bmap.c 33.6 KB
Newer Older
D
David Teigland 已提交
1 2
/*
 * Copyright (C) Sistina Software, Inc.  1997-2003 All rights reserved.
3
 * Copyright (C) 2004-2006 Red Hat, Inc.  All rights reserved.
D
David Teigland 已提交
4 5 6
 *
 * This copyrighted material is made available to anyone wishing to use,
 * modify, copy, or redistribute it subject to the terms and conditions
7
 * of the GNU General Public License version 2.
D
David Teigland 已提交
8 9 10 11 12
 */

#include <linux/spinlock.h>
#include <linux/completion.h>
#include <linux/buffer_head.h>
13
#include <linux/blkdev.h>
14
#include <linux/gfs2_ondisk.h>
15
#include <linux/crc32.h>
D
David Teigland 已提交
16 17

#include "gfs2.h"
18
#include "incore.h"
D
David Teigland 已提交
19 20 21 22 23 24
#include "bmap.h"
#include "glock.h"
#include "inode.h"
#include "meta_io.h"
#include "quota.h"
#include "rgrp.h"
25
#include "log.h"
26
#include "super.h"
D
David Teigland 已提交
27
#include "trans.h"
28
#include "dir.h"
29
#include "util.h"
S
Steven Whitehouse 已提交
30
#include "trace_gfs2.h"
D
David Teigland 已提交
31 32 33 34 35 36

/* This doesn't need to be that large as max 64 bit pointers in a 4k
 * block is 512, so __u16 is fine for that. It saves stack space to
 * keep it small.
 */
struct metapath {
37
	struct buffer_head *mp_bh[GFS2_MAX_META_HEIGHT];
D
David Teigland 已提交
38 39 40 41 42 43 44 45
	__u16 mp_list[GFS2_MAX_META_HEIGHT];
};

struct strip_mine {
	int sm_first;
	unsigned int sm_height;
};

46 47 48 49 50
/**
 * gfs2_unstuffer_page - unstuff a stuffed inode into a block cached by a page
 * @ip: the inode
 * @dibh: the dinode buffer
 * @block: the block number that was allocated
S
Steven Whitehouse 已提交
51
 * @page: The (optional) page. This is looked up if @page is NULL
52 53 54 55 56
 *
 * Returns: errno
 */

static int gfs2_unstuffer_page(struct gfs2_inode *ip, struct buffer_head *dibh,
57
			       u64 block, struct page *page)
58 59 60 61 62 63
{
	struct inode *inode = &ip->i_inode;
	struct buffer_head *bh;
	int release = 0;

	if (!page || page->index) {
64
		page = find_or_create_page(inode->i_mapping, 0, GFP_NOFS);
65 66 67 68 69 70 71
		if (!page)
			return -ENOMEM;
		release = 1;
	}

	if (!PageUptodate(page)) {
		void *kaddr = kmap(page);
72 73 74 75
		u64 dsize = i_size_read(inode);
 
		if (dsize > (dibh->b_size - sizeof(struct gfs2_dinode)))
			dsize = dibh->b_size - sizeof(struct gfs2_dinode);
76

77 78
		memcpy(kaddr, dibh->b_data + sizeof(struct gfs2_dinode), dsize);
		memset(kaddr + dsize, 0, PAGE_CACHE_SIZE - dsize);
79 80 81 82 83 84 85 86 87 88 89 90 91 92 93
		kunmap(page);

		SetPageUptodate(page);
	}

	if (!page_has_buffers(page))
		create_empty_buffers(page, 1 << inode->i_blkbits,
				     (1 << BH_Uptodate));

	bh = page_buffers(page);

	if (!buffer_mapped(bh))
		map_bh(bh, inode->i_sb, block);

	set_buffer_uptodate(bh);
94 95
	if (!gfs2_is_jdata(ip))
		mark_buffer_dirty(bh);
96
	if (!gfs2_is_writeback(ip))
97
		gfs2_trans_add_data(ip->i_gl, bh);
98 99 100 101 102 103 104 105 106

	if (release) {
		unlock_page(page);
		page_cache_release(page);
	}

	return 0;
}

D
David Teigland 已提交
107 108 109
/**
 * gfs2_unstuff_dinode - Unstuff a dinode when the data has grown too big
 * @ip: The GFS2 inode to unstuff
S
Steven Whitehouse 已提交
110
 * @page: The (optional) page. This is looked up if the @page is NULL
D
David Teigland 已提交
111 112 113 114 115 116 117
 *
 * This routine unstuffs a dinode and returns it to a "normal" state such
 * that the height can be grown in the traditional way.
 *
 * Returns: errno
 */

118
int gfs2_unstuff_dinode(struct gfs2_inode *ip, struct page *page)
D
David Teigland 已提交
119 120
{
	struct buffer_head *bh, *dibh;
121
	struct gfs2_dinode *di;
122
	u64 block = 0;
123
	int isdir = gfs2_is_dir(ip);
D
David Teigland 已提交
124 125 126 127 128 129 130
	int error;

	down_write(&ip->i_rw_mutex);

	error = gfs2_meta_inode_buffer(ip, &dibh);
	if (error)
		goto out;
131

S
Steven Whitehouse 已提交
132
	if (i_size_read(&ip->i_inode)) {
D
David Teigland 已提交
133 134 135
		/* Get a free block, fill it with the stuffed data,
		   and write it out to disk */

136
		unsigned int n = 1;
137
		error = gfs2_alloc_blocks(ip, &block, &n, 0, NULL);
138 139
		if (error)
			goto out_brelse;
140
		if (isdir) {
141
			gfs2_trans_add_unrevoke(GFS2_SB(&ip->i_inode), block, 1);
142
			error = gfs2_dir_get_new_buffer(ip, block, &bh);
D
David Teigland 已提交
143 144
			if (error)
				goto out_brelse;
145
			gfs2_buffer_copy_tail(bh, sizeof(struct gfs2_meta_header),
D
David Teigland 已提交
146 147 148
					      dibh, sizeof(struct gfs2_dinode));
			brelse(bh);
		} else {
149
			error = gfs2_unstuffer_page(ip, dibh, block, page);
D
David Teigland 已提交
150 151 152 153 154 155 156
			if (error)
				goto out_brelse;
		}
	}

	/*  Set up the pointer to the new block  */

157
	gfs2_trans_add_meta(ip->i_gl, dibh);
158
	di = (struct gfs2_dinode *)dibh->b_data;
D
David Teigland 已提交
159 160
	gfs2_buffer_clear_tail(dibh, sizeof(struct gfs2_dinode));

S
Steven Whitehouse 已提交
161
	if (i_size_read(&ip->i_inode)) {
162
		*(__be64 *)(di + 1) = cpu_to_be64(block);
163 164
		gfs2_add_inode_blocks(&ip->i_inode, 1);
		di->di_blocks = cpu_to_be64(gfs2_get_inode_blocks(&ip->i_inode));
D
David Teigland 已提交
165 166
	}

167
	ip->i_height = 1;
168
	di->di_height = cpu_to_be16(1);
D
David Teigland 已提交
169

170
out_brelse:
D
David Teigland 已提交
171
	brelse(dibh);
172
out:
D
David Teigland 已提交
173 174 175 176 177 178 179
	up_write(&ip->i_rw_mutex);
	return error;
}


/**
 * find_metapath - Find path through the metadata tree
180
 * @sdp: The superblock
D
David Teigland 已提交
181 182
 * @mp: The metapath to return the result in
 * @block: The disk block to look up
183
 * @height: The pre-calculated height of the metadata tree
D
David Teigland 已提交
184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237
 *
 *   This routine returns a struct metapath structure that defines a path
 *   through the metadata of inode "ip" to get to block "block".
 *
 *   Example:
 *   Given:  "ip" is a height 3 file, "offset" is 101342453, and this is a
 *   filesystem with a blocksize of 4096.
 *
 *   find_metapath() would return a struct metapath structure set to:
 *   mp_offset = 101342453, mp_height = 3, mp_list[0] = 0, mp_list[1] = 48,
 *   and mp_list[2] = 165.
 *
 *   That means that in order to get to the block containing the byte at
 *   offset 101342453, we would load the indirect block pointed to by pointer
 *   0 in the dinode.  We would then load the indirect block pointed to by
 *   pointer 48 in that indirect block.  We would then load the data block
 *   pointed to by pointer 165 in that indirect block.
 *
 *             ----------------------------------------
 *             | Dinode |                             |
 *             |        |                            4|
 *             |        |0 1 2 3 4 5                 9|
 *             |        |                            6|
 *             ----------------------------------------
 *                       |
 *                       |
 *                       V
 *             ----------------------------------------
 *             | Indirect Block                       |
 *             |                                     5|
 *             |            4 4 4 4 4 5 5            1|
 *             |0           5 6 7 8 9 0 1            2|
 *             ----------------------------------------
 *                                |
 *                                |
 *                                V
 *             ----------------------------------------
 *             | Indirect Block                       |
 *             |                         1 1 1 1 1   5|
 *             |                         6 6 6 6 6   1|
 *             |0                        3 4 5 6 7   2|
 *             ----------------------------------------
 *                                           |
 *                                           |
 *                                           V
 *             ----------------------------------------
 *             | Data block containing offset         |
 *             |            101342453                 |
 *             |                                      |
 *             |                                      |
 *             ----------------------------------------
 *
 */

238 239
static void find_metapath(const struct gfs2_sbd *sdp, u64 block,
			  struct metapath *mp, unsigned int height)
D
David Teigland 已提交
240 241 242
{
	unsigned int i;

243
	for (i = height; i--;)
B
Bob Peterson 已提交
244
		mp->mp_list[i] = do_div(block, sdp->sd_inptrs);
D
David Teigland 已提交
245 246 247

}

248
static inline unsigned int metapath_branch_start(const struct metapath *mp)
249
{
250 251 252
	if (mp->mp_list[0] == 0)
		return 2;
	return 1;
253 254
}

D
David Teigland 已提交
255 256 257 258 259 260 261 262 263 264
/**
 * metapointer - Return pointer to start of metadata in a buffer
 * @height: The metadata height (0 = dinode)
 * @mp: The metapath
 *
 * Return a pointer to the block number of the next height of the metadata
 * tree given a buffer containing the pointer to the current height of the
 * metadata tree.
 */

265
static inline __be64 *metapointer(unsigned int height, const struct metapath *mp)
D
David Teigland 已提交
266
{
267
	struct buffer_head *bh = mp->mp_bh[height];
D
David Teigland 已提交
268 269
	unsigned int head_size = (height > 0) ?
		sizeof(struct gfs2_meta_header) : sizeof(struct gfs2_dinode);
270
	return ((__be64 *)(bh->b_data + head_size)) + mp->mp_list[height];
D
David Teigland 已提交
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
static void gfs2_metapath_ra(struct gfs2_glock *gl,
			     const struct buffer_head *bh, const __be64 *pos)
{
	struct buffer_head *rabh;
	const __be64 *endp = (const __be64 *)(bh->b_data + bh->b_size);
	const __be64 *t;

	for (t = pos; t < endp; t++) {
		if (!*t)
			continue;

		rabh = gfs2_getbuf(gl, be64_to_cpu(*t), CREATE);
		if (trylock_buffer(rabh)) {
			if (!buffer_uptodate(rabh)) {
				rabh->b_end_io = end_buffer_read_sync;
				submit_bh(READA | REQ_META, rabh);
				continue;
			}
			unlock_buffer(rabh);
		}
		brelse(rabh);
	}
}

D
David Teigland 已提交
297
/**
298 299
 * lookup_metapath - Walk the metadata tree to a specific point
 * @ip: The inode
D
David Teigland 已提交
300 301
 * @mp: The metapath
 *
302 303 304 305 306 307 308 309
 * Assumes that the inode's buffer has already been looked up and
 * hooked onto mp->mp_bh[0] and that the metapath has been initialised
 * by find_metapath().
 *
 * If this function encounters part of the tree which has not been
 * allocated, it returns the current height of the tree at the point
 * at which it found the unallocated block. Blocks which are found are
 * added to the mp->mp_bh[] list.
D
David Teigland 已提交
310
 *
311
 * Returns: error or height of metadata tree
D
David Teigland 已提交
312 313
 */

314
static int lookup_metapath(struct gfs2_inode *ip, struct metapath *mp)
315 316 317
{
	unsigned int end_of_metadata = ip->i_height - 1;
	unsigned int x;
318 319
	__be64 *ptr;
	u64 dblock;
320
	int ret;
321 322

	for (x = 0; x < end_of_metadata; x++) {
323 324 325 326
		ptr = metapointer(x, mp);
		dblock = be64_to_cpu(*ptr);
		if (!dblock)
			return x + 1;
327

328
		ret = gfs2_meta_indirect_buffer(ip, x+1, dblock, &mp->mp_bh[x+1]);
329 330 331 332
		if (ret)
			return ret;
	}

333
	return ip->i_height;
334 335
}

336
static inline void release_metapath(struct metapath *mp)
337 338 339
{
	int i;

340 341 342 343 344
	for (i = 0; i < GFS2_MAX_META_HEIGHT; i++) {
		if (mp->mp_bh[i] == NULL)
			break;
		brelse(mp->mp_bh[i]);
	}
345 346
}

347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382
/**
 * gfs2_extent_length - Returns length of an extent of blocks
 * @start: Start of the buffer
 * @len: Length of the buffer in bytes
 * @ptr: Current position in the buffer
 * @limit: Max extent length to return (0 = unlimited)
 * @eob: Set to 1 if we hit "end of block"
 *
 * If the first block is zero (unallocated) it will return the number of
 * unallocated blocks in the extent, otherwise it will return the number
 * of contiguous blocks in the extent.
 *
 * Returns: The length of the extent (minimum of one block)
 */

static inline unsigned int gfs2_extent_length(void *start, unsigned int len, __be64 *ptr, unsigned limit, int *eob)
{
	const __be64 *end = (start + len);
	const __be64 *first = ptr;
	u64 d = be64_to_cpu(*ptr);

	*eob = 0;
	do {
		ptr++;
		if (ptr >= end)
			break;
		if (limit && --limit == 0)
			break;
		if (d)
			d++;
	} while(be64_to_cpu(*ptr) == d);
	if (ptr >= end)
		*eob = 1;
	return (ptr - first);
}

383
static inline void bmap_lock(struct gfs2_inode *ip, int create)
384 385 386 387 388 389 390
{
	if (create)
		down_write(&ip->i_rw_mutex);
	else
		down_read(&ip->i_rw_mutex);
}

391
static inline void bmap_unlock(struct gfs2_inode *ip, int create)
392 393 394 395 396 397 398
{
	if (create)
		up_write(&ip->i_rw_mutex);
	else
		up_read(&ip->i_rw_mutex);
}

399 400 401 402 403 404 405 406 407 408
static inline __be64 *gfs2_indirect_init(struct metapath *mp,
					 struct gfs2_glock *gl, unsigned int i,
					 unsigned offset, u64 bn)
{
	__be64 *ptr = (__be64 *)(mp->mp_bh[i - 1]->b_data +
		       ((i > 1) ? sizeof(struct gfs2_meta_header) :
				 sizeof(struct gfs2_dinode)));
	BUG_ON(i < 1);
	BUG_ON(mp->mp_bh[i] != NULL);
	mp->mp_bh[i] = gfs2_meta_new(gl, bn);
409
	gfs2_trans_add_meta(gl, mp->mp_bh[i]);
410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455
	gfs2_metatype_set(mp->mp_bh[i], GFS2_METATYPE_IN, GFS2_FORMAT_IN);
	gfs2_buffer_clear_tail(mp->mp_bh[i], sizeof(struct gfs2_meta_header));
	ptr += offset;
	*ptr = cpu_to_be64(bn);
	return ptr;
}

enum alloc_state {
	ALLOC_DATA = 0,
	ALLOC_GROW_DEPTH = 1,
	ALLOC_GROW_HEIGHT = 2,
	/* ALLOC_UNSTUFF = 3,   TBD and rather complicated */
};

/**
 * gfs2_bmap_alloc - Build a metadata tree of the requested height
 * @inode: The GFS2 inode
 * @lblock: The logical starting block of the extent
 * @bh_map: This is used to return the mapping details
 * @mp: The metapath
 * @sheight: The starting height (i.e. whats already mapped)
 * @height: The height to build to
 * @maxlen: The max number of data blocks to alloc
 *
 * In this routine we may have to alloc:
 *   i) Indirect blocks to grow the metadata tree height
 *  ii) Indirect blocks to fill in lower part of the metadata tree
 * iii) Data blocks
 *
 * The function is in two parts. The first part works out the total
 * number of blocks which we need. The second part does the actual
 * allocation asking for an extent at a time (if enough contiguous free
 * blocks are available, there will only be one request per bmap call)
 * and uses the state machine to initialise the blocks in order.
 *
 * Returns: errno on error
 */

static int gfs2_bmap_alloc(struct inode *inode, const sector_t lblock,
			   struct buffer_head *bh_map, struct metapath *mp,
			   const unsigned int sheight,
			   const unsigned int height,
			   const unsigned int maxlen)
{
	struct gfs2_inode *ip = GFS2_I(inode);
	struct gfs2_sbd *sdp = GFS2_SB(inode);
456
	struct super_block *sb = sdp->sd_vfs;
457 458
	struct buffer_head *dibh = mp->mp_bh[0];
	u64 bn, dblock = 0;
459
	unsigned n, i, blks, alloced = 0, iblks = 0, branch_start = 0;
460 461 462
	unsigned dblks = 0;
	unsigned ptrs_per_blk;
	const unsigned end_of_metadata = height - 1;
463
	int ret;
464 465 466 467 468 469 470 471
	int eob = 0;
	enum alloc_state state;
	__be64 *ptr;
	__be64 zero_bn = 0;

	BUG_ON(sheight < 1);
	BUG_ON(dibh == NULL);

472
	gfs2_trans_add_meta(ip->i_gl, dibh);
473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494

	if (height == sheight) {
		struct buffer_head *bh;
		/* Bottom indirect block exists, find unalloced extent size */
		ptr = metapointer(end_of_metadata, mp);
		bh = mp->mp_bh[end_of_metadata];
		dblks = gfs2_extent_length(bh->b_data, bh->b_size, ptr, maxlen,
					   &eob);
		BUG_ON(dblks < 1);
		state = ALLOC_DATA;
	} else {
		/* Need to allocate indirect blocks */
		ptrs_per_blk = height > 1 ? sdp->sd_inptrs : sdp->sd_diptrs;
		dblks = min(maxlen, ptrs_per_blk - mp->mp_list[end_of_metadata]);
		if (height == ip->i_height) {
			/* Writing into existing tree, extend tree down */
			iblks = height - sheight;
			state = ALLOC_GROW_DEPTH;
		} else {
			/* Building up tree height */
			state = ALLOC_GROW_HEIGHT;
			iblks = height - ip->i_height;
495 496
			branch_start = metapath_branch_start(mp);
			iblks += (height - branch_start);
497 498 499 500 501 502 503 504
		}
	}

	/* start of the second part of the function (state machine) */

	blks = dblks + iblks;
	i = sheight;
	do {
505
		int error;
506
		n = blks - alloced;
507
		error = gfs2_alloc_blocks(ip, &bn, &n, 0, NULL);
508 509
		if (error)
			return error;
510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534
		alloced += n;
		if (state != ALLOC_DATA || gfs2_is_jdata(ip))
			gfs2_trans_add_unrevoke(sdp, bn, n);
		switch (state) {
		/* Growing height of tree */
		case ALLOC_GROW_HEIGHT:
			if (i == 1) {
				ptr = (__be64 *)(dibh->b_data +
						 sizeof(struct gfs2_dinode));
				zero_bn = *ptr;
			}
			for (; i - 1 < height - ip->i_height && n > 0; i++, n--)
				gfs2_indirect_init(mp, ip->i_gl, i, 0, bn++);
			if (i - 1 == height - ip->i_height) {
				i--;
				gfs2_buffer_copy_tail(mp->mp_bh[i],
						sizeof(struct gfs2_meta_header),
						dibh, sizeof(struct gfs2_dinode));
				gfs2_buffer_clear_tail(dibh,
						sizeof(struct gfs2_dinode) +
						sizeof(__be64));
				ptr = (__be64 *)(mp->mp_bh[i]->b_data +
					sizeof(struct gfs2_meta_header));
				*ptr = zero_bn;
				state = ALLOC_GROW_DEPTH;
535
				for(i = branch_start; i < height; i++) {
536 537 538 539 540
					if (mp->mp_bh[i] == NULL)
						break;
					brelse(mp->mp_bh[i]);
					mp->mp_bh[i] = NULL;
				}
541
				i = branch_start;
542 543 544 545 546 547
			}
			if (n == 0)
				break;
		/* Branching from existing tree */
		case ALLOC_GROW_DEPTH:
			if (i > 1 && i < height)
548
				gfs2_trans_add_meta(ip->i_gl, mp->mp_bh[i-1]);
549 550 551 552 553 554 555 556 557 558 559
			for (; i < height && n > 0; i++, n--)
				gfs2_indirect_init(mp, ip->i_gl, i,
						   mp->mp_list[i-1], bn++);
			if (i == height)
				state = ALLOC_DATA;
			if (n == 0)
				break;
		/* Tree complete, adding data blocks */
		case ALLOC_DATA:
			BUG_ON(n > dblks);
			BUG_ON(mp->mp_bh[end_of_metadata] == NULL);
560
			gfs2_trans_add_meta(ip->i_gl, mp->mp_bh[end_of_metadata]);
561 562 563 564 565
			dblks = n;
			ptr = metapointer(end_of_metadata, mp);
			dblock = bn;
			while (n-- > 0)
				*ptr++ = cpu_to_be64(bn++);
566 567 568 569 570 571 572 573 574
			if (buffer_zeronew(bh_map)) {
				ret = sb_issue_zeroout(sb, dblock, dblks,
						       GFP_NOFS);
				if (ret) {
					fs_err(sdp,
					       "Failed to zero data buffers\n");
					clear_buffer_zeronew(bh_map);
				}
			}
575 576
			break;
		}
577
	} while ((state != ALLOC_DATA) || !dblock);
578 579 580 581 582 583 584 585 586 587

	ip->i_height = height;
	gfs2_add_inode_blocks(&ip->i_inode, alloced);
	gfs2_dinode_out(ip, mp->mp_bh[0]->b_data);
	map_bh(bh_map, inode->i_sb, dblock);
	bh_map->b_size = dblks << inode->i_blkbits;
	set_buffer_new(bh_map);
	return 0;
}

D
David Teigland 已提交
588
/**
589
 * gfs2_block_map - Map a block from an inode to a disk block
S
Steven Whitehouse 已提交
590
 * @inode: The inode
D
David Teigland 已提交
591
 * @lblock: The logical block number
592
 * @bh_map: The bh to be mapped
593
 * @create: True if its ok to alloc blocks to satify the request
D
David Teigland 已提交
594
 *
595 596 597
 * Sets buffer_mapped() if successful, sets buffer_boundary() if a
 * read of metadata will be required before the next block can be
 * mapped. Sets buffer_new() if new blocks were allocated.
D
David Teigland 已提交
598 599 600 601
 *
 * Returns: errno
 */

602 603
int gfs2_block_map(struct inode *inode, sector_t lblock,
		   struct buffer_head *bh_map, int create)
D
David Teigland 已提交
604
{
605 606
	struct gfs2_inode *ip = GFS2_I(inode);
	struct gfs2_sbd *sdp = GFS2_SB(inode);
607
	unsigned int bsize = sdp->sd_sb.sb_bsize;
608
	const unsigned int maxlen = bh_map->b_size >> inode->i_blkbits;
609
	const u64 *arr = sdp->sd_heightsize;
610 611 612 613 614 615 616 617
	__be64 *ptr;
	u64 size;
	struct metapath mp;
	int ret;
	int eob;
	unsigned int len;
	struct buffer_head *bh;
	u8 height;
S
Steven Whitehouse 已提交
618

619
	BUG_ON(maxlen == 0);
D
David Teigland 已提交
620

621
	memset(mp.mp_bh, 0, sizeof(mp.mp_bh));
622
	bmap_lock(ip, create);
623 624 625
	clear_buffer_mapped(bh_map);
	clear_buffer_new(bh_map);
	clear_buffer_boundary(bh_map);
S
Steven Whitehouse 已提交
626
	trace_gfs2_bmap(ip, bh_map, lblock, create, 1);
627 628 629 630
	if (gfs2_is_dir(ip)) {
		bsize = sdp->sd_jbsize;
		arr = sdp->sd_jheightsize;
	}
631

632 633 634
	ret = gfs2_meta_inode_buffer(ip, &mp.mp_bh[0]);
	if (ret)
		goto out;
D
David Teigland 已提交
635

636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660
	height = ip->i_height;
	size = (lblock + 1) * bsize;
	while (size > arr[height])
		height++;
	find_metapath(sdp, lblock, &mp, height);
	ret = 1;
	if (height > ip->i_height || gfs2_is_stuffed(ip))
		goto do_alloc;
	ret = lookup_metapath(ip, &mp);
	if (ret < 0)
		goto out;
	if (ret != ip->i_height)
		goto do_alloc;
	ptr = metapointer(ip->i_height - 1, &mp);
	if (*ptr == 0)
		goto do_alloc;
	map_bh(bh_map, inode->i_sb, be64_to_cpu(*ptr));
	bh = mp.mp_bh[ip->i_height - 1];
	len = gfs2_extent_length(bh->b_data, bh->b_size, ptr, maxlen, &eob);
	bh_map->b_size = (len << inode->i_blkbits);
	if (eob)
		set_buffer_boundary(bh_map);
	ret = 0;
out:
	release_metapath(&mp);
S
Steven Whitehouse 已提交
661
	trace_gfs2_bmap(ip, bh_map, lblock, create, ret);
662 663
	bmap_unlock(ip, create);
	return ret;
664

665 666 667 668 669 670
do_alloc:
	/* All allocations are done here, firstly check create flag */
	if (!create) {
		BUG_ON(gfs2_is_stuffed(ip));
		ret = 0;
		goto out;
D
David Teigland 已提交
671
	}
672 673 674 675

	/* At this point ret is the tree depth of already allocated blocks */
	ret = gfs2_bmap_alloc(inode, lblock, bh_map, &mp, ret, height, maxlen);
	goto out;
S
Steven Whitehouse 已提交
676 677
}

678 679 680
/*
 * Deprecated: do not use in new code
 */
S
Steven Whitehouse 已提交
681 682
int gfs2_extent_map(struct inode *inode, u64 lblock, int *new, u64 *dblock, unsigned *extlen)
{
683
	struct buffer_head bh = { .b_state = 0, .b_blocknr = 0 };
684
	int ret;
S
Steven Whitehouse 已提交
685 686 687 688 689 690
	int create = *new;

	BUG_ON(!extlen);
	BUG_ON(!dblock);
	BUG_ON(!new);

691
	bh.b_size = 1 << (inode->i_blkbits + (create ? 0 : 5));
692
	ret = gfs2_block_map(inode, lblock, &bh, create);
693 694 695 696 697 698 699
	*extlen = bh.b_size >> inode->i_blkbits;
	*dblock = bh.b_blocknr;
	if (buffer_new(&bh))
		*new = 1;
	else
		*new = 0;
	return ret;
D
David Teigland 已提交
700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715
}

/**
 * do_strip - Look for a layer a particular layer of the file and strip it off
 * @ip: the inode
 * @dibh: the dinode buffer
 * @bh: A buffer of pointers
 * @top: The first pointer in the buffer
 * @bottom: One more than the last pointer
 * @height: the height this buffer is at
 * @data: a pointer to a struct strip_mine
 *
 * Returns: errno
 */

static int do_strip(struct gfs2_inode *ip, struct buffer_head *dibh,
A
Al Viro 已提交
716
		    struct buffer_head *bh, __be64 *top, __be64 *bottom,
717
		    unsigned int height, struct strip_mine *sm)
D
David Teigland 已提交
718
{
719
	struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
D
David Teigland 已提交
720
	struct gfs2_rgrp_list rlist;
721
	u64 bn, bstart;
722
	u32 blen, btotal;
A
Al Viro 已提交
723
	__be64 *p;
D
David Teigland 已提交
724 725 726 727
	unsigned int rg_blocks = 0;
	int metadata;
	unsigned int revokes = 0;
	int x;
728 729 730 731 732
	int error;

	error = gfs2_rindex_update(sdp);
	if (error)
		return error;
D
David Teigland 已提交
733 734 735 736 737 738 739 740 741 742 743 744

	if (!*top)
		sm->sm_first = 0;

	if (height != sm->sm_height)
		return 0;

	if (sm->sm_first) {
		top++;
		sm->sm_first = 0;
	}

745
	metadata = (height != ip->i_height - 1);
D
David Teigland 已提交
746 747
	if (metadata)
		revokes = (height) ? sdp->sd_inptrs : sdp->sd_diptrs;
748 749
	else if (ip->i_depth)
		revokes = sdp->sd_inptrs;
D
David Teigland 已提交
750 751 752 753 754 755 756 757 758 759 760 761 762 763 764

	memset(&rlist, 0, sizeof(struct gfs2_rgrp_list));
	bstart = 0;
	blen = 0;

	for (p = top; p < bottom; p++) {
		if (!*p)
			continue;

		bn = be64_to_cpu(*p);

		if (bstart + blen == bn)
			blen++;
		else {
			if (bstart)
765
				gfs2_rlist_add(ip, &rlist, bstart);
D
David Teigland 已提交
766 767 768 769 770 771 772

			bstart = bn;
			blen = 1;
		}
	}

	if (bstart)
773
		gfs2_rlist_add(ip, &rlist, bstart);
D
David Teigland 已提交
774 775 776
	else
		goto out; /* Nothing to do */

777
	gfs2_rlist_alloc(&rlist, LM_ST_EXCLUSIVE);
D
David Teigland 已提交
778 779 780

	for (x = 0; x < rlist.rl_rgrps; x++) {
		struct gfs2_rgrpd *rgd;
781
		rgd = rlist.rl_ghs[x].gh_gl->gl_object;
782
		rg_blocks += rgd->rd_length;
D
David Teigland 已提交
783 784 785 786 787 788
	}

	error = gfs2_glock_nq_m(rlist.rl_rgrps, rlist.rl_ghs);
	if (error)
		goto out_rlist;

B
Bob Peterson 已提交
789
	if (gfs2_rs_active(ip->i_res)) /* needs to be done with the rgrp glock held */
790
		gfs2_rs_deltree(ip->i_res);
B
Bob Peterson 已提交
791

D
David Teigland 已提交
792 793 794 795 796 797 798 799
	error = gfs2_trans_begin(sdp, rg_blocks + RES_DINODE +
				 RES_INDIRECT + RES_STATFS + RES_QUOTA,
				 revokes);
	if (error)
		goto out_rg_gunlock;

	down_write(&ip->i_rw_mutex);

800 801
	gfs2_trans_add_meta(ip->i_gl, dibh);
	gfs2_trans_add_meta(ip->i_gl, bh);
D
David Teigland 已提交
802 803 804

	bstart = 0;
	blen = 0;
805
	btotal = 0;
D
David Teigland 已提交
806 807 808 809 810 811 812 813 814 815 816

	for (p = top; p < bottom; p++) {
		if (!*p)
			continue;

		bn = be64_to_cpu(*p);

		if (bstart + blen == bn)
			blen++;
		else {
			if (bstart) {
817
				__gfs2_free_blocks(ip, bstart, blen, metadata);
818
				btotal += blen;
D
David Teigland 已提交
819 820 821 822 823 824 825
			}

			bstart = bn;
			blen = 1;
		}

		*p = 0;
826
		gfs2_add_inode_blocks(&ip->i_inode, -1);
D
David Teigland 已提交
827 828
	}
	if (bstart) {
829
		__gfs2_free_blocks(ip, bstart, blen, metadata);
830
		btotal += blen;
D
David Teigland 已提交
831 832
	}

833 834 835 836
	gfs2_statfs_change(sdp, 0, +btotal, 0);
	gfs2_quota_change(ip, -(s64)btotal, ip->i_inode.i_uid,
			  ip->i_inode.i_gid);

837
	ip->i_inode.i_mtime = ip->i_inode.i_ctime = CURRENT_TIME;
D
David Teigland 已提交
838

839
	gfs2_dinode_out(ip, dibh->b_data);
D
David Teigland 已提交
840 841 842 843 844

	up_write(&ip->i_rw_mutex);

	gfs2_trans_end(sdp);

845
out_rg_gunlock:
D
David Teigland 已提交
846
	gfs2_glock_dq_m(rlist.rl_rgrps, rlist.rl_ghs);
847
out_rlist:
D
David Teigland 已提交
848
	gfs2_rlist_free(&rlist);
849
out:
D
David Teigland 已提交
850 851 852
	return error;
}

853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874
/**
 * recursive_scan - recursively scan through the end of a file
 * @ip: the inode
 * @dibh: the dinode buffer
 * @mp: the path through the metadata to the point to start
 * @height: the height the recursion is at
 * @block: the indirect block to look at
 * @first: 1 if this is the first block
 * @sm: data opaque to this function to pass to @bc
 *
 * When this is first called @height and @block should be zero and
 * @first should be 1.
 *
 * Returns: errno
 */

static int recursive_scan(struct gfs2_inode *ip, struct buffer_head *dibh,
			  struct metapath *mp, unsigned int height,
			  u64 block, int first, struct strip_mine *sm)
{
	struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
	struct buffer_head *bh = NULL;
875
	__be64 *top, *bottom;
876 877 878 879 880 881 882 883 884 885 886 887 888
	u64 bn;
	int error;
	int mh_size = sizeof(struct gfs2_meta_header);

	if (!height) {
		error = gfs2_meta_inode_buffer(ip, &bh);
		if (error)
			return error;
		dibh = bh;

		top = (__be64 *)(bh->b_data + sizeof(struct gfs2_dinode)) + mp->mp_list[0];
		bottom = (__be64 *)(bh->b_data + sizeof(struct gfs2_dinode)) + sdp->sd_diptrs;
	} else {
889
		error = gfs2_meta_indirect_buffer(ip, height, block, &bh);
890 891 892 893 894 895 896 897 898 899 900 901 902
		if (error)
			return error;

		top = (__be64 *)(bh->b_data + mh_size) +
				  (first ? mp->mp_list[height] : 0);

		bottom = (__be64 *)(bh->b_data + mh_size) + sdp->sd_inptrs;
	}

	error = do_strip(ip, dibh, bh, top, bottom, height, sm);
	if (error)
		goto out;

903 904
	if (height < ip->i_height - 1) {

905
		gfs2_metapath_ra(ip->i_gl, bh, top);
906

907 908 909 910 911 912 913 914 915 916 917
		for (; top < bottom; top++, first = 0) {
			if (!*top)
				continue;

			bn = be64_to_cpu(*top);

			error = recursive_scan(ip, dibh, mp, height + 1, bn,
					       first, sm);
			if (error)
				break;
		}
918
	}
919 920 921 922 923 924
out:
	brelse(bh);
	return error;
}


S
Steven Whitehouse 已提交
925 926 927 928 929
/**
 * gfs2_block_truncate_page - Deal with zeroing out data for truncate
 *
 * This is partly borrowed from ext3.
 */
S
Steven Whitehouse 已提交
930
static int gfs2_block_truncate_page(struct address_space *mapping, loff_t from)
S
Steven Whitehouse 已提交
931 932 933 934 935 936 937 938 939 940
{
	struct inode *inode = mapping->host;
	struct gfs2_inode *ip = GFS2_I(inode);
	unsigned long index = from >> PAGE_CACHE_SHIFT;
	unsigned offset = from & (PAGE_CACHE_SIZE-1);
	unsigned blocksize, iblock, length, pos;
	struct buffer_head *bh;
	struct page *page;
	int err;

941
	page = find_or_create_page(mapping, index, GFP_NOFS);
S
Steven Whitehouse 已提交
942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963
	if (!page)
		return 0;

	blocksize = inode->i_sb->s_blocksize;
	length = blocksize - (offset & (blocksize - 1));
	iblock = index << (PAGE_CACHE_SHIFT - inode->i_sb->s_blocksize_bits);

	if (!page_has_buffers(page))
		create_empty_buffers(page, blocksize, 0);

	/* Find the buffer that contains "offset" */
	bh = page_buffers(page);
	pos = blocksize;
	while (offset >= pos) {
		bh = bh->b_this_page;
		iblock++;
		pos += blocksize;
	}

	err = 0;

	if (!buffer_mapped(bh)) {
964
		gfs2_block_map(inode, iblock, bh, 0);
S
Steven Whitehouse 已提交
965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980
		/* unmapped? It's a hole - nothing to do */
		if (!buffer_mapped(bh))
			goto unlock;
	}

	/* Ok, it's mapped. Make sure it's up-to-date */
	if (PageUptodate(page))
		set_buffer_uptodate(bh);

	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;
981
		err = 0;
S
Steven Whitehouse 已提交
982 983
	}

984
	if (!gfs2_is_writeback(ip))
985
		gfs2_trans_add_data(ip->i_gl, bh);
S
Steven Whitehouse 已提交
986

987
	zero_user(page, offset, length);
988
	mark_buffer_dirty(bh);
S
Steven Whitehouse 已提交
989 990 991 992 993 994
unlock:
	unlock_page(page);
	page_cache_release(page);
	return err;
}

995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018
/**
 * gfs2_journaled_truncate - Wrapper for truncate_pagecache for jdata files
 * @inode: The inode being truncated
 * @oldsize: The original (larger) size
 * @newsize: The new smaller size
 *
 * With jdata files, we have to journal a revoke for each block which is
 * truncated. As a result, we need to split this into separate transactions
 * if the number of pages being truncated gets too large.
 */

#define GFS2_JTRUNC_REVOKES 8192

static int gfs2_journaled_truncate(struct inode *inode, u64 oldsize, u64 newsize)
{
	struct gfs2_sbd *sdp = GFS2_SB(inode);
	u64 max_chunk = GFS2_JTRUNC_REVOKES * sdp->sd_vfs->s_blocksize;
	u64 chunk;
	int error;

	while (oldsize != newsize) {
		chunk = oldsize - newsize;
		if (chunk > max_chunk)
			chunk = max_chunk;
1019
		truncate_pagecache(inode, oldsize - chunk);
1020 1021 1022 1023 1024 1025 1026 1027 1028 1029
		oldsize -= chunk;
		gfs2_trans_end(sdp);
		error = gfs2_trans_begin(sdp, RES_DINODE, GFS2_JTRUNC_REVOKES);
		if (error)
			return error;
	}

	return 0;
}

S
Steven Whitehouse 已提交
1030
static int trunc_start(struct inode *inode, u64 oldsize, u64 newsize)
D
David Teigland 已提交
1031
{
S
Steven Whitehouse 已提交
1032 1033 1034
	struct gfs2_inode *ip = GFS2_I(inode);
	struct gfs2_sbd *sdp = GFS2_SB(inode);
	struct address_space *mapping = inode->i_mapping;
D
David Teigland 已提交
1035 1036 1037 1038
	struct buffer_head *dibh;
	int journaled = gfs2_is_jdata(ip);
	int error;

1039 1040 1041 1042
	if (journaled)
		error = gfs2_trans_begin(sdp, RES_DINODE + RES_JDATA, GFS2_JTRUNC_REVOKES);
	else
		error = gfs2_trans_begin(sdp, RES_DINODE, 0);
D
David Teigland 已提交
1043 1044 1045 1046 1047 1048 1049
	if (error)
		return error;

	error = gfs2_meta_inode_buffer(ip, &dibh);
	if (error)
		goto out;

1050
	gfs2_trans_add_meta(ip->i_gl, dibh);
S
Steven Whitehouse 已提交
1051

D
David Teigland 已提交
1052
	if (gfs2_is_stuffed(ip)) {
S
Steven Whitehouse 已提交
1053
		gfs2_buffer_clear_tail(dibh, sizeof(struct gfs2_dinode) + newsize);
D
David Teigland 已提交
1054
	} else {
S
Steven Whitehouse 已提交
1055 1056 1057 1058
		if (newsize & (u64)(sdp->sd_sb.sb_bsize - 1)) {
			error = gfs2_block_truncate_page(mapping, newsize);
			if (error)
				goto out_brelse;
D
David Teigland 已提交
1059
		}
S
Steven Whitehouse 已提交
1060
		ip->i_diskflags |= GFS2_DIF_TRUNC_IN_PROG;
D
David Teigland 已提交
1061 1062
	}

S
Steven Whitehouse 已提交
1063 1064 1065
	i_size_write(inode, newsize);
	ip->i_inode.i_mtime = ip->i_inode.i_ctime = CURRENT_TIME;
	gfs2_dinode_out(ip, dibh->b_data);
D
David Teigland 已提交
1066

1067 1068 1069
	if (journaled)
		error = gfs2_journaled_truncate(inode, oldsize, newsize);
	else
1070
		truncate_pagecache(inode, newsize);
1071 1072 1073 1074 1075 1076

	if (error) {
		brelse(dibh);
		return error;
	}

S
Steven Whitehouse 已提交
1077 1078
out_brelse:
	brelse(dibh);
1079
out:
D
David Teigland 已提交
1080 1081 1082 1083
	gfs2_trans_end(sdp);
	return error;
}

1084
static int trunc_dealloc(struct gfs2_inode *ip, u64 size)
D
David Teigland 已提交
1085
{
1086
	struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
1087
	unsigned int height = ip->i_height;
1088
	u64 lblock;
D
David Teigland 已提交
1089 1090 1091 1092 1093
	struct metapath mp;
	int error;

	if (!size)
		lblock = 0;
1094
	else
1095
		lblock = (size - 1) >> sdp->sd_sb.sb_bsize_shift;
D
David Teigland 已提交
1096

1097
	find_metapath(sdp, lblock, &mp, ip->i_height);
1098 1099 1100
	error = gfs2_rindex_update(sdp);
	if (error)
		return error;
D
David Teigland 已提交
1101

1102
	error = gfs2_quota_hold(ip, NO_UID_QUOTA_CHANGE, NO_GID_QUOTA_CHANGE);
D
David Teigland 已提交
1103
	if (error)
1104
		return error;
D
David Teigland 已提交
1105 1106 1107 1108 1109 1110

	while (height--) {
		struct strip_mine sm;
		sm.sm_first = !!size;
		sm.sm_height = height;

1111
		error = recursive_scan(ip, NULL, &mp, 0, 0, 1, &sm);
D
David Teigland 已提交
1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122
		if (error)
			break;
	}

	gfs2_quota_unhold(ip);

	return error;
}

static int trunc_end(struct gfs2_inode *ip)
{
1123
	struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
D
David Teigland 已提交
1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136
	struct buffer_head *dibh;
	int error;

	error = gfs2_trans_begin(sdp, RES_DINODE, 0);
	if (error)
		return error;

	down_write(&ip->i_rw_mutex);

	error = gfs2_meta_inode_buffer(ip, &dibh);
	if (error)
		goto out;

S
Steven Whitehouse 已提交
1137
	if (!i_size_read(&ip->i_inode)) {
1138
		ip->i_height = 0;
1139
		ip->i_goal = ip->i_no_addr;
D
David Teigland 已提交
1140
		gfs2_buffer_clear_tail(dibh, sizeof(struct gfs2_dinode));
1141
		gfs2_ordered_del_inode(ip);
D
David Teigland 已提交
1142
	}
1143
	ip->i_inode.i_mtime = ip->i_inode.i_ctime = CURRENT_TIME;
1144
	ip->i_diskflags &= ~GFS2_DIF_TRUNC_IN_PROG;
D
David Teigland 已提交
1145

1146
	gfs2_trans_add_meta(ip->i_gl, dibh);
1147
	gfs2_dinode_out(ip, dibh->b_data);
D
David Teigland 已提交
1148 1149
	brelse(dibh);

1150
out:
D
David Teigland 已提交
1151 1152 1153 1154 1155 1156 1157
	up_write(&ip->i_rw_mutex);
	gfs2_trans_end(sdp);
	return error;
}

/**
 * do_shrink - make a file smaller
S
Steven Whitehouse 已提交
1158 1159 1160
 * @inode: the inode
 * @oldsize: the current inode size
 * @newsize: the size to make the file
D
David Teigland 已提交
1161
 *
S
Steven Whitehouse 已提交
1162 1163
 * Called with an exclusive lock on @inode. The @size must
 * be equal to or smaller than the current inode size.
D
David Teigland 已提交
1164 1165 1166 1167
 *
 * Returns: errno
 */

S
Steven Whitehouse 已提交
1168
static int do_shrink(struct inode *inode, u64 oldsize, u64 newsize)
D
David Teigland 已提交
1169
{
S
Steven Whitehouse 已提交
1170
	struct gfs2_inode *ip = GFS2_I(inode);
D
David Teigland 已提交
1171 1172
	int error;

S
Steven Whitehouse 已提交
1173
	error = trunc_start(inode, oldsize, newsize);
D
David Teigland 已提交
1174 1175
	if (error < 0)
		return error;
S
Steven Whitehouse 已提交
1176
	if (gfs2_is_stuffed(ip))
D
David Teigland 已提交
1177 1178
		return 0;

S
Steven Whitehouse 已提交
1179 1180
	error = trunc_dealloc(ip, newsize);
	if (error == 0)
D
David Teigland 已提交
1181 1182 1183 1184 1185
		error = trunc_end(ip);

	return error;
}

S
Steven Whitehouse 已提交
1186
void gfs2_trim_blocks(struct inode *inode)
W
Wendy Cheng 已提交
1187
{
S
Steven Whitehouse 已提交
1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208
	u64 size = inode->i_size;
	int ret;

	ret = do_shrink(inode, size, size);
	WARN_ON(ret != 0);
}

/**
 * do_grow - Touch and update inode size
 * @inode: The inode
 * @size: The new size
 *
 * This function updates the timestamps on the inode and
 * may also increase the size of the inode. This function
 * must not be called with @size any smaller than the current
 * inode size.
 *
 * Although it is not strictly required to unstuff files here,
 * earlier versions of GFS2 have a bug in the stuffed file reading
 * code which will result in a buffer overrun if the size is larger
 * than the max stuffed file size. In order to prevent this from
L
Lucas De Marchi 已提交
1209
 * occurring, such files are unstuffed, but in other cases we can
S
Steven Whitehouse 已提交
1210 1211 1212 1213 1214 1215 1216 1217 1218
 * just update the inode size directly.
 *
 * Returns: 0 on success, or -ve on error
 */

static int do_grow(struct inode *inode, u64 size)
{
	struct gfs2_inode *ip = GFS2_I(inode);
	struct gfs2_sbd *sdp = GFS2_SB(inode);
W
Wendy Cheng 已提交
1219 1220
	struct buffer_head *dibh;
	int error;
1221
	int unstuff = 0;
W
Wendy Cheng 已提交
1222

S
Steven Whitehouse 已提交
1223 1224 1225 1226
	if (gfs2_is_stuffed(ip) &&
	    (size > (sdp->sd_sb.sb_bsize - sizeof(struct gfs2_dinode)))) {
		error = gfs2_quota_lock_check(ip);
		if (error)
1227
			return error;
S
Steven Whitehouse 已提交
1228

S
Steven Whitehouse 已提交
1229
		error = gfs2_inplace_reserve(ip, 1, 0);
S
Steven Whitehouse 已提交
1230 1231
		if (error)
			goto do_grow_qunlock;
1232
		unstuff = 1;
S
Steven Whitehouse 已提交
1233 1234
	}

1235 1236 1237
	error = gfs2_trans_begin(sdp, RES_DINODE + RES_STATFS + RES_RG_BIT +
				 (sdp->sd_args.ar_quota == GFS2_QUOTA_OFF ?
				  0 : RES_QUOTA), 0);
W
Wendy Cheng 已提交
1238
	if (error)
S
Steven Whitehouse 已提交
1239
		goto do_grow_release;
W
Wendy Cheng 已提交
1240

1241
	if (unstuff) {
S
Steven Whitehouse 已提交
1242 1243 1244 1245
		error = gfs2_unstuff_dinode(ip, NULL);
		if (error)
			goto do_end_trans;
	}
W
Wendy Cheng 已提交
1246 1247 1248

	error = gfs2_meta_inode_buffer(ip, &dibh);
	if (error)
S
Steven Whitehouse 已提交
1249
		goto do_end_trans;
W
Wendy Cheng 已提交
1250

S
Steven Whitehouse 已提交
1251
	i_size_write(inode, size);
W
Wendy Cheng 已提交
1252
	ip->i_inode.i_mtime = ip->i_inode.i_ctime = CURRENT_TIME;
1253
	gfs2_trans_add_meta(ip->i_gl, dibh);
W
Wendy Cheng 已提交
1254 1255 1256
	gfs2_dinode_out(ip, dibh->b_data);
	brelse(dibh);

S
Steven Whitehouse 已提交
1257
do_end_trans:
W
Wendy Cheng 已提交
1258
	gfs2_trans_end(sdp);
S
Steven Whitehouse 已提交
1259
do_grow_release:
1260
	if (unstuff) {
S
Steven Whitehouse 已提交
1261 1262 1263 1264
		gfs2_inplace_release(ip);
do_grow_qunlock:
		gfs2_quota_unlock(ip);
	}
W
Wendy Cheng 已提交
1265 1266 1267
	return error;
}

D
David Teigland 已提交
1268
/**
S
Steven Whitehouse 已提交
1269 1270 1271
 * gfs2_setattr_size - make a file a given size
 * @inode: the inode
 * @newsize: the size to make the file
D
David Teigland 已提交
1272
 *
S
Steven Whitehouse 已提交
1273 1274 1275
 * The file size can grow, shrink, or stay the same size. This
 * is called holding i_mutex and an exclusive glock on the inode
 * in question.
D
David Teigland 已提交
1276 1277 1278 1279
 *
 * Returns: errno
 */

S
Steven Whitehouse 已提交
1280
int gfs2_setattr_size(struct inode *inode, u64 newsize)
D
David Teigland 已提交
1281
{
1282
	struct gfs2_inode *ip = GFS2_I(inode);
S
Steven Whitehouse 已提交
1283 1284
	int ret;
	u64 oldsize;
D
David Teigland 已提交
1285

S
Steven Whitehouse 已提交
1286
	BUG_ON(!S_ISREG(inode->i_mode));
D
David Teigland 已提交
1287

S
Steven Whitehouse 已提交
1288 1289 1290
	ret = inode_newsize_ok(inode, newsize);
	if (ret)
		return ret;
D
David Teigland 已提交
1291

1292 1293 1294 1295
	ret = get_write_access(inode);
	if (ret)
		return ret;

1296 1297
	inode_dio_wait(inode);

1298
	ret = gfs2_rs_alloc(ip);
1299
	if (ret)
1300
		goto out;
1301

S
Steven Whitehouse 已提交
1302
	oldsize = inode->i_size;
1303 1304 1305 1306
	if (newsize >= oldsize) {
		ret = do_grow(inode, newsize);
		goto out;
	}
S
Steven Whitehouse 已提交
1307

1308
	gfs2_rs_deltree(ip->i_res);
1309 1310 1311 1312
	ret = do_shrink(inode, oldsize, newsize);
out:
	put_write_access(inode);
	return ret;
D
David Teigland 已提交
1313 1314 1315 1316 1317
}

int gfs2_truncatei_resume(struct gfs2_inode *ip)
{
	int error;
S
Steven Whitehouse 已提交
1318
	error = trunc_dealloc(ip, i_size_read(&ip->i_inode));
D
David Teigland 已提交
1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334
	if (!error)
		error = trunc_end(ip);
	return error;
}

int gfs2_file_dealloc(struct gfs2_inode *ip)
{
	return trunc_dealloc(ip, 0);
}

/**
 * gfs2_write_alloc_required - figure out if a write will require an allocation
 * @ip: the file being written to
 * @offset: the offset to write to
 * @len: the number of bytes being written
 *
1335
 * Returns: 1 if an alloc is required, 0 otherwise
D
David Teigland 已提交
1336 1337
 */

1338
int gfs2_write_alloc_required(struct gfs2_inode *ip, u64 offset,
1339
			      unsigned int len)
D
David Teigland 已提交
1340
{
1341
	struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
1342 1343 1344
	struct buffer_head bh;
	unsigned int shift;
	u64 lblock, lblock_stop, size;
1345
	u64 end_of_file;
D
David Teigland 已提交
1346 1347 1348 1349 1350 1351 1352

	if (!len)
		return 0;

	if (gfs2_is_stuffed(ip)) {
		if (offset + len >
		    sdp->sd_sb.sb_bsize - sizeof(struct gfs2_dinode))
1353
			return 1;
D
David Teigland 已提交
1354 1355 1356
		return 0;
	}

1357
	shift = sdp->sd_sb.sb_bsize_shift;
1358
	BUG_ON(gfs2_is_dir(ip));
S
Steven Whitehouse 已提交
1359
	end_of_file = (i_size_read(&ip->i_inode) + sdp->sd_sb.sb_bsize - 1) >> shift;
1360 1361 1362
	lblock = offset >> shift;
	lblock_stop = (offset + len + sdp->sd_sb.sb_bsize - 1) >> shift;
	if (lblock_stop > end_of_file)
1363
		return 1;
D
David Teigland 已提交
1364

1365 1366 1367 1368 1369 1370
	size = (lblock_stop - lblock) << shift;
	do {
		bh.b_state = 0;
		bh.b_size = size;
		gfs2_block_map(&ip->i_inode, lblock, &bh, 0);
		if (!buffer_mapped(&bh))
1371
			return 1;
1372 1373 1374
		size -= bh.b_size;
		lblock += (bh.b_size >> ip->i_inode.i_blkbits);
	} while(size > 0);
D
David Teigland 已提交
1375 1376 1377 1378

	return 0;
}