2 * Copyright (c) 2020 Ori Bernstein
3 * Copyright (c) 2021, 2022 Stefan Sperling <stsp@openbsd.org>
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.
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.
18 #include <sys/types.h>
19 #include <sys/queue.h>
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"
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"
49 static const struct got_error *
50 get_base_object_id(struct got_object_id *base_id, struct got_packidx *packidx,
53 const struct got_error *err;
56 err = got_packidx_get_offset_idx(&idx, packidx, base_offset);
60 return got_error(GOT_ERR_BAD_PACKIDX);
62 return got_packidx_get_object_id(base_id, packidx, idx);
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;
73 struct got_ratelimit *rl;
74 got_cancel_cb cancel_cb;
78 static const struct got_error *
79 search_delta_for_object(struct got_object_id *id, void *data, void *arg)
81 const struct got_error *err;
82 struct search_deltas_arg *a = arg;
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;
91 err = a->cancel_cb(a->cancel_arg);
96 obj_idx = got_packidx_get_object_idx(a->packidx, id);
98 return NULL; /* object not present in our pack file */
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);
105 if (err->code == GOT_ERR_OBJ_TYPE)
106 return NULL; /* object not stored as a delta */
111 * If this is an offset delta we must determine the base
112 * object ID ourselves.
114 if (base_offset != 0) {
115 err = get_base_object_id(&base_id, a->packidx, base_offset);
120 if (got_object_idset_contains(a->idset, &base_id)) {
121 struct got_pack_meta *m, *base;
123 m = got_object_idset_get(a->idset, id);
125 err = got_error_msg(GOT_ERR_NO_OBJ,
126 "delta object not found");
130 base = got_object_idset_get(a->idset, &base_id);
132 err = got_error_msg(GOT_ERR_NO_OBJ,
133 "delta base object not found");
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");
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;
150 err = got_pack_add_meta(m, a->v);
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);
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)
173 const struct got_error *err = NULL;
174 struct search_deltas_arg sda;
179 err = got_pack_find_pack_for_reuse(packidx, repo);
183 if (*packidx == NULL)
186 err = got_pack_cache_pack_for_packidx(pack, *packidx, repo);
190 memset(&sda, 0, sizeof(sda));
194 sda.packidx = *packidx;
195 sda.ncolored = ncolored;
198 sda.ncommits = ncommits;
199 sda.progress_cb = progress_cb;
200 sda.progress_arg = progress_arg;
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);
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)
218 /* We do not need this optimized traversal while using direct I/O. */
219 *found_all_objects = 0;
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)
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;
238 while (!STAILQ_EMPTY(ids) && nskip != nqueued) {
242 err = cancel_cb(cancel_arg);
247 qid = STAILQ_FIRST(ids);
248 STAILQ_REMOVE_HEAD(ids, entry);
250 color = (intptr_t)qid->data;
251 if (color == COLOR_SKIP)
254 if (got_object_idset_contains(skip, &qid->id)) {
255 got_object_qid_free(qid);
259 if (color == COLOR_KEEP &&
260 got_object_idset_contains(keep, &qid->id)) {
261 got_object_qid_free(qid);
265 if (color == COLOR_DROP &&
266 got_object_idset_contains(drop, &qid->id)) {
267 got_object_qid_free(qid);
274 if (got_object_idset_contains(drop, &qid->id)) {
275 err = got_pack_paint_commit(qid, COLOR_SKIP);
280 err = got_object_idset_add(keep, &qid->id, NULL);
285 if (got_object_idset_contains(keep, &qid->id)) {
286 err = got_pack_paint_commit(qid, COLOR_SKIP);
291 err = got_object_idset_add(drop, &qid->id, NULL);
296 if (!got_object_idset_contains(skip, &qid->id)) {
297 err = got_object_idset_add(skip, &qid->id,
304 /* should not happen */
305 err = got_error_fmt(GOT_ERR_NOT_IMPL,
306 "%s invalid commit color %"PRIdPTR, __func__,
311 err = got_pack_report_progress(progress_cb, progress_arg, rl,
312 *ncolored, 0, 0, 0L, 0, 0, 0, 0);
316 err = got_object_open_as_commit(&commit, repo, &qid->id);
320 parents = got_object_commit_get_parent_ids(commit);
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,
330 if (color == COLOR_SKIP)
335 if (pack == NULL && (commit->flags & GOT_COMMIT_FLAG_PACKED)) {
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.
341 if (packidx == NULL) {
342 err = got_pack_find_pack_for_commit_painting(
343 &packidx, ids, nqueued, repo);
347 if (packidx != NULL) {
348 err = got_pack_cache_pack_for_packidx(&pack,
352 err = got_repo_pin_pack(repo, packidx, pack);
358 got_object_commit_close(commit);
361 got_object_qid_free(qid);
366 got_object_commit_close(commit);
367 got_object_qid_free(qid);
368 got_repo_unpin_pack(repo);