xref: /qemu/block/qcow2-cache.c (revision 737e150e)
1 /*
2  * L2/refcount table cache for the QCOW2 format
3  *
4  * Copyright (c) 2010 Kevin Wolf <kwolf@redhat.com>
5  *
6  * Permission is hereby granted, free of charge, to any person obtaining a copy
7  * of this software and associated documentation files (the "Software"), to deal
8  * in the Software without restriction, including without limitation the rights
9  * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
10  * copies of the Software, and to permit persons to whom the Software is
11  * furnished to do so, subject to the following conditions:
12  *
13  * The above copyright notice and this permission notice shall be included in
14  * all copies or substantial portions of the Software.
15  *
16  * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
17  * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
18  * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
19  * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
20  * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
21  * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
22  * THE SOFTWARE.
23  */
24 
25 #include "block/block_int.h"
26 #include "qemu-common.h"
27 #include "qcow2.h"
28 #include "trace.h"
29 
30 typedef struct Qcow2CachedTable {
31     void*   table;
32     int64_t offset;
33     bool    dirty;
34     int     cache_hits;
35     int     ref;
36 } Qcow2CachedTable;
37 
38 struct Qcow2Cache {
39     Qcow2CachedTable*       entries;
40     struct Qcow2Cache*      depends;
41     int                     size;
42     bool                    depends_on_flush;
43 };
44 
45 Qcow2Cache *qcow2_cache_create(BlockDriverState *bs, int num_tables)
46 {
47     BDRVQcowState *s = bs->opaque;
48     Qcow2Cache *c;
49     int i;
50 
51     c = g_malloc0(sizeof(*c));
52     c->size = num_tables;
53     c->entries = g_malloc0(sizeof(*c->entries) * num_tables);
54 
55     for (i = 0; i < c->size; i++) {
56         c->entries[i].table = qemu_blockalign(bs, s->cluster_size);
57     }
58 
59     return c;
60 }
61 
62 int qcow2_cache_destroy(BlockDriverState* bs, Qcow2Cache *c)
63 {
64     int i;
65 
66     for (i = 0; i < c->size; i++) {
67         assert(c->entries[i].ref == 0);
68         qemu_vfree(c->entries[i].table);
69     }
70 
71     g_free(c->entries);
72     g_free(c);
73 
74     return 0;
75 }
76 
77 static int qcow2_cache_flush_dependency(BlockDriverState *bs, Qcow2Cache *c)
78 {
79     int ret;
80 
81     ret = qcow2_cache_flush(bs, c->depends);
82     if (ret < 0) {
83         return ret;
84     }
85 
86     c->depends = NULL;
87     c->depends_on_flush = false;
88 
89     return 0;
90 }
91 
92 static int qcow2_cache_entry_flush(BlockDriverState *bs, Qcow2Cache *c, int i)
93 {
94     BDRVQcowState *s = bs->opaque;
95     int ret = 0;
96 
97     if (!c->entries[i].dirty || !c->entries[i].offset) {
98         return 0;
99     }
100 
101     trace_qcow2_cache_entry_flush(qemu_coroutine_self(),
102                                   c == s->l2_table_cache, i);
103 
104     if (c->depends) {
105         ret = qcow2_cache_flush_dependency(bs, c);
106     } else if (c->depends_on_flush) {
107         ret = bdrv_flush(bs->file);
108         if (ret >= 0) {
109             c->depends_on_flush = false;
110         }
111     }
112 
113     if (ret < 0) {
114         return ret;
115     }
116 
117     if (c == s->refcount_block_cache) {
118         BLKDBG_EVENT(bs->file, BLKDBG_REFBLOCK_UPDATE_PART);
119     } else if (c == s->l2_table_cache) {
120         BLKDBG_EVENT(bs->file, BLKDBG_L2_UPDATE);
121     }
122 
123     ret = bdrv_pwrite(bs->file, c->entries[i].offset, c->entries[i].table,
124         s->cluster_size);
125     if (ret < 0) {
126         return ret;
127     }
128 
129     c->entries[i].dirty = false;
130 
131     return 0;
132 }
133 
134 int qcow2_cache_flush(BlockDriverState *bs, Qcow2Cache *c)
135 {
136     BDRVQcowState *s = bs->opaque;
137     int result = 0;
138     int ret;
139     int i;
140 
141     trace_qcow2_cache_flush(qemu_coroutine_self(), c == s->l2_table_cache);
142 
143     for (i = 0; i < c->size; i++) {
144         ret = qcow2_cache_entry_flush(bs, c, i);
145         if (ret < 0 && result != -ENOSPC) {
146             result = ret;
147         }
148     }
149 
150     if (result == 0) {
151         ret = bdrv_flush(bs->file);
152         if (ret < 0) {
153             result = ret;
154         }
155     }
156 
157     return result;
158 }
159 
160 int qcow2_cache_set_dependency(BlockDriverState *bs, Qcow2Cache *c,
161     Qcow2Cache *dependency)
162 {
163     int ret;
164 
165     if (dependency->depends) {
166         ret = qcow2_cache_flush_dependency(bs, dependency);
167         if (ret < 0) {
168             return ret;
169         }
170     }
171 
172     if (c->depends && (c->depends != dependency)) {
173         ret = qcow2_cache_flush_dependency(bs, c);
174         if (ret < 0) {
175             return ret;
176         }
177     }
178 
179     c->depends = dependency;
180     return 0;
181 }
182 
183 void qcow2_cache_depends_on_flush(Qcow2Cache *c)
184 {
185     c->depends_on_flush = true;
186 }
187 
188 static int qcow2_cache_find_entry_to_replace(Qcow2Cache *c)
189 {
190     int i;
191     int min_count = INT_MAX;
192     int min_index = -1;
193 
194 
195     for (i = 0; i < c->size; i++) {
196         if (c->entries[i].ref) {
197             continue;
198         }
199 
200         if (c->entries[i].cache_hits < min_count) {
201             min_index = i;
202             min_count = c->entries[i].cache_hits;
203         }
204 
205         /* Give newer hits priority */
206         /* TODO Check how to optimize the replacement strategy */
207         c->entries[i].cache_hits /= 2;
208     }
209 
210     if (min_index == -1) {
211         /* This can't happen in current synchronous code, but leave the check
212          * here as a reminder for whoever starts using AIO with the cache */
213         abort();
214     }
215     return min_index;
216 }
217 
218 static int qcow2_cache_do_get(BlockDriverState *bs, Qcow2Cache *c,
219     uint64_t offset, void **table, bool read_from_disk)
220 {
221     BDRVQcowState *s = bs->opaque;
222     int i;
223     int ret;
224 
225     trace_qcow2_cache_get(qemu_coroutine_self(), c == s->l2_table_cache,
226                           offset, read_from_disk);
227 
228     /* Check if the table is already cached */
229     for (i = 0; i < c->size; i++) {
230         if (c->entries[i].offset == offset) {
231             goto found;
232         }
233     }
234 
235     /* If not, write a table back and replace it */
236     i = qcow2_cache_find_entry_to_replace(c);
237     trace_qcow2_cache_get_replace_entry(qemu_coroutine_self(),
238                                         c == s->l2_table_cache, i);
239     if (i < 0) {
240         return i;
241     }
242 
243     ret = qcow2_cache_entry_flush(bs, c, i);
244     if (ret < 0) {
245         return ret;
246     }
247 
248     trace_qcow2_cache_get_read(qemu_coroutine_self(),
249                                c == s->l2_table_cache, i);
250     c->entries[i].offset = 0;
251     if (read_from_disk) {
252         if (c == s->l2_table_cache) {
253             BLKDBG_EVENT(bs->file, BLKDBG_L2_LOAD);
254         }
255 
256         ret = bdrv_pread(bs->file, offset, c->entries[i].table, s->cluster_size);
257         if (ret < 0) {
258             return ret;
259         }
260     }
261 
262     /* Give the table some hits for the start so that it won't be replaced
263      * immediately. The number 32 is completely arbitrary. */
264     c->entries[i].cache_hits = 32;
265     c->entries[i].offset = offset;
266 
267     /* And return the right table */
268 found:
269     c->entries[i].cache_hits++;
270     c->entries[i].ref++;
271     *table = c->entries[i].table;
272 
273     trace_qcow2_cache_get_done(qemu_coroutine_self(),
274                                c == s->l2_table_cache, i);
275 
276     return 0;
277 }
278 
279 int qcow2_cache_get(BlockDriverState *bs, Qcow2Cache *c, uint64_t offset,
280     void **table)
281 {
282     return qcow2_cache_do_get(bs, c, offset, table, true);
283 }
284 
285 int qcow2_cache_get_empty(BlockDriverState *bs, Qcow2Cache *c, uint64_t offset,
286     void **table)
287 {
288     return qcow2_cache_do_get(bs, c, offset, table, false);
289 }
290 
291 int qcow2_cache_put(BlockDriverState *bs, Qcow2Cache *c, void **table)
292 {
293     int i;
294 
295     for (i = 0; i < c->size; i++) {
296         if (c->entries[i].table == *table) {
297             goto found;
298         }
299     }
300     return -ENOENT;
301 
302 found:
303     c->entries[i].ref--;
304     *table = NULL;
305 
306     assert(c->entries[i].ref >= 0);
307     return 0;
308 }
309 
310 void qcow2_cache_entry_mark_dirty(Qcow2Cache *c, void *table)
311 {
312     int i;
313 
314     for (i = 0; i < c->size; i++) {
315         if (c->entries[i].table == table) {
316             goto found;
317         }
318     }
319     abort();
320 
321 found:
322     c->entries[i].dirty = true;
323 }
324