bmap.c 28.6 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 13 14
 */

#include <linux/sched.h>
#include <linux/slab.h>
#include <linux/spinlock.h>
#include <linux/completion.h>
#include <linux/buffer_head.h>
15
#include <linux/gfs2_ondisk.h>
16
#include <linux/crc32.h>
17
#include <linux/lm_interface.h>
D
David Teigland 已提交
18 19

#include "gfs2.h"
20
#include "incore.h"
D
David Teigland 已提交
21 22 23 24 25 26 27
#include "bmap.h"
#include "glock.h"
#include "inode.h"
#include "meta_io.h"
#include "quota.h"
#include "rgrp.h"
#include "trans.h"
28
#include "dir.h"
29
#include "util.h"
S
Steven Whitehouse 已提交
30
#include "ops_address.h"
D
David Teigland 已提交
31 32 33 34 35 36 37 38 39 40

/* 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 {
	__u16 mp_list[GFS2_MAX_META_HEIGHT];
};

typedef int (*block_call_t) (struct gfs2_inode *ip, struct buffer_head *dibh,
A
Al Viro 已提交
41 42
			     struct buffer_head *bh, __be64 *top,
			     __be64 *bottom, unsigned int height,
D
David Teigland 已提交
43 44 45 46 47 48 49
			     void *data);

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

50 51 52 53 54 55 56 57 58 59 60
/**
 * 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
 * @private: any locked page held by the caller process
 *
 * Returns: errno
 */

static int gfs2_unstuffer_page(struct gfs2_inode *ip, struct buffer_head *dibh,
61
			       u64 block, struct page *page)
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
{
	struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
	struct inode *inode = &ip->i_inode;
	struct buffer_head *bh;
	int release = 0;

	if (!page || page->index) {
		page = grab_cache_page(inode->i_mapping, 0);
		if (!page)
			return -ENOMEM;
		release = 1;
	}

	if (!PageUptodate(page)) {
		void *kaddr = kmap(page);

		memcpy(kaddr, dibh->b_data + sizeof(struct gfs2_dinode),
		       ip->i_di.di_size);
		memset(kaddr + ip->i_di.di_size, 0,
		       PAGE_CACHE_SIZE - ip->i_di.di_size);
		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);
97
	if (sdp->sd_args.ar_data == GFS2_DATA_ORDERED || gfs2_is_jdata(ip))
98 99 100 101 102 103 104 105 106 107 108
		gfs2_trans_add_bh(ip->i_gl, bh, 0);
	mark_buffer_dirty(bh);

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

	return 0;
}

D
David Teigland 已提交
109 110 111 112 113 114 115 116 117 118 119 120
/**
 * gfs2_unstuff_dinode - Unstuff a dinode when the data has grown too big
 * @ip: The GFS2 inode to unstuff
 * @unstuffer: the routine that handles unstuffing a non-zero length file
 * @private: private data for the unstuffer
 *
 * 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
 */

121
int gfs2_unstuff_dinode(struct gfs2_inode *ip, struct page *page)
D
David Teigland 已提交
122 123
{
	struct buffer_head *bh, *dibh;
124
	struct gfs2_dinode *di;
125
	u64 block = 0;
126
	int isdir = gfs2_is_dir(ip);
D
David Teigland 已提交
127 128 129 130 131 132 133
	int error;

	down_write(&ip->i_rw_mutex);

	error = gfs2_meta_inode_buffer(ip, &dibh);
	if (error)
		goto out;
134

D
David Teigland 已提交
135 136 137 138
	if (ip->i_di.di_size) {
		/* Get a free block, fill it with the stuffed data,
		   and write it out to disk */

139
		if (isdir) {
D
David Teigland 已提交
140 141
			block = gfs2_alloc_meta(ip);

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
146 147 148 149 150
					      dibh, sizeof(struct gfs2_dinode));
			brelse(bh);
		} else {
			block = gfs2_alloc_data(ip);

151
			error = gfs2_unstuffer_page(ip, dibh, block, page);
D
David Teigland 已提交
152 153 154 155 156 157 158
			if (error)
				goto out_brelse;
		}
	}

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

159
	gfs2_trans_add_bh(ip->i_gl, dibh, 1);
160
	di = (struct gfs2_dinode *)dibh->b_data;
D
David Teigland 已提交
161 162 163
	gfs2_buffer_clear_tail(dibh, sizeof(struct gfs2_dinode));

	if (ip->i_di.di_size) {
164
		*(__be64 *)(di + 1) = cpu_to_be64(block);
D
David Teigland 已提交
165
		ip->i_di.di_blocks++;
166
		di->di_blocks = cpu_to_be64(ip->i_di.di_blocks);
D
David Teigland 已提交
167 168 169
	}

	ip->i_di.di_height = 1;
170
	di->di_height = cpu_to_be16(1);
D
David Teigland 已提交
171

172
out_brelse:
D
David Teigland 已提交
173
	brelse(dibh);
174
out:
D
David Teigland 已提交
175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191
	up_write(&ip->i_rw_mutex);
	return error;
}

/**
 * calc_tree_height - Calculate the height of a metadata tree
 * @ip: The GFS2 inode
 * @size: The proposed size of the file
 *
 * Work out how tall a metadata tree needs to be in order to accommodate a
 * file of a particular size. If size is less than the current size of
 * the inode, then the current size of the inode is used instead of the
 * supplied one.
 *
 * Returns: the height the tree should be
 */

192
static unsigned int calc_tree_height(struct gfs2_inode *ip, u64 size)
D
David Teigland 已提交
193
{
194
	struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
195
	u64 *arr;
D
David Teigland 已提交
196 197 198 199 200
	unsigned int max, height;

	if (ip->i_di.di_size > size)
		size = ip->i_di.di_size;

201
	if (gfs2_is_dir(ip)) {
D
David Teigland 已提交
202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224
		arr = sdp->sd_jheightsize;
		max = sdp->sd_max_jheight;
	} else {
		arr = sdp->sd_heightsize;
		max = sdp->sd_max_height;
	}

	for (height = 0; height < max; height++)
		if (arr[height] >= size)
			break;

	return height;
}

/**
 * build_height - Build a metadata tree of the requested height
 * @ip: The GFS2 inode
 * @height: The height to build to
 *
 *
 * Returns: errno
 */

225
static int build_height(struct inode *inode, unsigned height)
D
David Teigland 已提交
226
{
227
	struct gfs2_inode *ip = GFS2_I(inode);
228 229 230
	unsigned new_height = height - ip->i_di.di_height;
	struct buffer_head *dibh;
	struct buffer_head *blocks[GFS2_MAX_META_HEIGHT];
231
	struct gfs2_dinode *di;
D
David Teigland 已提交
232
	int error;
A
Al Viro 已提交
233
	__be64 *bp;
234 235
	u64 bn;
	unsigned n;
D
David Teigland 已提交
236

237 238
	if (height <= ip->i_di.di_height)
		return 0;
D
David Teigland 已提交
239

240 241 242
	error = gfs2_meta_inode_buffer(ip, &dibh);
	if (error)
		return error;
D
David Teigland 已提交
243

244 245 246 247 248
	for(n = 0; n < new_height; n++) {
		bn = gfs2_alloc_meta(ip);
		blocks[n] = gfs2_meta_new(ip->i_gl, bn);
		gfs2_trans_add_bh(ip->i_gl, blocks[n], 1);
	}
249

250 251 252 253 254
	n = 0;
	bn = blocks[0]->b_blocknr;
	if (new_height > 1) {
		for(; n < new_height-1; n++) {
			gfs2_metatype_set(blocks[n], GFS2_METATYPE_IN,
D
David Teigland 已提交
255
					  GFS2_FORMAT_IN);
256 257
			gfs2_buffer_clear_tail(blocks[n],
					       sizeof(struct gfs2_meta_header));
A
Al Viro 已提交
258
			bp = (__be64 *)(blocks[n]->b_data +
259 260 261 262
				     sizeof(struct gfs2_meta_header));
			*bp = cpu_to_be64(blocks[n+1]->b_blocknr);
			brelse(blocks[n]);
			blocks[n] = NULL;
D
David Teigland 已提交
263 264
		}
	}
265 266 267 268 269
	gfs2_metatype_set(blocks[n], GFS2_METATYPE_IN, GFS2_FORMAT_IN);
	gfs2_buffer_copy_tail(blocks[n], sizeof(struct gfs2_meta_header),
			      dibh, sizeof(struct gfs2_dinode));
	brelse(blocks[n]);
	gfs2_trans_add_bh(ip->i_gl, dibh, 1);
270
	di = (struct gfs2_dinode *)dibh->b_data;
271
	gfs2_buffer_clear_tail(dibh, sizeof(struct gfs2_dinode));
272
	*(__be64 *)(di + 1) = cpu_to_be64(bn);
273 274
	ip->i_di.di_height += new_height;
	ip->i_di.di_blocks += new_height;
275 276
	di->di_height = cpu_to_be16(ip->i_di.di_height);
	di->di_blocks = cpu_to_be64(ip->i_di.di_blocks);
277 278
	brelse(dibh);
	return error;
D
David Teigland 已提交
279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339
}

/**
 * find_metapath - Find path through the metadata tree
 * @ip: The inode pointer
 * @mp: The metapath to return the result in
 * @block: The disk block to look up
 *
 *   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                 |
 *             |                                      |
 *             |                                      |
 *             ----------------------------------------
 *
 */

340
static void find_metapath(struct gfs2_inode *ip, u64 block,
341
			  struct metapath *mp)
D
David Teigland 已提交
342
{
343
	struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
344
	u64 b = block;
D
David Teigland 已提交
345 346 347
	unsigned int i;

	for (i = ip->i_di.di_height; i--;)
348
		mp->mp_list[i] = do_div(b, sdp->sd_inptrs);
D
David Teigland 已提交
349 350 351 352 353 354 355 356 357 358 359 360 361 362

}

/**
 * metapointer - Return pointer to start of metadata in a buffer
 * @bh: The 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.
 */

A
Al Viro 已提交
363
static inline __be64 *metapointer(struct buffer_head *bh, int *boundary,
S
Steven Whitehouse 已提交
364
			       unsigned int height, const struct metapath *mp)
D
David Teigland 已提交
365 366 367
{
	unsigned int head_size = (height > 0) ?
		sizeof(struct gfs2_meta_header) : sizeof(struct gfs2_dinode);
A
Al Viro 已提交
368
	__be64 *ptr;
S
Steven Whitehouse 已提交
369
	*boundary = 0;
A
Al Viro 已提交
370 371
	ptr = ((__be64 *)(bh->b_data + head_size)) + mp->mp_list[height];
	if (ptr + 1 == (__be64 *)(bh->b_data + bh->b_size))
S
Steven Whitehouse 已提交
372 373
		*boundary = 1;
	return ptr;
D
David Teigland 已提交
374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391
}

/**
 * lookup_block - Get the next metadata block in metadata tree
 * @ip: The GFS2 inode
 * @bh: Buffer containing the pointers to metadata blocks
 * @height: The height of the tree (0 = dinode)
 * @mp: The metapath
 * @create: Non-zero if we may create a new meatdata block
 * @new: Used to indicate if we did create a new metadata block
 * @block: the returned disk block number
 *
 * Given a metatree, complete to a particular height, checks to see if the next
 * height of the tree exists. If not the next height of the tree is created.
 * The block number of the next height of the metadata tree is returned.
 *
 */

S
Steven Whitehouse 已提交
392 393
static int lookup_block(struct gfs2_inode *ip, struct buffer_head *bh,
			unsigned int height, struct metapath *mp, int create,
394
			int *new, u64 *block)
D
David Teigland 已提交
395
{
S
Steven Whitehouse 已提交
396
	int boundary;
A
Al Viro 已提交
397
	__be64 *ptr = metapointer(bh, &boundary, height, mp);
D
David Teigland 已提交
398 399 400

	if (*ptr) {
		*block = be64_to_cpu(*ptr);
S
Steven Whitehouse 已提交
401
		return boundary;
D
David Teigland 已提交
402 403 404 405 406
	}

	*block = 0;

	if (!create)
S
Steven Whitehouse 已提交
407
		return 0;
D
David Teigland 已提交
408

S
Steven Whitehouse 已提交
409
	if (height == ip->i_di.di_height - 1 && !gfs2_is_dir(ip))
D
David Teigland 已提交
410 411 412 413
		*block = gfs2_alloc_data(ip);
	else
		*block = gfs2_alloc_meta(ip);

414
	gfs2_trans_add_bh(ip->i_gl, bh, 1);
D
David Teigland 已提交
415 416 417 418 419

	*ptr = cpu_to_be64(*block);
	ip->i_di.di_blocks++;

	*new = 1;
S
Steven Whitehouse 已提交
420
	return 0;
D
David Teigland 已提交
421 422 423
}

/**
S
Steven Whitehouse 已提交
424 425
 * gfs2_block_pointers - Map a block from an inode to a disk block
 * @inode: The inode
D
David Teigland 已提交
426
 * @lblock: The logical block number
427
 * @map_bh: The bh to be mapped
S
Steven Whitehouse 已提交
428
 * @mp: metapath to use
D
David Teigland 已提交
429 430 431 432 433 434 435
 *
 * Find the block number on the current device which corresponds to an
 * inode's block. If the block had to be created, "new" will be set.
 *
 * Returns: errno
 */

436
static int gfs2_block_pointers(struct inode *inode, u64 lblock, int create,
437
			       struct buffer_head *bh_map, struct metapath *mp)
D
David Teigland 已提交
438
{
439 440
	struct gfs2_inode *ip = GFS2_I(inode);
	struct gfs2_sbd *sdp = GFS2_SB(inode);
D
David Teigland 已提交
441 442 443 444 445 446
	struct buffer_head *bh;
	unsigned int bsize;
	unsigned int height;
	unsigned int end_of_metadata;
	unsigned int x;
	int error = 0;
447 448 449
	int new = 0;
	u64 dblock = 0;
	int boundary;
450
	unsigned int maxlen = bh_map->b_size >> inode->i_blkbits;
D
David Teigland 已提交
451

S
Steven Whitehouse 已提交
452 453
	BUG_ON(maxlen == 0);

D
David Teigland 已提交
454
	if (gfs2_assert_warn(sdp, !gfs2_is_stuffed(ip)))
455
		return 0;
D
David Teigland 已提交
456

457
	bsize = gfs2_is_dir(ip) ? sdp->sd_jbsize : sdp->sd_sb.sb_bsize;
D
David Teigland 已提交
458 459 460 461

	height = calc_tree_height(ip, (lblock + 1) * bsize);
	if (ip->i_di.di_height < height) {
		if (!create)
462
			return 0;
D
David Teigland 已提交
463

464
		error = build_height(inode, height);
D
David Teigland 已提交
465
		if (error)
466
			return error;
D
David Teigland 已提交
467 468
	}

S
Steven Whitehouse 已提交
469
	find_metapath(ip, lblock, mp);
D
David Teigland 已提交
470 471 472 473
	end_of_metadata = ip->i_di.di_height - 1;

	error = gfs2_meta_inode_buffer(ip, &bh);
	if (error)
474
		return error;
D
David Teigland 已提交
475 476

	for (x = 0; x < end_of_metadata; x++) {
477
		lookup_block(ip, bh, x, mp, create, &new, &dblock);
D
David Teigland 已提交
478
		brelse(bh);
479 480
		if (!dblock)
			return 0;
D
David Teigland 已提交
481

482
		error = gfs2_meta_indirect_buffer(ip, x+1, dblock, new, &bh);
D
David Teigland 已提交
483
		if (error)
484
			return error;
D
David Teigland 已提交
485 486
	}

487 488 489 490 491 492 493 494 495 496 497 498 499 500
	boundary = lookup_block(ip, bh, end_of_metadata, mp, create, &new, &dblock);
	clear_buffer_mapped(bh_map);
	clear_buffer_new(bh_map);
	clear_buffer_boundary(bh_map);

	if (dblock) {
		map_bh(bh_map, inode->i_sb, dblock);
		if (boundary)
			set_buffer_boundary(bh);
		if (new) {
			struct buffer_head *dibh;
			error = gfs2_meta_inode_buffer(ip, &dibh);
			if (!error) {
				gfs2_trans_add_bh(ip->i_gl, dibh, 1);
501
				gfs2_dinode_out(ip, dibh->b_data);
502 503 504 505 506 507 508 509 510 511 512 513
				brelse(dibh);
			}
			set_buffer_new(bh_map);
			goto out_brelse;
		}
		while(--maxlen && !buffer_boundary(bh_map)) {
			u64 eblock;

			mp->mp_list[end_of_metadata]++;
			boundary = lookup_block(ip, bh, end_of_metadata, mp, 0, &new, &eblock);
			if (eblock != ++dblock)
				break;
514
			bh_map->b_size += (1 << inode->i_blkbits);
515 516
			if (boundary)
				set_buffer_boundary(bh_map);
D
David Teigland 已提交
517 518
		}
	}
519 520 521
out_brelse:
	brelse(bh);
	return 0;
S
Steven Whitehouse 已提交
522
}
D
David Teigland 已提交
523

S
Steven Whitehouse 已提交
524 525 526

static inline void bmap_lock(struct inode *inode, int create)
{
527
	struct gfs2_inode *ip = GFS2_I(inode);
S
Steven Whitehouse 已提交
528 529 530 531 532 533 534 535
	if (create)
		down_write(&ip->i_rw_mutex);
	else
		down_read(&ip->i_rw_mutex);
}

static inline void bmap_unlock(struct inode *inode, int create)
{
536
	struct gfs2_inode *ip = GFS2_I(inode);
D
David Teigland 已提交
537 538 539 540
	if (create)
		up_write(&ip->i_rw_mutex);
	else
		up_read(&ip->i_rw_mutex);
S
Steven Whitehouse 已提交
541
}
D
David Teigland 已提交
542

543
int gfs2_block_map(struct inode *inode, u64 lblock, int create,
544
		   struct buffer_head *bh)
S
Steven Whitehouse 已提交
545 546
{
	struct metapath mp;
547
	int ret;
S
Steven Whitehouse 已提交
548 549

	bmap_lock(inode, create);
550
	ret = gfs2_block_pointers(inode, lblock, create, bh, &mp);
S
Steven Whitehouse 已提交
551
	bmap_unlock(inode, create);
552
	return ret;
S
Steven Whitehouse 已提交
553 554 555 556 557
}

int gfs2_extent_map(struct inode *inode, u64 lblock, int *new, u64 *dblock, unsigned *extlen)
{
	struct metapath mp;
558
	struct buffer_head bh = { .b_state = 0, .b_blocknr = 0 };
559
	int ret;
S
Steven Whitehouse 已提交
560 561 562 563 564 565
	int create = *new;

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

566
	bh.b_size = 1 << (inode->i_blkbits + 5);
S
Steven Whitehouse 已提交
567
	bmap_lock(inode, create);
568
	ret = gfs2_block_pointers(inode, lblock, create, &bh, &mp);
S
Steven Whitehouse 已提交
569
	bmap_unlock(inode, create);
570 571 572 573 574 575 576
	*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
577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597
}

/**
 * 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
 * @bc: the call to make for each piece of metadata
 * @data: 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,
598
			  u64 block, int first, block_call_t bc,
D
David Teigland 已提交
599 600
			  void *data)
{
601
	struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
D
David Teigland 已提交
602
	struct buffer_head *bh = NULL;
A
Al Viro 已提交
603
	__be64 *top, *bottom;
604
	u64 bn;
D
David Teigland 已提交
605 606 607 608 609 610 611 612 613
	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;

A
Al Viro 已提交
614 615
		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;
D
David Teigland 已提交
616 617 618 619 620
	} else {
		error = gfs2_meta_indirect_buffer(ip, height, block, 0, &bh);
		if (error)
			return error;

A
Al Viro 已提交
621
		top = (__be64 *)(bh->b_data + mh_size) +
J
Jan Engelhardt 已提交
622
				  (first ? mp->mp_list[height] : 0);
D
David Teigland 已提交
623

A
Al Viro 已提交
624
		bottom = (__be64 *)(bh->b_data + mh_size) + sdp->sd_inptrs;
D
David Teigland 已提交
625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643
	}

	error = bc(ip, dibh, bh, top, bottom, height, data);
	if (error)
		goto out;

	if (height < ip->i_di.di_height - 1)
		for (; top < bottom; top++, first = 0) {
			if (!*top)
				continue;

			bn = be64_to_cpu(*top);

			error = recursive_scan(ip, dibh, mp, height + 1, bn,
					       first, bc, data);
			if (error)
				break;
		}

644
out:
D
David Teigland 已提交
645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662
	brelse(bh);
	return error;
}

/**
 * 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
663
		    struct buffer_head *bh, __be64 *top, __be64 *bottom,
D
David Teigland 已提交
664 665
		    unsigned int height, void *data)
{
666 667
	struct strip_mine *sm = data;
	struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
D
David Teigland 已提交
668
	struct gfs2_rgrp_list rlist;
669 670
	u64 bn, bstart;
	u32 blen;
A
Al Viro 已提交
671
	__be64 *p;
D
David Teigland 已提交
672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688
	unsigned int rg_blocks = 0;
	int metadata;
	unsigned int revokes = 0;
	int x;
	int error;

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

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

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

689
	metadata = (height != ip->i_di.di_height - 1);
D
David Teigland 已提交
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
	if (metadata)
		revokes = (height) ? sdp->sd_inptrs : sdp->sd_diptrs;

	error = gfs2_rindex_hold(sdp, &ip->i_alloc.al_ri_gh);
	if (error)
		return error;

	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)
				gfs2_rlist_add(sdp, &rlist, bstart);

			bstart = bn;
			blen = 1;
		}
	}

	if (bstart)
		gfs2_rlist_add(sdp, &rlist, bstart);
	else
		goto out; /* Nothing to do */

	gfs2_rlist_alloc(&rlist, LM_ST_EXCLUSIVE, 0);

	for (x = 0; x < rlist.rl_rgrps; x++) {
		struct gfs2_rgrpd *rgd;
727
		rgd = rlist.rl_ghs[x].gh_gl->gl_object;
D
David Teigland 已提交
728 729 730 731 732 733 734 735 736 737 738 739 740 741 742
		rg_blocks += rgd->rd_ri.ri_length;
	}

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

	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);

743 744
	gfs2_trans_add_bh(ip->i_gl, dibh, 1);
	gfs2_trans_add_bh(ip->i_gl, bh, 1);
D
David Teigland 已提交
745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782

	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) {
				if (metadata)
					gfs2_free_meta(ip, bstart, blen);
				else
					gfs2_free_data(ip, bstart, blen);
			}

			bstart = bn;
			blen = 1;
		}

		*p = 0;
		if (!ip->i_di.di_blocks)
			gfs2_consist_inode(ip);
		ip->i_di.di_blocks--;
	}
	if (bstart) {
		if (metadata)
			gfs2_free_meta(ip, bstart, blen);
		else
			gfs2_free_data(ip, bstart, blen);
	}

	ip->i_di.di_mtime = ip->i_di.di_ctime = get_seconds();

783
	gfs2_dinode_out(ip, dibh->b_data);
D
David Teigland 已提交
784 785 786 787 788

	up_write(&ip->i_rw_mutex);

	gfs2_trans_end(sdp);

789
out_rg_gunlock:
D
David Teigland 已提交
790
	gfs2_glock_dq_m(rlist.rl_rgrps, rlist.rl_ghs);
791
out_rlist:
D
David Teigland 已提交
792
	gfs2_rlist_free(&rlist);
793
out:
D
David Teigland 已提交
794 795 796 797 798 799 800 801 802 803 804 805 806 807
	gfs2_glock_dq_uninit(&ip->i_alloc.al_ri_gh);
	return error;
}

/**
 * do_grow - Make a file look bigger than it is
 * @ip: the inode
 * @size: the size to set the file to
 *
 * Called with an exclusive lock on @ip.
 *
 * Returns: errno
 */

808
static int do_grow(struct gfs2_inode *ip, u64 size)
D
David Teigland 已提交
809
{
810
	struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
D
David Teigland 已提交
811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839
	struct gfs2_alloc *al;
	struct buffer_head *dibh;
	unsigned int h;
	int error;

	al = gfs2_alloc_get(ip);

	error = gfs2_quota_lock(ip, NO_QUOTA_CHANGE, NO_QUOTA_CHANGE);
	if (error)
		goto out;

	error = gfs2_quota_check(ip, ip->i_di.di_uid, ip->i_di.di_gid);
	if (error)
		goto out_gunlock_q;

	al->al_requested = sdp->sd_max_height + RES_DATA;

	error = gfs2_inplace_reserve(ip);
	if (error)
		goto out_gunlock_q;

	error = gfs2_trans_begin(sdp,
			sdp->sd_max_height + al->al_rgd->rd_ri.ri_length +
			RES_JDATA + RES_DINODE + RES_STATFS + RES_QUOTA, 0);
	if (error)
		goto out_ipres;

	if (size > sdp->sd_sb.sb_bsize - sizeof(struct gfs2_dinode)) {
		if (gfs2_is_stuffed(ip)) {
840
			error = gfs2_unstuff_dinode(ip, NULL);
D
David Teigland 已提交
841 842 843 844 845 846 847
			if (error)
				goto out_end_trans;
		}

		h = calc_tree_height(ip, size);
		if (ip->i_di.di_height < h) {
			down_write(&ip->i_rw_mutex);
848
			error = build_height(&ip->i_inode, h);
D
David Teigland 已提交
849 850 851 852 853 854 855 856 857 858 859 860 861
			up_write(&ip->i_rw_mutex);
			if (error)
				goto out_end_trans;
		}
	}

	ip->i_di.di_size = size;
	ip->i_di.di_mtime = ip->i_di.di_ctime = get_seconds();

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

862
	gfs2_trans_add_bh(ip->i_gl, dibh, 1);
863
	gfs2_dinode_out(ip, dibh->b_data);
D
David Teigland 已提交
864 865
	brelse(dibh);

866
out_end_trans:
D
David Teigland 已提交
867
	gfs2_trans_end(sdp);
868
out_ipres:
D
David Teigland 已提交
869
	gfs2_inplace_release(ip);
870
out_gunlock_q:
D
David Teigland 已提交
871
	gfs2_quota_unlock(ip);
872
out:
D
David Teigland 已提交
873 874 875 876
	gfs2_alloc_put(ip);
	return error;
}

S
Steven Whitehouse 已提交
877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952

/**
 * gfs2_block_truncate_page - Deal with zeroing out data for truncate
 *
 * This is partly borrowed from ext3.
 */
static int gfs2_block_truncate_page(struct address_space *mapping)
{
	struct inode *inode = mapping->host;
	struct gfs2_inode *ip = GFS2_I(inode);
	struct gfs2_sbd *sdp = GFS2_SB(inode);
	loff_t from = inode->i_size;
	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;
	void *kaddr;
	int err;

	page = grab_cache_page(mapping, index);
	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)) {
		gfs2_get_block(inode, iblock, bh, 0);
		/* 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;
	}

	if (sdp->sd_args.ar_data == GFS2_DATA_ORDERED || gfs2_is_jdata(ip))
		gfs2_trans_add_bh(ip->i_gl, bh, 0);

	kaddr = kmap_atomic(page, KM_USER0);
	memset(kaddr + offset, 0, length);
	flush_dcache_page(page);
	kunmap_atomic(kaddr, KM_USER0);

unlock:
	unlock_page(page);
	page_cache_release(page);
	return err;
}

953
static int trunc_start(struct gfs2_inode *ip, u64 size)
D
David Teigland 已提交
954
{
955
	struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
D
David Teigland 已提交
956 957 958 959 960
	struct buffer_head *dibh;
	int journaled = gfs2_is_jdata(ip);
	int error;

	error = gfs2_trans_begin(sdp,
J
Jan Engelhardt 已提交
961
				 RES_DINODE + (journaled ? RES_JDATA : 0), 0);
D
David Teigland 已提交
962 963 964 965 966 967 968 969 970 971
	if (error)
		return error;

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

	if (gfs2_is_stuffed(ip)) {
		ip->i_di.di_size = size;
		ip->i_di.di_mtime = ip->i_di.di_ctime = get_seconds();
972
		gfs2_trans_add_bh(ip->i_gl, dibh, 1);
973
		gfs2_dinode_out(ip, dibh->b_data);
D
David Teigland 已提交
974 975 976 977
		gfs2_buffer_clear_tail(dibh, sizeof(struct gfs2_dinode) + size);
		error = 1;

	} else {
978
		if (size & (u64)(sdp->sd_sb.sb_bsize - 1))
979
			error = gfs2_block_truncate_page(ip->i_inode.i_mapping);
D
David Teigland 已提交
980 981 982 983 984

		if (!error) {
			ip->i_di.di_size = size;
			ip->i_di.di_mtime = ip->i_di.di_ctime = get_seconds();
			ip->i_di.di_flags |= GFS2_DIF_TRUNC_IN_PROG;
985
			gfs2_trans_add_bh(ip->i_gl, dibh, 1);
986
			gfs2_dinode_out(ip, dibh->b_data);
D
David Teigland 已提交
987 988 989 990 991
		}
	}

	brelse(dibh);

992
out:
D
David Teigland 已提交
993 994 995 996
	gfs2_trans_end(sdp);
	return error;
}

997
static int trunc_dealloc(struct gfs2_inode *ip, u64 size)
D
David Teigland 已提交
998 999
{
	unsigned int height = ip->i_di.di_height;
1000
	u64 lblock;
D
David Teigland 已提交
1001 1002 1003 1004 1005
	struct metapath mp;
	int error;

	if (!size)
		lblock = 0;
1006
	else
1007
		lblock = (size - 1) >> GFS2_SB(&ip->i_inode)->sd_sb.sb_bsize_shift;
D
David Teigland 已提交
1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027

	find_metapath(ip, lblock, &mp);
	gfs2_alloc_get(ip);

	error = gfs2_quota_hold(ip, NO_QUOTA_CHANGE, NO_QUOTA_CHANGE);
	if (error)
		goto out;

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

		error = recursive_scan(ip, NULL, &mp, 0, 0, 1, do_strip, &sm);
		if (error)
			break;
	}

	gfs2_quota_unhold(ip);

1028
out:
D
David Teigland 已提交
1029 1030 1031 1032 1033 1034
	gfs2_alloc_put(ip);
	return error;
}

static int trunc_end(struct gfs2_inode *ip)
{
1035
	struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
D
David Teigland 已提交
1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058
	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;

	if (!ip->i_di.di_size) {
		ip->i_di.di_height = 0;
		ip->i_di.di_goal_meta =
			ip->i_di.di_goal_data =
			ip->i_num.no_addr;
		gfs2_buffer_clear_tail(dibh, sizeof(struct gfs2_dinode));
	}
	ip->i_di.di_mtime = ip->i_di.di_ctime = get_seconds();
	ip->i_di.di_flags &= ~GFS2_DIF_TRUNC_IN_PROG;

1059
	gfs2_trans_add_bh(ip->i_gl, dibh, 1);
1060
	gfs2_dinode_out(ip, dibh->b_data);
D
David Teigland 已提交
1061 1062
	brelse(dibh);

1063
out:
D
David Teigland 已提交
1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079
	up_write(&ip->i_rw_mutex);
	gfs2_trans_end(sdp);
	return error;
}

/**
 * do_shrink - make a file smaller
 * @ip: the inode
 * @size: the size to make the file
 * @truncator: function to truncate the last partial block
 *
 * Called with an exclusive lock on @ip.
 *
 * Returns: errno
 */

1080
static int do_shrink(struct gfs2_inode *ip, u64 size)
D
David Teigland 已提交
1081 1082 1083
{
	int error;

1084
	error = trunc_start(ip, size);
D
David Teigland 已提交
1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097
	if (error < 0)
		return error;
	if (error > 0)
		return 0;

	error = trunc_dealloc(ip, size);
	if (!error)
		error = trunc_end(ip);

	return error;
}

/**
1098
 * gfs2_truncatei - make a file a given size
D
David Teigland 已提交
1099 1100 1101 1102 1103 1104 1105 1106 1107
 * @ip: the inode
 * @size: the size to make the file
 * @truncator: function to truncate the last partial block
 *
 * The file size can grow, shrink, or stay the same size.
 *
 * Returns: errno
 */

1108
int gfs2_truncatei(struct gfs2_inode *ip, u64 size)
D
David Teigland 已提交
1109 1110 1111
{
	int error;

1112
	if (gfs2_assert_warn(GFS2_SB(&ip->i_inode), S_ISREG(ip->i_di.di_mode)))
D
David Teigland 已提交
1113 1114 1115 1116 1117
		return -EINVAL;

	if (size > ip->i_di.di_size)
		error = do_grow(ip, size);
	else
1118
		error = do_shrink(ip, size);
D
David Teigland 已提交
1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148

	return error;
}

int gfs2_truncatei_resume(struct gfs2_inode *ip)
{
	int error;
	error = trunc_dealloc(ip, ip->i_di.di_size);
	if (!error)
		error = trunc_end(ip);
	return error;
}

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

/**
 * gfs2_write_calc_reserv - calculate number of blocks needed to write to a file
 * @ip: the file
 * @len: the number of bytes to be written to the file
 * @data_blocks: returns the number of data blocks required
 * @ind_blocks: returns the number of indirect blocks required
 *
 */

void gfs2_write_calc_reserv(struct gfs2_inode *ip, unsigned int len,
			    unsigned int *data_blocks, unsigned int *ind_blocks)
{
1149
	struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
D
David Teigland 已提交
1150 1151
	unsigned int tmp;

1152
	if (gfs2_is_dir(ip)) {
1153
		*data_blocks = DIV_ROUND_UP(len, sdp->sd_jbsize) + 2;
D
David Teigland 已提交
1154 1155 1156 1157 1158 1159 1160
		*ind_blocks = 3 * (sdp->sd_max_jheight - 1);
	} else {
		*data_blocks = (len >> sdp->sd_sb.sb_bsize_shift) + 3;
		*ind_blocks = 3 * (sdp->sd_max_height - 1);
	}

	for (tmp = *data_blocks; tmp > sdp->sd_diptrs;) {
1161
		tmp = DIV_ROUND_UP(tmp, sdp->sd_inptrs);
D
David Teigland 已提交
1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175
		*ind_blocks += tmp;
	}
}

/**
 * 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
 * @alloc_required: set to 1 if an alloc is required, 0 otherwise
 *
 * Returns: errno
 */

1176
int gfs2_write_alloc_required(struct gfs2_inode *ip, u64 offset,
D
David Teigland 已提交
1177 1178
			      unsigned int len, int *alloc_required)
{
1179
	struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
1180 1181
	u64 lblock, lblock_stop, dblock;
	u32 extlen;
D
David Teigland 已提交
1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196
	int new = 0;
	int error = 0;

	*alloc_required = 0;

	if (!len)
		return 0;

	if (gfs2_is_stuffed(ip)) {
		if (offset + len >
		    sdp->sd_sb.sb_bsize - sizeof(struct gfs2_dinode))
			*alloc_required = 1;
		return 0;
	}

1197
	if (gfs2_is_dir(ip)) {
D
David Teigland 已提交
1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209
		unsigned int bsize = sdp->sd_jbsize;
		lblock = offset;
		do_div(lblock, bsize);
		lblock_stop = offset + len + bsize - 1;
		do_div(lblock_stop, bsize);
	} else {
		unsigned int shift = sdp->sd_sb.sb_bsize_shift;
		lblock = offset >> shift;
		lblock_stop = (offset + len + sdp->sd_sb.sb_bsize - 1) >> shift;
	}

	for (; lblock < lblock_stop; lblock += extlen) {
1210
		error = gfs2_extent_map(&ip->i_inode, lblock, &new, &dblock, &extlen);
D
David Teigland 已提交
1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222
		if (error)
			return error;

		if (!dblock) {
			*alloc_required = 1;
			return 0;
		}
	}

	return 0;
}