Blob


1 /*
2 * Copyright (c) 2018 Stefan Sperling <stsp@openbsd.org>
3 *
4 * Permission to use, copy, modify, and distribute this software for any
5 * purpose with or without fee is hereby granted, provided that the above
6 * copyright notice and this permission notice appear in all copies.
7 *
8 * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
9 * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
10 * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
11 * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
12 * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
13 * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
14 * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
15 */
17 #include <sys/time.h>
18 #include <sys/resource.h>
20 #include <stdio.h>
21 #include <stdlib.h>
22 #include <stdint.h>
23 #include <string.h>
24 #include <limits.h>
25 #include <zlib.h>
27 #include "got_error.h"
28 #include "got_object.h"
30 #include "got_lib_delta.h"
31 #include "got_lib_inflate.h"
32 #include "got_lib_object.h"
33 #include "got_lib_object_idset.h"
34 #include "got_lib_object_cache.h"
36 /*
37 * XXX This should be reworked to track cache size and usage in bytes,
38 * rather than tracking N elements capped to a maximum element size.
39 */
40 #define GOT_OBJECT_CACHE_SIZE_OBJ 256
41 #define GOT_OBJECT_CACHE_SIZE_TREE 256
42 #define GOT_OBJECT_CACHE_SIZE_COMMIT 64
43 #define GOT_OBJECT_CACHE_SIZE_TAG 2048
44 #define GOT_OBJECT_CACHE_SIZE_RAW 64
45 #define GOT_OBJECT_CACHE_MAX_ELEM_SIZE 1048576 /* 1 MB */
47 const struct got_error *
48 got_object_cache_init(struct got_object_cache *cache,
49 enum got_object_cache_type type)
50 {
51 struct rlimit rl;
53 memset(cache, 0, sizeof(*cache));
55 cache->idset = got_object_idset_alloc();
56 if (cache->idset == NULL)
57 return got_error_from_errno("got_object_idset_alloc");
59 cache->type = type;
60 switch (type) {
61 case GOT_OBJECT_CACHE_TYPE_OBJ:
62 cache->size = GOT_OBJECT_CACHE_SIZE_OBJ;
63 break;
64 case GOT_OBJECT_CACHE_TYPE_TREE:
65 cache->size = GOT_OBJECT_CACHE_SIZE_TREE;
66 break;
67 case GOT_OBJECT_CACHE_TYPE_COMMIT:
68 cache->size = GOT_OBJECT_CACHE_SIZE_COMMIT;
69 break;
70 case GOT_OBJECT_CACHE_TYPE_TAG:
71 cache->size = GOT_OBJECT_CACHE_SIZE_TAG;
72 break;
73 case GOT_OBJECT_CACHE_TYPE_RAW:
74 if (getrlimit(RLIMIT_NOFILE, &rl) == -1)
75 return got_error_from_errno("getrlimit");
76 cache->size = GOT_OBJECT_CACHE_SIZE_RAW;
77 if (cache->size > rl.rlim_cur / 16)
78 cache->size = rl.rlim_cur / 16;
79 break;
80 }
81 return NULL;
82 }
84 size_t
85 get_size_obj(struct got_object *obj)
86 {
87 size_t size = sizeof(*obj);
88 struct got_delta *delta;
90 if ((obj->flags & GOT_OBJ_FLAG_DELTIFIED) == 0)
91 return size;
93 STAILQ_FOREACH(delta, &obj->deltas.entries, entry) {
94 if (SIZE_MAX - sizeof(*delta) < size)
95 return SIZE_MAX;
96 size += sizeof(*delta);
97 }
99 return size;
102 size_t
103 get_size_tree(struct got_tree_object *tree)
105 size_t size = sizeof(*tree);
107 size += sizeof(struct got_tree_entry) * tree->nentries;
108 return size;
111 size_t
112 get_size_commit(struct got_commit_object *commit)
114 size_t size = sizeof(*commit);
115 struct got_object_qid *qid;
117 size += sizeof(*commit->tree_id);
118 size += strlen(commit->author);
119 size += strlen(commit->committer);
120 size += strlen(commit->logmsg);
122 STAILQ_FOREACH(qid, &commit->parent_ids, entry)
123 size += sizeof(*qid) + sizeof(*qid->id);
125 return size;
128 size_t
129 get_size_tag(struct got_tag_object *tag)
131 size_t size = sizeof(*tag);
133 size += strlen(tag->tag);
134 size += strlen(tag->tagger);
135 size += strlen(tag->tagmsg);
137 return size;
140 size_t
141 get_size_raw(struct got_raw_object *raw)
143 return sizeof(*raw);
146 const struct got_error *
147 got_object_cache_add(struct got_object_cache *cache, struct got_object_id *id, void *item)
149 const struct got_error *err = NULL;
150 struct got_object_cache_entry *ce;
151 int nelem;
152 size_t size;
154 switch (cache->type) {
155 case GOT_OBJECT_CACHE_TYPE_OBJ:
156 size = get_size_obj((struct got_object *)item);
157 break;
158 case GOT_OBJECT_CACHE_TYPE_TREE:
159 size = get_size_tree((struct got_tree_object *)item);
160 break;
161 case GOT_OBJECT_CACHE_TYPE_COMMIT:
162 size = get_size_commit((struct got_commit_object *)item);
163 break;
164 case GOT_OBJECT_CACHE_TYPE_TAG:
165 size = get_size_tag((struct got_tag_object *)item);
166 break;
167 case GOT_OBJECT_CACHE_TYPE_RAW:
168 size = get_size_raw((struct got_raw_object *)item);
169 break;
170 default:
171 return got_error(GOT_ERR_OBJ_TYPE);
174 if (size > GOT_OBJECT_CACHE_MAX_ELEM_SIZE) {
175 #ifdef GOT_OBJ_CACHE_DEBUG
176 char *id_str;
177 if (got_object_id_str(&id_str, id) != NULL)
178 return got_error_from_errno("got_object_id_str");
179 fprintf(stderr, "%s: not caching ", getprogname());
180 switch (cache->type) {
181 case GOT_OBJECT_CACHE_TYPE_OBJ:
182 fprintf(stderr, "object");
183 break;
184 case GOT_OBJECT_CACHE_TYPE_TREE:
185 fprintf(stderr, "tree");
186 break;
187 case GOT_OBJECT_CACHE_TYPE_COMMIT:
188 fprintf(stderr, "commit");
189 break;
190 case GOT_OBJECT_CACHE_TYPE_TAG:
191 fprintf(stderr, "tag");
192 break;
193 case GOT_OBJECT_CACHE_TYPE_RAW:
194 fprintf(stderr, "raw");
195 break;
197 fprintf(stderr, " %s (%zd bytes; %zd MB)\n", id_str, size,
198 size/1024/1024);
199 free(id_str);
200 #endif
201 cache->cache_toolarge++;
202 return got_error(GOT_ERR_OBJ_TOO_LARGE);
205 nelem = got_object_idset_num_elements(cache->idset);
206 if (nelem >= cache->size) {
207 err = got_object_idset_remove((void **)&ce,
208 cache->idset, NULL);
209 if (err)
210 return err;
211 switch (cache->type) {
212 case GOT_OBJECT_CACHE_TYPE_OBJ:
213 got_object_close(ce->data.obj);
214 break;
215 case GOT_OBJECT_CACHE_TYPE_TREE:
216 got_object_tree_close(ce->data.tree);
217 break;
218 case GOT_OBJECT_CACHE_TYPE_COMMIT:
219 got_object_commit_close(ce->data.commit);
220 break;
221 case GOT_OBJECT_CACHE_TYPE_TAG:
222 got_object_tag_close(ce->data.tag);
223 break;
224 case GOT_OBJECT_CACHE_TYPE_RAW:
225 got_object_raw_close(ce->data.raw);
226 break;
228 free(ce);
229 cache->cache_evict++;
232 ce = malloc(sizeof(*ce));
233 if (ce == NULL)
234 return got_error_from_errno("malloc");
235 memcpy(&ce->id, id, sizeof(ce->id));
236 switch (cache->type) {
237 case GOT_OBJECT_CACHE_TYPE_OBJ:
238 ce->data.obj = (struct got_object *)item;
239 break;
240 case GOT_OBJECT_CACHE_TYPE_TREE:
241 ce->data.tree = (struct got_tree_object *)item;
242 break;
243 case GOT_OBJECT_CACHE_TYPE_COMMIT:
244 ce->data.commit = (struct got_commit_object *)item;
245 break;
246 case GOT_OBJECT_CACHE_TYPE_TAG:
247 ce->data.tag = (struct got_tag_object *)item;
248 break;
249 case GOT_OBJECT_CACHE_TYPE_RAW:
250 ce->data.raw = (struct got_raw_object *)item;
251 break;
254 err = got_object_idset_add(cache->idset, id, ce);
255 if (err)
256 free(ce);
257 return err;
260 void *
261 got_object_cache_get(struct got_object_cache *cache, struct got_object_id *id)
263 struct got_object_cache_entry *ce;
265 cache->cache_searches++;
266 ce = got_object_idset_get(cache->idset, id);
267 if (ce) {
268 cache->cache_hit++;
269 switch (cache->type) {
270 case GOT_OBJECT_CACHE_TYPE_OBJ:
271 return ce->data.obj;
272 case GOT_OBJECT_CACHE_TYPE_TREE:
273 return ce->data.tree;
274 case GOT_OBJECT_CACHE_TYPE_COMMIT:
275 return ce->data.commit;
276 case GOT_OBJECT_CACHE_TYPE_TAG:
277 return ce->data.tag;
278 case GOT_OBJECT_CACHE_TYPE_RAW:
279 return ce->data.raw;
283 cache->cache_miss++;
284 return NULL;
287 #ifdef GOT_OBJ_CACHE_DEBUG
288 static void
289 print_cache_stats(struct got_object_cache *cache, const char *name)
291 fprintf(stderr, "%s: %s cache: %d elements, %d searches, %d hits, "
292 "%d missed, %d evicted, %d too large\n", getprogname(), name,
293 got_object_idset_num_elements(cache->idset),
294 cache->cache_searches, cache->cache_hit,
295 cache->cache_miss, cache->cache_evict, cache->cache_toolarge);
298 const struct got_error *
299 check_refcount(struct got_object_id *id, void *data, void *arg)
301 struct got_object_cache *cache = arg;
302 struct got_object_cache_entry *ce = data;
303 struct got_object *obj;
304 struct got_tree_object *tree;
305 struct got_commit_object *commit;
306 struct got_tag_object *tag;
307 struct got_raw_object *raw;
308 char *id_str;
310 if (got_object_id_str(&id_str, id) != NULL)
311 return NULL;
313 switch (cache->type) {
314 case GOT_OBJECT_CACHE_TYPE_OBJ:
315 obj = ce->data.obj;
316 if (obj->refcnt == 1)
317 break;
318 fprintf(stderr, "object %s has %d unclaimed references\n",
319 id_str, obj->refcnt - 1);
320 break;
321 case GOT_OBJECT_CACHE_TYPE_TREE:
322 tree = ce->data.tree;
323 if (tree->refcnt == 1)
324 break;
325 fprintf(stderr, "tree %s has %d unclaimed references\n",
326 id_str, tree->refcnt - 1);
327 break;
328 case GOT_OBJECT_CACHE_TYPE_COMMIT:
329 commit = ce->data.commit;
330 if (commit->refcnt == 1)
331 break;
332 fprintf(stderr, "commit %s has %d unclaimed references\n",
333 id_str, commit->refcnt - 1);
334 break;
335 case GOT_OBJECT_CACHE_TYPE_TAG:
336 tag = ce->data.tag;
337 if (tag->refcnt == 1)
338 break;
339 fprintf(stderr, "tag %s has %d unclaimed references\n",
340 id_str, tag->refcnt - 1);
341 break;
342 case GOT_OBJECT_CACHE_TYPE_RAW:
343 raw = ce->data.raw;
344 if (raw->refcnt == 1)
345 break;
346 fprintf(stderr, "raw %s has %d unclaimed references\n",
347 id_str, raw->refcnt - 1);
348 break;
350 free(id_str);
351 return NULL;
353 #endif
355 void
356 got_object_cache_close(struct got_object_cache *cache)
358 #ifdef GOT_OBJ_CACHE_DEBUG
359 switch (cache->type) {
360 case GOT_OBJECT_CACHE_TYPE_OBJ:
361 print_cache_stats(cache, "object");
362 break;
363 case GOT_OBJECT_CACHE_TYPE_TREE:
364 print_cache_stats(cache, "tree");
365 break;
366 case GOT_OBJECT_CACHE_TYPE_COMMIT:
367 print_cache_stats(cache, "commit");
368 break;
369 case GOT_OBJECT_CACHE_TYPE_TAG:
370 print_cache_stats(cache, "tag");
371 break;
372 case GOT_OBJECT_CACHE_TYPE_RAW:
373 print_cache_stats(cache, "raw");
374 break;
377 got_object_idset_for_each(cache->idset, check_refcount, cache);
378 #endif
380 if (cache->idset) {
381 got_object_idset_free(cache->idset);
382 cache->idset = NULL;
384 cache->size = 0;