dir.c 52.4 KB
Newer Older
Linus Torvalds's avatar
Linus Torvalds committed
1 2 3
/**
 * dir.c - NTFS kernel directory operations. Part of the Linux-NTFS project.
 *
4
 * Copyright (c) 2001-2007 Anton Altaparmakov
Linus Torvalds's avatar
Linus Torvalds committed
5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34
 * Copyright (c) 2002 Richard Russon
 *
 * This program/include file is free software; you can redistribute it and/or
 * modify it under the terms of the GNU General Public License as published
 * by the Free Software Foundation; either version 2 of the License, or
 * (at your option) any later version.
 *
 * This program/include file is distributed in the hope that it will 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.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program (in the main directory of the Linux-NTFS
 * distribution in the file COPYING); if not, write to the Free Software
 * Foundation,Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
 */

#include <linux/buffer_head.h>

#include "dir.h"
#include "aops.h"
#include "attrib.h"
#include "mft.h"
#include "debug.h"
#include "ntfs.h"

/**
 * The little endian Unicode string $I30 as a global constant.
 */
35 36
ntfschar I30[5] = { cpu_to_le16('$'), cpu_to_le16('I'),
		cpu_to_le16('3'),	cpu_to_le16('0'), 0 };
Linus Torvalds's avatar
Linus Torvalds committed
37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70

/**
 * ntfs_lookup_inode_by_name - find an inode in a directory given its name
 * @dir_ni:	ntfs inode of the directory in which to search for the name
 * @uname:	Unicode name for which to search in the directory
 * @uname_len:	length of the name @uname in Unicode characters
 * @res:	return the found file name if necessary (see below)
 *
 * Look for an inode with name @uname in the directory with inode @dir_ni.
 * ntfs_lookup_inode_by_name() walks the contents of the directory looking for
 * the Unicode name. If the name is found in the directory, the corresponding
 * inode number (>= 0) is returned as a mft reference in cpu format, i.e. it
 * is a 64-bit number containing the sequence number.
 *
 * On error, a negative value is returned corresponding to the error code. In
 * particular if the inode is not found -ENOENT is returned. Note that you
 * can't just check the return value for being negative, you have to check the
 * inode number for being negative which you can extract using MREC(return
 * value).
 *
 * Note, @uname_len does not include the (optional) terminating NULL character.
 *
 * Note, we look for a case sensitive match first but we also look for a case
 * insensitive match at the same time. If we find a case insensitive match, we
 * save that for the case that we don't find an exact match, where we return
 * the case insensitive match and setup @res (which we allocate!) with the mft
 * reference, the file name type, length and with a copy of the little endian
 * Unicode file name itself. If we match a file name which is in the DOS name
 * space, we only return the mft reference and file name type in @res.
 * ntfs_lookup() then uses this to find the long file name in the inode itself.
 * This is to avoid polluting the dcache with short file names. We want them to
 * work but we don't care for how quickly one can access them. This also fixes
 * the dcache aliasing issues.
 *
71
 * Locking:  - Caller must hold i_mutex on the directory.
Linus Torvalds's avatar
Linus Torvalds committed
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184
 *	     - Each page cache page in the index allocation mapping must be
 *	       locked whilst being accessed otherwise we may find a corrupt
 *	       page due to it being under ->writepage at the moment which
 *	       applies the mst protection fixups before writing out and then
 *	       removes them again after the write is complete after which it 
 *	       unlocks the page.
 */
MFT_REF ntfs_lookup_inode_by_name(ntfs_inode *dir_ni, const ntfschar *uname,
		const int uname_len, ntfs_name **res)
{
	ntfs_volume *vol = dir_ni->vol;
	struct super_block *sb = vol->sb;
	MFT_RECORD *m;
	INDEX_ROOT *ir;
	INDEX_ENTRY *ie;
	INDEX_ALLOCATION *ia;
	u8 *index_end;
	u64 mref;
	ntfs_attr_search_ctx *ctx;
	int err, rc;
	VCN vcn, old_vcn;
	struct address_space *ia_mapping;
	struct page *page;
	u8 *kaddr;
	ntfs_name *name = NULL;

	BUG_ON(!S_ISDIR(VFS_I(dir_ni)->i_mode));
	BUG_ON(NInoAttr(dir_ni));
	/* Get hold of the mft record for the directory. */
	m = map_mft_record(dir_ni);
	if (IS_ERR(m)) {
		ntfs_error(sb, "map_mft_record() failed with error code %ld.",
				-PTR_ERR(m));
		return ERR_MREF(PTR_ERR(m));
	}
	ctx = ntfs_attr_get_search_ctx(dir_ni, m);
	if (unlikely(!ctx)) {
		err = -ENOMEM;
		goto err_out;
	}
	/* Find the index root attribute in the mft record. */
	err = ntfs_attr_lookup(AT_INDEX_ROOT, I30, 4, CASE_SENSITIVE, 0, NULL,
			0, ctx);
	if (unlikely(err)) {
		if (err == -ENOENT) {
			ntfs_error(sb, "Index root attribute missing in "
					"directory inode 0x%lx.",
					dir_ni->mft_no);
			err = -EIO;
		}
		goto err_out;
	}
	/* Get to the index root value (it's been verified in read_inode). */
	ir = (INDEX_ROOT*)((u8*)ctx->attr +
			le16_to_cpu(ctx->attr->data.resident.value_offset));
	index_end = (u8*)&ir->index + le32_to_cpu(ir->index.index_length);
	/* The first index entry. */
	ie = (INDEX_ENTRY*)((u8*)&ir->index +
			le32_to_cpu(ir->index.entries_offset));
	/*
	 * Loop until we exceed valid memory (corruption case) or until we
	 * reach the last entry.
	 */
	for (;; ie = (INDEX_ENTRY*)((u8*)ie + le16_to_cpu(ie->length))) {
		/* Bounds checks. */
		if ((u8*)ie < (u8*)ctx->mrec || (u8*)ie +
				sizeof(INDEX_ENTRY_HEADER) > index_end ||
				(u8*)ie + le16_to_cpu(ie->key_length) >
				index_end)
			goto dir_err_out;
		/*
		 * The last entry cannot contain a name. It can however contain
		 * a pointer to a child node in the B+tree so we just break out.
		 */
		if (ie->flags & INDEX_ENTRY_END)
			break;
		/*
		 * We perform a case sensitive comparison and if that matches
		 * we are done and return the mft reference of the inode (i.e.
		 * the inode number together with the sequence number for
		 * consistency checking). We convert it to cpu format before
		 * returning.
		 */
		if (ntfs_are_names_equal(uname, uname_len,
				(ntfschar*)&ie->key.file_name.file_name,
				ie->key.file_name.file_name_length,
				CASE_SENSITIVE, vol->upcase, vol->upcase_len)) {
found_it:
			/*
			 * We have a perfect match, so we don't need to care
			 * about having matched imperfectly before, so we can
			 * free name and set *res to NULL.
			 * However, if the perfect match is a short file name,
			 * we need to signal this through *res, so that
			 * ntfs_lookup() can fix dcache aliasing issues.
			 * As an optimization we just reuse an existing
			 * allocation of *res.
			 */
			if (ie->key.file_name.file_name_type == FILE_NAME_DOS) {
				if (!name) {
					name = kmalloc(sizeof(ntfs_name),
							GFP_NOFS);
					if (!name) {
						err = -ENOMEM;
						goto err_out;
					}
				}
				name->mref = le64_to_cpu(
						ie->data.dir.indexed_file);
				name->type = FILE_NAME_DOS;
				name->len = 0;
				*res = name;
			} else {
185
				kfree(name);
Linus Torvalds's avatar
Linus Torvalds committed
186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444
				*res = NULL;
			}
			mref = le64_to_cpu(ie->data.dir.indexed_file);
			ntfs_attr_put_search_ctx(ctx);
			unmap_mft_record(dir_ni);
			return mref;
		}
		/*
		 * For a case insensitive mount, we also perform a case
		 * insensitive comparison (provided the file name is not in the
		 * POSIX namespace). If the comparison matches, and the name is
		 * in the WIN32 namespace, we cache the filename in *res so
		 * that the caller, ntfs_lookup(), can work on it. If the
		 * comparison matches, and the name is in the DOS namespace, we
		 * only cache the mft reference and the file name type (we set
		 * the name length to zero for simplicity).
		 */
		if (!NVolCaseSensitive(vol) &&
				ie->key.file_name.file_name_type &&
				ntfs_are_names_equal(uname, uname_len,
				(ntfschar*)&ie->key.file_name.file_name,
				ie->key.file_name.file_name_length,
				IGNORE_CASE, vol->upcase, vol->upcase_len)) {
			int name_size = sizeof(ntfs_name);
			u8 type = ie->key.file_name.file_name_type;
			u8 len = ie->key.file_name.file_name_length;

			/* Only one case insensitive matching name allowed. */
			if (name) {
				ntfs_error(sb, "Found already allocated name "
						"in phase 1. Please run chkdsk "
						"and if that doesn't find any "
						"errors please report you saw "
						"this message to "
						"linux-ntfs-dev@lists."
						"sourceforge.net.");
				goto dir_err_out;
			}

			if (type != FILE_NAME_DOS)
				name_size += len * sizeof(ntfschar);
			name = kmalloc(name_size, GFP_NOFS);
			if (!name) {
				err = -ENOMEM;
				goto err_out;
			}
			name->mref = le64_to_cpu(ie->data.dir.indexed_file);
			name->type = type;
			if (type != FILE_NAME_DOS) {
				name->len = len;
				memcpy(name->name, ie->key.file_name.file_name,
						len * sizeof(ntfschar));
			} else
				name->len = 0;
			*res = name;
		}
		/*
		 * Not a perfect match, need to do full blown collation so we
		 * know which way in the B+tree we have to go.
		 */
		rc = ntfs_collate_names(uname, uname_len,
				(ntfschar*)&ie->key.file_name.file_name,
				ie->key.file_name.file_name_length, 1,
				IGNORE_CASE, vol->upcase, vol->upcase_len);
		/*
		 * If uname collates before the name of the current entry, there
		 * is definitely no such name in this index but we might need to
		 * descend into the B+tree so we just break out of the loop.
		 */
		if (rc == -1)
			break;
		/* The names are not equal, continue the search. */
		if (rc)
			continue;
		/*
		 * Names match with case insensitive comparison, now try the
		 * case sensitive comparison, which is required for proper
		 * collation.
		 */
		rc = ntfs_collate_names(uname, uname_len,
				(ntfschar*)&ie->key.file_name.file_name,
				ie->key.file_name.file_name_length, 1,
				CASE_SENSITIVE, vol->upcase, vol->upcase_len);
		if (rc == -1)
			break;
		if (rc)
			continue;
		/*
		 * Perfect match, this will never happen as the
		 * ntfs_are_names_equal() call will have gotten a match but we
		 * still treat it correctly.
		 */
		goto found_it;
	}
	/*
	 * We have finished with this index without success. Check for the
	 * presence of a child node and if not present return -ENOENT, unless
	 * we have got a matching name cached in name in which case return the
	 * mft reference associated with it.
	 */
	if (!(ie->flags & INDEX_ENTRY_NODE)) {
		if (name) {
			ntfs_attr_put_search_ctx(ctx);
			unmap_mft_record(dir_ni);
			return name->mref;
		}
		ntfs_debug("Entry not found.");
		err = -ENOENT;
		goto err_out;
	} /* Child node present, descend into it. */
	/* Consistency check: Verify that an index allocation exists. */
	if (!NInoIndexAllocPresent(dir_ni)) {
		ntfs_error(sb, "No index allocation attribute but index entry "
				"requires one. Directory inode 0x%lx is "
				"corrupt or driver bug.", dir_ni->mft_no);
		goto err_out;
	}
	/* Get the starting vcn of the index_block holding the child node. */
	vcn = sle64_to_cpup((sle64*)((u8*)ie + le16_to_cpu(ie->length) - 8));
	ia_mapping = VFS_I(dir_ni)->i_mapping;
	/*
	 * We are done with the index root and the mft record. Release them,
	 * otherwise we deadlock with ntfs_map_page().
	 */
	ntfs_attr_put_search_ctx(ctx);
	unmap_mft_record(dir_ni);
	m = NULL;
	ctx = NULL;
descend_into_child_node:
	/*
	 * Convert vcn to index into the index allocation attribute in units
	 * of PAGE_CACHE_SIZE and map the page cache page, reading it from
	 * disk if necessary.
	 */
	page = ntfs_map_page(ia_mapping, vcn <<
			dir_ni->itype.index.vcn_size_bits >> PAGE_CACHE_SHIFT);
	if (IS_ERR(page)) {
		ntfs_error(sb, "Failed to map directory index page, error %ld.",
				-PTR_ERR(page));
		err = PTR_ERR(page);
		goto err_out;
	}
	lock_page(page);
	kaddr = (u8*)page_address(page);
fast_descend_into_child_node:
	/* Get to the index allocation block. */
	ia = (INDEX_ALLOCATION*)(kaddr + ((vcn <<
			dir_ni->itype.index.vcn_size_bits) & ~PAGE_CACHE_MASK));
	/* Bounds checks. */
	if ((u8*)ia < kaddr || (u8*)ia > kaddr + PAGE_CACHE_SIZE) {
		ntfs_error(sb, "Out of bounds check failed. Corrupt directory "
				"inode 0x%lx or driver bug.", dir_ni->mft_no);
		goto unm_err_out;
	}
	/* Catch multi sector transfer fixup errors. */
	if (unlikely(!ntfs_is_indx_record(ia->magic))) {
		ntfs_error(sb, "Directory index record with vcn 0x%llx is "
				"corrupt.  Corrupt inode 0x%lx.  Run chkdsk.",
				(unsigned long long)vcn, dir_ni->mft_no);
		goto unm_err_out;
	}
	if (sle64_to_cpu(ia->index_block_vcn) != vcn) {
		ntfs_error(sb, "Actual VCN (0x%llx) of index buffer is "
				"different from expected VCN (0x%llx). "
				"Directory inode 0x%lx is corrupt or driver "
				"bug.", (unsigned long long)
				sle64_to_cpu(ia->index_block_vcn),
				(unsigned long long)vcn, dir_ni->mft_no);
		goto unm_err_out;
	}
	if (le32_to_cpu(ia->index.allocated_size) + 0x18 !=
			dir_ni->itype.index.block_size) {
		ntfs_error(sb, "Index buffer (VCN 0x%llx) of directory inode "
				"0x%lx has a size (%u) differing from the "
				"directory specified size (%u). Directory "
				"inode is corrupt or driver bug.",
				(unsigned long long)vcn, dir_ni->mft_no,
				le32_to_cpu(ia->index.allocated_size) + 0x18,
				dir_ni->itype.index.block_size);
		goto unm_err_out;
	}
	index_end = (u8*)ia + dir_ni->itype.index.block_size;
	if (index_end > kaddr + PAGE_CACHE_SIZE) {
		ntfs_error(sb, "Index buffer (VCN 0x%llx) of directory inode "
				"0x%lx crosses page boundary. Impossible! "
				"Cannot access! This is probably a bug in the "
				"driver.", (unsigned long long)vcn,
				dir_ni->mft_no);
		goto unm_err_out;
	}
	index_end = (u8*)&ia->index + le32_to_cpu(ia->index.index_length);
	if (index_end > (u8*)ia + dir_ni->itype.index.block_size) {
		ntfs_error(sb, "Size of index buffer (VCN 0x%llx) of directory "
				"inode 0x%lx exceeds maximum size.",
				(unsigned long long)vcn, dir_ni->mft_no);
		goto unm_err_out;
	}
	/* The first index entry. */
	ie = (INDEX_ENTRY*)((u8*)&ia->index +
			le32_to_cpu(ia->index.entries_offset));
	/*
	 * Iterate similar to above big loop but applied to index buffer, thus
	 * loop until we exceed valid memory (corruption case) or until we
	 * reach the last entry.
	 */
	for (;; ie = (INDEX_ENTRY*)((u8*)ie + le16_to_cpu(ie->length))) {
		/* Bounds check. */
		if ((u8*)ie < (u8*)ia || (u8*)ie +
				sizeof(INDEX_ENTRY_HEADER) > index_end ||
				(u8*)ie + le16_to_cpu(ie->key_length) >
				index_end) {
			ntfs_error(sb, "Index entry out of bounds in "
					"directory inode 0x%lx.",
					dir_ni->mft_no);
			goto unm_err_out;
		}
		/*
		 * The last entry cannot contain a name. It can however contain
		 * a pointer to a child node in the B+tree so we just break out.
		 */
		if (ie->flags & INDEX_ENTRY_END)
			break;
		/*
		 * We perform a case sensitive comparison and if that matches
		 * we are done and return the mft reference of the inode (i.e.
		 * the inode number together with the sequence number for
		 * consistency checking). We convert it to cpu format before
		 * returning.
		 */
		if (ntfs_are_names_equal(uname, uname_len,
				(ntfschar*)&ie->key.file_name.file_name,
				ie->key.file_name.file_name_length,
				CASE_SENSITIVE, vol->upcase, vol->upcase_len)) {
found_it2:
			/*
			 * We have a perfect match, so we don't need to care
			 * about having matched imperfectly before, so we can
			 * free name and set *res to NULL.
			 * However, if the perfect match is a short file name,
			 * we need to signal this through *res, so that
			 * ntfs_lookup() can fix dcache aliasing issues.
			 * As an optimization we just reuse an existing
			 * allocation of *res.
			 */
			if (ie->key.file_name.file_name_type == FILE_NAME_DOS) {
				if (!name) {
					name = kmalloc(sizeof(ntfs_name),
							GFP_NOFS);
					if (!name) {
						err = -ENOMEM;
						goto unm_err_out;
					}
				}
				name->mref = le64_to_cpu(
						ie->data.dir.indexed_file);
				name->type = FILE_NAME_DOS;
				name->len = 0;
				*res = name;
			} else {
445
				kfree(name);
Linus Torvalds's avatar
Linus Torvalds committed
446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609
				*res = NULL;
			}
			mref = le64_to_cpu(ie->data.dir.indexed_file);
			unlock_page(page);
			ntfs_unmap_page(page);
			return mref;
		}
		/*
		 * For a case insensitive mount, we also perform a case
		 * insensitive comparison (provided the file name is not in the
		 * POSIX namespace). If the comparison matches, and the name is
		 * in the WIN32 namespace, we cache the filename in *res so
		 * that the caller, ntfs_lookup(), can work on it. If the
		 * comparison matches, and the name is in the DOS namespace, we
		 * only cache the mft reference and the file name type (we set
		 * the name length to zero for simplicity).
		 */
		if (!NVolCaseSensitive(vol) &&
				ie->key.file_name.file_name_type &&
				ntfs_are_names_equal(uname, uname_len,
				(ntfschar*)&ie->key.file_name.file_name,
				ie->key.file_name.file_name_length,
				IGNORE_CASE, vol->upcase, vol->upcase_len)) {
			int name_size = sizeof(ntfs_name);
			u8 type = ie->key.file_name.file_name_type;
			u8 len = ie->key.file_name.file_name_length;

			/* Only one case insensitive matching name allowed. */
			if (name) {
				ntfs_error(sb, "Found already allocated name "
						"in phase 2. Please run chkdsk "
						"and if that doesn't find any "
						"errors please report you saw "
						"this message to "
						"linux-ntfs-dev@lists."
						"sourceforge.net.");
				unlock_page(page);
				ntfs_unmap_page(page);
				goto dir_err_out;
			}

			if (type != FILE_NAME_DOS)
				name_size += len * sizeof(ntfschar);
			name = kmalloc(name_size, GFP_NOFS);
			if (!name) {
				err = -ENOMEM;
				goto unm_err_out;
			}
			name->mref = le64_to_cpu(ie->data.dir.indexed_file);
			name->type = type;
			if (type != FILE_NAME_DOS) {
				name->len = len;
				memcpy(name->name, ie->key.file_name.file_name,
						len * sizeof(ntfschar));
			} else
				name->len = 0;
			*res = name;
		}
		/*
		 * Not a perfect match, need to do full blown collation so we
		 * know which way in the B+tree we have to go.
		 */
		rc = ntfs_collate_names(uname, uname_len,
				(ntfschar*)&ie->key.file_name.file_name,
				ie->key.file_name.file_name_length, 1,
				IGNORE_CASE, vol->upcase, vol->upcase_len);
		/*
		 * If uname collates before the name of the current entry, there
		 * is definitely no such name in this index but we might need to
		 * descend into the B+tree so we just break out of the loop.
		 */
		if (rc == -1)
			break;
		/* The names are not equal, continue the search. */
		if (rc)
			continue;
		/*
		 * Names match with case insensitive comparison, now try the
		 * case sensitive comparison, which is required for proper
		 * collation.
		 */
		rc = ntfs_collate_names(uname, uname_len,
				(ntfschar*)&ie->key.file_name.file_name,
				ie->key.file_name.file_name_length, 1,
				CASE_SENSITIVE, vol->upcase, vol->upcase_len);
		if (rc == -1)
			break;
		if (rc)
			continue;
		/*
		 * Perfect match, this will never happen as the
		 * ntfs_are_names_equal() call will have gotten a match but we
		 * still treat it correctly.
		 */
		goto found_it2;
	}
	/*
	 * We have finished with this index buffer without success. Check for
	 * the presence of a child node.
	 */
	if (ie->flags & INDEX_ENTRY_NODE) {
		if ((ia->index.flags & NODE_MASK) == LEAF_NODE) {
			ntfs_error(sb, "Index entry with child node found in "
					"a leaf node in directory inode 0x%lx.",
					dir_ni->mft_no);
			goto unm_err_out;
		}
		/* Child node present, descend into it. */
		old_vcn = vcn;
		vcn = sle64_to_cpup((sle64*)((u8*)ie +
				le16_to_cpu(ie->length) - 8));
		if (vcn >= 0) {
			/* If vcn is in the same page cache page as old_vcn we
			 * recycle the mapped page. */
			if (old_vcn << vol->cluster_size_bits >>
					PAGE_CACHE_SHIFT == vcn <<
					vol->cluster_size_bits >>
					PAGE_CACHE_SHIFT)
				goto fast_descend_into_child_node;
			unlock_page(page);
			ntfs_unmap_page(page);
			goto descend_into_child_node;
		}
		ntfs_error(sb, "Negative child node vcn in directory inode "
				"0x%lx.", dir_ni->mft_no);
		goto unm_err_out;
	}
	/*
	 * No child node present, return -ENOENT, unless we have got a matching
	 * name cached in name in which case return the mft reference
	 * associated with it.
	 */
	if (name) {
		unlock_page(page);
		ntfs_unmap_page(page);
		return name->mref;
	}
	ntfs_debug("Entry not found.");
	err = -ENOENT;
unm_err_out:
	unlock_page(page);
	ntfs_unmap_page(page);
err_out:
	if (!err)
		err = -EIO;
	if (ctx)
		ntfs_attr_put_search_ctx(ctx);
	if (m)
		unmap_mft_record(dir_ni);
	if (name) {
		kfree(name);
		*res = NULL;
	}
	return ERR_MREF(err);
dir_err_out:
	ntfs_error(sb, "Corrupt directory.  Aborting lookup.");
	goto err_out;
}

#if 0

// TODO: (AIA)
// The algorithm embedded in this code will be required for the time when we
// want to support adding of entries to directories, where we require correct
610
// collation of file names in order not to cause corruption of the filesystem.
Linus Torvalds's avatar
Linus Torvalds committed
611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 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 1051 1052

/**
 * ntfs_lookup_inode_by_name - find an inode in a directory given its name
 * @dir_ni:	ntfs inode of the directory in which to search for the name
 * @uname:	Unicode name for which to search in the directory
 * @uname_len:	length of the name @uname in Unicode characters
 *
 * Look for an inode with name @uname in the directory with inode @dir_ni.
 * ntfs_lookup_inode_by_name() walks the contents of the directory looking for
 * the Unicode name. If the name is found in the directory, the corresponding
 * inode number (>= 0) is returned as a mft reference in cpu format, i.e. it
 * is a 64-bit number containing the sequence number.
 *
 * On error, a negative value is returned corresponding to the error code. In
 * particular if the inode is not found -ENOENT is returned. Note that you
 * can't just check the return value for being negative, you have to check the
 * inode number for being negative which you can extract using MREC(return
 * value).
 *
 * Note, @uname_len does not include the (optional) terminating NULL character.
 */
u64 ntfs_lookup_inode_by_name(ntfs_inode *dir_ni, const ntfschar *uname,
		const int uname_len)
{
	ntfs_volume *vol = dir_ni->vol;
	struct super_block *sb = vol->sb;
	MFT_RECORD *m;
	INDEX_ROOT *ir;
	INDEX_ENTRY *ie;
	INDEX_ALLOCATION *ia;
	u8 *index_end;
	u64 mref;
	ntfs_attr_search_ctx *ctx;
	int err, rc;
	IGNORE_CASE_BOOL ic;
	VCN vcn, old_vcn;
	struct address_space *ia_mapping;
	struct page *page;
	u8 *kaddr;

	/* Get hold of the mft record for the directory. */
	m = map_mft_record(dir_ni);
	if (IS_ERR(m)) {
		ntfs_error(sb, "map_mft_record() failed with error code %ld.",
				-PTR_ERR(m));
		return ERR_MREF(PTR_ERR(m));
	}
	ctx = ntfs_attr_get_search_ctx(dir_ni, m);
	if (!ctx) {
		err = -ENOMEM;
		goto err_out;
	}
	/* Find the index root attribute in the mft record. */
	err = ntfs_attr_lookup(AT_INDEX_ROOT, I30, 4, CASE_SENSITIVE, 0, NULL,
			0, ctx);
	if (unlikely(err)) {
		if (err == -ENOENT) {
			ntfs_error(sb, "Index root attribute missing in "
					"directory inode 0x%lx.",
					dir_ni->mft_no);
			err = -EIO;
		}
		goto err_out;
	}
	/* Get to the index root value (it's been verified in read_inode). */
	ir = (INDEX_ROOT*)((u8*)ctx->attr +
			le16_to_cpu(ctx->attr->data.resident.value_offset));
	index_end = (u8*)&ir->index + le32_to_cpu(ir->index.index_length);
	/* The first index entry. */
	ie = (INDEX_ENTRY*)((u8*)&ir->index +
			le32_to_cpu(ir->index.entries_offset));
	/*
	 * Loop until we exceed valid memory (corruption case) or until we
	 * reach the last entry.
	 */
	for (;; ie = (INDEX_ENTRY*)((u8*)ie + le16_to_cpu(ie->length))) {
		/* Bounds checks. */
		if ((u8*)ie < (u8*)ctx->mrec || (u8*)ie +
				sizeof(INDEX_ENTRY_HEADER) > index_end ||
				(u8*)ie + le16_to_cpu(ie->key_length) >
				index_end)
			goto dir_err_out;
		/*
		 * The last entry cannot contain a name. It can however contain
		 * a pointer to a child node in the B+tree so we just break out.
		 */
		if (ie->flags & INDEX_ENTRY_END)
			break;
		/*
		 * If the current entry has a name type of POSIX, the name is
		 * case sensitive and not otherwise. This has the effect of us
		 * not being able to access any POSIX file names which collate
		 * after the non-POSIX one when they only differ in case, but
		 * anyone doing screwy stuff like that deserves to burn in
		 * hell... Doing that kind of stuff on NT4 actually causes
		 * corruption on the partition even when using SP6a and Linux
		 * is not involved at all.
		 */
		ic = ie->key.file_name.file_name_type ? IGNORE_CASE :
				CASE_SENSITIVE;
		/*
		 * If the names match perfectly, we are done and return the
		 * mft reference of the inode (i.e. the inode number together
		 * with the sequence number for consistency checking. We
		 * convert it to cpu format before returning.
		 */
		if (ntfs_are_names_equal(uname, uname_len,
				(ntfschar*)&ie->key.file_name.file_name,
				ie->key.file_name.file_name_length, ic,
				vol->upcase, vol->upcase_len)) {
found_it:
			mref = le64_to_cpu(ie->data.dir.indexed_file);
			ntfs_attr_put_search_ctx(ctx);
			unmap_mft_record(dir_ni);
			return mref;
		}
		/*
		 * Not a perfect match, need to do full blown collation so we
		 * know which way in the B+tree we have to go.
		 */
		rc = ntfs_collate_names(uname, uname_len,
				(ntfschar*)&ie->key.file_name.file_name,
				ie->key.file_name.file_name_length, 1,
				IGNORE_CASE, vol->upcase, vol->upcase_len);
		/*
		 * If uname collates before the name of the current entry, there
		 * is definitely no such name in this index but we might need to
		 * descend into the B+tree so we just break out of the loop.
		 */
		if (rc == -1)
			break;
		/* The names are not equal, continue the search. */
		if (rc)
			continue;
		/*
		 * Names match with case insensitive comparison, now try the
		 * case sensitive comparison, which is required for proper
		 * collation.
		 */
		rc = ntfs_collate_names(uname, uname_len,
				(ntfschar*)&ie->key.file_name.file_name,
				ie->key.file_name.file_name_length, 1,
				CASE_SENSITIVE, vol->upcase, vol->upcase_len);
		if (rc == -1)
			break;
		if (rc)
			continue;
		/*
		 * Perfect match, this will never happen as the
		 * ntfs_are_names_equal() call will have gotten a match but we
		 * still treat it correctly.
		 */
		goto found_it;
	}
	/*
	 * We have finished with this index without success. Check for the
	 * presence of a child node.
	 */
	if (!(ie->flags & INDEX_ENTRY_NODE)) {
		/* No child node, return -ENOENT. */
		err = -ENOENT;
		goto err_out;
	} /* Child node present, descend into it. */
	/* Consistency check: Verify that an index allocation exists. */
	if (!NInoIndexAllocPresent(dir_ni)) {
		ntfs_error(sb, "No index allocation attribute but index entry "
				"requires one. Directory inode 0x%lx is "
				"corrupt or driver bug.", dir_ni->mft_no);
		goto err_out;
	}
	/* Get the starting vcn of the index_block holding the child node. */
	vcn = sle64_to_cpup((u8*)ie + le16_to_cpu(ie->length) - 8);
	ia_mapping = VFS_I(dir_ni)->i_mapping;
	/*
	 * We are done with the index root and the mft record. Release them,
	 * otherwise we deadlock with ntfs_map_page().
	 */
	ntfs_attr_put_search_ctx(ctx);
	unmap_mft_record(dir_ni);
	m = NULL;
	ctx = NULL;
descend_into_child_node:
	/*
	 * Convert vcn to index into the index allocation attribute in units
	 * of PAGE_CACHE_SIZE and map the page cache page, reading it from
	 * disk if necessary.
	 */
	page = ntfs_map_page(ia_mapping, vcn <<
			dir_ni->itype.index.vcn_size_bits >> PAGE_CACHE_SHIFT);
	if (IS_ERR(page)) {
		ntfs_error(sb, "Failed to map directory index page, error %ld.",
				-PTR_ERR(page));
		err = PTR_ERR(page);
		goto err_out;
	}
	lock_page(page);
	kaddr = (u8*)page_address(page);
fast_descend_into_child_node:
	/* Get to the index allocation block. */
	ia = (INDEX_ALLOCATION*)(kaddr + ((vcn <<
			dir_ni->itype.index.vcn_size_bits) & ~PAGE_CACHE_MASK));
	/* Bounds checks. */
	if ((u8*)ia < kaddr || (u8*)ia > kaddr + PAGE_CACHE_SIZE) {
		ntfs_error(sb, "Out of bounds check failed. Corrupt directory "
				"inode 0x%lx or driver bug.", dir_ni->mft_no);
		goto unm_err_out;
	}
	/* Catch multi sector transfer fixup errors. */
	if (unlikely(!ntfs_is_indx_record(ia->magic))) {
		ntfs_error(sb, "Directory index record with vcn 0x%llx is "
				"corrupt.  Corrupt inode 0x%lx.  Run chkdsk.",
				(unsigned long long)vcn, dir_ni->mft_no);
		goto unm_err_out;
	}
	if (sle64_to_cpu(ia->index_block_vcn) != vcn) {
		ntfs_error(sb, "Actual VCN (0x%llx) of index buffer is "
				"different from expected VCN (0x%llx). "
				"Directory inode 0x%lx is corrupt or driver "
				"bug.", (unsigned long long)
				sle64_to_cpu(ia->index_block_vcn),
				(unsigned long long)vcn, dir_ni->mft_no);
		goto unm_err_out;
	}
	if (le32_to_cpu(ia->index.allocated_size) + 0x18 !=
			dir_ni->itype.index.block_size) {
		ntfs_error(sb, "Index buffer (VCN 0x%llx) of directory inode "
				"0x%lx has a size (%u) differing from the "
				"directory specified size (%u). Directory "
				"inode is corrupt or driver bug.",
				(unsigned long long)vcn, dir_ni->mft_no,
				le32_to_cpu(ia->index.allocated_size) + 0x18,
				dir_ni->itype.index.block_size);
		goto unm_err_out;
	}
	index_end = (u8*)ia + dir_ni->itype.index.block_size;
	if (index_end > kaddr + PAGE_CACHE_SIZE) {
		ntfs_error(sb, "Index buffer (VCN 0x%llx) of directory inode "
				"0x%lx crosses page boundary. Impossible! "
				"Cannot access! This is probably a bug in the "
				"driver.", (unsigned long long)vcn,
				dir_ni->mft_no);
		goto unm_err_out;
	}
	index_end = (u8*)&ia->index + le32_to_cpu(ia->index.index_length);
	if (index_end > (u8*)ia + dir_ni->itype.index.block_size) {
		ntfs_error(sb, "Size of index buffer (VCN 0x%llx) of directory "
				"inode 0x%lx exceeds maximum size.",
				(unsigned long long)vcn, dir_ni->mft_no);
		goto unm_err_out;
	}
	/* The first index entry. */
	ie = (INDEX_ENTRY*)((u8*)&ia->index +
			le32_to_cpu(ia->index.entries_offset));
	/*
	 * Iterate similar to above big loop but applied to index buffer, thus
	 * loop until we exceed valid memory (corruption case) or until we
	 * reach the last entry.
	 */
	for (;; ie = (INDEX_ENTRY*)((u8*)ie + le16_to_cpu(ie->length))) {
		/* Bounds check. */
		if ((u8*)ie < (u8*)ia || (u8*)ie +
				sizeof(INDEX_ENTRY_HEADER) > index_end ||
				(u8*)ie + le16_to_cpu(ie->key_length) >
				index_end) {
			ntfs_error(sb, "Index entry out of bounds in "
					"directory inode 0x%lx.",
					dir_ni->mft_no);
			goto unm_err_out;
		}
		/*
		 * The last entry cannot contain a name. It can however contain
		 * a pointer to a child node in the B+tree so we just break out.
		 */
		if (ie->flags & INDEX_ENTRY_END)
			break;
		/*
		 * If the current entry has a name type of POSIX, the name is
		 * case sensitive and not otherwise. This has the effect of us
		 * not being able to access any POSIX file names which collate
		 * after the non-POSIX one when they only differ in case, but
		 * anyone doing screwy stuff like that deserves to burn in
		 * hell... Doing that kind of stuff on NT4 actually causes
		 * corruption on the partition even when using SP6a and Linux
		 * is not involved at all.
		 */
		ic = ie->key.file_name.file_name_type ? IGNORE_CASE :
				CASE_SENSITIVE;
		/*
		 * If the names match perfectly, we are done and return the
		 * mft reference of the inode (i.e. the inode number together
		 * with the sequence number for consistency checking. We
		 * convert it to cpu format before returning.
		 */
		if (ntfs_are_names_equal(uname, uname_len,
				(ntfschar*)&ie->key.file_name.file_name,
				ie->key.file_name.file_name_length, ic,
				vol->upcase, vol->upcase_len)) {
found_it2:
			mref = le64_to_cpu(ie->data.dir.indexed_file);
			unlock_page(page);
			ntfs_unmap_page(page);
			return mref;
		}
		/*
		 * Not a perfect match, need to do full blown collation so we
		 * know which way in the B+tree we have to go.
		 */
		rc = ntfs_collate_names(uname, uname_len,
				(ntfschar*)&ie->key.file_name.file_name,
				ie->key.file_name.file_name_length, 1,
				IGNORE_CASE, vol->upcase, vol->upcase_len);
		/*
		 * If uname collates before the name of the current entry, there
		 * is definitely no such name in this index but we might need to
		 * descend into the B+tree so we just break out of the loop.
		 */
		if (rc == -1)
			break;
		/* The names are not equal, continue the search. */
		if (rc)
			continue;
		/*
		 * Names match with case insensitive comparison, now try the
		 * case sensitive comparison, which is required for proper
		 * collation.
		 */
		rc = ntfs_collate_names(uname, uname_len,
				(ntfschar*)&ie->key.file_name.file_name,
				ie->key.file_name.file_name_length, 1,
				CASE_SENSITIVE, vol->upcase, vol->upcase_len);
		if (rc == -1)
			break;
		if (rc)
			continue;
		/*
		 * Perfect match, this will never happen as the
		 * ntfs_are_names_equal() call will have gotten a match but we
		 * still treat it correctly.
		 */
		goto found_it2;
	}
	/*
	 * We have finished with this index buffer without success. Check for
	 * the presence of a child node.
	 */
	if (ie->flags & INDEX_ENTRY_NODE) {
		if ((ia->index.flags & NODE_MASK) == LEAF_NODE) {
			ntfs_error(sb, "Index entry with child node found in "
					"a leaf node in directory inode 0x%lx.",
					dir_ni->mft_no);
			goto unm_err_out;
		}
		/* Child node present, descend into it. */
		old_vcn = vcn;
		vcn = sle64_to_cpup((u8*)ie + le16_to_cpu(ie->length) - 8);
		if (vcn >= 0) {
			/* If vcn is in the same page cache page as old_vcn we
			 * recycle the mapped page. */
			if (old_vcn << vol->cluster_size_bits >>
					PAGE_CACHE_SHIFT == vcn <<
					vol->cluster_size_bits >>
					PAGE_CACHE_SHIFT)
				goto fast_descend_into_child_node;
			unlock_page(page);
			ntfs_unmap_page(page);
			goto descend_into_child_node;
		}
		ntfs_error(sb, "Negative child node vcn in directory inode "
				"0x%lx.", dir_ni->mft_no);
		goto unm_err_out;
	}
	/* No child node, return -ENOENT. */
	ntfs_debug("Entry not found.");
	err = -ENOENT;
unm_err_out:
	unlock_page(page);
	ntfs_unmap_page(page);
err_out:
	if (!err)
		err = -EIO;
	if (ctx)
		ntfs_attr_put_search_ctx(ctx);
	if (m)
		unmap_mft_record(dir_ni);
	return ERR_MREF(err);
dir_err_out:
	ntfs_error(sb, "Corrupt directory. Aborting lookup.");
	goto err_out;
}

#endif

/**
 * ntfs_filldir - ntfs specific filldir method
 * @vol:	current ntfs volume
 * @fpos:	position in the directory
 * @ndir:	ntfs inode of current directory
 * @ia_page:	page in which the index allocation buffer @ie is in resides
 * @ie:		current index entry
 * @name:	buffer to use for the converted name
 * @dirent:	vfs filldir callback context
 * @filldir:	vfs filldir callback
 *
 * Convert the Unicode @name to the loaded NLS and pass it to the @filldir
 * callback.
 *
 * If @ia_page is not NULL it is the locked page containing the index
 * allocation block containing the index entry @ie.
 *
 * Note, we drop (and then reacquire) the page lock on @ia_page across the
 * @filldir() call otherwise we would deadlock with NFSd when it calls ->lookup
 * since ntfs_lookup() will lock the same page.  As an optimization, we do not
 * retake the lock if we are returning a non-zero value as ntfs_readdir()
 * would need to drop the lock immediately anyway.
 */
static inline int ntfs_filldir(ntfs_volume *vol, loff_t fpos,
		ntfs_inode *ndir, struct page *ia_page, INDEX_ENTRY *ie,
		u8 *name, void *dirent, filldir_t filldir)
{
	unsigned long mref;
	int name_len, rc;
	unsigned dt_type;
	FILE_NAME_TYPE_FLAGS name_type;

	name_type = ie->key.file_name.file_name_type;
	if (name_type == FILE_NAME_DOS) {
		ntfs_debug("Skipping DOS name space entry.");
		return 0;
	}
	if (MREF_LE(ie->data.dir.indexed_file) == FILE_root) {
		ntfs_debug("Skipping root directory self reference entry.");
		return 0;
	}
	if (MREF_LE(ie->data.dir.indexed_file) < FILE_first_user &&
			!NVolShowSystemFiles(vol)) {
		ntfs_debug("Skipping system file.");
		return 0;
	}
	name_len = ntfs_ucstonls(vol, (ntfschar*)&ie->key.file_name.file_name,
			ie->key.file_name.file_name_length, &name,
			NTFS_MAX_NAME_LEN * NLS_MAX_CHARSET_SIZE + 1);
	if (name_len <= 0) {
1053 1054
		ntfs_warning(vol->sb, "Skipping unrepresentable inode 0x%llx.",
				(long long)MREF_LE(ie->data.dir.indexed_file));
Linus Torvalds's avatar
Linus Torvalds committed
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 1083 1084 1085 1086
		return 0;
	}
	if (ie->key.file_name.file_attributes &
			FILE_ATTR_DUP_FILE_NAME_INDEX_PRESENT)
		dt_type = DT_DIR;
	else
		dt_type = DT_REG;
	mref = MREF_LE(ie->data.dir.indexed_file);
	/*
	 * Drop the page lock otherwise we deadlock with NFS when it calls
	 * ->lookup since ntfs_lookup() will lock the same page.
	 */
	if (ia_page)
		unlock_page(ia_page);
	ntfs_debug("Calling filldir for %s with len %i, fpos 0x%llx, inode "
			"0x%lx, DT_%s.", name, name_len, fpos, mref,
			dt_type == DT_DIR ? "DIR" : "REG");
	rc = filldir(dirent, name, name_len, fpos, mref, dt_type);
	/* Relock the page but not if we are aborting ->readdir. */
	if (!rc && ia_page)
		lock_page(ia_page);
	return rc;
}

/*
 * We use the same basic approach as the old NTFS driver, i.e. we parse the
 * index root entries and then the index allocation entries that are marked
 * as in use in the index bitmap.
 *
 * While this will return the names in random order this doesn't matter for
 * ->readdir but OTOH results in a faster ->readdir.
 *
1087
 * VFS calls ->readdir without BKL but with i_mutex held. This protects the VFS
Linus Torvalds's avatar
Linus Torvalds committed
1088 1089 1090
 * parts (e.g. ->f_pos and ->i_size, and it also protects against directory
 * modifications).
 *
1091
 * Locking:  - Caller must hold i_mutex on the directory.
Linus Torvalds's avatar
Linus Torvalds committed
1092 1093 1094 1095 1096 1097 1098 1099 1100 1101
 *	     - Each page cache page in the index allocation mapping must be
 *	       locked whilst being accessed otherwise we may find a corrupt
 *	       page due to it being under ->writepage at the moment which
 *	       applies the mst protection fixups before writing out and then
 *	       removes them again after the write is complete after which it 
 *	       unlocks the page.
 */
static int ntfs_readdir(struct file *filp, void *dirent, filldir_t filldir)
{
	s64 ia_pos, ia_start, prev_ia_pos, bmp_pos;
1102
	loff_t fpos, i_size;
1103
	struct inode *bmp_vi, *vdir = filp->f_path.dentry->d_inode;
Linus Torvalds's avatar
Linus Torvalds committed
1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122
	struct super_block *sb = vdir->i_sb;
	ntfs_inode *ndir = NTFS_I(vdir);
	ntfs_volume *vol = NTFS_SB(sb);
	MFT_RECORD *m;
	INDEX_ROOT *ir = NULL;
	INDEX_ENTRY *ie;
	INDEX_ALLOCATION *ia;
	u8 *name = NULL;
	int rc, err, ir_pos, cur_bmp_pos;
	struct address_space *ia_mapping, *bmp_mapping;
	struct page *bmp_page = NULL, *ia_page = NULL;
	u8 *kaddr, *bmp, *index_end;
	ntfs_attr_search_ctx *ctx;

	fpos = filp->f_pos;
	ntfs_debug("Entering for inode 0x%lx, fpos 0x%llx.",
			vdir->i_ino, fpos);
	rc = err = 0;
	/* Are we at end of dir yet? */
1123 1124
	i_size = i_size_read(vdir);
	if (fpos >= i_size + vol->mft_record_size)
Linus Torvalds's avatar
Linus Torvalds committed
1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137
		goto done;
	/* Emulate . and .. for all directories. */
	if (!fpos) {
		ntfs_debug("Calling filldir for . with len 1, fpos 0x0, "
				"inode 0x%lx, DT_DIR.", vdir->i_ino);
		rc = filldir(dirent, ".", 1, fpos, vdir->i_ino, DT_DIR);
		if (rc)
			goto done;
		fpos++;
	}
	if (fpos == 1) {
		ntfs_debug("Calling filldir for .. with len 2, fpos 0x1, "
				"inode 0x%lx, DT_DIR.",
1138
				(unsigned long)parent_ino(filp->f_path.dentry));
Linus Torvalds's avatar
Linus Torvalds committed
1139
		rc = filldir(dirent, "..", 2, fpos,
1140
				parent_ino(filp->f_path.dentry), DT_DIR);
Linus Torvalds's avatar
Linus Torvalds committed
1141 1142 1143 1144 1145 1146 1147 1148 1149 1150
		if (rc)
			goto done;
		fpos++;
	}
	m = NULL;
	ctx = NULL;
	/*
	 * Allocate a buffer to store the current name being processed
	 * converted to format determined by current NLS.
	 */
1151
	name = kmalloc(NTFS_MAX_NAME_LEN * NLS_MAX_CHARSET_SIZE + 1, GFP_NOFS);
Linus Torvalds's avatar
Linus Torvalds committed
1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191
	if (unlikely(!name)) {
		err = -ENOMEM;
		goto err_out;
	}
	/* Are we jumping straight into the index allocation attribute? */
	if (fpos >= vol->mft_record_size)
		goto skip_index_root;
	/* Get hold of the mft record for the directory. */
	m = map_mft_record(ndir);
	if (IS_ERR(m)) {
		err = PTR_ERR(m);
		m = NULL;
		goto err_out;
	}
	ctx = ntfs_attr_get_search_ctx(ndir, m);
	if (unlikely(!ctx)) {
		err = -ENOMEM;
		goto err_out;
	}
	/* Get the offset into the index root attribute. */
	ir_pos = (s64)fpos;
	/* Find the index root attribute in the mft record. */
	err = ntfs_attr_lookup(AT_INDEX_ROOT, I30, 4, CASE_SENSITIVE, 0, NULL,
			0, ctx);
	if (unlikely(err)) {
		ntfs_error(sb, "Index root attribute missing in directory "
				"inode 0x%lx.", vdir->i_ino);
		goto err_out;
	}
	/*
	 * Copy the index root attribute value to a buffer so that we can put
	 * the search context and unmap the mft record before calling the
	 * filldir() callback.  We need to do this because of NFSd which calls
	 * ->lookup() from its filldir callback() and this causes NTFS to
	 * deadlock as ntfs_lookup() maps the mft record of the directory and
	 * we have got it mapped here already.  The only solution is for us to
	 * unmap the mft record here so that a call to ntfs_lookup() is able to
	 * map the mft record without deadlocking.
	 */
	rc = le32_to_cpu(ctx->attr->data.resident.value_length);
1192
	ir = kmalloc(rc, GFP_NOFS);
Linus Torvalds's avatar
Linus Torvalds committed
1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250
	if (unlikely(!ir)) {
		err = -ENOMEM;
		goto err_out;
	}
	/* Copy the index root value (it has been verified in read_inode). */
	memcpy(ir, (u8*)ctx->attr +
			le16_to_cpu(ctx->attr->data.resident.value_offset), rc);
	ntfs_attr_put_search_ctx(ctx);
	unmap_mft_record(ndir);
	ctx = NULL;
	m = NULL;
	index_end = (u8*)&ir->index + le32_to_cpu(ir->index.index_length);
	/* The first index entry. */
	ie = (INDEX_ENTRY*)((u8*)&ir->index +
			le32_to_cpu(ir->index.entries_offset));
	/*
	 * Loop until we exceed valid memory (corruption case) or until we
	 * reach the last entry or until filldir tells us it has had enough
	 * or signals an error (both covered by the rc test).
	 */
	for (;; ie = (INDEX_ENTRY*)((u8*)ie + le16_to_cpu(ie->length))) {
		ntfs_debug("In index root, offset 0x%zx.", (u8*)ie - (u8*)ir);
		/* Bounds checks. */
		if (unlikely((u8*)ie < (u8*)ir || (u8*)ie +
				sizeof(INDEX_ENTRY_HEADER) > index_end ||
				(u8*)ie + le16_to_cpu(ie->key_length) >
				index_end))
			goto err_out;
		/* The last entry cannot contain a name. */
		if (ie->flags & INDEX_ENTRY_END)
			break;
		/* Skip index root entry if continuing previous readdir. */
		if (ir_pos > (u8*)ie - (u8*)ir)
			continue;
		/* Advance the position even if going to skip the entry. */
		fpos = (u8*)ie - (u8*)ir;
		/* Submit the name to the filldir callback. */
		rc = ntfs_filldir(vol, fpos, ndir, NULL, ie, name, dirent,
				filldir);
		if (rc) {
			kfree(ir);
			goto abort;
		}
	}
	/* We are done with the index root and can free the buffer. */
	kfree(ir);
	ir = NULL;
	/* If there is no index allocation attribute we are finished. */
	if (!NInoIndexAllocPresent(ndir))
		goto EOD;
	/* Advance fpos to the beginning of the index allocation. */
	fpos = vol->mft_record_size;
skip_index_root:
	kaddr = NULL;
	prev_ia_pos = -1LL;
	/* Get the offset into the index allocation attribute. */
	ia_pos = (s64)fpos - vol->mft_record_size;
	ia_mapping = vdir->i_mapping;
1251 1252 1253 1254 1255 1256
	ntfs_debug("Inode 0x%lx, getting index bitmap.", vdir->i_ino);
	bmp_vi = ntfs_attr_iget(vdir, AT_BITMAP, I30, 4);
	if (IS_ERR(bmp_vi)) {
		ntfs_error(sb, "Failed to get bitmap attribute.");
		err = PTR_ERR(bmp_vi);
		goto err_out;
Linus Torvalds's avatar
Linus Torvalds committed
1257 1258 1259 1260
	}
	bmp_mapping = bmp_vi->i_mapping;
	/* Get the starting bitmap bit position and sanity check it. */
	bmp_pos = ia_pos >> ndir->itype.index.block_size_bits;
1261
	if (unlikely(bmp_pos >> 3 >= i_size_read(bmp_vi))) {
Linus Torvalds's avatar
Linus Torvalds committed
1262 1263
		ntfs_error(sb, "Current index allocation position exceeds "
				"index bitmap size.");
1264
		goto iput_err_out;
Linus Torvalds's avatar
Linus Torvalds committed
1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279
	}
	/* Get the starting bit position in the current bitmap page. */
	cur_bmp_pos = bmp_pos & ((PAGE_CACHE_SIZE * 8) - 1);
	bmp_pos &= ~(u64)((PAGE_CACHE_SIZE * 8) - 1);
get_next_bmp_page:
	ntfs_debug("Reading bitmap with page index 0x%llx, bit ofs 0x%llx",
			(unsigned long long)bmp_pos >> (3 + PAGE_CACHE_SHIFT),
			(unsigned long long)bmp_pos &
			(unsigned long long)((PAGE_CACHE_SIZE * 8) - 1));
	bmp_page = ntfs_map_page(bmp_mapping,
			bmp_pos >> (3 + PAGE_CACHE_SHIFT));
	if (IS_ERR(bmp_page)) {
		ntfs_error(sb, "Reading index bitmap failed.");
		err = PTR_ERR(bmp_page);
		bmp_page = NULL;
1280
		goto iput_err_out;
Linus Torvalds's avatar
Linus Torvalds committed
1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297
	}
	bmp = (u8*)page_address(bmp_page);
	/* Find next index block in use. */
	while (!(bmp[cur_bmp_pos >> 3] & (1 << (cur_bmp_pos & 7)))) {
find_next_index_buffer:
		cur_bmp_pos++;
		/*
		 * If we have reached the end of the bitmap page, get the next
		 * page, and put away the old one.
		 */
		if (unlikely((cur_bmp_pos >> 3) >= PAGE_CACHE_SIZE)) {
			ntfs_unmap_page(bmp_page);
			bmp_pos += PAGE_CACHE_SIZE * 8;
			cur_bmp_pos = 0;
			goto get_next_bmp_page;
		}
		/* If we have reached the end of the bitmap, we are done. */
1298
		if (unlikely(((bmp_pos + cur_bmp_pos) >> 3) >= i_size))
Linus Torvalds's avatar
Linus Torvalds committed
1299 1300 1301 1302 1303 1304 1305
			goto unm_EOD;
		ia_pos = (bmp_pos + cur_bmp_pos) <<
				ndir->itype.index.block_size_bits;
	}
	ntfs_debug("Handling index buffer 0x%llx.",
			(unsigned long long)bmp_pos + cur_bmp_pos);
	/* If the current index buffer is in the same page we reuse the page. */
1306 1307
	if ((prev_ia_pos & (s64)PAGE_CACHE_MASK) !=
			(ia_pos & (s64)PAGE_CACHE_MASK)) {
Linus Torvalds's avatar
Linus Torvalds committed
1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 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 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426
		prev_ia_pos = ia_pos;
		if (likely(ia_page != NULL)) {
			unlock_page(ia_page);
			ntfs_unmap_page(ia_page);
		}
		/*
		 * Map the page cache page containing the current ia_pos,
		 * reading it from disk if necessary.
		 */
		ia_page = ntfs_map_page(ia_mapping, ia_pos >> PAGE_CACHE_SHIFT);
		if (IS_ERR(ia_page)) {
			ntfs_error(sb, "Reading index allocation data failed.");
			err = PTR_ERR(ia_page);
			ia_page = NULL;
			goto err_out;
		}
		lock_page(ia_page);
		kaddr = (u8*)page_address(ia_page);
	}
	/* Get the current index buffer. */
	ia = (INDEX_ALLOCATION*)(kaddr + (ia_pos & ~PAGE_CACHE_MASK &
			~(s64)(ndir->itype.index.block_size - 1)));
	/* Bounds checks. */
	if (unlikely((u8*)ia < kaddr || (u8*)ia > kaddr + PAGE_CACHE_SIZE)) {
		ntfs_error(sb, "Out of bounds check failed. Corrupt directory "
				"inode 0x%lx or driver bug.", vdir->i_ino);
		goto err_out;
	}
	/* Catch multi sector transfer fixup errors. */
	if (unlikely(!ntfs_is_indx_record(ia->magic))) {
		ntfs_error(sb, "Directory index record with vcn 0x%llx is "
				"corrupt.  Corrupt inode 0x%lx.  Run chkdsk.",
				(unsigned long long)ia_pos >>
				ndir->itype.index.vcn_size_bits, vdir->i_ino);
		goto err_out;
	}
	if (unlikely(sle64_to_cpu(ia->index_block_vcn) != (ia_pos &
			~(s64)(ndir->itype.index.block_size - 1)) >>
			ndir->itype.index.vcn_size_bits)) {
		ntfs_error(sb, "Actual VCN (0x%llx) of index buffer is "
				"different from expected VCN (0x%llx). "
				"Directory inode 0x%lx is corrupt or driver "
				"bug. ", (unsigned long long)
				sle64_to_cpu(ia->index_block_vcn),
				(unsigned long long)ia_pos >>
				ndir->itype.index.vcn_size_bits, vdir->i_ino);
		goto err_out;
	}
	if (unlikely(le32_to_cpu(ia->index.allocated_size) + 0x18 !=
			ndir->itype.index.block_size)) {
		ntfs_error(sb, "Index buffer (VCN 0x%llx) of directory inode "
				"0x%lx has a size (%u) differing from the "
				"directory specified size (%u). Directory "
				"inode is corrupt or driver bug.",
				(unsigned long long)ia_pos >>
				ndir->itype.index.vcn_size_bits, vdir->i_ino,
				le32_to_cpu(ia->index.allocated_size) + 0x18,
				ndir->itype.index.block_size);
		goto err_out;
	}
	index_end = (u8*)ia + ndir->itype.index.block_size;
	if (unlikely(index_end > kaddr + PAGE_CACHE_SIZE)) {
		ntfs_error(sb, "Index buffer (VCN 0x%llx) of directory inode "
				"0x%lx crosses page boundary. Impossible! "
				"Cannot access! This is probably a bug in the "
				"driver.", (unsigned long long)ia_pos >>
				ndir->itype.index.vcn_size_bits, vdir->i_ino);
		goto err_out;
	}
	ia_start = ia_pos & ~(s64)(ndir->itype.index.block_size - 1);
	index_end = (u8*)&ia->index + le32_to_cpu(ia->index.index_length);
	if (unlikely(index_end > (u8*)ia + ndir->itype.index.block_size)) {
		ntfs_error(sb, "Size of index buffer (VCN 0x%llx) of directory "
				"inode 0x%lx exceeds maximum size.",
				(unsigned long long)ia_pos >>
				ndir->itype.index.vcn_size_bits, vdir->i_ino);
		goto err_out;
	}
	/* The first index entry in this index buffer. */
	ie = (INDEX_ENTRY*)((u8*)&ia->index +
			le32_to_cpu(ia->index.entries_offset));
	/*
	 * Loop until we exceed valid memory (corruption case) or until we
	 * reach the last entry or until filldir tells us it has had enough
	 * or signals an error (both covered by the rc test).
	 */
	for (;; ie = (INDEX_ENTRY*)((u8*)ie + le16_to_cpu(ie->length))) {
		ntfs_debug("In index allocation, offset 0x%llx.",
				(unsigned long long)ia_start +
				(unsigned long long)((u8*)ie - (u8*)ia));
		/* Bounds checks. */
		if (unlikely((u8*)ie < (u8*)ia || (u8*)ie +
				sizeof(INDEX_ENTRY_HEADER) > index_end ||
				(u8*)ie + le16_to_cpu(ie->key_length) >
				index_end))
			goto err_out;
		/* The last entry cannot contain a name. */
		if (ie->flags & INDEX_ENTRY_END)
			break;
		/* Skip index block entry if continuing previous readdir. */
		if (ia_pos - ia_start > (u8*)ie - (u8*)ia)
			continue;
		/* Advance the position even if going to skip the entry. */
		fpos = (u8*)ie - (u8*)ia +
				(sle64_to_cpu(ia->index_block_vcn) <<
				ndir->itype.index.vcn_size_bits) +
				vol->mft_record_size;
		/*
		 * Submit the name to the @filldir callback.  Note,
		 * ntfs_filldir() drops the lock on @ia_page but it retakes it
		 * before returning, unless a non-zero value is returned in
		 * which case the page is left unlocked.
		 */
		rc = ntfs_filldir(vol, fpos, ndir, ia_page, ie, name, dirent,
				filldir);
		if (rc) {
			/* @ia_page is already unlocked in this case. */
			ntfs_unmap_page(ia_page);
			ntfs_unmap_page(bmp_page);
1427
			iput(bmp_vi);
Linus Torvalds's avatar
Linus Torvalds committed
1428 1429 1430 1431 1432 1433 1434 1435 1436 1437
			goto abort;
		}
	}
	goto find_next_index_buffer;
unm_EOD:
	if (ia_page) {
		unlock_page(ia_page);
		ntfs_unmap_page(ia_page);
	}
	ntfs_unmap_page(bmp_page);
1438
	iput(bmp_vi);
Linus Torvalds's avatar
Linus Torvalds committed
1439 1440
EOD:
	/* We are finished, set fpos to EOD. */
1441
	fpos = i_size + vol->mft_record_size;
Linus Torvalds's avatar
Linus Torvalds committed
1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454
abort:
	kfree(name);
done:
#ifdef DEBUG
	if (!rc)
		ntfs_debug("EOD, fpos 0x%llx, returning 0.", fpos);
	else
		ntfs_debug("filldir returned %i, fpos 0x%llx, returning 0.",
				rc, fpos);
#endif
	filp->f_pos = fpos;
	return 0;
err_out:
1455
	if (bmp_page) {
Linus Torvalds's avatar
Linus Torvalds committed
1456
		ntfs_unmap_page(bmp_page);
1457 1458 1459
iput_err_out:
		iput(bmp_vi);
	}
Linus Torvalds's avatar
Linus Torvalds committed
1460 1461 1462 1463
	if (ia_page) {
		unlock_page(ia_page);
		ntfs_unmap_page(ia_page);
	}
1464 1465
	kfree(ir);
	kfree(name);
Linus Torvalds's avatar
Linus Torvalds committed
1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495
	if (ctx)
		ntfs_attr_put_search_ctx(ctx);
	if (m)
		unmap_mft_record(ndir);
	if (!err)
		err = -EIO;
	ntfs_debug("Failed. Returning error code %i.", -err);
	filp->f_pos = fpos;
	return err;
}

/**
 * ntfs_dir_open - called when an inode is about to be opened
 * @vi:		inode to be opened
 * @filp:	file structure describing the inode
 *
 * Limit directory size to the page cache limit on architectures where unsigned
 * long is 32-bits. This is the most we can do for now without overflowing the
 * page cache page index. Doing it this way means we don't run into problems
 * because of existing too large directories. It would be better to allow the
 * user to read the accessible part of the directory but I doubt very much
 * anyone is going to hit this check on a 32-bit architecture, so there is no
 * point in adding the extra complexity required to support this.
 *
 * On 64-bit architectures, the check is hopefully optimized away by the
 * compiler.
 */
static int ntfs_dir_open(struct inode *vi, struct file *filp)
{
	if (sizeof(unsigned long) < 8) {
1496
		if (i_size_read(vi) > MAX_LFS_FILESIZE)
Linus Torvalds's avatar
Linus Torvalds committed
1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521
			return -EFBIG;
	}
	return 0;
}

#ifdef NTFS_RW

/**
 * ntfs_dir_fsync - sync a directory to disk
 * @filp:	directory to be synced
 * @dentry:	dentry describing the directory to sync
 * @datasync:	if non-zero only flush user data and not metadata
 *
 * Data integrity sync of a directory to disk.  Used for fsync, fdatasync, and
 * msync system calls.  This function is based on file.c::ntfs_file_fsync().
 *
 * Write the mft record and all associated extent mft records as well as the
 * $INDEX_ALLOCATION and $BITMAP attributes and then sync the block device.
 *
 * If @datasync is true, we do not wait on the inode(s) to be written out
 * but we always wait on the page cache pages to be written out.
 *
 * Note: In the past @filp could be NULL so we ignore it as we don't need it
 * anyway.
 *
1522
 * Locking: Caller must hold i_mutex on the inode.
Linus Torvalds's avatar
Linus Torvalds committed
1523 1524 1525 1526 1527 1528 1529 1530 1531
 *
 * TODO: We should probably also write all attribute/index inodes associated
 * with this inode but since we have no simple way of getting to them we ignore
 * this problem for now.  We do write the $BITMAP attribute if it is present
 * which is the important one for a directory so things are not too bad.
 */
static int ntfs_dir_fsync(struct file *filp, struct dentry *dentry,
		int datasync)
{
1532
	struct inode *bmp_vi, *vi = dentry->d_inode;
Linus Torvalds's avatar
Linus Torvalds committed
1533
	int err, ret;
1534
	ntfs_attr na;
Linus Torvalds's avatar
Linus Torvalds committed
1535 1536 1537

	ntfs_debug("Entering for inode 0x%lx.", vi->i_ino);
	BUG_ON(!S_ISDIR(vi->i_mode));
1538 1539 1540 1541 1542 1543 1544 1545 1546 1547
	/* If the bitmap attribute inode is in memory sync it, too. */
	na.mft_no = vi->i_ino;
	na.type = AT_BITMAP;
	na.name = I30;
	na.name_len = 4;
	bmp_vi = ilookup5(vi->i_sb, vi->i_ino, (test_t)ntfs_test_inode, &na);
	if (bmp_vi) {
 		write_inode_now(bmp_vi, !datasync);
		iput(bmp_vi);
	}
Linus Torvalds's avatar
Linus Torvalds committed
1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562
	ret = ntfs_write_inode(vi, 1);
	write_inode_now(vi, !datasync);
	err = sync_blockdev(vi->i_sb->s_bdev);
	if (unlikely(err && !ret))
		ret = err;
	if (likely(!ret))
		ntfs_debug("Done.");
	else
		ntfs_warning(vi->i_sb, "Failed to f%ssync inode 0x%lx.  Error "
				"%u.", datasync ? "data" : "", vi->i_ino, -ret);
	return ret;
}

#endif /* NTFS_RW */

1563
const struct file_operations ntfs_dir_ops = {
Linus Torvalds's avatar
Linus Torvalds committed
1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575
	.llseek		= generic_file_llseek,	/* Seek inside directory. */
	.read		= generic_read_dir,	/* Return -EISDIR. */
	.readdir	= ntfs_readdir,		/* Read directory contents. */
#ifdef NTFS_RW
	.fsync		= ntfs_dir_fsync,	/* Sync a directory to disk. */
	/*.aio_fsync	= ,*/			/* Sync all outstanding async
						   i/o operations on a kiocb. */
#endif /* NTFS_RW */
	/*.ioctl	= ,*/			/* Perform function on the
						   mounted filesystem. */
	.open		= ntfs_dir_open,	/* Open directory. */
};