xfs_dir2_leaf.c 52.6 KB
Newer Older
L
Linus Torvalds 已提交
1
/*
2 3
 * Copyright (c) 2000-2003,2005 Silicon Graphics, Inc.
 * 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 28 29 30
#include "xfs_trans.h"
#include "xfs_sb.h"
#include "xfs_ag.h"
#include "xfs_dir.h"
#include "xfs_dir2.h"
#include "xfs_dmapi.h"
#include "xfs_mount.h"
31
#include "xfs_da_btree.h"
L
Linus Torvalds 已提交
32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55
#include "xfs_bmap_btree.h"
#include "xfs_attr_sf.h"
#include "xfs_dir_sf.h"
#include "xfs_dir2_sf.h"
#include "xfs_dinode.h"
#include "xfs_inode.h"
#include "xfs_bmap.h"
#include "xfs_dir2_data.h"
#include "xfs_dir2_leaf.h"
#include "xfs_dir2_block.h"
#include "xfs_dir2_node.h"
#include "xfs_dir2_trace.h"
#include "xfs_error.h"

/*
 * Local function declarations.
 */
#ifdef DEBUG
static void xfs_dir2_leaf_check(xfs_inode_t *dp, xfs_dabuf_t *bp);
#else
#define	xfs_dir2_leaf_check(dp, bp)
#endif
static int xfs_dir2_leaf_lookup_int(xfs_da_args_t *args, xfs_dabuf_t **lbpp,
				    int *indexp, xfs_dabuf_t **dbpp);
56 57
static void xfs_dir2_leaf_log_bests(struct xfs_trans *tp, struct xfs_dabuf *bp,
				    int first, int last);
58
static void xfs_dir2_leaf_log_tail(struct xfs_trans *tp, struct xfs_dabuf *bp);
59

L
Linus Torvalds 已提交
60 61 62 63 64 65 66 67 68

/*
 * Convert a block form directory to a leaf form directory.
 */
int						/* error */
xfs_dir2_block_to_leaf(
	xfs_da_args_t		*args,		/* operation arguments */
	xfs_dabuf_t		*dbp)		/* input block's buffer */
{
69
	__be16			*bestsp;	/* leaf's bestsp entries */
L
Linus Torvalds 已提交
70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113
	xfs_dablk_t		blkno;		/* leaf block's bno */
	xfs_dir2_block_t	*block;		/* block structure */
	xfs_dir2_leaf_entry_t	*blp;		/* block's leaf entries */
	xfs_dir2_block_tail_t	*btp;		/* block's tail */
	xfs_inode_t		*dp;		/* incore directory inode */
	int			error;		/* error return code */
	xfs_dabuf_t		*lbp;		/* leaf block's buffer */
	xfs_dir2_db_t		ldb;		/* leaf block's bno */
	xfs_dir2_leaf_t		*leaf;		/* leaf structure */
	xfs_dir2_leaf_tail_t	*ltp;		/* leaf's tail */
	xfs_mount_t		*mp;		/* filesystem mount point */
	int			needlog;	/* need to log block header */
	int			needscan;	/* need to rescan bestfree */
	xfs_trans_t		*tp;		/* transaction pointer */

	xfs_dir2_trace_args_b("block_to_leaf", args, dbp);
	dp = args->dp;
	mp = dp->i_mount;
	tp = args->trans;
	/*
	 * Add the leaf block to the inode.
	 * This interface will only put blocks in the leaf/node range.
	 * Since that's empty now, we'll get the root (block 0 in range).
	 */
	if ((error = xfs_da_grow_inode(args, &blkno))) {
		return error;
	}
	ldb = XFS_DIR2_DA_TO_DB(mp, blkno);
	ASSERT(ldb == XFS_DIR2_LEAF_FIRSTDB(mp));
	/*
	 * Initialize the leaf block, get a buffer for it.
	 */
	if ((error = xfs_dir2_leaf_init(args, ldb, &lbp, XFS_DIR2_LEAF1_MAGIC))) {
		return error;
	}
	ASSERT(lbp != NULL);
	leaf = lbp->data;
	block = dbp->data;
	xfs_dir2_data_check(dp, dbp);
	btp = XFS_DIR2_BLOCK_TAIL_P(mp, block);
	blp = XFS_DIR2_BLOCK_LEAF_P(btp);
	/*
	 * Set the counts in the leaf header.
	 */
114 115
	leaf->hdr.count = cpu_to_be16(be32_to_cpu(btp->count));
	leaf->hdr.stale = cpu_to_be16(be32_to_cpu(btp->stale));
L
Linus Torvalds 已提交
116 117 118 119
	/*
	 * Could compact these but I think we always do the conversion
	 * after squeezing out stale entries.
	 */
120
	memcpy(leaf->ents, blp, be32_to_cpu(btp->count) * sizeof(xfs_dir2_leaf_entry_t));
121
	xfs_dir2_leaf_log_ents(tp, lbp, 0, be16_to_cpu(leaf->hdr.count) - 1);
L
Linus Torvalds 已提交
122 123 124 125 126 127 128 129 130 131 132 133 134 135
	needscan = 0;
	needlog = 1;
	/*
	 * Make the space formerly occupied by the leaf entries and block
	 * tail be free.
	 */
	xfs_dir2_data_make_free(tp, dbp,
		(xfs_dir2_data_aoff_t)((char *)blp - (char *)block),
		(xfs_dir2_data_aoff_t)((char *)block + mp->m_dirblksize -
				       (char *)blp),
		&needlog, &needscan);
	/*
	 * Fix up the block header, make it a data block.
	 */
136
	block->hdr.magic = cpu_to_be32(XFS_DIR2_DATA_MAGIC);
L
Linus Torvalds 已提交
137 138 139 140 141 142 143
	if (needscan)
		xfs_dir2_data_freescan(mp, (xfs_dir2_data_t *)block, &needlog,
			NULL);
	/*
	 * Set up leaf tail and bests table.
	 */
	ltp = XFS_DIR2_LEAF_TAIL_P(mp, leaf);
144
	ltp->bestcount = cpu_to_be32(1);
L
Linus Torvalds 已提交
145
	bestsp = XFS_DIR2_LEAF_BESTS_P(ltp);
146
	bestsp[0] =  block->hdr.bestfree[0].length;
L
Linus Torvalds 已提交
147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165
	/*
	 * Log the data header and leaf bests table.
	 */
	if (needlog)
		xfs_dir2_data_log_header(tp, dbp);
	xfs_dir2_leaf_check(dp, lbp);
	xfs_dir2_data_check(dp, dbp);
	xfs_dir2_leaf_log_bests(tp, lbp, 0, 0);
	xfs_da_buf_done(lbp);
	return 0;
}

/*
 * Add an entry to a leaf form directory.
 */
int						/* error */
xfs_dir2_leaf_addname(
	xfs_da_args_t		*args)		/* operation arguments */
{
166
	__be16			*bestsp;	/* freespace table in leaf */
L
Linus Torvalds 已提交
167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189
	int			compact;	/* need to compact leaves */
	xfs_dir2_data_t		*data;		/* data block structure */
	xfs_dabuf_t		*dbp;		/* data block buffer */
	xfs_dir2_data_entry_t	*dep;		/* data block entry */
	xfs_inode_t		*dp;		/* incore directory inode */
	xfs_dir2_data_unused_t	*dup;		/* data unused entry */
	int			error;		/* error return value */
	int			grown;		/* allocated new data block */
	int			highstale;	/* index of next stale leaf */
	int			i;		/* temporary, index */
	int			index;		/* leaf table position */
	xfs_dabuf_t		*lbp;		/* leaf's buffer */
	xfs_dir2_leaf_t		*leaf;		/* leaf structure */
	int			length;		/* length of new entry */
	xfs_dir2_leaf_entry_t	*lep;		/* leaf entry table pointer */
	int			lfloglow;	/* low leaf logging index */
	int			lfloghigh;	/* high leaf logging index */
	int			lowstale;	/* index of prev stale leaf */
	xfs_dir2_leaf_tail_t	*ltp;		/* leaf tail pointer */
	xfs_mount_t		*mp;		/* filesystem mount point */
	int			needbytes;	/* leaf block bytes needed */
	int			needlog;	/* need to log data header */
	int			needscan;	/* need to rescan data free */
190
	__be16			*tagp;		/* end of data entry */
L
Linus Torvalds 已提交
191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224
	xfs_trans_t		*tp;		/* transaction pointer */
	xfs_dir2_db_t		use_block;	/* data block number */

	xfs_dir2_trace_args("leaf_addname", args);
	dp = args->dp;
	tp = args->trans;
	mp = dp->i_mount;
	/*
	 * Read the leaf block.
	 */
	error = xfs_da_read_buf(tp, dp, mp->m_dirleafblk, -1, &lbp,
		XFS_DATA_FORK);
	if (error) {
		return error;
	}
	ASSERT(lbp != NULL);
	/*
	 * Look up the entry by hash value and name.
	 * We know it's not there, our caller has already done a lookup.
	 * So the index is of the entry to insert in front of.
	 * But if there are dup hash values the index is of the first of those.
	 */
	index = xfs_dir2_leaf_search_hash(args, lbp);
	leaf = lbp->data;
	ltp = XFS_DIR2_LEAF_TAIL_P(mp, leaf);
	bestsp = XFS_DIR2_LEAF_BESTS_P(ltp);
	length = XFS_DIR2_DATA_ENTSIZE(args->namelen);
	/*
	 * See if there are any entries with the same hash value
	 * and space in their block for the new entry.
	 * This is good because it puts multiple same-hash value entries
	 * in a data block, improving the lookup of those entries.
	 */
	for (use_block = -1, lep = &leaf->ents[index];
225
	     index < be16_to_cpu(leaf->hdr.count) && be32_to_cpu(lep->hashval) == args->hashval;
L
Linus Torvalds 已提交
226
	     index++, lep++) {
227
		if (be32_to_cpu(lep->address) == XFS_DIR2_NULL_DATAPTR)
L
Linus Torvalds 已提交
228
			continue;
229
		i = XFS_DIR2_DATAPTR_TO_DB(mp, be32_to_cpu(lep->address));
230
		ASSERT(i < be32_to_cpu(ltp->bestcount));
231 232
		ASSERT(be16_to_cpu(bestsp[i]) != NULLDATAOFF);
		if (be16_to_cpu(bestsp[i]) >= length) {
L
Linus Torvalds 已提交
233 234 235 236 237 238 239 240
			use_block = i;
			break;
		}
	}
	/*
	 * Didn't find a block yet, linear search all the data blocks.
	 */
	if (use_block == -1) {
241
		for (i = 0; i < be32_to_cpu(ltp->bestcount); i++) {
L
Linus Torvalds 已提交
242 243 244
			/*
			 * Remember a block we see that's missing.
			 */
245
			if (be16_to_cpu(bestsp[i]) == NULLDATAOFF && use_block == -1)
L
Linus Torvalds 已提交
246
				use_block = i;
247
			else if (be16_to_cpu(bestsp[i]) >= length) {
L
Linus Torvalds 已提交
248 249 250 251 252 253 254 255 256 257 258 259 260 261 262
				use_block = i;
				break;
			}
		}
	}
	/*
	 * How many bytes do we need in the leaf block?
	 */
	needbytes =
		(leaf->hdr.stale ? 0 : (uint)sizeof(leaf->ents[0])) +
		(use_block != -1 ? 0 : (uint)sizeof(leaf->bests[0]));
	/*
	 * Now kill use_block if it refers to a missing block, so we
	 * can use it as an indication of allocation needed.
	 */
263
	if (use_block != -1 && be16_to_cpu(bestsp[use_block]) == NULLDATAOFF)
L
Linus Torvalds 已提交
264 265 266 267 268
		use_block = -1;
	/*
	 * If we don't have enough free bytes but we can make enough
	 * by compacting out stale entries, we'll do that.
	 */
269 270
	if ((char *)bestsp - (char *)&leaf->ents[be16_to_cpu(leaf->hdr.count)] < needbytes &&
	    be16_to_cpu(leaf->hdr.stale) > 1) {
L
Linus Torvalds 已提交
271 272 273 274 275 276
		compact = 1;
	}
	/*
	 * Otherwise if we don't have enough free bytes we need to
	 * convert to node form.
	 */
277
	else if ((char *)bestsp - (char *)&leaf->ents[be16_to_cpu(leaf->hdr.count)] <
L
Linus Torvalds 已提交
278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332
		 needbytes) {
		/*
		 * Just checking or no space reservation, give up.
		 */
		if (args->justcheck || args->total == 0) {
			xfs_da_brelse(tp, lbp);
			return XFS_ERROR(ENOSPC);
		}
		/*
		 * Convert to node form.
		 */
		error = xfs_dir2_leaf_to_node(args, lbp);
		xfs_da_buf_done(lbp);
		if (error)
			return error;
		/*
		 * Then add the new entry.
		 */
		return xfs_dir2_node_addname(args);
	}
	/*
	 * Otherwise it will fit without compaction.
	 */
	else
		compact = 0;
	/*
	 * If just checking, then it will fit unless we needed to allocate
	 * a new data block.
	 */
	if (args->justcheck) {
		xfs_da_brelse(tp, lbp);
		return use_block == -1 ? XFS_ERROR(ENOSPC) : 0;
	}
	/*
	 * If no allocations are allowed, return now before we've
	 * changed anything.
	 */
	if (args->total == 0 && use_block == -1) {
		xfs_da_brelse(tp, lbp);
		return XFS_ERROR(ENOSPC);
	}
	/*
	 * Need to compact the leaf entries, removing stale ones.
	 * Leave one stale entry behind - the one closest to our
	 * insertion index - and we'll shift that one to our insertion
	 * point later.
	 */
	if (compact) {
		xfs_dir2_leaf_compact_x1(lbp, &index, &lowstale, &highstale,
			&lfloglow, &lfloghigh);
	}
	/*
	 * There are stale entries, so we'll need log-low and log-high
	 * impossibly bad values later.
	 */
333 334
	else if (be16_to_cpu(leaf->hdr.stale)) {
		lfloglow = be16_to_cpu(leaf->hdr.count);
L
Linus Torvalds 已提交
335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360
		lfloghigh = -1;
	}
	/*
	 * If there was no data block space found, we need to allocate
	 * a new one.
	 */
	if (use_block == -1) {
		/*
		 * Add the new data block.
		 */
		if ((error = xfs_dir2_grow_inode(args, XFS_DIR2_DATA_SPACE,
				&use_block))) {
			xfs_da_brelse(tp, lbp);
			return error;
		}
		/*
		 * Initialize the block.
		 */
		if ((error = xfs_dir2_data_init(args, use_block, &dbp))) {
			xfs_da_brelse(tp, lbp);
			return error;
		}
		/*
		 * If we're adding a new data block on the end we need to
		 * extend the bests table.  Copy it up one entry.
		 */
361
		if (use_block >= be32_to_cpu(ltp->bestcount)) {
L
Linus Torvalds 已提交
362 363
			bestsp--;
			memmove(&bestsp[0], &bestsp[1],
364 365
				be32_to_cpu(ltp->bestcount) * sizeof(bestsp[0]));
			be32_add(&ltp->bestcount, 1);
L
Linus Torvalds 已提交
366
			xfs_dir2_leaf_log_tail(tp, lbp);
367
			xfs_dir2_leaf_log_bests(tp, lbp, 0, be32_to_cpu(ltp->bestcount) - 1);
L
Linus Torvalds 已提交
368 369 370 371 372 373 374
		}
		/*
		 * If we're filling in a previously empty block just log it.
		 */
		else
			xfs_dir2_leaf_log_bests(tp, lbp, use_block, use_block);
		data = dbp->data;
375
		bestsp[use_block] = data->hdr.bestfree[0].length;
L
Linus Torvalds 已提交
376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396
		grown = 1;
	}
	/*
	 * Already had space in some data block.
	 * Just read that one in.
	 */
	else {
		if ((error =
		    xfs_da_read_buf(tp, dp, XFS_DIR2_DB_TO_DA(mp, use_block),
			    -1, &dbp, XFS_DATA_FORK))) {
			xfs_da_brelse(tp, lbp);
			return error;
		}
		data = dbp->data;
		grown = 0;
	}
	xfs_dir2_data_check(dp, dbp);
	/*
	 * Point to the biggest freespace in our data block.
	 */
	dup = (xfs_dir2_data_unused_t *)
397
	      ((char *)data + be16_to_cpu(data->hdr.bestfree[0].offset));
398
	ASSERT(be16_to_cpu(dup->length) >= length);
L
Linus Torvalds 已提交
399 400 401 402 403 404 405 406 407 408 409
	needscan = needlog = 0;
	/*
	 * Mark the initial part of our freespace in use for the new entry.
	 */
	xfs_dir2_data_use_free(tp, dbp, dup,
		(xfs_dir2_data_aoff_t)((char *)dup - (char *)data), length,
		&needlog, &needscan);
	/*
	 * Initialize our new entry (at last).
	 */
	dep = (xfs_dir2_data_entry_t *)dup;
410
	dep->inumber = cpu_to_be64(args->inumber);
L
Linus Torvalds 已提交
411 412 413
	dep->namelen = args->namelen;
	memcpy(dep->name, args->name, dep->namelen);
	tagp = XFS_DIR2_DATA_ENTRY_TAG_P(dep);
414
	*tagp = cpu_to_be16((char *)dep - (char *)data);
L
Linus Torvalds 已提交
415 416 417 418 419 420 421 422 423 424 425 426 427 428 429
	/*
	 * Need to scan fix up the bestfree table.
	 */
	if (needscan)
		xfs_dir2_data_freescan(mp, data, &needlog, NULL);
	/*
	 * Need to log the data block's header.
	 */
	if (needlog)
		xfs_dir2_data_log_header(tp, dbp);
	xfs_dir2_data_log_entry(tp, dbp, dep);
	/*
	 * If the bests table needs to be changed, do it.
	 * Log the change unless we've already done that.
	 */
430
	if (be16_to_cpu(bestsp[use_block]) != be16_to_cpu(data->hdr.bestfree[0].length)) {
431
		bestsp[use_block] = data->hdr.bestfree[0].length;
L
Linus Torvalds 已提交
432 433 434 435 436 437 438 439 440 441 442
		if (!grown)
			xfs_dir2_leaf_log_bests(tp, lbp, use_block, use_block);
	}
	/*
	 * Now we need to make room to insert the leaf entry.
	 * If there are no stale entries, we just insert a hole at index.
	 */
	if (!leaf->hdr.stale) {
		/*
		 * lep is still good as the index leaf entry.
		 */
443
		if (index < be16_to_cpu(leaf->hdr.count))
L
Linus Torvalds 已提交
444
			memmove(lep + 1, lep,
445
				(be16_to_cpu(leaf->hdr.count) - index) * sizeof(*lep));
L
Linus Torvalds 已提交
446 447 448 449
		/*
		 * Record low and high logging indices for the leaf.
		 */
		lfloglow = index;
450 451
		lfloghigh = be16_to_cpu(leaf->hdr.count);
		be16_add(&leaf->hdr.count, 1);
L
Linus Torvalds 已提交
452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470
	}
	/*
	 * There are stale entries.
	 * We will use one of them for the new entry.
	 * It's probably not at the right location, so we'll have to
	 * shift some up or down first.
	 */
	else {
		/*
		 * If we didn't compact before, we need to find the nearest
		 * stale entries before and after our insertion point.
		 */
		if (compact == 0) {
			/*
			 * Find the first stale entry before the insertion
			 * point, if any.
			 */
			for (lowstale = index - 1;
			     lowstale >= 0 &&
471
				be32_to_cpu(leaf->ents[lowstale].address) !=
L
Linus Torvalds 已提交
472 473 474 475 476 477 478 479 480
				XFS_DIR2_NULL_DATAPTR;
			     lowstale--)
				continue;
			/*
			 * Find the next stale entry at or after the insertion
			 * point, if any.   Stop if we go so far that the
			 * lowstale entry would be better.
			 */
			for (highstale = index;
481
			     highstale < be16_to_cpu(leaf->hdr.count) &&
482
				be32_to_cpu(leaf->ents[highstale].address) !=
L
Linus Torvalds 已提交
483 484 485 486 487 488 489 490 491 492
				XFS_DIR2_NULL_DATAPTR &&
				(lowstale < 0 ||
				 index - lowstale - 1 >= highstale - index);
			     highstale++)
				continue;
		}
		/*
		 * If the low one is better, use it.
		 */
		if (lowstale >= 0 &&
493
		    (highstale == be16_to_cpu(leaf->hdr.count) ||
L
Linus Torvalds 已提交
494 495
		     index - lowstale - 1 < highstale - index)) {
			ASSERT(index - lowstale - 1 >= 0);
496
			ASSERT(be32_to_cpu(leaf->ents[lowstale].address) ==
L
Linus Torvalds 已提交
497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514
			       XFS_DIR2_NULL_DATAPTR);
			/*
			 * Copy entries up to cover the stale entry
			 * and make room for the new entry.
			 */
			if (index - lowstale - 1 > 0)
				memmove(&leaf->ents[lowstale],
					&leaf->ents[lowstale + 1],
					(index - lowstale - 1) * sizeof(*lep));
			lep = &leaf->ents[index - 1];
			lfloglow = MIN(lowstale, lfloglow);
			lfloghigh = MAX(index - 1, lfloghigh);
		}
		/*
		 * The high one is better, so use that one.
		 */
		else {
			ASSERT(highstale - index >= 0);
515
			ASSERT(be32_to_cpu(leaf->ents[highstale].address) ==
L
Linus Torvalds 已提交
516 517
			       XFS_DIR2_NULL_DATAPTR);
			/*
518
			 * Copy entries down to cover the stale entry
L
Linus Torvalds 已提交
519 520 521 522 523 524 525 526 527 528
			 * and make room for the new entry.
			 */
			if (highstale - index > 0)
				memmove(&leaf->ents[index + 1],
					&leaf->ents[index],
					(highstale - index) * sizeof(*lep));
			lep = &leaf->ents[index];
			lfloglow = MIN(index, lfloglow);
			lfloghigh = MAX(highstale, lfloghigh);
		}
529
		be16_add(&leaf->hdr.stale, -1);
L
Linus Torvalds 已提交
530 531 532 533
	}
	/*
	 * Fill in the new leaf entry.
	 */
534 535
	lep->hashval = cpu_to_be32(args->hashval);
	lep->address = cpu_to_be32(XFS_DIR2_DB_OFF_TO_DATAPTR(mp, use_block,
536
				be16_to_cpu(*tagp)));
L
Linus Torvalds 已提交
537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566
	/*
	 * Log the leaf fields and give up the buffers.
	 */
	xfs_dir2_leaf_log_header(tp, lbp);
	xfs_dir2_leaf_log_ents(tp, lbp, lfloglow, lfloghigh);
	xfs_dir2_leaf_check(dp, lbp);
	xfs_da_buf_done(lbp);
	xfs_dir2_data_check(dp, dbp);
	xfs_da_buf_done(dbp);
	return 0;
}

#ifdef DEBUG
/*
 * Check the internal consistency of a leaf1 block.
 * Pop an assert if something is wrong.
 */
void
xfs_dir2_leaf_check(
	xfs_inode_t		*dp,		/* incore directory inode */
	xfs_dabuf_t		*bp)		/* leaf's buffer */
{
	int			i;		/* leaf index */
	xfs_dir2_leaf_t		*leaf;		/* leaf structure */
	xfs_dir2_leaf_tail_t	*ltp;		/* leaf tail pointer */
	xfs_mount_t		*mp;		/* filesystem mount point */
	int			stale;		/* count of stale leaves */

	leaf = bp->data;
	mp = dp->i_mount;
567
	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_DIR2_LEAF1_MAGIC);
L
Linus Torvalds 已提交
568 569 570 571 572
	/*
	 * This value is not restrictive enough.
	 * Should factor in the size of the bests table as well.
	 * We can deduce a value for that from di_size.
	 */
573
	ASSERT(be16_to_cpu(leaf->hdr.count) <= XFS_DIR2_MAX_LEAF_ENTS(mp));
L
Linus Torvalds 已提交
574 575 576 577
	ltp = XFS_DIR2_LEAF_TAIL_P(mp, leaf);
	/*
	 * Leaves and bests don't overlap.
	 */
578
	ASSERT((char *)&leaf->ents[be16_to_cpu(leaf->hdr.count)] <=
L
Linus Torvalds 已提交
579 580 581 582
	       (char *)XFS_DIR2_LEAF_BESTS_P(ltp));
	/*
	 * Check hash value order, count stale entries.
	 */
583 584
	for (i = stale = 0; i < be16_to_cpu(leaf->hdr.count); i++) {
		if (i + 1 < be16_to_cpu(leaf->hdr.count))
585 586 587
			ASSERT(be32_to_cpu(leaf->ents[i].hashval) <=
			       be32_to_cpu(leaf->ents[i + 1].hashval));
		if (be32_to_cpu(leaf->ents[i].address) == XFS_DIR2_NULL_DATAPTR)
L
Linus Torvalds 已提交
588 589
			stale++;
	}
590
	ASSERT(be16_to_cpu(leaf->hdr.stale) == stale);
L
Linus Torvalds 已提交
591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614
}
#endif	/* DEBUG */

/*
 * Compact out any stale entries in the leaf.
 * Log the header and changed leaf entries, if any.
 */
void
xfs_dir2_leaf_compact(
	xfs_da_args_t	*args,		/* operation arguments */
	xfs_dabuf_t	*bp)		/* leaf buffer */
{
	int		from;		/* source leaf index */
	xfs_dir2_leaf_t	*leaf;		/* leaf structure */
	int		loglow;		/* first leaf entry to log */
	int		to;		/* target leaf index */

	leaf = bp->data;
	if (!leaf->hdr.stale) {
		return;
	}
	/*
	 * Compress out the stale entries in place.
	 */
615
	for (from = to = 0, loglow = -1; from < be16_to_cpu(leaf->hdr.count); from++) {
616
		if (be32_to_cpu(leaf->ents[from].address) == XFS_DIR2_NULL_DATAPTR)
L
Linus Torvalds 已提交
617 618 619 620 621 622 623 624 625 626 627 628 629 630
			continue;
		/*
		 * Only actually copy the entries that are different.
		 */
		if (from > to) {
			if (loglow == -1)
				loglow = to;
			leaf->ents[to] = leaf->ents[from];
		}
		to++;
	}
	/*
	 * Update and log the header, log the leaf entries.
	 */
631 632
	ASSERT(be16_to_cpu(leaf->hdr.stale) == from - to);
	be16_add(&leaf->hdr.count, -(be16_to_cpu(leaf->hdr.stale)));
L
Linus Torvalds 已提交
633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665
	leaf->hdr.stale = 0;
	xfs_dir2_leaf_log_header(args->trans, bp);
	if (loglow != -1)
		xfs_dir2_leaf_log_ents(args->trans, bp, loglow, to - 1);
}

/*
 * Compact the leaf entries, removing stale ones.
 * Leave one stale entry behind - the one closest to our
 * insertion index - and the caller will shift that one to our insertion
 * point later.
 * Return new insertion index, where the remaining stale entry is,
 * and leaf logging indices.
 */
void
xfs_dir2_leaf_compact_x1(
	xfs_dabuf_t	*bp,		/* leaf buffer */
	int		*indexp,	/* insertion index */
	int		*lowstalep,	/* out: stale entry before us */
	int		*highstalep,	/* out: stale entry after us */
	int		*lowlogp,	/* out: low log index */
	int		*highlogp)	/* out: high log index */
{
	int		from;		/* source copy index */
	int		highstale;	/* stale entry at/after index */
	int		index;		/* insertion index */
	int		keepstale;	/* source index of kept stale */
	xfs_dir2_leaf_t	*leaf;		/* leaf structure */
	int		lowstale;	/* stale entry before index */
	int		newindex=0;	/* new insertion index */
	int		to;		/* destination copy index */

	leaf = bp->data;
666
	ASSERT(be16_to_cpu(leaf->hdr.stale) > 1);
L
Linus Torvalds 已提交
667 668 669 670 671 672
	index = *indexp;
	/*
	 * Find the first stale entry before our index, if any.
	 */
	for (lowstale = index - 1;
	     lowstale >= 0 &&
673
		be32_to_cpu(leaf->ents[lowstale].address) != XFS_DIR2_NULL_DATAPTR;
L
Linus Torvalds 已提交
674 675 676 677 678 679 680
	     lowstale--)
		continue;
	/*
	 * Find the first stale entry at or after our index, if any.
	 * Stop if the answer would be worse than lowstale.
	 */
	for (highstale = index;
681
	     highstale < be16_to_cpu(leaf->hdr.count) &&
682
		be32_to_cpu(leaf->ents[highstale].address) != XFS_DIR2_NULL_DATAPTR &&
L
Linus Torvalds 已提交
683 684 685 686 687 688 689
		(lowstale < 0 || index - lowstale > highstale - index);
	     highstale++)
		continue;
	/*
	 * Pick the better of lowstale and highstale.
	 */
	if (lowstale >= 0 &&
690
	    (highstale == be16_to_cpu(leaf->hdr.count) ||
L
Linus Torvalds 已提交
691 692 693 694 695 696 697 698
	     index - lowstale <= highstale - index))
		keepstale = lowstale;
	else
		keepstale = highstale;
	/*
	 * Copy the entries in place, removing all the stale entries
	 * except keepstale.
	 */
699
	for (from = to = 0; from < be16_to_cpu(leaf->hdr.count); from++) {
L
Linus Torvalds 已提交
700 701 702 703 704 705
		/*
		 * Notice the new value of index.
		 */
		if (index == from)
			newindex = to;
		if (from != keepstale &&
706
		    be32_to_cpu(leaf->ents[from].address) == XFS_DIR2_NULL_DATAPTR) {
L
Linus Torvalds 已提交
707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733
			if (from == to)
				*lowlogp = to;
			continue;
		}
		/*
		 * Record the new keepstale value for the insertion.
		 */
		if (from == keepstale)
			lowstale = highstale = to;
		/*
		 * Copy only the entries that have moved.
		 */
		if (from > to)
			leaf->ents[to] = leaf->ents[from];
		to++;
	}
	ASSERT(from > to);
	/*
	 * If the insertion point was past the last entry,
	 * set the new insertion point accordingly.
	 */
	if (index == from)
		newindex = to;
	*indexp = newindex;
	/*
	 * Adjust the leaf header values.
	 */
734 735
	be16_add(&leaf->hdr.count, -(from - to));
	leaf->hdr.stale = cpu_to_be16(1);
L
Linus Torvalds 已提交
736 737 738 739 740 741 742
	/*
	 * Remember the low/high stale value only in the "right"
	 * direction.
	 */
	if (lowstale >= newindex)
		lowstale = -1;
	else
743 744
		highstale = be16_to_cpu(leaf->hdr.count);
	*highlogp = be16_to_cpu(leaf->hdr.count) - 1;
L
Linus Torvalds 已提交
745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769
	*lowstalep = lowstale;
	*highstalep = highstale;
}

/*
 * Getdents (readdir) for leaf and node directories.
 * This reads the data blocks only, so is the same for both forms.
 */
int						/* error */
xfs_dir2_leaf_getdents(
	xfs_trans_t		*tp,		/* transaction pointer */
	xfs_inode_t		*dp,		/* incore directory inode */
	uio_t			*uio,		/* I/O control & vectors */
	int			*eofp,		/* out: reached end of dir */
	xfs_dirent_t		*dbp,		/* caller's buffer */
	xfs_dir2_put_t		put)		/* ABI formatting routine */
{
	xfs_dabuf_t		*bp;		/* data block buffer */
	int			byteoff;	/* offset in current block */
	xfs_dir2_db_t		curdb;		/* db for current block */
	xfs_dir2_off_t		curoff;		/* current overall offset */
	xfs_dir2_data_t		*data;		/* data block structure */
	xfs_dir2_data_entry_t	*dep;		/* data entry */
	xfs_dir2_data_unused_t	*dup;		/* unused entry */
	int			eof;		/* reached end of directory */
770
	int			error = 0;	/* error return value */
L
Linus Torvalds 已提交
771 772 773 774 775 776 777 778 779 780 781
	int			i;		/* temporary loop index */
	int			j;		/* temporary loop index */
	int			length;		/* temporary length value */
	xfs_bmbt_irec_t		*map;		/* map vector for blocks */
	xfs_extlen_t		map_blocks;	/* number of fsbs in map */
	xfs_dablk_t		map_off;	/* last mapped file offset */
	int			map_size;	/* total entries in *map */
	int			map_valid;	/* valid entries in *map */
	xfs_mount_t		*mp;		/* filesystem mount point */
	xfs_dir2_off_t		newoff;		/* new curoff after new blk */
	int			nmap;		/* mappings to ask xfs_bmapi */
782 783
	xfs_dir2_put_args_t	*p;		/* formatting arg bundle */
	char			*ptr = NULL;	/* pointer to current data */
L
Linus Torvalds 已提交
784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800
	int			ra_current;	/* number of read-ahead blks */
	int			ra_index;	/* *map index for read-ahead */
	int			ra_offset;	/* map entry offset for ra */
	int			ra_want;	/* readahead count wanted */

	/*
	 * If the offset is at or past the largest allowed value,
	 * give up right away, return eof.
	 */
	if (uio->uio_offset >= XFS_DIR2_MAX_DATAPTR) {
		*eofp = 1;
		return 0;
	}
	mp = dp->i_mount;
	/*
	 * Setup formatting arguments.
	 */
801 802 803 804
	p = kmem_alloc(sizeof(*p), KM_SLEEP);
	p->dbp = dbp;
	p->put = put;
	p->uio = uio;
L
Linus Torvalds 已提交
805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886
	/*
	 * Set up to bmap a number of blocks based on the caller's
	 * buffer size, the directory block size, and the filesystem
	 * block size.
	 */
	map_size =
		howmany(uio->uio_resid + mp->m_dirblksize,
			mp->m_sb.sb_blocksize);
	map = kmem_alloc(map_size * sizeof(*map), KM_SLEEP);
	map_valid = ra_index = ra_offset = ra_current = map_blocks = 0;
	bp = NULL;
	eof = 1;
	/*
	 * Inside the loop we keep the main offset value as a byte offset
	 * in the directory file.
	 */
	curoff = XFS_DIR2_DATAPTR_TO_BYTE(mp, uio->uio_offset);
	/*
	 * Force this conversion through db so we truncate the offset
	 * down to get the start of the data block.
	 */
	map_off = XFS_DIR2_DB_TO_DA(mp, XFS_DIR2_BYTE_TO_DB(mp, curoff));
	/*
	 * Loop over directory entries until we reach the end offset.
	 * Get more blocks and readahead as necessary.
	 */
	while (curoff < XFS_DIR2_LEAF_OFFSET) {
		/*
		 * If we have no buffer, or we're off the end of the
		 * current buffer, need to get another one.
		 */
		if (!bp || ptr >= (char *)bp->data + mp->m_dirblksize) {
			/*
			 * If we have a buffer, we need to release it and
			 * take it out of the mapping.
			 */
			if (bp) {
				xfs_da_brelse(tp, bp);
				bp = NULL;
				map_blocks -= mp->m_dirblkfsbs;
				/*
				 * Loop to get rid of the extents for the
				 * directory block.
				 */
				for (i = mp->m_dirblkfsbs; i > 0; ) {
					j = MIN((int)map->br_blockcount, i);
					map->br_blockcount -= j;
					map->br_startblock += j;
					map->br_startoff += j;
					/*
					 * If mapping is done, pitch it from
					 * the table.
					 */
					if (!map->br_blockcount && --map_valid)
						memmove(&map[0], &map[1],
							sizeof(map[0]) *
							map_valid);
					i -= j;
				}
			}
			/*
			 * Recalculate the readahead blocks wanted.
			 */
			ra_want = howmany(uio->uio_resid + mp->m_dirblksize,
					  mp->m_sb.sb_blocksize) - 1;
			/*
			 * If we don't have as many as we want, and we haven't
			 * run out of data blocks, get some more mappings.
			 */
			if (1 + ra_want > map_blocks &&
			    map_off <
			    XFS_DIR2_BYTE_TO_DA(mp, XFS_DIR2_LEAF_OFFSET)) {
				/*
				 * Get more bmaps, fill in after the ones
				 * we already have in the table.
				 */
				nmap = map_size - map_valid;
				error = xfs_bmapi(tp, dp,
					map_off,
					XFS_DIR2_BYTE_TO_DA(mp,
						XFS_DIR2_LEAF_OFFSET) - map_off,
					XFS_BMAPI_METADATA, NULL, 0,
887
					&map[map_valid], &nmap, NULL, NULL);
L
Linus Torvalds 已提交
888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050
				/*
				 * Don't know if we should ignore this or
				 * try to return an error.
				 * The trouble with returning errors
				 * is that readdir will just stop without
				 * actually passing the error through.
				 */
				if (error)
					break;	/* XXX */
				/*
				 * If we got all the mappings we asked for,
				 * set the final map offset based on the
				 * last bmap value received.
				 * Otherwise, we've reached the end.
				 */
				if (nmap == map_size - map_valid)
					map_off =
					map[map_valid + nmap - 1].br_startoff +
					map[map_valid + nmap - 1].br_blockcount;
				else
					map_off =
						XFS_DIR2_BYTE_TO_DA(mp,
							XFS_DIR2_LEAF_OFFSET);
				/*
				 * Look for holes in the mapping, and
				 * eliminate them.  Count up the valid blocks.
				 */
				for (i = map_valid; i < map_valid + nmap; ) {
					if (map[i].br_startblock ==
					    HOLESTARTBLOCK) {
						nmap--;
						length = map_valid + nmap - i;
						if (length)
							memmove(&map[i],
								&map[i + 1],
								sizeof(map[i]) *
								length);
					} else {
						map_blocks +=
							map[i].br_blockcount;
						i++;
					}
				}
				map_valid += nmap;
			}
			/*
			 * No valid mappings, so no more data blocks.
			 */
			if (!map_valid) {
				curoff = XFS_DIR2_DA_TO_BYTE(mp, map_off);
				break;
			}
			/*
			 * Read the directory block starting at the first
			 * mapping.
			 */
			curdb = XFS_DIR2_DA_TO_DB(mp, map->br_startoff);
			error = xfs_da_read_buf(tp, dp, map->br_startoff,
				map->br_blockcount >= mp->m_dirblkfsbs ?
				    XFS_FSB_TO_DADDR(mp, map->br_startblock) :
				    -1,
				&bp, XFS_DATA_FORK);
			/*
			 * Should just skip over the data block instead
			 * of giving up.
			 */
			if (error)
				break;	/* XXX */
			/*
			 * Adjust the current amount of read-ahead: we just
			 * read a block that was previously ra.
			 */
			if (ra_current)
				ra_current -= mp->m_dirblkfsbs;
			/*
			 * Do we need more readahead?
			 */
			for (ra_index = ra_offset = i = 0;
			     ra_want > ra_current && i < map_blocks;
			     i += mp->m_dirblkfsbs) {
				ASSERT(ra_index < map_valid);
				/*
				 * Read-ahead a contiguous directory block.
				 */
				if (i > ra_current &&
				    map[ra_index].br_blockcount >=
				    mp->m_dirblkfsbs) {
					xfs_baread(mp->m_ddev_targp,
						XFS_FSB_TO_DADDR(mp,
						   map[ra_index].br_startblock +
						   ra_offset),
						(int)BTOBB(mp->m_dirblksize));
					ra_current = i;
				}
				/*
				 * Read-ahead a non-contiguous directory block.
				 * This doesn't use our mapping, but this
				 * is a very rare case.
				 */
				else if (i > ra_current) {
					(void)xfs_da_reada_buf(tp, dp,
						map[ra_index].br_startoff +
						ra_offset, XFS_DATA_FORK);
					ra_current = i;
				}
				/*
				 * Advance offset through the mapping table.
				 */
				for (j = 0; j < mp->m_dirblkfsbs; j++) {
					/*
					 * The rest of this extent but not
					 * more than a dir block.
					 */
					length = MIN(mp->m_dirblkfsbs,
						(int)(map[ra_index].br_blockcount -
						ra_offset));
					j += length;
					ra_offset += length;
					/*
					 * Advance to the next mapping if
					 * this one is used up.
					 */
					if (ra_offset ==
					    map[ra_index].br_blockcount) {
						ra_offset = 0;
						ra_index++;
					}
				}
			}
			/*
			 * Having done a read, we need to set a new offset.
			 */
			newoff = XFS_DIR2_DB_OFF_TO_BYTE(mp, curdb, 0);
			/*
			 * Start of the current block.
			 */
			if (curoff < newoff)
				curoff = newoff;
			/*
			 * Make sure we're in the right block.
			 */
			else if (curoff > newoff)
				ASSERT(XFS_DIR2_BYTE_TO_DB(mp, curoff) ==
				       curdb);
			data = bp->data;
			xfs_dir2_data_check(dp, bp);
			/*
			 * Find our position in the block.
			 */
			ptr = (char *)&data->u;
			byteoff = XFS_DIR2_BYTE_TO_OFF(mp, curoff);
			/*
			 * Skip past the header.
			 */
			if (byteoff == 0)
				curoff += (uint)sizeof(data->hdr);
			/*
			 * Skip past entries until we reach our offset.
			 */
			else {
				while ((char *)ptr - (char *)data < byteoff) {
					dup = (xfs_dir2_data_unused_t *)ptr;

1051
					if (be16_to_cpu(dup->freetag)
L
Linus Torvalds 已提交
1052 1053
						  == XFS_DIR2_DATA_FREE_TAG) {

1054
						length = be16_to_cpu(dup->length);
L
Linus Torvalds 已提交
1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082
						ptr += length;
						continue;
					}
					dep = (xfs_dir2_data_entry_t *)ptr;
					length =
					   XFS_DIR2_DATA_ENTSIZE(dep->namelen);
					ptr += length;
				}
				/*
				 * Now set our real offset.
				 */
				curoff =
					XFS_DIR2_DB_OFF_TO_BYTE(mp,
					    XFS_DIR2_BYTE_TO_DB(mp, curoff),
					    (char *)ptr - (char *)data);
				if (ptr >= (char *)data + mp->m_dirblksize) {
					continue;
				}
			}
		}
		/*
		 * We have a pointer to an entry.
		 * Is it a live one?
		 */
		dup = (xfs_dir2_data_unused_t *)ptr;
		/*
		 * No, it's unused, skip over it.
		 */
1083 1084
		if (be16_to_cpu(dup->freetag) == XFS_DIR2_DATA_FREE_TAG) {
			length = be16_to_cpu(dup->length);
L
Linus Torvalds 已提交
1085 1086 1087 1088 1089 1090 1091 1092 1093 1094
			ptr += length;
			curoff += length;
			continue;
		}

		/*
		 * Copy the entry into the putargs, and try formatting it.
		 */
		dep = (xfs_dir2_data_entry_t *)ptr;

1095
		p->namelen = dep->namelen;
L
Linus Torvalds 已提交
1096

1097
		length = XFS_DIR2_DATA_ENTSIZE(p->namelen);
L
Linus Torvalds 已提交
1098

1099
		p->cook = XFS_DIR2_BYTE_TO_DATAPTR(mp, curoff + length);
L
Linus Torvalds 已提交
1100

1101
		p->ino = be64_to_cpu(dep->inumber);
L
Linus Torvalds 已提交
1102
#if XFS_BIG_INUMS
1103
		p->ino += mp->m_inoadd;
L
Linus Torvalds 已提交
1104
#endif
1105
		p->name = (char *)dep->name;
L
Linus Torvalds 已提交
1106

1107
		error = p->put(p);
L
Linus Torvalds 已提交
1108 1109 1110 1111

		/*
		 * Won't fit.  Return to caller.
		 */
1112
		if (!p->done) {
L
Linus Torvalds 已提交
1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131
			eof = 0;
			break;
		}
		/*
		 * Advance to next entry in the block.
		 */
		ptr += length;
		curoff += length;
	}

	/*
	 * All done.  Set output offset value to current offset.
	 */
	*eofp = eof;
	if (curoff > XFS_DIR2_DATAPTR_TO_BYTE(mp, XFS_DIR2_MAX_DATAPTR))
		uio->uio_offset = XFS_DIR2_MAX_DATAPTR;
	else
		uio->uio_offset = XFS_DIR2_BYTE_TO_DATAPTR(mp, curoff);
	kmem_free(map, map_size * sizeof(*map));
1132
	kmem_free(p, sizeof(*p));
L
Linus Torvalds 已提交
1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174
	if (bp)
		xfs_da_brelse(tp, bp);
	return error;
}

/*
 * Initialize a new leaf block, leaf1 or leafn magic accepted.
 */
int
xfs_dir2_leaf_init(
	xfs_da_args_t		*args,		/* operation arguments */
	xfs_dir2_db_t		bno,		/* directory block number */
	xfs_dabuf_t		**bpp,		/* out: leaf buffer */
	int			magic)		/* magic number for block */
{
	xfs_dabuf_t		*bp;		/* leaf buffer */
	xfs_inode_t		*dp;		/* incore directory inode */
	int			error;		/* error return code */
	xfs_dir2_leaf_t		*leaf;		/* leaf structure */
	xfs_dir2_leaf_tail_t	*ltp;		/* leaf tail structure */
	xfs_mount_t		*mp;		/* filesystem mount point */
	xfs_trans_t		*tp;		/* transaction pointer */

	dp = args->dp;
	ASSERT(dp != NULL);
	tp = args->trans;
	mp = dp->i_mount;
	ASSERT(bno >= XFS_DIR2_LEAF_FIRSTDB(mp) &&
	       bno < XFS_DIR2_FREE_FIRSTDB(mp));
	/*
	 * Get the buffer for the block.
	 */
	error = xfs_da_get_buf(tp, dp, XFS_DIR2_DB_TO_DA(mp, bno), -1, &bp,
		XFS_DATA_FORK);
	if (error) {
		return error;
	}
	ASSERT(bp != NULL);
	leaf = bp->data;
	/*
	 * Initialize the header.
	 */
1175
	leaf->hdr.info.magic = cpu_to_be16(magic);
L
Linus Torvalds 已提交
1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197
	leaf->hdr.info.forw = 0;
	leaf->hdr.info.back = 0;
	leaf->hdr.count = 0;
	leaf->hdr.stale = 0;
	xfs_dir2_leaf_log_header(tp, bp);
	/*
	 * If it's a leaf-format directory initialize the tail.
	 * In this case our caller has the real bests table to copy into
	 * the block.
	 */
	if (magic == XFS_DIR2_LEAF1_MAGIC) {
		ltp = XFS_DIR2_LEAF_TAIL_P(mp, leaf);
		ltp->bestcount = 0;
		xfs_dir2_leaf_log_tail(tp, bp);
	}
	*bpp = bp;
	return 0;
}

/*
 * Log the bests entries indicated from a leaf1 block.
 */
1198
static void
L
Linus Torvalds 已提交
1199 1200 1201 1202 1203 1204
xfs_dir2_leaf_log_bests(
	xfs_trans_t		*tp,		/* transaction pointer */
	xfs_dabuf_t		*bp,		/* leaf buffer */
	int			first,		/* first entry to log */
	int			last)		/* last entry to log */
{
1205 1206
	__be16			*firstb;	/* pointer to first entry */
	__be16			*lastb;		/* pointer to last entry */
L
Linus Torvalds 已提交
1207 1208 1209 1210
	xfs_dir2_leaf_t		*leaf;		/* leaf structure */
	xfs_dir2_leaf_tail_t	*ltp;		/* leaf tail structure */

	leaf = bp->data;
1211
	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_DIR2_LEAF1_MAGIC);
L
Linus Torvalds 已提交
1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233
	ltp = XFS_DIR2_LEAF_TAIL_P(tp->t_mountp, leaf);
	firstb = XFS_DIR2_LEAF_BESTS_P(ltp) + first;
	lastb = XFS_DIR2_LEAF_BESTS_P(ltp) + last;
	xfs_da_log_buf(tp, bp, (uint)((char *)firstb - (char *)leaf),
		(uint)((char *)lastb - (char *)leaf + sizeof(*lastb) - 1));
}

/*
 * Log the leaf entries indicated from a leaf1 or leafn block.
 */
void
xfs_dir2_leaf_log_ents(
	xfs_trans_t		*tp,		/* transaction pointer */
	xfs_dabuf_t		*bp,		/* leaf buffer */
	int			first,		/* first entry to log */
	int			last)		/* last entry to log */
{
	xfs_dir2_leaf_entry_t	*firstlep;	/* pointer to first entry */
	xfs_dir2_leaf_entry_t	*lastlep;	/* pointer to last entry */
	xfs_dir2_leaf_t		*leaf;		/* leaf structure */

	leaf = bp->data;
1234 1235
	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_DIR2_LEAF1_MAGIC ||
	       be16_to_cpu(leaf->hdr.info.magic) == XFS_DIR2_LEAFN_MAGIC);
L
Linus Torvalds 已提交
1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252
	firstlep = &leaf->ents[first];
	lastlep = &leaf->ents[last];
	xfs_da_log_buf(tp, bp, (uint)((char *)firstlep - (char *)leaf),
		(uint)((char *)lastlep - (char *)leaf + sizeof(*lastlep) - 1));
}

/*
 * Log the header of the leaf1 or leafn block.
 */
void
xfs_dir2_leaf_log_header(
	xfs_trans_t		*tp,		/* transaction pointer */
	xfs_dabuf_t		*bp)		/* leaf buffer */
{
	xfs_dir2_leaf_t		*leaf;		/* leaf structure */

	leaf = bp->data;
1253 1254
	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_DIR2_LEAF1_MAGIC ||
	       be16_to_cpu(leaf->hdr.info.magic) == XFS_DIR2_LEAFN_MAGIC);
L
Linus Torvalds 已提交
1255 1256 1257 1258 1259 1260 1261
	xfs_da_log_buf(tp, bp, (uint)((char *)&leaf->hdr - (char *)leaf),
		(uint)(sizeof(leaf->hdr) - 1));
}

/*
 * Log the tail of the leaf1 block.
 */
1262
STATIC void
L
Linus Torvalds 已提交
1263 1264 1265 1266 1267 1268 1269 1270 1271 1272
xfs_dir2_leaf_log_tail(
	xfs_trans_t		*tp,		/* transaction pointer */
	xfs_dabuf_t		*bp)		/* leaf buffer */
{
	xfs_dir2_leaf_t		*leaf;		/* leaf structure */
	xfs_dir2_leaf_tail_t	*ltp;		/* leaf tail structure */
	xfs_mount_t		*mp;		/* filesystem mount point */

	mp = tp->t_mountp;
	leaf = bp->data;
1273
	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_DIR2_LEAF1_MAGIC);
L
Linus Torvalds 已提交
1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317
	ltp = XFS_DIR2_LEAF_TAIL_P(mp, leaf);
	xfs_da_log_buf(tp, bp, (uint)((char *)ltp - (char *)leaf),
		(uint)(mp->m_dirblksize - 1));
}

/*
 * Look up the entry referred to by args in the leaf format directory.
 * Most of the work is done by the xfs_dir2_leaf_lookup_int routine which
 * is also used by the node-format code.
 */
int
xfs_dir2_leaf_lookup(
	xfs_da_args_t		*args)		/* operation arguments */
{
	xfs_dabuf_t		*dbp;		/* data block buffer */
	xfs_dir2_data_entry_t	*dep;		/* data block entry */
	xfs_inode_t		*dp;		/* incore directory inode */
	int			error;		/* error return code */
	int			index;		/* found entry index */
	xfs_dabuf_t		*lbp;		/* leaf buffer */
	xfs_dir2_leaf_t		*leaf;		/* leaf structure */
	xfs_dir2_leaf_entry_t	*lep;		/* leaf entry */
	xfs_trans_t		*tp;		/* transaction pointer */

	xfs_dir2_trace_args("leaf_lookup", args);
	/*
	 * Look up name in the leaf block, returning both buffers and index.
	 */
	if ((error = xfs_dir2_leaf_lookup_int(args, &lbp, &index, &dbp))) {
		return error;
	}
	tp = args->trans;
	dp = args->dp;
	xfs_dir2_leaf_check(dp, lbp);
	leaf = lbp->data;
	/*
	 * Get to the leaf entry and contained data entry address.
	 */
	lep = &leaf->ents[index];
	/*
	 * Point to the data entry.
	 */
	dep = (xfs_dir2_data_entry_t *)
	      ((char *)dbp->data +
1318
	       XFS_DIR2_DATAPTR_TO_OFF(dp->i_mount, be32_to_cpu(lep->address)));
L
Linus Torvalds 已提交
1319 1320 1321
	/*
	 * Return the found inode number.
	 */
1322
	args->inumber = be64_to_cpu(dep->inumber);
L
Linus Torvalds 已提交
1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376
	xfs_da_brelse(tp, dbp);
	xfs_da_brelse(tp, lbp);
	return XFS_ERROR(EEXIST);
}

/*
 * Look up name/hash in the leaf block.
 * Fill in indexp with the found index, and dbpp with the data buffer.
 * If not found dbpp will be NULL, and ENOENT comes back.
 * lbpp will always be filled in with the leaf buffer unless there's an error.
 */
static int					/* error */
xfs_dir2_leaf_lookup_int(
	xfs_da_args_t		*args,		/* operation arguments */
	xfs_dabuf_t		**lbpp,		/* out: leaf buffer */
	int			*indexp,	/* out: index in leaf block */
	xfs_dabuf_t		**dbpp)		/* out: data buffer */
{
	xfs_dir2_db_t		curdb;		/* current data block number */
	xfs_dabuf_t		*dbp;		/* data buffer */
	xfs_dir2_data_entry_t	*dep;		/* data entry */
	xfs_inode_t		*dp;		/* incore directory inode */
	int			error;		/* error return code */
	int			index;		/* index in leaf block */
	xfs_dabuf_t		*lbp;		/* leaf buffer */
	xfs_dir2_leaf_entry_t	*lep;		/* leaf entry */
	xfs_dir2_leaf_t		*leaf;		/* leaf structure */
	xfs_mount_t		*mp;		/* filesystem mount point */
	xfs_dir2_db_t		newdb;		/* new data block number */
	xfs_trans_t		*tp;		/* transaction pointer */

	dp = args->dp;
	tp = args->trans;
	mp = dp->i_mount;
	/*
	 * Read the leaf block into the buffer.
	 */
	if ((error =
	    xfs_da_read_buf(tp, dp, mp->m_dirleafblk, -1, &lbp,
		    XFS_DATA_FORK))) {
		return error;
	}
	*lbpp = lbp;
	leaf = lbp->data;
	xfs_dir2_leaf_check(dp, lbp);
	/*
	 * Look for the first leaf entry with our hash value.
	 */
	index = xfs_dir2_leaf_search_hash(args, lbp);
	/*
	 * Loop over all the entries with the right hash value
	 * looking to match the name.
	 */
	for (lep = &leaf->ents[index], dbp = NULL, curdb = -1;
1377
	     index < be16_to_cpu(leaf->hdr.count) && be32_to_cpu(lep->hashval) == args->hashval;
L
Linus Torvalds 已提交
1378 1379 1380 1381
	     lep++, index++) {
		/*
		 * Skip over stale leaf entries.
		 */
1382
		if (be32_to_cpu(lep->address) == XFS_DIR2_NULL_DATAPTR)
L
Linus Torvalds 已提交
1383 1384 1385 1386
			continue;
		/*
		 * Get the new data block number.
		 */
1387
		newdb = XFS_DIR2_DATAPTR_TO_DB(mp, be32_to_cpu(lep->address));
L
Linus Torvalds 已提交
1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409
		/*
		 * If it's not the same as the old data block number,
		 * need to pitch the old one and read the new one.
		 */
		if (newdb != curdb) {
			if (dbp)
				xfs_da_brelse(tp, dbp);
			if ((error =
			    xfs_da_read_buf(tp, dp,
				    XFS_DIR2_DB_TO_DA(mp, newdb), -1, &dbp,
				    XFS_DATA_FORK))) {
				xfs_da_brelse(tp, lbp);
				return error;
			}
			xfs_dir2_data_check(dp, dbp);
			curdb = newdb;
		}
		/*
		 * Point to the data entry.
		 */
		dep = (xfs_dir2_data_entry_t *)
		      ((char *)dbp->data +
1410
		       XFS_DIR2_DATAPTR_TO_OFF(mp, be32_to_cpu(lep->address)));
L
Linus Torvalds 已提交
1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438
		/*
		 * If it matches then return it.
		 */
		if (dep->namelen == args->namelen &&
		    dep->name[0] == args->name[0] &&
		    memcmp(dep->name, args->name, args->namelen) == 0) {
			*dbpp = dbp;
			*indexp = index;
			return 0;
		}
	}
	/*
	 * No match found, return ENOENT.
	 */
	ASSERT(args->oknoent);
	if (dbp)
		xfs_da_brelse(tp, dbp);
	xfs_da_brelse(tp, lbp);
	return XFS_ERROR(ENOENT);
}

/*
 * Remove an entry from a leaf format directory.
 */
int						/* error */
xfs_dir2_leaf_removename(
	xfs_da_args_t		*args)		/* operation arguments */
{
1439
	__be16			*bestsp;	/* leaf block best freespace */
L
Linus Torvalds 已提交
1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474
	xfs_dir2_data_t		*data;		/* data block structure */
	xfs_dir2_db_t		db;		/* data block number */
	xfs_dabuf_t		*dbp;		/* data block buffer */
	xfs_dir2_data_entry_t	*dep;		/* data entry structure */
	xfs_inode_t		*dp;		/* incore directory inode */
	int			error;		/* error return code */
	xfs_dir2_db_t		i;		/* temporary data block # */
	int			index;		/* index into leaf entries */
	xfs_dabuf_t		*lbp;		/* leaf buffer */
	xfs_dir2_leaf_t		*leaf;		/* leaf structure */
	xfs_dir2_leaf_entry_t	*lep;		/* leaf entry */
	xfs_dir2_leaf_tail_t	*ltp;		/* leaf tail structure */
	xfs_mount_t		*mp;		/* filesystem mount point */
	int			needlog;	/* need to log data header */
	int			needscan;	/* need to rescan data frees */
	xfs_dir2_data_off_t	oldbest;	/* old value of best free */
	xfs_trans_t		*tp;		/* transaction pointer */

	xfs_dir2_trace_args("leaf_removename", args);
	/*
	 * Lookup the leaf entry, get the leaf and data blocks read in.
	 */
	if ((error = xfs_dir2_leaf_lookup_int(args, &lbp, &index, &dbp))) {
		return error;
	}
	dp = args->dp;
	tp = args->trans;
	mp = dp->i_mount;
	leaf = lbp->data;
	data = dbp->data;
	xfs_dir2_data_check(dp, dbp);
	/*
	 * Point to the leaf entry, use that to point to the data entry.
	 */
	lep = &leaf->ents[index];
1475
	db = XFS_DIR2_DATAPTR_TO_DB(mp, be32_to_cpu(lep->address));
L
Linus Torvalds 已提交
1476
	dep = (xfs_dir2_data_entry_t *)
1477
	      ((char *)data + XFS_DIR2_DATAPTR_TO_OFF(mp, be32_to_cpu(lep->address)));
L
Linus Torvalds 已提交
1478
	needscan = needlog = 0;
1479
	oldbest = be16_to_cpu(data->hdr.bestfree[0].length);
L
Linus Torvalds 已提交
1480 1481
	ltp = XFS_DIR2_LEAF_TAIL_P(mp, leaf);
	bestsp = XFS_DIR2_LEAF_BESTS_P(ltp);
1482
	ASSERT(be16_to_cpu(bestsp[db]) == oldbest);
L
Linus Torvalds 已提交
1483 1484 1485 1486 1487 1488 1489 1490 1491
	/*
	 * Mark the former data entry unused.
	 */
	xfs_dir2_data_make_free(tp, dbp,
		(xfs_dir2_data_aoff_t)((char *)dep - (char *)data),
		XFS_DIR2_DATA_ENTSIZE(dep->namelen), &needlog, &needscan);
	/*
	 * We just mark the leaf entry stale by putting a null in it.
	 */
1492
	be16_add(&leaf->hdr.stale, 1);
L
Linus Torvalds 已提交
1493
	xfs_dir2_leaf_log_header(tp, lbp);
1494
	lep->address = cpu_to_be32(XFS_DIR2_NULL_DATAPTR);
L
Linus Torvalds 已提交
1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507
	xfs_dir2_leaf_log_ents(tp, lbp, index, index);
	/*
	 * Scan the freespace in the data block again if necessary,
	 * log the data block header if necessary.
	 */
	if (needscan)
		xfs_dir2_data_freescan(mp, data, &needlog, NULL);
	if (needlog)
		xfs_dir2_data_log_header(tp, dbp);
	/*
	 * If the longest freespace in the data block has changed,
	 * put the new value in the bests table and log that.
	 */
1508 1509
	if (be16_to_cpu(data->hdr.bestfree[0].length) != oldbest) {
		bestsp[db] = data->hdr.bestfree[0].length;
L
Linus Torvalds 已提交
1510 1511 1512 1513 1514 1515
		xfs_dir2_leaf_log_bests(tp, lbp, db, db);
	}
	xfs_dir2_data_check(dp, dbp);
	/*
	 * If the data block is now empty then get rid of the data block.
	 */
1516
	if (be16_to_cpu(data->hdr.bestfree[0].length) ==
L
Linus Torvalds 已提交
1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538
	    mp->m_dirblksize - (uint)sizeof(data->hdr)) {
		ASSERT(db != mp->m_dirdatablk);
		if ((error = xfs_dir2_shrink_inode(args, db, dbp))) {
			/*
			 * Nope, can't get rid of it because it caused
			 * allocation of a bmap btree block to do so.
			 * Just go on, returning success, leaving the
			 * empty block in place.
			 */
			if (error == ENOSPC && args->total == 0) {
				xfs_da_buf_done(dbp);
				error = 0;
			}
			xfs_dir2_leaf_check(dp, lbp);
			xfs_da_buf_done(lbp);
			return error;
		}
		dbp = NULL;
		/*
		 * If this is the last data block then compact the
		 * bests table by getting rid of entries.
		 */
1539
		if (db == be32_to_cpu(ltp->bestcount) - 1) {
L
Linus Torvalds 已提交
1540 1541 1542 1543
			/*
			 * Look for the last active entry (i).
			 */
			for (i = db - 1; i > 0; i--) {
1544
				if (be16_to_cpu(bestsp[i]) != NULLDATAOFF)
L
Linus Torvalds 已提交
1545 1546 1547 1548 1549 1550 1551
					break;
			}
			/*
			 * Copy the table down so inactive entries at the
			 * end are removed.
			 */
			memmove(&bestsp[db - i], bestsp,
1552 1553
				(be32_to_cpu(ltp->bestcount) - (db - i)) * sizeof(*bestsp));
			be32_add(&ltp->bestcount, -(db - i));
L
Linus Torvalds 已提交
1554
			xfs_dir2_leaf_log_tail(tp, lbp);
1555
			xfs_dir2_leaf_log_bests(tp, lbp, 0, be32_to_cpu(ltp->bestcount) - 1);
L
Linus Torvalds 已提交
1556
		} else
1557
			bestsp[db] = cpu_to_be16(NULLDATAOFF);
L
Linus Torvalds 已提交
1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607
	}
	/*
	 * If the data block was not the first one, drop it.
	 */
	else if (db != mp->m_dirdatablk && dbp != NULL) {
		xfs_da_buf_done(dbp);
		dbp = NULL;
	}
	xfs_dir2_leaf_check(dp, lbp);
	/*
	 * See if we can convert to block form.
	 */
	return xfs_dir2_leaf_to_block(args, lbp, dbp);
}

/*
 * Replace the inode number in a leaf format directory entry.
 */
int						/* error */
xfs_dir2_leaf_replace(
	xfs_da_args_t		*args)		/* operation arguments */
{
	xfs_dabuf_t		*dbp;		/* data block buffer */
	xfs_dir2_data_entry_t	*dep;		/* data block entry */
	xfs_inode_t		*dp;		/* incore directory inode */
	int			error;		/* error return code */
	int			index;		/* index of leaf entry */
	xfs_dabuf_t		*lbp;		/* leaf buffer */
	xfs_dir2_leaf_t		*leaf;		/* leaf structure */
	xfs_dir2_leaf_entry_t	*lep;		/* leaf entry */
	xfs_trans_t		*tp;		/* transaction pointer */

	xfs_dir2_trace_args("leaf_replace", args);
	/*
	 * Look up the entry.
	 */
	if ((error = xfs_dir2_leaf_lookup_int(args, &lbp, &index, &dbp))) {
		return error;
	}
	dp = args->dp;
	leaf = lbp->data;
	/*
	 * Point to the leaf entry, get data address from it.
	 */
	lep = &leaf->ents[index];
	/*
	 * Point to the data entry.
	 */
	dep = (xfs_dir2_data_entry_t *)
	      ((char *)dbp->data +
1608
	       XFS_DIR2_DATAPTR_TO_OFF(dp->i_mount, be32_to_cpu(lep->address)));
1609
	ASSERT(args->inumber != be64_to_cpu(dep->inumber));
L
Linus Torvalds 已提交
1610 1611 1612
	/*
	 * Put the new inode number in, log it.
	 */
1613
	dep->inumber = cpu_to_be64(args->inumber);
L
Linus Torvalds 已提交
1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648
	tp = args->trans;
	xfs_dir2_data_log_entry(tp, dbp, dep);
	xfs_da_buf_done(dbp);
	xfs_dir2_leaf_check(dp, lbp);
	xfs_da_brelse(tp, lbp);
	return 0;
}

/*
 * Return index in the leaf block (lbp) which is either the first
 * one with this hash value, or if there are none, the insert point
 * for that hash value.
 */
int						/* index value */
xfs_dir2_leaf_search_hash(
	xfs_da_args_t		*args,		/* operation arguments */
	xfs_dabuf_t		*lbp)		/* leaf buffer */
{
	xfs_dahash_t		hash=0;		/* hash from this entry */
	xfs_dahash_t		hashwant;	/* hash value looking for */
	int			high;		/* high leaf index */
	int			low;		/* low leaf index */
	xfs_dir2_leaf_t		*leaf;		/* leaf structure */
	xfs_dir2_leaf_entry_t	*lep;		/* leaf entry */
	int			mid=0;		/* current leaf index */

	leaf = lbp->data;
#ifndef __KERNEL__
	if (!leaf->hdr.count)
		return 0;
#endif
	/*
	 * Note, the table cannot be empty, so we have to go through the loop.
	 * Binary search the leaf entries looking for our hash value.
	 */
1649
	for (lep = leaf->ents, low = 0, high = be16_to_cpu(leaf->hdr.count) - 1,
L
Linus Torvalds 已提交
1650 1651 1652
		hashwant = args->hashval;
	     low <= high; ) {
		mid = (low + high) >> 1;
1653
		if ((hash = be32_to_cpu(lep[mid].hashval)) == hashwant)
L
Linus Torvalds 已提交
1654 1655 1656 1657 1658 1659 1660 1661 1662 1663
			break;
		if (hash < hashwant)
			low = mid + 1;
		else
			high = mid - 1;
	}
	/*
	 * Found one, back up through all the equal hash values.
	 */
	if (hash == hashwant) {
1664
		while (mid > 0 && be32_to_cpu(lep[mid - 1].hashval) == hashwant) {
L
Linus Torvalds 已提交
1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685
			mid--;
		}
	}
	/*
	 * Need to point to an entry higher than ours.
	 */
	else if (hash < hashwant)
		mid++;
	return mid;
}

/*
 * Trim off a trailing data block.  We know it's empty since the leaf
 * freespace table says so.
 */
int						/* error */
xfs_dir2_leaf_trim_data(
	xfs_da_args_t		*args,		/* operation arguments */
	xfs_dabuf_t		*lbp,		/* leaf buffer */
	xfs_dir2_db_t		db)		/* data block number */
{
1686
	__be16			*bestsp;	/* leaf bests table */
L
Linus Torvalds 已提交
1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709
#ifdef DEBUG
	xfs_dir2_data_t		*data;		/* data block structure */
#endif
	xfs_dabuf_t		*dbp;		/* data block buffer */
	xfs_inode_t		*dp;		/* incore directory inode */
	int			error;		/* error return value */
	xfs_dir2_leaf_t		*leaf;		/* leaf structure */
	xfs_dir2_leaf_tail_t	*ltp;		/* leaf tail structure */
	xfs_mount_t		*mp;		/* filesystem mount point */
	xfs_trans_t		*tp;		/* transaction pointer */

	dp = args->dp;
	mp = dp->i_mount;
	tp = args->trans;
	/*
	 * Read the offending data block.  We need its buffer.
	 */
	if ((error = xfs_da_read_buf(tp, dp, XFS_DIR2_DB_TO_DA(mp, db), -1, &dbp,
			XFS_DATA_FORK))) {
		return error;
	}
#ifdef DEBUG
	data = dbp->data;
1710
	ASSERT(be32_to_cpu(data->hdr.magic) == XFS_DIR2_DATA_MAGIC);
L
Linus Torvalds 已提交
1711 1712 1713 1714 1715 1716 1717 1718
#endif
	/* this seems to be an error
	 * data is only valid if DEBUG is defined?
	 * RMC 09/08/1999
	 */

	leaf = lbp->data;
	ltp = XFS_DIR2_LEAF_TAIL_P(mp, leaf);
1719
	ASSERT(be16_to_cpu(data->hdr.bestfree[0].length) ==
L
Linus Torvalds 已提交
1720
	       mp->m_dirblksize - (uint)sizeof(data->hdr));
1721
	ASSERT(db == be32_to_cpu(ltp->bestcount) - 1);
L
Linus Torvalds 已提交
1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733
	/*
	 * Get rid of the data block.
	 */
	if ((error = xfs_dir2_shrink_inode(args, db, dbp))) {
		ASSERT(error != ENOSPC);
		xfs_da_brelse(tp, dbp);
		return error;
	}
	/*
	 * Eliminate the last bests entry from the table.
	 */
	bestsp = XFS_DIR2_LEAF_BESTS_P(ltp);
1734 1735
	be32_add(&ltp->bestcount, -1);
	memmove(&bestsp[1], &bestsp[0], be32_to_cpu(ltp->bestcount) * sizeof(*bestsp));
L
Linus Torvalds 已提交
1736
	xfs_dir2_leaf_log_tail(tp, lbp);
1737
	xfs_dir2_leaf_log_bests(tp, lbp, 0, be32_to_cpu(ltp->bestcount) - 1);
L
Linus Torvalds 已提交
1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808
	return 0;
}

/*
 * Convert node form directory to leaf form directory.
 * The root of the node form dir needs to already be a LEAFN block.
 * Just return if we can't do anything.
 */
int						/* error */
xfs_dir2_node_to_leaf(
	xfs_da_state_t		*state)		/* directory operation state */
{
	xfs_da_args_t		*args;		/* operation arguments */
	xfs_inode_t		*dp;		/* incore directory inode */
	int			error;		/* error return code */
	xfs_dabuf_t		*fbp;		/* buffer for freespace block */
	xfs_fileoff_t		fo;		/* freespace file offset */
	xfs_dir2_free_t		*free;		/* freespace structure */
	xfs_dabuf_t		*lbp;		/* buffer for leaf block */
	xfs_dir2_leaf_tail_t	*ltp;		/* tail of leaf structure */
	xfs_dir2_leaf_t		*leaf;		/* leaf structure */
	xfs_mount_t		*mp;		/* filesystem mount point */
	int			rval;		/* successful free trim? */
	xfs_trans_t		*tp;		/* transaction pointer */

	/*
	 * There's more than a leaf level in the btree, so there must
	 * be multiple leafn blocks.  Give up.
	 */
	if (state->path.active > 1)
		return 0;
	args = state->args;
	xfs_dir2_trace_args("node_to_leaf", args);
	mp = state->mp;
	dp = args->dp;
	tp = args->trans;
	/*
	 * Get the last offset in the file.
	 */
	if ((error = xfs_bmap_last_offset(tp, dp, &fo, XFS_DATA_FORK))) {
		return error;
	}
	fo -= mp->m_dirblkfsbs;
	/*
	 * If there are freespace blocks other than the first one,
	 * take this opportunity to remove trailing empty freespace blocks
	 * that may have been left behind during no-space-reservation
	 * operations.
	 */
	while (fo > mp->m_dirfreeblk) {
		if ((error = xfs_dir2_node_trim_free(args, fo, &rval))) {
			return error;
		}
		if (rval)
			fo -= mp->m_dirblkfsbs;
		else
			return 0;
	}
	/*
	 * Now find the block just before the freespace block.
	 */
	if ((error = xfs_bmap_last_before(tp, dp, &fo, XFS_DATA_FORK))) {
		return error;
	}
	/*
	 * If it's not the single leaf block, give up.
	 */
	if (XFS_FSB_TO_B(mp, fo) > XFS_DIR2_LEAF_OFFSET + mp->m_dirblksize)
		return 0;
	lbp = state->path.blk[0].bp;
	leaf = lbp->data;
1809
	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_DIR2_LEAFN_MAGIC);
L
Linus Torvalds 已提交
1810 1811 1812 1813 1814 1815 1816 1817
	/*
	 * Read the freespace block.
	 */
	if ((error = xfs_da_read_buf(tp, dp, mp->m_dirfreeblk, -1, &fbp,
			XFS_DATA_FORK))) {
		return error;
	}
	free = fbp->data;
1818
	ASSERT(be32_to_cpu(free->hdr.magic) == XFS_DIR2_FREE_MAGIC);
L
Linus Torvalds 已提交
1819 1820 1821 1822 1823 1824
	ASSERT(!free->hdr.firstdb);
	/*
	 * Now see if the leafn and free data will fit in a leaf1.
	 * If not, release the buffer and give up.
	 */
	if ((uint)sizeof(leaf->hdr) +
1825
	    (be16_to_cpu(leaf->hdr.count) - be16_to_cpu(leaf->hdr.stale)) * (uint)sizeof(leaf->ents[0]) +
1826
	    be32_to_cpu(free->hdr.nvalid) * (uint)sizeof(leaf->bests[0]) +
L
Linus Torvalds 已提交
1827 1828 1829 1830 1831 1832 1833 1834 1835
	    (uint)sizeof(leaf->tail) >
	    mp->m_dirblksize) {
		xfs_da_brelse(tp, fbp);
		return 0;
	}
	/*
	 * If the leaf has any stale entries in it, compress them out.
	 * The compact routine will log the header.
	 */
1836
	if (be16_to_cpu(leaf->hdr.stale))
L
Linus Torvalds 已提交
1837 1838 1839
		xfs_dir2_leaf_compact(args, lbp);
	else
		xfs_dir2_leaf_log_header(tp, lbp);
1840
	leaf->hdr.info.magic = cpu_to_be16(XFS_DIR2_LEAF1_MAGIC);
L
Linus Torvalds 已提交
1841 1842 1843 1844
	/*
	 * Set up the leaf tail from the freespace block.
	 */
	ltp = XFS_DIR2_LEAF_TAIL_P(mp, leaf);
1845
	ltp->bestcount = free->hdr.nvalid;
L
Linus Torvalds 已提交
1846 1847 1848 1849
	/*
	 * Set up the leaf bests table.
	 */
	memcpy(XFS_DIR2_LEAF_BESTS_P(ltp), free->bests,
1850 1851
		be32_to_cpu(ltp->bestcount) * sizeof(leaf->bests[0]));
	xfs_dir2_leaf_log_bests(tp, lbp, 0, be32_to_cpu(ltp->bestcount) - 1);
L
Linus Torvalds 已提交
1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877
	xfs_dir2_leaf_log_tail(tp, lbp);
	xfs_dir2_leaf_check(dp, lbp);
	/*
	 * Get rid of the freespace block.
	 */
	error = xfs_dir2_shrink_inode(args, XFS_DIR2_FREE_FIRSTDB(mp), fbp);
	if (error) {
		/*
		 * This can't fail here because it can only happen when
		 * punching out the middle of an extent, and this is an
		 * isolated block.
		 */
		ASSERT(error != ENOSPC);
		return error;
	}
	fbp = NULL;
	/*
	 * Now see if we can convert the single-leaf directory
	 * down to a block form directory.
	 * This routine always kills the dabuf for the leaf, so
	 * eliminate it from the path.
	 */
	error = xfs_dir2_leaf_to_block(args, lbp, NULL);
	state->path.blk[0].bp = NULL;
	return error;
}