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/queue.h>
19 #include <sys/resource.h>
21 #include <stdio.h>
22 #include <stdlib.h>
23 #include <stdint.h>
24 #include <string.h>
25 #include <limits.h>
26 #include <sha1.h>
27 #include <sha2.h>
28 #include <zlib.h>
30 #include "got_error.h"
31 #include "got_object.h"
33 #include "got_lib_delta.h"
34 #include "got_lib_inflate.h"
35 #include "got_lib_object.h"
36 #include "got_lib_object_idset.h"
37 #include "got_lib_object_cache.h"
39 /*
40 * XXX This should be reworked to track cache size and usage in bytes,
41 * rather than tracking N elements capped to a maximum element size.
42 */
43 #define GOT_OBJECT_CACHE_SIZE_OBJ 256
44 #define GOT_OBJECT_CACHE_SIZE_TREE 256
45 #define GOT_OBJECT_CACHE_SIZE_COMMIT 64
46 #define GOT_OBJECT_CACHE_SIZE_TAG 256
47 #define GOT_OBJECT_CACHE_SIZE_RAW 16
48 #define GOT_OBJECT_CACHE_MAX_ELEM_SIZE 1048576 /* 1 MB */
50 const struct got_error *
51 got_object_cache_init(struct got_object_cache *cache,
52 enum got_object_cache_type type)
53 {
54 struct rlimit rl;
56 memset(cache, 0, sizeof(*cache));
58 cache->idset = got_object_idset_alloc();
59 if (cache->idset == NULL)
60 return got_error_from_errno("got_object_idset_alloc");
62 cache->type = type;
63 switch (type) {
64 case GOT_OBJECT_CACHE_TYPE_OBJ:
65 cache->size = GOT_OBJECT_CACHE_SIZE_OBJ;
66 break;
67 case GOT_OBJECT_CACHE_TYPE_TREE:
68 cache->size = GOT_OBJECT_CACHE_SIZE_TREE;
69 break;
70 case GOT_OBJECT_CACHE_TYPE_COMMIT:
71 cache->size = GOT_OBJECT_CACHE_SIZE_COMMIT;
72 break;
73 case GOT_OBJECT_CACHE_TYPE_TAG:
74 cache->size = GOT_OBJECT_CACHE_SIZE_TAG;
75 break;
76 case GOT_OBJECT_CACHE_TYPE_RAW:
77 if (getrlimit(RLIMIT_NOFILE, &rl) == -1)
78 return got_error_from_errno("getrlimit");
79 cache->size = GOT_OBJECT_CACHE_SIZE_RAW;
80 if (cache->size > rl.rlim_cur / 16)
81 cache->size = rl.rlim_cur / 16;
82 break;
83 }
84 return NULL;
85 }
87 static size_t
88 get_size_obj(struct got_object *obj)
89 {
90 size_t size = sizeof(*obj);
91 struct got_delta *delta;
93 if ((obj->flags & GOT_OBJ_FLAG_DELTIFIED) == 0)
94 return size;
96 STAILQ_FOREACH(delta, &obj->deltas.entries, entry) {
97 if (SIZE_MAX - sizeof(*delta) < size)
98 return SIZE_MAX;
99 size += sizeof(*delta);
102 return size;
105 static size_t
106 get_size_tree(struct got_tree_object *tree)
108 size_t size = sizeof(*tree);
110 size += sizeof(struct got_tree_entry) * tree->nentries;
111 return size;
114 static size_t
115 get_size_commit(struct got_commit_object *commit)
117 size_t size = sizeof(*commit);
118 struct got_object_qid *qid;
120 size += sizeof(*commit->tree_id);
121 size += strlen(commit->author);
122 size += strlen(commit->committer);
123 size += strlen(commit->logmsg);
125 STAILQ_FOREACH(qid, &commit->parent_ids, entry)
126 size += sizeof(*qid) + sizeof(qid->id);
128 return size;
131 static size_t
132 get_size_tag(struct got_tag_object *tag)
134 size_t size = sizeof(*tag);
136 size += strlen(tag->tag);
137 size += strlen(tag->tagger);
138 size += strlen(tag->tagmsg);
140 return size;
143 static size_t
144 get_size_raw(struct got_raw_object *raw)
146 return sizeof(*raw);
149 const struct got_error *
150 got_object_cache_add(struct got_object_cache *cache, struct got_object_id *id,
151 void *item)
153 const struct got_error *err = NULL;
154 struct got_object_cache_entry *ce;
155 int nelem;
156 size_t size;
158 switch (cache->type) {
159 case GOT_OBJECT_CACHE_TYPE_OBJ:
160 size = get_size_obj((struct got_object *)item);
161 break;
162 case GOT_OBJECT_CACHE_TYPE_TREE:
163 size = get_size_tree((struct got_tree_object *)item);
164 break;
165 case GOT_OBJECT_CACHE_TYPE_COMMIT:
166 size = get_size_commit((struct got_commit_object *)item);
167 break;
168 case GOT_OBJECT_CACHE_TYPE_TAG:
169 size = get_size_tag((struct got_tag_object *)item);
170 break;
171 case GOT_OBJECT_CACHE_TYPE_RAW:
172 size = get_size_raw((struct got_raw_object *)item);
173 break;
174 default:
175 return got_error(GOT_ERR_OBJ_TYPE);
178 if (size > GOT_OBJECT_CACHE_MAX_ELEM_SIZE) {
179 #ifdef GOT_OBJ_CACHE_DEBUG
180 char *id_str;
181 if (got_object_id_str(&id_str, id) != NULL)
182 return got_error_from_errno("got_object_id_str");
183 fprintf(stderr, "%s: not caching ", getprogname());
184 switch (cache->type) {
185 case GOT_OBJECT_CACHE_TYPE_OBJ:
186 fprintf(stderr, "object");
187 break;
188 case GOT_OBJECT_CACHE_TYPE_TREE:
189 fprintf(stderr, "tree");
190 break;
191 case GOT_OBJECT_CACHE_TYPE_COMMIT:
192 fprintf(stderr, "commit");
193 break;
194 case GOT_OBJECT_CACHE_TYPE_TAG:
195 fprintf(stderr, "tag");
196 break;
197 case GOT_OBJECT_CACHE_TYPE_RAW:
198 fprintf(stderr, "raw");
199 break;
201 fprintf(stderr, " %s (%zd bytes; %zd MB)\n", id_str, size,
202 size/1024/1024);
203 free(id_str);
204 #endif
205 cache->cache_toolarge++;
206 return got_error(GOT_ERR_OBJ_TOO_LARGE);
209 nelem = got_object_idset_num_elements(cache->idset);
210 if (nelem >= cache->size) {
211 err = got_object_idset_remove((void **)&ce,
212 cache->idset, NULL);
213 if (err)
214 return err;
215 switch (cache->type) {
216 case GOT_OBJECT_CACHE_TYPE_OBJ:
217 got_object_close(ce->data.obj);
218 break;
219 case GOT_OBJECT_CACHE_TYPE_TREE:
220 got_object_tree_close(ce->data.tree);
221 break;
222 case GOT_OBJECT_CACHE_TYPE_COMMIT:
223 got_object_commit_close(ce->data.commit);
224 break;
225 case GOT_OBJECT_CACHE_TYPE_TAG:
226 got_object_tag_close(ce->data.tag);
227 break;
228 case GOT_OBJECT_CACHE_TYPE_RAW:
229 got_object_raw_close(ce->data.raw);
230 break;
232 memset(ce, 0, sizeof(*ce));
233 cache->cache_evict++;
234 } else {
235 ce = malloc(sizeof(*ce));
236 if (ce == NULL)
237 return got_error_from_errno("malloc");
240 memcpy(&ce->id, id, sizeof(ce->id));
241 switch (cache->type) {
242 case GOT_OBJECT_CACHE_TYPE_OBJ:
243 ce->data.obj = (struct got_object *)item;
244 break;
245 case GOT_OBJECT_CACHE_TYPE_TREE:
246 ce->data.tree = (struct got_tree_object *)item;
247 break;
248 case GOT_OBJECT_CACHE_TYPE_COMMIT:
249 ce->data.commit = (struct got_commit_object *)item;
250 break;
251 case GOT_OBJECT_CACHE_TYPE_TAG:
252 ce->data.tag = (struct got_tag_object *)item;
253 break;
254 case GOT_OBJECT_CACHE_TYPE_RAW:
255 ce->data.raw = (struct got_raw_object *)item;
256 break;
259 err = got_object_idset_add(cache->idset, id, ce);
260 if (err)
261 free(ce);
262 else if (size > cache->max_cached_size)
263 cache->max_cached_size = size;
264 return err;
267 void *
268 got_object_cache_get(struct got_object_cache *cache, struct got_object_id *id)
270 struct got_object_cache_entry *ce;
272 cache->cache_searches++;
273 ce = got_object_idset_get(cache->idset, id);
274 if (ce) {
275 cache->cache_hit++;
276 switch (cache->type) {
277 case GOT_OBJECT_CACHE_TYPE_OBJ:
278 return ce->data.obj;
279 case GOT_OBJECT_CACHE_TYPE_TREE:
280 return ce->data.tree;
281 case GOT_OBJECT_CACHE_TYPE_COMMIT:
282 return ce->data.commit;
283 case GOT_OBJECT_CACHE_TYPE_TAG:
284 return ce->data.tag;
285 case GOT_OBJECT_CACHE_TYPE_RAW:
286 return ce->data.raw;
290 cache->cache_miss++;
291 return NULL;
294 #ifdef GOT_OBJ_CACHE_DEBUG
295 static void
296 print_cache_stats(struct got_object_cache *cache, const char *name)
298 fprintf(stderr, "%s: %s cache: %d elements, %d searches, %d hits, "
299 "%d missed, %d evicted, %d too large, max cached %zd bytes\n",
300 getprogname(), name,
301 cache->idset ? got_object_idset_num_elements(cache->idset) : -1,
302 cache->cache_searches, cache->cache_hit,
303 cache->cache_miss, cache->cache_evict, cache->cache_toolarge,
304 cache->max_cached_size);
307 static const struct got_error *
308 check_refcount(struct got_object_id *id, void *data, void *arg)
310 struct got_object_cache *cache = arg;
311 struct got_object_cache_entry *ce = data;
312 struct got_object *obj;
313 struct got_tree_object *tree;
314 struct got_commit_object *commit;
315 struct got_tag_object *tag;
316 struct got_raw_object *raw;
317 char *id_str;
319 if (got_object_id_str(&id_str, id) != NULL)
320 return NULL;
322 switch (cache->type) {
323 case GOT_OBJECT_CACHE_TYPE_OBJ:
324 obj = ce->data.obj;
325 if (obj->refcnt == 1)
326 break;
327 fprintf(stderr, "object %s has %d unclaimed references\n",
328 id_str, obj->refcnt - 1);
329 break;
330 case GOT_OBJECT_CACHE_TYPE_TREE:
331 tree = ce->data.tree;
332 if (tree->refcnt == 1)
333 break;
334 fprintf(stderr, "tree %s has %d unclaimed references\n",
335 id_str, tree->refcnt - 1);
336 break;
337 case GOT_OBJECT_CACHE_TYPE_COMMIT:
338 commit = ce->data.commit;
339 if (commit->refcnt == 1)
340 break;
341 fprintf(stderr, "commit %s has %d unclaimed references\n",
342 id_str, commit->refcnt - 1);
343 break;
344 case GOT_OBJECT_CACHE_TYPE_TAG:
345 tag = ce->data.tag;
346 if (tag->refcnt == 1)
347 break;
348 fprintf(stderr, "tag %s has %d unclaimed references\n",
349 id_str, tag->refcnt - 1);
350 break;
351 case GOT_OBJECT_CACHE_TYPE_RAW:
352 raw = ce->data.raw;
353 if (raw->refcnt == 1)
354 break;
355 fprintf(stderr, "raw %s has %d unclaimed references\n",
356 id_str, raw->refcnt - 1);
357 break;
359 free(id_str);
360 return NULL;
362 #endif
364 static const struct got_error *
365 free_entry(struct got_object_id *id, void *data, void *arg)
367 struct got_object_cache *cache = arg;
368 struct got_object_cache_entry *ce = data;
370 switch (cache->type) {
371 case GOT_OBJECT_CACHE_TYPE_OBJ:
372 got_object_close(ce->data.obj);
373 break;
374 case GOT_OBJECT_CACHE_TYPE_TREE:
375 got_object_tree_close(ce->data.tree);
376 break;
377 case GOT_OBJECT_CACHE_TYPE_COMMIT:
378 got_object_commit_close(ce->data.commit);
379 break;
380 case GOT_OBJECT_CACHE_TYPE_TAG:
381 got_object_tag_close(ce->data.tag);
382 break;
383 case GOT_OBJECT_CACHE_TYPE_RAW:
384 got_object_raw_close(ce->data.raw);
385 break;
388 free(ce);
390 return NULL;
393 void
394 got_object_cache_close(struct got_object_cache *cache)
396 #ifdef GOT_OBJ_CACHE_DEBUG
397 switch (cache->type) {
398 case GOT_OBJECT_CACHE_TYPE_OBJ:
399 print_cache_stats(cache, "object");
400 break;
401 case GOT_OBJECT_CACHE_TYPE_TREE:
402 print_cache_stats(cache, "tree");
403 break;
404 case GOT_OBJECT_CACHE_TYPE_COMMIT:
405 print_cache_stats(cache, "commit");
406 break;
407 case GOT_OBJECT_CACHE_TYPE_TAG:
408 print_cache_stats(cache, "tag");
409 break;
410 case GOT_OBJECT_CACHE_TYPE_RAW:
411 print_cache_stats(cache, "raw");
412 break;
415 if (cache->idset)
416 got_object_idset_for_each(cache->idset, check_refcount, cache);
417 #endif
419 if (cache->idset) {
420 got_object_idset_for_each(cache->idset, free_entry, cache);
421 got_object_idset_free(cache->idset);
422 cache->idset = NULL;
424 cache->size = 0;