->length = XFS_FILBLKS_MIN(bma->length,
					bma->got.br_startoff - bma->offset);
L
Linus Torvalds 已提交
4269 4270
	}

4271
	/*
4272 4273 4274 4275
	 * Set the data type being allocated. For the data fork, the first data
	 * in the file is treated differently to all other allocations. For the
	 * attribute fork, we only need to ensure the allocated range is not on
	 * the busy list.
4276 4277
	 */
	if (!(bma->flags & XFS_BMAPI_METADATA)) {
4278 4279 4280 4281 4282 4283 4284
		bma->datatype = XFS_ALLOC_NOBUSY;
		if (whichfork == XFS_DATA_FORK) {
			if (bma->offset == 0)
				bma->datatype |= XFS_ALLOC_INITIAL_USER_DATA;
			else
				bma->datatype |= XFS_ALLOC_USERDATA;
		}
4285
		if (bma->flags & XFS_BMAPI_ZERO)
4286
			bma->datatype |= XFS_ALLOC_USERDATA_ZERO;
4287
	}
L
Linus Torvalds 已提交
4288

4289
	bma->minlen = (bma->flags & XFS_BMAPI_CONTIG) ? bma->length : 1;
C
Christoph Hellwig 已提交
4290

4291 4292 4293 4294 4295 4296 4297 4298 4299
	/*
	 * Only want to do the alignment at the eof if it is userdata and
	 * allocation length is larger than a stripe unit.
	 */
	if (mp->m_dalign && bma->length >= mp->m_dalign &&
	    !(bma->flags & XFS_BMAPI_METADATA) && whichfork == XFS_DATA_FORK) {
		error = xfs_bmap_isaeof(bma, whichfork);
		if (error)
			return error;
L
Linus Torvalds 已提交
4300
	}
4301

4302 4303
	error = xfs_bmap_alloc(bma);
	if (error)
L
Linus Torvalds 已提交
4304
		return error;
4305

4306
	if (bma->dfops->dop_low)
4307 4308 4309 4310
		bma->minleft = 0;
	if (bma->cur)
		bma->cur->bc_private.b.firstblock = *bma->firstblock;
	if (bma->blkno == NULLFSBLOCK)
L
Linus Torvalds 已提交
4311
		return 0;
4312 4313 4314
	if ((ifp->if_flags & XFS_IFBROOT) && !bma->cur) {
		bma->cur = xfs_bmbt_init_cursor(mp, bma->tp, bma->ip, whichfork);
		bma->cur->bc_private.b.firstblock = *bma->firstblock;
4315
		bma->cur->bc_private.b.dfops = bma->dfops;
L
Linus Torvalds 已提交
4316
	}
4317 4318 4319 4320 4321 4322 4323 4324 4325 4326 4327 4328 4329 4330
	/*
	 * Bump the number of extents we've allocated
	 * in this call.
	 */
	bma->nallocs++;

	if (bma->cur)
		bma->cur->bc_private.b.flags =
			bma->wasdel ? XFS_BTCUR_BPRV_WASDEL : 0;

	bma->got.br_startoff = bma->offset;
	bma->got.br_startblock = bma->blkno;
	bma->got.br_blockcount = bma->length;
	bma->got.br_state = XFS_EXT_NORM;
4331

L
Linus Torvalds 已提交
4332
	/*
4333 4334
	 * A wasdelay extent has been initialized, so shouldn't be flagged
	 * as unwritten.
L
Linus Torvalds 已提交
4335
	 */
4336 4337 4338 4339 4340 4341 4342 4343 4344 4345 4346 4347 4348 4349 4350 4351 4352 4353 4354 4355 4356 4357 4358 4359 4360 4361 4362 4363 4364 4365 4366 4367 4368 4369 4370 4371 4372 4373 4374 4375 4376 4377 4378 4379 4380 4381 4382 4383 4384 4385 4386 4387 4388 4389 4390 4391 4392 4393 4394 4395
	if (!bma->wasdel && (bma->flags & XFS_BMAPI_PREALLOC) &&
	    xfs_sb_version_hasextflgbit(&mp->m_sb))
		bma->got.br_state = XFS_EXT_UNWRITTEN;

	if (bma->wasdel)
		error = xfs_bmap_add_extent_delay_real(bma);
	else
		error = xfs_bmap_add_extent_hole_real(bma, whichfork);

	bma->logflags |= tmp_logflags;
	if (error)
		return error;

	/*
	 * Update our extent pointer, given that xfs_bmap_add_extent_delay_real
	 * or xfs_bmap_add_extent_hole_real might have merged it into one of
	 * the neighbouring ones.
	 */
	xfs_bmbt_get_all(xfs_iext_get_ext(ifp, bma->idx), &bma->got);

	ASSERT(bma->got.br_startoff <= bma->offset);
	ASSERT(bma->got.br_startoff + bma->got.br_blockcount >=
	       bma->offset + bma->length);
	ASSERT(bma->got.br_state == XFS_EXT_NORM ||
	       bma->got.br_state == XFS_EXT_UNWRITTEN);
	return 0;
}

STATIC int
xfs_bmapi_convert_unwritten(
	struct xfs_bmalloca	*bma,
	struct xfs_bmbt_irec	*mval,
	xfs_filblks_t		len,
	int			flags)
{
	int			whichfork = (flags & XFS_BMAPI_ATTRFORK) ?
						XFS_ATTR_FORK : XFS_DATA_FORK;
	struct xfs_ifork	*ifp = XFS_IFORK_PTR(bma->ip, whichfork);
	int			tmp_logflags = 0;
	int			error;

	/* check if we need to do unwritten->real conversion */
	if (mval->br_state == XFS_EXT_UNWRITTEN &&
	    (flags & XFS_BMAPI_PREALLOC))
		return 0;

	/* check if we need to do real->unwritten conversion */
	if (mval->br_state == XFS_EXT_NORM &&
	    (flags & (XFS_BMAPI_PREALLOC | XFS_BMAPI_CONVERT)) !=
			(XFS_BMAPI_PREALLOC | XFS_BMAPI_CONVERT))
		return 0;

	/*
	 * Modify (by adding) the state flag, if writing.
	 */
	ASSERT(mval->br_blockcount <= len);
	if ((ifp->if_flags & XFS_IFBROOT) && !bma->cur) {
		bma->cur = xfs_bmbt_init_cursor(bma->ip->i_mount, bma->tp,
					bma->ip, whichfork);
		bma->cur->bc_private.b.firstblock = *bma->firstblock;
4396
		bma->cur->bc_private.b.dfops = bma->dfops;
L
Linus Torvalds 已提交
4397
	}
4398 4399
	mval->br_state = (mval->br_state == XFS_EXT_UNWRITTEN)
				? XFS_EXT_NORM : XFS_EXT_UNWRITTEN;
4400

4401 4402 4403 4404 4405 4406 4407 4408 4409 4410 4411
	/*
	 * Before insertion into the bmbt, zero the range being converted
	 * if required.
	 */
	if (flags & XFS_BMAPI_ZERO) {
		error = xfs_zero_extent(bma->ip, mval->br_startblock,
					mval->br_blockcount);
		if (error)
			return error;
	}

4412
	error = xfs_bmap_add_extent_unwritten_real(bma->tp, bma->ip, &bma->idx,
4413
			&bma->cur, mval, bma->firstblock, bma->dfops,
4414
			&tmp_logflags);
4415 4416 4417 4418 4419 4420 4421 4422 4423
	/*
	 * Log the inode core unconditionally in the unwritten extent conversion
	 * path because the conversion might not have done so (e.g., if the
	 * extent count hasn't changed). We need to make sure the inode is dirty
	 * in the transaction for the sake of fsync(), even if nothing has
	 * changed, because fsync() will not force the log for this transaction
	 * unless it sees the inode pinned.
	 */
	bma->logflags |= tmp_logflags | XFS_ILOG_CORE;
4424 4425 4426 4427 4428 4429 4430 4431 4432 4433 4434 4435 4436 4437 4438
	if (error)
		return error;

	/*
	 * Update our extent pointer, given that
	 * xfs_bmap_add_extent_unwritten_real might have merged it into one
	 * of the neighbouring ones.
	 */
	xfs_bmbt_get_all(xfs_iext_get_ext(ifp, bma->idx), &bma->got);

	/*
	 * We may have combined previously unwritten space with written space,
	 * so generate another request.
	 */
	if (mval->br_blockcount < len)
D
Dave Chinner 已提交
4439
		return -EAGAIN;
4440 4441 4442 4443 4444 4445 4446 4447 4448 4449 4450 4451 4452 4453 4454 4455 4456 4457 4458 4459 4460 4461 4462 4463 4464 4465 4466
	return 0;
}

/*
 * Map file blocks to filesystem blocks, and allocate blocks or convert the
 * extent state if necessary.  Details behaviour is controlled by the flags
 * parameter.  Only allocates blocks from a single allocation group, to avoid
 * locking problems.
 *
 * The returned value in "firstblock" from the first call in a transaction
 * must be remembered and presented to subsequent calls in "firstblock".
 * An upper bound for the number of blocks to be allocated is supplied to
 * the first call in "total"; if no allocation group has that many free
 * blocks then the call will fail (return NULLFSBLOCK in "firstblock").
 */
int
xfs_bmapi_write(
	struct xfs_trans	*tp,		/* transaction pointer */
	struct xfs_inode	*ip,		/* incore inode */
	xfs_fileoff_t		bno,		/* starting file offs. mapped */
	xfs_filblks_t		len,		/* length to map in file */
	int			flags,		/* XFS_BMAPI_... */
	xfs_fsblock_t		*firstblock,	/* first allocated block
						   controls a.g. for allocs */
	xfs_extlen_t		total,		/* total blocks needed */
	struct xfs_bmbt_irec	*mval,		/* output: map values */
	int			*nmap,		/* i/o: mval size/count */
4467
	struct xfs_defer_ops	*dfops)		/* i/o: list extents to free */
4468 4469 4470
{
	struct xfs_mount	*mp = ip->i_mount;
	struct xfs_ifork	*ifp;
4471
	struct xfs_bmalloca	bma = { NULL };	/* args for xfs_bmap_alloc */
4472 4473 4474 4475 4476 4477 4478 4479 4480 4481 4482 4483 4484 4485 4486 4487 4488 4489 4490 4491 4492 4493
	xfs_fileoff_t		end;		/* end of mapped file region */
	int			eof;		/* after the end of extents */
	int			error;		/* error return */
	int			n;		/* current extent index */
	xfs_fileoff_t		obno;		/* old block number (offset) */
	int			whichfork;	/* data or attr fork */
	char			inhole;		/* current location is hole in file */
	char			wasdelay;	/* old extent was delayed */

#ifdef DEBUG
	xfs_fileoff_t		orig_bno;	/* original block number value */
	int			orig_flags;	/* original flags arg value */
	xfs_filblks_t		orig_len;	/* original value of len arg */
	struct xfs_bmbt_irec	*orig_mval;	/* original value of mval */
	int			orig_nmap;	/* original value of *nmap */

	orig_bno = bno;
	orig_len = len;
	orig_flags = flags;
	orig_mval = mval;
	orig_nmap = *nmap;
#endif
4494 4495
	whichfork = (flags & XFS_BMAPI_ATTRFORK) ?
		XFS_ATTR_FORK : XFS_DATA_FORK;
4496 4497 4498 4499 4500 4501

	ASSERT(*nmap >= 1);
	ASSERT(*nmap <= XFS_BMAP_MAX_NMAP);
	ASSERT(!(flags & XFS_BMAPI_IGSTATE));
	ASSERT(tp != NULL);
	ASSERT(len > 0);
4502
	ASSERT(XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_LOCAL);
4503
	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
4504 4505 4506
	ASSERT(!(flags & XFS_BMAPI_REMAP) || whichfork == XFS_DATA_FORK);
	ASSERT(!(flags & XFS_BMAPI_PREALLOC) || !(flags & XFS_BMAPI_REMAP));
	ASSERT(!(flags & XFS_BMAPI_CONVERT) || !(flags & XFS_BMAPI_REMAP));
4507

4508 4509 4510 4511 4512 4513 4514 4515 4516 4517 4518 4519
	/* zeroing is for currently only for data extents, not metadata */
	ASSERT((flags & (XFS_BMAPI_METADATA | XFS_BMAPI_ZERO)) !=
			(XFS_BMAPI_METADATA | XFS_BMAPI_ZERO));
	/*
	 * we can allocate unwritten extents or pre-zero allocated blocks,
	 * but it makes no sense to do both at once. This would result in
	 * zeroing the unwritten extent twice, but it still being an
	 * unwritten extent....
	 */
	ASSERT((flags & (XFS_BMAPI_PREALLOC | XFS_BMAPI_ZERO)) !=
			(XFS_BMAPI_PREALLOC | XFS_BMAPI_ZERO));

4520 4521
	if (unlikely(XFS_TEST_ERROR(
	    (XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_EXTENTS &&
4522
	     XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_BTREE),
4523 4524
	     mp, XFS_ERRTAG_BMAPIFORMAT, XFS_RANDOM_BMAPIFORMAT))) {
		XFS_ERROR_REPORT("xfs_bmapi_write", XFS_ERRLEVEL_LOW, mp);
D
Dave Chinner 已提交
4525
		return -EFSCORRUPTED;
4526 4527
	}

4528
	if (XFS_FORCED_SHUTDOWN(mp))
D
Dave Chinner 已提交
4529
		return -EIO;
4530 4531 4532

	ifp = XFS_IFORK_PTR(ip, whichfork);

4533
	XFS_STATS_INC(mp, xs_blk_mapw);
4534 4535 4536 4537 4538 4539 4540 4541 4542 4543 4544 4545 4546 4547 4548 4549 4550 4551 4552 4553 4554 4555 4556 4557 4558

	if (*firstblock == NULLFSBLOCK) {
		if (XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_BTREE)
			bma.minleft = be16_to_cpu(ifp->if_broot->bb_level) + 1;
		else
			bma.minleft = 1;
	} else {
		bma.minleft = 0;
	}

	if (!(ifp->if_flags & XFS_IFEXTENTS)) {
		error = xfs_iread_extents(tp, ip, whichfork);
		if (error)
			goto error0;
	}

	xfs_bmap_search_extents(ip, bno, whichfork, &eof, &bma.idx, &bma.got,
				&bma.prev);
	n = 0;
	end = bno + len;
	obno = bno;

	bma.tp = tp;
	bma.ip = ip;
	bma.total = total;
4559
	bma.datatype = 0;
4560
	bma.dfops = dfops;
4561 4562 4563 4564 4565 4566
	bma.firstblock = firstblock;

	while (bno < end && n < *nmap) {
		inhole = eof || bma.got.br_startoff > bno;
		wasdelay = !inhole && isnullstartblock(bma.got.br_startblock);

4567 4568 4569 4570 4571 4572
		/*
		 * Make sure we only reflink into a hole.
		 */
		if (flags & XFS_BMAPI_REMAP)
			ASSERT(inhole);

L
Linus Torvalds 已提交
4573
		/*
4574 4575
		 * First, deal with the hole before the allocated space
		 * that we found, if any.
L
Linus Torvalds 已提交
4576
		 */
4577 4578 4579 4580 4581 4582 4583
		if (inhole || wasdelay) {
			bma.eof = eof;
			bma.conv = !!(flags & XFS_BMAPI_CONVERT);
			bma.wasdel = wasdelay;
			bma.offset = bno;
			bma.flags = flags;

L
Linus Torvalds 已提交
4584
			/*
4585 4586 4587 4588 4589
			 * There's a 32/64 bit type mismatch between the
			 * allocation length request (which can be 64 bits in
			 * length) and the bma length request, which is
			 * xfs_extlen_t and therefore 32 bits. Hence we have to
			 * check for 32-bit overflows and handle them here.
L
Linus Torvalds 已提交
4590
			 */
4591 4592 4593 4594 4595 4596 4597 4598
			if (len > (xfs_filblks_t)MAXEXTLEN)
				bma.length = MAXEXTLEN;
			else
				bma.length = len;

			ASSERT(len > 0);
			ASSERT(bma.length > 0);
			error = xfs_bmapi_allocate(&bma);
L
Linus Torvalds 已提交
4599 4600
			if (error)
				goto error0;
4601 4602
			if (bma.blkno == NULLFSBLOCK)
				break;
L
Linus Torvalds 已提交
4603
		}
4604 4605 4606 4607 4608 4609 4610

		/* Deal with the allocated space we found.  */
		xfs_bmapi_trim_map(mval, &bma.got, &bno, len, obno,
							end, n, flags);

		/* Execute unwritten extent conversion if necessary */
		error = xfs_bmapi_convert_unwritten(&bma, mval, len, flags);
D
Dave Chinner 已提交
4611
		if (error == -EAGAIN)
4612 4613 4614 4615 4616 4617 4618 4619 4620 4621 4622 4623 4624 4625 4626 4627 4628 4629 4630 4631 4632 4633 4634 4635 4636 4637 4638 4639 4640 4641 4642 4643 4644 4645 4646 4647 4648 4649 4650 4651 4652 4653 4654 4655 4656 4657 4658 4659 4660 4661 4662 4663 4664 4665 4666 4667 4668 4669 4670 4671 4672 4673 4674 4675 4676 4677 4678 4679
			continue;
		if (error)
			goto error0;

		/* update the extent map to return */
		xfs_bmapi_update_map(&mval, &bno, &len, obno, end, &n, flags);

		/*
		 * If we're done, stop now.  Stop when we've allocated
		 * XFS_BMAP_MAX_NMAP extents no matter what.  Otherwise
		 * the transaction may get too big.
		 */
		if (bno >= end || n >= *nmap || bma.nallocs >= *nmap)
			break;

		/* Else go on to the next record. */
		bma.prev = bma.got;
		if (++bma.idx < ifp->if_bytes / sizeof(xfs_bmbt_rec_t)) {
			xfs_bmbt_get_all(xfs_iext_get_ext(ifp, bma.idx),
					 &bma.got);
		} else
			eof = 1;
	}
	*nmap = n;

	/*
	 * Transform from btree to extents, give it cur.
	 */
	if (xfs_bmap_wants_extents(ip, whichfork)) {
		int		tmp_logflags = 0;

		ASSERT(bma.cur);
		error = xfs_bmap_btree_to_extents(tp, ip, bma.cur,
			&tmp_logflags, whichfork);
		bma.logflags |= tmp_logflags;
		if (error)
			goto error0;
	}

	ASSERT(XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_BTREE ||
	       XFS_IFORK_NEXTENTS(ip, whichfork) >
		XFS_IFORK_MAXEXT(ip, whichfork));
	error = 0;
error0:
	/*
	 * Log everything.  Do this after conversion, there's no point in
	 * logging the extent records if we've converted to btree format.
	 */
	if ((bma.logflags & xfs_ilog_fext(whichfork)) &&
	    XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_EXTENTS)
		bma.logflags &= ~xfs_ilog_fext(whichfork);
	else if ((bma.logflags & xfs_ilog_fbroot(whichfork)) &&
		 XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_BTREE)
		bma.logflags &= ~xfs_ilog_fbroot(whichfork);
	/*
	 * Log whatever the flags say, even if error.  Otherwise we might miss
	 * detecting a case where the data is changed, there's an error,
	 * and it's not logged so we don't shutdown when we should.
	 */
	if (bma.logflags)
		xfs_trans_log_inode(tp, ip, bma.logflags);

	if (bma.cur) {
		if (!error) {
			ASSERT(*firstblock == NULLFSBLOCK ||
			       XFS_FSB_TO_AGNO(mp, *firstblock) ==
			       XFS_FSB_TO_AGNO(mp,
				       bma.cur->bc_private.b.firstblock) ||
4680
			       (dfops->dop_low &&
4681 4682 4683 4684
				XFS_FSB_TO_AGNO(mp, *firstblock) <
				XFS_FSB_TO_AGNO(mp,
					bma.cur->bc_private.b.firstblock)));
			*firstblock = bma.cur->bc_private.b.firstblock;
L
Linus Torvalds 已提交
4685
		}
4686 4687 4688 4689 4690 4691 4692 4693
		xfs_btree_del_cursor(bma.cur,
			error ? XFS_BTREE_ERROR : XFS_BTREE_NOERROR);
	}
	if (!error)
		xfs_bmap_validate_ret(orig_bno, orig_len, orig_flags, orig_mval,
			orig_nmap, *nmap);
	return error;
}
4694

4695 4696 4697 4698 4699 4700 4701
/*
 * When a delalloc extent is split (e.g., due to a hole punch), the original
 * indlen reservation must be shared across the two new extents that are left
 * behind.
 *
 * Given the original reservation and the worst case indlen for the two new
 * extents (as calculated by xfs_bmap_worst_indlen()), split the original
4702 4703 4704 4705
 * reservation fairly across the two new extents. If necessary, steal available
 * blocks from a deleted extent to make up a reservation deficiency (e.g., if
 * ores == 1). The number of stolen blocks is returned. The availability and
 * subsequent accounting of stolen blocks is the responsibility of the caller.
4706
 */
4707
static xfs_filblks_t
4708 4709 4710
xfs_bmap_split_indlen(
	xfs_filblks_t			ores,		/* original res. */
	xfs_filblks_t			*indlen1,	/* ext1 worst indlen */
4711 4712
	xfs_filblks_t			*indlen2,	/* ext2 worst indlen */
	xfs_filblks_t			avail)		/* stealable blocks */
4713 4714 4715 4716
{
	xfs_filblks_t			len1 = *indlen1;
	xfs_filblks_t			len2 = *indlen2;
	xfs_filblks_t			nres = len1 + len2; /* new total res. */
4717 4718 4719 4720 4721 4722 4723 4724 4725 4726 4727
	xfs_filblks_t			stolen = 0;

	/*
	 * Steal as many blocks as we can to try and satisfy the worst case
	 * indlen for both new extents.
	 */
	while (nres > ores && avail) {
		nres--;
		avail--;
		stolen++;
	}
4728 4729

	/*
4730 4731 4732 4733 4734
	 * The only blocks available are those reserved for the original
	 * extent and what we can steal from the extent being removed.
	 * If this still isn't enough to satisfy the combined
	 * requirements for the two new extents, skim blocks off of each
	 * of the new reservations until they match what is available.
4735 4736 4737 4738 4739 4740 4741 4742 4743 4744 4745 4746 4747 4748 4749 4750
	 */
	while (nres > ores) {
		if (len1) {
			len1--;
			nres--;
		}
		if (nres == ores)
			break;
		if (len2) {
			len2--;
			nres--;
		}
	}

	*indlen1 = len1;
	*indlen2 = len2;
4751 4752

	return stolen;
4753 4754
}

4755 4756 4757 4758 4759 4760 4761 4762 4763
/*
 * Called by xfs_bmapi to update file extent records and the btree
 * after removing space (or undoing a delayed allocation).
 */
STATIC int				/* error */
xfs_bmap_del_extent(
	xfs_inode_t		*ip,	/* incore inode pointer */
	xfs_trans_t		*tp,	/* current transaction pointer */
	xfs_extnum_t		*idx,	/* extent number to update/delete */
4764
	struct xfs_defer_ops	*dfops,	/* list of extents to be freed */
4765 4766 4767 4768 4769 4770 4771 4772 4773 4774 4775 4776 4777 4778 4779 4780 4781 4782 4783 4784 4785 4786 4787 4788 4789 4790 4791
	xfs_btree_cur_t		*cur,	/* if null, not a btree */
	xfs_bmbt_irec_t		*del,	/* data to remove from extents */
	int			*logflagsp, /* inode logging flags */
	int			whichfork) /* data or attr fork */
{
	xfs_filblks_t		da_new;	/* new delay-alloc indirect blocks */
	xfs_filblks_t		da_old;	/* old delay-alloc indirect blocks */
	xfs_fsblock_t		del_endblock=0;	/* first block past del */
	xfs_fileoff_t		del_endoff;	/* first offset past del */
	int			delay;	/* current block is delayed allocated */
	int			do_fx;	/* free extent at end of routine */
	xfs_bmbt_rec_host_t	*ep;	/* current extent entry pointer */
	int			error;	/* error return value */
	int			flags;	/* inode logging flags */
	xfs_bmbt_irec_t		got;	/* current extent entry */
	xfs_fileoff_t		got_endoff;	/* first offset past got */
	int			i;	/* temp state */
	xfs_ifork_t		*ifp;	/* inode fork pointer */
	xfs_mount_t		*mp;	/* mount structure */
	xfs_filblks_t		nblks;	/* quota/sb block count */
	xfs_bmbt_irec_t		new;	/* new record to be inserted */
	/* REFERENCED */
	uint			qfield;	/* quota field to update */
	xfs_filblks_t		temp;	/* for indirect length calculations */
	xfs_filblks_t		temp2;	/* for indirect length calculations */
	int			state = 0;

4792 4793
	mp = ip->i_mount;
	XFS_STATS_INC(mp, xs_del_exlist);
4794 4795 4796 4797 4798 4799 4800 4801 4802 4803 4804 4805 4806 4807 4808 4809 4810 4811 4812 4813 4814 4815 4816 4817 4818 4819 4820 4821 4822 4823 4824 4825 4826 4827 4828 4829 4830 4831 4832 4833 4834 4835 4836 4837 4838 4839

	if (whichfork == XFS_ATTR_FORK)
		state |= BMAP_ATTRFORK;

	ifp = XFS_IFORK_PTR(ip, whichfork);
	ASSERT((*idx >= 0) && (*idx < ifp->if_bytes /
		(uint)sizeof(xfs_bmbt_rec_t)));
	ASSERT(del->br_blockcount > 0);
	ep = xfs_iext_get_ext(ifp, *idx);
	xfs_bmbt_get_all(ep, &got);
	ASSERT(got.br_startoff <= del->br_startoff);
	del_endoff = del->br_startoff + del->br_blockcount;
	got_endoff = got.br_startoff + got.br_blockcount;
	ASSERT(got_endoff >= del_endoff);
	delay = isnullstartblock(got.br_startblock);
	ASSERT(isnullstartblock(del->br_startblock) == delay);
	flags = 0;
	qfield = 0;
	error = 0;
	/*
	 * If deleting a real allocation, must free up the disk space.
	 */
	if (!delay) {
		flags = XFS_ILOG_CORE;
		/*
		 * Realtime allocation.  Free it and record di_nblocks update.
		 */
		if (whichfork == XFS_DATA_FORK && XFS_IS_REALTIME_INODE(ip)) {
			xfs_fsblock_t	bno;
			xfs_filblks_t	len;

			ASSERT(do_mod(del->br_blockcount,
				      mp->m_sb.sb_rextsize) == 0);
			ASSERT(do_mod(del->br_startblock,
				      mp->m_sb.sb_rextsize) == 0);
			bno = del->br_startblock;
			len = del->br_blockcount;
			do_div(bno, mp->m_sb.sb_rextsize);
			do_div(len, mp->m_sb.sb_rextsize);
			error = xfs_rtfree_extent(tp, bno, (xfs_extlen_t)len);
			if (error)
				goto done;
			do_fx = 0;
			nblks = len * mp->m_sb.sb_rextsize;
			qfield = XFS_TRANS_DQ_RTBCOUNT;
		}
L
Linus Torvalds 已提交
4840
		/*
4841
		 * Ordinary allocation.
L
Linus Torvalds 已提交
4842
		 */
4843 4844 4845 4846
		else {
			do_fx = 1;
			nblks = del->br_blockcount;
			qfield = XFS_TRANS_DQ_BCOUNT;
L
Linus Torvalds 已提交
4847 4848
		}
		/*
4849
		 * Set up del_endblock and cur for later.
L
Linus Torvalds 已提交
4850
		 */
4851 4852 4853 4854 4855 4856
		del_endblock = del->br_startblock + del->br_blockcount;
		if (cur) {
			if ((error = xfs_bmbt_lookup_eq(cur, got.br_startoff,
					got.br_startblock, got.br_blockcount,
					&i)))
				goto done;
4857
			XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
L
Linus Torvalds 已提交
4858
		}
4859 4860 4861 4862 4863 4864
		da_old = da_new = 0;
	} else {
		da_old = startblockval(got.br_startblock);
		da_new = 0;
		nblks = 0;
		do_fx = 0;
L
Linus Torvalds 已提交
4865
	}
4866

L
Linus Torvalds 已提交
4867
	/*
4868 4869
	 * Set flag value to use in switch statement.
	 * Left-contig is 2, right-contig is 1.
L
Linus Torvalds 已提交
4870
	 */
4871 4872 4873 4874 4875 4876 4877 4878 4879 4880 4881 4882 4883 4884 4885 4886 4887 4888 4889 4890 4891
	switch (((got.br_startoff == del->br_startoff) << 1) |
		(got_endoff == del_endoff)) {
	case 3:
		/*
		 * Matches the whole extent.  Delete the entry.
		 */
		xfs_iext_remove(ip, *idx, 1,
				whichfork == XFS_ATTR_FORK ? BMAP_ATTRFORK : 0);
		--*idx;
		if (delay)
			break;

		XFS_IFORK_NEXT_SET(ip, whichfork,
			XFS_IFORK_NEXTENTS(ip, whichfork) - 1);
		flags |= XFS_ILOG_CORE;
		if (!cur) {
			flags |= xfs_ilog_fext(whichfork);
			break;
		}
		if ((error = xfs_btree_delete(cur, &i)))
			goto done;
4892
		XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
4893 4894 4895 4896 4897 4898 4899 4900 4901 4902 4903 4904 4905 4906 4907 4908 4909 4910 4911 4912 4913 4914 4915 4916 4917 4918 4919 4920 4921 4922 4923 4924 4925 4926 4927 4928 4929 4930 4931 4932 4933 4934 4935 4936 4937 4938 4939 4940 4941 4942 4943 4944 4945 4946 4947 4948 4949 4950 4951 4952 4953 4954 4955 4956 4957 4958 4959 4960 4961 4962 4963 4964 4965 4966 4967 4968 4969 4970 4971 4972 4973
		break;

	case 2:
		/*
		 * Deleting the first part of the extent.
		 */
		trace_xfs_bmap_pre_update(ip, *idx, state, _THIS_IP_);
		xfs_bmbt_set_startoff(ep, del_endoff);
		temp = got.br_blockcount - del->br_blockcount;
		xfs_bmbt_set_blockcount(ep, temp);
		if (delay) {
			temp = XFS_FILBLKS_MIN(xfs_bmap_worst_indlen(ip, temp),
				da_old);
			xfs_bmbt_set_startblock(ep, nullstartblock((int)temp));
			trace_xfs_bmap_post_update(ip, *idx, state, _THIS_IP_);
			da_new = temp;
			break;
		}
		xfs_bmbt_set_startblock(ep, del_endblock);
		trace_xfs_bmap_post_update(ip, *idx, state, _THIS_IP_);
		if (!cur) {
			flags |= xfs_ilog_fext(whichfork);
			break;
		}
		if ((error = xfs_bmbt_update(cur, del_endoff, del_endblock,
				got.br_blockcount - del->br_blockcount,
				got.br_state)))
			goto done;
		break;

	case 1:
		/*
		 * Deleting the last part of the extent.
		 */
		temp = got.br_blockcount - del->br_blockcount;
		trace_xfs_bmap_pre_update(ip, *idx, state, _THIS_IP_);
		xfs_bmbt_set_blockcount(ep, temp);
		if (delay) {
			temp = XFS_FILBLKS_MIN(xfs_bmap_worst_indlen(ip, temp),
				da_old);
			xfs_bmbt_set_startblock(ep, nullstartblock((int)temp));
			trace_xfs_bmap_post_update(ip, *idx, state, _THIS_IP_);
			da_new = temp;
			break;
		}
		trace_xfs_bmap_post_update(ip, *idx, state, _THIS_IP_);
		if (!cur) {
			flags |= xfs_ilog_fext(whichfork);
			break;
		}
		if ((error = xfs_bmbt_update(cur, got.br_startoff,
				got.br_startblock,
				got.br_blockcount - del->br_blockcount,
				got.br_state)))
			goto done;
		break;

	case 0:
		/*
		 * Deleting the middle of the extent.
		 */
		temp = del->br_startoff - got.br_startoff;
		trace_xfs_bmap_pre_update(ip, *idx, state, _THIS_IP_);
		xfs_bmbt_set_blockcount(ep, temp);
		new.br_startoff = del_endoff;
		temp2 = got_endoff - del_endoff;
		new.br_blockcount = temp2;
		new.br_state = got.br_state;
		if (!delay) {
			new.br_startblock = del_endblock;
			flags |= XFS_ILOG_CORE;
			if (cur) {
				if ((error = xfs_bmbt_update(cur,
						got.br_startoff,
						got.br_startblock, temp,
						got.br_state)))
					goto done;
				if ((error = xfs_btree_increment(cur, 0, &i)))
					goto done;
				cur->bc_rec.b = new;
				error = xfs_btree_insert(cur, &i);
D
Dave Chinner 已提交
4974
				if (error && error != -ENOSPC)
4975 4976 4977 4978 4979 4980 4981
					goto done;
				/*
				 * If get no-space back from btree insert,
				 * it tried a split, and we have a zero
				 * block reservation.
				 * Fix up our state and return the error.
				 */
D
Dave Chinner 已提交
4982
				if (error == -ENOSPC) {
4983 4984 4985 4986 4987 4988 4989 4990 4991
					/*
					 * Reset the cursor, don't trust
					 * it after any insert operation.
					 */
					if ((error = xfs_bmbt_lookup_eq(cur,
							got.br_startoff,
							got.br_startblock,
							temp, &i)))
						goto done;
4992 4993
					XFS_WANT_CORRUPTED_GOTO(mp,
								i == 1, done);
4994 4995 4996 4997 4998 4999 5000 5001 5002 5003 5004 5005 5006 5007 5008 5009 5010
					/*
					 * Update the btree record back
					 * to the original value.
					 */
					if ((error = xfs_bmbt_update(cur,
							got.br_startoff,
							got.br_startblock,
							got.br_blockcount,
							got.br_state)))
						goto done;
					/*
					 * Reset the extent record back
					 * to the original value.
					 */
					xfs_bmbt_set_blockcount(ep,
						got.br_blockcount);
					flags = 0;
D
Dave Chinner 已提交
5011
					error = -ENOSPC;
5012 5013
					goto done;
				}
5014
				XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
5015 5016 5017 5018 5019
			} else
				flags |= xfs_ilog_fext(whichfork);
			XFS_IFORK_NEXT_SET(ip, whichfork,
				XFS_IFORK_NEXTENTS(ip, whichfork) + 1);
		} else {
5020
			xfs_filblks_t	stolen;
5021
			ASSERT(whichfork == XFS_DATA_FORK);
5022 5023 5024

			/*
			 * Distribute the original indlen reservation across the
5025 5026 5027
			 * two new extents. Steal blocks from the deleted extent
			 * if necessary. Stealing blocks simply fudges the
			 * fdblocks accounting in xfs_bunmapi().
5028 5029 5030
			 */
			temp = xfs_bmap_worst_indlen(ip, got.br_blockcount);
			temp2 = xfs_bmap_worst_indlen(ip, new.br_blockcount);
5031 5032 5033 5034
			stolen = xfs_bmap_split_indlen(da_old, &temp, &temp2,
						       del->br_blockcount);
			da_new = temp + temp2 - stolen;
			del->br_blockcount -= stolen;
5035 5036

			/*
5037 5038
			 * Set the reservation for each extent. Warn if either
			 * is zero as this can lead to delalloc problems.
5039
			 */
5040
			WARN_ON_ONCE(!temp || !temp2);
5041 5042 5043 5044 5045 5046 5047
			xfs_bmbt_set_startblock(ep, nullstartblock((int)temp));
			new.br_startblock = nullstartblock((int)temp2);
		}
		trace_xfs_bmap_post_update(ip, *idx, state, _THIS_IP_);
		xfs_iext_insert(ip, *idx + 1, 1, &new, state);
		++*idx;
		break;
L
Linus Torvalds 已提交
5048
	}
5049 5050 5051 5052 5053 5054 5055 5056

	/* remove reverse mapping */
	if (!delay) {
		error = xfs_rmap_unmap_extent(mp, dfops, ip, whichfork, del);
		if (error)
			goto done;
	}

L
Linus Torvalds 已提交
5057
	/*
5058
	 * If we need to, add to list of extents to delete.
L
Linus Torvalds 已提交
5059
	 */
5060 5061 5062 5063 5064 5065 5066 5067 5068 5069
	if (do_fx) {
		if (xfs_is_reflink_inode(ip) && whichfork == XFS_DATA_FORK) {
			error = xfs_refcount_decrease_extent(mp, dfops, del);
			if (error)
				goto done;
		} else
			xfs_bmap_add_free(mp, dfops, del->br_startblock,
					del->br_blockcount, NULL);
	}

L
Linus Torvalds 已提交
5070
	/*
5071
	 * Adjust inode # blocks in the file.
L
Linus Torvalds 已提交
5072
	 */
5073 5074
	if (nblks)
		ip->i_d.di_nblocks -= nblks;
L
Linus Torvalds 已提交
5075
	/*
5076
	 * Adjust quota data.
L
Linus Torvalds 已提交
5077
	 */
5078 5079 5080 5081 5082 5083 5084 5085
	if (qfield)
		xfs_trans_mod_dquot_byino(tp, ip, qfield, (long)-nblks);

	/*
	 * Account for change in delayed indirect blocks.
	 * Nothing to do for disk quota accounting here.
	 */
	ASSERT(da_old >= da_new);
5086 5087
	if (da_old > da_new)
		xfs_mod_fdblocks(mp, (int64_t)(da_old - da_new), false);
5088 5089
done:
	*logflagsp = flags;
L
Linus Torvalds 已提交
5090 5091 5092
	return error;
}

5093
/*
5094 5095 5096 5097
 * Unmap (remove) blocks from a file.
 * If nexts is nonzero then the number of extents to remove is limited to
 * that value.  If not all extents in the block range can be removed then
 * *done is set.
5098
 */
5099 5100 5101 5102 5103 5104 5105 5106 5107 5108
int						/* error */
xfs_bunmapi(
	xfs_trans_t		*tp,		/* transaction pointer */
	struct xfs_inode	*ip,		/* incore inode */
	xfs_fileoff_t		bno,		/* starting offset to unmap */
	xfs_filblks_t		len,		/* length to unmap in file */
	int			flags,		/* misc flags */
	xfs_extnum_t		nexts,		/* number of extents max */
	xfs_fsblock_t		*firstblock,	/* first allocated block
						   controls a.g. for allocs */
5109
	struct xfs_defer_ops	*dfops,		/* i/o: list extents to free */
5110
	int			*done)		/* set if not done yet */
5111
{
5112 5113 5114 5115 5116 5117 5118
	xfs_btree_cur_t		*cur;		/* bmap btree cursor */
	xfs_bmbt_irec_t		del;		/* extent being deleted */
	int			eof;		/* is deleting at eof */
	xfs_bmbt_rec_host_t	*ep;		/* extent record pointer */
	int			error;		/* error return value */
	xfs_extnum_t		extno;		/* extent number in list */
	xfs_bmbt_irec_t		got;		/* current extent record */
E
Eric Sandeen 已提交
5119
	xfs_ifork_t		*ifp;		/* inode fork pointer */
5120 5121 5122 5123 5124 5125 5126 5127 5128 5129 5130 5131
	int			isrt;		/* freeing in rt area */
	xfs_extnum_t		lastx;		/* last extent index used */
	int			logflags;	/* transaction logging flags */
	xfs_extlen_t		mod;		/* rt extent offset */
	xfs_mount_t		*mp;		/* mount structure */
	xfs_extnum_t		nextents;	/* number of file extents */
	xfs_bmbt_irec_t		prev;		/* previous extent record */
	xfs_fileoff_t		start;		/* first file offset deleted */
	int			tmp_logflags;	/* partial logging flags */
	int			wasdel;		/* was a delayed alloc extent */
	int			whichfork;	/* data or attribute fork */
	xfs_fsblock_t		sum;
L
Linus Torvalds 已提交
5132

5133
	trace_xfs_bunmap(ip, bno, len, flags, _RET_IP_);
L
Linus Torvalds 已提交
5134

5135 5136 5137 5138 5139 5140 5141 5142
	whichfork = (flags & XFS_BMAPI_ATTRFORK) ?
		XFS_ATTR_FORK : XFS_DATA_FORK;
	ifp = XFS_IFORK_PTR(ip, whichfork);
	if (unlikely(
	    XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_EXTENTS &&
	    XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_BTREE)) {
		XFS_ERROR_REPORT("xfs_bunmapi", XFS_ERRLEVEL_LOW,
				 ip->i_mount);
D
Dave Chinner 已提交
5143
		return -EFSCORRUPTED;
L
Linus Torvalds 已提交
5144
	}
5145 5146
	mp = ip->i_mount;
	if (XFS_FORCED_SHUTDOWN(mp))
D
Dave Chinner 已提交
5147
		return -EIO;
L
Linus Torvalds 已提交
5148

5149
	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
5150 5151
	ASSERT(len > 0);
	ASSERT(nexts >= 0);
L
Linus Torvalds 已提交
5152

5153 5154 5155 5156 5157 5158 5159 5160
	if (!(ifp->if_flags & XFS_IFEXTENTS) &&
	    (error = xfs_iread_extents(tp, ip, whichfork)))
		return error;
	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
	if (nextents == 0) {
		*done = 1;
		return 0;
	}
5161
	XFS_STATS_INC(mp, xs_blk_unmap);
5162 5163 5164 5165 5166
	isrt = (whichfork == XFS_DATA_FORK) && XFS_IS_REALTIME_INODE(ip);
	start = bno;
	bno = start + len - 1;
	ep = xfs_bmap_search_extents(ip, bno, whichfork, &eof, &lastx, &got,
		&prev);
L
Linus Torvalds 已提交
5167

5168 5169 5170 5171 5172 5173 5174 5175 5176 5177 5178 5179 5180 5181
	/*
	 * Check to see if the given block number is past the end of the
	 * file, back up to the last block if so...
	 */
	if (eof) {
		ep = xfs_iext_get_ext(ifp, --lastx);
		xfs_bmbt_get_all(ep, &got);
		bno = got.br_startoff + got.br_blockcount - 1;
	}
	logflags = 0;
	if (ifp->if_flags & XFS_IFBROOT) {
		ASSERT(XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_BTREE);
		cur = xfs_bmbt_init_cursor(mp, tp, ip, whichfork);
		cur->bc_private.b.firstblock = *firstblock;
5182
		cur->bc_private.b.dfops = dfops;
5183 5184 5185 5186 5187 5188 5189 5190
		cur->bc_private.b.flags = 0;
	} else
		cur = NULL;

	if (isrt) {
		/*
		 * Synchronize by locking the bitmap inode.
		 */
5191
		xfs_ilock(mp->m_rbmip, XFS_ILOCK_EXCL|XFS_ILOCK_RTBITMAP);
5192
		xfs_trans_ijoin(tp, mp->m_rbmip, XFS_ILOCK_EXCL);
5193 5194
		xfs_ilock(mp->m_rsumip, XFS_ILOCK_EXCL|XFS_ILOCK_RTSUM);
		xfs_trans_ijoin(tp, mp->m_rsumip, XFS_ILOCK_EXCL);
5195
	}
5196

5197 5198 5199 5200 5201 5202 5203 5204 5205 5206 5207 5208 5209 5210 5211 5212 5213 5214 5215 5216 5217 5218 5219 5220 5221 5222 5223 5224 5225 5226 5227 5228 5229 5230 5231 5232 5233 5234 5235
	extno = 0;
	while (bno != (xfs_fileoff_t)-1 && bno >= start && lastx >= 0 &&
	       (nexts == 0 || extno < nexts)) {
		/*
		 * Is the found extent after a hole in which bno lives?
		 * Just back up to the previous extent, if so.
		 */
		if (got.br_startoff > bno) {
			if (--lastx < 0)
				break;
			ep = xfs_iext_get_ext(ifp, lastx);
			xfs_bmbt_get_all(ep, &got);
		}
		/*
		 * Is the last block of this extent before the range
		 * we're supposed to delete?  If so, we're done.
		 */
		bno = XFS_FILEOFF_MIN(bno,
			got.br_startoff + got.br_blockcount - 1);
		if (bno < start)
			break;
		/*
		 * Then deal with the (possibly delayed) allocated space
		 * we found.
		 */
		ASSERT(ep != NULL);
		del = got;
		wasdel = isnullstartblock(del.br_startblock);
		if (got.br_startoff < start) {
			del.br_startoff = start;
			del.br_blockcount -= start - got.br_startoff;
			if (!wasdel)
				del.br_startblock += start - got.br_startoff;
		}
		if (del.br_startoff + del.br_blockcount > bno + 1)
			del.br_blockcount = bno + 1 - del.br_startoff;
		sum = del.br_startblock + del.br_blockcount;
		if (isrt &&
		    (mod = do_mod(sum, mp->m_sb.sb_rextsize))) {
5236
			/*
5237 5238 5239 5240 5241
			 * Realtime extent not lined up at the end.
			 * The extent could have been split into written
			 * and unwritten pieces, or we could just be
			 * unmapping part of it.  But we can't really
			 * get rid of part of a realtime extent.
5242
			 */
5243 5244 5245 5246 5247 5248 5249 5250 5251 5252 5253 5254 5255 5256 5257
			if (del.br_state == XFS_EXT_UNWRITTEN ||
			    !xfs_sb_version_hasextflgbit(&mp->m_sb)) {
				/*
				 * This piece is unwritten, or we're not
				 * using unwritten extents.  Skip over it.
				 */
				ASSERT(bno >= mod);
				bno -= mod > del.br_blockcount ?
					del.br_blockcount : mod;
				if (bno < got.br_startoff) {
					if (--lastx >= 0)
						xfs_bmbt_get_all(xfs_iext_get_ext(
							ifp, lastx), &got);
				}
				continue;
L
Linus Torvalds 已提交
5258
			}
T
Tao Ma 已提交
5259
			/*
5260 5261
			 * It's written, turn it unwritten.
			 * This is better than zeroing it.
T
Tao Ma 已提交
5262
			 */
5263
			ASSERT(del.br_state == XFS_EXT_NORM);
5264
			ASSERT(tp->t_blk_res > 0);
5265 5266 5267 5268 5269 5270 5271 5272 5273 5274 5275
			/*
			 * If this spans a realtime extent boundary,
			 * chop it back to the start of the one we end at.
			 */
			if (del.br_blockcount > mod) {
				del.br_startoff += del.br_blockcount - mod;
				del.br_startblock += del.br_blockcount - mod;
				del.br_blockcount = mod;
			}
			del.br_state = XFS_EXT_UNWRITTEN;
			error = xfs_bmap_add_extent_unwritten_real(tp, ip,
5276
					&lastx, &cur, &del, firstblock, dfops,
5277 5278 5279 5280 5281 5282 5283 5284 5285 5286 5287 5288 5289 5290 5291 5292 5293 5294
					&logflags);
			if (error)
				goto error0;
			goto nodelete;
		}
		if (isrt && (mod = do_mod(del.br_startblock, mp->m_sb.sb_rextsize))) {
			/*
			 * Realtime extent is lined up at the end but not
			 * at the front.  We'll get rid of full extents if
			 * we can.
			 */
			mod = mp->m_sb.sb_rextsize - mod;
			if (del.br_blockcount > mod) {
				del.br_blockcount -= mod;
				del.br_startoff += mod;
				del.br_startblock += mod;
			} else if ((del.br_startoff == start &&
				    (del.br_state == XFS_EXT_UNWRITTEN ||
5295
				     tp->t_blk_res == 0)) ||
5296 5297 5298 5299 5300 5301 5302 5303 5304 5305 5306 5307 5308 5309
				   !xfs_sb_version_hasextflgbit(&mp->m_sb)) {
				/*
				 * Can't make it unwritten.  There isn't
				 * a full extent here so just skip it.
				 */
				ASSERT(bno >= del.br_blockcount);
				bno -= del.br_blockcount;
				if (got.br_startoff > bno) {
					if (--lastx >= 0) {
						ep = xfs_iext_get_ext(ifp,
								      lastx);
						xfs_bmbt_get_all(ep, &got);
					}
				}
T
Tao Ma 已提交
5310
				continue;
5311 5312 5313 5314 5315 5316 5317 5318 5319 5320 5321 5322 5323 5324 5325 5326 5327 5328 5329 5330 5331 5332 5333 5334
			} else if (del.br_state == XFS_EXT_UNWRITTEN) {
				/*
				 * This one is already unwritten.
				 * It must have a written left neighbor.
				 * Unwrite the killed part of that one and
				 * try again.
				 */
				ASSERT(lastx > 0);
				xfs_bmbt_get_all(xfs_iext_get_ext(ifp,
						lastx - 1), &prev);
				ASSERT(prev.br_state == XFS_EXT_NORM);
				ASSERT(!isnullstartblock(prev.br_startblock));
				ASSERT(del.br_startblock ==
				       prev.br_startblock + prev.br_blockcount);
				if (prev.br_startoff < start) {
					mod = start - prev.br_startoff;
					prev.br_blockcount -= mod;
					prev.br_startblock += mod;
					prev.br_startoff = start;
				}
				prev.br_state = XFS_EXT_UNWRITTEN;
				lastx--;
				error = xfs_bmap_add_extent_unwritten_real(tp,
						ip, &lastx, &cur, &prev,
5335
						firstblock, dfops, &logflags);
5336 5337 5338 5339 5340 5341 5342 5343
				if (error)
					goto error0;
				goto nodelete;
			} else {
				ASSERT(del.br_state == XFS_EXT_NORM);
				del.br_state = XFS_EXT_UNWRITTEN;
				error = xfs_bmap_add_extent_unwritten_real(tp,
						ip, &lastx, &cur, &del,
5344
						firstblock, dfops, &logflags);
5345 5346 5347
				if (error)
					goto error0;
				goto nodelete;
T
Tao Ma 已提交
5348
			}
L
Linus Torvalds 已提交
5349 5350
		}

5351 5352 5353 5354 5355 5356 5357 5358 5359 5360 5361
		/*
		 * If it's the case where the directory code is running
		 * with no block reservation, and the deleted block is in
		 * the middle of its extent, and the resulting insert
		 * of an extent would cause transformation to btree format,
		 * then reject it.  The calling code will then swap
		 * blocks around instead.
		 * We have to do this now, rather than waiting for the
		 * conversion to btree format, since the transaction
		 * will be dirty.
		 */
5362
		if (!wasdel && tp->t_blk_res == 0 &&
5363 5364 5365 5366 5367 5368
		    XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_EXTENTS &&
		    XFS_IFORK_NEXTENTS(ip, whichfork) >= /* Note the >= */
			XFS_IFORK_MAXEXT(ip, whichfork) &&
		    del.br_startoff > got.br_startoff &&
		    del.br_startoff + del.br_blockcount <
		    got.br_startoff + got.br_blockcount) {
D
Dave Chinner 已提交
5369
			error = -ENOSPC;
5370
			goto error0;
L
Linus Torvalds 已提交
5371
		}
5372 5373 5374 5375 5376 5377 5378 5379 5380 5381 5382 5383 5384 5385 5386 5387 5388 5389 5390 5391 5392 5393 5394 5395 5396 5397 5398 5399 5400 5401

		/*
		 * Unreserve quota and update realtime free space, if
		 * appropriate. If delayed allocation, update the inode delalloc
		 * counter now and wait to update the sb counters as
		 * xfs_bmap_del_extent() might need to borrow some blocks.
		 */
		if (wasdel) {
			ASSERT(startblockval(del.br_startblock) > 0);
			if (isrt) {
				xfs_filblks_t rtexts;

				rtexts = XFS_FSB_TO_B(mp, del.br_blockcount);
				do_div(rtexts, mp->m_sb.sb_rextsize);
				xfs_mod_frextents(mp, (int64_t)rtexts);
				(void)xfs_trans_reserve_quota_nblks(NULL,
					ip, -((long)del.br_blockcount), 0,
					XFS_QMOPT_RES_RTBLKS);
			} else {
				(void)xfs_trans_reserve_quota_nblks(NULL,
					ip, -((long)del.br_blockcount), 0,
					XFS_QMOPT_RES_REGBLKS);
			}
			ip->i_delayed_blks -= del.br_blockcount;
			if (cur)
				cur->bc_private.b.flags |=
					XFS_BTCUR_BPRV_WASDEL;
		} else if (cur)
			cur->bc_private.b.flags &= ~XFS_BTCUR_BPRV_WASDEL;

5402
		error = xfs_bmap_del_extent(ip, tp, &lastx, dfops, cur, &del,
5403 5404 5405 5406
				&tmp_logflags, whichfork);
		logflags |= tmp_logflags;
		if (error)
			goto error0;
5407 5408 5409 5410

		if (!isrt && wasdel)
			xfs_mod_fdblocks(mp, (int64_t)del.br_blockcount, false);

5411 5412
		bno = del.br_startoff - 1;
nodelete:
L
Linus Torvalds 已提交
5413
		/*
5414
		 * If not done go on to the next (previous) record.
L
Linus Torvalds 已提交
5415
		 */
5416 5417 5418 5419 5420 5421 5422 5423 5424
		if (bno != (xfs_fileoff_t)-1 && bno >= start) {
			if (lastx >= 0) {
				ep = xfs_iext_get_ext(ifp, lastx);
				if (xfs_bmbt_get_startoff(ep) > bno) {
					if (--lastx >= 0)
						ep = xfs_iext_get_ext(ifp,
								      lastx);
				}
				xfs_bmbt_get_all(ep, &got);
L
Linus Torvalds 已提交
5425
			}
5426
			extno++;
L
Linus Torvalds 已提交
5427 5428
		}
	}
5429
	*done = bno == (xfs_fileoff_t)-1 || bno < start || lastx < 0;
5430

L
Linus Torvalds 已提交
5431
	/*
5432
	 * Convert to a btree if necessary.
L
Linus Torvalds 已提交
5433
	 */
5434 5435
	if (xfs_bmap_needs_btree(ip, whichfork)) {
		ASSERT(cur == NULL);
5436
		error = xfs_bmap_extents_to_btree(tp, ip, firstblock, dfops,
5437 5438 5439 5440
			&cur, 0, &tmp_logflags, whichfork);
		logflags |= tmp_logflags;
		if (error)
			goto error0;
L
Linus Torvalds 已提交
5441 5442
	}
	/*
5443
	 * transform from btree to extents, give it cur
L
Linus Torvalds 已提交
5444
	 */
5445 5446 5447 5448 5449 5450 5451 5452
	else if (xfs_bmap_wants_extents(ip, whichfork)) {
		ASSERT(cur != NULL);
		error = xfs_bmap_btree_to_extents(tp, ip, cur, &tmp_logflags,
			whichfork);
		logflags |= tmp_logflags;
		if (error)
			goto error0;
	}
L
Linus Torvalds 已提交
5453
	/*
5454
	 * transform from extents to local?
L
Linus Torvalds 已提交
5455
	 */
5456 5457 5458 5459 5460 5461 5462 5463 5464 5465 5466 5467 5468 5469 5470 5471 5472 5473 5474 5475 5476 5477 5478 5479 5480 5481 5482 5483
	error = 0;
error0:
	/*
	 * Log everything.  Do this after conversion, there's no point in
	 * logging the extent records if we've converted to btree format.
	 */
	if ((logflags & xfs_ilog_fext(whichfork)) &&
	    XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_EXTENTS)
		logflags &= ~xfs_ilog_fext(whichfork);
	else if ((logflags & xfs_ilog_fbroot(whichfork)) &&
		 XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_BTREE)
		logflags &= ~xfs_ilog_fbroot(whichfork);
	/*
	 * Log inode even in the error case, if the transaction
	 * is dirty we'll need to shut down the filesystem.
	 */
	if (logflags)
		xfs_trans_log_inode(tp, ip, logflags);
	if (cur) {
		if (!error) {
			*firstblock = cur->bc_private.b.firstblock;
			cur->bc_private.b.allocated = 0;
		}
		xfs_btree_del_cursor(cur,
			error ? XFS_BTREE_ERROR : XFS_BTREE_NOERROR);
	}
	return error;
}
5484

5485 5486 5487 5488 5489 5490 5491 5492 5493 5494 5495 5496 5497 5498 5499 5500 5501 5502 5503 5504 5505 5506 5507 5508 5509 5510 5511 5512 5513 5514 5515 5516 5517 5518 5519 5520 5521 5522 5523 5524 5525 5526 5527 5528 5529 5530 5531 5532 5533 5534 5535
/*
 * Determine whether an extent shift can be accomplished by a merge with the
 * extent that precedes the target hole of the shift.
 */
STATIC bool
xfs_bmse_can_merge(
	struct xfs_bmbt_irec	*left,	/* preceding extent */
	struct xfs_bmbt_irec	*got,	/* current extent to shift */
	xfs_fileoff_t		shift)	/* shift fsb */
{
	xfs_fileoff_t		startoff;

	startoff = got->br_startoff - shift;

	/*
	 * The extent, once shifted, must be adjacent in-file and on-disk with
	 * the preceding extent.
	 */
	if ((left->br_startoff + left->br_blockcount != startoff) ||
	    (left->br_startblock + left->br_blockcount != got->br_startblock) ||
	    (left->br_state != got->br_state) ||
	    (left->br_blockcount + got->br_blockcount > MAXEXTLEN))
		return false;

	return true;
}

/*
 * A bmap extent shift adjusts the file offset of an extent to fill a preceding
 * hole in the file. If an extent shift would result in the extent being fully
 * adjacent to the extent that currently precedes the hole, we can merge with
 * the preceding extent rather than do the shift.
 *
 * This function assumes the caller has verified a shift-by-merge is possible
 * with the provided extents via xfs_bmse_can_merge().
 */
STATIC int
xfs_bmse_merge(
	struct xfs_inode		*ip,
	int				whichfork,
	xfs_fileoff_t			shift,		/* shift fsb */
	int				current_ext,	/* idx of gotp */
	struct xfs_bmbt_rec_host	*gotp,		/* extent to shift */
	struct xfs_bmbt_rec_host	*leftp,		/* preceding extent */
	struct xfs_btree_cur		*cur,
	int				*logflags)	/* output */
{
	struct xfs_bmbt_irec		got;
	struct xfs_bmbt_irec		left;
	xfs_filblks_t			blockcount;
	int				error, i;
5536
	struct xfs_mount		*mp = ip->i_mount;
5537 5538 5539 5540 5541 5542 5543 5544 5545 5546 5547 5548 5549 5550 5551 5552 5553 5554 5555 5556 5557 5558 5559 5560 5561 5562 5563 5564 5565 5566 5567 5568 5569

	xfs_bmbt_get_all(gotp, &got);
	xfs_bmbt_get_all(leftp, &left);
	blockcount = left.br_blockcount + got.br_blockcount;

	ASSERT(xfs_isilocked(ip, XFS_IOLOCK_EXCL));
	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
	ASSERT(xfs_bmse_can_merge(&left, &got, shift));

	/*
	 * Merge the in-core extents. Note that the host record pointers and
	 * current_ext index are invalid once the extent has been removed via
	 * xfs_iext_remove().
	 */
	xfs_bmbt_set_blockcount(leftp, blockcount);
	xfs_iext_remove(ip, current_ext, 1, 0);

	/*
	 * Update the on-disk extent count, the btree if necessary and log the
	 * inode.
	 */
	XFS_IFORK_NEXT_SET(ip, whichfork,
			   XFS_IFORK_NEXTENTS(ip, whichfork) - 1);
	*logflags |= XFS_ILOG_CORE;
	if (!cur) {
		*logflags |= XFS_ILOG_DEXT;
		return 0;
	}

	/* lookup and remove the extent to merge */
	error = xfs_bmbt_lookup_eq(cur, got.br_startoff, got.br_startblock,
				   got.br_blockcount, &i);
	if (error)
D
Dave Chinner 已提交
5570
		return error;
5571
	XFS_WANT_CORRUPTED_RETURN(mp, i == 1);
5572 5573 5574

	error = xfs_btree_delete(cur, &i);
	if (error)
D
Dave Chinner 已提交
5575
		return error;
5576
	XFS_WANT_CORRUPTED_RETURN(mp, i == 1);
5577 5578 5579 5580 5581

	/* lookup and update size of the previous extent */
	error = xfs_bmbt_lookup_eq(cur, left.br_startoff, left.br_startblock,
				   left.br_blockcount, &i);
	if (error)
D
Dave Chinner 已提交
5582
		return error;
5583
	XFS_WANT_CORRUPTED_RETURN(mp, i == 1);
5584 5585 5586

	left.br_blockcount = blockcount;

D
Dave Chinner 已提交
5587 5588
	return xfs_bmbt_update(cur, left.br_startoff, left.br_startblock,
			       left.br_blockcount, left.br_state);
5589 5590
}

5591 5592 5593 5594 5595 5596 5597 5598 5599 5600 5601
/*
 * Shift a single extent.
 */
STATIC int
xfs_bmse_shift_one(
	struct xfs_inode		*ip,
	int				whichfork,
	xfs_fileoff_t			offset_shift_fsb,
	int				*current_ext,
	struct xfs_bmbt_rec_host	*gotp,
	struct xfs_btree_cur		*cur,
5602
	int				*logflags,
5603 5604
	enum shift_direction		direction,
	struct xfs_defer_ops		*dfops)
5605 5606
{
	struct xfs_ifork		*ifp;
5607
	struct xfs_mount		*mp;
5608
	xfs_fileoff_t			startoff;
5609
	struct xfs_bmbt_rec_host	*adj_irecp;
5610
	struct xfs_bmbt_irec		got;
5611
	struct xfs_bmbt_irec		adj_irec;
5612 5613
	int				error;
	int				i;
5614
	int				total_extents;
5615

5616
	mp = ip->i_mount;
5617
	ifp = XFS_IFORK_PTR(ip, whichfork);
5618
	total_extents = ifp->if_bytes / sizeof(xfs_bmbt_rec_t);
5619 5620 5621

	xfs_bmbt_get_all(gotp, &got);

5622
	/* delalloc extents should be prevented by caller */
5623
	XFS_WANT_CORRUPTED_RETURN(mp, !isnullstartblock(got.br_startblock));
5624

5625 5626 5627 5628 5629 5630 5631 5632 5633 5634 5635 5636 5637 5638 5639 5640 5641 5642 5643
	if (direction == SHIFT_LEFT) {
		startoff = got.br_startoff - offset_shift_fsb;

		/*
		 * Check for merge if we've got an extent to the left,
		 * otherwise make sure there's enough room at the start
		 * of the file for the shift.
		 */
		if (!*current_ext) {
			if (got.br_startoff < offset_shift_fsb)
				return -EINVAL;
			goto update_current_ext;
		}
		/*
		 * grab the left extent and check for a large
		 * enough hole.
		 */
		adj_irecp = xfs_iext_get_ext(ifp, *current_ext - 1);
		xfs_bmbt_get_all(adj_irecp, &adj_irec);
5644

5645 5646
		if (startoff <
		    adj_irec.br_startoff + adj_irec.br_blockcount)
5647
			return -EINVAL;
5648

5649
		/* check whether to merge the extent or shift it down */
5650 5651
		if (xfs_bmse_can_merge(&adj_irec, &got,
				       offset_shift_fsb)) {
5652 5653 5654 5655 5656 5657 5658
			error = xfs_bmse_merge(ip, whichfork, offset_shift_fsb,
					       *current_ext, gotp, adj_irecp,
					       cur, logflags);
			if (error)
				return error;
			adj_irec = got;
			goto update_rmap;
5659
		}
5660 5661 5662 5663 5664 5665 5666 5667 5668 5669 5670 5671 5672 5673 5674 5675 5676 5677 5678 5679 5680 5681 5682 5683
	} else {
		startoff = got.br_startoff + offset_shift_fsb;
		/* nothing to move if this is the last extent */
		if (*current_ext >= (total_extents - 1))
			goto update_current_ext;
		/*
		 * If this is not the last extent in the file, make sure there
		 * is enough room between current extent and next extent for
		 * accommodating the shift.
		 */
		adj_irecp = xfs_iext_get_ext(ifp, *current_ext + 1);
		xfs_bmbt_get_all(adj_irecp, &adj_irec);
		if (startoff + got.br_blockcount > adj_irec.br_startoff)
			return -EINVAL;
		/*
		 * Unlike a left shift (which involves a hole punch),
		 * a right shift does not modify extent neighbors
		 * in any way. We should never find mergeable extents
		 * in this scenario. Check anyways and warn if we
		 * encounter two extents that could be one.
		 */
		if (xfs_bmse_can_merge(&got, &adj_irec, offset_shift_fsb))
			WARN_ON_ONCE(1);
	}
5684 5685 5686 5687
	/*
	 * Increment the extent index for the next iteration, update the start
	 * offset of the in-core extent and update the btree if applicable.
	 */
5688 5689 5690 5691 5692
update_current_ext:
	if (direction == SHIFT_LEFT)
		(*current_ext)++;
	else
		(*current_ext)--;
5693 5694
	xfs_bmbt_set_startoff(gotp, startoff);
	*logflags |= XFS_ILOG_CORE;
5695
	adj_irec = got;
5696 5697
	if (!cur) {
		*logflags |= XFS_ILOG_DEXT;
5698
		goto update_rmap;
5699 5700 5701 5702 5703 5704
	}

	error = xfs_bmbt_lookup_eq(cur, got.br_startoff, got.br_startblock,
				   got.br_blockcount, &i);
	if (error)
		return error;
5705
	XFS_WANT_CORRUPTED_RETURN(mp, i == 1);
5706 5707

	got.br_startoff = startoff;
5708 5709 5710 5711 5712 5713 5714 5715 5716 5717 5718 5719
	error = xfs_bmbt_update(cur, got.br_startoff, got.br_startblock,
			got.br_blockcount, got.br_state);
	if (error)
		return error;

update_rmap:
	/* update reverse mapping */
	error = xfs_rmap_unmap_extent(mp, dfops, ip, whichfork, &adj_irec);
	if (error)
		return error;
	adj_irec.br_startoff = startoff;
	return xfs_rmap_map_extent(mp, dfops, ip, whichfork, &adj_irec);
5720 5721
}

5722
/*
5723
 * Shift extent records to the left/right to cover/create a hole.
5724
 *
5725
 * The maximum number of extents to be shifted in a single operation is
5726
 * @num_exts. @stop_fsb specifies the file offset at which to stop shift and the
5727 5728 5729 5730
 * file offset where we've left off is returned in @next_fsb. @offset_shift_fsb
 * is the length by which each extent is shifted. If there is no hole to shift
 * the extents into, this will be considered invalid operation and we abort
 * immediately.
5731 5732 5733 5734 5735
 */
int
xfs_bmap_shift_extents(
	struct xfs_trans	*tp,
	struct xfs_inode	*ip,
5736
	xfs_fileoff_t		*next_fsb,
5737
	xfs_fileoff_t		offset_shift_fsb,
5738
	int			*done,
5739
	xfs_fileoff_t		stop_fsb,
5740
	xfs_fsblock_t		*firstblock,
5741
	struct xfs_defer_ops	*dfops,
5742
	enum shift_direction	direction,
5743 5744
	int			num_exts)
{
5745
	struct xfs_btree_cur		*cur = NULL;
5746 5747 5748 5749 5750
	struct xfs_bmbt_rec_host	*gotp;
	struct xfs_bmbt_irec            got;
	struct xfs_mount		*mp = ip->i_mount;
	struct xfs_ifork		*ifp;
	xfs_extnum_t			nexts = 0;
5751
	xfs_extnum_t			current_ext;
5752 5753
	xfs_extnum_t			total_extents;
	xfs_extnum_t			stop_extent;
5754 5755
	int				error = 0;
	int				whichfork = XFS_DATA_FORK;
5756
	int				logflags = 0;
5757 5758 5759 5760 5761 5762 5763

	if (unlikely(XFS_TEST_ERROR(
	    (XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_EXTENTS &&
	     XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_BTREE),
	     mp, XFS_ERRTAG_BMAPIFORMAT, XFS_RANDOM_BMAPIFORMAT))) {
		XFS_ERROR_REPORT("xfs_bmap_shift_extents",
				 XFS_ERRLEVEL_LOW, mp);
D
Dave Chinner 已提交
5764
		return -EFSCORRUPTED;
5765 5766 5767
	}

	if (XFS_FORCED_SHUTDOWN(mp))
D
Dave Chinner 已提交
5768
		return -EIO;
5769

5770 5771
	ASSERT(xfs_isilocked(ip, XFS_IOLOCK_EXCL));
	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
5772 5773
	ASSERT(direction == SHIFT_LEFT || direction == SHIFT_RIGHT);
	ASSERT(*next_fsb != NULLFSBLOCK || direction == SHIFT_RIGHT);
5774 5775 5776 5777 5778 5779 5780 5781 5782

	ifp = XFS_IFORK_PTR(ip, whichfork);
	if (!(ifp->if_flags & XFS_IFEXTENTS)) {
		/* Read in all the extents */
		error = xfs_iread_extents(tp, ip, whichfork);
		if (error)
			return error;
	}

5783 5784 5785
	if (ifp->if_flags & XFS_IFBROOT) {
		cur = xfs_bmbt_init_cursor(mp, tp, ip, whichfork);
		cur->bc_private.b.firstblock = *firstblock;
5786
		cur->bc_private.b.dfops = dfops;
5787 5788 5789
		cur->bc_private.b.flags = 0;
	}

5790 5791 5792 5793 5794 5795 5796 5797 5798 5799 5800 5801 5802 5803 5804 5805 5806 5807 5808 5809 5810 5811 5812 5813 5814 5815 5816 5817 5818 5819 5820 5821
	/*
	 * There may be delalloc extents in the data fork before the range we
	 * are collapsing out, so we cannot use the count of real extents here.
	 * Instead we have to calculate it from the incore fork.
	 */
	total_extents = ifp->if_bytes / sizeof(xfs_bmbt_rec_t);
	if (total_extents == 0) {
		*done = 1;
		goto del_cursor;
	}

	/*
	 * In case of first right shift, we need to initialize next_fsb
	 */
	if (*next_fsb == NULLFSBLOCK) {
		gotp = xfs_iext_get_ext(ifp, total_extents - 1);
		xfs_bmbt_get_all(gotp, &got);
		*next_fsb = got.br_startoff;
		if (stop_fsb > *next_fsb) {
			*done = 1;
			goto del_cursor;
		}
	}

	/* Lookup the extent index at which we have to stop */
	if (direction == SHIFT_RIGHT) {
		gotp = xfs_iext_bno_to_ext(ifp, stop_fsb, &stop_extent);
		/* Make stop_extent exclusive of shift range */
		stop_extent--;
	} else
		stop_extent = total_extents;

5822
	/*
5823 5824 5825 5826 5827
	 * Look up the extent index for the fsb where we start shifting. We can
	 * henceforth iterate with current_ext as extent list changes are locked
	 * out via ilock.
	 *
	 * gotp can be null in 2 cases: 1) if there are no extents or 2)
5828
	 * *next_fsb lies in a hole beyond which there are no extents. Either
5829
	 * way, we are done.
5830
	 */
5831
	gotp = xfs_iext_bno_to_ext(ifp, *next_fsb, &current_ext);
5832 5833
	if (!gotp) {
		*done = 1;
5834
		goto del_cursor;
5835 5836
	}

5837 5838 5839 5840 5841 5842 5843 5844
	/* some sanity checking before we finally start shifting extents */
	if ((direction == SHIFT_LEFT && current_ext >= stop_extent) ||
	     (direction == SHIFT_RIGHT && current_ext <= stop_extent)) {
		error = -EIO;
		goto del_cursor;
	}

	while (nexts++ < num_exts) {
5845
		error = xfs_bmse_shift_one(ip, whichfork, offset_shift_fsb,
5846
					   &current_ext, gotp, cur, &logflags,
5847
					   direction, dfops);
5848 5849
		if (error)
			goto del_cursor;
5850 5851 5852 5853 5854 5855 5856 5857
		/*
		 * If there was an extent merge during the shift, the extent
		 * count can change. Update the total and grade the next record.
		 */
		if (direction == SHIFT_LEFT) {
			total_extents = ifp->if_bytes / sizeof(xfs_bmbt_rec_t);
			stop_extent = total_extents;
		}
5858

5859 5860 5861
		if (current_ext == stop_extent) {
			*done = 1;
			*next_fsb = NULLFSBLOCK;
5862
			break;
5863
		}
5864
		gotp = xfs_iext_get_ext(ifp, current_ext);
5865 5866
	}

5867
	if (!*done) {
5868
		xfs_bmbt_get_all(gotp, &got);
5869
		*next_fsb = got.br_startoff;
5870
	}
5871 5872 5873 5874 5875 5876

del_cursor:
	if (cur)
		xfs_btree_del_cursor(cur,
			error ? XFS_BTREE_ERROR : XFS_BTREE_NOERROR);

5877 5878
	if (logflags)
		xfs_trans_log_inode(tp, ip, logflags);
5879

5880 5881
	return error;
}
5882 5883 5884 5885 5886 5887 5888 5889 5890 5891 5892 5893 5894

/*
 * Splits an extent into two extents at split_fsb block such that it is
 * the first block of the current_ext. @current_ext is a target extent
 * to be split. @split_fsb is a block where the extents is split.
 * If split_fsb lies in a hole or the first block of extents, just return 0.
 */
STATIC int
xfs_bmap_split_extent_at(
	struct xfs_trans	*tp,
	struct xfs_inode	*ip,
	xfs_fileoff_t		split_fsb,
	xfs_fsblock_t		*firstfsb,
5895
	struct xfs_defer_ops	*dfops)
5896 5897 5898 5899 5900 5901 5902 5903 5904 5905 5906 5907 5908 5909 5910 5911 5912 5913 5914 5915 5916 5917 5918 5919 5920 5921 5922 5923 5924 5925 5926 5927 5928 5929 5930 5931 5932 5933 5934 5935 5936 5937 5938 5939 5940 5941 5942 5943 5944 5945 5946 5947 5948 5949 5950 5951 5952 5953 5954 5955 5956
{
	int				whichfork = XFS_DATA_FORK;
	struct xfs_btree_cur		*cur = NULL;
	struct xfs_bmbt_rec_host	*gotp;
	struct xfs_bmbt_irec		got;
	struct xfs_bmbt_irec		new; /* split extent */
	struct xfs_mount		*mp = ip->i_mount;
	struct xfs_ifork		*ifp;
	xfs_fsblock_t			gotblkcnt; /* new block count for got */
	xfs_extnum_t			current_ext;
	int				error = 0;
	int				logflags = 0;
	int				i = 0;

	if (unlikely(XFS_TEST_ERROR(
	    (XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_EXTENTS &&
	     XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_BTREE),
	     mp, XFS_ERRTAG_BMAPIFORMAT, XFS_RANDOM_BMAPIFORMAT))) {
		XFS_ERROR_REPORT("xfs_bmap_split_extent_at",
				 XFS_ERRLEVEL_LOW, mp);
		return -EFSCORRUPTED;
	}

	if (XFS_FORCED_SHUTDOWN(mp))
		return -EIO;

	ifp = XFS_IFORK_PTR(ip, whichfork);
	if (!(ifp->if_flags & XFS_IFEXTENTS)) {
		/* Read in all the extents */
		error = xfs_iread_extents(tp, ip, whichfork);
		if (error)
			return error;
	}

	/*
	 * gotp can be null in 2 cases: 1) if there are no extents
	 * or 2) split_fsb lies in a hole beyond which there are
	 * no extents. Either way, we are done.
	 */
	gotp = xfs_iext_bno_to_ext(ifp, split_fsb, &current_ext);
	if (!gotp)
		return 0;

	xfs_bmbt_get_all(gotp, &got);

	/*
	 * Check split_fsb lies in a hole or the start boundary offset
	 * of the extent.
	 */
	if (got.br_startoff >= split_fsb)
		return 0;

	gotblkcnt = split_fsb - got.br_startoff;
	new.br_startoff = split_fsb;
	new.br_startblock = got.br_startblock + gotblkcnt;
	new.br_blockcount = got.br_blockcount - gotblkcnt;
	new.br_state = got.br_state;

	if (ifp->if_flags & XFS_IFBROOT) {
		cur = xfs_bmbt_init_cursor(mp, tp, ip, whichfork);
		cur->bc_private.b.firstblock = *firstfsb;
5957
		cur->bc_private.b.dfops = dfops;
5958 5959 5960 5961 5962 5963 5964 5965 5966 5967 5968 5969 5970 5971 5972 5973 5974 5975 5976 5977 5978 5979 5980 5981 5982 5983 5984 5985 5986 5987 5988 5989 5990 5991 5992 5993 5994 5995 5996 5997 5998 5999 6000 6001 6002 6003 6004 6005 6006 6007 6008 6009
		cur->bc_private.b.flags = 0;
		error = xfs_bmbt_lookup_eq(cur, got.br_startoff,
				got.br_startblock,
				got.br_blockcount,
				&i);
		if (error)
			goto del_cursor;
		XFS_WANT_CORRUPTED_GOTO(mp, i == 1, del_cursor);
	}

	xfs_bmbt_set_blockcount(gotp, gotblkcnt);
	got.br_blockcount = gotblkcnt;

	logflags = XFS_ILOG_CORE;
	if (cur) {
		error = xfs_bmbt_update(cur, got.br_startoff,
				got.br_startblock,
				got.br_blockcount,
				got.br_state);
		if (error)
			goto del_cursor;
	} else
		logflags |= XFS_ILOG_DEXT;

	/* Add new extent */
	current_ext++;
	xfs_iext_insert(ip, current_ext, 1, &new, 0);
	XFS_IFORK_NEXT_SET(ip, whichfork,
			   XFS_IFORK_NEXTENTS(ip, whichfork) + 1);

	if (cur) {
		error = xfs_bmbt_lookup_eq(cur, new.br_startoff,
				new.br_startblock, new.br_blockcount,
				&i);
		if (error)
			goto del_cursor;
		XFS_WANT_CORRUPTED_GOTO(mp, i == 0, del_cursor);
		cur->bc_rec.b.br_state = new.br_state;

		error = xfs_btree_insert(cur, &i);
		if (error)
			goto del_cursor;
		XFS_WANT_CORRUPTED_GOTO(mp, i == 1, del_cursor);
	}

	/*
	 * Convert to a btree if necessary.
	 */
	if (xfs_bmap_needs_btree(ip, whichfork)) {
		int tmp_logflags; /* partial log flag return val */

		ASSERT(cur == NULL);
6010
		error = xfs_bmap_extents_to_btree(tp, ip, firstfsb, dfops,
6011 6012 6013 6014 6015 6016 6017 6018 6019 6020 6021 6022 6023 6024 6025 6026 6027 6028 6029 6030 6031 6032 6033
				&cur, 0, &tmp_logflags, whichfork);
		logflags |= tmp_logflags;
	}

del_cursor:
	if (cur) {
		cur->bc_private.b.allocated = 0;
		xfs_btree_del_cursor(cur,
				error ? XFS_BTREE_ERROR : XFS_BTREE_NOERROR);
	}

	if (logflags)
		xfs_trans_log_inode(tp, ip, logflags);
	return error;
}

int
xfs_bmap_split_extent(
	struct xfs_inode        *ip,
	xfs_fileoff_t           split_fsb)
{
	struct xfs_mount        *mp = ip->i_mount;
	struct xfs_trans        *tp;
6034
	struct xfs_defer_ops    dfops;
6035 6036 6037
	xfs_fsblock_t           firstfsb;
	int                     error;

6038 6039 6040
	error = xfs_trans_alloc(mp, &M_RES(mp)->tr_write,
			XFS_DIOSTRAT_SPACE_RES(mp, 0), 0, 0, &tp);
	if (error)
6041 6042 6043 6044 6045
		return error;

	xfs_ilock(ip, XFS_ILOCK_EXCL);
	xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL);

6046
	xfs_defer_init(&dfops, &firstfsb);
6047 6048

	error = xfs_bmap_split_extent_at(tp, ip, split_fsb,
6049
			&firstfsb, &dfops);
6050 6051 6052
	if (error)
		goto out;

6053
	error = xfs_defer_finish(&tp, &dfops, NULL);
6054 6055 6056
	if (error)
		goto out;

6057
	return xfs_trans_commit(tp);
6058 6059

out:
6060
	xfs_defer_cancel(&dfops);
6061
	xfs_trans_cancel(tp);
6062 6063
	return error;
}