1// SPDX-License-Identifier: GPL-2.0-or-later
2/*
3 * Squashfs - a compressed read only filesystem for Linux
4 *
5 * Copyright (c) 2002, 2003, 2004, 2005, 2006, 2007, 2008
6 * Phillip Lougher <phillip@squashfs.org.uk>
7 *
8 * cache.c
9 */
10
11/*
12 * Blocks in Squashfs are compressed. To avoid repeatedly decompressing
13 * recently accessed data Squashfs uses two small metadata and fragment caches.
14 *
15 * This file implements a generic cache implementation used for both caches,
16 * plus functions layered ontop of the generic cache implementation to
17 * access the metadata and fragment caches.
18 *
19 * To avoid out of memory and fragmentation issues with vmalloc the cache
20 * uses sequences of kmalloced PAGE_SIZE buffers.
21 *
22 * It should be noted that the cache is not used for file datablocks, these
23 * are decompressed and cached in the page-cache in the normal way. The
24 * cache is only used to temporarily cache fragment and metadata blocks
25 * which have been read as as a result of a metadata (i.e. inode or
26 * directory) or fragment access. Because metadata and fragments are packed
27 * together into blocks (to gain greater compression) the read of a particular
28 * piece of metadata or fragment will retrieve other metadata/fragments which
29 * have been packed with it, these because of locality-of-reference may be read
30 * in the near future. Temporarily caching them ensures they are available for
31 * near future access without requiring an additional read and decompress.
32 */
33
34#include <linux/fs.h>
35#include <linux/vfs.h>
36#include <linux/slab.h>
37#include <linux/vmalloc.h>
38#include <linux/sched.h>
39#include <linux/spinlock.h>
40#include <linux/wait.h>
41#include <linux/pagemap.h>
42
43#include "squashfs_fs.h"
44#include "squashfs_fs_sb.h"
45#include "squashfs.h"
46#include "page_actor.h"
47
48/*
49 * Look-up block in cache, and increment usage count. If not in cache, read
50 * and decompress it from disk.
51 */
52struct squashfs_cache_entry *squashfs_cache_get(struct super_block *sb,
53 struct squashfs_cache *cache, u64 block, int length)
54{
55 int i, n;
56 struct squashfs_cache_entry *entry;
57
58 spin_lock(lock: &cache->lock);
59
60 while (1) {
61 for (i = cache->curr_blk, n = 0; n < cache->entries; n++) {
62 if (cache->entry[i].block == block) {
63 cache->curr_blk = i;
64 break;
65 }
66 i = (i + 1) % cache->entries;
67 }
68
69 if (n == cache->entries) {
70 /*
71 * Block not in cache, if all cache entries are used
72 * go to sleep waiting for one to become available.
73 */
74 if (cache->unused == 0) {
75 cache->num_waiters++;
76 spin_unlock(lock: &cache->lock);
77 wait_event(cache->wait_queue, cache->unused);
78 spin_lock(lock: &cache->lock);
79 cache->num_waiters--;
80 continue;
81 }
82
83 /*
84 * At least one unused cache entry. A simple
85 * round-robin strategy is used to choose the entry to
86 * be evicted from the cache.
87 */
88 i = cache->next_blk;
89 for (n = 0; n < cache->entries; n++) {
90 if (cache->entry[i].refcount == 0)
91 break;
92 i = (i + 1) % cache->entries;
93 }
94
95 cache->next_blk = (i + 1) % cache->entries;
96 entry = &cache->entry[i];
97
98 /*
99 * Initialise chosen cache entry, and fill it in from
100 * disk.
101 */
102 cache->unused--;
103 entry->block = block;
104 entry->refcount = 1;
105 entry->pending = 1;
106 entry->num_waiters = 0;
107 entry->error = 0;
108 spin_unlock(lock: &cache->lock);
109
110 entry->length = squashfs_read_data(sb, block, length,
111 &entry->next_index, entry->actor);
112
113 spin_lock(lock: &cache->lock);
114
115 if (entry->length < 0)
116 entry->error = entry->length;
117
118 entry->pending = 0;
119
120 /*
121 * While filling this entry one or more other processes
122 * have looked it up in the cache, and have slept
123 * waiting for it to become available.
124 */
125 if (entry->num_waiters) {
126 spin_unlock(lock: &cache->lock);
127 wake_up_all(&entry->wait_queue);
128 } else
129 spin_unlock(lock: &cache->lock);
130
131 goto out;
132 }
133
134 /*
135 * Block already in cache. Increment refcount so it doesn't
136 * get reused until we're finished with it, if it was
137 * previously unused there's one less cache entry available
138 * for reuse.
139 */
140 entry = &cache->entry[i];
141 if (entry->refcount == 0)
142 cache->unused--;
143 entry->refcount++;
144
145 /*
146 * If the entry is currently being filled in by another process
147 * go to sleep waiting for it to become available.
148 */
149 if (entry->pending) {
150 entry->num_waiters++;
151 spin_unlock(lock: &cache->lock);
152 wait_event(entry->wait_queue, !entry->pending);
153 } else
154 spin_unlock(lock: &cache->lock);
155
156 goto out;
157 }
158
159out:
160 TRACE("Got %s %d, start block %lld, refcount %d, error %d\n",
161 cache->name, i, entry->block, entry->refcount, entry->error);
162
163 if (entry->error)
164 ERROR("Unable to read %s cache entry [%llx]\n", cache->name,
165 block);
166 return entry;
167}
168
169
170/*
171 * Release cache entry, once usage count is zero it can be reused.
172 */
173void squashfs_cache_put(struct squashfs_cache_entry *entry)
174{
175 struct squashfs_cache *cache = entry->cache;
176
177 spin_lock(lock: &cache->lock);
178 entry->refcount--;
179 if (entry->refcount == 0) {
180 cache->unused++;
181 /*
182 * If there's any processes waiting for a block to become
183 * available, wake one up.
184 */
185 if (cache->num_waiters) {
186 spin_unlock(lock: &cache->lock);
187 wake_up(&cache->wait_queue);
188 return;
189 }
190 }
191 spin_unlock(lock: &cache->lock);
192}
193
194/*
195 * Delete cache reclaiming all kmalloced buffers.
196 */
197void squashfs_cache_delete(struct squashfs_cache *cache)
198{
199 int i, j;
200
201 if (cache == NULL)
202 return;
203
204 for (i = 0; i < cache->entries; i++) {
205 if (cache->entry[i].data) {
206 for (j = 0; j < cache->pages; j++)
207 kfree(objp: cache->entry[i].data[j]);
208 kfree(objp: cache->entry[i].data);
209 }
210 kfree(objp: cache->entry[i].actor);
211 }
212
213 kfree(objp: cache->entry);
214 kfree(objp: cache);
215}
216
217
218/*
219 * Initialise cache allocating the specified number of entries, each of
220 * size block_size. To avoid vmalloc fragmentation issues each entry
221 * is allocated as a sequence of kmalloced PAGE_SIZE buffers.
222 */
223struct squashfs_cache *squashfs_cache_init(char *name, int entries,
224 int block_size)
225{
226 int i, j;
227 struct squashfs_cache *cache = kzalloc(size: sizeof(*cache), GFP_KERNEL);
228
229 if (cache == NULL) {
230 ERROR("Failed to allocate %s cache\n", name);
231 return NULL;
232 }
233
234 cache->entry = kcalloc(n: entries, size: sizeof(*(cache->entry)), GFP_KERNEL);
235 if (cache->entry == NULL) {
236 ERROR("Failed to allocate %s cache\n", name);
237 goto cleanup;
238 }
239
240 cache->curr_blk = 0;
241 cache->next_blk = 0;
242 cache->unused = entries;
243 cache->entries = entries;
244 cache->block_size = block_size;
245 cache->pages = block_size >> PAGE_SHIFT;
246 cache->pages = cache->pages ? cache->pages : 1;
247 cache->name = name;
248 cache->num_waiters = 0;
249 spin_lock_init(&cache->lock);
250 init_waitqueue_head(&cache->wait_queue);
251
252 for (i = 0; i < entries; i++) {
253 struct squashfs_cache_entry *entry = &cache->entry[i];
254
255 init_waitqueue_head(&cache->entry[i].wait_queue);
256 entry->cache = cache;
257 entry->block = SQUASHFS_INVALID_BLK;
258 entry->data = kcalloc(n: cache->pages, size: sizeof(void *), GFP_KERNEL);
259 if (entry->data == NULL) {
260 ERROR("Failed to allocate %s cache entry\n", name);
261 goto cleanup;
262 }
263
264 for (j = 0; j < cache->pages; j++) {
265 entry->data[j] = kmalloc(PAGE_SIZE, GFP_KERNEL);
266 if (entry->data[j] == NULL) {
267 ERROR("Failed to allocate %s buffer\n", name);
268 goto cleanup;
269 }
270 }
271
272 entry->actor = squashfs_page_actor_init(buffer: entry->data,
273 pages: cache->pages, length: 0);
274 if (entry->actor == NULL) {
275 ERROR("Failed to allocate %s cache entry\n", name);
276 goto cleanup;
277 }
278 }
279
280 return cache;
281
282cleanup:
283 squashfs_cache_delete(cache);
284 return NULL;
285}
286
287
288/*
289 * Copy up to length bytes from cache entry to buffer starting at offset bytes
290 * into the cache entry. If there's not length bytes then copy the number of
291 * bytes available. In all cases return the number of bytes copied.
292 */
293int squashfs_copy_data(void *buffer, struct squashfs_cache_entry *entry,
294 int offset, int length)
295{
296 int remaining = length;
297
298 if (length == 0)
299 return 0;
300 else if (buffer == NULL)
301 return min(length, entry->length - offset);
302
303 while (offset < entry->length) {
304 void *buff = entry->data[offset / PAGE_SIZE]
305 + (offset % PAGE_SIZE);
306 int bytes = min_t(int, entry->length - offset,
307 PAGE_SIZE - (offset % PAGE_SIZE));
308
309 if (bytes >= remaining) {
310 memcpy(buffer, buff, remaining);
311 remaining = 0;
312 break;
313 }
314
315 memcpy(buffer, buff, bytes);
316 buffer += bytes;
317 remaining -= bytes;
318 offset += bytes;
319 }
320
321 return length - remaining;
322}
323
324
325/*
326 * Read length bytes from metadata position <block, offset> (block is the
327 * start of the compressed block on disk, and offset is the offset into
328 * the block once decompressed). Data is packed into consecutive blocks,
329 * and length bytes may require reading more than one block.
330 */
331int squashfs_read_metadata(struct super_block *sb, void *buffer,
332 u64 *block, int *offset, int length)
333{
334 struct squashfs_sb_info *msblk = sb->s_fs_info;
335 int bytes, res = length;
336 struct squashfs_cache_entry *entry;
337
338 TRACE("Entered squashfs_read_metadata [%llx:%x]\n", *block, *offset);
339
340 if (unlikely(length < 0))
341 return -EIO;
342
343 while (length) {
344 entry = squashfs_cache_get(sb, cache: msblk->block_cache, block: *block, length: 0);
345 if (entry->error) {
346 res = entry->error;
347 goto error;
348 } else if (*offset >= entry->length) {
349 res = -EIO;
350 goto error;
351 }
352
353 bytes = squashfs_copy_data(buffer, entry, offset: *offset, length);
354 if (buffer)
355 buffer += bytes;
356 length -= bytes;
357 *offset += bytes;
358
359 if (*offset == entry->length) {
360 *block = entry->next_index;
361 *offset = 0;
362 }
363
364 squashfs_cache_put(entry);
365 }
366
367 return res;
368
369error:
370 squashfs_cache_put(entry);
371 return res;
372}
373
374
375/*
376 * Look-up in the fragmment cache the fragment located at <start_block> in the
377 * filesystem. If necessary read and decompress it from disk.
378 */
379struct squashfs_cache_entry *squashfs_get_fragment(struct super_block *sb,
380 u64 start_block, int length)
381{
382 struct squashfs_sb_info *msblk = sb->s_fs_info;
383
384 return squashfs_cache_get(sb, cache: msblk->fragment_cache, block: start_block,
385 length);
386}
387
388
389/*
390 * Read and decompress the datablock located at <start_block> in the
391 * filesystem. The cache is used here to avoid duplicating locking and
392 * read/decompress code.
393 */
394struct squashfs_cache_entry *squashfs_get_datablock(struct super_block *sb,
395 u64 start_block, int length)
396{
397 struct squashfs_sb_info *msblk = sb->s_fs_info;
398
399 return squashfs_cache_get(sb, cache: msblk->read_page, block: start_block, length);
400}
401
402
403/*
404 * Read a filesystem table (uncompressed sequence of bytes) from disk
405 */
406void *squashfs_read_table(struct super_block *sb, u64 block, int length)
407{
408 int pages = (length + PAGE_SIZE - 1) >> PAGE_SHIFT;
409 int i, res;
410 void *table, *buffer, **data;
411 struct squashfs_page_actor *actor;
412
413 table = buffer = kmalloc(size: length, GFP_KERNEL);
414 if (table == NULL)
415 return ERR_PTR(error: -ENOMEM);
416
417 data = kcalloc(n: pages, size: sizeof(void *), GFP_KERNEL);
418 if (data == NULL) {
419 res = -ENOMEM;
420 goto failed;
421 }
422
423 actor = squashfs_page_actor_init(buffer: data, pages, length);
424 if (actor == NULL) {
425 res = -ENOMEM;
426 goto failed2;
427 }
428
429 for (i = 0; i < pages; i++, buffer += PAGE_SIZE)
430 data[i] = buffer;
431
432 res = squashfs_read_data(sb, block, length |
433 SQUASHFS_COMPRESSED_BIT_BLOCK, NULL, actor);
434
435 kfree(objp: data);
436 kfree(objp: actor);
437
438 if (res < 0)
439 goto failed;
440
441 return table;
442
443failed2:
444 kfree(objp: data);
445failed:
446 kfree(objp: table);
447 return ERR_PTR(error: res);
448}
449

source code of linux/fs/squashfs/cache.c