xfs_bmap.c 181.2 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"
L
Linus Torvalds 已提交
20
#include "xfs_types.h"
21
#include "xfs_bit.h"
L
Linus Torvalds 已提交
22
#include "xfs_log.h"
23
#include "xfs_inum.h"
L
Linus Torvalds 已提交
24 25 26 27
#include "xfs_trans.h"
#include "xfs_sb.h"
#include "xfs_ag.h"
#include "xfs_dir2.h"
28
#include "xfs_da_btree.h"
L
Linus Torvalds 已提交
29
#include "xfs_bmap_btree.h"
30
#include "xfs_alloc_btree.h"
L
Linus Torvalds 已提交
31 32 33 34
#include "xfs_ialloc_btree.h"
#include "xfs_dir2_sf.h"
#include "xfs_dinode.h"
#include "xfs_inode.h"
35 36
#include "xfs_btree.h"
#include "xfs_mount.h"
L
Linus Torvalds 已提交
37
#include "xfs_itable.h"
38 39 40
#include "xfs_dir2_data.h"
#include "xfs_dir2_leaf.h"
#include "xfs_dir2_block.h"
41
#include "xfs_inode_item.h"
L
Linus Torvalds 已提交
42 43 44 45 46
#include "xfs_extfree_item.h"
#include "xfs_alloc.h"
#include "xfs_bmap.h"
#include "xfs_rtalloc.h"
#include "xfs_error.h"
47
#include "xfs_attr_leaf.h"
L
Linus Torvalds 已提交
48 49 50 51
#include "xfs_rw.h"
#include "xfs_quota.h"
#include "xfs_trans_space.h"
#include "xfs_buf_item.h"
52
#include "xfs_filestream.h"
53
#include "xfs_vnodeops.h"
C
Christoph Hellwig 已提交
54
#include "xfs_trace.h"
L
Linus Torvalds 已提交
55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91


#ifdef DEBUG
STATIC void
xfs_bmap_check_leaf_extents(xfs_btree_cur_t *cur, xfs_inode_t *ip, int whichfork);
#endif

kmem_zone_t		*xfs_bmap_free_item_zone;

/*
 * Prototypes for internal bmap routines.
 */


/*
 * 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 */
	xfs_bmap_free_t		*flist,		/* blocks to free at commit */
	int			*flags);	/* inode logging flags */

/*
 * Called from xfs_bmap_add_attrfork to handle local format files.
 */
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 */
	xfs_bmap_free_t		*flist,		/* blocks to free at commit */
	int			*flags);	/* inode logging flags */

/*
92
 * Called by xfs_bmapi to update file extent records and the btree
L
Linus Torvalds 已提交
93 94 95 96 97 98 99
 * after allocating space (or doing a delayed allocation).
 */
STATIC int				/* error */
xfs_bmap_add_extent(
	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 */
100
	xfs_bmbt_irec_t		*new,	/* new data to add to file extents */
L
Linus Torvalds 已提交
101 102 103 104 105 106 107 108 109 110 111 112 113 114 115
	xfs_fsblock_t		*first,	/* pointer to firstblock variable */
	xfs_bmap_free_t		*flist,	/* list of extents to be freed */
	int			*logflagsp, /* inode logging flags */
	int			whichfork, /* data or attr fork */
	int			rsvd);	/* OK to allocate reserved blocks */

/*
 * Called by xfs_bmap_add_extent to handle cases converting a delayed
 * allocation to a real allocation.
 */
STATIC int				/* error */
xfs_bmap_add_extent_delay_real(
	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 */
116
	xfs_bmbt_irec_t		*new,	/* new data to add to file extents */
L
Linus Torvalds 已提交
117 118 119 120 121 122 123 124 125 126 127 128 129 130
	xfs_filblks_t		*dnew,	/* new delayed-alloc indirect blocks */
	xfs_fsblock_t		*first,	/* pointer to firstblock variable */
	xfs_bmap_free_t		*flist,	/* list of extents to be freed */
	int			*logflagsp, /* inode logging flags */
	int			rsvd);	/* OK to allocate reserved blocks */

/*
 * Called by xfs_bmap_add_extent to handle cases converting a hole
 * to a delayed allocation.
 */
STATIC int				/* error */
xfs_bmap_add_extent_hole_delay(
	xfs_inode_t		*ip,	/* incore inode pointer */
	xfs_extnum_t		idx,	/* extent number to update/insert */
131
	xfs_bmbt_irec_t		*new,	/* new data to add to file extents */
L
Linus Torvalds 已提交
132 133 134 135 136 137 138 139 140 141 142 143
	int			*logflagsp,/* inode logging flags */
	int			rsvd);	/* OK to allocate reserved blocks */

/*
 * Called by xfs_bmap_add_extent to handle cases converting a hole
 * to a real allocation.
 */
STATIC int				/* error */
xfs_bmap_add_extent_hole_real(
	xfs_inode_t		*ip,	/* incore inode pointer */
	xfs_extnum_t		idx,	/* extent number to update/insert */
	xfs_btree_cur_t		*cur,	/* if null, not a btree */
144
	xfs_bmbt_irec_t		*new,	/* new data to add to file extents */
L
Linus Torvalds 已提交
145 146 147 148 149 150 151 152 153 154 155 156
	int			*logflagsp, /* inode logging flags */
	int			whichfork); /* data or attr fork */

/*
 * Called by xfs_bmap_add_extent to handle cases converting an unwritten
 * allocation to a real allocation or vice versa.
 */
STATIC int				/* error */
xfs_bmap_add_extent_unwritten_real(
	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 */
157
	xfs_bmbt_irec_t		*new,	/* new data to add to file extents */
158
	int			*logflagsp); /* inode logging flags */
L
Linus Torvalds 已提交
159 160 161 162 163 164 165 166 167 168 169 170

/*
 * 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				/* error */
xfs_bmap_alloc(
	xfs_bmalloca_t		*ap);	/* bmap alloc argument struct */

/*
 * Transform a btree format file with only one leaf node, where the
 * extents list will fit in the inode, into an extents format file.
171
 * Since the file extents are already in-core, all we have to do is
L
Linus Torvalds 已提交
172 173 174 175 176 177 178 179 180 181 182
 * 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 */

/*
183
 * Called by xfs_bmapi to update file extent records and the btree
L
Linus Torvalds 已提交
184 185 186 187 188 189 190 191 192
 * 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 trans pointer */
	xfs_extnum_t		idx,	/* extent number to update/insert */
	xfs_bmap_free_t		*flist,	/* list of extents to be freed */
	xfs_btree_cur_t		*cur,	/* if null, not a btree */
193
	xfs_bmbt_irec_t		*new,	/* new data to add to file extents */
L
Linus Torvalds 已提交
194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243
	int			*logflagsp,/* inode logging flags */
	int			whichfork, /* data or attr fork */
	int			rsvd);	 /* OK to allocate reserved blocks */

/*
 * Remove the entry "free" from the free item list.  Prev points to the
 * previous entry, unless "free" is the head of the list.
 */
STATIC void
xfs_bmap_del_free(
	xfs_bmap_free_t		*flist,	/* free item list header */
	xfs_bmap_free_item_t	*prev,	/* previous item on list, if any */
	xfs_bmap_free_item_t	*free);	/* list item to be freed */

/*
 * 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 */
	xfs_bmap_free_t		*flist,		/* blocks freed in xaction */
	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 */

/*
 * Convert a local file to an extents file.
 * This code is sort of bogus, since the file data needs to get
 * logged so it won't be lost.  The bmap-level manipulations are ok, though.
 */
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);	/* data or attr fork */

/*
 * 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.
 */
244
STATIC xfs_bmbt_rec_host_t *		/* pointer to found extent entry */
L
Linus Torvalds 已提交
245 246 247 248 249 250 251 252 253
xfs_bmap_search_extents(
	xfs_inode_t	*ip,		/* incore inode pointer */
	xfs_fileoff_t	bno,		/* block number searched for */
	int		whichfork,	/* 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 */

254 255 256 257 258 259 260 261 262 263 264 265 266
/*
 * 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.
 */
STATIC int				/* error */
xfs_bmap_isaeof(
	xfs_inode_t	*ip,		/* incore inode pointer */
	xfs_fileoff_t   off,		/* file offset in fsblocks */
	int             whichfork,	/* data or attribute fork */
	char		*aeof);		/* return value */

L
Linus Torvalds 已提交
267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296
/*
 * Compute the worst-case number of indirect blocks that will be used
 * for ip's delayed extent of length "len".
 */
STATIC xfs_filblks_t
xfs_bmap_worst_indlen(
	xfs_inode_t		*ip,	/* incore inode pointer */
	xfs_filblks_t		len);	/* delayed extent length */

#ifdef DEBUG
/*
 * Perform various validation checks on the values being returned
 * from xfs_bmapi().
 */
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);
#else
#define	xfs_bmap_validate_ret(bno,len,flags,mval,onmap,nmap)
#endif /* DEBUG */

STATIC int
xfs_bmap_count_tree(
	xfs_mount_t     *mp,
	xfs_trans_t     *tp,
297
	xfs_ifork_t	*ifp,
L
Linus Torvalds 已提交
298 299 300 301
	xfs_fsblock_t   blockno,
	int             levelin,
	int		*count);

302
STATIC void
L
Linus Torvalds 已提交
303
xfs_bmap_count_leaves(
304 305
	xfs_ifork_t		*ifp,
	xfs_extnum_t		idx,
L
Linus Torvalds 已提交
306 307 308
	int			numrecs,
	int			*count);

309
STATIC void
310
xfs_bmap_disk_count_leaves(
311
	struct xfs_mount	*mp,
312
	struct xfs_btree_block	*block,
313 314 315
	int			numrecs,
	int			*count);

L
Linus Torvalds 已提交
316 317 318 319
/*
 * Bmap internal routines.
 */

320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347
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);
}

348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365
/*
* Update the record referred to by cur to the value given
 * 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);
}
366

L
Linus Torvalds 已提交
367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386
/*
 * 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 */
	xfs_bmap_free_t		*flist,		/* blocks to free at commit */
	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 */

	mp = ip->i_mount;
	if (ip->i_df.if_broot_bytes <= XFS_IFORK_DSIZE(ip))
		*flags |= XFS_ILOG_DBROOT;
	else {
387
		cur = xfs_bmbt_init_cursor(mp, tp, ip, XFS_DATA_FORK);
L
Linus Torvalds 已提交
388 389 390 391
		cur->bc_private.b.flist = flist;
		cur->bc_private.b.firstblock = *firstblock;
		if ((error = xfs_bmbt_lookup_ge(cur, 0, 0, 0, &stat)))
			goto error0;
392 393
		/* must be at least one entry */
		XFS_WANT_CORRUPTED_GOTO(stat == 1, error0);
394
		if ((error = xfs_btree_new_iroot(cur, flags, &stat)))
L
Linus Torvalds 已提交
395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462
			goto error0;
		if (stat == 0) {
			xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR);
			return XFS_ERROR(ENOSPC);
		}
		*firstblock = cur->bc_private.b.firstblock;
		cur->bc_private.b.allocated = 0;
		xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR);
	}
	return 0;
error0:
	xfs_btree_del_cursor(cur, XFS_BTREE_ERROR);
	return error;
}

/*
 * 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 */
	xfs_bmap_free_t		*flist,		/* blocks to free at commit */
	int			*flags)		/* inode logging flags */
{
	xfs_btree_cur_t		*cur;		/* bmap btree cursor */
	int			error;		/* error return value */

	if (ip->i_d.di_nextents * sizeof(xfs_bmbt_rec_t) <= XFS_IFORK_DSIZE(ip))
		return 0;
	cur = NULL;
	error = xfs_bmap_extents_to_btree(tp, ip, firstblock, flist, &cur, 0,
		flags, XFS_DATA_FORK);
	if (cur) {
		cur->bc_private.b.allocated = 0;
		xfs_btree_del_cursor(cur,
			error ? XFS_BTREE_ERROR : XFS_BTREE_NOERROR);
	}
	return error;
}

/*
 * Called from xfs_bmap_add_attrfork to handle local format files.
 */
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 */
	xfs_bmap_free_t		*flist,		/* blocks to free at commit */
	int			*flags)		/* inode logging flags */
{
	xfs_da_args_t		dargs;		/* args for dir/attr code */
	int			error;		/* error return value */
	xfs_mount_t		*mp;		/* mount structure pointer */

	if (ip->i_df.if_bytes <= XFS_IFORK_DSIZE(ip))
		return 0;
	if ((ip->i_d.di_mode & S_IFMT) == S_IFDIR) {
		mp = ip->i_mount;
		memset(&dargs, 0, sizeof(dargs));
		dargs.dp = ip;
		dargs.firstblock = firstblock;
		dargs.flist = flist;
		dargs.total = mp->m_dirblkfsbs;
		dargs.whichfork = XFS_DATA_FORK;
		dargs.trans = tp;
463
		error = xfs_dir2_sf_to_block(&dargs);
L
Linus Torvalds 已提交
464 465 466 467 468 469 470
	} else
		error = xfs_bmap_local_to_extents(tp, ip, firstblock, 1, flags,
			XFS_DATA_FORK);
	return error;
}

/*
471
 * Called by xfs_bmapi to update file extent records and the btree
L
Linus Torvalds 已提交
472 473 474 475 476 477 478
 * after allocating space (or doing a delayed allocation).
 */
STATIC int				/* error */
xfs_bmap_add_extent(
	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 */
479
	xfs_bmbt_irec_t		*new,	/* new data to add to file extents */
L
Linus Torvalds 已提交
480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506
	xfs_fsblock_t		*first,	/* pointer to firstblock variable */
	xfs_bmap_free_t		*flist,	/* list of extents to be freed */
	int			*logflagsp, /* inode logging flags */
	int			whichfork, /* data or attr fork */
	int			rsvd)	/* OK to use reserved data blocks */
{
	xfs_btree_cur_t		*cur;	/* btree cursor or null */
	xfs_filblks_t		da_new; /* new count del alloc blocks used */
	xfs_filblks_t		da_old; /* old count del alloc blocks used */
	int			error;	/* error return value */
	xfs_ifork_t		*ifp;	/* inode fork ptr */
	int			logflags; /* returned value */
	xfs_extnum_t		nextents; /* number of extents in file now */

	XFS_STATS_INC(xs_add_exlist);
	cur = *curp;
	ifp = XFS_IFORK_PTR(ip, whichfork);
	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
	ASSERT(idx <= nextents);
	da_old = da_new = 0;
	error = 0;
	/*
	 * This is the first extent added to a new/empty file.
	 * Special case this one, so other routines get to assume there are
	 * already extents in the list.
	 */
	if (nextents == 0) {
507 508 509
		xfs_iext_insert(ip, 0, 1, new,
				whichfork == XFS_ATTR_FORK ? BMAP_ATTRFORK : 0);

L
Linus Torvalds 已提交
510 511
		ASSERT(cur == NULL);
		ifp->if_lastex = 0;
512
		if (!isnullstartblock(new->br_startblock)) {
L
Linus Torvalds 已提交
513
			XFS_IFORK_NEXT_SET(ip, whichfork, 1);
514
			logflags = XFS_ILOG_CORE | xfs_ilog_fext(whichfork);
L
Linus Torvalds 已提交
515 516 517 518 519 520
		} else
			logflags = 0;
	}
	/*
	 * Any kind of new delayed allocation goes here.
	 */
521
	else if (isnullstartblock(new->br_startblock)) {
L
Linus Torvalds 已提交
522 523 524
		if (cur)
			ASSERT((cur->bc_private.b.flags &
				XFS_BTCUR_BPRV_WASDEL) == 0);
525
		if ((error = xfs_bmap_add_extent_hole_delay(ip, idx, new,
526
				&logflags, rsvd)))
L
Linus Torvalds 已提交
527 528 529 530 531 532 533 534 535 536
			goto done;
	}
	/*
	 * Real allocation off the end of the file.
	 */
	else if (idx == nextents) {
		if (cur)
			ASSERT((cur->bc_private.b.flags &
				XFS_BTCUR_BPRV_WASDEL) == 0);
		if ((error = xfs_bmap_add_extent_hole_real(ip, idx, cur, new,
537
				&logflags, whichfork)))
L
Linus Torvalds 已提交
538 539 540 541 542 543 544
			goto done;
	} else {
		xfs_bmbt_irec_t	prev;	/* old extent at offset idx */

		/*
		 * Get the record referred to by idx.
		 */
545
		xfs_bmbt_get_all(xfs_iext_get_ext(ifp, idx), &prev);
L
Linus Torvalds 已提交
546 547 548 549 550 551
		/*
		 * If it's a real allocation record, and the new allocation ends
		 * after the start of the referred to record, then we're filling
		 * in a delayed or unwritten allocation with a real one, or
		 * converting real back to unwritten.
		 */
552
		if (!isnullstartblock(new->br_startblock) &&
L
Linus Torvalds 已提交
553 554
		    new->br_startoff + new->br_blockcount > prev.br_startoff) {
			if (prev.br_state != XFS_EXT_UNWRITTEN &&
555 556
			    isnullstartblock(prev.br_startblock)) {
				da_old = startblockval(prev.br_startblock);
L
Linus Torvalds 已提交
557 558 559 560 561
				if (cur)
					ASSERT(cur->bc_private.b.flags &
						XFS_BTCUR_BPRV_WASDEL);
				if ((error = xfs_bmap_add_extent_delay_real(ip,
					idx, &cur, new, &da_new, first, flist,
562
					&logflags, rsvd)))
L
Linus Torvalds 已提交
563 564 565 566
					goto done;
			} else if (new->br_state == XFS_EXT_NORM) {
				ASSERT(new->br_state == XFS_EXT_NORM);
				if ((error = xfs_bmap_add_extent_unwritten_real(
567
					ip, idx, &cur, new, &logflags)))
L
Linus Torvalds 已提交
568 569 570 571
					goto done;
			} else {
				ASSERT(new->br_state == XFS_EXT_UNWRITTEN);
				if ((error = xfs_bmap_add_extent_unwritten_real(
572
					ip, idx, &cur, new, &logflags)))
L
Linus Torvalds 已提交
573 574 575 576 577 578 579 580 581 582 583 584
					goto done;
			}
			ASSERT(*curp == cur || *curp == NULL);
		}
		/*
		 * Otherwise we're filling in a hole with an allocation.
		 */
		else {
			if (cur)
				ASSERT((cur->bc_private.b.flags &
					XFS_BTCUR_BPRV_WASDEL) == 0);
			if ((error = xfs_bmap_add_extent_hole_real(ip, idx, cur,
585
					new, &logflags, whichfork)))
L
Linus Torvalds 已提交
586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617
				goto done;
		}
	}

	ASSERT(*curp == cur || *curp == NULL);
	/*
	 * Convert to a btree if necessary.
	 */
	if (XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_EXTENTS &&
	    XFS_IFORK_NEXTENTS(ip, whichfork) > ifp->if_ext_max) {
		int	tmp_logflags;	/* partial log flag return val */

		ASSERT(cur == NULL);
		error = xfs_bmap_extents_to_btree(ip->i_transp, ip, first,
			flist, &cur, da_old > 0, &tmp_logflags, whichfork);
		logflags |= tmp_logflags;
		if (error)
			goto done;
	}
	/*
	 * Adjust for changes in reserved delayed indirect blocks.
	 * Nothing to do for disk quotas here.
	 */
	if (da_old || da_new) {
		xfs_filblks_t	nblks;

		nblks = da_new;
		if (cur)
			nblks += cur->bc_private.b.allocated;
		ASSERT(nblks <= da_old);
		if (nblks < da_old)
			xfs_mod_incore_sb(ip->i_mount, XFS_SBS_FDBLOCKS,
618
				(int64_t)(da_old - nblks), rsvd);
L
Linus Torvalds 已提交
619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644
	}
	/*
	 * Clear out the allocated field, done with it now in any case.
	 */
	if (cur) {
		cur->bc_private.b.allocated = 0;
		*curp = cur;
	}
done:
#ifdef DEBUG
	if (!error)
		xfs_bmap_check_leaf_extents(*curp, ip, whichfork);
#endif
	*logflagsp = logflags;
	return error;
}

/*
 * Called by xfs_bmap_add_extent to handle cases converting a delayed
 * allocation to a real allocation.
 */
STATIC int				/* error */
xfs_bmap_add_extent_delay_real(
	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 */
645
	xfs_bmbt_irec_t		*new,	/* new data to add to file extents */
L
Linus Torvalds 已提交
646 647 648 649 650 651 652 653
	xfs_filblks_t		*dnew,	/* new delayed-alloc indirect blocks */
	xfs_fsblock_t		*first,	/* pointer to firstblock variable */
	xfs_bmap_free_t		*flist,	/* list of extents to be freed */
	int			*logflagsp, /* inode logging flags */
	int			rsvd)	/* OK to use reserved data block allocation */
{
	xfs_btree_cur_t		*cur;	/* btree cursor */
	int			diff;	/* temp value */
654
	xfs_bmbt_rec_host_t	*ep;	/* extent entry for idx */
L
Linus Torvalds 已提交
655 656
	int			error;	/* error return value */
	int			i;	/* temp state */
657
	xfs_ifork_t		*ifp;	/* inode fork pointer */
L
Linus Torvalds 已提交
658 659 660 661 662
	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 */
663 664
	xfs_filblks_t		temp=0;	/* value for dnew calculations */
	xfs_filblks_t		temp2=0;/* value for dnew calculations */
L
Linus Torvalds 已提交
665 666 667 668 669 670 671 672 673 674
	int			tmp_rval;	/* partial logging flags */

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

	/*
	 * Set up a bunch of variables to make the tests simpler.
	 */
	cur = *curp;
675 676
	ifp = XFS_IFORK_PTR(ip, XFS_DATA_FORK);
	ep = xfs_iext_get_ext(ifp, idx);
L
Linus Torvalds 已提交
677 678 679 680
	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);
681

L
Linus Torvalds 已提交
682 683 684 685
	/*
	 * Set flags determining what part of the previous delayed allocation
	 * extent is being replaced by a real allocation.
	 */
686 687 688 689 690
	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
691 692 693 694
	/*
	 * Check and set flags if this segment has a left neighbor.
	 * Don't set contiguous if the combined extent would be too large.
	 */
695 696
	if (idx > 0) {
		state |= BMAP_LEFT_VALID;
697
		xfs_bmbt_get_all(xfs_iext_get_ext(ifp, idx - 1), &LEFT);
698 699 700

		if (isnullstartblock(LEFT.br_startblock))
			state |= BMAP_LEFT_DELAY;
L
Linus Torvalds 已提交
701
	}
702 703 704 705 706 707 708 709

	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;

L
Linus Torvalds 已提交
710 711 712 713 714
	/*
	 * 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.
	 */
715 716
	if (idx < ip->i_df.if_bytes / (uint)sizeof(xfs_bmbt_rec_t) - 1) {
		state |= BMAP_RIGHT_VALID;
717
		xfs_bmbt_get_all(xfs_iext_get_ext(ifp, idx + 1), &RIGHT);
718 719 720

		if (isnullstartblock(RIGHT.br_startblock))
			state |= BMAP_RIGHT_DELAY;
L
Linus Torvalds 已提交
721
	}
722 723 724 725 726 727 728 729 730 731 732 733 734 735

	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;

L
Linus Torvalds 已提交
736 737 738 739
	error = 0;
	/*
	 * Switch out based on the FILLING and CONTIG state bits.
	 */
740 741 742 743
	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
744 745 746 747
		/*
		 * Filling in all of a previously delayed allocation extent.
		 * The left and right neighbors are both contiguous with new.
		 */
C
Christoph Hellwig 已提交
748
		trace_xfs_bmap_pre_update(ip, idx - 1, state, _THIS_IP_);
749
		xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, idx - 1),
L
Linus Torvalds 已提交
750 751
			LEFT.br_blockcount + PREV.br_blockcount +
			RIGHT.br_blockcount);
C
Christoph Hellwig 已提交
752 753
		trace_xfs_bmap_post_update(ip, idx - 1, state, _THIS_IP_);

754
		xfs_iext_remove(ip, idx, 2, state);
L
Linus Torvalds 已提交
755 756 757 758 759 760 761 762 763 764
		ip->i_df.if_lastex = idx - 1;
		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;
765
			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
766
			if ((error = xfs_btree_delete(cur, &i)))
L
Linus Torvalds 已提交
767
				goto done;
768
			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
769
			if ((error = xfs_btree_decrement(cur, 0, &i)))
L
Linus Torvalds 已提交
770
				goto done;
771
			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
L
Linus Torvalds 已提交
772 773 774 775 776 777 778 779 780 781
			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;
		}
		*dnew = 0;
		break;

782
	case BMAP_LEFT_FILLING | BMAP_RIGHT_FILLING | BMAP_LEFT_CONTIG:
L
Linus Torvalds 已提交
783 784 785 786
		/*
		 * Filling in all of a previously delayed allocation extent.
		 * The left neighbor is contiguous, the right is not.
		 */
C
Christoph Hellwig 已提交
787
		trace_xfs_bmap_pre_update(ip, idx - 1, state, _THIS_IP_);
788
		xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, idx - 1),
L
Linus Torvalds 已提交
789
			LEFT.br_blockcount + PREV.br_blockcount);
C
Christoph Hellwig 已提交
790 791
		trace_xfs_bmap_post_update(ip, idx - 1, state, _THIS_IP_);

L
Linus Torvalds 已提交
792
		ip->i_df.if_lastex = idx - 1;
793
		xfs_iext_remove(ip, idx, 1, state);
L
Linus Torvalds 已提交
794 795 796 797 798 799 800 801
		if (cur == NULL)
			rval = XFS_ILOG_DEXT;
		else {
			rval = 0;
			if ((error = xfs_bmbt_lookup_eq(cur, LEFT.br_startoff,
					LEFT.br_startblock, LEFT.br_blockcount,
					&i)))
				goto done;
802
			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
L
Linus Torvalds 已提交
803 804 805 806 807 808 809 810 811
			if ((error = xfs_bmbt_update(cur, LEFT.br_startoff,
					LEFT.br_startblock,
					LEFT.br_blockcount +
					PREV.br_blockcount, LEFT.br_state)))
				goto done;
		}
		*dnew = 0;
		break;

812
	case BMAP_LEFT_FILLING | BMAP_RIGHT_FILLING | BMAP_RIGHT_CONTIG:
L
Linus Torvalds 已提交
813 814 815 816
		/*
		 * Filling in all of a previously delayed allocation extent.
		 * The right neighbor is contiguous, the left is not.
		 */
C
Christoph Hellwig 已提交
817
		trace_xfs_bmap_pre_update(ip, idx, state, _THIS_IP_);
L
Linus Torvalds 已提交
818 819 820
		xfs_bmbt_set_startblock(ep, new->br_startblock);
		xfs_bmbt_set_blockcount(ep,
			PREV.br_blockcount + RIGHT.br_blockcount);
C
Christoph Hellwig 已提交
821 822
		trace_xfs_bmap_post_update(ip, idx, state, _THIS_IP_);

L
Linus Torvalds 已提交
823
		ip->i_df.if_lastex = idx;
824
		xfs_iext_remove(ip, idx + 1, 1, state);
L
Linus Torvalds 已提交
825 826 827 828 829 830 831 832
		if (cur == NULL)
			rval = XFS_ILOG_DEXT;
		else {
			rval = 0;
			if ((error = xfs_bmbt_lookup_eq(cur, RIGHT.br_startoff,
					RIGHT.br_startblock,
					RIGHT.br_blockcount, &i)))
				goto done;
833
			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
L
Linus Torvalds 已提交
834 835 836 837 838 839 840 841 842
			if ((error = xfs_bmbt_update(cur, PREV.br_startoff,
					new->br_startblock,
					PREV.br_blockcount +
					RIGHT.br_blockcount, PREV.br_state)))
				goto done;
		}
		*dnew = 0;
		break;

843
	case BMAP_LEFT_FILLING | BMAP_RIGHT_FILLING:
L
Linus Torvalds 已提交
844 845 846 847 848
		/*
		 * Filling in all of a previously delayed allocation extent.
		 * Neither the left nor right neighbors are contiguous with
		 * the new one.
		 */
C
Christoph Hellwig 已提交
849
		trace_xfs_bmap_pre_update(ip, idx, state, _THIS_IP_);
L
Linus Torvalds 已提交
850
		xfs_bmbt_set_startblock(ep, new->br_startblock);
C
Christoph Hellwig 已提交
851 852
		trace_xfs_bmap_post_update(ip, idx, state, _THIS_IP_);

L
Linus Torvalds 已提交
853 854 855 856 857 858 859 860 861 862
		ip->i_df.if_lastex = idx;
		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, new->br_startoff,
					new->br_startblock, new->br_blockcount,
					&i)))
				goto done;
863
			XFS_WANT_CORRUPTED_GOTO(i == 0, done);
L
Linus Torvalds 已提交
864
			cur->bc_rec.b.br_state = XFS_EXT_NORM;
865
			if ((error = xfs_btree_insert(cur, &i)))
L
Linus Torvalds 已提交
866
				goto done;
867
			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
L
Linus Torvalds 已提交
868 869 870 871
		}
		*dnew = 0;
		break;

872
	case BMAP_LEFT_FILLING | BMAP_LEFT_CONTIG:
L
Linus Torvalds 已提交
873 874 875 876
		/*
		 * Filling in the first part of a previous delayed allocation.
		 * The left neighbor is contiguous.
		 */
C
Christoph Hellwig 已提交
877
		trace_xfs_bmap_pre_update(ip, idx - 1, state, _THIS_IP_);
878
		xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, idx - 1),
L
Linus Torvalds 已提交
879 880 881
			LEFT.br_blockcount + new->br_blockcount);
		xfs_bmbt_set_startoff(ep,
			PREV.br_startoff + new->br_blockcount);
C
Christoph Hellwig 已提交
882 883
		trace_xfs_bmap_post_update(ip, idx - 1, state, _THIS_IP_);

L
Linus Torvalds 已提交
884
		temp = PREV.br_blockcount - new->br_blockcount;
C
Christoph Hellwig 已提交
885
		trace_xfs_bmap_pre_update(ip, idx, state, _THIS_IP_);
L
Linus Torvalds 已提交
886 887 888 889 890 891 892 893 894 895
		xfs_bmbt_set_blockcount(ep, temp);
		ip->i_df.if_lastex = idx - 1;
		if (cur == NULL)
			rval = XFS_ILOG_DEXT;
		else {
			rval = 0;
			if ((error = xfs_bmbt_lookup_eq(cur, LEFT.br_startoff,
					LEFT.br_startblock, LEFT.br_blockcount,
					&i)))
				goto done;
896
			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
L
Linus Torvalds 已提交
897 898 899 900 901 902 903 904
			if ((error = xfs_bmbt_update(cur, LEFT.br_startoff,
					LEFT.br_startblock,
					LEFT.br_blockcount +
					new->br_blockcount,
					LEFT.br_state)))
				goto done;
		}
		temp = XFS_FILBLKS_MIN(xfs_bmap_worst_indlen(ip, temp),
905 906
			startblockval(PREV.br_startblock));
		xfs_bmbt_set_startblock(ep, nullstartblock((int)temp));
C
Christoph Hellwig 已提交
907
		trace_xfs_bmap_post_update(ip, idx, state, _THIS_IP_);
L
Linus Torvalds 已提交
908 909 910
		*dnew = temp;
		break;

911
	case BMAP_LEFT_FILLING:
L
Linus Torvalds 已提交
912 913 914 915
		/*
		 * Filling in the first part of a previous delayed allocation.
		 * The left neighbor is not contiguous.
		 */
C
Christoph Hellwig 已提交
916
		trace_xfs_bmap_pre_update(ip, idx, state, _THIS_IP_);
L
Linus Torvalds 已提交
917 918 919
		xfs_bmbt_set_startoff(ep, new_endoff);
		temp = PREV.br_blockcount - new->br_blockcount;
		xfs_bmbt_set_blockcount(ep, temp);
920
		xfs_iext_insert(ip, idx, 1, new, state);
L
Linus Torvalds 已提交
921 922 923 924 925 926 927 928 929 930
		ip->i_df.if_lastex = idx;
		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, new->br_startoff,
					new->br_startblock, new->br_blockcount,
					&i)))
				goto done;
931
			XFS_WANT_CORRUPTED_GOTO(i == 0, done);
L
Linus Torvalds 已提交
932
			cur->bc_rec.b.br_state = XFS_EXT_NORM;
933
			if ((error = xfs_btree_insert(cur, &i)))
L
Linus Torvalds 已提交
934
				goto done;
935
			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
L
Linus Torvalds 已提交
936 937 938 939 940 941 942 943 944 945 946
		}
		if (ip->i_d.di_format == XFS_DINODE_FMT_EXTENTS &&
		    ip->i_d.di_nextents > ip->i_df.if_ext_max) {
			error = xfs_bmap_extents_to_btree(ip->i_transp, ip,
					first, flist, &cur, 1, &tmp_rval,
					XFS_DATA_FORK);
			rval |= tmp_rval;
			if (error)
				goto done;
		}
		temp = XFS_FILBLKS_MIN(xfs_bmap_worst_indlen(ip, temp),
947
			startblockval(PREV.br_startblock) -
L
Linus Torvalds 已提交
948
			(cur ? cur->bc_private.b.allocated : 0));
949
		ep = xfs_iext_get_ext(ifp, idx + 1);
950
		xfs_bmbt_set_startblock(ep, nullstartblock((int)temp));
C
Christoph Hellwig 已提交
951
		trace_xfs_bmap_post_update(ip, idx + 1, state, _THIS_IP_);
L
Linus Torvalds 已提交
952 953 954
		*dnew = temp;
		break;

955
	case BMAP_RIGHT_FILLING | BMAP_RIGHT_CONTIG:
L
Linus Torvalds 已提交
956 957 958 959 960
		/*
		 * Filling in the last part of a previous delayed allocation.
		 * The right neighbor is contiguous with the new allocation.
		 */
		temp = PREV.br_blockcount - new->br_blockcount;
C
Christoph Hellwig 已提交
961 962
		trace_xfs_bmap_pre_update(ip, idx, state, _THIS_IP_);
		trace_xfs_bmap_pre_update(ip, idx + 1, state, _THIS_IP_);
L
Linus Torvalds 已提交
963
		xfs_bmbt_set_blockcount(ep, temp);
964 965
		xfs_bmbt_set_allf(xfs_iext_get_ext(ifp, idx + 1),
			new->br_startoff, new->br_startblock,
L
Linus Torvalds 已提交
966 967
			new->br_blockcount + RIGHT.br_blockcount,
			RIGHT.br_state);
C
Christoph Hellwig 已提交
968
		trace_xfs_bmap_post_update(ip, idx + 1, state, _THIS_IP_);
L
Linus Torvalds 已提交
969 970 971 972 973 974 975 976 977
		ip->i_df.if_lastex = idx + 1;
		if (cur == NULL)
			rval = XFS_ILOG_DEXT;
		else {
			rval = 0;
			if ((error = xfs_bmbt_lookup_eq(cur, RIGHT.br_startoff,
					RIGHT.br_startblock,
					RIGHT.br_blockcount, &i)))
				goto done;
978
			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
L
Linus Torvalds 已提交
979 980 981 982 983 984 985 986
			if ((error = xfs_bmbt_update(cur, new->br_startoff,
					new->br_startblock,
					new->br_blockcount +
					RIGHT.br_blockcount,
					RIGHT.br_state)))
				goto done;
		}
		temp = XFS_FILBLKS_MIN(xfs_bmap_worst_indlen(ip, temp),
987 988
			startblockval(PREV.br_startblock));
		xfs_bmbt_set_startblock(ep, nullstartblock((int)temp));
C
Christoph Hellwig 已提交
989
		trace_xfs_bmap_post_update(ip, idx, state, _THIS_IP_);
L
Linus Torvalds 已提交
990 991 992
		*dnew = temp;
		break;

993
	case BMAP_RIGHT_FILLING:
L
Linus Torvalds 已提交
994 995 996 997 998
		/*
		 * Filling in the last part of a previous delayed allocation.
		 * The right neighbor is not contiguous.
		 */
		temp = PREV.br_blockcount - new->br_blockcount;
C
Christoph Hellwig 已提交
999
		trace_xfs_bmap_pre_update(ip, idx, state, _THIS_IP_);
L
Linus Torvalds 已提交
1000
		xfs_bmbt_set_blockcount(ep, temp);
1001
		xfs_iext_insert(ip, idx + 1, 1, new, state);
L
Linus Torvalds 已提交
1002 1003 1004 1005 1006 1007 1008 1009 1010 1011
		ip->i_df.if_lastex = idx + 1;
		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, new->br_startoff,
					new->br_startblock, new->br_blockcount,
					&i)))
				goto done;
1012
			XFS_WANT_CORRUPTED_GOTO(i == 0, done);
L
Linus Torvalds 已提交
1013
			cur->bc_rec.b.br_state = XFS_EXT_NORM;
1014
			if ((error = xfs_btree_insert(cur, &i)))
L
Linus Torvalds 已提交
1015
				goto done;
1016
			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
L
Linus Torvalds 已提交
1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027
		}
		if (ip->i_d.di_format == XFS_DINODE_FMT_EXTENTS &&
		    ip->i_d.di_nextents > ip->i_df.if_ext_max) {
			error = xfs_bmap_extents_to_btree(ip->i_transp, ip,
				first, flist, &cur, 1, &tmp_rval,
				XFS_DATA_FORK);
			rval |= tmp_rval;
			if (error)
				goto done;
		}
		temp = XFS_FILBLKS_MIN(xfs_bmap_worst_indlen(ip, temp),
1028
			startblockval(PREV.br_startblock) -
L
Linus Torvalds 已提交
1029
			(cur ? cur->bc_private.b.allocated : 0));
1030
		ep = xfs_iext_get_ext(ifp, idx);
1031
		xfs_bmbt_set_startblock(ep, nullstartblock((int)temp));
C
Christoph Hellwig 已提交
1032
		trace_xfs_bmap_post_update(ip, idx, state, _THIS_IP_);
L
Linus Torvalds 已提交
1033 1034 1035 1036 1037 1038 1039 1040 1041 1042
		*dnew = temp;
		break;

	case 0:
		/*
		 * Filling in the middle part of a previous delayed allocation.
		 * Contiguity is impossible here.
		 * This case is avoided almost all the time.
		 */
		temp = new->br_startoff - PREV.br_startoff;
C
Christoph Hellwig 已提交
1043
		trace_xfs_bmap_pre_update(ip, idx, 0, _THIS_IP_);
L
Linus Torvalds 已提交
1044 1045
		xfs_bmbt_set_blockcount(ep, temp);
		r[0] = *new;
1046 1047
		r[1].br_state = PREV.br_state;
		r[1].br_startblock = 0;
L
Linus Torvalds 已提交
1048 1049 1050
		r[1].br_startoff = new_endoff;
		temp2 = PREV.br_startoff + PREV.br_blockcount - new_endoff;
		r[1].br_blockcount = temp2;
1051
		xfs_iext_insert(ip, idx + 1, 2, &r[0], state);
L
Linus Torvalds 已提交
1052 1053 1054 1055 1056 1057 1058 1059 1060 1061
		ip->i_df.if_lastex = idx + 1;
		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, new->br_startoff,
					new->br_startblock, new->br_blockcount,
					&i)))
				goto done;
1062
			XFS_WANT_CORRUPTED_GOTO(i == 0, done);
L
Linus Torvalds 已提交
1063
			cur->bc_rec.b.br_state = XFS_EXT_NORM;
1064
			if ((error = xfs_btree_insert(cur, &i)))
L
Linus Torvalds 已提交
1065
				goto done;
1066
			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
L
Linus Torvalds 已提交
1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078
		}
		if (ip->i_d.di_format == XFS_DINODE_FMT_EXTENTS &&
		    ip->i_d.di_nextents > ip->i_df.if_ext_max) {
			error = xfs_bmap_extents_to_btree(ip->i_transp, ip,
					first, flist, &cur, 1, &tmp_rval,
					XFS_DATA_FORK);
			rval |= tmp_rval;
			if (error)
				goto done;
		}
		temp = xfs_bmap_worst_indlen(ip, temp);
		temp2 = xfs_bmap_worst_indlen(ip, temp2);
1079
		diff = (int)(temp + temp2 - startblockval(PREV.br_startblock) -
L
Linus Torvalds 已提交
1080 1081
			(cur ? cur->bc_private.b.allocated : 0));
		if (diff > 0 &&
1082
		    xfs_mod_incore_sb(ip->i_mount, XFS_SBS_FDBLOCKS, -((int64_t)diff), rsvd)) {
L
Linus Torvalds 已提交
1083 1084 1085 1086 1087 1088 1089 1090 1091 1092
			/*
			 * Ick gross gag me with a spoon.
			 */
			ASSERT(0);	/* want to see if this ever happens! */
			while (diff > 0) {
				if (temp) {
					temp--;
					diff--;
					if (!diff ||
					    !xfs_mod_incore_sb(ip->i_mount,
1093
						    XFS_SBS_FDBLOCKS, -((int64_t)diff), rsvd))
L
Linus Torvalds 已提交
1094 1095 1096 1097 1098 1099 1100
						break;
				}
				if (temp2) {
					temp2--;
					diff--;
					if (!diff ||
					    !xfs_mod_incore_sb(ip->i_mount,
1101
						    XFS_SBS_FDBLOCKS, -((int64_t)diff), rsvd))
L
Linus Torvalds 已提交
1102 1103 1104 1105
						break;
				}
			}
		}
1106
		ep = xfs_iext_get_ext(ifp, idx);
1107
		xfs_bmbt_set_startblock(ep, nullstartblock((int)temp));
C
Christoph Hellwig 已提交
1108 1109
		trace_xfs_bmap_post_update(ip, idx, state, _THIS_IP_);
		trace_xfs_bmap_pre_update(ip, idx + 2, state, _THIS_IP_);
1110
		xfs_bmbt_set_startblock(xfs_iext_get_ext(ifp, idx + 2),
1111
			nullstartblock((int)temp2));
C
Christoph Hellwig 已提交
1112
		trace_xfs_bmap_post_update(ip, idx + 2, state, _THIS_IP_);
L
Linus Torvalds 已提交
1113 1114 1115
		*dnew = temp + temp2;
		break;

1116 1117 1118 1119 1120 1121 1122
	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:
L
Linus Torvalds 已提交
1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145
		/*
		 * These cases are all impossible.
		 */
		ASSERT(0);
	}
	*curp = cur;
done:
	*logflagsp = rval;
	return error;
#undef	LEFT
#undef	RIGHT
#undef	PREV
}

/*
 * Called by xfs_bmap_add_extent to handle cases converting an unwritten
 * allocation to a real allocation or vice versa.
 */
STATIC int				/* error */
xfs_bmap_add_extent_unwritten_real(
	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 */
1146
	xfs_bmbt_irec_t		*new,	/* new data to add to file extents */
1147
	int			*logflagsp) /* inode logging flags */
L
Linus Torvalds 已提交
1148 1149
{
	xfs_btree_cur_t		*cur;	/* btree cursor */
1150
	xfs_bmbt_rec_host_t	*ep;	/* extent entry for idx */
L
Linus Torvalds 已提交
1151 1152
	int			error;	/* error return value */
	int			i;	/* temp state */
1153
	xfs_ifork_t		*ifp;	/* inode fork pointer */
L
Linus Torvalds 已提交
1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169
	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 */

#define	LEFT		r[0]
#define	RIGHT		r[1]
#define	PREV		r[2]
	/*
	 * Set up a bunch of variables to make the tests simpler.
	 */
	error = 0;
	cur = *curp;
1170 1171
	ifp = XFS_IFORK_PTR(ip, XFS_DATA_FORK);
	ep = xfs_iext_get_ext(ifp, idx);
L
Linus Torvalds 已提交
1172 1173 1174 1175 1176 1177 1178 1179
	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);
1180

L
Linus Torvalds 已提交
1181 1182 1183 1184
	/*
	 * Set flags determining what part of the previous oldext allocation
	 * extent is being replaced by a newext allocation.
	 */
1185 1186 1187 1188 1189
	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
1190 1191 1192 1193
	/*
	 * Check and set flags if this segment has a left neighbor.
	 * Don't set contiguous if the combined extent would be too large.
	 */
1194 1195
	if (idx > 0) {
		state |= BMAP_LEFT_VALID;
1196
		xfs_bmbt_get_all(xfs_iext_get_ext(ifp, idx - 1), &LEFT);
1197 1198 1199

		if (isnullstartblock(LEFT.br_startblock))
			state |= BMAP_LEFT_DELAY;
L
Linus Torvalds 已提交
1200
	}
1201 1202 1203 1204 1205 1206 1207 1208

	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
1209 1210 1211 1212 1213
	/*
	 * 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.
	 */
1214 1215
	if (idx < ip->i_df.if_bytes / (uint)sizeof(xfs_bmbt_rec_t) - 1) {
		state |= BMAP_RIGHT_VALID;
1216
		xfs_bmbt_get_all(xfs_iext_get_ext(ifp, idx + 1), &RIGHT);
1217 1218
		if (isnullstartblock(RIGHT.br_startblock))
			state |= BMAP_RIGHT_DELAY;
L
Linus Torvalds 已提交
1219
	}
1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233

	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;

L
Linus Torvalds 已提交
1234 1235 1236
	/*
	 * Switch out based on the FILLING and CONTIG state bits.
	 */
1237 1238 1239 1240
	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
1241 1242 1243 1244
		/*
		 * Setting all of a previous oldext extent to newext.
		 * The left and right neighbors are both contiguous with new.
		 */
C
Christoph Hellwig 已提交
1245
		trace_xfs_bmap_pre_update(ip, idx - 1, state, _THIS_IP_);
1246
		xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, idx - 1),
L
Linus Torvalds 已提交
1247 1248
			LEFT.br_blockcount + PREV.br_blockcount +
			RIGHT.br_blockcount);
C
Christoph Hellwig 已提交
1249 1250
		trace_xfs_bmap_post_update(ip, idx - 1, state, _THIS_IP_);

1251
		xfs_iext_remove(ip, idx, 2, state);
L
Linus Torvalds 已提交
1252 1253 1254 1255 1256 1257 1258 1259 1260 1261
		ip->i_df.if_lastex = idx - 1;
		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;
1262
			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
1263
			if ((error = xfs_btree_delete(cur, &i)))
L
Linus Torvalds 已提交
1264
				goto done;
1265
			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
1266
			if ((error = xfs_btree_decrement(cur, 0, &i)))
L
Linus Torvalds 已提交
1267
				goto done;
1268
			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
1269
			if ((error = xfs_btree_delete(cur, &i)))
L
Linus Torvalds 已提交
1270
				goto done;
1271
			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
1272
			if ((error = xfs_btree_decrement(cur, 0, &i)))
L
Linus Torvalds 已提交
1273
				goto done;
1274
			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
L
Linus Torvalds 已提交
1275 1276 1277 1278 1279 1280 1281 1282
			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;

1283
	case BMAP_LEFT_FILLING | BMAP_RIGHT_FILLING | BMAP_LEFT_CONTIG:
L
Linus Torvalds 已提交
1284 1285 1286 1287
		/*
		 * Setting all of a previous oldext extent to newext.
		 * The left neighbor is contiguous, the right is not.
		 */
C
Christoph Hellwig 已提交
1288
		trace_xfs_bmap_pre_update(ip, idx - 1, state, _THIS_IP_);
1289
		xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, idx - 1),
L
Linus Torvalds 已提交
1290
			LEFT.br_blockcount + PREV.br_blockcount);
C
Christoph Hellwig 已提交
1291 1292
		trace_xfs_bmap_post_update(ip, idx - 1, state, _THIS_IP_);

L
Linus Torvalds 已提交
1293
		ip->i_df.if_lastex = idx - 1;
1294
		xfs_iext_remove(ip, idx, 1, state);
L
Linus Torvalds 已提交
1295 1296 1297 1298 1299 1300 1301 1302 1303
		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;
1304
			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
1305
			if ((error = xfs_btree_delete(cur, &i)))
L
Linus Torvalds 已提交
1306
				goto done;
1307
			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
1308
			if ((error = xfs_btree_decrement(cur, 0, &i)))
L
Linus Torvalds 已提交
1309
				goto done;
1310
			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
L
Linus Torvalds 已提交
1311 1312 1313 1314 1315 1316 1317 1318
			if ((error = xfs_bmbt_update(cur, LEFT.br_startoff,
				LEFT.br_startblock,
				LEFT.br_blockcount + PREV.br_blockcount,
				LEFT.br_state)))
				goto done;
		}
		break;

1319
	case BMAP_LEFT_FILLING | BMAP_RIGHT_FILLING | BMAP_RIGHT_CONTIG:
L
Linus Torvalds 已提交
1320 1321 1322 1323
		/*
		 * Setting all of a previous oldext extent to newext.
		 * The right neighbor is contiguous, the left is not.
		 */
C
Christoph Hellwig 已提交
1324
		trace_xfs_bmap_pre_update(ip, idx, state, _THIS_IP_);
L
Linus Torvalds 已提交
1325 1326 1327
		xfs_bmbt_set_blockcount(ep,
			PREV.br_blockcount + RIGHT.br_blockcount);
		xfs_bmbt_set_state(ep, newext);
C
Christoph Hellwig 已提交
1328
		trace_xfs_bmap_post_update(ip, idx, state, _THIS_IP_);
L
Linus Torvalds 已提交
1329
		ip->i_df.if_lastex = idx;
1330
		xfs_iext_remove(ip, idx + 1, 1, state);
L
Linus Torvalds 已提交
1331 1332 1333 1334 1335 1336 1337 1338 1339
		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;
1340
			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
1341
			if ((error = xfs_btree_delete(cur, &i)))
L
Linus Torvalds 已提交
1342
				goto done;
1343
			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
1344
			if ((error = xfs_btree_decrement(cur, 0, &i)))
L
Linus Torvalds 已提交
1345
				goto done;
1346
			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
L
Linus Torvalds 已提交
1347 1348 1349 1350 1351 1352 1353 1354
			if ((error = xfs_bmbt_update(cur, new->br_startoff,
				new->br_startblock,
				new->br_blockcount + RIGHT.br_blockcount,
				newext)))
				goto done;
		}
		break;

1355
	case BMAP_LEFT_FILLING | BMAP_RIGHT_FILLING:
L
Linus Torvalds 已提交
1356 1357 1358 1359 1360
		/*
		 * Setting all of a previous oldext extent to newext.
		 * Neither the left nor right neighbors are contiguous with
		 * the new one.
		 */
C
Christoph Hellwig 已提交
1361
		trace_xfs_bmap_pre_update(ip, idx, state, _THIS_IP_);
L
Linus Torvalds 已提交
1362
		xfs_bmbt_set_state(ep, newext);
C
Christoph Hellwig 已提交
1363 1364
		trace_xfs_bmap_post_update(ip, idx, state, _THIS_IP_);

L
Linus Torvalds 已提交
1365 1366 1367 1368 1369 1370 1371 1372 1373
		ip->i_df.if_lastex = idx;
		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;
1374
			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
L
Linus Torvalds 已提交
1375 1376 1377 1378 1379 1380 1381
			if ((error = xfs_bmbt_update(cur, new->br_startoff,
				new->br_startblock, new->br_blockcount,
				newext)))
				goto done;
		}
		break;

1382
	case BMAP_LEFT_FILLING | BMAP_LEFT_CONTIG:
L
Linus Torvalds 已提交
1383 1384 1385 1386
		/*
		 * Setting the first part of a previous oldext extent to newext.
		 * The left neighbor is contiguous.
		 */
C
Christoph Hellwig 已提交
1387
		trace_xfs_bmap_pre_update(ip, idx - 1, state, _THIS_IP_);
1388
		xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, idx - 1),
L
Linus Torvalds 已提交
1389 1390 1391
			LEFT.br_blockcount + new->br_blockcount);
		xfs_bmbt_set_startoff(ep,
			PREV.br_startoff + new->br_blockcount);
C
Christoph Hellwig 已提交
1392 1393 1394
		trace_xfs_bmap_post_update(ip, idx - 1, state, _THIS_IP_);

		trace_xfs_bmap_pre_update(ip, idx, state, _THIS_IP_);
L
Linus Torvalds 已提交
1395 1396 1397 1398
		xfs_bmbt_set_startblock(ep,
			new->br_startblock + new->br_blockcount);
		xfs_bmbt_set_blockcount(ep,
			PREV.br_blockcount - new->br_blockcount);
C
Christoph Hellwig 已提交
1399 1400
		trace_xfs_bmap_post_update(ip, idx, state, _THIS_IP_);

L
Linus Torvalds 已提交
1401 1402 1403 1404 1405 1406 1407 1408 1409
		ip->i_df.if_lastex = idx - 1;
		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;
1410
			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
L
Linus Torvalds 已提交
1411 1412 1413 1414 1415 1416
			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;
1417
			if ((error = xfs_btree_decrement(cur, 0, &i)))
L
Linus Torvalds 已提交
1418 1419 1420 1421 1422 1423 1424 1425 1426
				goto done;
			if (xfs_bmbt_update(cur, LEFT.br_startoff,
				LEFT.br_startblock,
				LEFT.br_blockcount + new->br_blockcount,
				LEFT.br_state))
				goto done;
		}
		break;

1427
	case BMAP_LEFT_FILLING:
L
Linus Torvalds 已提交
1428 1429 1430 1431
		/*
		 * Setting the first part of a previous oldext extent to newext.
		 * The left neighbor is not contiguous.
		 */
C
Christoph Hellwig 已提交
1432
		trace_xfs_bmap_pre_update(ip, idx, state, _THIS_IP_);
L
Linus Torvalds 已提交
1433 1434 1435 1436 1437 1438
		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);
C
Christoph Hellwig 已提交
1439 1440
		trace_xfs_bmap_post_update(ip, idx, state, _THIS_IP_);

1441
		xfs_iext_insert(ip, idx, 1, new, state);
L
Linus Torvalds 已提交
1442 1443 1444 1445 1446 1447 1448 1449 1450 1451
		ip->i_df.if_lastex = idx;
		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;
1452
			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
L
Linus Torvalds 已提交
1453 1454 1455 1456 1457 1458 1459
			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;
1460
			if ((error = xfs_btree_insert(cur, &i)))
L
Linus Torvalds 已提交
1461
				goto done;
1462
			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
L
Linus Torvalds 已提交
1463 1464 1465
		}
		break;

1466
	case BMAP_RIGHT_FILLING | BMAP_RIGHT_CONTIG:
L
Linus Torvalds 已提交
1467 1468 1469 1470
		/*
		 * Setting the last part of a previous oldext extent to newext.
		 * The right neighbor is contiguous with the new allocation.
		 */
C
Christoph Hellwig 已提交
1471 1472
		trace_xfs_bmap_pre_update(ip, idx, state, _THIS_IP_);
		trace_xfs_bmap_pre_update(ip, idx + 1, state, _THIS_IP_);
L
Linus Torvalds 已提交
1473 1474
		xfs_bmbt_set_blockcount(ep,
			PREV.br_blockcount - new->br_blockcount);
C
Christoph Hellwig 已提交
1475
		trace_xfs_bmap_post_update(ip, idx, state, _THIS_IP_);
1476 1477
		xfs_bmbt_set_allf(xfs_iext_get_ext(ifp, idx + 1),
			new->br_startoff, new->br_startblock,
L
Linus Torvalds 已提交
1478
			new->br_blockcount + RIGHT.br_blockcount, newext);
C
Christoph Hellwig 已提交
1479 1480
		trace_xfs_bmap_post_update(ip, idx + 1, state, _THIS_IP_);

L
Linus Torvalds 已提交
1481 1482 1483 1484 1485 1486 1487 1488 1489
		ip->i_df.if_lastex = idx + 1;
		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;
1490
			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
L
Linus Torvalds 已提交
1491 1492 1493 1494 1495
			if ((error = xfs_bmbt_update(cur, PREV.br_startoff,
				PREV.br_startblock,
				PREV.br_blockcount - new->br_blockcount,
				oldext)))
				goto done;
1496
			if ((error = xfs_btree_increment(cur, 0, &i)))
L
Linus Torvalds 已提交
1497 1498 1499 1500 1501 1502 1503 1504 1505
				goto done;
			if ((error = xfs_bmbt_update(cur, new->br_startoff,
				new->br_startblock,
				new->br_blockcount + RIGHT.br_blockcount,
				newext)))
				goto done;
		}
		break;

1506
	case BMAP_RIGHT_FILLING:
L
Linus Torvalds 已提交
1507 1508 1509 1510
		/*
		 * Setting the last part of a previous oldext extent to newext.
		 * The right neighbor is not contiguous.
		 */
C
Christoph Hellwig 已提交
1511
		trace_xfs_bmap_pre_update(ip, idx, state, _THIS_IP_);
L
Linus Torvalds 已提交
1512 1513
		xfs_bmbt_set_blockcount(ep,
			PREV.br_blockcount - new->br_blockcount);
C
Christoph Hellwig 已提交
1514 1515
		trace_xfs_bmap_post_update(ip, idx, state, _THIS_IP_);

1516
		xfs_iext_insert(ip, idx + 1, 1, new, state);
L
Linus Torvalds 已提交
1517 1518 1519 1520 1521 1522 1523 1524 1525 1526
		ip->i_df.if_lastex = idx + 1;
		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;
1527
			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
L
Linus Torvalds 已提交
1528 1529 1530 1531 1532 1533 1534 1535 1536
			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;
1537
			XFS_WANT_CORRUPTED_GOTO(i == 0, done);
L
Linus Torvalds 已提交
1538
			cur->bc_rec.b.br_state = XFS_EXT_NORM;
1539
			if ((error = xfs_btree_insert(cur, &i)))
L
Linus Torvalds 已提交
1540
				goto done;
1541
			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
L
Linus Torvalds 已提交
1542 1543 1544 1545 1546 1547 1548 1549 1550
		}
		break;

	case 0:
		/*
		 * Setting the middle part of a previous oldext extent to
		 * newext.  Contiguity is impossible here.
		 * One extent becomes three extents.
		 */
C
Christoph Hellwig 已提交
1551
		trace_xfs_bmap_pre_update(ip, idx, state, _THIS_IP_);
L
Linus Torvalds 已提交
1552 1553
		xfs_bmbt_set_blockcount(ep,
			new->br_startoff - PREV.br_startoff);
C
Christoph Hellwig 已提交
1554 1555
		trace_xfs_bmap_post_update(ip, idx, state, _THIS_IP_);

L
Linus Torvalds 已提交
1556 1557 1558 1559 1560 1561
		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;
1562
		xfs_iext_insert(ip, idx + 1, 2, &r[0], state);
L
Linus Torvalds 已提交
1563 1564 1565 1566 1567 1568 1569 1570 1571 1572
		ip->i_df.if_lastex = idx + 1;
		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;
1573
			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
L
Linus Torvalds 已提交
1574 1575 1576 1577 1578 1579 1580
			/* 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;
1581 1582
			cur->bc_rec.b.br_blockcount =
				new->br_startoff - PREV.br_startoff;
1583
			if ((error = xfs_btree_insert(cur, &i)))
L
Linus Torvalds 已提交
1584
				goto done;
1585
			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
1586 1587 1588 1589 1590 1591 1592 1593
			/*
			 * 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)))
L
Linus Torvalds 已提交
1594
				goto done;
1595
			XFS_WANT_CORRUPTED_GOTO(i == 0, done);
L
Linus Torvalds 已提交
1596
			/* new middle extent - newext */
1597
			cur->bc_rec.b.br_state = new->br_state;
1598
			if ((error = xfs_btree_insert(cur, &i)))
L
Linus Torvalds 已提交
1599
				goto done;
1600
			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
L
Linus Torvalds 已提交
1601 1602 1603
		}
		break;

1604 1605 1606 1607 1608 1609 1610
	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:
L
Linus Torvalds 已提交
1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633
		/*
		 * These cases are all impossible.
		 */
		ASSERT(0);
	}
	*curp = cur;
done:
	*logflagsp = rval;
	return error;
#undef	LEFT
#undef	RIGHT
#undef	PREV
}

/*
 * Called by xfs_bmap_add_extent to handle cases converting a hole
 * to a delayed allocation.
 */
/*ARGSUSED*/
STATIC int				/* error */
xfs_bmap_add_extent_hole_delay(
	xfs_inode_t		*ip,	/* incore inode pointer */
	xfs_extnum_t		idx,	/* extent number to update/insert */
1634
	xfs_bmbt_irec_t		*new,	/* new data to add to file extents */
L
Linus Torvalds 已提交
1635 1636 1637
	int			*logflagsp, /* inode logging flags */
	int			rsvd)		/* OK to allocate reserved blocks */
{
1638
	xfs_bmbt_rec_host_t	*ep;	/* extent record for idx */
1639
	xfs_ifork_t		*ifp;	/* inode fork pointer */
L
Linus Torvalds 已提交
1640 1641 1642 1643 1644
	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 */
1645
	xfs_filblks_t		temp=0;	/* temp for indirect calculations */
L
Linus Torvalds 已提交
1646

1647 1648
	ifp = XFS_IFORK_PTR(ip, XFS_DATA_FORK);
	ep = xfs_iext_get_ext(ifp, idx);
L
Linus Torvalds 已提交
1649
	state = 0;
1650
	ASSERT(isnullstartblock(new->br_startblock));
1651

L
Linus Torvalds 已提交
1652 1653 1654
	/*
	 * Check and set flags if this segment has a left neighbor
	 */
1655 1656
	if (idx > 0) {
		state |= BMAP_LEFT_VALID;
1657
		xfs_bmbt_get_all(xfs_iext_get_ext(ifp, idx - 1), &left);
1658 1659 1660

		if (isnullstartblock(left.br_startblock))
			state |= BMAP_LEFT_DELAY;
L
Linus Torvalds 已提交
1661
	}
1662

L
Linus Torvalds 已提交
1663 1664 1665 1666
	/*
	 * Check and set flags if the current (right) segment exists.
	 * If it doesn't exist, we're converting the hole at end-of-file.
	 */
1667 1668
	if (idx < ip->i_df.if_bytes / (uint)sizeof(xfs_bmbt_rec_t)) {
		state |= BMAP_RIGHT_VALID;
L
Linus Torvalds 已提交
1669
		xfs_bmbt_get_all(ep, &right);
1670 1671 1672

		if (isnullstartblock(right.br_startblock))
			state |= BMAP_RIGHT_DELAY;
L
Linus Torvalds 已提交
1673
	}
1674

L
Linus Torvalds 已提交
1675 1676 1677 1678
	/*
	 * Set contiguity flags on the left and right neighbors.
	 * Don't let extents get too large, even if the pieces are contiguous.
	 */
1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691
	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;

L
Linus Torvalds 已提交
1692 1693 1694
	/*
	 * Switch out based on the contiguity flags.
	 */
1695 1696
	switch (state & (BMAP_LEFT_CONTIG | BMAP_RIGHT_CONTIG)) {
	case BMAP_LEFT_CONTIG | BMAP_RIGHT_CONTIG:
L
Linus Torvalds 已提交
1697 1698 1699
		/*
		 * New allocation is contiguous with delayed allocations
		 * on the left and on the right.
1700
		 * Merge all three into a single extent record.
L
Linus Torvalds 已提交
1701 1702 1703
		 */
		temp = left.br_blockcount + new->br_blockcount +
			right.br_blockcount;
C
Christoph Hellwig 已提交
1704 1705

		trace_xfs_bmap_pre_update(ip, idx - 1, state, _THIS_IP_);
1706
		xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, idx - 1), temp);
1707 1708 1709
		oldlen = startblockval(left.br_startblock) +
			startblockval(new->br_startblock) +
			startblockval(right.br_startblock);
L
Linus Torvalds 已提交
1710
		newlen = xfs_bmap_worst_indlen(ip, temp);
1711
		xfs_bmbt_set_startblock(xfs_iext_get_ext(ifp, idx - 1),
1712
			nullstartblock((int)newlen));
C
Christoph Hellwig 已提交
1713 1714
		trace_xfs_bmap_post_update(ip, idx - 1, state, _THIS_IP_);

1715
		xfs_iext_remove(ip, idx, 1, state);
L
Linus Torvalds 已提交
1716 1717 1718
		ip->i_df.if_lastex = idx - 1;
		break;

1719
	case BMAP_LEFT_CONTIG:
L
Linus Torvalds 已提交
1720 1721 1722 1723 1724 1725
		/*
		 * New allocation is contiguous with a delayed allocation
		 * on the left.
		 * Merge the new allocation with the left neighbor.
		 */
		temp = left.br_blockcount + new->br_blockcount;
C
Christoph Hellwig 已提交
1726
		trace_xfs_bmap_pre_update(ip, idx - 1, state, _THIS_IP_);
1727
		xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, idx - 1), temp);
1728 1729
		oldlen = startblockval(left.br_startblock) +
			startblockval(new->br_startblock);
L
Linus Torvalds 已提交
1730
		newlen = xfs_bmap_worst_indlen(ip, temp);
1731
		xfs_bmbt_set_startblock(xfs_iext_get_ext(ifp, idx - 1),
1732
			nullstartblock((int)newlen));
C
Christoph Hellwig 已提交
1733 1734
		trace_xfs_bmap_post_update(ip, idx - 1, state, _THIS_IP_);

L
Linus Torvalds 已提交
1735 1736 1737
		ip->i_df.if_lastex = idx - 1;
		break;

1738
	case BMAP_RIGHT_CONTIG:
L
Linus Torvalds 已提交
1739 1740 1741 1742 1743
		/*
		 * New allocation is contiguous with a delayed allocation
		 * on the right.
		 * Merge the new allocation with the right neighbor.
		 */
C
Christoph Hellwig 已提交
1744
		trace_xfs_bmap_pre_update(ip, idx, state, _THIS_IP_);
L
Linus Torvalds 已提交
1745
		temp = new->br_blockcount + right.br_blockcount;
1746 1747
		oldlen = startblockval(new->br_startblock) +
			startblockval(right.br_startblock);
L
Linus Torvalds 已提交
1748 1749
		newlen = xfs_bmap_worst_indlen(ip, temp);
		xfs_bmbt_set_allf(ep, new->br_startoff,
1750
			nullstartblock((int)newlen), temp, right.br_state);
C
Christoph Hellwig 已提交
1751 1752
		trace_xfs_bmap_post_update(ip, idx, state, _THIS_IP_);

L
Linus Torvalds 已提交
1753 1754 1755 1756 1757 1758 1759 1760 1761 1762
		ip->i_df.if_lastex = idx;
		break;

	case 0:
		/*
		 * New allocation is not contiguous with another
		 * delayed allocation.
		 * Insert a new entry.
		 */
		oldlen = newlen = 0;
1763
		xfs_iext_insert(ip, idx, 1, new, state);
L
Linus Torvalds 已提交
1764 1765 1766 1767 1768 1769
		ip->i_df.if_lastex = idx;
		break;
	}
	if (oldlen != newlen) {
		ASSERT(oldlen > newlen);
		xfs_mod_incore_sb(ip->i_mount, XFS_SBS_FDBLOCKS,
1770
			(int64_t)(oldlen - newlen), rsvd);
L
Linus Torvalds 已提交
1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787
		/*
		 * Nothing to do for disk quota accounting here.
		 */
	}
	*logflagsp = 0;
	return 0;
}

/*
 * Called by xfs_bmap_add_extent to handle cases converting a hole
 * to a real allocation.
 */
STATIC int				/* error */
xfs_bmap_add_extent_hole_real(
	xfs_inode_t		*ip,	/* incore inode pointer */
	xfs_extnum_t		idx,	/* extent number to update/insert */
	xfs_btree_cur_t		*cur,	/* if null, not a btree */
1788
	xfs_bmbt_irec_t		*new,	/* new data to add to file extents */
L
Linus Torvalds 已提交
1789 1790 1791
	int			*logflagsp, /* inode logging flags */
	int			whichfork) /* data or attr fork */
{
1792
	xfs_bmbt_rec_host_t	*ep;	/* pointer to extent entry ins. point */
L
Linus Torvalds 已提交
1793 1794 1795 1796 1797
	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 */
1798
	int			rval=0;	/* return value (logging flags) */
L
Linus Torvalds 已提交
1799 1800 1801 1802
	int			state;	/* state bits, accessed thru macros */

	ifp = XFS_IFORK_PTR(ip, whichfork);
	ASSERT(idx <= ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t));
1803
	ep = xfs_iext_get_ext(ifp, idx);
L
Linus Torvalds 已提交
1804
	state = 0;
1805

1806 1807 1808
	if (whichfork == XFS_ATTR_FORK)
		state |= BMAP_ATTRFORK;

L
Linus Torvalds 已提交
1809 1810 1811
	/*
	 * Check and set flags if this segment has a left neighbor.
	 */
1812 1813
	if (idx > 0) {
		state |= BMAP_LEFT_VALID;
1814
		xfs_bmbt_get_all(xfs_iext_get_ext(ifp, idx - 1), &left);
1815 1816
		if (isnullstartblock(left.br_startblock))
			state |= BMAP_LEFT_DELAY;
L
Linus Torvalds 已提交
1817
	}
1818

L
Linus Torvalds 已提交
1819 1820 1821 1822
	/*
	 * Check and set flags if this segment has a current value.
	 * Not true if we're inserting into the "hole" at eof.
	 */
1823 1824
	if (idx < ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t)) {
		state |= BMAP_RIGHT_VALID;
L
Linus Torvalds 已提交
1825
		xfs_bmbt_get_all(ep, &right);
1826 1827
		if (isnullstartblock(right.br_startblock))
			state |= BMAP_RIGHT_DELAY;
L
Linus Torvalds 已提交
1828
	}
1829

L
Linus Torvalds 已提交
1830 1831 1832 1833
	/*
	 * We're inserting a real allocation between "left" and "right".
	 * Set the contiguity flags.  Don't let extents get too large.
	 */
1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849
	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;

	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;
L
Linus Torvalds 已提交
1850

1851
	error = 0;
L
Linus Torvalds 已提交
1852 1853 1854
	/*
	 * Select which case we're in here, and implement it.
	 */
1855 1856
	switch (state & (BMAP_LEFT_CONTIG | BMAP_RIGHT_CONTIG)) {
	case BMAP_LEFT_CONTIG | BMAP_RIGHT_CONTIG:
L
Linus Torvalds 已提交
1857 1858 1859
		/*
		 * New allocation is contiguous with real allocations on the
		 * left and on the right.
1860
		 * Merge all three into a single extent record.
L
Linus Torvalds 已提交
1861
		 */
C
Christoph Hellwig 已提交
1862
		trace_xfs_bmap_pre_update(ip, idx - 1, state, _THIS_IP_);
1863
		xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, idx - 1),
L
Linus Torvalds 已提交
1864 1865
			left.br_blockcount + new->br_blockcount +
			right.br_blockcount);
C
Christoph Hellwig 已提交
1866 1867
		trace_xfs_bmap_post_update(ip, idx - 1, state, _THIS_IP_);

1868
		xfs_iext_remove(ip, idx, 1, state);
L
Linus Torvalds 已提交
1869 1870 1871 1872
		ifp->if_lastex = idx - 1;
		XFS_IFORK_NEXT_SET(ip, whichfork,
			XFS_IFORK_NEXTENTS(ip, whichfork) - 1);
		if (cur == NULL) {
1873
			rval = XFS_ILOG_CORE | xfs_ilog_fext(whichfork);
1874 1875 1876 1877 1878 1879 1880
		} else {
			rval = XFS_ILOG_CORE;
			if ((error = xfs_bmbt_lookup_eq(cur,
					right.br_startoff,
					right.br_startblock,
					right.br_blockcount, &i)))
				goto done;
1881
			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
1882
			if ((error = xfs_btree_delete(cur, &i)))
1883
				goto done;
1884
			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
1885
			if ((error = xfs_btree_decrement(cur, 0, &i)))
1886
				goto done;
1887
			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
1888 1889 1890 1891 1892 1893 1894
			if ((error = xfs_bmbt_update(cur, left.br_startoff,
					left.br_startblock,
					left.br_blockcount +
						new->br_blockcount +
						right.br_blockcount,
					left.br_state)))
				goto done;
L
Linus Torvalds 已提交
1895
		}
1896
		break;
L
Linus Torvalds 已提交
1897

1898
	case BMAP_LEFT_CONTIG:
L
Linus Torvalds 已提交
1899 1900 1901 1902 1903
		/*
		 * New allocation is contiguous with a real allocation
		 * on the left.
		 * Merge the new allocation with the left neighbor.
		 */
C
Christoph Hellwig 已提交
1904
		trace_xfs_bmap_pre_update(ip, idx - 1, state, _THIS_IP_);
1905
		xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, idx - 1),
L
Linus Torvalds 已提交
1906
			left.br_blockcount + new->br_blockcount);
C
Christoph Hellwig 已提交
1907 1908
		trace_xfs_bmap_post_update(ip, idx - 1, state, _THIS_IP_);

L
Linus Torvalds 已提交
1909 1910
		ifp->if_lastex = idx - 1;
		if (cur == NULL) {
1911
			rval = xfs_ilog_fext(whichfork);
1912 1913 1914 1915 1916 1917 1918
		} else {
			rval = 0;
			if ((error = xfs_bmbt_lookup_eq(cur,
					left.br_startoff,
					left.br_startblock,
					left.br_blockcount, &i)))
				goto done;
1919
			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
1920 1921 1922 1923 1924 1925
			if ((error = xfs_bmbt_update(cur, left.br_startoff,
					left.br_startblock,
					left.br_blockcount +
						new->br_blockcount,
					left.br_state)))
				goto done;
L
Linus Torvalds 已提交
1926
		}
1927
		break;
L
Linus Torvalds 已提交
1928

1929
	case BMAP_RIGHT_CONTIG:
L
Linus Torvalds 已提交
1930 1931 1932 1933 1934
		/*
		 * New allocation is contiguous with a real allocation
		 * on the right.
		 * Merge the new allocation with the right neighbor.
		 */
C
Christoph Hellwig 已提交
1935
		trace_xfs_bmap_pre_update(ip, idx, state, _THIS_IP_);
L
Linus Torvalds 已提交
1936 1937 1938
		xfs_bmbt_set_allf(ep, new->br_startoff, new->br_startblock,
			new->br_blockcount + right.br_blockcount,
			right.br_state);
C
Christoph Hellwig 已提交
1939 1940
		trace_xfs_bmap_post_update(ip, idx, state, _THIS_IP_);

L
Linus Torvalds 已提交
1941 1942
		ifp->if_lastex = idx;
		if (cur == NULL) {
1943
			rval = xfs_ilog_fext(whichfork);
1944 1945 1946 1947 1948 1949 1950
		} else {
			rval = 0;
			if ((error = xfs_bmbt_lookup_eq(cur,
					right.br_startoff,
					right.br_startblock,
					right.br_blockcount, &i)))
				goto done;
1951
			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
1952 1953 1954 1955 1956 1957
			if ((error = xfs_bmbt_update(cur, new->br_startoff,
					new->br_startblock,
					new->br_blockcount +
						right.br_blockcount,
					right.br_state)))
				goto done;
L
Linus Torvalds 已提交
1958
		}
1959
		break;
L
Linus Torvalds 已提交
1960 1961 1962 1963 1964 1965 1966

	case 0:
		/*
		 * New allocation is not contiguous with another
		 * real allocation.
		 * Insert a new entry.
		 */
1967
		xfs_iext_insert(ip, idx, 1, new, state);
L
Linus Torvalds 已提交
1968 1969 1970 1971
		ifp->if_lastex = idx;
		XFS_IFORK_NEXT_SET(ip, whichfork,
			XFS_IFORK_NEXTENTS(ip, whichfork) + 1);
		if (cur == NULL) {
1972
			rval = XFS_ILOG_CORE | xfs_ilog_fext(whichfork);
1973 1974 1975 1976 1977 1978 1979
		} else {
			rval = XFS_ILOG_CORE;
			if ((error = xfs_bmbt_lookup_eq(cur,
					new->br_startoff,
					new->br_startblock,
					new->br_blockcount, &i)))
				goto done;
1980
			XFS_WANT_CORRUPTED_GOTO(i == 0, done);
1981
			cur->bc_rec.b.br_state = new->br_state;
1982
			if ((error = xfs_btree_insert(cur, &i)))
1983
				goto done;
1984
			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
L
Linus Torvalds 已提交
1985
		}
1986
		break;
L
Linus Torvalds 已提交
1987
	}
1988 1989 1990
done:
	*logflagsp = rval;
	return error;
L
Linus Torvalds 已提交
1991 1992
}

1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 2097 2098 2099 2100 2101 2102 2103 2104 2105 2106 2107 2108 2109 2110 2111 2112 2113 2114 2115 2116 2117 2118 2119 2120 2121 2122 2123 2124 2125 2126 2127 2128 2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140 2141 2142 2143 2144 2145 2146 2147 2148 2149 2150 2151 2152 2153 2154 2155
/*
 * Adjust the size of the new extent based on di_extsize and rt extsize.
 */
STATIC int
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 */
{
	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 */

	if (convert)
		return 0;

	orig_off = align_off = *offp;
	orig_alen = align_alen = *lenp;
	orig_end = orig_off + orig_alen;

	/*
	 * If this request overlaps an existing extent, then don't
	 * attempt to perform any additional alignment.
	 */
	if (!delay && !eof &&
	    (orig_off >= gotp->br_startoff) &&
	    (orig_end <= gotp->br_startoff + gotp->br_blockcount)) {
		return 0;
	}

	/*
	 * 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.
	 */
	temp = do_mod(orig_off, extsz);
	if (temp) {
		align_alen += temp;
		align_off -= temp;
	}
	/*
	 * Same adjustment for the end of the requested area.
	 */
	if ((temp = (align_alen % extsz))) {
		align_alen += extsz - temp;
	}
	/*
	 * If the previous block overlaps with this proposed allocation
	 * then move the start forward without adjusting the length.
	 */
	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;
	}

	/*
	 * 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))) {
		/*
		 * We're not covering the original request, or
		 * we won't be able to once we fix the length.
		 */
		if (orig_off < align_off ||
		    orig_end > align_off + align_alen ||
		    align_alen - temp < orig_alen)
			return XFS_ERROR(EINVAL);
		/*
		 * Try to fix it by moving the start up.
		 */
		if (align_off + temp <= orig_off) {
			align_alen -= temp;
			align_off += temp;
		}
		/*
		 * 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;
		}
		/*
		 * Result doesn't cover the request, fail it.
		 */
		if (orig_off < align_off || orig_end > align_off + align_alen)
			return XFS_ERROR(EINVAL);
	} else {
		ASSERT(orig_off >= align_off);
		ASSERT(orig_end <= align_off + align_alen);
	}

#ifdef DEBUG
	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

	*lenp = align_alen;
	*offp = align_off;
	return 0;
}

L
Linus Torvalds 已提交
2156 2157
#define XFS_ALLOC_GAP_UNITS	4

2158
STATIC void
2159
xfs_bmap_adjacent(
L
Linus Torvalds 已提交
2160 2161 2162 2163 2164 2165 2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180
	xfs_bmalloca_t	*ap)		/* bmap alloc argument struct */
{
	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 */

#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)

	mp = ap->ip->i_mount;
	nullfb = ap->firstblock == NULLFSBLOCK;
	rt = XFS_IS_REALTIME_INODE(ap->ip) && ap->userdata;
	fb_agno = nullfb ? NULLAGNUMBER : XFS_FSB_TO_AGNO(mp, ap->firstblock);
	/*
	 * If allocating at eof, and there's a previous real block,
M
Malcolm Parsons 已提交
2181
	 * try to use its last block as our starting point.
L
Linus Torvalds 已提交
2182 2183
	 */
	if (ap->eof && ap->prevp->br_startoff != NULLFILEOFF &&
2184
	    !isnullstartblock(ap->prevp->br_startblock) &&
L
Linus Torvalds 已提交
2185 2186 2187 2188 2189 2190 2191 2192 2193 2194 2195 2196 2197 2198 2199 2200 2201 2202 2203 2204 2205 2206 2207 2208 2209 2210 2211 2212
	    ISVALID(ap->prevp->br_startblock + ap->prevp->br_blockcount,
		    ap->prevp->br_startblock)) {
		ap->rval = ap->prevp->br_startblock + ap->prevp->br_blockcount;
		/*
		 * Adjust for the gap between prevp and us.
		 */
		adjust = ap->off -
			(ap->prevp->br_startoff + ap->prevp->br_blockcount);
		if (adjust &&
		    ISVALID(ap->rval + adjust, ap->prevp->br_startblock))
			ap->rval += 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->prevp->br_startoff != NULLFILEOFF &&
2213
		    !isnullstartblock(ap->prevp->br_startblock) &&
L
Linus Torvalds 已提交
2214 2215 2216 2217 2218 2219 2220 2221 2222 2223 2224 2225 2226 2227 2228 2229 2230 2231 2232 2233 2234 2235 2236 2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250 2251 2252 2253
		    (prevbno = ap->prevp->br_startblock +
			       ap->prevp->br_blockcount) &&
		    ISVALID(prevbno, ap->prevp->br_startblock)) {
			/*
			 * Calculate gap to end of previous block.
			 */
			adjust = prevdiff = ap->off -
				(ap->prevp->br_startoff +
				 ap->prevp->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->alen &&
			    ISVALID(prevbno + prevdiff,
				    ap->prevp->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;
		}
		/*
		 * 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.
		 */
2254
		if (!isnullstartblock(ap->gotp->br_startblock)) {
L
Linus Torvalds 已提交
2255 2256 2257 2258 2259 2260 2261 2262 2263 2264 2265 2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277 2278 2279 2280 2281 2282 2283 2284 2285 2286 2287 2288 2289 2290 2291 2292 2293 2294 2295 2296 2297 2298 2299 2300 2301 2302
			/*
			 * Calculate gap to start of next block.
			 */
			adjust = gotdiff = ap->gotp->br_startoff - ap->off;
			/*
			 * Figure the startblock based on the next block's
			 * start and the gap size.
			 */
			gotbno = ap->gotp->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->alen &&
			    ISVALID(gotbno - gotdiff, gotbno))
				gotbno -= adjust;
			else if (ISVALID(gotbno - ap->alen, gotbno)) {
				gotbno -= ap->alen;
				gotdiff += adjust - ap->alen;
			} 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->rval is already set (to 0 or the inode block).
		 */
		if (prevbno != NULLFSBLOCK && gotbno != NULLFSBLOCK)
			ap->rval = prevdiff <= gotdiff ? prevbno : gotbno;
		else if (prevbno != NULLFSBLOCK)
			ap->rval = prevbno;
		else if (gotbno != NULLFSBLOCK)
			ap->rval = gotbno;
	}
2303 2304 2305 2306 2307 2308 2309 2310 2311 2312 2313 2314 2315 2316 2317 2318
#undef ISVALID
}

STATIC int
xfs_bmap_rtalloc(
	xfs_bmalloca_t	*ap)		/* bmap alloc argument struct */
{
	xfs_alloctype_t	atype = 0;	/* type for allocation routines */
	int		error;		/* error return value */
	xfs_mount_t	*mp;		/* mount point structure */
	xfs_extlen_t	prod = 0;	/* product factor for allocators */
	xfs_extlen_t	ralen = 0;	/* realtime allocation length */
	xfs_extlen_t	align;		/* minimum allocation alignment */
	xfs_rtblock_t	rtb;

	mp = ap->ip->i_mount;
2319
	align = xfs_get_extsz_hint(ap->ip);
2320 2321 2322 2323 2324 2325 2326 2327 2328 2329 2330 2331 2332 2333 2334 2335 2336 2337 2338 2339 2340 2341 2342 2343 2344 2345 2346 2347 2348 2349 2350 2351 2352
	prod = align / mp->m_sb.sb_rextsize;
	error = xfs_bmap_extsize_align(mp, ap->gotp, ap->prevp,
					align, 1, ap->eof, 0,
					ap->conv, &ap->off, &ap->alen);
	if (error)
		return error;
	ASSERT(ap->alen);
	ASSERT(ap->alen % mp->m_sb.sb_rextsize == 0);

	/*
	 * If the offset & length are not perfectly aligned
	 * then kill prod, it will just get us in trouble.
	 */
	if (do_mod(ap->off, align) || ap->alen % align)
		prod = 1;
	/*
	 * Set ralen to be the actual requested length in rtextents.
	 */
	ralen = ap->alen / mp->m_sb.sb_rextsize;
	/*
	 * If the old value was close enough to MAXEXTLEN that
	 * we rounded up to it, cut it back so it's valid again.
	 * Note that if it's a really large request (bigger than
	 * MAXEXTLEN), we don't hear about that number, and can't
	 * adjust the starting point to match it.
	 */
	if (ralen * mp->m_sb.sb_rextsize >= MAXEXTLEN)
		ralen = MAXEXTLEN / mp->m_sb.sb_rextsize;
	/*
	 * If it's an allocation to an empty file at offset 0,
	 * pick an extent that will space things out in the rt area.
	 */
	if (ap->eof && ap->off == 0) {
2353 2354
		xfs_rtblock_t uninitialized_var(rtx); /* realtime extent no */

2355 2356 2357 2358 2359 2360 2361 2362 2363 2364 2365 2366 2367 2368 2369 2370 2371 2372 2373 2374 2375 2376 2377 2378 2379 2380 2381 2382 2383 2384 2385 2386 2387 2388 2389 2390 2391 2392
		error = xfs_rtpick_extent(mp, ap->tp, ralen, &rtx);
		if (error)
			return error;
		ap->rval = rtx * mp->m_sb.sb_rextsize;
	} else {
		ap->rval = 0;
	}

	xfs_bmap_adjacent(ap);

	/*
	 * Realtime allocation, done through xfs_rtallocate_extent.
	 */
	atype = ap->rval == 0 ?  XFS_ALLOCTYPE_ANY_AG : XFS_ALLOCTYPE_NEAR_BNO;
	do_div(ap->rval, mp->m_sb.sb_rextsize);
	rtb = ap->rval;
	ap->alen = ralen;
	if ((error = xfs_rtallocate_extent(ap->tp, ap->rval, 1, ap->alen,
				&ralen, atype, ap->wasdel, prod, &rtb)))
		return error;
	if (rtb == NULLFSBLOCK && prod > 1 &&
	    (error = xfs_rtallocate_extent(ap->tp, ap->rval, 1,
					   ap->alen, &ralen, atype,
					   ap->wasdel, 1, &rtb)))
		return error;
	ap->rval = rtb;
	if (ap->rval != NULLFSBLOCK) {
		ap->rval *= mp->m_sb.sb_rextsize;
		ralen *= mp->m_sb.sb_rextsize;
		ap->alen = ralen;
		ap->ip->i_d.di_nblocks += ralen;
		xfs_trans_log_inode(ap->tp, ap->ip, XFS_ILOG_CORE);
		if (ap->wasdel)
			ap->ip->i_delayed_blks -= ralen;
		/*
		 * Adjust the disk quota also. This was reserved
		 * earlier.
		 */
C
Christoph Hellwig 已提交
2393
		xfs_trans_mod_dquot_byino(ap->tp, ap->ip,
2394 2395 2396 2397 2398 2399 2400 2401
			ap->wasdel ? XFS_TRANS_DQ_DELRTBCOUNT :
					XFS_TRANS_DQ_RTBCOUNT, (long) ralen);
	} else {
		ap->alen = 0;
	}
	return 0;
}

C
Christoph Hellwig 已提交
2402 2403 2404 2405 2406 2407 2408 2409 2410 2411 2412 2413 2414 2415 2416 2417 2418 2419 2420 2421 2422 2423 2424 2425 2426 2427 2428 2429 2430 2431 2432 2433 2434 2435 2436 2437 2438 2439 2440 2441 2442 2443 2444 2445 2446 2447 2448 2449 2450 2451 2452 2453 2454 2455 2456 2457 2458 2459 2460 2461 2462 2463 2464 2465 2466 2467 2468 2469 2470 2471 2472 2473 2474 2475 2476 2477 2478 2479 2480 2481 2482 2483 2484 2485 2486 2487 2488 2489 2490 2491 2492 2493 2494 2495 2496 2497 2498 2499 2500 2501 2502 2503 2504 2505 2506 2507 2508 2509 2510 2511 2512 2513 2514 2515 2516
STATIC int
xfs_bmap_btalloc_nullfb(
	struct xfs_bmalloca	*ap,
	struct xfs_alloc_arg	*args,
	xfs_extlen_t		*blen)
{
	struct xfs_mount	*mp = ap->ip->i_mount;
	struct xfs_perag	*pag;
	xfs_agnumber_t		ag, startag;
	int			notinit = 0;
	int			error;

	if (ap->userdata && xfs_inode_is_filestream(ap->ip))
		args->type = XFS_ALLOCTYPE_NEAR_BNO;
	else
		args->type = XFS_ALLOCTYPE_START_BNO;
	args->total = ap->total;

	/*
	 * 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.
	 */
	startag = ag = XFS_FSB_TO_AGNO(mp, args->fsbno);
	if (startag == NULLAGNUMBER)
		startag = ag = 0;

	pag = xfs_perag_get(mp, ag);
	while (*blen < ap->alen) {
		if (!pag->pagf_init) {
			error = xfs_alloc_pagf_init(mp, args->tp, ag,
						    XFS_ALLOC_FLAG_TRYLOCK);
			if (error) {
				xfs_perag_put(pag);
				return error;
			}
		}

		/*
		 * See xfs_alloc_fix_freelist...
		 */
		if (pag->pagf_init) {
			xfs_extlen_t	longest;
			longest = xfs_alloc_longest_free_extent(mp, pag);
			if (*blen < longest)
				*blen = longest;
		} else
			notinit = 1;

		if (xfs_inode_is_filestream(ap->ip)) {
			if (*blen >= ap->alen)
				break;

			if (ap->userdata) {
				/*
				 * If startag is an invalid AG, we've
				 * come here once before and
				 * xfs_filestream_new_ag picked the
				 * best currently available.
				 *
				 * Don't continue looping, since we
				 * could loop forever.
				 */
				if (startag == NULLAGNUMBER)
					break;

				error = xfs_filestream_new_ag(ap, &ag);
				xfs_perag_put(pag);
				if (error)
					return error;

				/* loop again to set 'blen'*/
				startag = NULLAGNUMBER;
				pag = xfs_perag_get(mp, ag);
				continue;
			}
		}
		if (++ag == mp->m_sb.sb_agcount)
			ag = 0;
		if (ag == startag)
			break;
		xfs_perag_put(pag);
		pag = xfs_perag_get(mp, ag);
	}
	xfs_perag_put(pag);

	/*
	 * Since the above loop did a BUF_TRYLOCK, it is
	 * possible that there is space for this request.
	 */
	if (notinit || *blen < ap->minlen)
		args->minlen = ap->minlen;
	/*
	 * If the best seen length is less than the request
	 * length, use the best as the minimum.
	 */
	else if (*blen < ap->alen)
		args->minlen = *blen;
	/*
	 * Otherwise we've seen an extent as big as alen,
	 * use that as the minimum.
	 */
	else
		args->minlen = ap->alen;

	/*
	 * set the failure fallback case to look in the selected
	 * AG as the stream may have moved.
	 */
	if (xfs_inode_is_filestream(ap->ip))
		ap->rval = args->fsbno = XFS_AGB_TO_FSB(mp, ag, 0);

	return 0;
}

2517 2518 2519 2520 2521 2522 2523 2524
STATIC int
xfs_bmap_btalloc(
	xfs_bmalloca_t	*ap)		/* bmap alloc argument struct */
{
	xfs_mount_t	*mp;		/* mount point structure */
	xfs_alloctype_t	atype = 0;	/* type for allocation routines */
	xfs_extlen_t	align;		/* minimum allocation alignment */
	xfs_agnumber_t	fb_agno;	/* ag number of ap->firstblock */
C
Christoph Hellwig 已提交
2525
	xfs_agnumber_t	ag;
2526 2527 2528 2529 2530 2531 2532 2533 2534
	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;

	mp = ap->ip->i_mount;
2535
	align = ap->userdata ? xfs_get_extsz_hint(ap->ip) : 0;
2536 2537 2538 2539 2540 2541 2542 2543 2544
	if (unlikely(align)) {
		error = xfs_bmap_extsize_align(mp, ap->gotp, ap->prevp,
						align, 0, ap->eof, 0, ap->conv,
						&ap->off, &ap->alen);
		ASSERT(!error);
		ASSERT(ap->alen);
	}
	nullfb = ap->firstblock == NULLFSBLOCK;
	fb_agno = nullfb ? NULLAGNUMBER : XFS_FSB_TO_AGNO(mp, ap->firstblock);
2545 2546 2547 2548 2549 2550 2551 2552 2553
	if (nullfb) {
		if (ap->userdata && xfs_inode_is_filestream(ap->ip)) {
			ag = xfs_filestream_lookup_ag(ap->ip);
			ag = (ag != NULLAGNUMBER) ? ag : 0;
			ap->rval = XFS_AGB_TO_FSB(mp, ag, 0);
		} else {
			ap->rval = XFS_INO_TO_FSB(mp, ap->ip->i_ino);
		}
	} else
2554 2555 2556 2557
		ap->rval = ap->firstblock;

	xfs_bmap_adjacent(ap);

L
Linus Torvalds 已提交
2558 2559 2560 2561
	/*
	 * If allowed, use ap->rval; otherwise must use firstblock since
	 * it's in the right allocation group.
	 */
2562
	if (nullfb || XFS_FSB_TO_AGNO(mp, ap->rval) == fb_agno)
L
Linus Torvalds 已提交
2563 2564 2565 2566
		;
	else
		ap->rval = ap->firstblock;
	/*
2567
	 * Normal allocation, done through xfs_alloc_vextent.
L
Linus Torvalds 已提交
2568
	 */
2569 2570 2571 2572 2573
	tryagain = isaligned = 0;
	args.tp = ap->tp;
	args.mp = mp;
	args.fsbno = ap->rval;
	args.maxlen = MIN(ap->alen, mp->m_sb.sb_agblocks);
2574
	args.firstblock = ap->firstblock;
2575 2576
	blen = 0;
	if (nullfb) {
C
Christoph Hellwig 已提交
2577 2578 2579
		error = xfs_bmap_btalloc_nullfb(ap, &args, &blen);
		if (error)
			return error;
2580
	} else if (ap->low) {
2581 2582 2583 2584
		if (xfs_inode_is_filestream(ap->ip))
			args.type = XFS_ALLOCTYPE_FIRST_AG;
		else
			args.type = XFS_ALLOCTYPE_START_BNO;
2585 2586 2587 2588 2589 2590
		args.total = args.minlen = ap->minlen;
	} else {
		args.type = XFS_ALLOCTYPE_NEAR_BNO;
		args.total = ap->total;
		args.minlen = ap->minlen;
	}
2591 2592 2593
	/* apply extent size hints if obtained earlier */
	if (unlikely(align)) {
		args.prod = align;
2594 2595
		if ((args.mod = (xfs_extlen_t)do_mod(ap->off, args.prod)))
			args.mod = (xfs_extlen_t)(args.prod - args.mod);
2596
	} else if (mp->m_sb.sb_blocksize >= PAGE_CACHE_SIZE) {
2597 2598 2599
		args.prod = 1;
		args.mod = 0;
	} else {
2600
		args.prod = PAGE_CACHE_SIZE >> mp->m_sb.sb_blocklog;
2601 2602
		if ((args.mod = (xfs_extlen_t)(do_mod(ap->off, args.prod))))
			args.mod = (xfs_extlen_t)(args.prod - args.mod);
L
Linus Torvalds 已提交
2603 2604
	}
	/*
2605 2606 2607 2608 2609 2610 2611
	 * 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.
L
Linus Torvalds 已提交
2612
	 */
2613 2614 2615 2616 2617
	if (!ap->low && ap->aeof) {
		if (!ap->off) {
			args.alignment = mp->m_dalign;
			atype = args.type;
			isaligned = 1;
L
Linus Torvalds 已提交
2618
			/*
2619
			 * Adjust for alignment
L
Linus Torvalds 已提交
2620
			 */
2621 2622 2623 2624
			if (blen > args.alignment && blen <= ap->alen)
				args.minlen = blen - args.alignment;
			args.minalignslop = 0;
		} else {
L
Linus Torvalds 已提交
2625
			/*
2626 2627 2628
			 * First try an exact bno allocation.
			 * If it fails then do a near or start bno
			 * allocation with alignment turned on.
L
Linus Torvalds 已提交
2629
			 */
2630 2631 2632 2633
			atype = args.type;
			tryagain = 1;
			args.type = XFS_ALLOCTYPE_THIS_BNO;
			args.alignment = 1;
L
Linus Torvalds 已提交
2634
			/*
2635 2636 2637 2638
			 * 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.
L
Linus Torvalds 已提交
2639
			 */
2640 2641
			if (blen > mp->m_dalign && blen <= ap->alen)
				nextminlen = blen - mp->m_dalign;
L
Linus Torvalds 已提交
2642
			else
2643 2644 2645 2646 2647 2648 2649
				nextminlen = args.minlen;
			if (nextminlen + mp->m_dalign > args.minlen + 1)
				args.minalignslop =
					nextminlen + mp->m_dalign -
					args.minlen - 1;
			else
				args.minalignslop = 0;
L
Linus Torvalds 已提交
2650
		}
2651 2652 2653 2654 2655 2656 2657 2658 2659 2660 2661
	} else {
		args.alignment = 1;
		args.minalignslop = 0;
	}
	args.minleft = ap->minleft;
	args.wasdel = ap->wasdel;
	args.isfl = 0;
	args.userdata = ap->userdata;
	if ((error = xfs_alloc_vextent(&args)))
		return error;
	if (tryagain && args.fsbno == NULLFSBLOCK) {
L
Linus Torvalds 已提交
2662
		/*
2663 2664
		 * Exact allocation failed. Now try with alignment
		 * turned on.
L
Linus Torvalds 已提交
2665
		 */
2666 2667 2668 2669 2670 2671
		args.type = atype;
		args.fsbno = ap->rval;
		args.alignment = mp->m_dalign;
		args.minlen = nextminlen;
		args.minalignslop = 0;
		isaligned = 1;
L
Linus Torvalds 已提交
2672 2673
		if ((error = xfs_alloc_vextent(&args)))
			return error;
2674 2675 2676 2677 2678 2679 2680 2681 2682 2683 2684 2685 2686 2687 2688 2689 2690 2691 2692 2693 2694 2695 2696 2697 2698 2699 2700 2701 2702 2703 2704 2705 2706 2707 2708 2709 2710 2711 2712 2713 2714 2715
	}
	if (isaligned && args.fsbno == NULLFSBLOCK) {
		/*
		 * allocation failed, so turn off alignment and
		 * try again.
		 */
		args.type = atype;
		args.fsbno = ap->rval;
		args.alignment = 0;
		if ((error = xfs_alloc_vextent(&args)))
			return error;
	}
	if (args.fsbno == NULLFSBLOCK && nullfb &&
	    args.minlen > ap->minlen) {
		args.minlen = ap->minlen;
		args.type = XFS_ALLOCTYPE_START_BNO;
		args.fsbno = ap->rval;
		if ((error = xfs_alloc_vextent(&args)))
			return error;
	}
	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;
		ap->low = 1;
	}
	if (args.fsbno != NULLFSBLOCK) {
		ap->firstblock = ap->rval = args.fsbno;
		ASSERT(nullfb || fb_agno == args.agno ||
		       (ap->low && fb_agno < args.agno));
		ap->alen = 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.
		 */
C
Christoph Hellwig 已提交
2716
		xfs_trans_mod_dquot_byino(ap->tp, ap->ip,
2717 2718 2719 2720 2721 2722
			ap->wasdel ? XFS_TRANS_DQ_DELBCOUNT :
					XFS_TRANS_DQ_BCOUNT,
			(long) args.len);
	} else {
		ap->rval = NULLFSBLOCK;
		ap->alen = 0;
L
Linus Torvalds 已提交
2723 2724
	}
	return 0;
2725 2726 2727 2728 2729 2730 2731 2732 2733 2734
}

/*
 * 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(
	xfs_bmalloca_t	*ap)		/* bmap alloc argument struct */
{
2735
	if (XFS_IS_REALTIME_INODE(ap->ip) && ap->userdata)
2736 2737
		return xfs_bmap_rtalloc(ap);
	return xfs_bmap_btalloc(ap);
L
Linus Torvalds 已提交
2738 2739 2740 2741 2742
}

/*
 * Transform a btree format file with only one leaf node, where the
 * extents list will fit in the inode, into an extents format file.
2743
 * Since the file extents are already in-core, all we have to do is
L
Linus Torvalds 已提交
2744 2745 2746 2747 2748 2749 2750 2751 2752 2753 2754
 * 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 */
2755
	struct xfs_btree_block	*cblock;/* child btree block */
L
Linus Torvalds 已提交
2756 2757 2758 2759 2760
	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 */
2761
	__be64			*pp;	/* ptr to block address */
2762
	struct xfs_btree_block	*rblock;/* root btree block */
L
Linus Torvalds 已提交
2763

2764
	mp = ip->i_mount;
L
Linus Torvalds 已提交
2765 2766 2767 2768
	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;
2769 2770
	ASSERT(be16_to_cpu(rblock->bb_level) == 1);
	ASSERT(be16_to_cpu(rblock->bb_numrecs) == 1);
2771 2772
	ASSERT(xfs_bmbt_maxrecs(mp, ifp->if_broot_bytes, 0) == 1);
	pp = XFS_BMAP_BROOT_PTR_ADDR(mp, rblock, 1, ifp->if_broot_bytes);
2773
	cbno = be64_to_cpu(*pp);
L
Linus Torvalds 已提交
2774 2775
	*logflagsp = 0;
#ifdef DEBUG
2776
	if ((error = xfs_btree_check_lptr(cur, cbno, 1)))
L
Linus Torvalds 已提交
2777 2778 2779 2780 2781
		return error;
#endif
	if ((error = xfs_btree_read_bufl(mp, tp, cbno, 0, &cbp,
			XFS_BMAP_BTREE_REF)))
		return error;
2782 2783
	cblock = XFS_BUF_TO_BLOCK(cbp);
	if ((error = xfs_btree_check_block(cur, cblock, 0, cbp)))
L
Linus Torvalds 已提交
2784 2785 2786
		return error;
	xfs_bmap_add_free(cbno, 1, cur->bc_private.b.flist, mp);
	ip->i_d.di_nblocks--;
C
Christoph Hellwig 已提交
2787
	xfs_trans_mod_dquot_byino(tp, ip, XFS_TRANS_DQ_BCOUNT, -1L);
L
Linus Torvalds 已提交
2788 2789 2790 2791 2792 2793 2794
	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);
2795
	*logflagsp = XFS_ILOG_CORE | xfs_ilog_fext(whichfork);
L
Linus Torvalds 已提交
2796 2797 2798 2799
	return 0;
}

/*
2800
 * Called by xfs_bmapi to update file extent records and the btree
L
Linus Torvalds 已提交
2801 2802 2803 2804 2805 2806 2807 2808 2809
 * 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 */
	xfs_bmap_free_t		*flist,	/* list of extents to be freed */
	xfs_btree_cur_t		*cur,	/* if null, not a btree */
2810
	xfs_bmbt_irec_t		*del,	/* data to remove from extents */
L
Linus Torvalds 已提交
2811 2812 2813 2814 2815 2816 2817 2818 2819 2820
	int			*logflagsp, /* inode logging flags */
	int			whichfork, /* data or attr fork */
	int			rsvd)	/* OK to allocate reserved blocks */
{
	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 */
2821
	xfs_bmbt_rec_host_t	*ep;	/* current extent entry pointer */
L
Linus Torvalds 已提交
2822 2823 2824 2825 2826 2827 2828 2829 2830 2831 2832 2833 2834
	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 */
C
Christoph Hellwig 已提交
2835
	int			state = 0;
L
Linus Torvalds 已提交
2836 2837

	XFS_STATS_INC(xs_del_exlist);
C
Christoph Hellwig 已提交
2838 2839 2840 2841

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

L
Linus Torvalds 已提交
2842 2843
	mp = ip->i_mount;
	ifp = XFS_IFORK_PTR(ip, whichfork);
2844 2845
	ASSERT((idx >= 0) && (idx < ifp->if_bytes /
		(uint)sizeof(xfs_bmbt_rec_t)));
L
Linus Torvalds 已提交
2846
	ASSERT(del->br_blockcount > 0);
2847
	ep = xfs_iext_get_ext(ifp, idx);
L
Linus Torvalds 已提交
2848 2849 2850 2851 2852
	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);
2853 2854
	delay = isnullstartblock(got.br_startblock);
	ASSERT(isnullstartblock(del->br_startblock) == delay);
L
Linus Torvalds 已提交
2855 2856 2857 2858 2859 2860 2861 2862 2863 2864 2865
	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.
		 */
2866
		if (whichfork == XFS_DATA_FORK && XFS_IS_REALTIME_INODE(ip)) {
L
Linus Torvalds 已提交
2867 2868 2869 2870 2871 2872 2873 2874 2875 2876 2877 2878 2879 2880 2881 2882 2883 2884 2885 2886 2887 2888 2889 2890 2891 2892 2893 2894 2895 2896 2897 2898 2899 2900 2901
			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);
			if ((error = xfs_rtfree_extent(ip->i_transp, bno,
					(xfs_extlen_t)len)))
				goto done;
			do_fx = 0;
			nblks = len * mp->m_sb.sb_rextsize;
			qfield = XFS_TRANS_DQ_RTBCOUNT;
		}
		/*
		 * Ordinary allocation.
		 */
		else {
			do_fx = 1;
			nblks = del->br_blockcount;
			qfield = XFS_TRANS_DQ_BCOUNT;
		}
		/*
		 * Set up del_endblock and cur for later.
		 */
		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;
2902
			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
L
Linus Torvalds 已提交
2903 2904 2905
		}
		da_old = da_new = 0;
	} else {
2906
		da_old = startblockval(got.br_startblock);
L
Linus Torvalds 已提交
2907 2908 2909 2910 2911 2912 2913 2914 2915 2916 2917 2918 2919 2920
		da_new = 0;
		nblks = 0;
		do_fx = 0;
	}
	/*
	 * Set flag value to use in switch statement.
	 * Left-contig is 2, right-contig is 1.
	 */
	switch (((got.br_startoff == del->br_startoff) << 1) |
		(got_endoff == del_endoff)) {
	case 3:
		/*
		 * Matches the whole extent.  Delete the entry.
		 */
2921 2922
		xfs_iext_remove(ip, idx, 1,
				whichfork == XFS_ATTR_FORK ? BMAP_ATTRFORK : 0);
L
Linus Torvalds 已提交
2923 2924 2925 2926 2927 2928 2929
		ifp->if_lastex = idx;
		if (delay)
			break;
		XFS_IFORK_NEXT_SET(ip, whichfork,
			XFS_IFORK_NEXTENTS(ip, whichfork) - 1);
		flags |= XFS_ILOG_CORE;
		if (!cur) {
2930
			flags |= xfs_ilog_fext(whichfork);
L
Linus Torvalds 已提交
2931 2932
			break;
		}
2933
		if ((error = xfs_btree_delete(cur, &i)))
L
Linus Torvalds 已提交
2934
			goto done;
2935
		XFS_WANT_CORRUPTED_GOTO(i == 1, done);
L
Linus Torvalds 已提交
2936 2937 2938 2939 2940 2941
		break;

	case 2:
		/*
		 * Deleting the first part of the extent.
		 */
C
Christoph Hellwig 已提交
2942
		trace_xfs_bmap_pre_update(ip, idx, state, _THIS_IP_);
L
Linus Torvalds 已提交
2943 2944 2945 2946 2947 2948 2949
		xfs_bmbt_set_startoff(ep, del_endoff);
		temp = got.br_blockcount - del->br_blockcount;
		xfs_bmbt_set_blockcount(ep, temp);
		ifp->if_lastex = idx;
		if (delay) {
			temp = XFS_FILBLKS_MIN(xfs_bmap_worst_indlen(ip, temp),
				da_old);
2950
			xfs_bmbt_set_startblock(ep, nullstartblock((int)temp));
C
Christoph Hellwig 已提交
2951
			trace_xfs_bmap_post_update(ip, idx, state, _THIS_IP_);
L
Linus Torvalds 已提交
2952 2953 2954 2955
			da_new = temp;
			break;
		}
		xfs_bmbt_set_startblock(ep, del_endblock);
C
Christoph Hellwig 已提交
2956
		trace_xfs_bmap_post_update(ip, idx, state, _THIS_IP_);
L
Linus Torvalds 已提交
2957
		if (!cur) {
2958
			flags |= xfs_ilog_fext(whichfork);
L
Linus Torvalds 已提交
2959 2960 2961 2962 2963 2964 2965 2966 2967 2968 2969 2970 2971
			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;
C
Christoph Hellwig 已提交
2972
		trace_xfs_bmap_pre_update(ip, idx, state, _THIS_IP_);
L
Linus Torvalds 已提交
2973 2974 2975 2976 2977
		xfs_bmbt_set_blockcount(ep, temp);
		ifp->if_lastex = idx;
		if (delay) {
			temp = XFS_FILBLKS_MIN(xfs_bmap_worst_indlen(ip, temp),
				da_old);
2978
			xfs_bmbt_set_startblock(ep, nullstartblock((int)temp));
C
Christoph Hellwig 已提交
2979
			trace_xfs_bmap_post_update(ip, idx, state, _THIS_IP_);
L
Linus Torvalds 已提交
2980 2981 2982
			da_new = temp;
			break;
		}
C
Christoph Hellwig 已提交
2983
		trace_xfs_bmap_post_update(ip, idx, state, _THIS_IP_);
L
Linus Torvalds 已提交
2984
		if (!cur) {
2985
			flags |= xfs_ilog_fext(whichfork);
L
Linus Torvalds 已提交
2986 2987 2988 2989 2990 2991 2992 2993 2994 2995 2996 2997 2998 2999
			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;
C
Christoph Hellwig 已提交
3000
		trace_xfs_bmap_pre_update(ip, idx, state, _THIS_IP_);
L
Linus Torvalds 已提交
3001 3002 3003 3004 3005 3006 3007 3008 3009 3010 3011 3012 3013 3014
		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;
3015
				if ((error = xfs_btree_increment(cur, 0, &i)))
L
Linus Torvalds 已提交
3016 3017
					goto done;
				cur->bc_rec.b = new;
3018
				error = xfs_btree_insert(cur, &i);
L
Linus Torvalds 已提交
3019 3020 3021 3022 3023 3024 3025 3026 3027 3028 3029 3030 3031 3032 3033 3034 3035 3036
				if (error && error != ENOSPC)
					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.
				 */
				if (error == ENOSPC) {
					/*
					 * 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;
3037
					XFS_WANT_CORRUPTED_GOTO(i == 1, done);
L
Linus Torvalds 已提交
3038 3039 3040 3041 3042 3043 3044 3045 3046 3047 3048 3049 3050 3051 3052 3053 3054 3055 3056 3057
					/*
					 * 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;
					error = XFS_ERROR(ENOSPC);
					goto done;
				}
3058
				XFS_WANT_CORRUPTED_GOTO(i == 1, done);
L
Linus Torvalds 已提交
3059
			} else
3060
				flags |= xfs_ilog_fext(whichfork);
L
Linus Torvalds 已提交
3061 3062 3063 3064 3065
			XFS_IFORK_NEXT_SET(ip, whichfork,
				XFS_IFORK_NEXTENTS(ip, whichfork) + 1);
		} else {
			ASSERT(whichfork == XFS_DATA_FORK);
			temp = xfs_bmap_worst_indlen(ip, temp);
3066
			xfs_bmbt_set_startblock(ep, nullstartblock((int)temp));
L
Linus Torvalds 已提交
3067
			temp2 = xfs_bmap_worst_indlen(ip, temp2);
3068
			new.br_startblock = nullstartblock((int)temp2);
L
Linus Torvalds 已提交
3069 3070 3071 3072 3073 3074
			da_new = temp + temp2;
			while (da_new > da_old) {
				if (temp) {
					temp--;
					da_new--;
					xfs_bmbt_set_startblock(ep,
3075
						nullstartblock((int)temp));
L
Linus Torvalds 已提交
3076 3077 3078 3079 3080 3081 3082
				}
				if (da_new == da_old)
					break;
				if (temp2) {
					temp2--;
					da_new--;
					new.br_startblock =
3083
						nullstartblock((int)temp2);
L
Linus Torvalds 已提交
3084 3085 3086
				}
			}
		}
C
Christoph Hellwig 已提交
3087 3088
		trace_xfs_bmap_post_update(ip, idx, state, _THIS_IP_);
		xfs_iext_insert(ip, idx + 1, 1, &new, state);
L
Linus Torvalds 已提交
3089 3090 3091 3092 3093 3094 3095 3096 3097 3098 3099 3100 3101 3102 3103 3104 3105 3106
		ifp->if_lastex = idx + 1;
		break;
	}
	/*
	 * If we need to, add to list of extents to delete.
	 */
	if (do_fx)
		xfs_bmap_add_free(del->br_startblock, del->br_blockcount, flist,
			mp);
	/*
	 * Adjust inode # blocks in the file.
	 */
	if (nblks)
		ip->i_d.di_nblocks -= nblks;
	/*
	 * Adjust quota data.
	 */
	if (qfield)
C
Christoph Hellwig 已提交
3107
		xfs_trans_mod_dquot_byino(tp, ip, qfield, (long)-nblks);
L
Linus Torvalds 已提交
3108 3109 3110 3111 3112 3113 3114

	/*
	 * Account for change in delayed indirect blocks.
	 * Nothing to do for disk quota accounting here.
	 */
	ASSERT(da_old >= da_new);
	if (da_old > da_new)
3115
		xfs_mod_incore_sb(mp, XFS_SBS_FDBLOCKS, (int64_t)(da_old - da_new),
L
Linus Torvalds 已提交
3116 3117 3118 3119 3120 3121 3122 3123 3124 3125 3126 3127 3128 3129 3130 3131 3132 3133 3134 3135 3136 3137 3138 3139 3140 3141 3142 3143 3144 3145 3146 3147 3148 3149 3150 3151 3152 3153 3154
			rsvd);
done:
	*logflagsp = flags;
	return error;
}

/*
 * Remove the entry "free" from the free item list.  Prev points to the
 * previous entry, unless "free" is the head of the list.
 */
STATIC void
xfs_bmap_del_free(
	xfs_bmap_free_t		*flist,	/* free item list header */
	xfs_bmap_free_item_t	*prev,	/* previous item on list, if any */
	xfs_bmap_free_item_t	*free)	/* list item to be freed */
{
	if (prev)
		prev->xbfi_next = free->xbfi_next;
	else
		flist->xbf_first = free->xbfi_next;
	flist->xbf_count--;
	kmem_zone_free(xfs_bmap_free_item_zone, free);
}

/*
 * 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 */
	xfs_bmap_free_t		*flist,		/* blocks freed in xaction */
	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 */
{
3155
	struct xfs_btree_block	*ablock;	/* allocated (child) bt block */
L
Linus Torvalds 已提交
3156 3157 3158
	xfs_buf_t		*abp;		/* buffer for ablock */
	xfs_alloc_arg_t		args;		/* allocation arguments */
	xfs_bmbt_rec_t		*arp;		/* child record pointer */
3159
	struct xfs_btree_block	*block;		/* btree root block */
L
Linus Torvalds 已提交
3160
	xfs_btree_cur_t		*cur;		/* bmap btree cursor */
3161
	xfs_bmbt_rec_host_t	*ep;		/* extent record pointer */
L
Linus Torvalds 已提交
3162
	int			error;		/* error return value */
3163
	xfs_extnum_t		i, cnt;		/* extent record index */
L
Linus Torvalds 已提交
3164 3165 3166
	xfs_ifork_t		*ifp;		/* inode fork pointer */
	xfs_bmbt_key_t		*kp;		/* root block key pointer */
	xfs_mount_t		*mp;		/* mount structure */
3167
	xfs_extnum_t		nextents;	/* number of file extents */
L
Linus Torvalds 已提交
3168 3169 3170 3171 3172 3173 3174 3175 3176 3177 3178
	xfs_bmbt_ptr_t		*pp;		/* root block address pointer */

	ifp = XFS_IFORK_PTR(ip, whichfork);
	ASSERT(XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_EXTENTS);
	ASSERT(ifp->if_ext_max ==
	       XFS_IFORK_SIZE(ip, whichfork) / (uint)sizeof(xfs_bmbt_rec_t));
	/*
	 * Make space in the inode incore.
	 */
	xfs_iroot_realloc(ip, 1, whichfork);
	ifp->if_flags |= XFS_IFBROOT;
3179

L
Linus Torvalds 已提交
3180 3181 3182 3183
	/*
	 * Fill in the root.
	 */
	block = ifp->if_broot;
3184 3185 3186
	block->bb_magic = cpu_to_be32(XFS_BMAP_MAGIC);
	block->bb_level = cpu_to_be16(1);
	block->bb_numrecs = cpu_to_be16(1);
3187 3188 3189
	block->bb_u.l.bb_leftsib = cpu_to_be64(NULLDFSBNO);
	block->bb_u.l.bb_rightsib = cpu_to_be64(NULLDFSBNO);

L
Linus Torvalds 已提交
3190 3191 3192 3193
	/*
	 * Need a cursor.  Can't allocate until bb_level is filled in.
	 */
	mp = ip->i_mount;
3194
	cur = xfs_bmbt_init_cursor(mp, tp, ip, whichfork);
L
Linus Torvalds 已提交
3195 3196 3197 3198 3199 3200 3201 3202 3203
	cur->bc_private.b.firstblock = *firstblock;
	cur->bc_private.b.flist = flist;
	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);
	args.tp = tp;
	args.mp = mp;
3204
	args.firstblock = *firstblock;
L
Linus Torvalds 已提交
3205 3206 3207 3208 3209 3210 3211 3212 3213 3214 3215 3216 3217 3218 3219 3220 3221 3222 3223 3224 3225 3226 3227 3228 3229 3230 3231 3232 3233 3234 3235
	if (*firstblock == NULLFSBLOCK) {
		args.type = XFS_ALLOCTYPE_START_BNO;
		args.fsbno = XFS_INO_TO_FSB(mp, ip->i_ino);
	} else if (flist->xbf_low) {
		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.total = args.minleft = args.alignment = args.mod = args.isfl =
		args.minalignslop = 0;
	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) ||
	       (flist->xbf_low &&
		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++;
C
Christoph Hellwig 已提交
3236
	xfs_trans_mod_dquot_byino(tp, ip, XFS_TRANS_DQ_BCOUNT, 1L);
L
Linus Torvalds 已提交
3237 3238 3239 3240
	abp = xfs_btree_get_bufl(mp, tp, args.fsbno, 0);
	/*
	 * Fill in the child block.
	 */
3241
	ablock = XFS_BUF_TO_BLOCK(abp);
3242
	ablock->bb_magic = cpu_to_be32(XFS_BMAP_MAGIC);
L
Linus Torvalds 已提交
3243
	ablock->bb_level = 0;
3244 3245
	ablock->bb_u.l.bb_leftsib = cpu_to_be64(NULLDFSBNO);
	ablock->bb_u.l.bb_rightsib = cpu_to_be64(NULLDFSBNO);
3246
	arp = XFS_BMBT_REC_ADDR(mp, ablock, 1);
L
Linus Torvalds 已提交
3247
	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
3248 3249
	for (cnt = i = 0; i < nextents; i++) {
		ep = xfs_iext_get_ext(ifp, i);
3250
		if (!isnullstartblock(xfs_bmbt_get_startblock(ep))) {
3251 3252
			arp->l0 = cpu_to_be64(ep->l0);
			arp->l1 = cpu_to_be64(ep->l1);
L
Linus Torvalds 已提交
3253 3254 3255
			arp++; cnt++;
		}
	}
3256
	ASSERT(cnt == XFS_IFORK_NEXTENTS(ip, whichfork));
3257 3258
	xfs_btree_set_numrecs(ablock, cnt);

L
Linus Torvalds 已提交
3259 3260 3261
	/*
	 * Fill in the root key and pointer.
	 */
3262 3263
	kp = XFS_BMBT_KEY_ADDR(mp, block, 1);
	arp = XFS_BMBT_REC_ADDR(mp, ablock, 1);
3264
	kp->br_startoff = cpu_to_be64(xfs_bmbt_disk_get_startoff(arp));
3265 3266
	pp = XFS_BMBT_PTR_ADDR(mp, block, 1, xfs_bmbt_get_maxrecs(cur,
						be16_to_cpu(block->bb_level)));
3267
	*pp = cpu_to_be64(args.fsbno);
3268

L
Linus Torvalds 已提交
3269 3270 3271 3272
	/*
	 * Do all this logging at the end so that
	 * the root is at the right level.
	 */
3273 3274
	xfs_btree_log_block(cur, abp, XFS_BB_ALL_BITS);
	xfs_btree_log_recs(cur, abp, 1, be16_to_cpu(ablock->bb_numrecs));
L
Linus Torvalds 已提交
3275 3276
	ASSERT(*curp == NULL);
	*curp = cur;
3277
	*logflagsp = XFS_ILOG_CORE | xfs_ilog_fbroot(whichfork);
L
Linus Torvalds 已提交
3278 3279 3280
	return 0;
}

C
Christoph Hellwig 已提交
3281 3282 3283 3284 3285 3286 3287 3288 3289 3290 3291 3292 3293 3294 3295 3296 3297 3298 3299 3300 3301
/*
 * Calculate the default attribute fork offset for newly created inodes.
 */
uint
xfs_default_attroffset(
	struct xfs_inode	*ip)
{
	struct xfs_mount	*mp = ip->i_mount;
	uint			offset;

	if (mp->m_sb.sb_inodesize == 256) {
		offset = XFS_LITINO(mp) -
				XFS_BMDR_SPACE_CALC(MINABTPTRS);
	} else {
		offset = XFS_BMDR_SPACE_CALC(6 * MINABTPTRS);
	}

	ASSERT(offset < XFS_LITINO(mp));
	return offset;
}

3302 3303 3304 3305 3306 3307 3308 3309 3310 3311 3312 3313
/*
 * 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.
 */
STATIC void
xfs_bmap_forkoff_reset(
	xfs_mount_t	*mp,
	xfs_inode_t	*ip,
	int		whichfork)
{
	if (whichfork == XFS_ATTR_FORK &&
C
Christoph Hellwig 已提交
3314 3315 3316 3317 3318 3319 3320 3321 3322 3323 3324 3325
	    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;
			ip->i_df.if_ext_max =
				XFS_IFORK_DSIZE(ip) / sizeof(xfs_bmbt_rec_t);
			ip->i_afp->if_ext_max =
				XFS_IFORK_ASIZE(ip) / sizeof(xfs_bmbt_rec_t);
		}
3326 3327 3328
	}
}

L
Linus Torvalds 已提交
3329 3330 3331 3332 3333 3334 3335 3336 3337 3338 3339 3340 3341 3342 3343 3344 3345 3346 3347 3348 3349 3350 3351 3352 3353 3354 3355 3356 3357 3358 3359
/*
 * 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).
 */
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)	/* data or attr fork */
{
	int		error;		/* error return value */
	int		flags;		/* logging flags returned */
	xfs_ifork_t	*ifp;		/* inode fork pointer */

	/*
	 * 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.
	 */
	ASSERT(!((ip->i_d.di_mode & S_IFMT) == S_IFREG &&
		 whichfork == XFS_DATA_FORK));
	ifp = XFS_IFORK_PTR(ip, whichfork);
	ASSERT(XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_LOCAL);
	flags = 0;
	error = 0;
	if (ifp->if_bytes) {
		xfs_alloc_arg_t	args;	/* allocation arguments */
3360
		xfs_buf_t	*bp;	/* buffer for extent block */
3361
		xfs_bmbt_rec_host_t *ep;/* extent record pointer */
L
Linus Torvalds 已提交
3362 3363 3364

		args.tp = tp;
		args.mp = ip->i_mount;
3365
		args.firstblock = *firstblock;
3366 3367
		ASSERT((ifp->if_flags &
			(XFS_IFINLINE|XFS_IFEXTENTS|XFS_IFEXTIREC)) == XFS_IFINLINE);
L
Linus Torvalds 已提交
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
		/*
		 * 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;
		} else {
			args.fsbno = *firstblock;
			args.type = XFS_ALLOCTYPE_NEAR_BNO;
		}
		args.total = total;
		args.mod = args.minleft = args.alignment = args.wasdel =
			args.isfl = args.minalignslop = 0;
		args.minlen = args.maxlen = args.prod = 1;
		if ((error = xfs_alloc_vextent(&args)))
			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);
		memcpy((char *)XFS_BUF_PTR(bp), ifp->if_u1.if_data,
			ifp->if_bytes);
		xfs_trans_log_buf(tp, bp, 0, ifp->if_bytes - 1);
3395
		xfs_bmap_forkoff_reset(args.mp, ip, whichfork);
L
Linus Torvalds 已提交
3396
		xfs_idata_realloc(ip, -ifp->if_bytes, whichfork);
3397 3398
		xfs_iext_add(ifp, 0, 1);
		ep = xfs_iext_get_ext(ifp, 0);
L
Linus Torvalds 已提交
3399
		xfs_bmbt_set_allf(ep, 0, args.fsbno, 1, XFS_EXT_NORM);
C
Christoph Hellwig 已提交
3400 3401 3402
		trace_xfs_bmap_post_update(ip, 0,
				whichfork == XFS_ATTR_FORK ? BMAP_ATTRFORK : 0,
				_THIS_IP_);
L
Linus Torvalds 已提交
3403 3404
		XFS_IFORK_NEXT_SET(ip, whichfork, 1);
		ip->i_d.di_nblocks = 1;
C
Christoph Hellwig 已提交
3405
		xfs_trans_mod_dquot_byino(tp, ip,
L
Linus Torvalds 已提交
3406
			XFS_TRANS_DQ_BCOUNT, 1L);
3407
		flags |= xfs_ilog_fext(whichfork);
3408
	} else {
L
Linus Torvalds 已提交
3409
		ASSERT(XFS_IFORK_NEXTENTS(ip, whichfork) == 0);
3410 3411
		xfs_bmap_forkoff_reset(ip->i_mount, ip, whichfork);
	}
L
Linus Torvalds 已提交
3412 3413 3414 3415 3416 3417 3418 3419 3420
	ifp->if_flags &= ~XFS_IFINLINE;
	ifp->if_flags |= XFS_IFEXTENTS;
	XFS_IFORK_FMT_SET(ip, whichfork, XFS_DINODE_FMT_EXTENTS);
	flags |= XFS_ILOG_CORE;
done:
	*logflagsp = flags;
	return error;
}

3421
/*
3422 3423 3424 3425 3426
 * 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.
3427
 */
3428
STATIC xfs_bmbt_rec_host_t *		/* pointer to found extent entry */
3429 3430 3431 3432 3433 3434 3435 3436
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 */
{
3437
	xfs_bmbt_rec_host_t *ep;		/* extent record pointer */
3438 3439 3440
	xfs_extnum_t	lastx;		/* last extent index */

	/*
3441 3442
	 * Initialize the extent entry structure to catch access to
	 * uninitialized br_startblock field.
3443
	 */
3444 3445 3446 3447 3448 3449 3450 3451 3452 3453 3454 3455 3456
	gotp->br_startoff = 0xffa5a5a5a5a5a5a5LL;
	gotp->br_blockcount = 0xa55a5a5a5a5a5a5aLL;
	gotp->br_state = XFS_EXT_INVALID;
#if XFS_BIG_BLKNOS
	gotp->br_startblock = 0xffffa5a5a5a5a5a5LL;
#else
	gotp->br_startblock = 0xffffa5a5;
#endif
	prevp->br_startoff = NULLFILEOFF;

	ep = xfs_iext_bno_to_ext(ifp, bno, &lastx);
	if (lastx > 0) {
		xfs_bmbt_get_all(xfs_iext_get_ext(ifp, lastx - 1), prevp);
3457
	}
3458 3459 3460 3461 3462 3463 3464 3465 3466
	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;
3467
	}
3468
	*lastxp = lastx;
3469 3470 3471
	return ep;
}

L
Linus Torvalds 已提交
3472 3473 3474 3475 3476 3477 3478
/*
 * 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.
 */
3479
STATIC xfs_bmbt_rec_host_t *                 /* pointer to found extent entry */
L
Linus Torvalds 已提交
3480 3481 3482
xfs_bmap_search_extents(
	xfs_inode_t     *ip,            /* incore inode pointer */
	xfs_fileoff_t   bno,            /* block number searched for */
3483
	int             fork,      	/* data or attr fork */
L
Linus Torvalds 已提交
3484 3485 3486 3487 3488 3489
	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_ifork_t	*ifp;		/* inode fork pointer */
3490
	xfs_bmbt_rec_host_t  *ep;            /* extent record pointer */
L
Linus Torvalds 已提交
3491 3492

	XFS_STATS_INC(xs_look_exlist);
3493
	ifp = XFS_IFORK_PTR(ip, fork);
L
Linus Torvalds 已提交
3494

3495 3496
	ep = xfs_bmap_search_multi_extents(ifp, bno, eofp, lastxp, gotp, prevp);

3497 3498 3499 3500 3501 3502 3503
	if (unlikely(!(gotp->br_startblock) && (*lastxp != NULLEXTNUM) &&
		     !(XFS_IS_REALTIME_INODE(ip) && fork == XFS_DATA_FORK))) {
		xfs_cmn_err(XFS_PTAG_FSBLOCK_ZERO, CE_ALERT, ip->i_mount,
				"Access to block zero in inode %llu "
				"start_block: %llx start_off: %llx "
				"blkcnt: %llx extent-state: %x lastx: %x\n",
			(unsigned long long)ip->i_ino,
3504 3505 3506
			(unsigned long long)gotp->br_startblock,
			(unsigned long long)gotp->br_startoff,
			(unsigned long long)gotp->br_blockcount,
3507 3508 3509 3510 3511 3512
			gotp->br_state, *lastxp);
		*lastxp = NULLEXTNUM;
		*eofp = 1;
		return NULL;
	}
	return ep;
L
Linus Torvalds 已提交
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 3545 3546 3547 3548 3549 3550 3551 3552
}

/*
 * Compute the worst-case number of indirect blocks that will be used
 * for ip's delayed extent of length "len".
 */
STATIC xfs_filblks_t
xfs_bmap_worst_indlen(
	xfs_inode_t	*ip,		/* incore inode pointer */
	xfs_filblks_t	len)		/* delayed extent length */
{
	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 */

	mp = ip->i_mount;
	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];
	}
	return rval;
}

/*
 * 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 */
3553 3554
	int			size,		/* space new attribute needs */
	int			rsvd)		/* xact may use reserved blks */
L
Linus Torvalds 已提交
3555 3556
{
	xfs_fsblock_t		firstblock;	/* 1st block/ag allocated */
3557
	xfs_bmap_free_t		flist;		/* freed extent records */
L
Linus Torvalds 已提交
3558 3559
	xfs_mount_t		*mp;		/* mount structure */
	xfs_trans_t		*tp;		/* transaction pointer */
3560 3561 3562 3563 3564
	int			blks;		/* space reservation */
	int			version = 1;	/* superblock attr version */
	int			committed;	/* xaction was committed */
	int			logflags;	/* logging flags */
	int			error;		/* error return value */
L
Linus Torvalds 已提交
3565

3566
	ASSERT(XFS_IFORK_Q(ip) == 0);
L
Linus Torvalds 已提交
3567 3568
	ASSERT(ip->i_df.if_ext_max ==
	       XFS_IFORK_DSIZE(ip) / (uint)sizeof(xfs_bmbt_rec_t));
3569

L
Linus Torvalds 已提交
3570 3571 3572 3573 3574 3575 3576 3577 3578 3579
	mp = ip->i_mount;
	ASSERT(!XFS_NOT_DQATTACHED(mp, ip));
	tp = xfs_trans_alloc(mp, XFS_TRANS_ADDAFORK);
	blks = XFS_ADDAFORK_SPACE_RES(mp);
	if (rsvd)
		tp->t_flags |= XFS_TRANS_RESERVE;
	if ((error = xfs_trans_reserve(tp, blks, XFS_ADDAFORK_LOG_RES(mp), 0,
			XFS_TRANS_PERM_LOG_RES, XFS_ADDAFORK_LOG_COUNT)))
		goto error0;
	xfs_ilock(ip, XFS_ILOCK_EXCL);
C
Christoph Hellwig 已提交
3580
	error = xfs_trans_reserve_quota_nblks(tp, ip, blks, 0, rsvd ?
L
Linus Torvalds 已提交
3581 3582 3583 3584 3585 3586 3587 3588 3589 3590 3591 3592 3593 3594 3595 3596 3597
			XFS_QMOPT_RES_REGBLKS | XFS_QMOPT_FORCE_RES :
			XFS_QMOPT_RES_REGBLKS);
	if (error) {
		xfs_iunlock(ip, XFS_ILOCK_EXCL);
		xfs_trans_cancel(tp, XFS_TRANS_RELEASE_LOG_RES);
		return error;
	}
	if (XFS_IFORK_Q(ip))
		goto error1;
	if (ip->i_d.di_aformat != XFS_DINODE_FMT_EXTENTS) {
		/*
		 * For inodes coming from pre-6.2 filesystems.
		 */
		ASSERT(ip->i_d.di_aformat == 0);
		ip->i_d.di_aformat = XFS_DINODE_FMT_EXTENTS;
	}
	ASSERT(ip->i_d.di_anextents == 0);
3598 3599

	xfs_trans_ijoin_ref(tp, ip, XFS_ILOCK_EXCL);
L
Linus Torvalds 已提交
3600
	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
3601

L
Linus Torvalds 已提交
3602 3603 3604 3605 3606 3607 3608 3609 3610 3611
	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:
3612 3613
		ip->i_d.di_forkoff = xfs_attr_shortform_bytesfit(ip, size);
		if (!ip->i_d.di_forkoff)
C
Christoph Hellwig 已提交
3614
			ip->i_d.di_forkoff = xfs_default_attroffset(ip) >> 3;
3615
		else if (mp->m_flags & XFS_MOUNT_ATTR2)
3616
			version = 2;
L
Linus Torvalds 已提交
3617 3618 3619 3620 3621 3622 3623 3624 3625 3626 3627 3628 3629 3630
		break;
	default:
		ASSERT(0);
		error = XFS_ERROR(EINVAL);
		goto error1;
	}
	ip->i_df.if_ext_max =
		XFS_IFORK_DSIZE(ip) / (uint)sizeof(xfs_bmbt_rec_t);
	ASSERT(ip->i_afp == NULL);
	ip->i_afp = kmem_zone_zalloc(xfs_ifork_zone, KM_SLEEP);
	ip->i_afp->if_ext_max =
		XFS_IFORK_ASIZE(ip) / (uint)sizeof(xfs_bmbt_rec_t);
	ip->i_afp->if_flags = XFS_IFEXTENTS;
	logflags = 0;
3631
	xfs_bmap_init(&flist, &firstblock);
L
Linus Torvalds 已提交
3632 3633 3634 3635 3636 3637 3638 3639 3640 3641 3642 3643 3644 3645 3646 3647 3648 3649 3650 3651 3652
	switch (ip->i_d.di_format) {
	case XFS_DINODE_FMT_LOCAL:
		error = xfs_bmap_add_attrfork_local(tp, ip, &firstblock, &flist,
			&logflags);
		break;
	case XFS_DINODE_FMT_EXTENTS:
		error = xfs_bmap_add_attrfork_extents(tp, ip, &firstblock,
			&flist, &logflags);
		break;
	case XFS_DINODE_FMT_BTREE:
		error = xfs_bmap_add_attrfork_btree(tp, ip, &firstblock, &flist,
			&logflags);
		break;
	default:
		error = 0;
		break;
	}
	if (logflags)
		xfs_trans_log_inode(tp, ip, logflags);
	if (error)
		goto error2;
3653 3654
	if (!xfs_sb_version_hasattr(&mp->m_sb) ||
	   (!xfs_sb_version_hasattr2(&mp->m_sb) && version == 2)) {
3655 3656
		__int64_t sbfields = 0;

E
Eric Sandeen 已提交
3657
		spin_lock(&mp->m_sb_lock);
3658 3659
		if (!xfs_sb_version_hasattr(&mp->m_sb)) {
			xfs_sb_version_addattr(&mp->m_sb);
3660
			sbfields |= XFS_SB_VERSIONNUM;
3661
		}
3662 3663
		if (!xfs_sb_version_hasattr2(&mp->m_sb) && version == 2) {
			xfs_sb_version_addattr2(&mp->m_sb);
3664
			sbfields |= (XFS_SB_VERSIONNUM | XFS_SB_FEATURES2);
3665
		}
3666
		if (sbfields) {
E
Eric Sandeen 已提交
3667
			spin_unlock(&mp->m_sb_lock);
3668
			xfs_mod_sb(tp, sbfields);
L
Linus Torvalds 已提交
3669
		} else
E
Eric Sandeen 已提交
3670
			spin_unlock(&mp->m_sb_lock);
L
Linus Torvalds 已提交
3671
	}
3672
	if ((error = xfs_bmap_finish(&tp, &flist, &committed)))
L
Linus Torvalds 已提交
3673
		goto error2;
3674
	error = xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES);
L
Linus Torvalds 已提交
3675 3676 3677 3678 3679 3680 3681 3682 3683 3684 3685 3686 3687 3688 3689 3690 3691 3692 3693 3694 3695 3696 3697 3698 3699 3700 3701 3702 3703 3704 3705 3706 3707 3708 3709 3710
	ASSERT(ip->i_df.if_ext_max ==
	       XFS_IFORK_DSIZE(ip) / (uint)sizeof(xfs_bmbt_rec_t));
	return error;
error2:
	xfs_bmap_cancel(&flist);
error1:
	xfs_iunlock(ip, XFS_ILOCK_EXCL);
error0:
	xfs_trans_cancel(tp, XFS_TRANS_RELEASE_LOG_RES|XFS_TRANS_ABORT);
	ASSERT(ip->i_df.if_ext_max ==
	       XFS_IFORK_DSIZE(ip) / (uint)sizeof(xfs_bmbt_rec_t));
	return error;
}

/*
 * Add the extent to the list of extents to be free at transaction end.
 * The list is maintained sorted (by block number).
 */
/* ARGSUSED */
void
xfs_bmap_add_free(
	xfs_fsblock_t		bno,		/* fs block number of extent */
	xfs_filblks_t		len,		/* length of extent */
	xfs_bmap_free_t		*flist,		/* list of extents */
	xfs_mount_t		*mp)		/* mount point structure */
{
	xfs_bmap_free_item_t	*cur;		/* current (next) element */
	xfs_bmap_free_item_t	*new;		/* new element */
	xfs_bmap_free_item_t	*prev;		/* previous element */
#ifdef DEBUG
	xfs_agnumber_t		agno;
	xfs_agblock_t		agbno;

	ASSERT(bno != NULLFSBLOCK);
	ASSERT(len > 0);
	ASSERT(len <= MAXEXTLEN);
3711
	ASSERT(!isnullstartblock(bno));
L
Linus Torvalds 已提交
3712 3713 3714 3715 3716 3717 3718 3719 3720 3721 3722 3723 3724 3725 3726 3727 3728 3729 3730 3731 3732 3733 3734 3735 3736 3737 3738 3739 3740 3741 3742 3743 3744 3745 3746 3747 3748 3749 3750 3751 3752 3753 3754 3755 3756 3757 3758
	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);
	new = kmem_zone_alloc(xfs_bmap_free_item_zone, KM_SLEEP);
	new->xbfi_startblock = bno;
	new->xbfi_blockcount = (xfs_extlen_t)len;
	for (prev = NULL, cur = flist->xbf_first;
	     cur != NULL;
	     prev = cur, cur = cur->xbfi_next) {
		if (cur->xbfi_startblock >= bno)
			break;
	}
	if (prev)
		prev->xbfi_next = new;
	else
		flist->xbf_first = new;
	new->xbfi_next = cur;
	flist->xbf_count++;
}

/*
 * Compute and fill in the value of the maximum depth of a bmap btree
 * in this filesystem.  Done once, during mount.
 */
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 */

	/*
	 * 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).
3759 3760
	 *
	 * Note that we can no longer assume that if we are in ATTR1 that
C
Christoph Hellwig 已提交
3761 3762 3763 3764 3765 3766
	 * 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.
L
Linus Torvalds 已提交
3767
	 */
3768 3769
	if (whichfork == XFS_DATA_FORK) {
		maxleafents = MAXEXTNUM;
3770
		sz = XFS_BMDR_SPACE_CALC(MINDBTPTRS);
3771 3772
	} else {
		maxleafents = MAXAEXTNUM;
3773
		sz = XFS_BMDR_SPACE_CALC(MINABTPTRS);
3774
	}
3775
	maxrootrecs = xfs_bmdr_maxrecs(mp, sz, 0);
L
Linus Torvalds 已提交
3776 3777 3778 3779 3780 3781 3782 3783 3784 3785 3786 3787 3788 3789 3790 3791 3792 3793 3794 3795 3796 3797 3798 3799 3800 3801 3802 3803 3804 3805 3806 3807 3808 3809 3810 3811 3812 3813
	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;
}

/*
 * Routine to be called at transaction's end by xfs_bmapi, xfs_bunmapi
 * caller.  Frees all the extents that need freeing, which must be done
 * last due to locking considerations.  We never free any extents in
 * the first transaction.  This is to allow the caller to make the first
 * transaction a synchronous one so that the pointers to the data being
 * broken in this transaction will be permanent before the data is actually
 * freed.  This is necessary to prevent blocks from being reallocated
 * and written to before the free and reallocation are actually permanent.
 * We do not just make the first transaction synchronous here, because
 * there are more efficient ways to gain the same protection in some cases
 * (see the file truncation code).
 *
 * Return 1 if the given transaction was committed and a new one
 * started, and 0 otherwise in the committed parameter.
 */
/*ARGSUSED*/
int						/* error */
xfs_bmap_finish(
	xfs_trans_t		**tp,		/* transaction pointer addr */
	xfs_bmap_free_t		*flist,		/* i/o: list extents to free */
	int			*committed)	/* xact committed or not */
{
	xfs_efd_log_item_t	*efd;		/* extent free data */
	xfs_efi_log_item_t	*efi;		/* extent free intention */
	int			error;		/* error return value */
3814
	xfs_bmap_free_item_t	*free;		/* free extent item */
L
Linus Torvalds 已提交
3815 3816 3817 3818 3819 3820 3821 3822 3823 3824 3825 3826 3827 3828 3829 3830 3831 3832 3833
	unsigned int		logres;		/* new log reservation */
	unsigned int		logcount;	/* new log count */
	xfs_mount_t		*mp;		/* filesystem mount structure */
	xfs_bmap_free_item_t	*next;		/* next item on free list */
	xfs_trans_t		*ntp;		/* new transaction pointer */

	ASSERT((*tp)->t_flags & XFS_TRANS_PERM_LOG_RES);
	if (flist->xbf_count == 0) {
		*committed = 0;
		return 0;
	}
	ntp = *tp;
	efi = xfs_trans_get_efi(ntp, flist->xbf_count);
	for (free = flist->xbf_first; free; free = free->xbfi_next)
		xfs_trans_log_efi_extent(ntp, efi, free->xbfi_startblock,
			free->xbfi_blockcount);
	logres = ntp->t_log_res;
	logcount = ntp->t_log_count;
	ntp = xfs_trans_dup(*tp);
3834
	error = xfs_trans_commit(*tp, 0);
L
Linus Torvalds 已提交
3835 3836 3837 3838 3839 3840
	*tp = ntp;
	*committed = 1;
	/*
	 * We have a new transaction, so we should return committed=1,
	 * even though we're returning an error.
	 */
3841
	if (error)
L
Linus Torvalds 已提交
3842
		return error;
3843 3844 3845 3846 3847 3848 3849

	/*
	 * transaction commit worked ok so we can drop the extra ticket
	 * reference that we gained in xfs_trans_dup()
	 */
	xfs_log_ticket_put(ntp->t_ticket);

L
Linus Torvalds 已提交
3850 3851 3852 3853 3854 3855 3856 3857 3858 3859 3860 3861 3862 3863 3864 3865 3866 3867 3868 3869
	if ((error = xfs_trans_reserve(ntp, 0, logres, 0, XFS_TRANS_PERM_LOG_RES,
			logcount)))
		return error;
	efd = xfs_trans_get_efd(ntp, efi, flist->xbf_count);
	for (free = flist->xbf_first; free != NULL; free = next) {
		next = free->xbfi_next;
		if ((error = xfs_free_extent(ntp, free->xbfi_startblock,
				free->xbfi_blockcount))) {
			/*
			 * The bmap free list will be cleaned up at a
			 * higher level.  The EFI will be canceled when
			 * this transaction is aborted.
			 * Need to force shutdown here to make sure it
			 * happens, since this transaction may not be
			 * dirty yet.
			 */
			mp = ntp->t_mountp;
			if (!XFS_FORCED_SHUTDOWN(mp))
				xfs_force_shutdown(mp,
						   (error == EFSCORRUPTED) ?
3870 3871
						   SHUTDOWN_CORRUPT_INCORE :
						   SHUTDOWN_META_IO_ERROR);
L
Linus Torvalds 已提交
3872 3873 3874 3875 3876 3877 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
			return error;
		}
		xfs_trans_log_efd_extent(ntp, efd, free->xbfi_startblock,
			free->xbfi_blockcount);
		xfs_bmap_del_free(flist, NULL, free);
	}
	return 0;
}

/*
 * Free up any items left in the list.
 */
void
xfs_bmap_cancel(
	xfs_bmap_free_t		*flist)	/* list of bmap_free_items */
{
	xfs_bmap_free_item_t	*free;	/* free list item */
	xfs_bmap_free_item_t	*next;

	if (flist->xbf_count == 0)
		return;
	ASSERT(flist->xbf_first != NULL);
	for (free = flist->xbf_first; free; free = next) {
		next = free->xbfi_next;
		xfs_bmap_del_free(flist, NULL, free);
	}
	ASSERT(flist->xbf_count == 0);
}

/*
 * 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 */
3917
	int		idx;			/* extent record index */
L
Linus Torvalds 已提交
3918 3919 3920 3921 3922 3923 3924 3925 3926 3927 3928 3929 3930 3931 3932 3933 3934 3935 3936 3937
	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;
	}
	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);
3938
	for (idx = 0, lastaddr = 0, max = lowest; idx < nextents; idx++) {
3939
		xfs_bmbt_rec_host_t *ep = xfs_iext_get_ext(ifp, idx);
L
Linus Torvalds 已提交
3940 3941 3942 3943 3944 3945 3946 3947 3948 3949 3950 3951 3952 3953 3954 3955 3956 3957
		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);
	}
	*first_unused = max;
	return 0;
}

/*
 * Returns the file-relative block number of the last block + 1 before
 * last_block (input value) in the file.
3958 3959
 * 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.
L
Linus Torvalds 已提交
3960 3961 3962 3963 3964 3965 3966 3967 3968 3969
 */
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 */
3970
	xfs_bmbt_rec_host_t *ep;		/* pointer to last extent */
L
Linus Torvalds 已提交
3971 3972 3973 3974 3975 3976 3977 3978 3979 3980 3981 3982 3983 3984 3985 3986 3987 3988 3989 3990 3991 3992 3993 3994 3995 3996 3997 3998 3999 4000 4001 4002 4003 4004 4005
	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)
	       return XFS_ERROR(EIO);
	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;
		else
			*last_block = prev.br_startoff + prev.br_blockcount;
	}
	/*
	 * Otherwise *last_block is already the right answer.
	 */
	return 0;
}

/*
 * Returns the file-relative block number of the first block past eof in
4006 4007
 * 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.
L
Linus Torvalds 已提交
4008 4009 4010 4011 4012 4013 4014 4015
 */
int						/* error */
xfs_bmap_last_offset(
	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 */
{
4016
	xfs_bmbt_rec_host_t *ep;		/* pointer to last extent */
L
Linus Torvalds 已提交
4017 4018 4019 4020 4021 4022 4023 4024 4025 4026 4027 4028 4029 4030 4031 4032 4033 4034 4035 4036 4037
	int		error;			/* error return value */
	xfs_ifork_t	*ifp;			/* inode fork pointer */
	xfs_extnum_t	nextents;		/* number of extent entries */

	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)
	       return XFS_ERROR(EIO);
	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;
	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
	if (!nextents) {
		*last_block = 0;
		return 0;
	}
4038
	ep = xfs_iext_get_ext(ifp, nextents - 1);
L
Linus Torvalds 已提交
4039 4040 4041 4042 4043 4044 4045 4046 4047 4048 4049 4050 4051 4052
	*last_block = xfs_bmbt_get_startoff(ep) + xfs_bmbt_get_blockcount(ep);
	return 0;
}

/*
 * 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 */
{
4053
	xfs_bmbt_rec_host_t *ep;	/* ptr to fork's extent */
L
Linus Torvalds 已提交
4054 4055 4056 4057 4058
	xfs_ifork_t	*ifp;		/* inode fork pointer */
	int		rval;		/* return value */
	xfs_bmbt_irec_t	s;		/* internal version of extent */

#ifndef DEBUG
4059 4060 4061 4062 4063
	if (whichfork == XFS_DATA_FORK) {
		return ((ip->i_d.di_mode & S_IFMT) == S_IFREG) ?
			(ip->i_size == ip->i_mount->m_sb.sb_blocksize) :
			(ip->i_d.di_size == ip->i_mount->m_sb.sb_blocksize);
	}
L
Linus Torvalds 已提交
4064 4065 4066 4067 4068 4069 4070
#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);
4071
	ep = xfs_iext_get_ext(ifp, 0);
L
Linus Torvalds 已提交
4072 4073 4074
	xfs_bmbt_get_all(ep, &s);
	rval = s.br_startoff == 0 && s.br_blockcount == 1;
	if (rval && whichfork == XFS_DATA_FORK)
4075
		ASSERT(ip->i_size == ip->i_mount->m_sb.sb_blocksize);
L
Linus Torvalds 已提交
4076 4077 4078
	return rval;
}

4079 4080 4081 4082 4083 4084 4085 4086 4087 4088 4089 4090 4091 4092 4093 4094
STATIC int
xfs_bmap_sanity_check(
	struct xfs_mount	*mp,
	struct xfs_buf		*bp,
	int			level)
{
	struct xfs_btree_block  *block = XFS_BUF_TO_BLOCK(bp);

	if (be32_to_cpu(block->bb_magic) != XFS_BMAP_MAGIC ||
	    be16_to_cpu(block->bb_level) != level ||
	    be16_to_cpu(block->bb_numrecs) == 0 ||
	    be16_to_cpu(block->bb_numrecs) > mp->m_bmap_dmxr[level != 0])
		return 0;
	return 1;
}

L
Linus Torvalds 已提交
4095 4096 4097 4098 4099 4100 4101 4102 4103 4104 4105 4106
/*
 * 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 */
{
4107
	struct xfs_btree_block	*block;	/* current btree block */
L
Linus Torvalds 已提交
4108 4109 4110 4111 4112 4113 4114 4115
	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 */
4116
	__be64			*pp;	/* pointer to block address */
L
Linus Torvalds 已提交
4117 4118 4119 4120 4121 4122 4123 4124 4125 4126 4127 4128
	/* REFERENCED */
	xfs_extnum_t		room;	/* number of entries there's room for */

	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;
	/*
	 * Root level must use BMAP_BROOT_PTR_ADDR macro to get ptr out.
	 */
4129 4130
	level = be16_to_cpu(block->bb_level);
	ASSERT(level > 0);
4131
	pp = XFS_BMAP_BROOT_PTR_ADDR(mp, block, 1, ifp->if_broot_bytes);
4132 4133 4134 4135
	bno = be64_to_cpu(*pp);
	ASSERT(bno != NULLDFSBNO);
	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
4136 4137 4138 4139 4140 4141 4142 4143
	/*
	 * Go down the tree until leaf level is reached, following the first
	 * pointer (leftmost) at each level.
	 */
	while (level-- > 0) {
		if ((error = xfs_btree_read_bufl(mp, tp, bno, 0, &bp,
				XFS_BMAP_BTREE_REF)))
			return error;
4144
		block = XFS_BUF_TO_BLOCK(bp);
L
Linus Torvalds 已提交
4145
		XFS_WANT_CORRUPTED_GOTO(
4146
			xfs_bmap_sanity_check(mp, bp, level),
L
Linus Torvalds 已提交
4147 4148 4149
			error0);
		if (level == 0)
			break;
4150
		pp = XFS_BMBT_PTR_ADDR(mp, block, 1, mp->m_bmap_dmxr[1]);
4151 4152
		bno = be64_to_cpu(*pp);
		XFS_WANT_CORRUPTED_GOTO(XFS_FSB_SANITY_CHECK(mp, bno), error0);
L
Linus Torvalds 已提交
4153 4154 4155 4156 4157
		xfs_trans_brelse(tp, bp);
	}
	/*
	 * Here with bp and block set to the leftmost leaf node in the tree.
	 */
4158
	room = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
L
Linus Torvalds 已提交
4159 4160
	i = 0;
	/*
4161
	 * Loop over all leaf nodes.  Copy information to the extent records.
L
Linus Torvalds 已提交
4162 4163
	 */
	for (;;) {
4164
		xfs_bmbt_rec_t	*frp;
L
Linus Torvalds 已提交
4165 4166
		xfs_fsblock_t	nextbno;
		xfs_extnum_t	num_recs;
4167
		xfs_extnum_t	start;
L
Linus Torvalds 已提交
4168 4169


4170
		num_recs = xfs_btree_get_numrecs(block);
L
Linus Torvalds 已提交
4171 4172
		if (unlikely(i + num_recs > room)) {
			ASSERT(i + num_recs <= room);
4173 4174
			xfs_fs_repair_cmn_err(CE_WARN, ip->i_mount,
				"corrupt dinode %Lu, (btree extents).",
L
Linus Torvalds 已提交
4175 4176 4177 4178 4179 4180 4181
				(unsigned long long) ip->i_ino);
			XFS_ERROR_REPORT("xfs_bmap_read_extents(1)",
					 XFS_ERRLEVEL_LOW,
					ip->i_mount);
			goto error0;
		}
		XFS_WANT_CORRUPTED_GOTO(
4182
			xfs_bmap_sanity_check(mp, bp, 0),
L
Linus Torvalds 已提交
4183 4184 4185 4186
			error0);
		/*
		 * Read-ahead the next leaf block, if any.
		 */
4187
		nextbno = be64_to_cpu(block->bb_u.l.bb_rightsib);
L
Linus Torvalds 已提交
4188 4189 4190
		if (nextbno != NULLFSBLOCK)
			xfs_btree_reada_bufl(mp, nextbno, 1);
		/*
4191
		 * Copy records into the extent records.
L
Linus Torvalds 已提交
4192
		 */
4193
		frp = XFS_BMBT_REC_ADDR(mp, block, 1);
4194 4195
		start = i;
		for (j = 0; j < num_recs; j++, i++, frp++) {
4196
			xfs_bmbt_rec_host_t *trp = xfs_iext_get_ext(ifp, i);
4197 4198
			trp->l0 = be64_to_cpu(frp->l0);
			trp->l1 = be64_to_cpu(frp->l1);
L
Linus Torvalds 已提交
4199 4200 4201 4202 4203 4204 4205
		}
		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.
			 */
4206 4207
			if (unlikely(xfs_check_nostate_extents(ifp,
					start, num_recs))) {
L
Linus Torvalds 已提交
4208 4209 4210 4211 4212 4213 4214 4215 4216 4217 4218 4219 4220 4221 4222 4223
				XFS_ERROR_REPORT("xfs_bmap_read_extents(2)",
						 XFS_ERRLEVEL_LOW,
						 ip->i_mount);
				goto error0;
			}
		}
		xfs_trans_brelse(tp, bp);
		bno = nextbno;
		/*
		 * If we've reached the end, stop.
		 */
		if (bno == NULLFSBLOCK)
			break;
		if ((error = xfs_btree_read_bufl(mp, tp, bno, 0, &bp,
				XFS_BMAP_BTREE_REF)))
			return error;
4224
		block = XFS_BUF_TO_BLOCK(bp);
L
Linus Torvalds 已提交
4225
	}
4226
	ASSERT(i == (ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t)));
L
Linus Torvalds 已提交
4227
	ASSERT(i == XFS_IFORK_NEXTENTS(ip, whichfork));
4228
	XFS_BMAP_TRACE_EXLIST(ip, i, whichfork);
L
Linus Torvalds 已提交
4229 4230 4231 4232 4233 4234
	return 0;
error0:
	xfs_trans_brelse(tp, bp);
	return XFS_ERROR(EFSCORRUPTED);
}

C
Christoph Hellwig 已提交
4235
#ifdef DEBUG
L
Linus Torvalds 已提交
4236
/*
4237
 * Add bmap trace insert entries for all the contents of the extent records.
L
Linus Torvalds 已提交
4238 4239 4240 4241 4242
 */
void
xfs_bmap_trace_exlist(
	xfs_inode_t	*ip,		/* incore inode pointer */
	xfs_extnum_t	cnt,		/* count of entries in the list */
C
Christoph Hellwig 已提交
4243 4244
	int		whichfork,	/* data or attr fork */
	unsigned long	caller_ip)
L
Linus Torvalds 已提交
4245
{
4246
	xfs_extnum_t	idx;		/* extent record index */
L
Linus Torvalds 已提交
4247
	xfs_ifork_t	*ifp;		/* inode fork pointer */
C
Christoph Hellwig 已提交
4248 4249 4250 4251
	int		state = 0;

	if (whichfork == XFS_ATTR_FORK)
		state |= BMAP_ATTRFORK;
L
Linus Torvalds 已提交
4252 4253

	ifp = XFS_IFORK_PTR(ip, whichfork);
4254
	ASSERT(cnt == (ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t)));
C
Christoph Hellwig 已提交
4255 4256
	for (idx = 0; idx < cnt; idx++)
		trace_xfs_extlist(ip, idx, whichfork, caller_ip);
L
Linus Torvalds 已提交
4257 4258 4259 4260 4261 4262 4263 4264 4265 4266 4267 4268 4269 4270 4271 4272 4273 4274 4275 4276 4277 4278 4279 4280 4281 4282 4283 4284 4285 4286 4287 4288 4289 4290 4291 4292 4293 4294 4295 4296 4297 4298 4299 4300 4301 4302 4303 4304 4305 4306 4307 4308 4309 4310 4311 4312 4313 4314 4315 4316 4317 4318 4319 4320 4321 4322 4323 4324 4325 4326 4327
}

/*
 * Validate that the bmbt_irecs being returned from bmapi are valid
 * given the callers original parameters.  Specifically check the
 * ranges of the returned irecs to ensure that they only extent beyond
 * 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 */

	ASSERT(ret_nmap <= nmap);

	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);
		if ((flags & XFS_BMAPI_WRITE) && !(flags & XFS_BMAPI_DELAY))
			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);
	}
}
#endif /* DEBUG */


/*
 * Map file blocks to filesystem blocks.
 * File range is given by the bno/len pair.
 * Adds blocks to file if a write ("flags & XFS_BMAPI_WRITE" set)
 * into a hole or past eof.
 * 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					/* error */
xfs_bmapi(
	xfs_trans_t	*tp,		/* transaction pointer */
	xfs_inode_t	*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 */
	xfs_bmbt_irec_t	*mval,		/* output: map values */
	int		*nmap,		/* i/o: mval size/count */
4328
	xfs_bmap_free_t	*flist)		/* i/o: list extents to free */
L
Linus Torvalds 已提交
4329 4330 4331 4332
{
	xfs_fsblock_t	abno;		/* allocated block number */
	xfs_extlen_t	alen;		/* allocated extent length */
	xfs_fileoff_t	aoff;		/* allocated file offset */
4333
	xfs_bmalloca_t	bma = { 0 };	/* args for xfs_bmap_alloc */
L
Linus Torvalds 已提交
4334 4335
	xfs_btree_cur_t	*cur;		/* bmap btree cursor */
	xfs_fileoff_t	end;		/* end of mapped file region */
4336
	int		eof;		/* we've hit the end of extents */
4337
	xfs_bmbt_rec_host_t *ep;	/* extent record pointer */
L
Linus Torvalds 已提交
4338
	int		error;		/* error return */
4339
	xfs_bmbt_irec_t	got;		/* current file extent record */
L
Linus Torvalds 已提交
4340 4341 4342 4343 4344 4345 4346 4347
	xfs_ifork_t	*ifp;		/* inode fork pointer */
	xfs_extlen_t	indlen;		/* indirect blocks length */
	xfs_extnum_t	lastx;		/* last useful extent number */
	int		logflags;	/* flags for transaction logging */
	xfs_extlen_t	minleft;	/* min blocks left after allocation */
	xfs_extlen_t	minlen;		/* min allocation size */
	xfs_mount_t	*mp;		/* xfs mount structure */
	int		n;		/* current extent index */
M
Malcolm Parsons 已提交
4348
	int		nallocs;	/* number of extents alloc'd */
L
Linus Torvalds 已提交
4349 4350
	xfs_extnum_t	nextents;	/* number of extents in file */
	xfs_fileoff_t	obno;		/* old block number (offset) */
4351
	xfs_bmbt_irec_t	prev;		/* previous file extent record */
L
Linus Torvalds 已提交
4352
	int		tmp_logflags;	/* temp flags holder */
4353 4354
	int		whichfork;	/* data or attr fork */
	char		inhole;		/* current location is hole in file */
L
Linus Torvalds 已提交
4355 4356
	char		wasdelay;	/* old extent was delayed */
	char		wr;		/* this is a write request */
4357
	char		rt;		/* this is a realtime file */
L
Linus Torvalds 已提交
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
#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 */
	xfs_bmbt_irec_t	*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
	ASSERT(*nmap >= 1);
	ASSERT(*nmap <= XFS_BMAP_MAX_NMAP || !(flags & XFS_BMAPI_WRITE));
	whichfork = (flags & XFS_BMAPI_ATTRFORK) ?
		XFS_ATTR_FORK : XFS_DATA_FORK;
	mp = ip->i_mount;
	if (unlikely(XFS_TEST_ERROR(
	    (XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_EXTENTS &&
	     XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_BTREE &&
	     XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_LOCAL),
	     mp, XFS_ERRTAG_BMAPIFORMAT, XFS_RANDOM_BMAPIFORMAT))) {
		XFS_ERROR_REPORT("xfs_bmapi", XFS_ERRLEVEL_LOW, mp);
		return XFS_ERROR(EFSCORRUPTED);
	}
	if (XFS_FORCED_SHUTDOWN(mp))
		return XFS_ERROR(EIO);
4386
	rt = (whichfork == XFS_DATA_FORK) && XFS_IS_REALTIME_INODE(ip);
L
Linus Torvalds 已提交
4387 4388 4389 4390 4391 4392 4393 4394
	ifp = XFS_IFORK_PTR(ip, whichfork);
	ASSERT(ifp->if_ext_max ==
	       XFS_IFORK_SIZE(ip, whichfork) / (uint)sizeof(xfs_bmbt_rec_t));
	if ((wr = (flags & XFS_BMAPI_WRITE)) != 0)
		XFS_STATS_INC(xs_blk_mapw);
	else
		XFS_STATS_INC(xs_blk_mapr);
	/*
4395
	 * IGSTATE flag is used to combine extents which
L
Linus Torvalds 已提交
4396 4397 4398 4399 4400 4401 4402 4403 4404 4405 4406 4407 4408 4409 4410
	 * differ only due to the state of the extents.
	 * This technique is used from xfs_getbmap()
	 * when the caller does not wish to see the
	 * separation (which is the default).
	 *
	 * This technique is also used when writing a
	 * buffer which has been partially written,
	 * (usually by being flushed during a chunkread),
	 * to ensure one write takes place. This also
	 * prevents a change in the xfs inode extents at
	 * this time, intentionally. This change occurs
	 * on completion of the write operation, in
	 * xfs_strat_comp(), where the xfs_bmapi() call
	 * is transactioned, and the extents combined.
	 */
4411 4412 4413
	if ((flags & XFS_BMAPI_IGSTATE) && wr)	/* if writing unwritten space */
		wr = 0;				/* no allocations are allowed */
	ASSERT(wr || !(flags & XFS_BMAPI_DELAY));
L
Linus Torvalds 已提交
4414 4415 4416 4417 4418 4419 4420 4421 4422 4423 4424
	logflags = 0;
	nallocs = 0;
	cur = NULL;
	if (XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_LOCAL) {
		ASSERT(wr && tp);
		if ((error = xfs_bmap_local_to_extents(tp, ip,
				firstblock, total, &logflags, whichfork)))
			goto error0;
	}
	if (wr && *firstblock == NULLFSBLOCK) {
		if (XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_BTREE)
4425
			minleft = be16_to_cpu(ifp->if_broot->bb_level) + 1;
L
Linus Torvalds 已提交
4426 4427 4428 4429 4430 4431 4432 4433 4434 4435 4436 4437 4438 4439
		else
			minleft = 1;
	} else
		minleft = 0;
	if (!(ifp->if_flags & XFS_IFEXTENTS) &&
	    (error = xfs_iread_extents(tp, ip, whichfork)))
		goto error0;
	ep = xfs_bmap_search_extents(ip, bno, whichfork, &eof, &lastx, &got,
		&prev);
	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
	n = 0;
	end = bno + len;
	obno = bno;
	bma.ip = NULL;
4440

L
Linus Torvalds 已提交
4441 4442 4443 4444 4445 4446 4447 4448
	while (bno < end && n < *nmap) {
		/*
		 * Reading past eof, act as though there's a hole
		 * up to end.
		 */
		if (eof && !wr)
			got.br_startoff = end;
		inhole = eof || got.br_startoff > bno;
4449
		wasdelay = wr && !inhole && !(flags & XFS_BMAPI_DELAY) &&
4450
			isnullstartblock(got.br_startblock);
L
Linus Torvalds 已提交
4451 4452 4453 4454 4455 4456 4457 4458 4459 4460
		/*
		 * First, deal with the hole before the allocated space
		 * that we found, if any.
		 */
		if (wr && (inhole || wasdelay)) {
			/*
			 * For the wasdelay case, we could also just
			 * allocate the stuff asked for in this bmap call
			 * but that wouldn't be as good.
			 */
4461
			if (wasdelay) {
L
Linus Torvalds 已提交
4462 4463 4464
				alen = (xfs_extlen_t)got.br_blockcount;
				aoff = got.br_startoff;
				if (lastx != NULLEXTNUM && lastx) {
4465
					ep = xfs_iext_get_ext(ifp, lastx - 1);
L
Linus Torvalds 已提交
4466 4467 4468 4469 4470 4471 4472 4473 4474 4475 4476
					xfs_bmbt_get_all(ep, &prev);
				}
			} else {
				alen = (xfs_extlen_t)
					XFS_FILBLKS_MIN(len, MAXEXTLEN);
				if (!eof)
					alen = (xfs_extlen_t)
						XFS_FILBLKS_MIN(alen,
							got.br_startoff - bno);
				aoff = bno;
			}
4477 4478
			minlen = (flags & XFS_BMAPI_CONTIG) ? alen : 1;
			if (flags & XFS_BMAPI_DELAY) {
4479
				xfs_extlen_t	extsz;
4480 4481

				/* Figure out the extent size, adjust alen */
4482
				extsz = xfs_get_extsz_hint(ip);
4483 4484 4485
				if (extsz) {
					error = xfs_bmap_extsize_align(mp,
							&got, &prev, extsz,
4486 4487 4488
							rt, eof,
							flags&XFS_BMAPI_DELAY,
							flags&XFS_BMAPI_CONVERT,
4489 4490
							&aoff, &alen);
					ASSERT(!error);
4491 4492
				}

4493 4494 4495
				if (rt)
					extsz = alen / mp->m_sb.sb_rextsize;

L
Linus Torvalds 已提交
4496 4497 4498
				/*
				 * Make a transaction-less quota reservation for
				 * delayed allocation blocks. This number gets
4499 4500
				 * adjusted later.  We return if we haven't
				 * allocated blocks already inside this loop.
L
Linus Torvalds 已提交
4501
				 */
C
Christoph Hellwig 已提交
4502 4503
				error = xfs_trans_reserve_quota_nblks(
						NULL, ip, (long)alen, 0,
4504
						rt ? XFS_QMOPT_RES_RTBLKS :
C
Christoph Hellwig 已提交
4505 4506
						     XFS_QMOPT_RES_REGBLKS);
				if (error) {
L
Linus Torvalds 已提交
4507 4508 4509
					if (n == 0) {
						*nmap = 0;
						ASSERT(cur == NULL);
4510
						return error;
L
Linus Torvalds 已提交
4511 4512 4513 4514 4515 4516 4517 4518
					}
					break;
				}

				/*
				 * Split changing sb for alen and indlen since
				 * they could be coming from different places.
				 */
4519 4520 4521
				indlen = (xfs_extlen_t)
					xfs_bmap_worst_indlen(ip, alen);
				ASSERT(indlen > 0);
L
Linus Torvalds 已提交
4522

4523
				if (rt) {
4524 4525
					error = xfs_mod_incore_sb(mp,
							XFS_SBS_FREXTENTS,
4526
							-((int64_t)extsz), (flags &
4527
							XFS_BMAPI_RSVBLOCKS));
4528
				} else {
4529 4530
					error = xfs_mod_incore_sb(mp,
							XFS_SBS_FDBLOCKS,
4531
							-((int64_t)alen), (flags &
4532
							XFS_BMAPI_RSVBLOCKS));
4533
				}
4534
				if (!error) {
4535 4536
					error = xfs_mod_incore_sb(mp,
							XFS_SBS_FDBLOCKS,
4537
							-((int64_t)indlen), (flags &
4538
							XFS_BMAPI_RSVBLOCKS));
4539 4540
					if (error && rt)
						xfs_mod_incore_sb(mp,
4541
							XFS_SBS_FREXTENTS,
4542
							(int64_t)extsz, (flags &
4543
							XFS_BMAPI_RSVBLOCKS));
4544 4545
					else if (error)
						xfs_mod_incore_sb(mp,
4546
							XFS_SBS_FDBLOCKS,
4547
							(int64_t)alen, (flags &
4548
							XFS_BMAPI_RSVBLOCKS));
4549
				}
4550 4551

				if (error) {
4552
					if (XFS_IS_QUOTA_ON(mp))
4553
						/* unreserve the blocks now */
4554
						(void)
C
Christoph Hellwig 已提交
4555 4556
						xfs_trans_unreserve_quota_nblks(
							NULL, ip,
4557 4558 4559
							(long)alen, 0, rt ?
							XFS_QMOPT_RES_RTBLKS :
							XFS_QMOPT_RES_REGBLKS);
L
Linus Torvalds 已提交
4560 4561
					break;
				}
4562

L
Linus Torvalds 已提交
4563
				ip->i_delayed_blks += alen;
4564
				abno = nullstartblock(indlen);
L
Linus Torvalds 已提交
4565 4566 4567 4568 4569 4570 4571 4572 4573 4574 4575 4576 4577 4578 4579 4580
			} else {
				/*
				 * If first time, allocate and fill in
				 * once-only bma fields.
				 */
				if (bma.ip == NULL) {
					bma.tp = tp;
					bma.ip = ip;
					bma.prevp = &prev;
					bma.gotp = &got;
					bma.total = total;
					bma.userdata = 0;
				}
				/* Indicate if this is the first user data
				 * in the file, or just any user data.
				 */
4581
				if (!(flags & XFS_BMAPI_METADATA)) {
L
Linus Torvalds 已提交
4582 4583 4584 4585 4586 4587 4588 4589 4590 4591 4592
					bma.userdata = (aoff == 0) ?
						XFS_ALLOC_INITIAL_USER_DATA :
						XFS_ALLOC_USERDATA;
				}
				/*
				 * Fill in changeable bma fields.
				 */
				bma.eof = eof;
				bma.firstblock = *firstblock;
				bma.alen = alen;
				bma.off = aoff;
4593
				bma.conv = !!(flags & XFS_BMAPI_CONVERT);
L
Linus Torvalds 已提交
4594 4595 4596 4597 4598 4599 4600 4601 4602 4603
				bma.wasdel = wasdelay;
				bma.minlen = minlen;
				bma.low = flist->xbf_low;
				bma.minleft = minleft;
				/*
				 * 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 && alen >= mp->m_dalign &&
4604 4605
				    (!(flags & XFS_BMAPI_METADATA)) &&
				    (whichfork == XFS_DATA_FORK)) {
L
Linus Torvalds 已提交
4606 4607 4608 4609 4610 4611 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
					if ((error = xfs_bmap_isaeof(ip, aoff,
							whichfork, &bma.aeof)))
						goto error0;
				} else
					bma.aeof = 0;
				/*
				 * Call allocator.
				 */
				if ((error = xfs_bmap_alloc(&bma)))
					goto error0;
				/*
				 * Copy out result fields.
				 */
				abno = bma.rval;
				if ((flist->xbf_low = bma.low))
					minleft = 0;
				alen = bma.alen;
				aoff = bma.off;
				ASSERT(*firstblock == NULLFSBLOCK ||
				       XFS_FSB_TO_AGNO(mp, *firstblock) ==
				       XFS_FSB_TO_AGNO(mp, bma.firstblock) ||
				       (flist->xbf_low &&
					XFS_FSB_TO_AGNO(mp, *firstblock) <
					XFS_FSB_TO_AGNO(mp, bma.firstblock)));
				*firstblock = bma.firstblock;
				if (cur)
					cur->bc_private.b.firstblock =
						*firstblock;
				if (abno == NULLFSBLOCK)
					break;
				if ((ifp->if_flags & XFS_IFBROOT) && !cur) {
4637
					cur = xfs_bmbt_init_cursor(mp, tp,
L
Linus Torvalds 已提交
4638 4639 4640 4641 4642 4643 4644 4645 4646 4647 4648 4649 4650 4651 4652 4653 4654 4655 4656 4657 4658 4659 4660
						ip, whichfork);
					cur->bc_private.b.firstblock =
						*firstblock;
					cur->bc_private.b.flist = flist;
				}
				/*
				 * Bump the number of extents we've allocated
				 * in this call.
				 */
				nallocs++;
			}
			if (cur)
				cur->bc_private.b.flags =
					wasdelay ? XFS_BTCUR_BPRV_WASDEL : 0;
			got.br_startoff = aoff;
			got.br_startblock = abno;
			got.br_blockcount = alen;
			got.br_state = XFS_EXT_NORM;	/* assume normal */
			/*
			 * Determine state of extent, and the filesystem.
			 * A wasdelay extent has been initialized, so
			 * shouldn't be flagged as unwritten.
			 */
4661
			if (wr && xfs_sb_version_hasextflgbit(&mp->m_sb)) {
L
Linus Torvalds 已提交
4662 4663 4664 4665
				if (!wasdelay && (flags & XFS_BMAPI_PREALLOC))
					got.br_state = XFS_EXT_UNWRITTEN;
			}
			error = xfs_bmap_add_extent(ip, lastx, &cur, &got,
4666
				firstblock, flist, &tmp_logflags,
4667
				whichfork, (flags & XFS_BMAPI_RSVBLOCKS));
L
Linus Torvalds 已提交
4668 4669 4670 4671
			logflags |= tmp_logflags;
			if (error)
				goto error0;
			lastx = ifp->if_lastex;
4672
			ep = xfs_iext_get_ext(ifp, lastx);
L
Linus Torvalds 已提交
4673 4674 4675 4676 4677 4678
			nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
			xfs_bmbt_get_all(ep, &got);
			ASSERT(got.br_startoff <= aoff);
			ASSERT(got.br_startoff + got.br_blockcount >=
				aoff + alen);
#ifdef DEBUG
4679
			if (flags & XFS_BMAPI_DELAY) {
4680 4681
				ASSERT(isnullstartblock(got.br_startblock));
				ASSERT(startblockval(got.br_startblock) > 0);
L
Linus Torvalds 已提交
4682 4683 4684 4685 4686 4687 4688 4689 4690 4691 4692 4693 4694 4695 4696 4697 4698 4699 4700 4701 4702 4703 4704 4705 4706 4707
			}
			ASSERT(got.br_state == XFS_EXT_NORM ||
			       got.br_state == XFS_EXT_UNWRITTEN);
#endif
			/*
			 * Fall down into the found allocated space case.
			 */
		} else if (inhole) {
			/*
			 * 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;
		}
		/*
		 * Then deal with the allocated space we found.
		 */
		ASSERT(ep != NULL);
4708 4709
		if (!(flags & XFS_BMAPI_ENTIRE) &&
		    (got.br_startoff + got.br_blockcount > obno)) {
L
Linus Torvalds 已提交
4710 4711 4712 4713 4714
			if (obno > bno)
				bno = obno;
			ASSERT((bno >= obno) || (n == 0));
			ASSERT(bno < end);
			mval->br_startoff = bno;
4715
			if (isnullstartblock(got.br_startblock)) {
4716
				ASSERT(!wr || (flags & XFS_BMAPI_DELAY));
L
Linus Torvalds 已提交
4717 4718 4719 4720 4721 4722 4723 4724 4725 4726 4727 4728 4729 4730 4731 4732 4733 4734 4735 4736
				mval->br_startblock = DELAYSTARTBLOCK;
			} else
				mval->br_startblock =
					got.br_startblock +
					(bno - got.br_startoff);
			/*
			 * 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.
			 */
			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);
		} else {
			*mval = got;
4737
			if (isnullstartblock(mval->br_startblock)) {
4738
				ASSERT(!wr || (flags & XFS_BMAPI_DELAY));
L
Linus Torvalds 已提交
4739 4740 4741 4742 4743 4744 4745 4746 4747 4748 4749 4750 4751 4752 4753
				mval->br_startblock = DELAYSTARTBLOCK;
			}
		}

		/*
		 * Check if writing previously allocated but
		 * unwritten extents.
		 */
		if (wr && mval->br_state == XFS_EXT_UNWRITTEN &&
		    ((flags & (XFS_BMAPI_PREALLOC|XFS_BMAPI_DELAY)) == 0)) {
			/*
			 * Modify (by adding) the state flag, if writing.
			 */
			ASSERT(mval->br_blockcount <= len);
			if ((ifp->if_flags & XFS_IFBROOT) && !cur) {
4754 4755
				cur = xfs_bmbt_init_cursor(mp,
					tp, ip, whichfork);
L
Linus Torvalds 已提交
4756 4757 4758 4759 4760 4761
				cur->bc_private.b.firstblock =
					*firstblock;
				cur->bc_private.b.flist = flist;
			}
			mval->br_state = XFS_EXT_NORM;
			error = xfs_bmap_add_extent(ip, lastx, &cur, mval,
4762
				firstblock, flist, &tmp_logflags,
4763
				whichfork, (flags & XFS_BMAPI_RSVBLOCKS));
L
Linus Torvalds 已提交
4764 4765 4766 4767
			logflags |= tmp_logflags;
			if (error)
				goto error0;
			lastx = ifp->if_lastex;
4768
			ep = xfs_iext_get_ext(ifp, lastx);
L
Linus Torvalds 已提交
4769 4770 4771 4772 4773 4774 4775 4776 4777 4778 4779
			nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
			xfs_bmbt_get_all(ep, &got);
			/*
			 * We may have combined previously unwritten
			 * space with written space, so generate
			 * another request.
			 */
			if (mval->br_blockcount < len)
				continue;
		}

4780
		ASSERT((flags & XFS_BMAPI_ENTIRE) ||
L
Linus Torvalds 已提交
4781
		       ((mval->br_startoff + mval->br_blockcount) <= end));
4782 4783
		ASSERT((flags & XFS_BMAPI_ENTIRE) ||
		       (mval->br_blockcount <= len) ||
L
Linus Torvalds 已提交
4784 4785 4786 4787 4788 4789 4790 4791 4792 4793 4794 4795 4796 4797
		       (mval->br_startoff < obno));
		bno = mval->br_startoff + mval->br_blockcount;
		len = end - bno;
		if (n > 0 && mval->br_startoff == mval[-1].br_startoff) {
			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 &&
4798 4799
			   ((flags & XFS_BMAPI_IGSTATE) ||
				mval[-1].br_state == mval->br_state)) {
L
Linus Torvalds 已提交
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
			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++;
		}
		/*
		 * 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 || nallocs >= *nmap)
			break;
		/*
		 * Else go on to the next record.
		 */
4826
		ep = xfs_iext_get_ext(ifp, ++lastx);
4827 4828
		prev = got;
		if (lastx >= nextents)
L
Linus Torvalds 已提交
4829
			eof = 1;
4830
		else
L
Linus Torvalds 已提交
4831 4832 4833 4834 4835 4836 4837 4838 4839 4840 4841 4842 4843 4844 4845 4846 4847 4848 4849 4850 4851 4852 4853 4854
			xfs_bmbt_get_all(ep, &got);
	}
	ifp->if_lastex = lastx;
	*nmap = n;
	/*
	 * Transform from btree to extents, give it cur.
	 */
	if (tp && XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_BTREE &&
	    XFS_IFORK_NEXTENTS(ip, whichfork) <= ifp->if_ext_max) {
		ASSERT(wr && cur);
		error = xfs_bmap_btree_to_extents(tp, ip, cur,
			&tmp_logflags, whichfork);
		logflags |= tmp_logflags;
		if (error)
			goto error0;
	}
	ASSERT(ifp->if_ext_max ==
	       XFS_IFORK_SIZE(ip, whichfork) / (uint)sizeof(xfs_bmbt_rec_t));
	ASSERT(XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_BTREE ||
	       XFS_IFORK_NEXTENTS(ip, whichfork) > ifp->if_ext_max);
	error = 0;
error0:
	/*
	 * Log everything.  Do this after conversion, there's no point in
4855
	 * logging the extent records if we've converted to btree format.
L
Linus Torvalds 已提交
4856
	 */
4857
	if ((logflags & xfs_ilog_fext(whichfork)) &&
L
Linus Torvalds 已提交
4858
	    XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_EXTENTS)
4859 4860
		logflags &= ~xfs_ilog_fext(whichfork);
	else if ((logflags & xfs_ilog_fbroot(whichfork)) &&
L
Linus Torvalds 已提交
4861
		 XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_BTREE)
4862
		logflags &= ~xfs_ilog_fbroot(whichfork);
L
Linus Torvalds 已提交
4863 4864 4865 4866 4867 4868 4869 4870 4871 4872 4873 4874 4875 4876 4877 4878 4879 4880 4881 4882 4883 4884 4885 4886 4887 4888 4889 4890 4891 4892 4893 4894 4895 4896 4897 4898 4899 4900 4901 4902 4903 4904 4905 4906 4907
	/*
	 * 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 (logflags) {
		ASSERT(tp && wr);
		xfs_trans_log_inode(tp, ip, logflags);
	}
	if (cur) {
		if (!error) {
			ASSERT(*firstblock == NULLFSBLOCK ||
			       XFS_FSB_TO_AGNO(mp, *firstblock) ==
			       XFS_FSB_TO_AGNO(mp,
				       cur->bc_private.b.firstblock) ||
			       (flist->xbf_low &&
				XFS_FSB_TO_AGNO(mp, *firstblock) <
				XFS_FSB_TO_AGNO(mp,
					cur->bc_private.b.firstblock)));
			*firstblock = cur->bc_private.b.firstblock;
		}
		xfs_btree_del_cursor(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;
}

/*
 * Map file blocks to filesystem blocks, simple version.
 * One block (extent) only, read-only.
 * For flags, only the XFS_BMAPI_ATTRFORK flag is examined.
 * For the other flag values, the effect is as if XFS_BMAPI_METADATA
 * was set and all the others were clear.
 */
int						/* error */
xfs_bmapi_single(
	xfs_trans_t	*tp,		/* transaction pointer */
	xfs_inode_t	*ip,		/* incore inode */
	int		whichfork,	/* data or attr fork */
	xfs_fsblock_t	*fsb,		/* output: mapped block */
	xfs_fileoff_t	bno)		/* starting file offs. mapped */
{
4908
	int		eof;		/* we've hit the end of extents */
L
Linus Torvalds 已提交
4909
	int		error;		/* error return */
4910
	xfs_bmbt_irec_t	got;		/* current file extent record */
L
Linus Torvalds 已提交
4911 4912
	xfs_ifork_t	*ifp;		/* inode fork pointer */
	xfs_extnum_t	lastx;		/* last useful extent number */
4913
	xfs_bmbt_irec_t	prev;		/* previous file extent record */
L
Linus Torvalds 已提交
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

	ifp = XFS_IFORK_PTR(ip, whichfork);
	if (unlikely(
	    XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_BTREE &&
	    XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_EXTENTS)) {
	       XFS_ERROR_REPORT("xfs_bmapi_single", XFS_ERRLEVEL_LOW,
				ip->i_mount);
	       return XFS_ERROR(EFSCORRUPTED);
	}
	if (XFS_FORCED_SHUTDOWN(ip->i_mount))
		return XFS_ERROR(EIO);
	XFS_STATS_INC(xs_blk_mapr);
	if (!(ifp->if_flags & XFS_IFEXTENTS) &&
	    (error = xfs_iread_extents(tp, ip, whichfork)))
		return error;
	(void)xfs_bmap_search_extents(ip, bno, whichfork, &eof, &lastx, &got,
		&prev);
	/*
	 * Reading past eof, act as though there's a hole
	 * up to end.
	 */
	if (eof || got.br_startoff > bno) {
		*fsb = NULLFSBLOCK;
		return 0;
	}
4939
	ASSERT(!isnullstartblock(got.br_startblock));
L
Linus Torvalds 已提交
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
	ASSERT(bno < got.br_startoff + got.br_blockcount);
	*fsb = got.br_startblock + (bno - got.br_startoff);
	ifp->if_lastex = lastx;
	return 0;
}

/*
 * 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.
 */
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 */
	xfs_bmap_free_t		*flist,		/* i/o: list extents to free */
	int			*done)		/* set if not done yet */
{
	xfs_btree_cur_t		*cur;		/* bmap btree cursor */
	xfs_bmbt_irec_t		del;		/* extent being deleted */
	int			eof;		/* is deleting at eof */
4968
	xfs_bmbt_rec_host_t	*ep;		/* extent record pointer */
L
Linus Torvalds 已提交
4969 4970
	int			error;		/* error return value */
	xfs_extnum_t		extno;		/* extent number in list */
4971
	xfs_bmbt_irec_t		got;		/* current extent record */
L
Linus Torvalds 已提交
4972 4973 4974 4975 4976 4977
	xfs_ifork_t		*ifp;		/* inode fork pointer */
	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 */
4978 4979
	xfs_extnum_t		nextents;	/* number of file extents */
	xfs_bmbt_irec_t		prev;		/* previous extent record */
L
Linus Torvalds 已提交
4980 4981 4982 4983 4984 4985 4986
	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 */
	int			rsvd;		/* OK to allocate reserved blocks */
	xfs_fsblock_t		sum;

C
Christoph Hellwig 已提交
4987 4988
	trace_xfs_bunmap(ip, bno, len, flags, _RET_IP_);

L
Linus Torvalds 已提交
4989 4990 4991 4992 4993 4994 4995 4996 4997 4998 4999 5000 5001 5002 5003 5004 5005 5006 5007 5008 5009 5010 5011 5012 5013 5014 5015
	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);
		return XFS_ERROR(EFSCORRUPTED);
	}
	mp = ip->i_mount;
	if (XFS_FORCED_SHUTDOWN(mp))
		return XFS_ERROR(EIO);
	rsvd = (flags & XFS_BMAPI_RSVBLOCKS) != 0;
	ASSERT(len > 0);
	ASSERT(nexts >= 0);
	ASSERT(ifp->if_ext_max ==
	       XFS_IFORK_SIZE(ip, whichfork) / (uint)sizeof(xfs_bmbt_rec_t));
	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;
	}
	XFS_STATS_INC(xs_blk_unmap);
5016
	isrt = (whichfork == XFS_DATA_FORK) && XFS_IS_REALTIME_INODE(ip);
L
Linus Torvalds 已提交
5017 5018 5019 5020
	start = bno;
	bno = start + len - 1;
	ep = xfs_bmap_search_extents(ip, bno, whichfork, &eof, &lastx, &got,
		&prev);
5021

L
Linus Torvalds 已提交
5022 5023 5024 5025 5026
	/*
	 * 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) {
5027
		ep = xfs_iext_get_ext(ifp, --lastx);
L
Linus Torvalds 已提交
5028 5029 5030 5031 5032 5033
		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);
5034
		cur = xfs_bmbt_init_cursor(mp, tp, ip, whichfork);
L
Linus Torvalds 已提交
5035 5036 5037 5038 5039 5040 5041 5042 5043 5044 5045 5046 5047 5048 5049
		cur->bc_private.b.firstblock = *firstblock;
		cur->bc_private.b.flist = flist;
		cur->bc_private.b.flags = 0;
	} else
		cur = NULL;
	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;
5050
			ep = xfs_iext_get_ext(ifp, lastx);
L
Linus Torvalds 已提交
5051 5052 5053 5054 5055 5056 5057 5058 5059 5060 5061 5062 5063 5064 5065 5066
			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;
5067
		wasdel = isnullstartblock(del.br_startblock);
L
Linus Torvalds 已提交
5068 5069 5070 5071 5072 5073 5074 5075 5076 5077 5078 5079 5080 5081 5082 5083 5084 5085 5086
		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))) {
			/*
			 * 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.
			 */
			if (del.br_state == XFS_EXT_UNWRITTEN ||
5087
			    !xfs_sb_version_hasextflgbit(&mp->m_sb)) {
L
Linus Torvalds 已提交
5088 5089 5090 5091 5092 5093 5094 5095 5096
				/*
				 * 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)
5097 5098
						xfs_bmbt_get_all(xfs_iext_get_ext(
							ifp, lastx), &got);
L
Linus Torvalds 已提交
5099 5100 5101 5102 5103 5104 5105 5106 5107 5108 5109 5110 5111 5112 5113 5114 5115 5116 5117 5118
				}
				continue;
			}
			/*
			 * It's written, turn it unwritten.
			 * This is better than zeroing it.
			 */
			ASSERT(del.br_state == XFS_EXT_NORM);
			ASSERT(xfs_trans_get_block_res(tp) > 0);
			/*
			 * 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(ip, lastx, &cur, &del,
5119
				firstblock, flist, &logflags,
5120
				XFS_DATA_FORK, 0);
L
Linus Torvalds 已提交
5121 5122 5123 5124 5125 5126 5127 5128 5129 5130 5131 5132 5133 5134 5135 5136 5137 5138
			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 ||
				     xfs_trans_get_block_res(tp) == 0)) ||
5139
				   !xfs_sb_version_hasextflgbit(&mp->m_sb)) {
L
Linus Torvalds 已提交
5140 5141 5142 5143 5144 5145 5146 5147 5148 5149 5150 5151 5152 5153 5154 5155 5156 5157 5158
				/*
				 * 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 (bno < got.br_startoff) {
					if (--lastx >= 0)
						xfs_bmbt_get_all(--ep, &got);
				}
				continue;
			} 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);
5159 5160
				xfs_bmbt_get_all(xfs_iext_get_ext(ifp,
						lastx - 1), &prev);
L
Linus Torvalds 已提交
5161
				ASSERT(prev.br_state == XFS_EXT_NORM);
5162
				ASSERT(!isnullstartblock(prev.br_startblock));
L
Linus Torvalds 已提交
5163 5164 5165 5166 5167 5168 5169 5170 5171 5172 5173
				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;
				error = xfs_bmap_add_extent(ip, lastx - 1, &cur,
					&prev, firstblock, flist, &logflags,
5174
					XFS_DATA_FORK, 0);
L
Linus Torvalds 已提交
5175 5176 5177 5178 5179 5180 5181 5182
				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(ip, lastx, &cur,
					&del, firstblock, flist, &logflags,
5183
					XFS_DATA_FORK, 0);
L
Linus Torvalds 已提交
5184 5185 5186 5187 5188 5189
				if (error)
					goto error0;
				goto nodelete;
			}
		}
		if (wasdel) {
5190
			ASSERT(startblockval(del.br_startblock) > 0);
5191
			/* Update realtime/data freespace, unreserve quota */
5192 5193 5194 5195 5196 5197
			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_incore_sb(mp, XFS_SBS_FREXTENTS,
5198
						(int64_t)rtexts, rsvd);
C
Christoph Hellwig 已提交
5199 5200
				(void)xfs_trans_reserve_quota_nblks(NULL,
					ip, -((long)del.br_blockcount), 0,
5201 5202 5203
					XFS_QMOPT_RES_RTBLKS);
			} else {
				xfs_mod_incore_sb(mp, XFS_SBS_FDBLOCKS,
5204
						(int64_t)del.br_blockcount, rsvd);
C
Christoph Hellwig 已提交
5205 5206
				(void)xfs_trans_reserve_quota_nblks(NULL,
					ip, -((long)del.br_blockcount), 0,
L
Linus Torvalds 已提交
5207
					XFS_QMOPT_RES_REGBLKS);
5208
			}
L
Linus Torvalds 已提交
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
			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;
		/*
		 * 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.
		 */
		if (!wasdel && xfs_trans_get_block_res(tp) == 0 &&
		    XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_EXTENTS &&
		    XFS_IFORK_NEXTENTS(ip, whichfork) >= ifp->if_ext_max &&
		    del.br_startoff > got.br_startoff &&
		    del.br_startoff + del.br_blockcount <
		    got.br_startoff + got.br_blockcount) {
			error = XFS_ERROR(ENOSPC);
			goto error0;
		}
		error = xfs_bmap_del_extent(ip, tp, lastx, flist, cur, &del,
5236
				&tmp_logflags, whichfork, rsvd);
L
Linus Torvalds 已提交
5237 5238 5239 5240 5241 5242 5243 5244 5245 5246
		logflags |= tmp_logflags;
		if (error)
			goto error0;
		bno = del.br_startoff - 1;
nodelete:
		lastx = ifp->if_lastex;
		/*
		 * If not done go on to the next (previous) record.
		 * Reset ep in case the extents array was re-alloced.
		 */
5247
		ep = xfs_iext_get_ext(ifp, lastx);
L
Linus Torvalds 已提交
5248 5249 5250
		if (bno != (xfs_fileoff_t)-1 && bno >= start) {
			if (lastx >= XFS_IFORK_NEXTENTS(ip, whichfork) ||
			    xfs_bmbt_get_startoff(ep) > bno) {
5251 5252
				if (--lastx >= 0)
					ep = xfs_iext_get_ext(ifp, lastx);
L
Linus Torvalds 已提交
5253 5254 5255 5256 5257 5258 5259 5260 5261 5262 5263 5264 5265 5266 5267 5268 5269 5270 5271 5272 5273 5274 5275 5276 5277 5278 5279 5280 5281 5282 5283 5284 5285 5286 5287 5288 5289 5290 5291 5292 5293 5294 5295
			}
			if (lastx >= 0)
				xfs_bmbt_get_all(ep, &got);
			extno++;
		}
	}
	ifp->if_lastex = lastx;
	*done = bno == (xfs_fileoff_t)-1 || bno < start || lastx < 0;
	ASSERT(ifp->if_ext_max ==
	       XFS_IFORK_SIZE(ip, whichfork) / (uint)sizeof(xfs_bmbt_rec_t));
	/*
	 * Convert to a btree if necessary.
	 */
	if (XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_EXTENTS &&
	    XFS_IFORK_NEXTENTS(ip, whichfork) > ifp->if_ext_max) {
		ASSERT(cur == NULL);
		error = xfs_bmap_extents_to_btree(tp, ip, firstblock, flist,
			&cur, 0, &tmp_logflags, whichfork);
		logflags |= tmp_logflags;
		if (error)
			goto error0;
	}
	/*
	 * transform from btree to extents, give it cur
	 */
	else if (XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_BTREE &&
		 XFS_IFORK_NEXTENTS(ip, whichfork) <= ifp->if_ext_max) {
		ASSERT(cur != NULL);
		error = xfs_bmap_btree_to_extents(tp, ip, cur, &tmp_logflags,
			whichfork);
		logflags |= tmp_logflags;
		if (error)
			goto error0;
	}
	/*
	 * transform from extents to local?
	 */
	ASSERT(ifp->if_ext_max ==
	       XFS_IFORK_SIZE(ip, whichfork) / (uint)sizeof(xfs_bmbt_rec_t));
	error = 0;
error0:
	/*
	 * Log everything.  Do this after conversion, there's no point in
5296
	 * logging the extent records if we've converted to btree format.
L
Linus Torvalds 已提交
5297
	 */
5298
	if ((logflags & xfs_ilog_fext(whichfork)) &&
L
Linus Torvalds 已提交
5299
	    XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_EXTENTS)
5300 5301
		logflags &= ~xfs_ilog_fext(whichfork);
	else if ((logflags & xfs_ilog_fbroot(whichfork)) &&
L
Linus Torvalds 已提交
5302
		 XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_BTREE)
5303
		logflags &= ~xfs_ilog_fbroot(whichfork);
L
Linus Torvalds 已提交
5304 5305 5306 5307 5308 5309 5310 5311 5312 5313 5314 5315 5316 5317 5318 5319 5320
	/*
	 * 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;
}

5321 5322 5323 5324 5325 5326
/*
 * returns 1 for success, 0 if we failed to map the extent.
 */
STATIC int
xfs_getbmapx_fix_eof_hole(
	xfs_inode_t		*ip,		/* xfs incore inode pointer */
5327
	struct getbmapx		*out,		/* output structure */
5328
	int			prealloced,	/* this is a file with
5329
						 * preallocated data space */
5330 5331 5332 5333 5334
	__int64_t		end,		/* last block requested */
	xfs_fsblock_t		startblock)
{
	__int64_t		fixlen;
	xfs_mount_t		*mp;		/* file system mount point */
E
Eric Sandeen 已提交
5335 5336 5337
	xfs_ifork_t		*ifp;		/* inode fork pointer */
	xfs_extnum_t		lastx;		/* last extent pointer */
	xfs_fileoff_t		fileblock;
5338 5339 5340 5341 5342 5343 5344 5345 5346 5347 5348 5349 5350

	if (startblock == HOLESTARTBLOCK) {
		mp = ip->i_mount;
		out->bmv_block = -1;
		fixlen = XFS_FSB_TO_BB(mp, XFS_B_TO_FSB(mp, ip->i_size));
		fixlen -= out->bmv_offset;
		if (prealloced && out->bmv_offset + out->bmv_length == end) {
			/* Came to hole at EOF. Trim it. */
			if (fixlen <= 0)
				return 0;
			out->bmv_length = fixlen;
		}
	} else {
E
Eric Sandeen 已提交
5351 5352 5353
		if (startblock == DELAYSTARTBLOCK)
			out->bmv_block = -2;
		else
5354
			out->bmv_block = xfs_fsb_to_db(ip, startblock);
E
Eric Sandeen 已提交
5355 5356 5357 5358 5359
		fileblock = XFS_BB_TO_FSB(ip->i_mount, out->bmv_offset);
		ifp = XFS_IFORK_PTR(ip, XFS_DATA_FORK);
		if (xfs_iext_bno_to_ext(ifp, fileblock, &lastx) &&
		   (lastx == (ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t))-1))
			out->bmv_oflags |= BMV_OF_LAST;
5360 5361 5362 5363 5364
	}

	return 1;
}

L
Linus Torvalds 已提交
5365
/*
5366 5367 5368 5369 5370
 * Get inode's extents as described in bmv, and format for output.
 * Calls formatter to fill the user's buffer until all extents
 * are mapped, until the passed-in bmv->bmv_count slots have
 * been filled, or until the formatter short-circuits the loop,
 * if it is tracking filled-in extents on its own.
L
Linus Torvalds 已提交
5371 5372 5373
 */
int						/* error code */
xfs_getbmap(
5374
	xfs_inode_t		*ip,
5375 5376 5377
	struct getbmapx		*bmv,		/* user bmap structure */
	xfs_bmap_format_t	formatter,	/* format to user */
	void			*arg)		/* formatter arg */
L
Linus Torvalds 已提交
5378 5379
{
	__int64_t		bmvend;		/* last block requested */
5380
	int			error = 0;	/* return value */
L
Linus Torvalds 已提交
5381 5382 5383 5384 5385 5386 5387 5388 5389
	__int64_t		fixlen;		/* length for -1 case */
	int			i;		/* extent number */
	int			lock;		/* lock state */
	xfs_bmbt_irec_t		*map;		/* buffer for user's data */
	xfs_mount_t		*mp;		/* file system mount point */
	int			nex;		/* # of user extents can do */
	int			nexleft;	/* # of user extents left */
	int			subnex;		/* # of bmapi's can do */
	int			nmap;		/* number of map entries */
5390
	struct getbmapx		*out;		/* output structure */
L
Linus Torvalds 已提交
5391 5392 5393
	int			whichfork;	/* data or attr fork */
	int			prealloced;	/* this is a file with
						 * preallocated data space */
5394
	int			iflags;		/* interface flags */
L
Linus Torvalds 已提交
5395
	int			bmapi_flags;	/* flags for xfs_bmapi */
5396
	int			cur_ext = 0;
L
Linus Torvalds 已提交
5397 5398

	mp = ip->i_mount;
5399 5400
	iflags = bmv->bmv_iflags;
	whichfork = iflags & BMV_IF_ATTRFORK ? XFS_ATTR_FORK : XFS_DATA_FORK;
L
Linus Torvalds 已提交
5401 5402 5403 5404 5405 5406 5407 5408 5409 5410 5411 5412 5413 5414

	if (whichfork == XFS_ATTR_FORK) {
		if (XFS_IFORK_Q(ip)) {
			if (ip->i_d.di_aformat != XFS_DINODE_FMT_EXTENTS &&
			    ip->i_d.di_aformat != XFS_DINODE_FMT_BTREE &&
			    ip->i_d.di_aformat != XFS_DINODE_FMT_LOCAL)
				return XFS_ERROR(EINVAL);
		} else if (unlikely(
			   ip->i_d.di_aformat != 0 &&
			   ip->i_d.di_aformat != XFS_DINODE_FMT_EXTENTS)) {
			XFS_ERROR_REPORT("xfs_getbmap", XFS_ERRLEVEL_LOW,
					 ip->i_mount);
			return XFS_ERROR(EFSCORRUPTED);
		}
5415 5416 5417 5418 5419 5420 5421 5422 5423

		prealloced = 0;
		fixlen = 1LL << 32;
	} else {
		if (ip->i_d.di_format != XFS_DINODE_FMT_EXTENTS &&
		    ip->i_d.di_format != XFS_DINODE_FMT_BTREE &&
		    ip->i_d.di_format != XFS_DINODE_FMT_LOCAL)
			return XFS_ERROR(EINVAL);

5424
		if (xfs_get_extsz_hint(ip) ||
5425
		    ip->i_d.di_flags & (XFS_DIFLAG_PREALLOC|XFS_DIFLAG_APPEND)){
L
Linus Torvalds 已提交
5426 5427 5428 5429
			prealloced = 1;
			fixlen = XFS_MAXIOFFSET(mp);
		} else {
			prealloced = 0;
5430
			fixlen = ip->i_size;
L
Linus Torvalds 已提交
5431 5432 5433 5434 5435
		}
	}

	if (bmv->bmv_length == -1) {
		fixlen = XFS_FSB_TO_BB(mp, XFS_B_TO_FSB(mp, fixlen));
5436 5437 5438
		bmv->bmv_length =
			max_t(__int64_t, fixlen - bmv->bmv_offset, 0);
	} else if (bmv->bmv_length == 0) {
L
Linus Torvalds 已提交
5439 5440
		bmv->bmv_entries = 0;
		return 0;
5441 5442
	} else if (bmv->bmv_length < 0) {
		return XFS_ERROR(EINVAL);
L
Linus Torvalds 已提交
5443
	}
5444

L
Linus Torvalds 已提交
5445 5446 5447 5448 5449
	nex = bmv->bmv_count - 1;
	if (nex <= 0)
		return XFS_ERROR(EINVAL);
	bmvend = bmv->bmv_offset + bmv->bmv_length;

5450 5451 5452 5453 5454 5455 5456

	if (bmv->bmv_count > ULONG_MAX / sizeof(struct getbmapx))
		return XFS_ERROR(ENOMEM);
	out = kmem_zalloc(bmv->bmv_count * sizeof(struct getbmapx), KM_MAYFAIL);
	if (!out)
		return XFS_ERROR(ENOMEM);

L
Linus Torvalds 已提交
5457
	xfs_ilock(ip, XFS_IOLOCK_SHARED);
5458 5459 5460 5461 5462
	if (whichfork == XFS_DATA_FORK && !(iflags & BMV_IF_DELALLOC)) {
		if (ip->i_delayed_blks || ip->i_size > ip->i_d.di_size) {
			error = xfs_flush_pages(ip, 0, -1, 0, FI_REMAPF);
			if (error)
				goto out_unlock_iolock;
5463
		}
L
Linus Torvalds 已提交
5464

5465 5466
		ASSERT(ip->i_delayed_blks == 0);
	}
L
Linus Torvalds 已提交
5467 5468 5469 5470 5471 5472 5473 5474 5475 5476

	lock = xfs_ilock_map_shared(ip);

	/*
	 * Don't let nex be bigger than the number of extents
	 * we can have assuming alternating holes and real extents.
	 */
	if (nex > XFS_IFORK_NEXTENTS(ip, whichfork) * 2 + 1)
		nex = XFS_IFORK_NEXTENTS(ip, whichfork) * 2 + 1;

5477 5478 5479
	bmapi_flags = xfs_bmapi_aflag(whichfork);
	if (!(iflags & BMV_IF_PREALLOC))
		bmapi_flags |= XFS_BMAPI_IGSTATE;
L
Linus Torvalds 已提交
5480 5481 5482 5483

	/*
	 * Allocate enough space to handle "subnex" maps at a time.
	 */
5484
	error = ENOMEM;
L
Linus Torvalds 已提交
5485
	subnex = 16;
5486
	map = kmem_alloc(subnex * sizeof(*map), KM_MAYFAIL | KM_NOFS);
5487 5488
	if (!map)
		goto out_unlock_ilock;
L
Linus Torvalds 已提交
5489 5490 5491

	bmv->bmv_entries = 0;

5492 5493 5494 5495
	if (XFS_IFORK_NEXTENTS(ip, whichfork) == 0 &&
	    (whichfork == XFS_ATTR_FORK || !(iflags & BMV_IF_DELALLOC))) {
		error = 0;
		goto out_free_map;
L
Linus Torvalds 已提交
5496 5497 5498 5499 5500 5501 5502 5503
	}

	nexleft = nex;

	do {
		nmap = (nexleft > subnex) ? subnex : nexleft;
		error = xfs_bmapi(NULL, ip, XFS_BB_TO_FSBT(mp, bmv->bmv_offset),
				  XFS_BB_TO_FSB(mp, bmv->bmv_length),
5504
				  bmapi_flags, NULL, 0, map, &nmap,
5505
				  NULL);
L
Linus Torvalds 已提交
5506
		if (error)
5507
			goto out_free_map;
L
Linus Torvalds 已提交
5508 5509 5510
		ASSERT(nmap <= subnex);

		for (i = 0; i < nmap && nexleft && bmv->bmv_length; i++) {
5511
			out[cur_ext].bmv_oflags = 0;
E
Eric Sandeen 已提交
5512
			if (map[i].br_state == XFS_EXT_UNWRITTEN)
5513
				out[cur_ext].bmv_oflags |= BMV_OF_PREALLOC;
E
Eric Sandeen 已提交
5514
			else if (map[i].br_startblock == DELAYSTARTBLOCK)
5515 5516 5517 5518 5519 5520 5521
				out[cur_ext].bmv_oflags |= BMV_OF_DELALLOC;
			out[cur_ext].bmv_offset =
				XFS_FSB_TO_BB(mp, map[i].br_startoff);
			out[cur_ext].bmv_length =
				XFS_FSB_TO_BB(mp, map[i].br_blockcount);
			out[cur_ext].bmv_unused1 = 0;
			out[cur_ext].bmv_unused2 = 0;
E
Eric Sandeen 已提交
5522 5523
			ASSERT(((iflags & BMV_IF_DELALLOC) != 0) ||
			      (map[i].br_startblock != DELAYSTARTBLOCK));
5524
                        if (map[i].br_startblock == HOLESTARTBLOCK &&
5525 5526
			    whichfork == XFS_ATTR_FORK) {
				/* came to the end of attribute fork */
5527
				out[cur_ext].bmv_oflags |= BMV_OF_LAST;
5528
				goto out_free_map;
L
Linus Torvalds 已提交
5529
			}
5530

5531 5532 5533
			if (!xfs_getbmapx_fix_eof_hole(ip, &out[cur_ext],
					prealloced, bmvend,
					map[i].br_startblock))
5534 5535 5536
				goto out_free_map;

			bmv->bmv_offset =
5537 5538
				out[cur_ext].bmv_offset +
				out[cur_ext].bmv_length;
5539 5540
			bmv->bmv_length =
				max_t(__int64_t, 0, bmvend - bmv->bmv_offset);
T
Tao Ma 已提交
5541 5542 5543 5544 5545 5546 5547 5548 5549 5550 5551 5552 5553

			/*
			 * In case we don't want to return the hole,
			 * don't increase cur_ext so that we can reuse
			 * it in the next loop.
			 */
			if ((iflags & BMV_IF_NO_HOLES) &&
			    map[i].br_startblock == HOLESTARTBLOCK) {
				memset(&out[cur_ext], 0, sizeof(out[cur_ext]));
				continue;
			}

			nexleft--;
5554
			bmv->bmv_entries++;
5555
			cur_ext++;
L
Linus Torvalds 已提交
5556 5557 5558
		}
	} while (nmap && nexleft && bmv->bmv_length);

5559 5560 5561
 out_free_map:
	kmem_free(map);
 out_unlock_ilock:
L
Linus Torvalds 已提交
5562
	xfs_iunlock_map_shared(ip, lock);
5563
 out_unlock_iolock:
L
Linus Torvalds 已提交
5564
	xfs_iunlock(ip, XFS_IOLOCK_SHARED);
5565 5566 5567 5568 5569 5570 5571 5572 5573 5574

	for (i = 0; i < cur_ext; i++) {
		int full = 0;	/* user array is full */

		/* format results & advance arg */
		error = formatter(&arg, &out[i], &full);
		if (error || full)
			break;
	}

5575
	kmem_free(out);
L
Linus Torvalds 已提交
5576 5577 5578 5579 5580 5581 5582 5583 5584
	return error;
}

/*
 * 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.
 */
5585
STATIC int				/* error */
L
Linus Torvalds 已提交
5586 5587 5588 5589 5590 5591 5592 5593
xfs_bmap_isaeof(
	xfs_inode_t	*ip,		/* incore inode pointer */
	xfs_fileoff_t   off,		/* file offset in fsblocks */
	int             whichfork,	/* data or attribute fork */
	char		*aeof)		/* return value */
{
	int		error;		/* error return value */
	xfs_ifork_t	*ifp;		/* inode fork pointer */
5594
	xfs_bmbt_rec_host_t *lastrec;	/* extent record pointer */
5595 5596
	xfs_extnum_t	nextents;	/* number of file extents */
	xfs_bmbt_irec_t	s;		/* expanded extent record */
L
Linus Torvalds 已提交
5597 5598 5599 5600 5601 5602 5603 5604 5605 5606 5607 5608 5609 5610

	ASSERT(whichfork == XFS_DATA_FORK);
	ifp = XFS_IFORK_PTR(ip, whichfork);
	if (!(ifp->if_flags & XFS_IFEXTENTS) &&
	    (error = xfs_iread_extents(NULL, ip, whichfork)))
		return error;
	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
	if (nextents == 0) {
		*aeof = 1;
		return 0;
	}
	/*
	 * Go to the last extent
	 */
5611
	lastrec = xfs_iext_get_ext(ifp, nextents - 1);
L
Linus Torvalds 已提交
5612 5613 5614 5615 5616 5617 5618
	xfs_bmbt_get_all(lastrec, &s);
	/*
	 * Check we are allocating in the last extent (for delayed allocations)
	 * or past the last extent for non-delayed allocations.
	 */
	*aeof = (off >= s.br_startoff &&
		 off < s.br_startoff + s.br_blockcount &&
5619
		 isnullstartblock(s.br_startblock)) ||
L
Linus Torvalds 已提交
5620 5621 5622 5623 5624 5625 5626 5627 5628 5629 5630 5631 5632 5633 5634 5635 5636 5637
		off >= s.br_startoff + s.br_blockcount;
	return 0;
}

/*
 * Check if the endoff is outside the last extent. If so the caller will grow
 * the allocation to a stripe unit boundary.
 */
int					/* error */
xfs_bmap_eof(
	xfs_inode_t	*ip,		/* incore inode pointer */
	xfs_fileoff_t	endoff,		/* file offset in fsblocks */
	int		whichfork,	/* data or attribute fork */
	int		*eof)		/* result value */
{
	xfs_fsblock_t	blockcount;	/* extent block count */
	int		error;		/* error return value */
	xfs_ifork_t	*ifp;		/* inode fork pointer */
5638
	xfs_bmbt_rec_host_t *lastrec;	/* extent record pointer */
5639
	xfs_extnum_t	nextents;	/* number of file extents */
L
Linus Torvalds 已提交
5640 5641 5642 5643 5644 5645 5646 5647 5648 5649 5650 5651 5652 5653 5654
	xfs_fileoff_t	startoff;	/* extent starting file offset */

	ASSERT(whichfork == XFS_DATA_FORK);
	ifp = XFS_IFORK_PTR(ip, whichfork);
	if (!(ifp->if_flags & XFS_IFEXTENTS) &&
	    (error = xfs_iread_extents(NULL, ip, whichfork)))
		return error;
	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
	if (nextents == 0) {
		*eof = 1;
		return 0;
	}
	/*
	 * Go to the last extent
	 */
5655
	lastrec = xfs_iext_get_ext(ifp, nextents - 1);
L
Linus Torvalds 已提交
5656 5657 5658 5659 5660 5661 5662
	startoff = xfs_bmbt_get_startoff(lastrec);
	blockcount = xfs_bmbt_get_blockcount(lastrec);
	*eof = endoff >= startoff + blockcount;
	return 0;
}

#ifdef DEBUG
5663
STATIC struct xfs_buf *
L
Linus Torvalds 已提交
5664
xfs_bmap_get_bp(
5665
	struct xfs_btree_cur	*cur,
L
Linus Torvalds 已提交
5666 5667
	xfs_fsblock_t		bno)
{
5668 5669
	struct xfs_log_item_desc *lidp;
	int			i;
L
Linus Torvalds 已提交
5670 5671

	if (!cur)
5672 5673 5674 5675 5676 5677 5678
		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
5679 5680
	}

5681 5682
	/* 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) {
5683
		struct xfs_buf_log_item	*bip;
5684 5685 5686 5687
		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;
L
Linus Torvalds 已提交
5688
	}
5689

5690
	return NULL;
L
Linus Torvalds 已提交
5691 5692
}

H
Hannes Eder 已提交
5693
STATIC void
L
Linus Torvalds 已提交
5694
xfs_check_block(
5695
	struct xfs_btree_block	*block,
L
Linus Torvalds 已提交
5696 5697 5698 5699 5700
	xfs_mount_t		*mp,
	int			root,
	short			sz)
{
	int			i, j, dmxr;
5701
	__be64			*pp, *thispa;	/* pointer to block address */
L
Linus Torvalds 已提交
5702 5703
	xfs_bmbt_key_t		*prevp, *keyp;

5704
	ASSERT(be16_to_cpu(block->bb_level) > 0);
L
Linus Torvalds 已提交
5705 5706

	prevp = NULL;
5707
	for( i = 1; i <= xfs_btree_get_numrecs(block); i++) {
L
Linus Torvalds 已提交
5708
		dmxr = mp->m_bmap_dmxr[0];
5709
		keyp = XFS_BMBT_KEY_ADDR(mp, block, i);
L
Linus Torvalds 已提交
5710 5711

		if (prevp) {
5712 5713
			ASSERT(be64_to_cpu(prevp->br_startoff) <
			       be64_to_cpu(keyp->br_startoff));
L
Linus Torvalds 已提交
5714 5715 5716 5717 5718 5719
		}
		prevp = keyp;

		/*
		 * Compare the block numbers to see if there are dups.
		 */
5720
		if (root)
5721
			pp = XFS_BMAP_BROOT_PTR_ADDR(mp, block, i, sz);
5722 5723 5724
		else
			pp = XFS_BMBT_PTR_ADDR(mp, block, i, dmxr);

5725
		for (j = i+1; j <= be16_to_cpu(block->bb_numrecs); j++) {
5726
			if (root)
5727
				thispa = XFS_BMAP_BROOT_PTR_ADDR(mp, block, j, sz);
5728 5729
			else
				thispa = XFS_BMBT_PTR_ADDR(mp, block, j, dmxr);
5730
			if (*thispa == *pp) {
L
Linus Torvalds 已提交
5731
				cmn_err(CE_WARN, "%s: thispa(%d) == pp(%d) %Ld",
5732
					__func__, j, i,
5733
					(unsigned long long)be64_to_cpu(*thispa));
L
Linus Torvalds 已提交
5734
				panic("%s: ptrs are equal in node\n",
5735
					__func__);
L
Linus Torvalds 已提交
5736 5737 5738 5739 5740 5741 5742 5743 5744 5745 5746 5747 5748 5749 5750 5751
			}
		}
	}
}

/*
 * Check that the extents for the inode ip are in the right order in all
 * btree leaves.
 */

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 */
{
5752
	struct xfs_btree_block	*block;	/* current btree block */
L
Linus Torvalds 已提交
5753 5754 5755
	xfs_fsblock_t		bno;	/* block # of "block" */
	xfs_buf_t		*bp;	/* buffer for "block" */
	int			error;	/* error return value */
5756
	xfs_extnum_t		i=0, j;	/* index into the extents list */
L
Linus Torvalds 已提交
5757 5758 5759
	xfs_ifork_t		*ifp;	/* fork structure */
	int			level;	/* btree level, for checking */
	xfs_mount_t		*mp;	/* file system mount structure */
5760
	__be64			*pp;	/* pointer to block address */
5761
	xfs_bmbt_rec_t		*ep;	/* pointer to current extent */
5762
	xfs_bmbt_rec_t		last = {0, 0}; /* last extent in prev block */
5763
	xfs_bmbt_rec_t		*nextp;	/* pointer to next extent */
L
Linus Torvalds 已提交
5764 5765 5766 5767 5768 5769 5770 5771 5772 5773 5774 5775 5776
	int			bp_release = 0;

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

	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.
	 */
5777 5778
	level = be16_to_cpu(block->bb_level);
	ASSERT(level > 0);
L
Linus Torvalds 已提交
5779
	xfs_check_block(block, mp, 1, ifp->if_broot_bytes);
5780
	pp = XFS_BMAP_BROOT_PTR_ADDR(mp, block, 1, ifp->if_broot_bytes);
5781 5782 5783 5784 5785 5786
	bno = be64_to_cpu(*pp);

	ASSERT(bno != NULLDFSBNO);
	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
5787 5788 5789 5790 5791 5792 5793 5794 5795 5796 5797 5798 5799 5800 5801
	/*
	 * 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 = xfs_bmap_get_bp(cur, XFS_FSB_TO_DADDR(mp, bno));
		if (bp) {
			bp_release = 0;
		} else {
			bp_release = 1;
		}
		if (!bp && (error = xfs_btree_read_bufl(mp, NULL, bno, 0, &bp,
				XFS_BMAP_BTREE_REF)))
			goto error_norelse;
5802
		block = XFS_BUF_TO_BLOCK(bp);
L
Linus Torvalds 已提交
5803
		XFS_WANT_CORRUPTED_GOTO(
5804
			xfs_bmap_sanity_check(mp, bp, level),
L
Linus Torvalds 已提交
5805 5806 5807 5808 5809 5810 5811 5812 5813 5814
			error0);
		if (level == 0)
			break;

		/*
		 * Check this block for basic sanity (increasing keys and
		 * no duplicate blocks).
		 */

		xfs_check_block(block, mp, 0, 0);
5815
		pp = XFS_BMBT_PTR_ADDR(mp, block, 1, mp->m_bmap_dmxr[1]);
5816 5817
		bno = be64_to_cpu(*pp);
		XFS_WANT_CORRUPTED_GOTO(XFS_FSB_SANITY_CHECK(mp, bno), error0);
L
Linus Torvalds 已提交
5818 5819 5820 5821 5822 5823 5824 5825 5826 5827 5828 5829 5830 5831 5832 5833 5834 5835 5836
		if (bp_release) {
			bp_release = 0;
			xfs_trans_brelse(NULL, bp);
		}
	}

	/*
	 * 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;


5837
		num_recs = xfs_btree_get_numrecs(block);
L
Linus Torvalds 已提交
5838 5839 5840 5841 5842

		/*
		 * Read-ahead the next leaf block, if any.
		 */

5843
		nextbno = be64_to_cpu(block->bb_u.l.bb_rightsib);
L
Linus Torvalds 已提交
5844 5845 5846 5847 5848 5849 5850

		/*
		 * 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.
		 */

5851
		ep = XFS_BMBT_REC_ADDR(mp, block, 1);
5852
		if (i) {
5853 5854 5855
			ASSERT(xfs_bmbt_disk_get_startoff(&last) +
			       xfs_bmbt_disk_get_blockcount(&last) <=
			       xfs_bmbt_disk_get_startoff(ep));
5856
		}
5857
		for (j = 1; j < num_recs; j++) {
5858
			nextp = XFS_BMBT_REC_ADDR(mp, block, j + 1);
5859 5860 5861
			ASSERT(xfs_bmbt_disk_get_startoff(ep) +
			       xfs_bmbt_disk_get_blockcount(ep) <=
			       xfs_bmbt_disk_get_startoff(nextp));
5862
			ep = nextp;
L
Linus Torvalds 已提交
5863 5864
		}

5865
		last = *ep;
L
Linus Torvalds 已提交
5866 5867 5868 5869 5870 5871 5872 5873 5874 5875 5876 5877 5878 5879 5880 5881 5882 5883 5884 5885 5886
		i += num_recs;
		if (bp_release) {
			bp_release = 0;
			xfs_trans_brelse(NULL, bp);
		}
		bno = nextbno;
		/*
		 * If we've reached the end, stop.
		 */
		if (bno == NULLFSBLOCK)
			break;

		bp = xfs_bmap_get_bp(cur, XFS_FSB_TO_DADDR(mp, bno));
		if (bp) {
			bp_release = 0;
		} else {
			bp_release = 1;
		}
		if (!bp && (error = xfs_btree_read_bufl(mp, NULL, bno, 0, &bp,
				XFS_BMAP_BTREE_REF)))
			goto error_norelse;
5887
		block = XFS_BUF_TO_BLOCK(bp);
L
Linus Torvalds 已提交
5888 5889 5890 5891 5892 5893 5894 5895
	}
	if (bp_release) {
		bp_release = 0;
		xfs_trans_brelse(NULL, bp);
	}
	return;

error0:
5896
	cmn_err(CE_WARN, "%s: at error0", __func__);
L
Linus Torvalds 已提交
5897 5898 5899 5900
	if (bp_release)
		xfs_trans_brelse(NULL, bp);
error_norelse:
	cmn_err(CE_WARN, "%s: BAD after btree leaves for %d extents",
5901 5902
		__func__, i);
	panic("%s: CORRUPTED BTREE OR SOMETHING", __func__);
L
Linus Torvalds 已提交
5903 5904 5905 5906 5907 5908 5909 5910 5911 5912 5913 5914 5915 5916
	return;
}
#endif

/*
 * Count fsblocks of the given fork.
 */
int						/* error */
xfs_bmap_count_blocks(
	xfs_trans_t		*tp,		/* transaction pointer */
	xfs_inode_t		*ip,		/* incore inode */
	int			whichfork,	/* data or attr fork */
	int			*count)		/* out: count of blocks */
{
5917
	struct xfs_btree_block	*block;	/* current btree block */
L
Linus Torvalds 已提交
5918 5919 5920 5921
	xfs_fsblock_t		bno;	/* block # of "block" */
	xfs_ifork_t		*ifp;	/* fork structure */
	int			level;	/* btree level, for checking */
	xfs_mount_t		*mp;	/* file system mount structure */
5922
	__be64			*pp;	/* pointer to block address */
L
Linus Torvalds 已提交
5923 5924 5925 5926 5927

	bno = NULLFSBLOCK;
	mp = ip->i_mount;
	ifp = XFS_IFORK_PTR(ip, whichfork);
	if ( XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_EXTENTS ) {
5928
		xfs_bmap_count_leaves(ifp, 0,
L
Linus Torvalds 已提交
5929
			ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t),
5930
			count);
L
Linus Torvalds 已提交
5931 5932 5933 5934 5935 5936 5937
		return 0;
	}

	/*
	 * Root level must use BMAP_BROOT_PTR_ADDR macro to get ptr out.
	 */
	block = ifp->if_broot;
5938 5939
	level = be16_to_cpu(block->bb_level);
	ASSERT(level > 0);
5940
	pp = XFS_BMAP_BROOT_PTR_ADDR(mp, block, 1, ifp->if_broot_bytes);
5941 5942 5943 5944
	bno = be64_to_cpu(*pp);
	ASSERT(bno != NULLDFSBNO);
	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
5945

5946
	if (unlikely(xfs_bmap_count_tree(mp, tp, ifp, bno, level, count) < 0)) {
L
Linus Torvalds 已提交
5947 5948 5949 5950 5951 5952 5953 5954 5955 5956 5957 5958
		XFS_ERROR_REPORT("xfs_bmap_count_blocks(2)", XFS_ERRLEVEL_LOW,
				 mp);
		return XFS_ERROR(EFSCORRUPTED);
	}

	return 0;
}

/*
 * Recursively walks each level of a btree
 * to count total fsblocks is use.
 */
D
David Chinner 已提交
5959
STATIC int                                     /* error */
L
Linus Torvalds 已提交
5960 5961 5962
xfs_bmap_count_tree(
	xfs_mount_t     *mp,            /* file system mount point */
	xfs_trans_t     *tp,            /* transaction pointer */
5963
	xfs_ifork_t	*ifp,		/* inode fork pointer */
L
Linus Torvalds 已提交
5964 5965 5966 5967 5968 5969 5970
	xfs_fsblock_t   blockno,	/* file system block number */
	int             levelin,	/* level in btree */
	int		*count)		/* Count of blocks */
{
	int			error;
	xfs_buf_t		*bp, *nbp;
	int			level = levelin;
5971
	__be64			*pp;
L
Linus Torvalds 已提交
5972 5973
	xfs_fsblock_t           bno = blockno;
	xfs_fsblock_t		nextbno;
5974
	struct xfs_btree_block	*block, *nextblock;
L
Linus Torvalds 已提交
5975 5976 5977 5978 5979
	int			numrecs;

	if ((error = xfs_btree_read_bufl(mp, tp, bno, 0, &bp, XFS_BMAP_BTREE_REF)))
		return error;
	*count += 1;
5980
	block = XFS_BUF_TO_BLOCK(bp);
L
Linus Torvalds 已提交
5981 5982

	if (--level) {
M
Malcolm Parsons 已提交
5983
		/* Not at node above leaves, count this level of nodes */
5984
		nextbno = be64_to_cpu(block->bb_u.l.bb_rightsib);
L
Linus Torvalds 已提交
5985 5986 5987 5988 5989
		while (nextbno != NULLFSBLOCK) {
			if ((error = xfs_btree_read_bufl(mp, tp, nextbno,
				0, &nbp, XFS_BMAP_BTREE_REF)))
				return error;
			*count += 1;
5990 5991
			nextblock = XFS_BUF_TO_BLOCK(nbp);
			nextbno = be64_to_cpu(nextblock->bb_u.l.bb_rightsib);
L
Linus Torvalds 已提交
5992 5993 5994 5995
			xfs_trans_brelse(tp, nbp);
		}

		/* Dive to the next level */
5996
		pp = XFS_BMBT_PTR_ADDR(mp, block, 1, mp->m_bmap_dmxr[1]);
5997
		bno = be64_to_cpu(*pp);
L
Linus Torvalds 已提交
5998
		if (unlikely((error =
5999
		     xfs_bmap_count_tree(mp, tp, ifp, bno, level, count)) < 0)) {
L
Linus Torvalds 已提交
6000 6001 6002 6003 6004 6005 6006 6007 6008
			xfs_trans_brelse(tp, bp);
			XFS_ERROR_REPORT("xfs_bmap_count_tree(1)",
					 XFS_ERRLEVEL_LOW, mp);
			return XFS_ERROR(EFSCORRUPTED);
		}
		xfs_trans_brelse(tp, bp);
	} else {
		/* count all level 1 nodes and their leaves */
		for (;;) {
6009
			nextbno = be64_to_cpu(block->bb_u.l.bb_rightsib);
6010
			numrecs = be16_to_cpu(block->bb_numrecs);
6011
			xfs_bmap_disk_count_leaves(mp, block, numrecs, count);
L
Linus Torvalds 已提交
6012 6013 6014 6015 6016 6017 6018 6019
			xfs_trans_brelse(tp, bp);
			if (nextbno == NULLFSBLOCK)
				break;
			bno = nextbno;
			if ((error = xfs_btree_read_bufl(mp, tp, bno, 0, &bp,
				XFS_BMAP_BTREE_REF)))
				return error;
			*count += 1;
6020
			block = XFS_BUF_TO_BLOCK(bp);
L
Linus Torvalds 已提交
6021 6022 6023 6024 6025 6026
		}
	}
	return 0;
}

/*
6027
 * Count leaf blocks given a range of extent records.
L
Linus Torvalds 已提交
6028
 */
6029
STATIC void
L
Linus Torvalds 已提交
6030
xfs_bmap_count_leaves(
6031 6032
	xfs_ifork_t		*ifp,
	xfs_extnum_t		idx,
L
Linus Torvalds 已提交
6033 6034 6035 6036 6037
	int			numrecs,
	int			*count)
{
	int		b;

6038
	for (b = 0; b < numrecs; b++) {
6039
		xfs_bmbt_rec_host_t *frp = xfs_iext_get_ext(ifp, idx + b);
6040
		*count += xfs_bmbt_get_blockcount(frp);
6041
	}
6042 6043 6044
}

/*
6045 6046
 * Count leaf blocks given a range of extent records originally
 * in btree format.
6047
 */
6048
STATIC void
6049
xfs_bmap_disk_count_leaves(
6050
	struct xfs_mount	*mp,
6051
	struct xfs_btree_block	*block,
6052 6053 6054 6055
	int			numrecs,
	int			*count)
{
	int		b;
6056
	xfs_bmbt_rec_t	*frp;
6057

6058
	for (b = 1; b <= numrecs; b++) {
6059
		frp = XFS_BMBT_REC_ADDR(mp, block, b);
L
Linus Torvalds 已提交
6060
		*count += xfs_bmbt_disk_get_blockcount(frp);
6061
	}
L
Linus Torvalds 已提交
6062
}