file.c 14.2 KB
Newer Older
1 2 3 4
/*
 * Squashfs - a compressed read only filesystem for Linux
 *
 * Copyright (c) 2002, 2003, 2004, 2005, 2006, 2007, 2008
5
 * Phillip Lougher <phillip@squashfs.org.uk>
6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 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
 *
 * This program is free software; you can redistribute it and/or
 * modify it under the terms of the GNU General Public License
 * as published by the Free Software Foundation; either version 2,
 * or (at your option) any later version.
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program; if not, write to the Free Software
 * Foundation, 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
 *
 * file.c
 */

/*
 * This file contains code for handling regular files.  A regular file
 * consists of a sequence of contiguous compressed blocks, and/or a
 * compressed fragment block (tail-end packed block).   The compressed size
 * of each datablock is stored in a block list contained within the
 * file inode (itself stored in one or more compressed metadata blocks).
 *
 * To speed up access to datablocks when reading 'large' files (256 Mbytes or
 * larger), the code implements an index cache that caches the mapping from
 * block index to datablock location on disk.
 *
 * The index cache allows Squashfs to handle large files (up to 1.75 TiB) while
 * retaining a simple and space-efficient block list on disk.  The cache
 * is split into slots, caching up to eight 224 GiB files (128 KiB blocks).
 * Larger files use multiple slots, with 1.75 TiB files using all 8 slots.
 * The index cache is designed to be memory efficient, and by default uses
 * 16 KiB.
 */

#include <linux/fs.h>
#include <linux/vfs.h>
#include <linux/kernel.h>
#include <linux/slab.h>
#include <linux/string.h>
#include <linux/pagemap.h>
#include <linux/mutex.h>

#include "squashfs_fs.h"
#include "squashfs_fs_sb.h"
#include "squashfs_fs_i.h"
#include "squashfs.h"

/*
 * Locate cache slot in range [offset, index] for specified inode.  If
 * there's more than one return the slot closest to index.
 */
static struct meta_index *locate_meta_index(struct inode *inode, int offset,
				int index)
{
	struct meta_index *meta = NULL;
	struct squashfs_sb_info *msblk = inode->i_sb->s_fs_info;
	int i;

	mutex_lock(&msblk->meta_index_mutex);

	TRACE("locate_meta_index: index %d, offset %d\n", index, offset);

	if (msblk->meta_index == NULL)
		goto not_allocated;

	for (i = 0; i < SQUASHFS_META_SLOTS; i++) {
		if (msblk->meta_index[i].inode_number == inode->i_ino &&
				msblk->meta_index[i].offset >= offset &&
				msblk->meta_index[i].offset <= index &&
				msblk->meta_index[i].locked == 0) {
			TRACE("locate_meta_index: entry %d, offset %d\n", i,
					msblk->meta_index[i].offset);
			meta = &msblk->meta_index[i];
			offset = meta->offset;
		}
	}

	if (meta)
		meta->locked = 1;

not_allocated:
	mutex_unlock(&msblk->meta_index_mutex);

	return meta;
}


/*
 * Find and initialise an empty cache slot for index offset.
 */
static struct meta_index *empty_meta_index(struct inode *inode, int offset,
				int skip)
{
	struct squashfs_sb_info *msblk = inode->i_sb->s_fs_info;
	struct meta_index *meta = NULL;
	int i;

	mutex_lock(&msblk->meta_index_mutex);

	TRACE("empty_meta_index: offset %d, skip %d\n", offset, skip);

	if (msblk->meta_index == NULL) {
		/*
		 * First time cache index has been used, allocate and
		 * initialise.  The cache index could be allocated at
		 * mount time but doing it here means it is allocated only
		 * if a 'large' file is read.
		 */
		msblk->meta_index = kcalloc(SQUASHFS_META_SLOTS,
			sizeof(*(msblk->meta_index)), GFP_KERNEL);
		if (msblk->meta_index == NULL) {
			ERROR("Failed to allocate meta_index\n");
			goto failed;
		}
		for (i = 0; i < SQUASHFS_META_SLOTS; i++) {
			msblk->meta_index[i].inode_number = 0;
			msblk->meta_index[i].locked = 0;
		}
		msblk->next_meta_index = 0;
	}

	for (i = SQUASHFS_META_SLOTS; i &&
			msblk->meta_index[msblk->next_meta_index].locked; i--)
		msblk->next_meta_index = (msblk->next_meta_index + 1) %
			SQUASHFS_META_SLOTS;

	if (i == 0) {
		TRACE("empty_meta_index: failed!\n");
		goto failed;
	}

	TRACE("empty_meta_index: returned meta entry %d, %p\n",
			msblk->next_meta_index,
			&msblk->meta_index[msblk->next_meta_index]);

	meta = &msblk->meta_index[msblk->next_meta_index];
	msblk->next_meta_index = (msblk->next_meta_index + 1) %
			SQUASHFS_META_SLOTS;

	meta->inode_number = inode->i_ino;
	meta->offset = offset;
	meta->skip = skip;
	meta->entries = 0;
	meta->locked = 1;

failed:
	mutex_unlock(&msblk->meta_index_mutex);
	return meta;
}


static void release_meta_index(struct inode *inode, struct meta_index *meta)
{
	struct squashfs_sb_info *msblk = inode->i_sb->s_fs_info;
	mutex_lock(&msblk->meta_index_mutex);
	meta->locked = 0;
	mutex_unlock(&msblk->meta_index_mutex);
}


/*
 * Read the next n blocks from the block list, starting from
 * metadata block <start_block, offset>.
 */
static long long read_indexes(struct super_block *sb, int n,
				u64 *start_block, int *offset)
{
	int err, i;
	long long block = 0;
178
	__le32 *blist = kmalloc(PAGE_SIZE, GFP_KERNEL);
179 180 181 182 183 184 185

	if (blist == NULL) {
		ERROR("read_indexes: Failed to allocate block_list\n");
		return -ENOMEM;
	}

	while (n) {
186
		int blocks = min_t(int, n, PAGE_SIZE >> 2);
187 188 189 190 191 192 193 194 195 196

		err = squashfs_read_metadata(sb, blist, start_block,
				offset, blocks << 2);
		if (err < 0) {
			ERROR("read_indexes: reading block [%llx:%x]\n",
				*start_block, *offset);
			goto failure;
		}

		for (i = 0; i < blocks; i++) {
197 198 199 200 201
			int size = squashfs_block_size(blist[i]);
			if (size < 0) {
				err = size;
				goto failure;
			}
202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373
			block += SQUASHFS_COMPRESSED_SIZE_BLOCK(size);
		}
		n -= blocks;
	}

	kfree(blist);
	return block;

failure:
	kfree(blist);
	return err;
}


/*
 * Each cache index slot has SQUASHFS_META_ENTRIES, each of which
 * can cache one index -> datablock/blocklist-block mapping.  We wish
 * to distribute these over the length of the file, entry[0] maps index x,
 * entry[1] maps index x + skip, entry[2] maps index x + 2 * skip, and so on.
 * The larger the file, the greater the skip factor.  The skip factor is
 * limited to the size of the metadata cache (SQUASHFS_CACHED_BLKS) to ensure
 * the number of metadata blocks that need to be read fits into the cache.
 * If the skip factor is limited in this way then the file will use multiple
 * slots.
 */
static inline int calculate_skip(int blocks)
{
	int skip = blocks / ((SQUASHFS_META_ENTRIES + 1)
		 * SQUASHFS_META_INDEXES);
	return min(SQUASHFS_CACHED_BLKS - 1, skip + 1);
}


/*
 * Search and grow the index cache for the specified inode, returning the
 * on-disk locations of the datablock and block list metadata block
 * <index_block, index_offset> for index (scaled to nearest cache index).
 */
static int fill_meta_index(struct inode *inode, int index,
		u64 *index_block, int *index_offset, u64 *data_block)
{
	struct squashfs_sb_info *msblk = inode->i_sb->s_fs_info;
	int skip = calculate_skip(i_size_read(inode) >> msblk->block_log);
	int offset = 0;
	struct meta_index *meta;
	struct meta_entry *meta_entry;
	u64 cur_index_block = squashfs_i(inode)->block_list_start;
	int cur_offset = squashfs_i(inode)->offset;
	u64 cur_data_block = squashfs_i(inode)->start;
	int err, i;

	/*
	 * Scale index to cache index (cache slot entry)
	 */
	index /= SQUASHFS_META_INDEXES * skip;

	while (offset < index) {
		meta = locate_meta_index(inode, offset + 1, index);

		if (meta == NULL) {
			meta = empty_meta_index(inode, offset + 1, skip);
			if (meta == NULL)
				goto all_done;
		} else {
			offset = index < meta->offset + meta->entries ? index :
				meta->offset + meta->entries - 1;
			meta_entry = &meta->meta_entry[offset - meta->offset];
			cur_index_block = meta_entry->index_block +
				msblk->inode_table;
			cur_offset = meta_entry->offset;
			cur_data_block = meta_entry->data_block;
			TRACE("get_meta_index: offset %d, meta->offset %d, "
				"meta->entries %d\n", offset, meta->offset,
				meta->entries);
			TRACE("get_meta_index: index_block 0x%llx, offset 0x%x"
				" data_block 0x%llx\n", cur_index_block,
				cur_offset, cur_data_block);
		}

		/*
		 * If necessary grow cache slot by reading block list.  Cache
		 * slot is extended up to index or to the end of the slot, in
		 * which case further slots will be used.
		 */
		for (i = meta->offset + meta->entries; i <= index &&
				i < meta->offset + SQUASHFS_META_ENTRIES; i++) {
			int blocks = skip * SQUASHFS_META_INDEXES;
			long long res = read_indexes(inode->i_sb, blocks,
					&cur_index_block, &cur_offset);

			if (res < 0) {
				if (meta->entries == 0)
					/*
					 * Don't leave an empty slot on read
					 * error allocated to this inode...
					 */
					meta->inode_number = 0;
				err = res;
				goto failed;
			}

			cur_data_block += res;
			meta_entry = &meta->meta_entry[i - meta->offset];
			meta_entry->index_block = cur_index_block -
				msblk->inode_table;
			meta_entry->offset = cur_offset;
			meta_entry->data_block = cur_data_block;
			meta->entries++;
			offset++;
		}

		TRACE("get_meta_index: meta->offset %d, meta->entries %d\n",
				meta->offset, meta->entries);

		release_meta_index(inode, meta);
	}

all_done:
	*index_block = cur_index_block;
	*index_offset = cur_offset;
	*data_block = cur_data_block;

	/*
	 * Scale cache index (cache slot entry) to index
	 */
	return offset * SQUASHFS_META_INDEXES * skip;

failed:
	release_meta_index(inode, meta);
	return err;
}


/*
 * Get the on-disk location and compressed size of the datablock
 * specified by index.  Fill_meta_index() does most of the work.
 */
static int read_blocklist(struct inode *inode, int index, u64 *block)
{
	u64 start;
	long long blks;
	int offset;
	__le32 size;
	int res = fill_meta_index(inode, index, &start, &offset, block);

	TRACE("read_blocklist: res %d, index %d, start 0x%llx, offset"
		       " 0x%x, block 0x%llx\n", res, index, start, offset,
			*block);

	if (res < 0)
		return res;

	/*
	 * res contains the index of the mapping returned by fill_meta_index(),
	 * this will likely be less than the desired index (because the
	 * meta_index cache works at a higher granularity).  Read any
	 * extra block indexes needed.
	 */
	if (res < index) {
		blks = read_indexes(inode->i_sb, index - res, &start, &offset);
		if (blks < 0)
			return (int) blks;
		*block += blks;
	}

	/*
	 * Read length of block specified by index.
	 */
	res = squashfs_read_metadata(inode->i_sb, &size, &start, &offset,
			sizeof(size));
	if (res < 0)
		return res;
374
	return squashfs_block_size(size);
375 376
}

377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393
void squashfs_fill_page(struct page *page, struct squashfs_cache_entry *buffer, int offset, int avail)
{
	int copied;
	void *pageaddr;

	pageaddr = kmap_atomic(page);
	copied = squashfs_copy_data(pageaddr, buffer, offset, avail);
	memset(pageaddr + copied, 0, PAGE_SIZE - copied);
	kunmap_atomic(pageaddr);

	flush_dcache_page(page);
	if (copied == avail)
		SetPageUptodate(page);
	else
		SetPageError(page);
}

394 395 396
/* Copy data into page cache  */
void squashfs_copy_cache(struct page *page, struct squashfs_cache_entry *buffer,
	int bytes, int offset)
397 398 399
{
	struct inode *inode = page->mapping->host;
	struct squashfs_sb_info *msblk = inode->i_sb->s_fs_info;
400
	int i, mask = (1 << (msblk->block_log - PAGE_SHIFT)) - 1;
401
	int start_index = page->index & ~mask, end_index = start_index | mask;
402 403 404

	/*
	 * Loop copying datablock into pages.  As the datablock likely covers
405
	 * many PAGE_SIZE pages (default block size is 128 KiB) explicitly
406 407 408 409
	 * grab the pages from the page cache, except for the page that we've
	 * been called to fill.
	 */
	for (i = start_index; i <= end_index && bytes > 0; i++,
410
			bytes -= PAGE_SIZE, offset += PAGE_SIZE) {
411
		struct page *push_page;
412
		int avail = buffer ? min_t(int, bytes, PAGE_SIZE) : 0;
413 414 415 416 417 418 419 420 421 422 423 424

		TRACE("bytes %d, i %d, available_bytes %d\n", bytes, i, avail);

		push_page = (i == page->index) ? page :
			grab_cache_page_nowait(page->mapping, i);

		if (!push_page)
			continue;

		if (PageUptodate(push_page))
			goto skip_page;

425
		squashfs_fill_page(push_page, buffer, offset, avail);
426 427 428
skip_page:
		unlock_page(push_page);
		if (i != page->index)
429
			put_page(push_page);
430
	}
431 432 433
}

/* Read datablock stored packed inside a fragment (tail-end packed block) */
434
static int squashfs_readpage_fragment(struct page *page, int expected)
435 436 437 438 439 440 441 442 443 444 445 446
{
	struct inode *inode = page->mapping->host;
	struct squashfs_cache_entry *buffer = squashfs_get_fragment(inode->i_sb,
		squashfs_i(inode)->fragment_block,
		squashfs_i(inode)->fragment_size);
	int res = buffer->error;

	if (res)
		ERROR("Unable to read page, block %llx, size %x\n",
			squashfs_i(inode)->fragment_block,
			squashfs_i(inode)->fragment_size);
	else
447
		squashfs_copy_cache(page, buffer, expected,
448 449 450 451 452
			squashfs_i(inode)->fragment_offset);

	squashfs_cache_put(buffer);
	return res;
}
453

454
static int squashfs_readpage_sparse(struct page *page, int expected)
455
{
456
	squashfs_copy_cache(page, NULL, expected, 0);
457
	return 0;
458 459 460 461 462 463
}

static int squashfs_readpage(struct file *file, struct page *page)
{
	struct inode *inode = page->mapping->host;
	struct squashfs_sb_info *msblk = inode->i_sb->s_fs_info;
464
	int index = page->index >> (msblk->block_log - PAGE_SHIFT);
465
	int file_end = i_size_read(inode) >> msblk->block_log;
466 467 468
	int expected = index == file_end ?
			(i_size_read(inode) & (msblk->block_size - 1)) :
			 msblk->block_size;
469 470 471 472 473 474
	int res;
	void *pageaddr;

	TRACE("Entered squashfs_readpage, page index %lx, start block %llx\n",
				page->index, squashfs_i(inode)->start);

475 476
	if (page->index >= ((i_size_read(inode) + PAGE_SIZE - 1) >>
					PAGE_SHIFT))
477 478 479 480 481 482 483 484 485 486
		goto out;

	if (index < file_end || squashfs_i(inode)->fragment_block ==
					SQUASHFS_INVALID_BLK) {
		u64 block = 0;
		int bsize = read_blocklist(inode, index, &block);
		if (bsize < 0)
			goto error_out;

		if (bsize == 0)
487
			res = squashfs_readpage_sparse(page, expected);
488
		else
489
			res = squashfs_readpage_block(page, block, bsize, expected);
490
	} else
491
		res = squashfs_readpage_fragment(page, expected);
492 493 494

	if (!res)
		return 0;
495 496 497 498

error_out:
	SetPageError(page);
out:
499
	pageaddr = kmap_atomic(page);
500
	memset(pageaddr, 0, PAGE_SIZE);
501
	kunmap_atomic(pageaddr);
502 503 504 505 506 507 508 509 510 511 512 513
	flush_dcache_page(page);
	if (!PageError(page))
		SetPageUptodate(page);
	unlock_page(page);

	return 0;
}


const struct address_space_operations squashfs_aops = {
	.readpage = squashfs_readpage
};