inode.c 83.4 KB
Newer Older
Linus Torvalds's avatar
Linus Torvalds committed
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
/*
 *  linux/fs/ext3/inode.c
 *
 * Copyright (C) 1992, 1993, 1994, 1995
 * Remy Card (card@masi.ibp.fr)
 * Laboratoire MASI - Institut Blaise Pascal
 * Universite Pierre et Marie Curie (Paris VI)
 *
 *  from
 *
 *  linux/fs/minix/inode.c
 *
 *  Copyright (C) 1991, 1992  Linus Torvalds
 *
 *  Goal-directed block allocation by Stephen Tweedie
 * 	(sct@redhat.com), 1993, 1998
 *  Big-endian to little-endian byte-swapping/bitmaps by
 *        David S. Miller (davem@caip.rutgers.edu), 1995
 *  64-bit file support on 64-bit platforms by Jakub Jelinek
 * 	(jj@sunsite.ms.mff.cuni.cz)
 *
 *  Assorted race fixes, rewrite of ext3_get_block() by Al Viro, 2000
 */

25
#include <linux/module.h>
Linus Torvalds's avatar
Linus Torvalds committed
26
#include <linux/fs.h>
27
#include <linux/time.h>
Linus Torvalds's avatar
Linus Torvalds committed
28 29 30 31
#include <linux/ext3_jbd.h>
#include <linux/jbd.h>
#include <linux/smp_lock.h>
#include <linux/highuid.h>
32
#include <linux/pagemap.h>
Linus Torvalds's avatar
Linus Torvalds committed
33
#include <linux/quotaops.h>
34
#include <linux/string.h>
35
#include <linux/buffer_head.h>
36
#include <linux/mpage.h>
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 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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201

/*
 * SEARCH_FROM_ZERO forces each block allocation to search from the start
 * of the filesystem.  This is to force rapid reallocation of recently-freed
 * blocks.  The file fragmentation is horrendous.
 */
#undef SEARCH_FROM_ZERO

/* The ext3 forget function must perform a revoke if we are freeing data
 * which has been journaled.  Metadata (eg. indirect blocks) must be
 * revoked in all cases. 
 *
 * "bh" may be NULL: a metadata block may have been freed from memory
 * but there may still be a record of it in the journal, and that record
 * still needs to be revoked.
 */

static int ext3_forget(handle_t *handle, int is_metadata,
		       struct inode *inode, struct buffer_head *bh,
		       int blocknr)
{
	int err;

	BUFFER_TRACE(bh, "enter");

	jbd_debug(4, "forgetting bh %p: is_metadata = %d, mode %o, "
		  "data mode %lx\n",
		  bh, is_metadata, inode->i_mode,
		  test_opt(inode->i_sb, DATA_FLAGS));
	
	/* Never use the revoke function if we are doing full data
	 * journaling: there is no need to, and a V1 superblock won't
	 * support it.  Otherwise, only skip the revoke on un-journaled
	 * data blocks. */

	if (test_opt(inode->i_sb, DATA_FLAGS) == EXT3_MOUNT_JOURNAL_DATA ||
	    (!is_metadata && !ext3_should_journal_data(inode))) {
		if (bh) {
			BUFFER_TRACE(bh, "call journal_forget");
			ext3_journal_forget(handle, bh);
		}
		return 0;
	}

	/*
	 * data!=journal && (is_metadata || should_journal_data(inode))
	 */
	BUFFER_TRACE(bh, "call ext3_journal_revoke");
	err = ext3_journal_revoke(handle, blocknr, bh);
	if (err)
		ext3_abort(inode->i_sb, __FUNCTION__,
			   "error %d when attempting revoke", err);
	BUFFER_TRACE(bh, "exit");
	return err;
}

/* 
 * Truncate transactions can be complex and absolutely huge.  So we need to
 * be able to restart the transaction at a conventient checkpoint to make
 * sure we don't overflow the journal.
 *
 * start_transaction gets us a new handle for a truncate transaction,
 * and extend_transaction tries to extend the existing one a bit.  If
 * extend fails, we need to propagate the failure up and restart the
 * transaction in the top-level truncate loop. --sct 
 */

static handle_t *start_transaction(struct inode *inode) 
{
	long needed;
	handle_t *result;
	
	needed = inode->i_blocks;
	if (needed > EXT3_MAX_TRANS_DATA) 
		needed = EXT3_MAX_TRANS_DATA;
	
	result = ext3_journal_start(inode, EXT3_DATA_TRANS_BLOCKS + needed);
	if (!IS_ERR(result))
		return result;
	
	ext3_std_error(inode->i_sb, PTR_ERR(result));
	return result;
}

/*
 * Try to extend this transaction for the purposes of truncation.
 *
 * Returns 0 if we managed to create more room.  If we can't create more
 * room, and the transaction must be restarted we return 1.
 */
static int try_to_extend_transaction(handle_t *handle, struct inode *inode)
{
	long needed;
	
	if (handle->h_buffer_credits > EXT3_RESERVE_TRANS_BLOCKS)
		return 0;
	needed = inode->i_blocks;
	if (needed > EXT3_MAX_TRANS_DATA) 
		needed = EXT3_MAX_TRANS_DATA;
	if (!ext3_journal_extend(handle, EXT3_RESERVE_TRANS_BLOCKS + needed))
		return 0;
	return 1;
}

/*
 * Restart the transaction associated with *handle.  This does a commit,
 * so before we call here everything must be consistently dirtied against
 * this transaction.
 */
static int ext3_journal_test_restart(handle_t *handle, struct inode *inode)
{
	long needed = inode->i_blocks;
	if (needed > EXT3_MAX_TRANS_DATA) 
		needed = EXT3_MAX_TRANS_DATA;
	jbd_debug(2, "restarting handle %p\n", handle);
	return ext3_journal_restart(handle, EXT3_DATA_TRANS_BLOCKS + needed);
}

/*
 * Called at each iput()
 */
void ext3_put_inode (struct inode * inode)
{
	ext3_discard_prealloc (inode);
}

/*
 * Called at the last iput() if i_nlink is zero.
 */
void ext3_delete_inode (struct inode * inode)
{
	handle_t *handle;
	
	if (is_bad_inode(inode) ||
	    inode->i_ino == EXT3_ACL_IDX_INO ||
	    inode->i_ino == EXT3_ACL_DATA_INO)
		goto no_delete;

	lock_kernel();
	handle = start_transaction(inode);
	if (IS_ERR(handle)) {
		/* If we're going to skip the normal cleanup, we still
		 * need to make sure that the in-core orphan linked list
		 * is properly cleaned up. */
		ext3_orphan_del(NULL, inode);

		ext3_std_error(inode->i_sb, PTR_ERR(handle));
		unlock_kernel();
		goto no_delete;
	}
	
	if (IS_SYNC(inode))
		handle->h_sync = 1;
	inode->i_size = 0;
	if (inode->i_blocks)
		ext3_truncate(inode);
	/*
	 * Kill off the orphan record which ext3_truncate created.
	 * AKPM: I think this can be inside the above `if'.
	 * Note that ext3_orphan_del() has to be able to cope with the
	 * deletion of a non-existent orphan - this is because we don't
	 * know if ext3_truncate() actually created an orphan record.
	 * (Well, we could do this if we need to, but heck - it works)
	 */
	ext3_orphan_del(handle, inode);
Linus Torvalds's avatar
Linus Torvalds committed
202
	EXT3_I(inode)->i_dtime	= CURRENT_TIME;
Linus Torvalds's avatar
Linus Torvalds committed
203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225

	/* 
	 * One subtle ordering requirement: if anything has gone wrong
	 * (transaction abort, IO errors, whatever), then we can still
	 * do these next steps (the fs will already have been marked as
	 * having errors), but we can't free the inode if the mark_dirty
	 * fails.  
	 */
	if (ext3_mark_inode_dirty(handle, inode))
		/* If that failed, just do the required in-core inode clear. */
		clear_inode(inode);
	else
		ext3_free_inode(handle, inode);
	ext3_journal_stop(handle, inode);
	unlock_kernel();
	return;
no_delete:
	clear_inode(inode);	/* We must guarantee clearing of inode... */
}

void ext3_discard_prealloc (struct inode * inode)
{
#ifdef EXT3_PREALLOCATE
Linus Torvalds's avatar
Linus Torvalds committed
226
	struct ext3_inode_info *ei = EXT3_I(inode);
Linus Torvalds's avatar
Linus Torvalds committed
227 228
	lock_kernel();
	/* Writer: ->i_prealloc* */
Linus Torvalds's avatar
Linus Torvalds committed
229 230 231 232 233
	if (ei->i_prealloc_count) {
		unsigned short total = ei->i_prealloc_count;
		unsigned long block = ei->i_prealloc_block;
		ei->i_prealloc_count = 0;
		ei->i_prealloc_block = 0;
Linus Torvalds's avatar
Linus Torvalds committed
234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249
		/* Writer: end */
		ext3_free_blocks (inode, block, total);
	}
	unlock_kernel();
#endif
}

static int ext3_alloc_block (handle_t *handle,
			struct inode * inode, unsigned long goal, int *err)
{
#ifdef EXT3FS_DEBUG
	static unsigned long alloc_hits = 0, alloc_attempts = 0;
#endif
	unsigned long result;

#ifdef EXT3_PREALLOCATE
Linus Torvalds's avatar
Linus Torvalds committed
250
	struct ext3_inode_info *ei = EXT3_I(inode);
Linus Torvalds's avatar
Linus Torvalds committed
251
	/* Writer: ->i_prealloc* */
Linus Torvalds's avatar
Linus Torvalds committed
252 253 254
	if (ei->i_prealloc_count &&
	    (goal == ei->i_prealloc_block ||
	     goal + 1 == ei->i_prealloc_block))
Linus Torvalds's avatar
Linus Torvalds committed
255
	{
Linus Torvalds's avatar
Linus Torvalds committed
256 257
		result = ei->i_prealloc_block++;
		ei->i_prealloc_count--;
Linus Torvalds's avatar
Linus Torvalds committed
258 259 260 261 262 263 264 265 266
		/* Writer: end */
		ext3_debug ("preallocation hit (%lu/%lu).\n",
			    ++alloc_hits, ++alloc_attempts);
	} else {
		ext3_discard_prealloc (inode);
		ext3_debug ("preallocation miss (%lu/%lu).\n",
			    alloc_hits, ++alloc_attempts);
		if (S_ISREG(inode->i_mode))
			result = ext3_new_block (inode, goal, 
Linus Torvalds's avatar
Linus Torvalds committed
267 268
				 &ei->i_prealloc_count,
				 &ei->i_prealloc_block, err);
Linus Torvalds's avatar
Linus Torvalds committed
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
		else
			result = ext3_new_block (inode, goal, 0, 0, err);
		/*
		 * AKPM: this is somewhat sticky.  I'm not surprised it was
		 * disabled in 2.2's ext3.  Need to integrate b_committed_data
		 * guarding with preallocation, if indeed preallocation is
		 * effective.
		 */
	}
#else
	result = ext3_new_block (handle, inode, goal, 0, 0, err);
#endif
	return result;
}


typedef struct {
	u32	*p;
	u32	key;
	struct buffer_head *bh;
} Indirect;

static inline void add_chain(Indirect *p, struct buffer_head *bh, u32 *v)
{
	p->key = *(p->p = v);
	p->bh = bh;
}

static inline int verify_chain(Indirect *from, Indirect *to)
{
	while (from <= to && from->key == *from->p)
		from++;
	return (from > to);
}

/**
 *	ext3_block_to_path - parse the block number into array of offsets
 *	@inode: inode in question (we are only interested in its superblock)
 *	@i_block: block number to be parsed
 *	@offsets: array to store the offsets in
309 310
 *      @boundary: set this non-zero if the referred-to block is likely to be
 *             followed (on disk) by an indirect block.
Linus Torvalds's avatar
Linus Torvalds committed
311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334
 *
 *	To store the locations of file's data ext3 uses a data structure common
 *	for UNIX filesystems - tree of pointers anchored in the inode, with
 *	data blocks at leaves and indirect blocks in intermediate nodes.
 *	This function translates the block number into path in that tree -
 *	return value is the path length and @offsets[n] is the offset of
 *	pointer to (n+1)th node in the nth one. If @block is out of range
 *	(negative or too large) warning is printed and zero returned.
 *
 *	Note: function doesn't find node addresses, so no IO is needed. All
 *	we need to know is the capacity of indirect blocks (taken from the
 *	inode->i_sb).
 */

/*
 * Portability note: the last comparison (check that we fit into triple
 * indirect block) is spelled differently, because otherwise on an
 * architecture with 32-bit longs and 8Kb pages we might get into trouble
 * if our filesystem had 8Kb blocks. We might use long long, but that would
 * kill us on x86. Oh, well, at least the sign propagation does not matter -
 * i_block would have to be negative in the very beginning, so we would not
 * get there at all.
 */

335 336
static int ext3_block_to_path(struct inode *inode,
			long i_block, int offsets[4], int *boundary)
Linus Torvalds's avatar
Linus Torvalds committed
337 338 339 340 341 342 343
{
	int ptrs = EXT3_ADDR_PER_BLOCK(inode->i_sb);
	int ptrs_bits = EXT3_ADDR_PER_BLOCK_BITS(inode->i_sb);
	const long direct_blocks = EXT3_NDIR_BLOCKS,
		indirect_blocks = ptrs,
		double_blocks = (1 << (ptrs_bits * 2));
	int n = 0;
344
	int final = 0;
Linus Torvalds's avatar
Linus Torvalds committed
345 346 347 348 349

	if (i_block < 0) {
		ext3_warning (inode->i_sb, "ext3_block_to_path", "block < 0");
	} else if (i_block < direct_blocks) {
		offsets[n++] = i_block;
350
		final = direct_blocks;
Linus Torvalds's avatar
Linus Torvalds committed
351 352 353
	} else if ( (i_block -= direct_blocks) < indirect_blocks) {
		offsets[n++] = EXT3_IND_BLOCK;
		offsets[n++] = i_block;
354
		final = ptrs;
Linus Torvalds's avatar
Linus Torvalds committed
355 356 357 358
	} else if ((i_block -= indirect_blocks) < double_blocks) {
		offsets[n++] = EXT3_DIND_BLOCK;
		offsets[n++] = i_block >> ptrs_bits;
		offsets[n++] = i_block & (ptrs - 1);
359
		final = ptrs;
Linus Torvalds's avatar
Linus Torvalds committed
360 361 362 363 364
	} else if (((i_block -= double_blocks) >> (ptrs_bits * 2)) < ptrs) {
		offsets[n++] = EXT3_TIND_BLOCK;
		offsets[n++] = i_block >> (ptrs_bits * 2);
		offsets[n++] = (i_block >> ptrs_bits) & (ptrs - 1);
		offsets[n++] = i_block & (ptrs - 1);
365
		final = ptrs;
Linus Torvalds's avatar
Linus Torvalds committed
366 367 368
	} else {
		ext3_warning (inode->i_sb, "ext3_block_to_path", "block > big");
	}
369 370
	if (boundary)
		*boundary = (i_block & (ptrs - 1)) == (final - 1);
Linus Torvalds's avatar
Linus Torvalds committed
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
	return n;
}

/**
 *	ext3_get_branch - read the chain of indirect blocks leading to data
 *	@inode: inode in question
 *	@depth: depth of the chain (1 - direct pointer, etc.)
 *	@offsets: offsets of pointers in inode/indirect blocks
 *	@chain: place to store the result
 *	@err: here we store the error value
 *
 *	Function fills the array of triples <key, p, bh> and returns %NULL
 *	if everything went OK or the pointer to the last filled triple
 *	(incomplete one) otherwise. Upon the return chain[i].key contains
 *	the number of (i+1)-th block in the chain (as it is stored in memory,
 *	i.e. little-endian 32-bit), chain[i].p contains the address of that
 *	number (it points into struct inode for i==0 and into the bh->b_data
 *	for i>0) and chain[i].bh points to the buffer_head of i-th indirect
 *	block for i>0 and NULL for i==0. In other words, it holds the block
 *	numbers of the chain, addresses they were taken from (and where we can
 *	verify that chain did not change) and buffer_heads hosting these
 *	numbers.
 *
 *	Function stops when it stumbles upon zero pointer (absent block)
 *		(pointer to last triple returned, *@err == 0)
 *	or when it gets an IO error reading an indirect block
 *		(ditto, *@err == -EIO)
 *	or when it notices that chain had been changed while it was reading
 *		(ditto, *@err == -EAGAIN)
 *	or when it reads all @depth-1 indirect blocks successfully and finds
 *	the whole chain, all way to the data (returns %NULL, *err == 0).
 */
static Indirect *ext3_get_branch(struct inode *inode, int depth, int *offsets,
				 Indirect chain[4], int *err)
{
Linus Torvalds's avatar
Linus Torvalds committed
406
	struct super_block *sb = inode->i_sb;
Linus Torvalds's avatar
Linus Torvalds committed
407 408 409 410 411
	Indirect *p = chain;
	struct buffer_head *bh;

	*err = 0;
	/* i_data is not going away, no lock needed */
Linus Torvalds's avatar
Linus Torvalds committed
412
	add_chain (chain, NULL, EXT3_I(inode)->i_data + *offsets);
Linus Torvalds's avatar
Linus Torvalds committed
413 414 415
	if (!p->key)
		goto no_block;
	while (--depth) {
Linus Torvalds's avatar
Linus Torvalds committed
416
		bh = sb_bread(sb, le32_to_cpu(p->key));
Linus Torvalds's avatar
Linus Torvalds committed
417 418 419 420 421 422 423 424 425 426 427 428 429
		if (!bh)
			goto failure;
		/* Reader: pointers */
		if (!verify_chain(chain, p))
			goto changed;
		add_chain(++p, bh, (u32*)bh->b_data + *++offsets);
		/* Reader: end */
		if (!p->key)
			goto no_block;
	}
	return NULL;

changed:
Alexander Viro's avatar
Alexander Viro committed
430
	brelse(bh);
Linus Torvalds's avatar
Linus Torvalds committed
431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455
	*err = -EAGAIN;
	goto no_block;
failure:
	*err = -EIO;
no_block:
	return p;
}

/**
 *	ext3_find_near - find a place for allocation with sufficient locality
 *	@inode: owner
 *	@ind: descriptor of indirect block.
 *
 *	This function returns the prefered place for block allocation.
 *	It is used when heuristic for sequential allocation fails.
 *	Rules are:
 *	  + if there is a block to the left of our position - allocate near it.
 *	  + if pointer will live in indirect block - allocate near that block.
 *	  + if pointer will live in inode - allocate in the same
 *	    cylinder group. 
 *	Caller must make sure that @ind is valid and will stay that way.
 */

static inline unsigned long ext3_find_near(struct inode *inode, Indirect *ind)
{
Linus Torvalds's avatar
Linus Torvalds committed
456 457
	struct ext3_inode_info *ei = EXT3_I(inode);
	u32 *start = ind->bh ? (u32*) ind->bh->b_data : ei->i_data;
Linus Torvalds's avatar
Linus Torvalds committed
458 459 460 461 462 463 464 465 466 467 468 469 470 471 472
	u32 *p;

	/* Try to find previous block */
	for (p = ind->p - 1; p >= start; p--)
		if (*p)
			return le32_to_cpu(*p);

	/* No such thing, so let's try location of indirect block */
	if (ind->bh)
		return ind->bh->b_blocknr;

	/*
	 * It is going to be refered from inode itself? OK, just put it into
	 * the same cylinder group then.
	 */
Linus Torvalds's avatar
Linus Torvalds committed
473
	return (ei->i_block_group * EXT3_BLOCKS_PER_GROUP(inode->i_sb)) +
474
	       le32_to_cpu(EXT3_SB(inode->i_sb)->s_es->s_first_data_block);
Linus Torvalds's avatar
Linus Torvalds committed
475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492
}

/**
 *	ext3_find_goal - find a prefered place for allocation.
 *	@inode: owner
 *	@block:  block we want
 *	@chain:  chain of indirect blocks
 *	@partial: pointer to the last triple within a chain
 *	@goal:	place to store the result.
 *
 *	Normally this function find the prefered place for block allocation,
 *	stores it in *@goal and returns zero. If the branch had been changed
 *	under us we return -EAGAIN.
 */

static int ext3_find_goal(struct inode *inode, long block, Indirect chain[4],
			  Indirect *partial, unsigned long *goal)
{
Linus Torvalds's avatar
Linus Torvalds committed
493
	struct ext3_inode_info *ei = EXT3_I(inode);
Linus Torvalds's avatar
Linus Torvalds committed
494
	/* Writer: ->i_next_alloc* */
Linus Torvalds's avatar
Linus Torvalds committed
495 496 497
	if (block == ei->i_next_alloc_block + 1) {
		ei->i_next_alloc_block++;
		ei->i_next_alloc_goal++;
Linus Torvalds's avatar
Linus Torvalds committed
498 499
	}
#ifdef SEARCH_FROM_ZERO
Linus Torvalds's avatar
Linus Torvalds committed
500 501
	ei->i_next_alloc_block = 0;
	ei->i_next_alloc_goal = 0;
Linus Torvalds's avatar
Linus Torvalds committed
502 503 504 505 506 507 508 509
#endif
	/* Writer: end */
	/* Reader: pointers, ->i_next_alloc* */
	if (verify_chain(chain, partial)) {
		/*
		 * try the heuristic for sequential allocation,
		 * failing that at least try to get decent locality.
		 */
Linus Torvalds's avatar
Linus Torvalds committed
510 511
		if (block == ei->i_next_alloc_block)
			*goal = ei->i_next_alloc_goal;
Linus Torvalds's avatar
Linus Torvalds committed
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
		if (!*goal)
			*goal = ext3_find_near(inode, partial);
#ifdef SEARCH_FROM_ZERO
		*goal = 0;
#endif
		return 0;
	}
	/* Reader: end */
	return -EAGAIN;
}

/**
 *	ext3_alloc_branch - allocate and set up a chain of blocks.
 *	@inode: owner
 *	@num: depth of the chain (number of blocks to allocate)
 *	@offsets: offsets (in the blocks) to store the pointers to next.
 *	@branch: place to store the chain in.
 *
 *	This function allocates @num blocks, zeroes out all but the last one,
 *	links them into chain and (if we are synchronous) writes them to disk.
 *	In other words, it prepares a branch that can be spliced onto the
 *	inode. It stores the information about that chain in the branch[], in
 *	the same format as ext3_get_branch() would do. We are calling it after
 *	we had read the existing part of chain and partial points to the last
 *	triple of that (one with zero ->key). Upon the exit we have the same
 *	picture as after the successful ext3_get_block(), excpet that in one
 *	place chain is disconnected - *branch->p is still zero (we did not
 *	set the last link), but branch->key contains the number that should
 *	be placed into *branch->p to fill that gap.
 *
 *	If allocation fails we free all blocks we've allocated (and forget
 *	their buffer_heads) and return the error value the from failed
 *	ext3_alloc_block() (normally -ENOSPC). Otherwise we set the chain
 *	as described above and return 0.
 */

static int ext3_alloc_branch(handle_t *handle, struct inode *inode,
			     int num,
			     unsigned long goal,
			     int *offsets,
			     Indirect *branch)
{
	int blocksize = inode->i_sb->s_blocksize;
	int n = 0, keys = 0;
	int err = 0;
	int i;
	int parent = ext3_alloc_block(handle, inode, goal, &err);

	branch[0].key = cpu_to_le32(parent);
	if (parent) {
		for (n = 1; n < num; n++) {
			struct buffer_head *bh;
			/* Allocate the next block */
			int nr = ext3_alloc_block(handle, inode, parent, &err);
			if (!nr)
				break;
			branch[n].key = cpu_to_le32(nr);
			keys = n+1;
			
			/*
			 * Get buffer_head for parent block, zero it out
			 * and set the pointer to new one, then send
			 * parent to disk.  
			 */
Linus Torvalds's avatar
Linus Torvalds committed
576
			bh = sb_getblk(inode->i_sb, parent);
Linus Torvalds's avatar
Linus Torvalds committed
577 578 579 580 581 582 583 584 585 586 587 588 589 590
			branch[n].bh = bh;
			lock_buffer(bh);
			BUFFER_TRACE(bh, "call get_create_access");
			err = ext3_journal_get_create_access(handle, bh);
			if (err) {
				unlock_buffer(bh);
				brelse(bh);
				break;
			}

			memset(bh->b_data, 0, blocksize);
			branch[n].p = (u32*) bh->b_data + offsets[n];
			*branch[n].p = branch[n].key;
			BUFFER_TRACE(bh, "marking uptodate");
Andrew Morton's avatar
Andrew Morton committed
591
			set_buffer_uptodate(bh);
Linus Torvalds's avatar
Linus Torvalds committed
592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636
			unlock_buffer(bh);

			BUFFER_TRACE(bh, "call ext3_journal_dirty_metadata");
			err = ext3_journal_dirty_metadata(handle, bh);
			if (err)
				break;
			
			parent = nr;
		}
	}
	if (n == num)
		return 0;

	/* Allocation failed, free what we already allocated */
	for (i = 1; i < keys; i++) {
		BUFFER_TRACE(branch[i].bh, "call journal_forget");
		ext3_journal_forget(handle, branch[i].bh);
	}
	for (i = 0; i < keys; i++)
		ext3_free_blocks(handle, inode, le32_to_cpu(branch[i].key), 1);
	return err;
}

/**
 *	ext3_splice_branch - splice the allocated branch onto inode.
 *	@inode: owner
 *	@block: (logical) number of block we are adding
 *	@chain: chain of indirect blocks (with a missing link - see
 *		ext3_alloc_branch)
 *	@where: location of missing link
 *	@num:   number of blocks we are adding
 *
 *	This function verifies that chain (up to the missing link) had not
 *	changed, fills the missing link and does all housekeeping needed in
 *	inode (->i_blocks, etc.). In case of success we end up with the full
 *	chain to new block and return 0. Otherwise (== chain had been changed)
 *	we free the new blocks (forgetting their buffer_heads, indeed) and
 *	return -EAGAIN.
 */

static int ext3_splice_branch(handle_t *handle, struct inode *inode, long block,
			      Indirect chain[4], Indirect *where, int num)
{
	int i;
	int err = 0;
Linus Torvalds's avatar
Linus Torvalds committed
637
	struct ext3_inode_info *ei = EXT3_I(inode);
Linus Torvalds's avatar
Linus Torvalds committed
638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659

	/*
	 * If we're splicing into a [td]indirect block (as opposed to the
	 * inode) then we need to get write access to the [td]indirect block
	 * before the splice.
	 */
	if (where->bh) {
		BUFFER_TRACE(where->bh, "get_write_access");
		err = ext3_journal_get_write_access(handle, where->bh);
		if (err)
			goto err_out;
	}
	/* Verify that place we are splicing to is still there and vacant */

	/* Writer: pointers, ->i_next_alloc* */
	if (!verify_chain(chain, where-1) || *where->p)
		/* Writer: end */
		goto changed;

	/* That's it */

	*where->p = where->key;
Linus Torvalds's avatar
Linus Torvalds committed
660 661
	ei->i_next_alloc_block = block;
	ei->i_next_alloc_goal = le32_to_cpu(where[num-1].key);
Linus Torvalds's avatar
Linus Torvalds committed
662
#ifdef SEARCH_FROM_ZERO
Linus Torvalds's avatar
Linus Torvalds committed
663 664
	ei->i_next_alloc_block = 0;
	ei->i_next_alloc_goal = 0;
Linus Torvalds's avatar
Linus Torvalds committed
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
#endif
	/* Writer: end */

	/* We are done with atomic stuff, now do the rest of housekeeping */

	inode->i_ctime = CURRENT_TIME;
	ext3_mark_inode_dirty(handle, inode);

	/* had we spliced it onto indirect block? */
	if (where->bh) {
		/*
		 * akpm: If we spliced it onto an indirect block, we haven't
		 * altered the inode.  Note however that if it is being spliced
		 * onto an indirect block at the very end of the file (the
		 * file is growing) then we *will* alter the inode to reflect
		 * the new i_size.  But that is not done here - it is done in
		 * generic_commit_write->__mark_inode_dirty->ext3_dirty_inode.
		 */
		jbd_debug(5, "splicing indirect only\n");
		BUFFER_TRACE(where->bh, "call ext3_journal_dirty_metadata");
		err = ext3_journal_dirty_metadata(handle, where->bh);
		if (err) 
			goto err_out;
	} else {
		/*
		 * OK, we spliced it into the inode itself on a direct block.
		 * Inode was dirtied above.
		 */
		jbd_debug(5, "splicing direct\n");
	}
	return err;

changed:
	/*
	 * AKPM: if where[i].bh isn't part of the current updating
	 * transaction then we explode nastily.  Test this code path.
	 */
	jbd_debug(1, "the chain changed: try again\n");
	err = -EAGAIN;
	
err_out:
	for (i = 1; i < num; i++) {
		BUFFER_TRACE(where[i].bh, "call journal_forget");
		ext3_journal_forget(handle, where[i].bh);
	}
	/* For the normal collision cleanup case, we free up the blocks.
	 * On genuine filesystem errors we don't even think about doing
	 * that. */
	if (err == -EAGAIN)
		for (i = 0; i < num; i++)
			ext3_free_blocks(handle, inode, 
					 le32_to_cpu(where[i].key), 1);
	return err;
}

/*
 * Allocation strategy is simple: if we have to allocate something, we will
 * have to go the whole way to leaf. So let's do it before attaching anything
 * to tree, set linkage between the newborn blocks, write them if sync is
 * required, recheck the path, free and repeat if check fails, otherwise
 * set the last missing link (that will protect us from any truncate-generated
 * removals - all blocks on the path are immune now) and possibly force the
 * write on the parent block.
 * That has a nice additional property: no special recovery from the failed
 * allocations is needed - we simply release blocks and do not touch anything
 * reachable from inode.
 *
 * akpm: `handle' can be NULL if create == 0.
Linus Torvalds's avatar
Linus Torvalds committed
733 734
 *
 * The BKL may not be held on entry here.  Be sure to take it early.
Linus Torvalds's avatar
Linus Torvalds committed
735 736
 */

Andrew Morton's avatar
Andrew Morton committed
737 738 739
static int
ext3_get_block_handle(handle_t *handle, struct inode *inode, sector_t iblock,
		struct buffer_head *bh_result, int create, int extend_disksize)
Linus Torvalds's avatar
Linus Torvalds committed
740 741 742 743 744 745 746
{
	int err = -EIO;
	int offsets[4];
	Indirect chain[4];
	Indirect *partial;
	unsigned long goal;
	int left;
747 748
	int boundary = 0;
	int depth = ext3_block_to_path(inode, iblock, offsets, &boundary);
Linus Torvalds's avatar
Linus Torvalds committed
749
	struct ext3_inode_info *ei = EXT3_I(inode);
Linus Torvalds's avatar
Linus Torvalds committed
750 751 752 753 754 755 756 757 758 759 760 761 762
	loff_t new_size;

	J_ASSERT(handle != NULL || create == 0);

	if (depth == 0)
		goto out;

	lock_kernel();
reread:
	partial = ext3_get_branch(inode, depth, offsets, chain, &err);

	/* Simplest case - block found, no allocation needed */
	if (!partial) {
Andrew Morton's avatar
Andrew Morton committed
763
		clear_buffer_new(bh_result);
Linus Torvalds's avatar
Linus Torvalds committed
764
got_it:
Linus Torvalds's avatar
Linus Torvalds committed
765
		map_bh(bh_result, inode->i_sb, le32_to_cpu(chain[depth-1].key));
766 767
		if (boundary)
			set_buffer_boundary(bh_result);
Linus Torvalds's avatar
Linus Torvalds committed
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
		/* Clean up and exit */
		partial = chain+depth-1; /* the whole chain */
		goto cleanup;
	}

	/* Next simple case - plain lookup or failed read of indirect block */
	if (!create || err == -EIO) {
cleanup:
		while (partial > chain) {
			BUFFER_TRACE(partial->bh, "call brelse");
			brelse(partial->bh);
			partial--;
		}
		BUFFER_TRACE(bh_result, "returned");
		unlock_kernel();
out:
		return err;
	}

	/*
	 * Indirect block might be removed by truncate while we were
	 * reading it. Handling of that case (forget what we've got and
	 * reread) is taken out of the main path.
	 */
	if (err == -EAGAIN)
		goto changed;

	if (ext3_find_goal(inode, iblock, chain, partial, &goal) < 0)
		goto changed;

	left = (chain + depth) - partial;

	/*
	 * Block out ext3_truncate while we alter the tree
	 */
Linus Torvalds's avatar
Linus Torvalds committed
803
	down_read(&ei->truncate_sem);
Linus Torvalds's avatar
Linus Torvalds committed
804 805 806 807 808 809 810 811 812 813 814
	err = ext3_alloc_branch(handle, inode, left, goal,
					offsets+(partial-chain), partial);

	/* The ext3_splice_branch call will free and forget any buffers
	 * on the new chain if there is a failure, but that risks using
	 * up transaction credits, especially for bitmaps where the
	 * credits cannot be returned.  Can we handle this somehow?  We
	 * may need to return -EAGAIN upwards in the worst case.  --sct */
	if (!err)
		err = ext3_splice_branch(handle, inode, iblock, chain,
					 partial, left);
Linus Torvalds's avatar
Linus Torvalds committed
815
	up_read(&ei->truncate_sem);
Linus Torvalds's avatar
Linus Torvalds committed
816 817 818 819 820
	if (err == -EAGAIN)
		goto changed;
	if (err)
		goto cleanup;

Andrew Morton's avatar
Andrew Morton committed
821 822 823 824 825 826 827 828 829 830 831
	if (extend_disksize) {
		/*
		 * This is not racy against ext3_truncate's modification of
		 * i_disksize because VM/VFS ensures that the file cannot be
		 * extended while truncate is in progress.  It is racy between
		 * multiple parallel instances of get_block, but we have BKL.
		 */
		new_size = inode->i_size;
		if (new_size > ei->i_disksize)
			ei->i_disksize = new_size;
	}
Andrew Morton's avatar
Andrew Morton committed
832
	set_buffer_new(bh_result);
Linus Torvalds's avatar
Linus Torvalds committed
833 834 835 836 837 838 839 840 841 842 843 844
	goto got_it;

changed:
	while (partial > chain) {
		jbd_debug(1, "buffer chain changed, retrying\n");
		BUFFER_TRACE(partial->bh, "brelsing");
		brelse(partial->bh);
		partial--;
	}
	goto reread;
}

Linus Torvalds's avatar
Linus Torvalds committed
845
static int ext3_get_block(struct inode *inode, sector_t iblock,
Linus Torvalds's avatar
Linus Torvalds committed
846 847 848 849 850 851 852 853 854
			struct buffer_head *bh_result, int create)
{
	handle_t *handle = 0;
	int ret;

	if (create) {
		handle = ext3_journal_current_handle();
		J_ASSERT(handle != 0);
	}
Andrew Morton's avatar
Andrew Morton committed
855 856
	ret = ext3_get_block_handle(handle, inode, iblock,
				bh_result, create, 1);
Linus Torvalds's avatar
Linus Torvalds committed
857 858 859
	return ret;
}

Andrew Morton's avatar
Andrew Morton committed
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
#define DIO_CREDITS (EXT3_RESERVE_TRANS_BLOCKS + 32)

static int
ext3_direct_io_get_blocks(struct inode *inode, sector_t iblock,
		unsigned long max_blocks, struct buffer_head *bh_result,
		int create)
{
	handle_t *handle = journal_current_handle();
	int ret = 0;

	lock_kernel();
	if (handle && handle->h_buffer_credits <= EXT3_RESERVE_TRANS_BLOCKS) {
		/*
		 * Getting low on buffer credits...
		 */
		if (!ext3_journal_extend(handle, DIO_CREDITS)) {
			/*
			 * Couldn't extend the transaction.  Start a new one
			 */
			ret = ext3_journal_restart(handle, DIO_CREDITS);
		}
	}
	if (ret == 0)
		ret = ext3_get_block_handle(handle, inode, iblock,
					bh_result, create, 0);
	if (ret == 0)
		bh_result->b_size = (1 << inode->i_blkbits);
	unlock_kernel();
	return ret;
}


Linus Torvalds's avatar
Linus Torvalds committed
892 893 894 895 896 897 898 899 900 901 902 903 904 905
/*
 * `handle' can be NULL if create is zero
 */
struct buffer_head *ext3_getblk(handle_t *handle, struct inode * inode,
				long block, int create, int * errp)
{
	struct buffer_head dummy;
	int fatal = 0, err;
	
	J_ASSERT(handle != NULL || create == 0);

	dummy.b_state = 0;
	dummy.b_blocknr = -1000;
	buffer_trace_init(&dummy.b_history);
Andrew Morton's avatar
Andrew Morton committed
906
	*errp = ext3_get_block_handle(handle, inode, block, &dummy, create, 1);
Linus Torvalds's avatar
Linus Torvalds committed
907 908
	if (!*errp && buffer_mapped(&dummy)) {
		struct buffer_head *bh;
Linus Torvalds's avatar
Linus Torvalds committed
909
		bh = sb_getblk(inode->i_sb, dummy.b_blocknr);
Linus Torvalds's avatar
Linus Torvalds committed
910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926
		if (buffer_new(&dummy)) {
			J_ASSERT(create != 0);
			J_ASSERT(handle != 0);

			/* Now that we do not always journal data, we
			   should keep in mind whether this should
			   always journal the new buffer as metadata.
			   For now, regular file writes use
			   ext3_get_block instead, so it's not a
			   problem. */
			lock_kernel();
			lock_buffer(bh);
			BUFFER_TRACE(bh, "call get_create_access");
			fatal = ext3_journal_get_create_access(handle, bh);
			if (!fatal) {
				memset(bh->b_data, 0,
				       inode->i_sb->s_blocksize);
Andrew Morton's avatar
Andrew Morton committed
927
				set_buffer_uptodate(bh);
Linus Torvalds's avatar
Linus Torvalds committed
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
			}
			unlock_buffer(bh);
			BUFFER_TRACE(bh, "call ext3_journal_dirty_metadata");
			err = ext3_journal_dirty_metadata(handle, bh);
			if (!fatal) fatal = err;
			unlock_kernel();
		} else {
			BUFFER_TRACE(bh, "not a new buffer");
		}
		if (fatal) {
			*errp = fatal;
			brelse(bh);
			bh = NULL;
		}
		return bh;
	}
	return NULL;
}

struct buffer_head *ext3_bread(handle_t *handle, struct inode * inode,
			       int block, int create, int *err)
{
	struct buffer_head * bh;
	int prev_blocks;

	prev_blocks = inode->i_blocks;

	bh = ext3_getblk (handle, inode, block, create, err);
	if (!bh)
		return bh;
#ifdef EXT3_PREALLOCATE
	/*
	 * If the inode has grown, and this is a directory, then use a few
	 * more of the preallocated blocks to keep directory fragmentation
	 * down.  The preallocated blocks are guaranteed to be contiguous.
	 */
	if (create &&
	    S_ISDIR(inode->i_mode) &&
	    inode->i_blocks > prev_blocks &&
	    EXT3_HAS_COMPAT_FEATURE(inode->i_sb,
				    EXT3_FEATURE_COMPAT_DIR_PREALLOC)) {
		int i;
		struct buffer_head *tmp_bh;

		for (i = 1;
Linus Torvalds's avatar
Linus Torvalds committed
973
		     EXT3_I(inode)->i_prealloc_count &&
Linus Torvalds's avatar
Linus Torvalds committed
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 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067
		     i < EXT3_SB(inode->i_sb)->s_es->s_prealloc_dir_blocks;
		     i++) {
			/*
			 * ext3_getblk will zero out the contents of the
			 * directory for us
			 */
			tmp_bh = ext3_getblk(handle, inode,
						block+i, create, err);
			if (!tmp_bh) {
				brelse (bh);
				return 0;
			}
			brelse (tmp_bh);
		}
	}
#endif
	if (buffer_uptodate(bh))
		return bh;
	ll_rw_block (READ, 1, &bh);
	wait_on_buffer (bh);
	if (buffer_uptodate(bh))
		return bh;
	brelse (bh);
	*err = -EIO;
	return NULL;
}

static int walk_page_buffers(	handle_t *handle,
				struct buffer_head *head,
				unsigned from,
				unsigned to,
				int *partial,
				int (*fn)(	handle_t *handle,
						struct buffer_head *bh))
{
	struct buffer_head *bh;
	unsigned block_start, block_end;
	unsigned blocksize = head->b_size;
	int err, ret = 0;

	for (	bh = head, block_start = 0;
		ret == 0 && (bh != head || !block_start);
	    	block_start = block_end, bh = bh->b_this_page)
	{
		block_end = block_start + blocksize;
		if (block_end <= from || block_start >= to) {
			if (partial && !buffer_uptodate(bh))
				*partial = 1;
			continue;
		}
		err = (*fn)(handle, bh);
		if (!ret)
			ret = err;
	}
	return ret;
}

/*
 * To preserve ordering, it is essential that the hole instantiation and
 * the data write be encapsulated in a single transaction.  We cannot
 * close off a transaction and start a new one between the ext3_get_block()
 * and the commit_write().  So doing the journal_start at the start of
 * prepare_write() is the right place.
 *
 * Also, this function can nest inside ext3_writepage() ->
 * block_write_full_page(). In that case, we *know* that ext3_writepage()
 * has generated enough buffer credits to do the whole page.  So we won't
 * block on the journal in that case, which is good, because the caller may
 * be PF_MEMALLOC.
 *
 * By accident, ext3 can be reentered when a transaction is open via
 * quota file writes.  If we were to commit the transaction while thus
 * reentered, there can be a deadlock - we would be holding a quota
 * lock, and the commit would never complete if another thread had a
 * transaction open and was blocking on the quota lock - a ranking
 * violation.
 *
 * So what we do is to rely on the fact that journal_stop/journal_start
 * will _not_ run commit under these circumstances because handle->h_ref
 * is elevated.  We'll still have enough credits for the tiny quotafile
 * write.  
 */

static int do_journal_get_write_access(handle_t *handle, 
				       struct buffer_head *bh)
{
	return ext3_journal_get_write_access(handle, bh);
}

static int ext3_prepare_write(struct file *file, struct page *page,
			      unsigned from, unsigned to)
{
	struct inode *inode = page->mapping->host;
	int ret, needed_blocks = ext3_writepage_trans_blocks(inode);
1068
	handle_t *handle;
Linus Torvalds's avatar
Linus Torvalds committed
1069 1070 1071 1072 1073 1074 1075

	lock_kernel();
	handle = ext3_journal_start(inode, needed_blocks);
	if (IS_ERR(handle)) {
		ret = PTR_ERR(handle);
		goto out;
	}
Linus Torvalds's avatar
Linus Torvalds committed
1076
	unlock_kernel();
Linus Torvalds's avatar
Linus Torvalds committed
1077
	ret = block_prepare_write(page, from, to, ext3_get_block);
Linus Torvalds's avatar
Linus Torvalds committed
1078
	lock_kernel();
Linus Torvalds's avatar
Linus Torvalds committed
1079 1080 1081
	if (ret != 0)
		goto prepare_write_failed;

Linus Torvalds's avatar
Linus Torvalds committed
1082
	if (ext3_should_journal_data(inode)) {
1083
		ret = walk_page_buffers(handle, page_buffers(page),
Linus Torvalds's avatar
Linus Torvalds committed
1084
				from, to, NULL, do_journal_get_write_access);
Linus Torvalds's avatar
Linus Torvalds committed
1085
	}
Linus Torvalds's avatar
Linus Torvalds committed
1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112
prepare_write_failed:
	if (ret)
		ext3_journal_stop(handle, inode);
out:
	unlock_kernel();
	return ret;
}

static int journal_dirty_sync_data(handle_t *handle, struct buffer_head *bh)
{
	return ext3_journal_dirty_data(handle, bh, 0);
}

/*
 * For ext3_writepage().  We also brelse() the buffer to account for
 * the bget() which ext3_writepage() performs.
 */
static int journal_dirty_async_data(handle_t *handle, struct buffer_head *bh)
{
	int ret = ext3_journal_dirty_data(handle, bh, 1);
	__brelse(bh);
	return ret;
}

/* For commit_write() in data=journal mode */
static int commit_write_fn(handle_t *handle, struct buffer_head *bh)
{
Andrew Morton's avatar
Andrew Morton committed
1113
	set_buffer_uptodate(bh);
Linus Torvalds's avatar
Linus Torvalds committed
1114 1115 1116 1117 1118
	return ext3_journal_dirty_metadata(handle, bh);
}

/*
 * We need to pick up the new inode size which generic_commit_write gave us
1119
 * `file' can be NULL - eg, when called from page_symlink().
Linus Torvalds's avatar
Linus Torvalds committed
1120
 *
1121 1122
 * ext3 never places buffers on inode->i_mapping->private_list.  metadata
 * buffers are managed internally.
Linus Torvalds's avatar
Linus Torvalds committed
1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139
 */

static int ext3_commit_write(struct file *file, struct page *page,
			     unsigned from, unsigned to)
{
	handle_t *handle = ext3_journal_current_handle();
	struct inode *inode = page->mapping->host;
	int ret = 0, ret2;

	lock_kernel();
	if (ext3_should_journal_data(inode)) {
		/*
		 * Here we duplicate the generic_commit_write() functionality
		 */
		int partial = 0;
		loff_t pos = ((loff_t)page->index << PAGE_CACHE_SHIFT) + to;

1140
		ret = walk_page_buffers(handle, page_buffers(page),
Linus Torvalds's avatar
Linus Torvalds committed
1141 1142 1143 1144 1145
			from, to, &partial, commit_write_fn);
		if (!partial)
			SetPageUptodate(page);
		if (pos > inode->i_size)
			inode->i_size = pos;
Linus Torvalds's avatar
Linus Torvalds committed
1146
		EXT3_I(inode)->i_state |= EXT3_STATE_JDATA;
Linus Torvalds's avatar
Linus Torvalds committed
1147 1148
	} else {
		if (ext3_should_order_data(inode)) {
1149
			ret = walk_page_buffers(handle, page_buffers(page),
Linus Torvalds's avatar
Linus Torvalds committed
1150 1151 1152 1153
				from, to, NULL, journal_dirty_sync_data);
		}
		/* Be careful here if generic_commit_write becomes a
		 * required invocation after block_prepare_write. */
1154
		if (ret == 0)
Linus Torvalds's avatar
Linus Torvalds committed
1155 1156
			ret = generic_commit_write(file, page, from, to);
	}
Linus Torvalds's avatar
Linus Torvalds committed
1157 1158
	if (inode->i_size > EXT3_I(inode)->i_disksize) {
		EXT3_I(inode)->i_disksize = inode->i_size;
Linus Torvalds's avatar
Linus Torvalds committed
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
		ret2 = ext3_mark_inode_dirty(handle, inode);
		if (!ret) 
			ret = ret2;
	}
	ret2 = ext3_journal_stop(handle, inode);
	unlock_kernel();
	if (!ret)
		ret = ret2;
	return ret;
}

/* 
 * bmap() is special.  It gets used by applications such as lilo and by
 * the swapper to find the on-disk block of a specific piece of data.
 *
 * Naturally, this is dangerous if the block concerned is still in the
 * journal.  If somebody makes a swapfile on an ext3 data-journaling
 * filesystem and enables swap, then they may get a nasty shock when the
 * data getting swapped to that swapfile suddenly gets overwritten by
 * the original zero's written out previously to the journal and
 * awaiting writeback in the kernel's buffer cache. 
 *
 * So, if we see any bmap calls here on a modified, data-journaled file,
 * take extra steps to flush any blocks which might be in the cache. 
 */
static int ext3_bmap(struct address_space *mapping, long block)
{
	struct inode *inode = mapping->host;
	journal_t *journal;
	int err;
	
Linus Torvalds's avatar
Linus Torvalds committed
1190
	if (EXT3_I(inode)->i_state & EXT3_STATE_JDATA) {
Linus Torvalds's avatar
Linus Torvalds committed
1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208
		/* 
		 * This is a REALLY heavyweight approach, but the use of
		 * bmap on dirty files is expected to be extremely rare:
		 * only if we run lilo or swapon on a freshly made file
		 * do we expect this to happen. 
		 *
		 * (bmap requires CAP_SYS_RAWIO so this does not
		 * represent an unprivileged user DOS attack --- we'd be
		 * in trouble if mortal users could trigger this path at
		 * will.) 
		 *
		 * NB. EXT3_STATE_JDATA is not set on files other than
		 * regular files.  If somebody wants to bmap a directory
		 * or symlink and gets confused because the buffer
		 * hasn't yet been flushed to disk, they deserve
		 * everything they get.
		 */
		
Linus Torvalds's avatar
Linus Torvalds committed
1209
		EXT3_I(inode)->i_state &= ~EXT3_STATE_JDATA;
Linus Torvalds's avatar
Linus Torvalds committed
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 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279
		journal = EXT3_JOURNAL(inode);
		journal_lock_updates(journal);
		err = journal_flush(journal);
		journal_unlock_updates(journal);
		
		if (err)
			return 0;
	}
	
	return generic_block_bmap(mapping,block,ext3_get_block);
}

static int bget_one(handle_t *handle, struct buffer_head *bh)
{
	atomic_inc(&bh->b_count);
	return 0;
}

/*
 * Note that we always start a transaction even if we're not journalling
 * data.  This is to preserve ordering: any hole instantiation within
 * __block_write_full_page -> ext3_get_block() should be journalled
 * along with the data so we don't crash and then get metadata which
 * refers to old data.
 *
 * In all journalling modes block_write_full_page() will start the I/O.
 *
 * Problem:
 *
 *	ext3_writepage() -> kmalloc() -> __alloc_pages() -> page_launder() ->
 *		ext3_writepage()
 *
 * Similar for:
 *
 *	ext3_file_write() -> generic_file_write() -> __alloc_pages() -> ...
 *
 * Same applies to ext3_get_block().  We will deadlock on various things like
 * lock_journal and i_truncate_sem.
 *
 * Setting PF_MEMALLOC here doesn't work - too many internal memory
 * allocations fail.
 *
 * 16May01: If we're reentered then journal_current_handle() will be
 *	    non-zero. We simply *return*.
 *
 * 1 July 2001: @@@ FIXME:
 *   In journalled data mode, a data buffer may be metadata against the
 *   current transaction.  But the same file is part of a shared mapping
 *   and someone does a writepage() on it.
 *
 *   We will move the buffer onto the async_data list, but *after* it has
 *   been dirtied. So there's a small window where we have dirty data on
 *   BJ_Metadata.
 *
 *   Note that this only applies to the last partial page in the file.  The
 *   bit which block_write_full_page() uses prepare/commit for.  (That's
 *   broken code anyway: it's wrong for msync()).
 *
 *   It's a rare case: affects the final partial page, for journalled data
 *   where the file is subject to bith write() and writepage() in the same
 *   transction.  To fix it we'll need a custom block_write_full_page().
 *   We'll probably need that anyway for journalling writepage() output.
 *
 * We don't honour synchronous mounts for writepage().  That would be
 * disastrous.  Any write() or metadata operation will sync the fs for
 * us.
 */
static int ext3_writepage(struct page *page)
{
	struct inode *inode = page->mapping->host;
1280
	struct buffer_head *page_bufs;
Linus Torvalds's avatar
Linus Torvalds committed
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
	handle_t *handle = NULL;
	int ret = 0, err;
	int needed;
	int order_data;

	J_ASSERT(PageLocked(page));
	
	/*
	 * We give up here if we're reentered, because it might be
	 * for a different filesystem.  One *could* look for a
	 * nested transaction opportunity.
	 */
	lock_kernel();
	if (ext3_journal_current_handle())
		goto out_fail;

	needed = ext3_writepage_trans_blocks(inode);
	if (current->flags & PF_MEMALLOC)
		handle = ext3_journal_try_start(inode, needed);
	else
		handle = ext3_journal_start(inode, needed);
				
	if (IS_ERR(handle)) {
		ret = PTR_ERR(handle);
		goto out_fail;
	}

	order_data = ext3_should_order_data(inode) ||
			ext3_should_journal_data(inode);

	unlock_kernel();

1313
	page_bufs = NULL;	/* Purely to prevent compiler warning */
Linus Torvalds's avatar
Linus Torvalds committed
1314 1315 1316

	/* bget() all the buffers */
	if (order_data) {
1317
		if (!page_has_buffers(page)) {
Andrew Morton's avatar
Andrew Morton committed
1318
			if (!PageUptodate(page))
1319 1320 1321 1322 1323
				buffer_error();
			create_empty_buffers(page,
				inode->i_sb->s_blocksize,
				(1 << BH_Dirty)|(1 << BH_Uptodate));
		}
1324 1325
		page_bufs = page_buffers(page);
		walk_page_buffers(handle, page_bufs, 0,
Linus Torvalds's avatar
Linus Torvalds committed
1326 1327 1328 1329 1330 1331 1332 1333
				PAGE_CACHE_SIZE, NULL, bget_one);
	}

	ret = block_write_full_page(page, ext3_get_block);

	/*
	 * The page can become unlocked at any point now, and
	 * truncate can then come in and change things.  So we
1334
	 * can't touch *page from now on.  But *page_bufs is
Linus Torvalds's avatar
Linus Torvalds committed
1335 1336 1337 1338 1339 1340 1341 1342
	 * safe due to elevated refcount.
	 */

	handle = ext3_journal_current_handle();
	lock_kernel();

	/* And attach them to the current transaction */
	if (order_data) {
1343
		err = walk_page_buffers(handle, page_bufs,
Linus Torvalds's avatar
Linus Torvalds committed
1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357
			0, PAGE_CACHE_SIZE, NULL, journal_dirty_async_data);
		if (!ret)
			ret = err;
	}

	err = ext3_journal_stop(handle, inode);
	if (!ret)
		ret = err;
	unlock_kernel();
	return ret;

out_fail:
	
	unlock_kernel();
1358 1359 1360

	/*
	 * We have to fail this writepage to avoid cross-fs transactions.
1361 1362 1363
	 * Return EAGAIN so the caller will the page back on
	 * mapping->dirty_pages.  The page's buffers' dirty state will be left
	 * as-is.
1364
	 */
1365
	ret = -EAGAIN;
Andrew Morton's avatar
Andrew Morton committed
1366
	unlock_page(page);
Linus Torvalds's avatar
Linus Torvalds committed
1367 1368 1369 1370 1371
	return ret;
}

static int ext3_readpage(struct file *file, struct page *page)
{
1372
	return mpage_readpage(page, ext3_get_block);
Linus Torvalds's avatar
Linus Torvalds committed
1373 1374
}

1375 1376 1377 1378 1379 1380
static int
ext3_readpages(struct address_space *mapping,
		struct list_head *pages, unsigned nr_pages)
{
	return mpage_readpages(mapping, pages, nr_pages, ext3_get_block);
}
Linus Torvalds's avatar
Linus Torvalds committed
1381

1382
static int ext3_invalidatepage(struct page *page, unsigned long offset)
Linus Torvalds's avatar
Linus Torvalds committed
1383 1384
{
	journal_t *journal = EXT3_JOURNAL(page->mapping->host);
1385
	return journal_invalidatepage(journal, page, offset);
Linus Torvalds's avatar
Linus Torvalds committed
1386 1387 1388 1389 1390 1391 1392 1393
}

static int ext3_releasepage(struct page *page, int wait)
{
	journal_t *journal = EXT3_JOURNAL(page->mapping->host);
	return journal_try_to_free_buffers(journal, page, wait);
}

Andrew Morton's avatar
Andrew Morton committed
1394 1395 1396 1397 1398 1399 1400 1401
/*
 * If the O_DIRECT write will extend the file then add this inode to the
 * orphan list.  So recovery will truncate it back to the original size
 * if the machine crashes during the write.
 *
 * If the O_DIRECT write is intantiating holes inside i_size and the machine
 * crashes then stale disk data _may_ be exposed inside the file.
 */
Andrew Morton's avatar
Andrew Morton committed
1402 1403 1404
static int ext3_direct_IO(int rw, struct inode *inode,
			const struct iovec *iov, loff_t offset,
			unsigned long nr_segs)
Andrew Morton's avatar
Andrew Morton committed
1405 1406 1407 1408 1409
{
	struct ext3_inode_info *ei = EXT3_I(inode);
	handle_t *handle = NULL;
	int ret;
	int orphan = 0;
Andrew Morton's avatar
Andrew Morton committed
1410
	size_t count = iov_length(iov, nr_segs);
Andrew Morton's avatar
Andrew Morton committed
1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432

	if (rw == WRITE) {
		loff_t final_size = offset + count;

		lock_kernel();
		handle = ext3_journal_start(inode, DIO_CREDITS);
		unlock_kernel();
		if (IS_ERR(handle)) {
			ret = PTR_ERR(handle);
			goto out;
		}
		if (final_size > inode->i_size) {
			lock_kernel();
			ret = ext3_orphan_add(handle, inode);
			unlock_kernel();
			if (ret)
				goto out_stop;
			orphan = 1;
			ei->i_disksize = inode->i_size;
		}
	}

Andrew Morton's avatar
Andrew Morton committed
1433 1434
	ret = generic_direct_IO(rw, inode, iov, offset,
				nr_segs, ext3_direct_io_get_blocks);
Andrew Morton's avatar
Andrew Morton committed
1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460

out_stop:
	if (handle) {
		int err;

		lock_kernel();
		if (orphan) 
			ext3_orphan_del(handle, inode);
		if (orphan && ret > 0) {
			loff_t end = offset + ret;
			if (end > inode->i_size) {
				ei->i_disksize = end;
				inode->i_size = end;
				err = ext3_mark_inode_dirty(handle, inode);
				if (!ret) 
					ret = err;
			}
		}
		err = ext3_journal_stop(handle, inode);
		if (ret == 0)
			ret = err;
		unlock_kernel();
	}
out:
	return ret;
}
Linus Torvalds's avatar
Linus Torvalds committed
1461 1462

struct address_space_operations ext3_aops = {
Andrew Morton's avatar
Andrew Morton committed
1463 1464 1465
	.readpage	= ext3_readpage,	/* BKL not held.  Don't need */
	.readpages	= ext3_readpages,	/* BKL not held.  Don't need */
	.writepage	= ext3_writepage,	/* BKL not held.  We take it */
1466 1467 1468 1469 1470 1471
	.sync_page	= block_sync_page,
	.prepare_write	= ext3_prepare_write,	/* BKL not held.  We take it */
	.commit_write	= ext3_commit_write,	/* BKL not held.  We take it */
	.bmap		= ext3_bmap,		/* BKL held */
	.invalidatepage	= ext3_invalidatepage,	/* BKL not held.  Don't need */
	.releasepage	= ext3_releasepage,	/* BKL not held.  Don't need */
Andrew Morton's avatar
Andrew Morton committed
1472
	.direct_IO	= ext3_direct_IO,	/* BKL not held.  Don't need */
Linus Torvalds's avatar
Linus Torvalds committed
1473 1474
};

1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490
/* For writeback mode, we can use mpage_writepages() */

static int
ext3_writepages(struct address_space *mapping, int *nr_to_write)
{
	int ret;
	int err;

	ret = write_mapping_buffers(mapping);
	err = mpage_writepages(mapping, nr_to_write, ext3_get_block);
	if (!ret)
		ret = err;
	return ret;
}

struct address_space_operations ext3_writeback_aops = {
1491 1492 1493
	.readpage	= ext3_readpage,	/* BKL not held.  Don't need */
	.readpages	= ext3_readpages,	/* BKL not held.  Don't need */
	.writepage	= ext3_writepage,	/* BKL not held.  We take it */
1494 1495 1496 1497 1498 1499 1500
	.writepages	= ext3_writepages,	/* BKL not held.  Don't need */
	.sync_page	= block_sync_page,
	.prepare_write	= ext3_prepare_write,	/* BKL not held.  We take it */
	.commit_write	= ext3_commit_write,	/* BKL not held.  We take it */
	.bmap		= ext3_bmap,		/* BKL held */
	.invalidatepage	= ext3_invalidatepage,	/* BKL not held.  Don't need */
	.releasepage	= ext3_releasepage,	/* BKL not held.  Don't need */
Andrew Morton's avatar
Andrew Morton committed
1501
	.direct_IO	= ext3_direct_IO,	/* BKL not held.  Don't need */
1502 1503
};

Linus Torvalds's avatar
Linus Torvalds committed
1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519
/*
 * ext3_block_truncate_page() zeroes out a mapping from file offset `from'
 * up to the end of the block which corresponds to `from'.
 * This required during truncate. We need to physically zero the tail end
 * of that block so it doesn't yield old data if the file is later grown.
 */
static int ext3_block_truncate_page(handle_t *handle,
		struct address_space *mapping, loff_t from)
{
	unsigned long index = from >> PAGE_CACHE_SHIFT;
	unsigned offset = from & (PAGE_CACHE_SIZE-1);
	unsigned blocksize, iblock, length, pos;
	struct inode *inode = mapping->host;
	struct page *page;
	struct buffer_head *bh;
	int err;
1520
	void *kaddr;
Linus Torvalds's avatar
Linus Torvalds committed
1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536

	blocksize = inode->i_sb->s_blocksize;
	length = offset & (blocksize - 1);

	/* Block boundary? Nothing to do */
	if (!length)
		return 0;

	length = blocksize - length;
	iblock = index << (PAGE_CACHE_SHIFT - inode->i_sb->s_blocksize_bits);

	page = grab_cache_page(mapping, index);
	err = -ENOMEM;
	if (!page)
		goto out;

1537
	if (!page_has_buffers(page))
1538
		create_empty_buffers(page, blocksize, 0);
Linus Torvalds's avatar
Linus Torvalds committed
1539 1540

	/* Find the buffer that contains "offset" */
1541
	bh = page_buffers(page);
Linus Torvalds's avatar
Linus Torvalds committed
1542 1543 1544 1545 1546 1547 1548 1549 1550 1551
	pos = blocksize;
	while (offset >= pos) {
		bh = bh->b_this_page;
		iblock++;
		pos += blocksize;
	}

	err = 0;
	if (!buffer_mapped(bh)) {
		ext3_get_block(inode, iblock, bh, 0);
1552
		/* unmapped? It's a hole - nothing to do */
Linus Torvalds's avatar
Linus Torvalds committed
1553 1554 1555 1556 1557
		if (!buffer_mapped(bh))
			goto unlock;
	}

	/* Ok, it's mapped. Make sure it's up-to-date */
Andrew Morton's avatar
Andrew Morton committed
1558
	if (PageUptodate(page))
Andrew Morton's avatar
Andrew Morton committed
1559
		set_buffer_uptodate(bh);
Linus Torvalds's avatar
Linus Torvalds committed
1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575

	if (!buffer_uptodate(bh)) {
		err = -EIO;
		ll_rw_block(READ, 1, &bh);
		wait_on_buffer(bh);
		/* Uhhuh. Read error. Complain and punt. */
		if (!buffer_uptodate(bh))
			goto unlock;
	}

	if (ext3_should_journal_data(inode)) {
		BUFFER_TRACE(bh, "get write access");
		err = ext3_journal_get_write_access(handle, bh);
		if (err)
			goto unlock;
	}
1576 1577 1578

	kaddr = kmap_atomic(page, KM_USER0);
	memset(kaddr + offset, 0, length);
Linus Torvalds's avatar
Linus Torvalds committed
1579
	flush_dcache_page(page);
1580
	kunmap_atomic(kaddr, KM_USER0);
Linus Torvalds's avatar
Linus Torvalds committed
1581 1582 1583 1584 1585 1586 1587 1588 1589

	BUFFER_TRACE(bh, "zeroed end of block");

	err = 0;
	if (ext3_should_journal_data(inode)) {
		err = ext3_journal_dirty_metadata(handle, bh);
	} else {
		if (ext3_should_order_data(inode))
			err = ext3_journal_dirty_data(handle, bh, 0);
1590
		mark_buffer_dirty(bh);
Linus Torvalds's avatar
Linus Torvalds committed
1591 1592 1593
	}

unlock:
Andrew Morton's avatar
Andrew Morton committed
1594
	unlock_page(page);
Linus Torvalds's avatar
Linus Torvalds committed
1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 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 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720
	page_cache_release(page);
out:
	return err;
}

/*
 * Probably it should be a library function... search for first non-zero word
 * or memcmp with zero_page, whatever is better for particular architecture.
 * Linus?
 */
static inline int all_zeroes(u32 *p, u32 *q)
{
	while (p < q)
		if (*p++)
			return 0;
	return 1;
}

/**
 *	ext3_find_shared - find the indirect blocks for partial truncation.
 *	@inode:	  inode in question
 *	@depth:	  depth of the affected branch
 *	@offsets: offsets of pointers in that branch (see ext3_block_to_path)
 *	@chain:	  place to store the pointers to partial indirect blocks
 *	@top:	  place to the (detached) top of branch
 *
 *	This is a helper function used by ext3_truncate().
 *
 *	When we do truncate() we may have to clean the ends of several
 *	indirect blocks but leave the blocks themselves alive. Block is
 *	partially truncated if some data below the new i_size is refered
 *	from it (and it is on the path to the first completely truncated
 *	data block, indeed).  We have to free the top of that path along
 *	with everything to the right of the path. Since no allocation
 *	past the truncation point is possible until ext3_truncate()
 *	finishes, we may safely do the latter, but top of branch may
 *	require special attention - pageout below the truncation point
 *	might try to populate it.
 *
 *	We atomically detach the top of branch from the tree, store the
 *	block number of its root in *@top, pointers to buffer_heads of
 *	partially truncated blocks - in @chain[].bh and pointers to
 *	their last elements that should not be removed - in
 *	@chain[].p. Return value is the pointer to last filled element
 *	of @chain.
 *
 *	The work left to caller to do the actual freeing of subtrees:
 *		a) free the subtree starting from *@top
 *		b) free the subtrees whose roots are stored in
 *			(@chain[i].p+1 .. end of @chain[i].bh->b_data)
 *		c) free the subtrees growing from the inode past the @chain[0].
 *			(no partially truncated stuff there).  */

static Indirect *ext3_find_shared(struct inode *inode,
				int depth,
				int offsets[4],
				Indirect chain[4],
				u32 *top)
{
	Indirect *partial, *p;
	int k, err;

	*top = 0;
	/* Make k index the deepest non-null offest + 1 */
	for (k = depth; k > 1 && !offsets[k-1]; k--)
		;
	partial = ext3_get_branch(inode, k, offsets, chain, &err);
	/* Writer: pointers */
	if (!partial)
		partial = chain + k-1;
	/*
	 * If the branch acquired continuation since we've looked at it -
	 * fine, it should all survive and (new) top doesn't belong to us.
	 */
	if (!partial->key && *partial->p)
		/* Writer: end */
		goto no_top;
	for (p=partial; p>chain && all_zeroes((u32*)p->bh->b_data,p->p); p--)
		;
	/*
	 * OK, we've found the last block that must survive. The rest of our
	 * branch should be detached before unlocking. However, if that rest
	 * of branch is all ours and does not grow immediately from the inode
	 * it's easier to cheat and just decrement partial->p.
	 */
	if (p == chain + k - 1 && p > chain) {
		p->p--;
	} else {
		*top = *p->p;
		/* Nope, don't do this in ext3.  Must leave the tree intact */
#if 0
		*p->p = 0;
#endif
	}
	/* Writer: end */

	while(partial > p)
	{
		brelse(partial->bh);
		partial--;
	}
no_top:
	return partial;
}

/*
 * Zero a number of block pointers in either an inode or an indirect block.
 * If we restart the transaction we must again get write access to the
 * indirect block for further modification.
 *
 * We release `count' blocks on disk, but (last - first) may be greater
 * than `count' because there can be holes in there.
 */
static void
ext3_clear_blocks(handle_t *handle, struct inode *inode, struct buffer_head *bh,
		unsigned long block_to_free, unsigned long count,
		u32 *first, u32 *last)
{
	u32 *p;
	if (try_to_extend_transaction(handle, inode)) {
		if (bh) {
			BUFFER_TRACE(bh, "call ext3_journal_dirty_metadata");
			ext3_journal_dirty_metadata(handle, bh);
		}
		ext3_mark_inode_dirty(handle, inode);
		ext3_journal_test_restart(handle, inode);
Andrew Morton's avatar
Andrew Morton committed
1721 1722 1723 1724
		if (bh) {
			BUFFER_TRACE(bh, "retaking write access");
			ext3_journal_get_write_access(handle, bh);
		}
Linus Torvalds's avatar
Linus Torvalds committed
1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740
	}

	/*
	 * Any buffers which are on the journal will be in memory. We find
	 * them on the hash table so journal_revoke() will run journal_forget()
	 * on them.  We've already detached each block from the file, so
	 * bforget() in journal_forget() should be safe.
	 *
	 * AKPM: turn on bforget in journal_forget()!!!
	 */
	for (p = first; p < last; p++) {
		u32 nr = le32_to_cpu(*p);
		if (nr) {
			struct buffer_head *bh;

			*p = 0;
1741
			bh = sb_find_get_block(inode->i_sb, nr);
Linus Torvalds's avatar
Linus Torvalds committed
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 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853
			ext3_forget(handle, 0, inode, bh, nr);
		}
	}

	ext3_free_blocks(handle, inode, block_to_free, count);
}

/**
 * ext3_free_data - free a list of data blocks
 * @handle:	handle for this transaction
 * @inode:	inode we are dealing with
 * @this_bh:	indirect buffer_head which contains *@first and *@last
 * @first:	array of block numbers
 * @last:	points immediately past the end of array
 *
 * We are freeing all blocks refered from that array (numbers are stored as
 * little-endian 32-bit) and updating @inode->i_blocks appropriately.
 *
 * We accumulate contiguous runs of blocks to free.  Conveniently, if these
 * blocks are contiguous then releasing them at one time will only affect one
 * or two bitmap blocks (+ group descriptor(s) and superblock) and we won't
 * actually use a lot of journal space.
 *
 * @this_bh will be %NULL if @first and @last point into the inode's direct
 * block pointers.
 */
static void ext3_free_data(handle_t *handle, struct inode *inode,
			   struct buffer_head *this_bh, u32 *first, u32 *last)
{
	unsigned long block_to_free = 0;    /* Starting block # of a run */
	unsigned long count = 0;	    /* Number of blocks in the run */ 
	u32 *block_to_free_p = NULL;	    /* Pointer into inode/ind
					       corresponding to
					       block_to_free */
	unsigned long nr;		    /* Current block # */
	u32 *p;				    /* Pointer into inode/ind
					       for current block */
	int err;

	if (this_bh) {				/* For indirect block */
		BUFFER_TRACE(this_bh, "get_write_access");
		err = ext3_journal_get_write_access(handle, this_bh);
		/* Important: if we can't update the indirect pointers
		 * to the blocks, we can't free them. */
		if (err)
			return;
	}

	for (p = first; p < last; p++) {
		nr = le32_to_cpu(*p);
		if (nr) {
			/* accumulate blocks to free if they're contiguous */
			if (count == 0) {
				block_to_free = nr;
				block_to_free_p = p;
				count = 1;
			} else if (nr == block_to_free + count) {
				count++;
			} else {
				ext3_clear_blocks(handle, inode, this_bh, 
						  block_to_free,
						  count, block_to_free_p, p);
				block_to_free = nr;
				block_to_free_p = p;
				count = 1;
			}
		}
	}

	if (count > 0)
		ext3_clear_blocks(handle, inode, this_bh, block_to_free,
				  count, block_to_free_p, p);

	if (this_bh) {
		BUFFER_TRACE(this_bh, "call ext3_journal_dirty_metadata");
		ext3_journal_dirty_metadata(handle, this_bh);
	}
}

/**
 *	ext3_free_branches - free an array of branches
 *	@handle: JBD handle for this transaction
 *	@inode:	inode we are dealing with
 *	@parent_bh: the buffer_head which contains *@first and *@last
 *	@first:	array of block numbers
 *	@last:	pointer immediately past the end of array
 *	@depth:	depth of the branches to free
 *
 *	We are freeing all blocks refered from these branches (numbers are
 *	stored as little-endian 32-bit) and updating @inode->i_blocks
 *	appropriately.
 */
static void ext3_free_branches(handle_t *handle, struct inode *inode,
			       struct buffer_head *parent_bh,
			       u32 *first, u32 *last, int depth)
{
	unsigned long nr;
	u32 *p;

	if (is_handle_aborted(handle))
		return;
	
	if (depth--) {
		struct buffer_head *bh;
		int addr_per_block = EXT3_ADDR_PER_BLOCK(inode->i_sb);
		p = last;
		while (--p >= first) {
			nr = le32_to_cpu(*p);
			if (!nr)
				continue;		/* A hole */

			/* Go read the buffer for the next level down */
Linus Torvalds's avatar
Linus Torvalds committed
1854
			bh = sb_bread(inode->i_sb, nr);
Linus Torvalds's avatar
Linus Torvalds committed
1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973

			/*
			 * A read failure? Report error and clear slot
			 * (should be rare).
			 */
			if (!bh) {
				ext3_error(inode->i_sb, "ext3_free_branches",
					   "Read failure, inode=%ld, block=%ld",
					   inode->i_ino, nr);
				continue;
			}

			/* This zaps the entire block.  Bottom up. */
			BUFFER_TRACE(bh, "free child branches");
			ext3_free_branches(handle, inode, bh, (u32*)bh->b_data,
					   (u32*)bh->b_data + addr_per_block,
					   depth);

			/*
			 * We've probably journalled the indirect block several
			 * times during the truncate.  But it's no longer
			 * needed and we now drop it from the transaction via
			 * journal_revoke().
			 *
			 * That's easy if it's exclusively part of this
			 * transaction.  But if it's part of the committing
			 * transaction then journal_forget() will simply
			 * brelse() it.  That means that if the underlying
			 * block is reallocated in ext3_get_block(),
			 * unmap_underlying_metadata() will find this block
			 * and will try to get rid of it.  damn, damn.
			 *
			 * If this block has already been committed to the
			 * journal, a revoke record will be written.  And
			 * revoke records must be emitted *before* clearing
			 * this block's bit in the bitmaps.
			 */
			ext3_forget(handle, 1, inode, bh, bh->b_blocknr);

			/*
			 * Everything below this this pointer has been
			 * released.  Now let this top-of-subtree go.
			 *
			 * We want the freeing of this indirect block to be
			 * atomic in the journal with the updating of the
			 * bitmap block which owns it.  So make some room in
			 * the journal.
			 *
			 * We zero the parent pointer *after* freeing its
			 * pointee in the bitmaps, so if extend_transaction()
			 * for some reason fails to put the bitmap changes and
			 * the release into the same transaction, recovery
			 * will merely complain about releasing a free block,
			 * rather than leaking blocks.
			 */
			if (is_handle_aborted(handle))
				return;
			if (try_to_extend_transaction(handle, inode)) {
				ext3_mark_inode_dirty(handle, inode);
				ext3_journal_test_restart(handle, inode);
			}

			ext3_free_blocks(handle, inode, nr, 1);

			if (parent_bh) {
				/*
				 * The block which we have just freed is
				 * pointed to by an indirect block: journal it
				 */
				BUFFER_TRACE(parent_bh, "get_write_access");
				if (!ext3_journal_get_write_access(handle,
								   parent_bh)){
					*p = 0;
					BUFFER_TRACE(parent_bh,
					"call ext3_journal_dirty_metadata");
					ext3_journal_dirty_metadata(handle, 
								    parent_bh);
				}
			}
		}
	} else {
		/* We have reached the bottom of the tree. */
		BUFFER_TRACE(parent_bh, "free data blocks");
		ext3_free_data(handle, inode, parent_bh, first, last);
	}
}

/*
 * ext3_truncate()
 *
 * We block out ext3_get_block() block instantiations across the entire
 * transaction, and VFS/VM ensures that ext3_truncate() cannot run
 * simultaneously on behalf of the same inode.
 *
 * As we work through the truncate and commmit bits of it to the journal there
 * is one core, guiding principle: the file's tree must always be consistent on
 * disk.  We must be able to restart the truncate after a crash.
 *
 * The file's tree may be transiently inconsistent in memory (although it
 * probably isn't), but whenever we close off and commit a journal transaction,
 * the contents of (the filesystem + the journal) must be consistent and
 * restartable.  It's pretty simple, really: bottom up, right to left (although
 * left-to-right works OK too).
 *
 * Note that at recovery time, journal replay occurs *before* the restart of
 * truncate against the orphan inode list.
 *
 * The committed inode has the new, desired i_size (which is the same as
 * i_disksize in this case).  After a crash, ext3_orphan_cleanup() will see
 * that this inode's truncate did not complete and it will again call
 * ext3_truncate() to have another go.  So there will be instantiated blocks
 * to the right of the truncation point in a crashed ext3 filesystem.  But
 * that's fine - as long as they are linked from the inode, the post-crash
 * ext3_truncate() run will find them and release them.
 */

void ext3_truncate(struct inode * inode)
{
	handle_t *handle;
Linus Torvalds's avatar
Linus Torvalds committed
1974 1975
	struct ext3_inode_info *ei = EXT3_I(inode);
	u32 *i_data = ei->i_data;
Linus Torvalds's avatar
Linus Torvalds committed
1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990
	int addr_per_block = EXT3_ADDR_PER_BLOCK(inode->i_sb);
	int offsets[4];
	Indirect chain[4];
	Indirect *partial;
	int nr = 0;
	int n;
	long last_block;
	unsigned blocksize;

	if (!(S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode) ||
	    S_ISLNK(inode->i_mode)))
		return;
	if (IS_APPEND(inode) || IS_IMMUTABLE(inode))
		return;

1991
	lock_kernel();
Linus Torvalds's avatar
Linus Torvalds committed
1992 1993 1994
	ext3_discard_prealloc(inode);

	handle = start_transaction(inode);
1995 1996
	if (IS_ERR(handle)) {
		unlock_kernel();
Linus Torvalds's avatar
Linus Torvalds committed
1997
		return;		/* AKPM: return what? */
1998
	}
Linus Torvalds's avatar
Linus Torvalds committed
1999 2000 2001 2002 2003 2004 2005 2006

	blocksize = inode->i_sb->s_blocksize;
	last_block = (inode->i_size + blocksize-1)
					>> EXT3_BLOCK_SIZE_BITS(inode->i_sb);

	ext3_block_truncate_page(handle, inode->i_mapping, inode->i_size);
		

2007
	n = ext3_block_to_path(inode, last_block, offsets, NULL);
Linus Torvalds's avatar
Linus Torvalds committed
2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029
	if (n == 0)
		goto out_stop;	/* error */

	/*
	 * OK.  This truncate is going to happen.  We add the inode to the
	 * orphan list, so that if this truncate spans multiple transactions,
	 * and we crash, we will resume the truncate when the filesystem
	 * recovers.  It also marks the inode dirty, to catch the new size.
	 *
	 * Implication: the file must always be in a sane, consistent
	 * truncatable state while each transaction commits.
	 */
	if (ext3_orphan_add(handle, inode))
		goto out_stop;

	/*
	 * The orphan list entry will now protect us from any crash which
	 * occurs before the truncate completes, so it is now safe to propagate
	 * the new, shorter inode size (held for now in i_size) into the
	 * on-disk inode. We do this via i_disksize, which is the value which
	 * ext3 *really* writes onto the disk inode.
	 */
Linus Torvalds's avatar
Linus Torvalds committed
2030
	ei->i_disksize = inode->i_size;
Linus Torvalds's avatar
Linus Torvalds committed
2031 2032 2033 2034 2035

	/*
	 * From here we block out all ext3_get_block() callers who want to
	 * modify the block allocation tree.
	 */
Linus Torvalds's avatar
Linus Torvalds committed
2036
	down_write(&ei->truncate_sem);
Linus Torvalds's avatar
Linus Torvalds committed
2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 2097 2098 2099

	if (n == 1) {		/* direct blocks */
		ext3_free_data(handle, inode, NULL, i_data+offsets[0],
			       i_data + EXT3_NDIR_BLOCKS);
		goto do_indirects;
	}

	partial = ext3_find_shared(inode, n, offsets, chain, &nr);
	/* Kill the top of shared branch (not detached) */
	if (nr) {
		if (partial == chain) {
			/* Shared branch grows from the inode */
			ext3_free_branches(handle, inode, NULL,
					   &nr, &nr+1, (chain+n-1) - partial);
			*partial->p = 0;
			/*
			 * We mark the inode dirty prior to restart,
			 * and prior to stop.  No need for it here.
			 */
		} else {
			/* Shared branch grows from an indirect block */
			BUFFER_TRACE(partial->bh, "get_write_access");
			ext3_free_branches(handle, inode, partial->bh,
					partial->p,
					partial->p+1, (chain+n-1) - partial);
		}
	}
	/* Clear the ends of indirect blocks on the shared branch */
	while (partial > chain) {
		ext3_free_branches(handle, inode, partial->bh, partial->p + 1,
				   (u32*)partial->bh->b_data + addr_per_block,
				   (chain+n-1) - partial);
		BUFFER_TRACE(partial->bh, "call brelse");
		brelse (partial->bh);
		partial--;
	}
do_indirects:
	/* Kill the remaining (whole) subtrees */
	switch (offsets[0]) {
		default:
			nr = i_data[EXT3_IND_BLOCK];
			if (nr) {
				ext3_free_branches(handle, inode, NULL,
						   &nr, &nr+1, 1);
				i_data[EXT3_IND_BLOCK] = 0;
			}
		case EXT3_IND_BLOCK:
			nr = i_data[EXT3_DIND_BLOCK];
			if (nr) {
				ext3_free_branches(handle, inode, NULL,
						   &nr, &nr+1, 2);
				i_data[EXT3_DIND_BLOCK] = 0;
			}
		case EXT3_DIND_BLOCK:
			nr = i_data[EXT3_TIND_BLOCK];
			if (nr) {
				ext3_free_branches(handle, inode, NULL,
						   &nr, &nr+1, 3);
				i_data[EXT3_TIND_BLOCK] = 0;
			}
		case EXT3_TIND_BLOCK:
			;
	}
Linus Torvalds's avatar
Linus Torvalds committed
2100
	up_write(&ei->truncate_sem);
Linus Torvalds's avatar
Linus Torvalds committed
2101 2102 2103 2104 2105 2106 2107 2108 2109 2110 2111 2112 2113 2114 2115 2116 2117 2118 2119
	inode->i_mtime = inode->i_ctime = CURRENT_TIME;
	ext3_mark_inode_dirty(handle, inode);

	/* In a multi-transaction truncate, we only make the final
	 * transaction synchronous */
	if (IS_SYNC(inode))
		handle->h_sync = 1;
out_stop:
	/*
	 * If this was a simple ftruncate(), and the file will remain alive
	 * then we need to clear up the orphan record which we created above.
	 * However, if this was a real unlink then we were called by
	 * ext3_delete_inode(), and we allow that function to clean up the
	 * orphan info for us.
	 */
	if (inode->i_nlink)
		ext3_orphan_del(handle, inode);

	ext3_journal_stop(handle, inode);
2120
	unlock_kernel();
Linus Torvalds's avatar
Linus Torvalds committed
2121 2122 2123 2124 2125 2126 2127 2128 2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140 2141 2142 2143
}

/* 
 * ext3_get_inode_loc returns with an extra refcount against the
 * inode's underlying buffer_head on success. 
 */

int ext3_get_inode_loc (struct inode *inode, struct ext3_iloc *iloc)
{
	struct buffer_head *bh = 0;
	unsigned long block;
	unsigned long block_group;
	unsigned long group_desc;
	unsigned long desc;
	unsigned long offset;
	struct ext3_group_desc * gdp;
		
	if ((inode->i_ino != EXT3_ROOT_INO &&
		inode->i_ino != EXT3_ACL_IDX_INO &&
		inode->i_ino != EXT3_ACL_DATA_INO &&
		inode->i_ino != EXT3_JOURNAL_INO &&
		inode->i_ino < EXT3_FIRST_INO(inode->i_sb)) ||
		inode->i_ino > le32_to_cpu(
2144
			EXT3_SB(inode->i_sb)->s_es->s_inodes_count)) {
Linus Torvalds's avatar
Linus Torvalds committed
2145 2146 2147 2148 2149
		ext3_error (inode->i_sb, "ext3_get_inode_loc",
			    "bad inode number: %lu", inode->i_ino);
		goto bad_inode;
	}
	block_group = (inode->i_ino - 1) / EXT3_INODES_PER_GROUP(inode->i_sb);
2150
	if (block_group >= EXT3_SB(inode->i_sb)->s_groups_count) {
Linus Torvalds's avatar
Linus Torvalds committed
2151 2152 2153 2154 2155 2156
		ext3_error (inode->i_sb, "ext3_get_inode_loc",
			    "group >= groups count");
		goto bad_inode;
	}
	group_desc = block_group >> EXT3_DESC_PER_BLOCK_BITS(inode->i_sb);
	desc = block_group & (EXT3_DESC_PER_BLOCK(inode->i_sb) - 1);
2157
	bh = EXT3_SB(inode->i_sb)->s_group_desc[group_desc];
Linus Torvalds's avatar
Linus Torvalds committed
2158 2159 2160 2161 2162 2163 2164 2165 2166 2167 2168 2169 2170 2171
	if (!bh) {
		ext3_error (inode->i_sb, "ext3_get_inode_loc",
			    "Descriptor not loaded");
		goto bad_inode;
	}

	gdp = (struct ext3_group_desc *) bh->b_data;
	/*
	 * Figure out the offset within the block group inode table
	 */
	offset = ((inode->i_ino - 1) % EXT3_INODES_PER_GROUP(inode->i_sb)) *
		EXT3_INODE_SIZE(inode->i_sb);
	block = le32_to_cpu(gdp[desc].bg_inode_table) +
		(offset >> EXT3_BLOCK_SIZE_BITS(inode->i_sb));
Linus Torvalds's avatar
Linus Torvalds committed
2172
	if (!(bh = sb_bread(inode->i_sb, block))) {
Linus Torvalds's avatar
Linus Torvalds committed
2173 2174 2175 2176 2177 2178 2179 2180 2181 2182 2183 2184 2185 2186 2187 2188 2189 2190 2191 2192 2193
		ext3_error (inode->i_sb, "ext3_get_inode_loc",
			    "unable to read inode block - "
			    "inode=%lu, block=%lu", inode->i_ino, block);
		goto bad_inode;
	}
	offset &= (EXT3_BLOCK_SIZE(inode->i_sb) - 1);

	iloc->bh = bh;
	iloc->raw_inode = (struct ext3_inode *) (bh->b_data + offset);
	iloc->block_group = block_group;
	
	return 0;
	
 bad_inode:
	return -EIO;
}

void ext3_read_inode(struct inode * inode)
{
	struct ext3_iloc iloc;
	struct ext3_inode *raw_inode;
Linus Torvalds's avatar
Linus Torvalds committed
2194
	struct ext3_inode_info *ei = EXT3_I(inode);
Linus Torvalds's avatar
Linus Torvalds committed
2195 2196 2197 2198 2199 2200 2201 2202 2203 2204 2205 2206 2207 2208 2209 2210 2211 2212 2213
	struct buffer_head *bh;
	int block;
	
	if(ext3_get_inode_loc(inode, &iloc))
		goto bad_inode;
	bh = iloc.bh;
	raw_inode = iloc.raw_inode;
	inode->i_mode = le16_to_cpu(raw_inode->i_mode);
	inode->i_uid = (uid_t)le16_to_cpu(raw_inode->i_uid_low);
	inode->i_gid = (gid_t)le16_to_cpu(raw_inode->i_gid_low);
	if(!(test_opt (inode->i_sb, NO_UID32))) {
		inode->i_uid |= le16_to_cpu(raw_inode->i_uid_high) << 16;
		inode->i_gid |= le16_to_cpu(raw_inode->i_gid_high) << 16;
	}
	inode->i_nlink = le16_to_cpu(raw_inode->i_links_count);
	inode->i_size = le32_to_cpu(raw_inode->i_size);
	inode->i_atime = le32_to_cpu(raw_inode->i_atime);
	inode->i_ctime = le32_to_cpu(raw_inode->i_ctime);
	inode->i_mtime = le32_to_cpu(raw_inode->i_mtime);
Linus Torvalds's avatar
Linus Torvalds committed
2214 2215 2216 2217 2218 2219

	ei->i_state = 0;
	ei->i_next_alloc_block = 0;
	ei->i_next_alloc_goal = 0;
	ei->i_dir_start_lookup = 0;
	ei->i_dtime = le32_to_cpu(raw_inode->i_dtime);
Linus Torvalds's avatar
Linus Torvalds committed
2220 2221 2222 2223 2224 2225 2226
	/* We now have enough fields to check if the inode was active or not.
	 * This is needed because nfsd might try to access dead inodes
	 * the test is that same one that e2fsck uses
	 * NeilBrown 1999oct15
	 */
	if (inode->i_nlink == 0) {
		if (inode->i_mode == 0 ||
2227
		    !(EXT3_SB(inode->i_sb)->s_mount_state & EXT3_ORPHAN_FS)) {
Linus Torvalds's avatar
Linus Torvalds committed
2228 2229 2230 2231 2232 2233 2234 2235 2236 2237 2238 2239 2240 2241
			/* this inode is deleted */
			brelse (bh);
			goto bad_inode;
		}
		/* The only unlinked inodes we let through here have
		 * valid i_mode and are being read by the orphan
		 * recovery code: that's fine, we're about to complete
		 * the process of deleting those. */
	}
	inode->i_blksize = PAGE_SIZE;	/* This is the optimal IO size
					 * (for stat), not the fs block
					 * size */  
	inode->i_blocks = le32_to_cpu(raw_inode->i_blocks);
	inode->i_version = ++event;
Linus Torvalds's avatar
Linus Torvalds committed
2242
	ei->i_flags = le32_to_cpu(raw_inode->i_flags);
Linus Torvalds's avatar
Linus Torvalds committed
2243
#ifdef EXT3_FRAGMENTS
Linus Torvalds's avatar
Linus Torvalds committed
2244 2245 2246
	ei->i_faddr = le32_to_cpu(raw_inode->i_faddr);
	ei->i_frag_no = raw_inode->i_frag;
	ei->i_frag_size = raw_inode->i_fsize;
Linus Torvalds's avatar
Linus Torvalds committed
2247
#endif
Linus Torvalds's avatar
Linus Torvalds committed
2248
	ei->i_file_acl = le32_to_cpu(raw_inode->i_file_acl);
Linus Torvalds's avatar
Linus Torvalds committed
2249
	if (!S_ISREG(inode->i_mode)) {
Linus Torvalds's avatar
Linus Torvalds committed
2250
		ei->i_dir_acl = le32_to_cpu(raw_inode->i_dir_acl);
Linus Torvalds's avatar
Linus Torvalds committed
2251 2252 2253 2254
	} else {
		inode->i_size |=
			((__u64)le32_to_cpu(raw_inode->i_size_high)) << 32;
	}
Linus Torvalds's avatar
Linus Torvalds committed
2255
	ei->i_disksize = inode->i_size;
Linus Torvalds's avatar
Linus Torvalds committed
2256 2257
	inode->i_generation = le32_to_cpu(raw_inode->i_generation);
#ifdef EXT3_PREALLOCATE
Linus Torvalds's avatar
Linus Torvalds committed
2258
	ei->i_prealloc_count = 0;
Linus Torvalds's avatar
Linus Torvalds committed
2259
#endif
Linus Torvalds's avatar
Linus Torvalds committed
2260
	ei->i_block_group = iloc.block_group;
Linus Torvalds's avatar
Linus Torvalds committed
2261 2262 2263 2264 2265 2266

	/*
	 * NOTE! The in-memory inode i_data array is in little-endian order
	 * even on big-endian machines: we do NOT byteswap the block numbers!
	 */
	for (block = 0; block < EXT3_N_BLOCKS; block++)
Linus Torvalds's avatar
Linus Torvalds committed
2267 2268
		ei->i_data[block] = iloc.raw_inode->i_block[block];
	INIT_LIST_HEAD(&ei->i_orphan);
Linus Torvalds's avatar
Linus Torvalds committed
2269 2270 2271 2272 2273 2274 2275 2276 2277

	brelse (iloc.bh);

	if (inode->i_ino == EXT3_ACL_IDX_INO ||
	    inode->i_ino == EXT3_ACL_DATA_INO)
		/* Nothing to do */ ;
	else if (S_ISREG(inode->i_mode)) {
		inode->i_op = &ext3_file_inode_operations;
		inode->i_fop = &ext3_file_operations;
2278 2279 2280 2281
		if (ext3_should_writeback_data(inode))
			inode->i_mapping->a_ops = &ext3_writeback_aops;
		else
			inode->i_mapping->a_ops = &ext3_aops;
Linus Torvalds's avatar
Linus Torvalds committed
2282 2283 2284 2285 2286 2287 2288 2289
	} else if (S_ISDIR(inode->i_mode)) {
		inode->i_op = &ext3_dir_inode_operations;
		inode->i_fop = &ext3_dir_operations;
	} else if (S_ISLNK(inode->i_mode)) {
		if (!inode->i_blocks)
			inode->i_op = &ext3_fast_symlink_inode_operations;
		else {
			inode->i_op = &page_symlink_inode_operations;
2290 2291 2292 2293
			if (ext3_should_writeback_data(inode))
				inode->i_mapping->a_ops = &ext3_writeback_aops;
			else
				inode->i_mapping->a_ops = &ext3_aops;
Linus Torvalds's avatar
Linus Torvalds committed
2294 2295 2296 2297
		}
	} else 
		init_special_inode(inode, inode->i_mode,
				   le32_to_cpu(iloc.raw_inode->i_block[0]));
Andrew Morton's avatar
Andrew Morton committed
2298
	if (ei->i_flags & EXT3_SYNC_FL)
Linus Torvalds's avatar
Linus Torvalds committed
2299
		inode->i_flags |= S_SYNC;
Andrew Morton's avatar
Andrew Morton committed
2300
	if (ei->i_flags & EXT3_APPEND_FL)
Linus Torvalds's avatar
Linus Torvalds committed
2301
		inode->i_flags |= S_APPEND;
Andrew Morton's avatar
Andrew Morton committed
2302
	if (ei->i_flags & EXT3_IMMUTABLE_FL)
Linus Torvalds's avatar
Linus Torvalds committed
2303
		inode->i_flags |= S_IMMUTABLE;
Andrew Morton's avatar
Andrew Morton committed
2304
	if (ei->i_flags & EXT3_NOATIME_FL)
Linus Torvalds's avatar
Linus Torvalds committed
2305 2306 2307 2308 2309 2310 2311 2312 2313 2314 2315 2316 2317 2318 2319 2320 2321 2322 2323
		inode->i_flags |= S_NOATIME;
	return;
	
bad_inode:
	make_bad_inode(inode);
	return;
}

/*
 * Post the struct inode info into an on-disk inode location in the
 * buffer-cache.  This gobbles the caller's reference to the
 * buffer_head in the inode location struct.  
 */

static int ext3_do_update_inode(handle_t *handle, 
				struct inode *inode, 
				struct ext3_iloc *iloc)
{
	struct ext3_inode *raw_inode = iloc->raw_inode;
Linus Torvalds's avatar
Linus Torvalds committed
2324
	struct ext3_inode_info *ei = EXT3_I(inode);
Linus Torvalds's avatar
Linus Torvalds committed
2325 2326 2327 2328 2329 2330 2331 2332 2333 2334 2335 2336 2337 2338 2339 2340 2341
	struct buffer_head *bh = iloc->bh;
	int err = 0, rc, block;

	if (handle) {
		BUFFER_TRACE(bh, "get_write_access");
		err = ext3_journal_get_write_access(handle, bh);
		if (err)
			goto out_brelse;
	}
	raw_inode->i_mode = cpu_to_le16(inode->i_mode);
	if(!(test_opt(inode->i_sb, NO_UID32))) {
		raw_inode->i_uid_low = cpu_to_le16(low_16_bits(inode->i_uid));
		raw_inode->i_gid_low = cpu_to_le16(low_16_bits(inode->i_gid));
/*
 * Fix up interoperability with old kernels. Otherwise, old inodes get
 * re-used with the upper 16 bits of the uid/gid intact
 */
Linus Torvalds's avatar
Linus Torvalds committed
2342
		if(!ei->i_dtime) {
Linus Torvalds's avatar
Linus Torvalds committed
2343 2344 2345 2346 2347 2348 2349 2350 2351 2352 2353 2354 2355 2356 2357 2358 2359
			raw_inode->i_uid_high =
				cpu_to_le16(high_16_bits(inode->i_uid));
			raw_inode->i_gid_high =
				cpu_to_le16(high_16_bits(inode->i_gid));
		} else {
			raw_inode->i_uid_high = 0;
			raw_inode->i_gid_high = 0;
		}
	} else {
		raw_inode->i_uid_low =
			cpu_to_le16(fs_high2lowuid(inode->i_uid));
		raw_inode->i_gid_low =
			cpu_to_le16(fs_high2lowgid(inode->i_gid));
		raw_inode->i_uid_high = 0;
		raw_inode->i_gid_high = 0;
	}
	raw_inode->i_links_count = cpu_to_le16(inode->i_nlink);
Linus Torvalds's avatar
Linus Torvalds committed
2360
	raw_inode->i_size = cpu_to_le32(ei->i_disksize);
Linus Torvalds's avatar
Linus Torvalds committed
2361 2362 2363 2364
	raw_inode->i_atime = cpu_to_le32(inode->i_atime);
	raw_inode->i_ctime = cpu_to_le32(inode->i_ctime);
	raw_inode->i_mtime = cpu_to_le32(inode->i_mtime);
	raw_inode->i_blocks = cpu_to_le32(inode->i_blocks);
Linus Torvalds's avatar
Linus Torvalds committed
2365 2366
	raw_inode->i_dtime = cpu_to_le32(ei->i_dtime);
	raw_inode->i_flags = cpu_to_le32(ei->i_flags);
Linus Torvalds's avatar
Linus Torvalds committed
2367
#ifdef EXT3_FRAGMENTS
Linus Torvalds's avatar
Linus Torvalds committed
2368 2369 2370
	raw_inode->i_faddr = cpu_to_le32(ei->i_faddr);
	raw_inode->i_frag = ei->i_frag_no;
	raw_inode->i_fsize = ei->i_frag_size;
Linus Torvalds's avatar
Linus Torvalds committed
2371 2372 2373 2374
#else
	/* If we are not tracking these fields in the in-memory inode,
	 * then preserve them on disk, but still initialise them to zero
	 * for new inodes. */
Linus Torvalds's avatar
Linus Torvalds committed
2375
	if (ei->i_state & EXT3_STATE_NEW) {
Linus Torvalds's avatar
Linus Torvalds committed
2376 2377 2378 2379 2380
		raw_inode->i_faddr = 0;
		raw_inode->i_frag = 0;
		raw_inode->i_fsize = 0;
	}
#endif
Linus Torvalds's avatar
Linus Torvalds committed
2381
	raw_inode->i_file_acl = cpu_to_le32(ei->i_file_acl);
Linus Torvalds's avatar
Linus Torvalds committed
2382
	if (!S_ISREG(inode->i_mode)) {
Linus Torvalds's avatar
Linus Torvalds committed
2383
		raw_inode->i_dir_acl = cpu_to_le32(ei->i_dir_acl);
Linus Torvalds's avatar
Linus Torvalds committed
2384 2385
	} else {
		raw_inode->i_size_high =
Linus Torvalds's avatar
Linus Torvalds committed
2386 2387
			cpu_to_le32(ei->i_disksize >> 32);
		if (ei->i_disksize > 0x7fffffffULL) {
Linus Torvalds's avatar
Linus Torvalds committed
2388 2389 2390 2391 2392 2393 2394 2395 2396
			struct super_block *sb = inode->i_sb;
			if (!EXT3_HAS_RO_COMPAT_FEATURE(sb,
					EXT3_FEATURE_RO_COMPAT_LARGE_FILE) ||
			    EXT3_SB(sb)->s_es->s_rev_level ==
					cpu_to_le32(EXT3_GOOD_OLD_REV)) {
			       /* If this is the first large file
				* created, add a flag to the superblock.
				*/
				err = ext3_journal_get_write_access(handle,
2397
						EXT3_SB(sb)->s_sbh);
Linus Torvalds's avatar
Linus Torvalds committed
2398 2399 2400 2401 2402 2403 2404 2405
				if (err)
					goto out_brelse;
				ext3_update_dynamic_rev(sb);
				EXT3_SET_RO_COMPAT_FEATURE(sb,
					EXT3_FEATURE_RO_COMPAT_LARGE_FILE);
				sb->s_dirt = 1;
				handle->h_sync = 1;
				err = ext3_journal_dirty_metadata(handle,
2406
						EXT3_SB(sb)->s_sbh);
Linus Torvalds's avatar
Linus Torvalds committed
2407 2408 2409 2410 2411 2412 2413 2414
			}
		}
	}
	raw_inode->i_generation = le32_to_cpu(inode->i_generation);
	if (S_ISCHR(inode->i_mode) || S_ISBLK(inode->i_mode))
		raw_inode->i_block[0] =
			cpu_to_le32(kdev_t_to_nr(inode->i_rdev));
	else for (block = 0; block < EXT3_N_BLOCKS; block++)
Linus Torvalds's avatar
Linus Torvalds committed
2415
		raw_inode->i_block[block] = ei->i_data[block];
Linus Torvalds's avatar
Linus Torvalds committed
2416 2417 2418 2419 2420

	BUFFER_TRACE(bh, "call ext3_journal_dirty_metadata");
	rc = ext3_journal_dirty_metadata(handle, bh);
	if (!err)
		err = rc;
Linus Torvalds's avatar
Linus Torvalds committed
2421
	ei->i_state &= ~EXT3_STATE_NEW;
Linus Torvalds's avatar
Linus Torvalds committed
2422 2423 2424 2425 2426 2427 2428 2429 2430 2431 2432 2433 2434 2435 2436 2437 2438 2439 2440 2441 2442 2443 2444 2445 2446 2447 2448 2449 2450 2451 2452 2453 2454 2455 2456 2457 2458 2459 2460 2461 2462 2463 2464 2465 2466 2467 2468 2469 2470 2471 2472 2473 2474 2475 2476 2477 2478 2479 2480 2481 2482 2483 2484 2485 2486 2487 2488 2489 2490 2491 2492 2493 2494 2495 2496 2497 2498 2499 2500 2501 2502 2503 2504

out_brelse:
	brelse (bh);
	ext3_std_error(inode->i_sb, err);
	return err;
}

/*
 * ext3_write_inode()
 *
 * We are called from a few places:
 *
 * - Within generic_file_write() for O_SYNC files.
 *   Here, there will be no transaction running. We wait for any running
 *   trasnaction to commit.
 *
 * - Within sys_sync(), kupdate and such.
 *   We wait on commit, if tol to.
 *
 * - Within prune_icache() (PF_MEMALLOC == true)
 *   Here we simply return.  We can't afford to block kswapd on the
 *   journal commit.
 *
 * In all cases it is actually safe for us to return without doing anything,
 * because the inode has been copied into a raw inode buffer in
 * ext3_mark_inode_dirty().  This is a correctness thing for O_SYNC and for
 * knfsd.
 *
 * Note that we are absolutely dependent upon all inode dirtiers doing the
 * right thing: they *must* call mark_inode_dirty() after dirtying info in
 * which we are interested.
 *
 * It would be a bug for them to not do this.  The code:
 *
 *	mark_inode_dirty(inode)
 *	stuff();
 *	inode->i_size = expr;
 *
 * is in error because a kswapd-driven write_inode() could occur while
 * `stuff()' is running, and the new i_size will be lost.  Plus the inode
 * will no longer be on the superblock's dirty inode list.
 */
void ext3_write_inode(struct inode *inode, int wait)
{
	if (current->flags & PF_MEMALLOC)
		return;

	if (ext3_journal_current_handle()) {
		jbd_debug(0, "called recursively, non-PF_MEMALLOC!\n");
		return;
	}

	if (!wait)
		return;

	ext3_force_commit(inode->i_sb);	
}

/*
 * ext3_setattr()
 *
 * Called from notify_change.
 *
 * We want to trap VFS attempts to truncate the file as soon as
 * possible.  In particular, we want to make sure that when the VFS
 * shrinks i_size, we put the inode on the orphan list and modify
 * i_disksize immediately, so that during the subsequent flushing of
 * dirty pages and freeing of disk blocks, we can guarantee that any
 * commit will leave the blocks being flushed in an unused state on
 * disk.  (On recovery, the inode will get truncated and the blocks will
 * be freed, so we have a strong guarantee that no future commit will
 * leave these blocks visible to the user.)  
 *
 * This is only needed for regular files.  rmdir() has its own path, and
 * we can never truncate a direcory except on final unlink (at which
 * point i_nlink is zero so recovery is easy.)
 *
 * Called with the BKL.  
 */

int ext3_setattr(struct dentry *dentry, struct iattr *attr)
{
	struct inode *inode = dentry->d_inode;
Linus Torvalds's avatar
Linus Torvalds committed
2505 2506
	int error, rc = 0;
	const unsigned int ia_valid = attr->ia_valid;
Linus Torvalds's avatar
Linus Torvalds committed
2507 2508 2509 2510

	error = inode_change_ok(inode, attr);
	if (error)
		return error;
Linus Torvalds's avatar
Linus Torvalds committed
2511 2512 2513 2514 2515 2516 2517 2518

	if ((ia_valid & ATTR_UID && attr->ia_uid != inode->i_uid) ||
		(ia_valid & ATTR_GID && attr->ia_gid != inode->i_gid)) {
		error = DQUOT_TRANSFER(inode, attr) ? -EDQUOT : 0;
		if (error)
			return error;
	}

2519 2520
	lock_kernel();

Linus Torvalds's avatar
Linus Torvalds committed
2521 2522 2523 2524 2525 2526 2527 2528 2529 2530
	if (attr->ia_valid & ATTR_SIZE && attr->ia_size < inode->i_size) {
		handle_t *handle;

		handle = ext3_journal_start(inode, 3);
		if (IS_ERR(handle)) {
			error = PTR_ERR(handle);
			goto err_out;
		}
		
		error = ext3_orphan_add(handle, inode);
Linus Torvalds's avatar
Linus Torvalds committed
2531
		EXT3_I(inode)->i_disksize = attr->ia_size;
Linus Torvalds's avatar
Linus Torvalds committed
2532 2533 2534 2535 2536 2537
		rc = ext3_mark_inode_dirty(handle, inode);
		if (!error)
			error = rc;
		ext3_journal_stop(handle, inode);
	}
	
Linus Torvalds's avatar
Linus Torvalds committed
2538
	rc = inode_setattr(inode, attr);
Linus Torvalds's avatar
Linus Torvalds committed
2539 2540 2541 2542 2543 2544 2545 2546 2547

	/* If inode_setattr's call to ext3_truncate failed to get a
	 * transaction handle at all, we need to clean up the in-core
	 * orphan list manually. */
	if (inode->i_nlink)
		ext3_orphan_del(NULL, inode);

err_out:
	ext3_std_error(inode->i_sb, error);
2548
	unlock_kernel();
Linus Torvalds's avatar
Linus Torvalds committed
2549 2550 2551
	if (!error)
		error = rc;
	return error;
Linus Torvalds's avatar
Linus Torvalds committed
2552 2553 2554 2555 2556 2557 2558 2559 2560 2561 2562 2563 2564 2565 2566 2567 2568 2569 2570 2571 2572 2573 2574 2575 2576 2577 2578 2579 2580 2581 2582 2583 2584 2585 2586 2587 2588 2589 2590 2591 2592 2593 2594 2595 2596 2597 2598 2599 2600 2601 2602 2603 2604 2605 2606 2607 2608 2609 2610 2611 2612 2613
}


/*
 * akpm: how many blocks doth make a writepage()?
 *
 * With N blocks per page, it may be:
 * N data blocks
 * 2 indirect block
 * 2 dindirect
 * 1 tindirect
 * N+5 bitmap blocks (from the above)
 * N+5 group descriptor summary blocks
 * 1 inode block
 * 1 superblock.
 * 2 * EXT3_SINGLEDATA_TRANS_BLOCKS for the quote files
 *
 * 3 * (N + 5) + 2 + 2 * EXT3_SINGLEDATA_TRANS_BLOCKS
 *
 * With ordered or writeback data it's the same, less the N data blocks.
 *
 * If the inode's direct blocks can hold an integral number of pages then a
 * page cannot straddle two indirect blocks, and we can only touch one indirect
 * and dindirect block, and the "5" above becomes "3".
 *
 * This still overestimates under most circumstances.  If we were to pass the
 * start and end offsets in here as well we could do block_to_path() on each
 * block and work out the exact number of indirects which are touched.  Pah.
 */

int ext3_writepage_trans_blocks(struct inode *inode)
{
	int bpp = ext3_journal_blocks_per_page(inode);
	int indirects = (EXT3_NDIR_BLOCKS % bpp) ? 5 : 3;
	int ret;
	
	if (ext3_should_journal_data(inode))
		ret = 3 * (bpp + indirects) + 2;
	else
		ret = 2 * (bpp + indirects) + 2;

#ifdef CONFIG_QUOTA
	ret += 2 * EXT3_SINGLEDATA_TRANS_BLOCKS;
#endif

	return ret;
}

int
ext3_mark_iloc_dirty(handle_t *handle, 
		     struct inode *inode,
		     struct ext3_iloc *iloc)
{
	int err = 0;

	if (handle) {
		/* the do_update_inode consumes one bh->b_count */
		atomic_inc(&iloc->bh->b_count);
		err = ext3_do_update_inode(handle, inode, iloc);
		/* ext3_do_update_inode() does journal_dirty_metadata */
		brelse(iloc->bh);
	} else {
2614
		printk(KERN_EMERG "%s: called with no handle!\n", __FUNCTION__);
Linus Torvalds's avatar
Linus Torvalds committed
2615 2616 2617 2618 2619 2620 2621 2622 2623 2624 2625 2626 2627 2628 2629 2630 2631 2632 2633 2634 2635 2636 2637 2638 2639 2640 2641 2642 2643 2644 2645 2646 2647 2648 2649 2650 2651 2652 2653 2654 2655 2656 2657 2658 2659 2660 2661 2662 2663 2664 2665 2666 2667 2668 2669 2670 2671 2672 2673 2674 2675 2676 2677 2678 2679 2680 2681 2682 2683 2684 2685 2686 2687 2688 2689 2690 2691 2692 2693 2694 2695 2696 2697 2698 2699 2700 2701
	}
	return err;
}

/* 
 * On success, We end up with an outstanding reference count against
 * iloc->bh.  This _must_ be cleaned up later. 
 */

int
ext3_reserve_inode_write(handle_t *handle, struct inode *inode, 
			 struct ext3_iloc *iloc)
{
	int err = 0;
	if (handle) {
		err = ext3_get_inode_loc(inode, iloc);
		if (!err) {
			BUFFER_TRACE(iloc->bh, "get_write_access");
			err = ext3_journal_get_write_access(handle, iloc->bh);
			if (err) {
				brelse(iloc->bh);
				iloc->bh = NULL;
			}
		}
	}
	ext3_std_error(inode->i_sb, err);
	return err;
}

/*
 * akpm: What we do here is to mark the in-core inode as clean
 * with respect to inode dirtiness (it may still be data-dirty).
 * This means that the in-core inode may be reaped by prune_icache
 * without having to perform any I/O.  This is a very good thing,
 * because *any* task may call prune_icache - even ones which
 * have a transaction open against a different journal.
 *
 * Is this cheating?  Not really.  Sure, we haven't written the
 * inode out, but prune_icache isn't a user-visible syncing function.
 * Whenever the user wants stuff synced (sys_sync, sys_msync, sys_fsync)
 * we start and wait on commits.
 *
 * Is this efficient/effective?  Well, we're being nice to the system
 * by cleaning up our inodes proactively so they can be reaped
 * without I/O.  But we are potentially leaving up to five seconds'
 * worth of inodes floating about which prune_icache wants us to
 * write out.  One way to fix that would be to get prune_icache()
 * to do a write_super() to free up some memory.  It has the desired
 * effect.
 */
int ext3_mark_inode_dirty(handle_t *handle, struct inode *inode)
{
	struct ext3_iloc iloc;
	int err;

	err = ext3_reserve_inode_write(handle, inode, &iloc);
	if (!err)
		err = ext3_mark_iloc_dirty(handle, inode, &iloc);
	return err;
}

/*
 * akpm: ext3_dirty_inode() is called from __mark_inode_dirty()
 *
 * We're really interested in the case where a file is being extended.
 * i_size has been changed by generic_commit_write() and we thus need
 * to include the updated inode in the current transaction.
 *
 * Also, DQUOT_ALLOC_SPACE() will always dirty the inode when blocks
 * are allocated to the file.
 *
 * If the inode is marked synchronous, we don't honour that here - doing
 * so would cause a commit on atime updates, which we don't bother doing.
 * We handle synchronous inodes at the highest possible level.
 */
void ext3_dirty_inode(struct inode *inode)
{
	handle_t *current_handle = ext3_journal_current_handle();
	handle_t *handle;

	lock_kernel();
	handle = ext3_journal_start(inode, 1);
	if (IS_ERR(handle))
		goto out;
	if (current_handle &&
		current_handle->h_transaction != handle->h_transaction) {
		/* This task has a transaction open against a different fs */
2702 2703
		printk(KERN_EMERG "%s: transactions do not match!\n",
		       __FUNCTION__);
Linus Torvalds's avatar
Linus Torvalds committed
2704 2705 2706 2707 2708 2709 2710 2711 2712 2713 2714 2715 2716 2717 2718 2719 2720 2721 2722 2723 2724 2725 2726 2727 2728 2729 2730 2731 2732 2733 2734 2735 2736 2737 2738 2739 2740 2741 2742 2743 2744 2745 2746 2747 2748 2749 2750 2751 2752 2753 2754 2755 2756 2757 2758 2759 2760 2761 2762 2763 2764 2765 2766 2767 2768 2769 2770 2771 2772 2773 2774 2775
	} else {
		jbd_debug(5, "marking dirty.  outer handle=%p\n",
				current_handle);
		ext3_mark_inode_dirty(handle, inode);
	}
	ext3_journal_stop(handle, inode);
out:
	unlock_kernel();
}

#ifdef AKPM
/* 
 * Bind an inode's backing buffer_head into this transaction, to prevent
 * it from being flushed to disk early.  Unlike
 * ext3_reserve_inode_write, this leaves behind no bh reference and
 * returns no iloc structure, so the caller needs to repeat the iloc
 * lookup to mark the inode dirty later.
 */
static inline int
ext3_pin_inode(handle_t *handle, struct inode *inode)
{
	struct ext3_iloc iloc;
	
	int err = 0;
	if (handle) {
		err = ext3_get_inode_loc(inode, &iloc);
		if (!err) {
			BUFFER_TRACE(iloc.bh, "get_write_access");
			err = journal_get_write_access(handle, iloc.bh);
			if (!err)
				err = ext3_journal_dirty_metadata(handle, 
								  iloc.bh);
			brelse(iloc.bh);
		}
	}
	ext3_std_error(inode->i_sb, err);
	return err;
}
#endif

int ext3_change_inode_journal_flag(struct inode *inode, int val)
{
	journal_t *journal;
	handle_t *handle;
	int err;

	/*
	 * We have to be very careful here: changing a data block's
	 * journaling status dynamically is dangerous.  If we write a
	 * data block to the journal, change the status and then delete
	 * that block, we risk forgetting to revoke the old log record
	 * from the journal and so a subsequent replay can corrupt data.
	 * So, first we make sure that the journal is empty and that
	 * nobody is changing anything.
	 */

	journal = EXT3_JOURNAL(inode);
	if (is_journal_aborted(journal) || IS_RDONLY(inode))
		return -EROFS;
	
	journal_lock_updates(journal);
	journal_flush(journal);

	/*
	 * OK, there are no updates running now, and all cached data is
	 * synced to disk.  We are now in a completely consistent state
	 * which doesn't have anything in the journal, and we know that
	 * no filesystem updates are running, so it is safe to modify
	 * the inode's in-core data-journaling state flag now.
	 */

	if (val)
Linus Torvalds's avatar
Linus Torvalds committed
2776
		EXT3_I(inode)->i_flags |= EXT3_JOURNAL_DATA_FL;
Linus Torvalds's avatar
Linus Torvalds committed
2777
	else
Linus Torvalds's avatar
Linus Torvalds committed
2778
		EXT3_I(inode)->i_flags &= ~EXT3_JOURNAL_DATA_FL;
Linus Torvalds's avatar
Linus Torvalds committed
2779 2780 2781 2782 2783 2784 2785 2786 2787 2788 2789 2790 2791 2792 2793 2794 2795 2796 2797 2798 2799 2800 2801 2802 2803 2804 2805 2806 2807 2808 2809 2810 2811 2812 2813 2814 2815 2816 2817 2818 2819 2820 2821 2822 2823 2824 2825 2826 2827 2828 2829 2830 2831 2832 2833 2834 2835 2836 2837 2838 2839 2840 2841 2842 2843 2844 2845 2846 2847 2848 2849 2850 2851 2852

	journal_unlock_updates(journal);

	/* Finally we can mark the inode as dirty. */

	handle = ext3_journal_start(inode, 1);
	if (IS_ERR(handle))
		return PTR_ERR(handle);

	err = ext3_mark_inode_dirty(handle, inode);
	handle->h_sync = 1;
	ext3_journal_stop(handle, inode);
	ext3_std_error(inode->i_sb, err);
	
	return err;
}


/*
 * ext3_aops_journal_start().
 *
 * <This function died, but the comment lives on>
 *
 * We need to take the inode semaphore *outside* the
 * journal_start/journal_stop.  Otherwise, a different task could do a
 * wait_for_commit() while holding ->i_sem, which deadlocks.  The rule
 * is: transaction open/closes are considered to be a locking operation
 * and they nest *inside* ->i_sem.
 * ----------------------------------------------------------------------------
 * Possible problem:
 *	ext3_file_write()
 *	-> generic_file_write()
 *	   -> __alloc_pages()
 *	      -> page_launder()
 *		 -> ext3_writepage()
 *
 * And the writepage can be on a different fs while we have a
 * transaction open against this one!  Bad.
 *
 * I tried making the task PF_MEMALLOC here, but that simply results in
 * 0-order allocation failures passed back to generic_file_write().
 * Instead, we rely on the reentrancy protection in ext3_writepage().
 * ----------------------------------------------------------------------------
 * When we do the journal_start() here we don't really need to reserve
 * any blocks - we won't need any until we hit ext3_prepare_write(),
 * which does all the needed journal extending.  However!  There is a
 * problem with quotas:
 *
 * Thread 1:
 * sys_sync
 * ->sync_dquots
 *   ->commit_dquot
 *     ->lock_dquot
 *     ->write_dquot
 *       ->ext3_file_write
 *         ->journal_start
 *         ->ext3_prepare_write
 *           ->journal_extend
 *           ->journal_start
 * Thread 2:
 * ext3_create		(for example)
 * ->ext3_new_inode
 *   ->dquot_initialize
 *     ->lock_dquot
 *
 * Deadlock.  Thread 1's journal_start blocks because thread 2 has a
 * transaction open.  Thread 2's transaction will never close because
 * thread 2 is stuck waiting for the dquot lock.
 *
 * So.  We must ensure that thread 1 *never* needs to extend the journal
 * for quota writes.  We do that by reserving enough journal blocks
 * here, in ext3_aops_journal_start() to ensure that the forthcoming "see if we
 * need to extend" test in ext3_prepare_write() succeeds.  
 */