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