1 /*
2  * Copyright (c) 2020 Ori Bernstein
3  * Copyright (c) 2021, 2022 Stefan Sperling <stsp@openbsd.org>
4  *
5  * Permission to use, copy, modify, and distribute this software for any
6  * purpose with or without fee is hereby granted, provided that the above
7  * copyright notice and this permission notice appear in all copies.
8  *
9  * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
10  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
11  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
12  * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
13  * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
14  * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
15  * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
16  */
17 
18 #include <sys/types.h>
19 #include <sys/queue.h>
20 #include <sys/tree.h>
21 #include <sys/uio.h>
22 
23 #include <sha1.h>
24 #include <sha2.h>
25 #include <limits.h>
26 #include <stdio.h>
27 #include <stdint.h>
28 #include <stdlib.h>
29 #include <string.h>
30 #include <time.h>
31 #include <imsg.h>
32 #include <inttypes.h>
33 #include <unistd.h>
34 
35 #include "got_error.h"
36 #include "got_cancel.h"
37 #include "got_object.h"
38 #include "got_reference.h"
39 #include "got_repository_admin.h"
40 #include "got_path.h"
41 
42 #include "got_lib_delta.h"
43 #include "got_lib_object.h"
44 #include "got_lib_object_cache.h"
45 #include "got_lib_object_idset.h"
46 #include "got_lib_ratelimit.h"
47 #include "got_lib_pack.h"
48 #include "got_lib_pack_create.h"
49 #include "got_lib_repository.h"
50 
51 static const struct got_error *
52 get_base_object_id(struct got_object_id *base_id, struct got_packidx *packidx,
53     off_t base_offset)
54 {
55 	const struct got_error *err;
56 	int idx;
57 
58 	err = got_packidx_get_offset_idx(&idx, packidx, base_offset);
59 	if (err)
60 		return err;
61 	if (idx == -1)
62 		return got_error(GOT_ERR_BAD_PACKIDX);
63 
64 	return got_packidx_get_object_id(base_id, packidx, idx);
65 }
66 
67 struct search_deltas_arg {
68 	struct got_pack_metavec *v;
69 	struct got_packidx *packidx;
70 	struct got_pack *pack;
71 	struct got_object_idset *idset;
72 	int ncolored, nfound, ntrees, ncommits;
73 	got_pack_progress_cb progress_cb;
74 	void *progress_arg;
75 	struct got_ratelimit *rl;
76 	got_cancel_cb cancel_cb;
77 	void *cancel_arg;
78 };
79 
80 static const struct got_error *
81 search_delta_for_object(struct got_object_id *id, void *data, void *arg)
82 {
83 	const struct got_error *err;
84 	struct search_deltas_arg *a = arg;
85 	int obj_idx;
86 	uint8_t *delta_buf = NULL;
87 	uint64_t base_size, result_size;
88 	size_t delta_size, delta_compressed_size;
89 	off_t delta_offset, delta_data_offset, base_offset;
90 	struct got_object_id base_id;
91 
92 	if (a->cancel_cb) {
93 		err = a->cancel_cb(a->cancel_arg);
94 		if (err)
95 			return err;
96 	}
97 
98 	obj_idx = got_packidx_get_object_idx(a->packidx, id);
99 	if (obj_idx == -1)
100 		return NULL; /* object not present in our pack file */
101 
102 	err = got_packfile_extract_raw_delta(&delta_buf, &delta_size,
103 	    &delta_compressed_size, &delta_offset, &delta_data_offset,
104 	    &base_offset, &base_id, &base_size, &result_size,
105 	    a->pack, a->packidx, obj_idx);
106 	if (err) {
107 		if (err->code == GOT_ERR_OBJ_TYPE)
108 			return NULL; /* object not stored as a delta */
109 		return err;
110 	}
111 
112 	/*
113 	 * If this is an offset delta we must determine the base
114 	 * object ID ourselves.
115 	 */
116 	if (base_offset != 0) {
117 		err = get_base_object_id(&base_id, a->packidx, base_offset);
118 		if (err)
119 			goto done;
120 	}
121 
122 	if (got_object_idset_contains(a->idset, &base_id)) {
123 		struct got_pack_meta *m, *base;
124 
125 		m = got_object_idset_get(a->idset, id);
126 		if (m == NULL) {
127 			err = got_error_msg(GOT_ERR_NO_OBJ,
128 			    "delta object not found");
129 			goto done;
130 		}
131 
132 		base = got_object_idset_get(a->idset, &base_id);
133 		if (m == NULL) {
134 			err = got_error_msg(GOT_ERR_NO_OBJ,
135 			    "delta base object not found");
136 			goto done;
137 		}
138 
139 		m->base_obj_id = got_object_id_dup(&base_id);
140 		if (m->base_obj_id == NULL) {
141 			err = got_error_from_errno("got_object_id_dup");
142 			goto done;
143 		}
144 
145 		m->prev = base;
146 		m->size = result_size;
147 		m->delta_len = delta_size;
148 		m->delta_compressed_len = delta_compressed_size;
149 		m->reused_delta_offset = delta_data_offset;
150 		m->delta_offset = 0;
151 
152 		err = got_pack_add_meta(m, a->v);
153 		if (err)
154 			goto done;
155 
156 		err = got_pack_report_progress(a->progress_cb, a->progress_arg,
157 		    a->rl, a->ncolored, a->nfound, a->ntrees, 0L, a->ncommits,
158 		    got_object_idset_num_elements(a->idset), a->v->nmeta, 0);
159 		if (err)
160 			goto done;
161 	}
162 done:
163 	free(delta_buf);
164 	return err;
165 }
166 
167 const struct got_error *
168 got_pack_search_deltas(struct got_packidx **packidx, struct got_pack **pack,
169     struct got_pack_metavec *v, struct got_object_idset *idset,
170     int ncolored, int nfound, int ntrees, int ncommits,
171     struct got_repository *repo,
172     got_pack_progress_cb progress_cb, void *progress_arg,
173     struct got_ratelimit *rl, got_cancel_cb cancel_cb, void *cancel_arg)
174 {
175 	const struct got_error *err = NULL;
176 	struct search_deltas_arg sda;
177 
178 	*packidx = NULL;
179 	*pack = NULL;
180 
181 	err = got_pack_find_pack_for_reuse(packidx, repo);
182 	if (err)
183 		return err;
184 
185 	if (*packidx == NULL)
186 		return NULL;
187 
188 	err = got_pack_cache_pack_for_packidx(pack, *packidx, repo);
189 	if (err)
190 		return err;
191 
192 	memset(&sda, 0, sizeof(sda));
193 	sda.v = v;
194 	sda.idset = idset;
195 	sda.pack = *pack;
196 	sda.packidx = *packidx;
197 	sda.ncolored = ncolored;
198 	sda.nfound = nfound;
199 	sda.ntrees = ntrees;
200 	sda.ncommits = ncommits;
201 	sda.progress_cb = progress_cb;
202 	sda.progress_arg = progress_arg;
203 	sda.rl = rl;
204 	sda.cancel_cb = cancel_cb;
205 	sda.cancel_arg = cancel_arg;
206 	return got_object_idset_for_each(idset, search_delta_for_object, &sda);
207 }
208 
209 const struct got_error *
210 got_pack_load_packed_object_ids(int *found_all_objects,
211     struct got_object_id **ours, int nours,
212     struct got_object_id **theirs, int ntheirs,
213     int want_meta, uint32_t seed, struct got_object_idset *idset,
214     struct got_object_idset *idset_exclude, int loose_obj_only,
215     struct got_repository *repo, struct got_packidx *packidx,
216     int *ncolored, int *nfound, int *ntrees,
217     got_pack_progress_cb progress_cb, void *progress_arg,
218     struct got_ratelimit *rl, got_cancel_cb cancel_cb, void *cancel_arg)
219 {
220 	/* We do not need this optimized traversal while using direct I/O. */
221 	*found_all_objects = 0;
222 	return NULL;
223 }
224 
225 const struct got_error *
226 got_pack_paint_commits(int *ncolored, struct got_object_id_queue *ids, int nids,
227     struct got_object_idset *keep, struct got_object_idset *drop,
228     struct got_object_idset *skip, struct got_repository *repo,
229     got_pack_progress_cb progress_cb, void *progress_arg,
230     struct got_ratelimit *rl, got_cancel_cb cancel_cb, void *cancel_arg)
231 {
232 	const struct got_error *err = NULL;
233 	struct got_commit_object *commit = NULL;
234 	struct got_packidx *packidx = NULL;
235 	struct got_pack *pack = NULL;
236 	const struct got_object_id_queue *parents;
237 	struct got_object_qid *qid = NULL;
238 	int nqueued = nids, nskip = 0;
239 
240 	while (!STAILQ_EMPTY(ids) && nskip != nqueued) {
241 		intptr_t color;
242 
243 		if (cancel_cb) {
244 			err = cancel_cb(cancel_arg);
245 			if (err)
246 				break;
247 		}
248 
249 		qid = STAILQ_FIRST(ids);
250 		STAILQ_REMOVE_HEAD(ids, entry);
251 		nqueued--;
252 		color = (intptr_t)qid->data;
253 		if (color == COLOR_SKIP)
254 			nskip--;
255 
256 		if (got_object_idset_contains(skip, &qid->id)) {
257 			got_object_qid_free(qid);
258 			qid = NULL;
259 			continue;
260 		}
261 		if (color == COLOR_KEEP &&
262 		    got_object_idset_contains(keep, &qid->id)) {
263 			got_object_qid_free(qid);
264 			qid = NULL;
265 			continue;
266 		}
267 		if (color == COLOR_DROP &&
268 		    got_object_idset_contains(drop, &qid->id)) {
269 			got_object_qid_free(qid);
270 			qid = NULL;
271 			continue;
272 		}
273 
274 		switch (color) {
275 		case COLOR_KEEP:
276 			if (got_object_idset_contains(drop, &qid->id)) {
277 				err = got_pack_paint_commit(qid, COLOR_SKIP);
278 				if (err)
279 					goto done;
280 			} else
281 				(*ncolored)++;
282 			err = got_object_idset_add(keep, &qid->id, NULL);
283 			if (err)
284 				goto done;
285 			break;
286 		case COLOR_DROP:
287 			if (got_object_idset_contains(keep, &qid->id)) {
288 				err = got_pack_paint_commit(qid, COLOR_SKIP);
289 				if (err)
290 					goto done;
291 			} else
292 				(*ncolored)++;
293 			err = got_object_idset_add(drop, &qid->id, NULL);
294 			if (err)
295 				goto done;
296 			break;
297 		case COLOR_SKIP:
298 			if (!got_object_idset_contains(skip, &qid->id)) {
299 				err = got_object_idset_add(skip, &qid->id,
300 				    NULL);
301 				if (err)
302 					goto done;
303 			}
304 			break;
305 		default:
306 			/* should not happen */
307 			err = got_error_fmt(GOT_ERR_NOT_IMPL,
308 			    "%s invalid commit color %"PRIdPTR, __func__,
309 			    color);
310 			goto done;
311 		}
312 
313 		err = got_pack_report_progress(progress_cb, progress_arg, rl,
314 		    *ncolored, 0, 0, 0L, 0, 0, 0, 0);
315 		if (err)
316 			break;
317 
318 		err = got_object_open_as_commit(&commit, repo, &qid->id);
319 		if (err)
320 			break;
321 
322 		parents = got_object_commit_get_parent_ids(commit);
323 		if (parents) {
324 			struct got_object_qid *pid;
325 			color = (intptr_t)qid->data;
326 			STAILQ_FOREACH(pid, parents, entry) {
327 				err = got_pack_queue_commit_id(ids, &pid->id,
328 				    color, repo);
329 				if (err)
330 					break;
331 				nqueued++;
332 				if (color == COLOR_SKIP)
333 					nskip++;
334 			}
335 		}
336 
337 		if (pack == NULL && (commit->flags & GOT_COMMIT_FLAG_PACKED)) {
338 			/*
339 			 * We now know that at least one pack file exists.
340 			 * Pin a suitable pack to ensure it remains cached
341 			 * while we are churning through commit history.
342 			 */
343 			if (packidx == NULL) {
344 				err = got_pack_find_pack_for_commit_painting(
345 				    &packidx, ids, nqueued, repo);
346 				if (err)
347 					goto done;
348 			}
349 			if (packidx != NULL) {
350 				err = got_pack_cache_pack_for_packidx(&pack,
351 				    packidx, repo);
352 				if (err)
353 					goto done;
354 				err = got_repo_pin_pack(repo, packidx, pack);
355 				if (err)
356 					goto done;
357 			}
358 		}
359 
360 		got_object_commit_close(commit);
361 		commit = NULL;
362 
363 		got_object_qid_free(qid);
364 		qid = NULL;
365 	}
366 done:
367 	if (commit)
368 		got_object_commit_close(commit);
369 	got_object_qid_free(qid);
370 	got_repo_unpin_pack(repo);
371 	return err;
372 }