lib/pack_create_io.c

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