4 #include "git-compat-util.h"
12 #include "sha1-lookup.h"
13 #include "commit-graph.h"
14 #include "object-store.h"
17 #include "replace-object.h"
20 #define GRAPH_SIGNATURE 0x43475048 /* "CGPH" */
21 #define GRAPH_CHUNKID_OIDFANOUT 0x4f494446 /* "OIDF" */
22 #define GRAPH_CHUNKID_OIDLOOKUP 0x4f49444c /* "OIDL" */
23 #define GRAPH_CHUNKID_DATA 0x43444154 /* "CDAT" */
24 #define GRAPH_CHUNKID_EXTRAEDGES 0x45444745 /* "EDGE" */
26 #define GRAPH_DATA_WIDTH (the_hash_algo->rawsz + 16)
28 #define GRAPH_VERSION_1 0x1
29 #define GRAPH_VERSION GRAPH_VERSION_1
31 #define GRAPH_EXTRA_EDGES_NEEDED 0x80000000
32 #define GRAPH_EDGE_LAST_MASK 0x7fffffff
33 #define GRAPH_PARENT_NONE 0x70000000
35 #define GRAPH_LAST_EDGE 0x80000000
37 #define GRAPH_HEADER_SIZE 8
38 #define GRAPH_FANOUT_SIZE (4 * 256)
39 #define GRAPH_CHUNKLOOKUP_WIDTH 12
40 #define GRAPH_MIN_SIZE (GRAPH_HEADER_SIZE + 4 * GRAPH_CHUNKLOOKUP_WIDTH \
41 + GRAPH_FANOUT_SIZE + the_hash_algo->rawsz)
43 char *get_commit_graph_filename(const char *obj_dir)
45 return xstrfmt("%s/info/commit-graph", obj_dir);
48 static uint8_t oid_version(void)
53 static struct commit_graph *alloc_commit_graph(void)
55 struct commit_graph *g = xcalloc(1, sizeof(*g));
61 extern int read_replace_refs;
63 static int commit_graph_compatible(struct repository *r)
68 if (read_replace_refs) {
69 prepare_replace_object(r);
70 if (hashmap_get_size(&r->objects->replace_map->map))
74 prepare_commit_graft(r);
75 if (r->parsed_objects && r->parsed_objects->grafts_nr)
77 if (is_repository_shallow(r))
83 int open_commit_graph(const char *graph_file, int *fd, struct stat *st)
85 *fd = git_open(graph_file);
95 struct commit_graph *load_commit_graph_one_fd_st(int fd, struct stat *st)
99 struct commit_graph *ret;
101 graph_size = xsize_t(st->st_size);
103 if (graph_size < GRAPH_MIN_SIZE) {
105 error(_("commit-graph file is too small"));
108 graph_map = xmmap(NULL, graph_size, PROT_READ, MAP_PRIVATE, fd, 0);
109 ret = parse_commit_graph(graph_map, fd, graph_size);
112 munmap(graph_map, graph_size);
119 static int verify_commit_graph_lite(struct commit_graph *g)
122 * Basic validation shared between parse_commit_graph()
123 * which'll be called every time the graph is used, and the
124 * much more expensive verify_commit_graph() used by
125 * "commit-graph verify".
127 * There should only be very basic checks here to ensure that
128 * we don't e.g. segfault in fill_commit_in_graph(), but
129 * because this is a very hot codepath nothing that e.g. loops
130 * over g->num_commits, or runs a checksum on the commit-graph
133 if (!g->chunk_oid_fanout) {
134 error("commit-graph is missing the OID Fanout chunk");
137 if (!g->chunk_oid_lookup) {
138 error("commit-graph is missing the OID Lookup chunk");
141 if (!g->chunk_commit_data) {
142 error("commit-graph is missing the Commit Data chunk");
149 struct commit_graph *parse_commit_graph(void *graph_map, int fd,
152 const unsigned char *data, *chunk_lookup;
154 struct commit_graph *graph;
155 uint64_t last_chunk_offset;
156 uint32_t last_chunk_id;
157 uint32_t graph_signature;
158 unsigned char graph_version, hash_version;
163 if (graph_size < GRAPH_MIN_SIZE)
166 data = (const unsigned char *)graph_map;
168 graph_signature = get_be32(data);
169 if (graph_signature != GRAPH_SIGNATURE) {
170 error(_("commit-graph signature %X does not match signature %X"),
171 graph_signature, GRAPH_SIGNATURE);
175 graph_version = *(unsigned char*)(data + 4);
176 if (graph_version != GRAPH_VERSION) {
177 error(_("commit-graph version %X does not match version %X"),
178 graph_version, GRAPH_VERSION);
182 hash_version = *(unsigned char*)(data + 5);
183 if (hash_version != oid_version()) {
184 error(_("commit-graph hash version %X does not match version %X"),
185 hash_version, oid_version());
189 graph = alloc_commit_graph();
191 graph->hash_len = the_hash_algo->rawsz;
192 graph->num_chunks = *(unsigned char*)(data + 6);
193 graph->graph_fd = fd;
194 graph->data = graph_map;
195 graph->data_len = graph_size;
198 last_chunk_offset = 8;
199 chunk_lookup = data + 8;
200 for (i = 0; i < graph->num_chunks; i++) {
202 uint64_t chunk_offset;
203 int chunk_repeated = 0;
205 if (data + graph_size - chunk_lookup <
206 GRAPH_CHUNKLOOKUP_WIDTH) {
207 error(_("commit-graph chunk lookup table entry missing; file may be incomplete"));
212 chunk_id = get_be32(chunk_lookup + 0);
213 chunk_offset = get_be64(chunk_lookup + 4);
215 chunk_lookup += GRAPH_CHUNKLOOKUP_WIDTH;
217 if (chunk_offset > graph_size - the_hash_algo->rawsz) {
218 error(_("commit-graph improper chunk offset %08x%08x"), (uint32_t)(chunk_offset >> 32),
219 (uint32_t)chunk_offset);
225 case GRAPH_CHUNKID_OIDFANOUT:
226 if (graph->chunk_oid_fanout)
229 graph->chunk_oid_fanout = (uint32_t*)(data + chunk_offset);
232 case GRAPH_CHUNKID_OIDLOOKUP:
233 if (graph->chunk_oid_lookup)
236 graph->chunk_oid_lookup = data + chunk_offset;
239 case GRAPH_CHUNKID_DATA:
240 if (graph->chunk_commit_data)
243 graph->chunk_commit_data = data + chunk_offset;
246 case GRAPH_CHUNKID_EXTRAEDGES:
247 if (graph->chunk_extra_edges)
250 graph->chunk_extra_edges = data + chunk_offset;
254 if (chunk_repeated) {
255 error(_("commit-graph chunk id %08x appears multiple times"), chunk_id);
260 if (last_chunk_id == GRAPH_CHUNKID_OIDLOOKUP)
262 graph->num_commits = (chunk_offset - last_chunk_offset)
266 last_chunk_id = chunk_id;
267 last_chunk_offset = chunk_offset;
270 if (verify_commit_graph_lite(graph)) {
278 static struct commit_graph *load_commit_graph_one(const char *graph_file)
283 int open_ok = open_commit_graph(graph_file, &fd, &st);
288 return load_commit_graph_one_fd_st(fd, &st);
291 static void prepare_commit_graph_one(struct repository *r, const char *obj_dir)
295 if (r->objects->commit_graph)
298 graph_name = get_commit_graph_filename(obj_dir);
299 r->objects->commit_graph =
300 load_commit_graph_one(graph_name);
302 FREE_AND_NULL(graph_name);
306 * Return 1 if commit_graph is non-NULL, and 0 otherwise.
308 * On the first invocation, this function attemps to load the commit
309 * graph if the_repository is configured to have one.
311 static int prepare_commit_graph(struct repository *r)
313 struct object_directory *odb;
316 if (git_env_bool(GIT_TEST_COMMIT_GRAPH_DIE_ON_LOAD, 0))
317 die("dying as requested by the '%s' variable on commit-graph load!",
318 GIT_TEST_COMMIT_GRAPH_DIE_ON_LOAD);
320 if (r->objects->commit_graph_attempted)
321 return !!r->objects->commit_graph;
322 r->objects->commit_graph_attempted = 1;
324 if (!git_env_bool(GIT_TEST_COMMIT_GRAPH, 0) &&
325 (repo_config_get_bool(r, "core.commitgraph", &config_value) ||
328 * This repository is not configured to use commit graphs, so
329 * do not load one. (But report commit_graph_attempted anyway
330 * so that commit graph loading is not attempted again for this
335 if (!commit_graph_compatible(r))
339 for (odb = r->objects->odb;
340 !r->objects->commit_graph && odb;
342 prepare_commit_graph_one(r, odb->path);
343 return !!r->objects->commit_graph;
346 int generation_numbers_enabled(struct repository *r)
348 uint32_t first_generation;
349 struct commit_graph *g;
350 if (!prepare_commit_graph(r))
353 g = r->objects->commit_graph;
358 first_generation = get_be32(g->chunk_commit_data +
359 g->hash_len + 8) >> 2;
361 return !!first_generation;
364 void close_commit_graph(struct raw_object_store *o)
366 free_commit_graph(o->commit_graph);
367 o->commit_graph = NULL;
370 static int bsearch_graph(struct commit_graph *g, struct object_id *oid, uint32_t *pos)
372 return bsearch_hash(oid->hash, g->chunk_oid_fanout,
373 g->chunk_oid_lookup, g->hash_len, pos);
376 static struct commit_list **insert_parent_or_die(struct repository *r,
377 struct commit_graph *g,
379 struct commit_list **pptr)
382 struct object_id oid;
384 if (pos >= g->num_commits)
385 die("invalid parent position %"PRIu64, pos);
387 hashcpy(oid.hash, g->chunk_oid_lookup + g->hash_len * pos);
388 c = lookup_commit(r, &oid);
390 die(_("could not find commit %s"), oid_to_hex(&oid));
392 return &commit_list_insert(c, pptr)->next;
395 static void fill_commit_graph_info(struct commit *item, struct commit_graph *g, uint32_t pos)
397 const unsigned char *commit_data = g->chunk_commit_data + GRAPH_DATA_WIDTH * pos;
398 item->graph_pos = pos;
399 item->generation = get_be32(commit_data + g->hash_len + 8) >> 2;
402 static inline void set_commit_tree(struct commit *c, struct tree *t)
407 static int fill_commit_in_graph(struct repository *r,
409 struct commit_graph *g, uint32_t pos)
412 uint32_t *parent_data_ptr;
413 uint64_t date_low, date_high;
414 struct commit_list **pptr;
415 const unsigned char *commit_data = g->chunk_commit_data + (g->hash_len + 16) * pos;
417 item->object.parsed = 1;
418 item->graph_pos = pos;
420 set_commit_tree(item, NULL);
422 date_high = get_be32(commit_data + g->hash_len + 8) & 0x3;
423 date_low = get_be32(commit_data + g->hash_len + 12);
424 item->date = (timestamp_t)((date_high << 32) | date_low);
426 item->generation = get_be32(commit_data + g->hash_len + 8) >> 2;
428 pptr = &item->parents;
430 edge_value = get_be32(commit_data + g->hash_len);
431 if (edge_value == GRAPH_PARENT_NONE)
433 pptr = insert_parent_or_die(r, g, edge_value, pptr);
435 edge_value = get_be32(commit_data + g->hash_len + 4);
436 if (edge_value == GRAPH_PARENT_NONE)
438 if (!(edge_value & GRAPH_EXTRA_EDGES_NEEDED)) {
439 pptr = insert_parent_or_die(r, g, edge_value, pptr);
443 parent_data_ptr = (uint32_t*)(g->chunk_extra_edges +
444 4 * (uint64_t)(edge_value & GRAPH_EDGE_LAST_MASK));
446 edge_value = get_be32(parent_data_ptr);
447 pptr = insert_parent_or_die(r, g,
448 edge_value & GRAPH_EDGE_LAST_MASK,
451 } while (!(edge_value & GRAPH_LAST_EDGE));
456 static int find_commit_in_graph(struct commit *item, struct commit_graph *g, uint32_t *pos)
458 if (item->graph_pos != COMMIT_NOT_FROM_GRAPH) {
459 *pos = item->graph_pos;
462 return bsearch_graph(g, &(item->object.oid), pos);
466 static int parse_commit_in_graph_one(struct repository *r,
467 struct commit_graph *g,
472 if (item->object.parsed)
475 if (find_commit_in_graph(item, g, &pos))
476 return fill_commit_in_graph(r, item, g, pos);
481 int parse_commit_in_graph(struct repository *r, struct commit *item)
483 if (!prepare_commit_graph(r))
485 return parse_commit_in_graph_one(r, r->objects->commit_graph, item);
488 void load_commit_graph_info(struct repository *r, struct commit *item)
491 if (!prepare_commit_graph(r))
493 if (find_commit_in_graph(item, r->objects->commit_graph, &pos))
494 fill_commit_graph_info(item, r->objects->commit_graph, pos);
497 static struct tree *load_tree_for_commit(struct repository *r,
498 struct commit_graph *g,
501 struct object_id oid;
502 const unsigned char *commit_data = g->chunk_commit_data +
503 GRAPH_DATA_WIDTH * (c->graph_pos);
505 hashcpy(oid.hash, commit_data);
506 set_commit_tree(c, lookup_tree(r, &oid));
508 return c->maybe_tree;
511 static struct tree *get_commit_tree_in_graph_one(struct repository *r,
512 struct commit_graph *g,
513 const struct commit *c)
516 return c->maybe_tree;
517 if (c->graph_pos == COMMIT_NOT_FROM_GRAPH)
518 BUG("get_commit_tree_in_graph_one called from non-commit-graph commit");
520 return load_tree_for_commit(r, g, (struct commit *)c);
523 struct tree *get_commit_tree_in_graph(struct repository *r, const struct commit *c)
525 return get_commit_tree_in_graph_one(r, r->objects->commit_graph, c);
528 struct packed_commit_list {
529 struct commit **list;
534 struct packed_oid_list {
535 struct object_id *list;
540 struct write_commit_graph_context {
541 struct repository *r;
544 struct packed_oid_list oids;
545 struct packed_commit_list commits;
547 unsigned long approx_nr_objects;
548 struct progress *progress;
550 uint64_t progress_cnt;
555 static void write_graph_chunk_fanout(struct hashfile *f,
556 struct write_commit_graph_context *ctx)
559 struct commit **list = ctx->commits.list;
562 * Write the first-level table (the list is sorted,
563 * but we use a 256-entry lookup to be able to avoid
564 * having to do eight extra binary search iterations).
566 for (i = 0; i < 256; i++) {
567 while (count < ctx->commits.nr) {
568 if ((*list)->object.oid.hash[0] != i)
570 display_progress(ctx->progress, ++ctx->progress_cnt);
575 hashwrite_be32(f, count);
579 static void write_graph_chunk_oids(struct hashfile *f, int hash_len,
580 struct write_commit_graph_context *ctx)
582 struct commit **list = ctx->commits.list;
584 for (count = 0; count < ctx->commits.nr; count++, list++) {
585 display_progress(ctx->progress, ++ctx->progress_cnt);
586 hashwrite(f, (*list)->object.oid.hash, (int)hash_len);
590 static const unsigned char *commit_to_sha1(size_t index, void *table)
592 struct commit **commits = table;
593 return commits[index]->object.oid.hash;
596 static void write_graph_chunk_data(struct hashfile *f, int hash_len,
597 struct write_commit_graph_context *ctx)
599 struct commit **list = ctx->commits.list;
600 struct commit **last = ctx->commits.list + ctx->commits.nr;
601 uint32_t num_extra_edges = 0;
603 while (list < last) {
604 struct commit_list *parent;
606 uint32_t packedDate[2];
607 display_progress(ctx->progress, ++ctx->progress_cnt);
609 parse_commit_no_graph(*list);
610 hashwrite(f, get_commit_tree_oid(*list)->hash, hash_len);
612 parent = (*list)->parents;
615 edge_value = GRAPH_PARENT_NONE;
617 edge_value = sha1_pos(parent->item->object.oid.hash,
623 BUG("missing parent %s for commit %s",
624 oid_to_hex(&parent->item->object.oid),
625 oid_to_hex(&(*list)->object.oid));
628 hashwrite_be32(f, edge_value);
631 parent = parent->next;
634 edge_value = GRAPH_PARENT_NONE;
635 else if (parent->next)
636 edge_value = GRAPH_EXTRA_EDGES_NEEDED | num_extra_edges;
638 edge_value = sha1_pos(parent->item->object.oid.hash,
643 BUG("missing parent %s for commit %s",
644 oid_to_hex(&parent->item->object.oid),
645 oid_to_hex(&(*list)->object.oid));
648 hashwrite_be32(f, edge_value);
650 if (edge_value & GRAPH_EXTRA_EDGES_NEEDED) {
653 parent = parent->next;
657 if (sizeof((*list)->date) > 4)
658 packedDate[0] = htonl(((*list)->date >> 32) & 0x3);
662 packedDate[0] |= htonl((*list)->generation << 2);
664 packedDate[1] = htonl((*list)->date);
665 hashwrite(f, packedDate, 8);
671 static void write_graph_chunk_extra_edges(struct hashfile *f,
672 struct write_commit_graph_context *ctx)
674 struct commit **list = ctx->commits.list;
675 struct commit **last = ctx->commits.list + ctx->commits.nr;
676 struct commit_list *parent;
678 while (list < last) {
681 display_progress(ctx->progress, ++ctx->progress_cnt);
683 for (parent = (*list)->parents; num_parents < 3 && parent;
684 parent = parent->next)
687 if (num_parents <= 2) {
692 /* Since num_parents > 2, this initializer is safe. */
693 for (parent = (*list)->parents->next; parent; parent = parent->next) {
694 int edge_value = sha1_pos(parent->item->object.oid.hash,
700 BUG("missing parent %s for commit %s",
701 oid_to_hex(&parent->item->object.oid),
702 oid_to_hex(&(*list)->object.oid));
703 else if (!parent->next)
704 edge_value |= GRAPH_LAST_EDGE;
706 hashwrite_be32(f, edge_value);
713 static int commit_compare(const void *_a, const void *_b)
715 const struct object_id *a = (const struct object_id *)_a;
716 const struct object_id *b = (const struct object_id *)_b;
720 static int add_packed_commits(const struct object_id *oid,
721 struct packed_git *pack,
725 struct write_commit_graph_context *ctx = (struct write_commit_graph_context*)data;
726 enum object_type type;
727 off_t offset = nth_packed_object_offset(pack, pos);
728 struct object_info oi = OBJECT_INFO_INIT;
731 display_progress(ctx->progress, ++ctx->progress_done);
734 if (packed_object_info(ctx->r, pack, offset, &oi) < 0)
735 die(_("unable to get type of object %s"), oid_to_hex(oid));
737 if (type != OBJ_COMMIT)
740 ALLOC_GROW(ctx->oids.list, ctx->oids.nr + 1, ctx->oids.alloc);
741 oidcpy(&(ctx->oids.list[ctx->oids.nr]), oid);
747 static void add_missing_parents(struct write_commit_graph_context *ctx, struct commit *commit)
749 struct commit_list *parent;
750 for (parent = commit->parents; parent; parent = parent->next) {
751 if (!(parent->item->object.flags & UNINTERESTING)) {
752 ALLOC_GROW(ctx->oids.list, ctx->oids.nr + 1, ctx->oids.alloc);
753 oidcpy(&ctx->oids.list[ctx->oids.nr], &(parent->item->object.oid));
755 parent->item->object.flags |= UNINTERESTING;
760 static void close_reachable(struct write_commit_graph_context *ctx)
763 struct commit *commit;
765 if (ctx->report_progress)
766 ctx->progress = start_delayed_progress(
767 _("Loading known commits in commit graph"),
769 for (i = 0; i < ctx->oids.nr; i++) {
770 display_progress(ctx->progress, i + 1);
771 commit = lookup_commit(ctx->r, &ctx->oids.list[i]);
773 commit->object.flags |= UNINTERESTING;
775 stop_progress(&ctx->progress);
778 * As this loop runs, ctx->oids.nr may grow, but not more
779 * than the number of missing commits in the reachable
782 if (ctx->report_progress)
783 ctx->progress = start_delayed_progress(
784 _("Expanding reachable commits in commit graph"),
786 for (i = 0; i < ctx->oids.nr; i++) {
787 display_progress(ctx->progress, i + 1);
788 commit = lookup_commit(ctx->r, &ctx->oids.list[i]);
790 if (commit && !parse_commit_no_graph(commit))
791 add_missing_parents(ctx, commit);
793 stop_progress(&ctx->progress);
795 if (ctx->report_progress)
796 ctx->progress = start_delayed_progress(
797 _("Clearing commit marks in commit graph"),
799 for (i = 0; i < ctx->oids.nr; i++) {
800 display_progress(ctx->progress, i + 1);
801 commit = lookup_commit(ctx->r, &ctx->oids.list[i]);
804 commit->object.flags &= ~UNINTERESTING;
806 stop_progress(&ctx->progress);
809 static void compute_generation_numbers(struct write_commit_graph_context *ctx)
812 struct commit_list *list = NULL;
814 if (ctx->report_progress)
815 ctx->progress = start_progress(
816 _("Computing commit graph generation numbers"),
818 for (i = 0; i < ctx->commits.nr; i++) {
819 display_progress(ctx->progress, i + 1);
820 if (ctx->commits.list[i]->generation != GENERATION_NUMBER_INFINITY &&
821 ctx->commits.list[i]->generation != GENERATION_NUMBER_ZERO)
824 commit_list_insert(ctx->commits.list[i], &list);
826 struct commit *current = list->item;
827 struct commit_list *parent;
828 int all_parents_computed = 1;
829 uint32_t max_generation = 0;
831 for (parent = current->parents; parent; parent = parent->next) {
832 if (parent->item->generation == GENERATION_NUMBER_INFINITY ||
833 parent->item->generation == GENERATION_NUMBER_ZERO) {
834 all_parents_computed = 0;
835 commit_list_insert(parent->item, &list);
837 } else if (parent->item->generation > max_generation) {
838 max_generation = parent->item->generation;
842 if (all_parents_computed) {
843 current->generation = max_generation + 1;
846 if (current->generation > GENERATION_NUMBER_MAX)
847 current->generation = GENERATION_NUMBER_MAX;
851 stop_progress(&ctx->progress);
854 static int add_ref_to_list(const char *refname,
855 const struct object_id *oid,
856 int flags, void *cb_data)
858 struct string_list *list = (struct string_list *)cb_data;
860 string_list_append(list, oid_to_hex(oid));
864 int write_commit_graph_reachable(const char *obj_dir, unsigned int flags)
866 struct string_list list = STRING_LIST_INIT_DUP;
869 for_each_ref(add_ref_to_list, &list);
870 result = write_commit_graph(obj_dir, NULL, &list,
873 string_list_clear(&list, 0);
877 static int fill_oids_from_packs(struct write_commit_graph_context *ctx,
878 struct string_list *pack_indexes)
881 struct strbuf progress_title = STRBUF_INIT;
882 struct strbuf packname = STRBUF_INIT;
885 strbuf_addf(&packname, "%s/pack/", ctx->obj_dir);
886 dirlen = packname.len;
887 if (ctx->report_progress) {
888 strbuf_addf(&progress_title,
889 Q_("Finding commits for commit graph in %d pack",
890 "Finding commits for commit graph in %d packs",
893 ctx->progress = start_delayed_progress(progress_title.buf, 0);
894 ctx->progress_done = 0;
896 for (i = 0; i < pack_indexes->nr; i++) {
897 struct packed_git *p;
898 strbuf_setlen(&packname, dirlen);
899 strbuf_addstr(&packname, pack_indexes->items[i].string);
900 p = add_packed_git(packname.buf, packname.len, 1);
902 error(_("error adding pack %s"), packname.buf);
905 if (open_pack_index(p)) {
906 error(_("error opening index for %s"), packname.buf);
909 for_each_object_in_pack(p, add_packed_commits, ctx,
910 FOR_EACH_OBJECT_PACK_ORDER);
915 stop_progress(&ctx->progress);
916 strbuf_reset(&progress_title);
917 strbuf_release(&packname);
922 static void fill_oids_from_commit_hex(struct write_commit_graph_context *ctx,
923 struct string_list *commit_hex)
926 struct strbuf progress_title = STRBUF_INIT;
928 if (ctx->report_progress) {
929 strbuf_addf(&progress_title,
930 Q_("Finding commits for commit graph from %d ref",
931 "Finding commits for commit graph from %d refs",
934 ctx->progress = start_delayed_progress(
938 for (i = 0; i < commit_hex->nr; i++) {
940 struct object_id oid;
941 struct commit *result;
943 display_progress(ctx->progress, i + 1);
944 if (commit_hex->items[i].string &&
945 parse_oid_hex(commit_hex->items[i].string, &oid, &end))
948 result = lookup_commit_reference_gently(ctx->r, &oid, 1);
951 ALLOC_GROW(ctx->oids.list, ctx->oids.nr + 1, ctx->oids.alloc);
952 oidcpy(&ctx->oids.list[ctx->oids.nr], &(result->object.oid));
956 stop_progress(&ctx->progress);
957 strbuf_release(&progress_title);
960 static void fill_oids_from_all_packs(struct write_commit_graph_context *ctx)
962 if (ctx->report_progress)
963 ctx->progress = start_delayed_progress(
964 _("Finding commits for commit graph among packed objects"),
965 ctx->approx_nr_objects);
966 for_each_packed_object(add_packed_commits, ctx,
967 FOR_EACH_OBJECT_PACK_ORDER);
968 if (ctx->progress_done < ctx->approx_nr_objects)
969 display_progress(ctx->progress, ctx->approx_nr_objects);
970 stop_progress(&ctx->progress);
973 static uint32_t count_distinct_commits(struct write_commit_graph_context *ctx)
975 uint32_t i, count_distinct = 1;
977 if (ctx->report_progress)
978 ctx->progress = start_delayed_progress(
979 _("Counting distinct commits in commit graph"),
981 display_progress(ctx->progress, 0); /* TODO: Measure QSORT() progress */
982 QSORT(ctx->oids.list, ctx->oids.nr, commit_compare);
984 for (i = 1; i < ctx->oids.nr; i++) {
985 display_progress(ctx->progress, i + 1);
986 if (!oideq(&ctx->oids.list[i - 1], &ctx->oids.list[i]))
989 stop_progress(&ctx->progress);
991 return count_distinct;
994 static void copy_oids_to_commits(struct write_commit_graph_context *ctx)
997 struct commit_list *parent;
999 ctx->num_extra_edges = 0;
1000 if (ctx->report_progress)
1001 ctx->progress = start_delayed_progress(
1002 _("Finding extra edges in commit graph"),
1004 for (i = 0; i < ctx->oids.nr; i++) {
1005 int num_parents = 0;
1006 display_progress(ctx->progress, i + 1);
1007 if (i > 0 && oideq(&ctx->oids.list[i - 1], &ctx->oids.list[i]))
1010 ctx->commits.list[ctx->commits.nr] = lookup_commit(ctx->r, &ctx->oids.list[i]);
1011 parse_commit_no_graph(ctx->commits.list[ctx->commits.nr]);
1013 for (parent = ctx->commits.list[ctx->commits.nr]->parents;
1014 parent; parent = parent->next)
1017 if (num_parents > 2)
1018 ctx->num_extra_edges += num_parents - 1;
1022 stop_progress(&ctx->progress);
1025 static int write_commit_graph_file(struct write_commit_graph_context *ctx)
1029 struct lock_file lk = LOCK_INIT;
1030 uint32_t chunk_ids[5];
1031 uint64_t chunk_offsets[5];
1032 const unsigned hashsz = the_hash_algo->rawsz;
1033 struct strbuf progress_title = STRBUF_INIT;
1034 int num_chunks = ctx->num_extra_edges ? 4 : 3;
1036 ctx->graph_name = get_commit_graph_filename(ctx->obj_dir);
1037 if (safe_create_leading_directories(ctx->graph_name)) {
1038 UNLEAK(ctx->graph_name);
1039 error(_("unable to create leading directories of %s"),
1044 hold_lock_file_for_update(&lk, ctx->graph_name, LOCK_DIE_ON_ERROR);
1045 f = hashfd(lk.tempfile->fd, lk.tempfile->filename.buf);
1047 hashwrite_be32(f, GRAPH_SIGNATURE);
1049 hashwrite_u8(f, GRAPH_VERSION);
1050 hashwrite_u8(f, oid_version());
1051 hashwrite_u8(f, num_chunks);
1052 hashwrite_u8(f, 0); /* unused padding byte */
1054 chunk_ids[0] = GRAPH_CHUNKID_OIDFANOUT;
1055 chunk_ids[1] = GRAPH_CHUNKID_OIDLOOKUP;
1056 chunk_ids[2] = GRAPH_CHUNKID_DATA;
1057 if (ctx->num_extra_edges)
1058 chunk_ids[3] = GRAPH_CHUNKID_EXTRAEDGES;
1063 chunk_offsets[0] = 8 + (num_chunks + 1) * GRAPH_CHUNKLOOKUP_WIDTH;
1064 chunk_offsets[1] = chunk_offsets[0] + GRAPH_FANOUT_SIZE;
1065 chunk_offsets[2] = chunk_offsets[1] + hashsz * ctx->commits.nr;
1066 chunk_offsets[3] = chunk_offsets[2] + (hashsz + 16) * ctx->commits.nr;
1067 chunk_offsets[4] = chunk_offsets[3] + 4 * ctx->num_extra_edges;
1069 for (i = 0; i <= num_chunks; i++) {
1070 uint32_t chunk_write[3];
1072 chunk_write[0] = htonl(chunk_ids[i]);
1073 chunk_write[1] = htonl(chunk_offsets[i] >> 32);
1074 chunk_write[2] = htonl(chunk_offsets[i] & 0xffffffff);
1075 hashwrite(f, chunk_write, 12);
1078 if (ctx->report_progress) {
1079 strbuf_addf(&progress_title,
1080 Q_("Writing out commit graph in %d pass",
1081 "Writing out commit graph in %d passes",
1084 ctx->progress = start_delayed_progress(
1086 num_chunks * ctx->commits.nr);
1088 write_graph_chunk_fanout(f, ctx);
1089 write_graph_chunk_oids(f, hashsz, ctx);
1090 write_graph_chunk_data(f, hashsz, ctx);
1091 if (ctx->num_extra_edges)
1092 write_graph_chunk_extra_edges(f, ctx);
1093 stop_progress(&ctx->progress);
1094 strbuf_release(&progress_title);
1096 close_commit_graph(ctx->r->objects);
1097 finalize_hashfile(f, NULL, CSUM_HASH_IN_STREAM | CSUM_FSYNC);
1098 commit_lock_file(&lk);
1103 int write_commit_graph(const char *obj_dir,
1104 struct string_list *pack_indexes,
1105 struct string_list *commit_hex,
1108 struct write_commit_graph_context *ctx;
1109 uint32_t i, count_distinct = 0;
1112 if (!commit_graph_compatible(the_repository))
1115 ctx = xcalloc(1, sizeof(struct write_commit_graph_context));
1116 ctx->r = the_repository;
1117 ctx->obj_dir = obj_dir;
1118 ctx->append = flags & COMMIT_GRAPH_APPEND ? 1 : 0;
1119 ctx->report_progress = flags & COMMIT_GRAPH_PROGRESS ? 1 : 0;
1121 ctx->approx_nr_objects = approximate_object_count();
1122 ctx->oids.alloc = ctx->approx_nr_objects / 32;
1125 prepare_commit_graph_one(ctx->r, ctx->obj_dir);
1126 if (ctx->r->objects->commit_graph)
1127 ctx->oids.alloc += ctx->r->objects->commit_graph->num_commits;
1130 if (ctx->oids.alloc < 1024)
1131 ctx->oids.alloc = 1024;
1132 ALLOC_ARRAY(ctx->oids.list, ctx->oids.alloc);
1134 if (ctx->append && ctx->r->objects->commit_graph) {
1135 struct commit_graph *g = ctx->r->objects->commit_graph;
1136 for (i = 0; i < g->num_commits; i++) {
1137 const unsigned char *hash = g->chunk_oid_lookup + g->hash_len * i;
1138 hashcpy(ctx->oids.list[ctx->oids.nr++].hash, hash);
1143 if ((res = fill_oids_from_packs(ctx, pack_indexes)))
1148 fill_oids_from_commit_hex(ctx, commit_hex);
1150 if (!pack_indexes && !commit_hex)
1151 fill_oids_from_all_packs(ctx);
1153 close_reachable(ctx);
1155 count_distinct = count_distinct_commits(ctx);
1157 if (count_distinct >= GRAPH_EDGE_LAST_MASK) {
1158 error(_("the commit graph format cannot write %d commits"), count_distinct);
1163 ctx->commits.alloc = count_distinct;
1164 ALLOC_ARRAY(ctx->commits.list, ctx->commits.alloc);
1166 copy_oids_to_commits(ctx);
1168 if (ctx->commits.nr >= GRAPH_EDGE_LAST_MASK) {
1169 error(_("too many commits to write graph"));
1174 compute_generation_numbers(ctx);
1176 res = write_commit_graph_file(ctx);
1179 free(ctx->graph_name);
1180 free(ctx->commits.list);
1181 free(ctx->oids.list);
1187 #define VERIFY_COMMIT_GRAPH_ERROR_HASH 2
1188 static int verify_commit_graph_error;
1190 static void graph_report(const char *fmt, ...)
1194 verify_commit_graph_error = 1;
1196 vfprintf(stderr, fmt, ap);
1197 fprintf(stderr, "\n");
1201 #define GENERATION_ZERO_EXISTS 1
1202 #define GENERATION_NUMBER_EXISTS 2
1204 int verify_commit_graph(struct repository *r, struct commit_graph *g)
1206 uint32_t i, cur_fanout_pos = 0;
1207 struct object_id prev_oid, cur_oid, checksum;
1208 int generation_zero = 0;
1211 struct progress *progress = NULL;
1214 graph_report("no commit-graph file loaded");
1218 verify_commit_graph_error = verify_commit_graph_lite(g);
1219 if (verify_commit_graph_error)
1220 return verify_commit_graph_error;
1222 devnull = open("/dev/null", O_WRONLY);
1223 f = hashfd(devnull, NULL);
1224 hashwrite(f, g->data, g->data_len - g->hash_len);
1225 finalize_hashfile(f, checksum.hash, CSUM_CLOSE);
1226 if (!hasheq(checksum.hash, g->data + g->data_len - g->hash_len)) {
1227 graph_report(_("the commit-graph file has incorrect checksum and is likely corrupt"));
1228 verify_commit_graph_error = VERIFY_COMMIT_GRAPH_ERROR_HASH;
1231 for (i = 0; i < g->num_commits; i++) {
1232 struct commit *graph_commit;
1234 hashcpy(cur_oid.hash, g->chunk_oid_lookup + g->hash_len * i);
1236 if (i && oidcmp(&prev_oid, &cur_oid) >= 0)
1237 graph_report(_("commit-graph has incorrect OID order: %s then %s"),
1238 oid_to_hex(&prev_oid),
1239 oid_to_hex(&cur_oid));
1241 oidcpy(&prev_oid, &cur_oid);
1243 while (cur_oid.hash[0] > cur_fanout_pos) {
1244 uint32_t fanout_value = get_be32(g->chunk_oid_fanout + cur_fanout_pos);
1246 if (i != fanout_value)
1247 graph_report(_("commit-graph has incorrect fanout value: fanout[%d] = %u != %u"),
1248 cur_fanout_pos, fanout_value, i);
1252 graph_commit = lookup_commit(r, &cur_oid);
1253 if (!parse_commit_in_graph_one(r, g, graph_commit))
1254 graph_report(_("failed to parse commit %s from commit-graph"),
1255 oid_to_hex(&cur_oid));
1258 while (cur_fanout_pos < 256) {
1259 uint32_t fanout_value = get_be32(g->chunk_oid_fanout + cur_fanout_pos);
1261 if (g->num_commits != fanout_value)
1262 graph_report(_("commit-graph has incorrect fanout value: fanout[%d] = %u != %u"),
1263 cur_fanout_pos, fanout_value, i);
1268 if (verify_commit_graph_error & ~VERIFY_COMMIT_GRAPH_ERROR_HASH)
1269 return verify_commit_graph_error;
1271 progress = start_progress(_("Verifying commits in commit graph"),
1273 for (i = 0; i < g->num_commits; i++) {
1274 struct commit *graph_commit, *odb_commit;
1275 struct commit_list *graph_parents, *odb_parents;
1276 uint32_t max_generation = 0;
1278 display_progress(progress, i + 1);
1279 hashcpy(cur_oid.hash, g->chunk_oid_lookup + g->hash_len * i);
1281 graph_commit = lookup_commit(r, &cur_oid);
1282 odb_commit = (struct commit *)create_object(r, cur_oid.hash, alloc_commit_node(r));
1283 if (parse_commit_internal(odb_commit, 0, 0)) {
1284 graph_report(_("failed to parse commit %s from object database for commit-graph"),
1285 oid_to_hex(&cur_oid));
1289 if (!oideq(&get_commit_tree_in_graph_one(r, g, graph_commit)->object.oid,
1290 get_commit_tree_oid(odb_commit)))
1291 graph_report(_("root tree OID for commit %s in commit-graph is %s != %s"),
1292 oid_to_hex(&cur_oid),
1293 oid_to_hex(get_commit_tree_oid(graph_commit)),
1294 oid_to_hex(get_commit_tree_oid(odb_commit)));
1296 graph_parents = graph_commit->parents;
1297 odb_parents = odb_commit->parents;
1299 while (graph_parents) {
1300 if (odb_parents == NULL) {
1301 graph_report(_("commit-graph parent list for commit %s is too long"),
1302 oid_to_hex(&cur_oid));
1306 if (!oideq(&graph_parents->item->object.oid, &odb_parents->item->object.oid))
1307 graph_report(_("commit-graph parent for %s is %s != %s"),
1308 oid_to_hex(&cur_oid),
1309 oid_to_hex(&graph_parents->item->object.oid),
1310 oid_to_hex(&odb_parents->item->object.oid));
1312 if (graph_parents->item->generation > max_generation)
1313 max_generation = graph_parents->item->generation;
1315 graph_parents = graph_parents->next;
1316 odb_parents = odb_parents->next;
1319 if (odb_parents != NULL)
1320 graph_report(_("commit-graph parent list for commit %s terminates early"),
1321 oid_to_hex(&cur_oid));
1323 if (!graph_commit->generation) {
1324 if (generation_zero == GENERATION_NUMBER_EXISTS)
1325 graph_report(_("commit-graph has generation number zero for commit %s, but non-zero elsewhere"),
1326 oid_to_hex(&cur_oid));
1327 generation_zero = GENERATION_ZERO_EXISTS;
1328 } else if (generation_zero == GENERATION_ZERO_EXISTS)
1329 graph_report(_("commit-graph has non-zero generation number for commit %s, but zero elsewhere"),
1330 oid_to_hex(&cur_oid));
1332 if (generation_zero == GENERATION_ZERO_EXISTS)
1336 * If one of our parents has generation GENERATION_NUMBER_MAX, then
1337 * our generation is also GENERATION_NUMBER_MAX. Decrement to avoid
1338 * extra logic in the following condition.
1340 if (max_generation == GENERATION_NUMBER_MAX)
1343 if (graph_commit->generation != max_generation + 1)
1344 graph_report(_("commit-graph generation for commit %s is %u != %u"),
1345 oid_to_hex(&cur_oid),
1346 graph_commit->generation,
1347 max_generation + 1);
1349 if (graph_commit->date != odb_commit->date)
1350 graph_report(_("commit date for commit %s in commit-graph is %"PRItime" != %"PRItime),
1351 oid_to_hex(&cur_oid),
1355 stop_progress(&progress);
1357 return verify_commit_graph_error;
1360 void free_commit_graph(struct commit_graph *g)
1364 if (g->graph_fd >= 0) {
1365 munmap((void *)g->data, g->data_len);