summaryrefslogtreecommitdiff
path: root/firmware/common/disk_cache.c
diff options
context:
space:
mode:
authorMichael Sevakis <jethead71@rockbox.org>2013-08-05 22:02:45 -0400
committerMichael Sevakis <jethead71@rockbox.org>2014-08-30 03:48:23 +0200
commit7d1a47cf13726c95ac46027156cc12dd9da5b855 (patch)
treeeb20d07656806479a8e1fea25887a490ea30d1d8 /firmware/common/disk_cache.c
parent95a4c3afcd53a1f8b835dec33de51f9c304de4d9 (diff)
downloadrockbox-7d1a47cf13726c95ac46027156cc12dd9da5b855.tar.gz
rockbox-7d1a47cf13726c95ac46027156cc12dd9da5b855.zip
Rewrite filesystem code (WIP)
This patch redoes the filesystem code from the FAT driver up to the clipboard code in onplay.c. Not every aspect of this is finished therefore it is still "WIP". I don't wish to do too much at once (haha!). What is left to do is get dircache back in the sim and find an implementation for the dircache indicies in the tagcache and playlist code or do something else that has the same benefit. Leaving these out for now does not make anything unusable. All the basics are done. Phone app code should probably get vetted (and app path handling just plain rewritten as environment expansions); the SDL app and Android run well. Main things addressed: 1) Thread safety: There is none right now in the trunk code. Most of what currently works is luck when multiple threads are involved or multiple descriptors to the same file are open. 2) POSIX compliance: Many of the functions behave nothing like their counterparts on a host system. This leads to inconsistent code or very different behavior from native to hosted. One huge offender was rename(). Going point by point would fill a book. 3) Actual running RAM usage: Many targets will use less RAM and less stack space (some more RAM because I upped the number of cache buffers for large memory). There's very little memory lying fallow in rarely-used areas (see 'Key core changes' below). Also, all targets may open the same number of directory streams whereas before those with less than 8MB RAM were limited to 8, not 12 implying those targets will save slightly less. 4) Performance: The test_disk plugin shows markedly improved performance, particularly in the area of (uncached) directory scanning, due partly to more optimal directory reading and to a better sector cache algorithm. Uncached times tend to be better while there is a bit of a slowdown in dircache due to it being a bit heavier of an implementation. It's not noticeable by a human as far as I can say. Key core changes: 1) Files and directories share core code and data structures. 2) The filesystem code knows which descriptors refer to same file. This ensures that changes from one stream are appropriately reflected in every open descriptor for that file (fileobj_mgr.c). 3) File and directory cache buffers are borrowed from the main sector cache. This means that when they are not in use by a file, they are not wasted, but used for the cache. Most of the time, only a few of them are needed. It also means that adding more file and directory handles is less expensive. All one must do in ensure a large enough cache to borrow from. 4) Relative path components are supported and the namespace is unified. It does not support full relative paths to an implied current directory; what is does support is use of "." and "..". Adding the former would not be very difficult. The namespace is unified in the sense that volumes may be specified several times along with relative parts, e.g.: "/<0>/foo/../../<1>/bar" :<=> "/<1>/bar". 5) Stack usage is down due to sharing of data, static allocation and less duplication of strings on the stack. This requires more serialization than I would like but since the number of threads is limited to a low number, the tradoff in favor of the stack seems reasonable. 6) Separates and heirarchicalizes (sic) the SIM and APP filesystem code. SIM path and volume handling is just like the target. Some aspects of the APP file code get more straightforward (e.g. no path hashing is needed). Dircache: Deserves its own section. Dircache is new but pays homage to the old. The old one was not compatible and so it, since it got redone, does all the stuff it always should have done such as: 1) It may be update and used at any time during the build process. No longer has one to wait for it to finish building to do basic file management (create, remove, rename, etc.). 2) It does not need to be either fully scanned or completely disabled; it can be incomplete (i.e. overfilled, missing paths), still be of benefit and be correct. 3) Handles mounting and dismounting of individual volumes which means a full rebuild is not needed just because you pop a new SD card in the slot. Now, because it reuses its freed entry data, may rebuild only that volume. 4) Much more fundamental to the file code. When it is built, it is the keeper of the master file list whether enabled or not ("disabled" is just a state of the cache). Its must always to ready to be started and bind all streams opened prior to being enabled. 5) Maintains any short filenames in OEM format which means that it does not need to be rebuilt when changing the default codepage. Miscellaneous Compatibility: 1) Update any other code that would otherwise not work such as the hotswap mounting code in various card drivers. 2) File management: Clipboard needed updating because of the behavioral changes. Still needs a little more work on some finer points. 3) Remove now-obsolete functionality such as the mutex's "no preempt" flag (which was only for the prior FAT driver). 4) struct dirinfo uses time_t rather than raw FAT directory entry time fields. I plan to follow up on genericizing everything there (i.e. no FAT attributes). 5) unicode.c needed some redoing so that the file code does not try try to load codepages during a scan, which is actually a problem with the current code. The default codepage, if any is required, is now kept in RAM separarately (bufalloced) from codepages specified to iso_decode() (which must not be bufalloced because the conversion may be done by playback threads). Brings with it some additional reusable core code: 1) Revised file functions: Reusable code that does things such as safe path concatenation and parsing without buffer limitations or data duplication. Variants that copy or alter the input path may be based off these. To do: 1) Put dircache functionality back in the sim. Treating it internally as a different kind of file system seems the best approach at this time. 2) Restore use of dircache indexes in the playlist and database or something effectively the same. Since the cache doesn't have to be complete in order to be used, not getting a hit on the cache doesn't unambiguously say if the path exists or not. Change-Id: Ia30f3082a136253e3a0eae0784e3091d138915c8 Reviewed-on: http://gerrit.rockbox.org/566 Reviewed-by: Michael Sevakis <jethead71@rockbox.org> Tested: Michael Sevakis <jethead71@rockbox.org>
Diffstat (limited to 'firmware/common/disk_cache.c')
-rw-r--r--firmware/common/disk_cache.c343
1 files changed, 343 insertions, 0 deletions
diff --git a/firmware/common/disk_cache.c b/firmware/common/disk_cache.c
new file mode 100644
index 0000000000..0e842e7796
--- /dev/null
+++ b/firmware/common/disk_cache.c
@@ -0,0 +1,343 @@
1/***************************************************************************
2 * __________ __ ___.
3 * Open \______ \ ____ ____ | | _\_ |__ _______ ___
4 * Source | _// _ \_/ ___\| |/ /| __ \ / _ \ \/ /
5 * Jukebox | | ( <_> ) \___| < | \_\ ( <_> > < <
6 * Firmware |____|_ /\____/ \___ >__|_ \|___ /\____/__/\_ \
7 * \/ \/ \/ \/ \/
8 * $Id$
9 *
10 * Copyright (C) 2014 by Michael Sevakis
11 *
12 * This program is free software; you can redistribute it and/or
13 * modify it under the terms of the GNU General Public License
14 * as published by the Free Software Foundation; either version 2
15 * of the License, or (at your option) any later version.
16 *
17 * This software is distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY
18 * KIND, either express or implied.
19 *
20 ****************************************************************************/
21#include "config.h"
22#include "debug.h"
23#include "system.h"
24#include "linked_list.h"
25#include "disk_cache.h"
26#include "fat.h" /* for SECTOR_SIZE */
27#include "bitarray.h"
28
29/* Cache: LRU cache with separately-chained hashtable
30 *
31 * Each entry of the map is the mapped location of the hashed sector value
32 * where each bit in each map entry indicates which corresponding cache
33 * entries are occupied by sector values that collide in that map entry.
34 *
35 * Each volume is given its own bit map.
36 *
37 * To probe for a specific key, each bit in the map entry must be examined,
38 * its position used as an index into the cache_entry array and the actual
39 * sector information compared for that cache entry. If the search exhausts
40 * all bits, the sector is not cached.
41 *
42 * To avoid long chains, the map entry count should be much greater than the
43 * number of cache entries. Since the cache is an LRU design, no buffer entry
44 * in the array is intrinsically associated with any particular sector number
45 * or volume.
46 *
47 * Example 6-sector cache with 8-entry map:
48 * cache entry 543210
49 * cache map 100000 <- sector number hashes into map
50 * 000000
51 * 000100
52 * 000000
53 * 010000
54 * 000000
55 * 001001 <- collision
56 * 000000
57 * volume map 111101 <- entry usage by the volume (OR of all map entries)
58 */
59
60enum dce_flags /* flags for each cache entry */
61{
62 DCE_INUSE = 0x01, /* entry in use and valid */
63 DCE_DIRTY = 0x02, /* entry is dirty in need of writeback */
64 DCE_BUF = 0x04, /* entry is being used as a general buffer */
65};
66
67struct disk_cache_entry
68{
69 struct lldc_node node; /* LRU list links */
70 unsigned char flags; /* entry flags */
71#ifdef HAVE_MULTIVOLUME
72 unsigned char volume; /* volume of sector */
73#endif
74 unsigned long sector; /* cached disk sector number */
75};
76
77BITARRAY_TYPE_DECLARE(cache_map_entry_t, cache_map, DC_NUM_ENTRIES)
78
79static inline unsigned int map_sector(unsigned long sector)
80{
81 /* keep sector hash simple for now */
82 return sector % DC_MAP_NUM_ENTRIES;
83}
84
85static struct lldc_head cache_lru; /* LRU cache list (head = LRU item) */
86static struct disk_cache_entry cache_entry[DC_NUM_ENTRIES];
87static cache_map_entry_t cache_map_entry[NUM_VOLUMES][DC_MAP_NUM_ENTRIES];
88static cache_map_entry_t cache_vol_map[NUM_VOLUMES] IBSS_ATTR;
89static uint8_t cache_buffer[DC_NUM_ENTRIES][DC_CACHE_BUFSIZE] CACHEALIGN_ATTR;
90struct mutex disk_cache_mutex SHAREDBSS_ATTR;
91
92#define CACHE_MAP_ENTRY(volume, mapnum) \
93 cache_map_entry[IF_MV_VOL(volume)][mapnum]
94#define CACHE_VOL_MAP(volume) \
95 cache_vol_map[IF_MV_VOL(volume)]
96
97#define DCE_LRU() ((struct disk_cache_entry *)cache_lru.head)
98#define DCE_NEXT(fce) ((struct disk_cache_entry *)(fce)->node.next)
99#define NODE_DCE(node) ((struct disk_cache_entry *)(node))
100
101/* get the cache index from a pointer to a buffer */
102#define DCIDX_FROM_BUF(buf) \
103 ((uint8_t (*)[DC_CACHE_BUFSIZE])(buf) - cache_buffer)
104
105#define DCIDX_FROM_DCE(dce) \
106 ((dce) - cache_entry)
107
108/* set the in-use bit in the map */
109static inline void cache_bitmap_set_bit(int volume, unsigned int mapnum,
110 unsigned int bitnum)
111{
112 cache_map_set_bit(&CACHE_MAP_ENTRY(volume, mapnum), bitnum);
113 cache_map_set_bit(&CACHE_VOL_MAP(volume), bitnum);
114 (void)volume;
115}
116
117/* clear the in-use bit in the map */
118static inline void cache_bitmap_clear_bit(int volume, unsigned int mapnum,
119 unsigned int bitnum)
120{
121 cache_map_clear_bit(&CACHE_MAP_ENTRY(volume, mapnum), bitnum);
122 cache_map_clear_bit(&CACHE_VOL_MAP(volume), bitnum);
123 (void)volume;
124}
125
126/* make entry MRU by moving it to the list tail */
127static inline void touch_cache_entry(struct disk_cache_entry *which)
128{
129 struct lldc_node *lru = cache_lru.head;
130 struct lldc_node *node = &which->node;
131
132 if (node == lru->prev) /* already MRU */
133 ; /**/
134 else if (node == lru) /* is the LRU? just rotate list */
135 cache_lru.head = lru->next;
136 else /* somewhere else; move it */
137 {
138 lldc_remove(&cache_lru, node);
139 lldc_insert_last(&cache_lru, node);
140 }
141}
142
143/* remove LRU entry from the cache list to use as a buffer */
144static struct disk_cache_entry * cache_remove_lru_entry(void)
145{
146 struct lldc_node *lru = cache_lru.head;
147
148 /* at least one is reserved for client */
149 if (lru == lru->next)
150 return NULL;
151
152 /* remove it; next-LRU becomes the LRU */
153 lldc_remove(&cache_lru, lru);
154 return NODE_DCE(lru);
155}
156
157/* return entry to the cache list and set it LRU */
158static void cache_return_lru_entry(struct disk_cache_entry *fce)
159{
160 lldc_insert_first(&cache_lru, &fce->node);
161}
162
163/* discard the entry's data and mark it unused */
164static inline void cache_discard_entry(struct disk_cache_entry *dce,
165 unsigned int index)
166{
167 cache_bitmap_clear_bit(IF_MV_VOL(dce->volume), map_sector(dce->sector),
168 index);
169 dce->flags = 0;
170}
171
172/* search the cache for the specified sector, returning a buffer, either
173 to the specified sector, if it exists, or a new/evicted entry that must
174 be filled */
175void * dc_cache_probe(IF_MV(int volume,) unsigned long sector,
176 unsigned int *flagsp)
177{
178 unsigned int mapnum = map_sector(sector);
179
180 FOR_EACH_BITARRAY_SET_BIT(&CACHE_MAP_ENTRY(volume, mapnum), index)
181 {
182 struct disk_cache_entry *dce = &cache_entry[index];
183
184 if (dce->sector == sector)
185 {
186 *flagsp = DCE_INUSE;
187 touch_cache_entry(dce);
188 return cache_buffer[index];
189 }
190 }
191
192 /* sector not found so the LRU is the victim */
193 struct disk_cache_entry *dce = DCE_LRU();
194 cache_lru.head = dce->node.next;
195
196 unsigned int index = DCIDX_FROM_DCE(dce);
197 void *buf = cache_buffer[index];
198 unsigned int old_flags = dce->flags;
199
200 if (old_flags)
201 {
202 int old_volume = IF_MV_VOL(dce->volume);
203 unsigned long sector = dce->sector;
204 unsigned int old_mapnum = map_sector(sector);
205
206 if (old_flags & DCE_DIRTY)
207 dc_writeback_callback(IF_MV(old_volume,) sector, buf);
208
209 if (mapnum == old_mapnum IF_MV( && volume == old_volume ))
210 goto finish_setup;
211
212 cache_bitmap_clear_bit(old_volume, old_mapnum, index);
213 }
214
215 cache_bitmap_set_bit(IF_MV_VOL(volume), mapnum, index);
216
217finish_setup:
218 dce->flags = DCE_INUSE;
219#ifdef HAVE_MULTIVOLUME
220 dce->volume = volume;
221#endif
222 dce->sector = sector;
223
224 *flagsp = 0;
225 return buf;
226}
227
228/* mark in-use cache entry as dirty by buffer */
229void dc_dirty_buf(void *buf)
230{
231 unsigned int index = DCIDX_FROM_BUF(buf);
232
233 if (index >= DC_NUM_ENTRIES)
234 return;
235
236 /* dirt remains, sticky until flushed */
237 struct disk_cache_entry *fce = &cache_entry[index];
238 if (fce->flags & DCE_INUSE)
239 fce->flags |= DCE_DIRTY;
240}
241
242/* discard in-use cache entry by buffer */
243void dc_discard_buf(void *buf)
244{
245 unsigned int index = DCIDX_FROM_BUF(buf);
246
247 if (index >= DC_NUM_ENTRIES)
248 return;
249
250 struct disk_cache_entry *dce = &cache_entry[index];
251 if (dce->flags & DCE_INUSE)
252 cache_discard_entry(dce, index);
253}
254
255/* commit all dirty cache entries to storage for a specified volume */
256void dc_commit_all(IF_MV_NONVOID(int volume))
257{
258 DEBUGF("dc_commit_all()\n");
259
260 FOR_EACH_BITARRAY_SET_BIT(&CACHE_VOL_MAP(volume), index)
261 {
262 struct disk_cache_entry *dce = &cache_entry[index];
263 unsigned int flags = dce->flags;
264
265 if (flags & DCE_DIRTY)
266 {
267 dc_writeback_callback(IF_MV(volume,) dce->sector,
268 cache_buffer[index]);
269 dce->flags = flags & ~DCE_DIRTY;
270 }
271 }
272}
273
274/* discard all cache entries from the specified volume */
275void dc_discard_all(IF_MV_NONVOID(int volume))
276{
277 DEBUGF("dc_discard_all()\n");
278
279 FOR_EACH_BITARRAY_SET_BIT(&CACHE_VOL_MAP(volume), index)
280 cache_discard_entry(&cache_entry[index], index);
281}
282
283/* expropriate a buffer from the cache */
284void * dc_get_buffer(void)
285{
286 dc_lock_cache();
287
288 void *buf = NULL;
289 struct disk_cache_entry *dce = cache_remove_lru_entry();
290
291 if (dce)
292 {
293 unsigned int index = DCIDX_FROM_DCE(dce);
294 unsigned int flags = dce->flags;
295
296 buf = cache_buffer[index];
297
298 if (flags)
299 {
300 /* must first commit this sector if dirty */
301 if (flags & DCE_DIRTY)
302 dc_writeback_callback(IF_MV(dce->volume,) dce->sector, buf);
303
304 cache_discard_entry(dce, index);
305 }
306
307 dce->flags = DCE_BUF;
308 }
309 /* cache is out of buffers */
310
311 dc_unlock_cache();
312 return buf;
313}
314
315/* return buffer to the cache by buffer */
316void dc_release_buffer(void *buf)
317{
318 unsigned int index = DCIDX_FROM_BUF(buf);
319
320 if (index >= DC_NUM_ENTRIES)
321 return;
322
323 dc_lock_cache();
324
325 struct disk_cache_entry *dce = &cache_entry[index];
326
327 if (dce->flags & DCE_BUF)
328 {
329 dce->flags = 0;
330 cache_return_lru_entry(dce);
331 }
332
333 dc_unlock_cache();
334}
335
336/* one-time init at startup */
337void dc_init(void)
338{
339 mutex_init(&disk_cache_mutex);
340 lldc_init(&cache_lru);
341 for (unsigned int i = 0; i < DC_NUM_ENTRIES; i++)
342 lldc_insert_last(&cache_lru, &cache_entry[i].node);
343}