diff options
Diffstat (limited to 'builtin/pack-objects.c')
-rw-r--r-- | builtin/pack-objects.c | 369 |
1 files changed, 257 insertions, 112 deletions
diff --git a/builtin/pack-objects.c b/builtin/pack-objects.c index 787ae10288..02aa6ee480 100644 --- a/builtin/pack-objects.c +++ b/builtin/pack-objects.c @@ -92,11 +92,16 @@ static struct progress *progress_state; static struct packed_git *reuse_packfile; static uint32_t reuse_packfile_objects; -static off_t reuse_packfile_offset; +static struct bitmap *reuse_packfile_bitmap; static int use_bitmap_index_default = 1; static int use_bitmap_index = -1; -static int write_bitmap_index; +static int allow_pack_reuse = 1; +static enum { + WRITE_BITMAP_FALSE = 0, + WRITE_BITMAP_QUIET, + WRITE_BITMAP_TRUE, +} write_bitmap_index; static uint16_t write_bitmap_options = BITMAP_OPT_HASH_CACHE; static int exclude_promisor_objects; @@ -159,7 +164,7 @@ static void *get_delta(struct object_entry *entry) delta_buf = diff_delta(base_buf, base_size, buf, size, &delta_size, 0); /* - * We succesfully computed this delta once but dropped it for + * We successfully computed this delta once but dropped it for * memory reasons. Something is very wrong if this time we * recompute and create a different delta. */ @@ -299,7 +304,8 @@ static unsigned long write_no_reuse_object(struct hashfile *f, struct object_ent if (!usable_delta) { if (oe_type(entry) == OBJ_BLOB && oe_size_greater_than(&to_pack, entry, big_file_threshold) && - (st = open_istream(&entry->idx.oid, &type, &size, NULL)) != NULL) + (st = open_istream(the_repository, &entry->idx.oid, &type, + &size, NULL)) != NULL) buf = NULL; else { buf = read_object_file(&entry->idx.oid, &type, &size); @@ -606,12 +612,12 @@ static int mark_tagged(const char *path, const struct object_id *oid, int flag, void *cb_data) { struct object_id peeled; - struct object_entry *entry = packlist_find(&to_pack, oid->hash, NULL); + struct object_entry *entry = packlist_find(&to_pack, oid); if (entry) entry->tagged = 1; if (!peel_ref(path, &peeled)) { - entry = packlist_find(&to_pack, peeled.hash, NULL); + entry = packlist_find(&to_pack, &peeled); if (entry) entry->tagged = 1; } @@ -780,57 +786,186 @@ static struct object_entry **compute_write_order(void) return wo; } -static off_t write_reused_pack(struct hashfile *f) + +/* + * A reused set of objects. All objects in a chunk have the same + * relative position in the original packfile and the generated + * packfile. + */ + +static struct reused_chunk { + /* The offset of the first object of this chunk in the original + * packfile. */ + off_t original; + /* The offset of the first object of this chunk in the generated + * packfile minus "original". */ + off_t difference; +} *reused_chunks; +static int reused_chunks_nr; +static int reused_chunks_alloc; + +static void record_reused_object(off_t where, off_t offset) +{ + if (reused_chunks_nr && reused_chunks[reused_chunks_nr-1].difference == offset) + return; + + ALLOC_GROW(reused_chunks, reused_chunks_nr + 1, + reused_chunks_alloc); + reused_chunks[reused_chunks_nr].original = where; + reused_chunks[reused_chunks_nr].difference = offset; + reused_chunks_nr++; +} + +/* + * Binary search to find the chunk that "where" is in. Note + * that we're not looking for an exact match, just the first + * chunk that contains it (which implicitly ends at the start + * of the next chunk. + */ +static off_t find_reused_offset(off_t where) +{ + int lo = 0, hi = reused_chunks_nr; + while (lo < hi) { + int mi = lo + ((hi - lo) / 2); + if (where == reused_chunks[mi].original) + return reused_chunks[mi].difference; + if (where < reused_chunks[mi].original) + hi = mi; + else + lo = mi + 1; + } + + /* + * The first chunk starts at zero, so we can't have gone below + * there. + */ + assert(lo); + return reused_chunks[lo-1].difference; +} + +static void write_reused_pack_one(size_t pos, struct hashfile *out, + struct pack_window **w_curs) { - unsigned char buffer[8192]; - off_t to_write, total; - int fd; + off_t offset, next, cur; + enum object_type type; + unsigned long size; - if (!is_pack_valid(reuse_packfile)) - die(_("packfile is invalid: %s"), reuse_packfile->pack_name); + offset = reuse_packfile->revindex[pos].offset; + next = reuse_packfile->revindex[pos + 1].offset; - fd = git_open(reuse_packfile->pack_name); - if (fd < 0) - die_errno(_("unable to open packfile for reuse: %s"), - reuse_packfile->pack_name); + record_reused_object(offset, offset - hashfile_total(out)); - if (lseek(fd, sizeof(struct pack_header), SEEK_SET) == -1) - die_errno(_("unable to seek in reused packfile")); + cur = offset; + type = unpack_object_header(reuse_packfile, w_curs, &cur, &size); + assert(type >= 0); - if (reuse_packfile_offset < 0) - reuse_packfile_offset = reuse_packfile->pack_size - the_hash_algo->rawsz; + if (type == OBJ_OFS_DELTA) { + off_t base_offset; + off_t fixup; - total = to_write = reuse_packfile_offset - sizeof(struct pack_header); + unsigned char header[MAX_PACK_OBJECT_HEADER]; + unsigned len; - while (to_write) { - int read_pack = xread(fd, buffer, sizeof(buffer)); + base_offset = get_delta_base(reuse_packfile, w_curs, &cur, type, offset); + assert(base_offset != 0); - if (read_pack <= 0) - die_errno(_("unable to read from reused packfile")); + /* Convert to REF_DELTA if we must... */ + if (!allow_ofs_delta) { + int base_pos = find_revindex_position(reuse_packfile, base_offset); + struct object_id base_oid; - if (read_pack > to_write) - read_pack = to_write; + nth_packed_object_id(&base_oid, reuse_packfile, + reuse_packfile->revindex[base_pos].nr); - hashwrite(f, buffer, read_pack); - to_write -= read_pack; + len = encode_in_pack_object_header(header, sizeof(header), + OBJ_REF_DELTA, size); + hashwrite(out, header, len); + hashwrite(out, base_oid.hash, 20); + copy_pack_data(out, reuse_packfile, w_curs, cur, next - cur); + return; + } + + /* Otherwise see if we need to rewrite the offset... */ + fixup = find_reused_offset(offset) - + find_reused_offset(base_offset); + if (fixup) { + unsigned char ofs_header[10]; + unsigned i, ofs_len; + off_t ofs = offset - base_offset - fixup; + + len = encode_in_pack_object_header(header, sizeof(header), + OBJ_OFS_DELTA, size); + + i = sizeof(ofs_header) - 1; + ofs_header[i] = ofs & 127; + while (ofs >>= 7) + ofs_header[--i] = 128 | (--ofs & 127); + + ofs_len = sizeof(ofs_header) - i; + + hashwrite(out, header, len); + hashwrite(out, ofs_header + sizeof(ofs_header) - ofs_len, ofs_len); + copy_pack_data(out, reuse_packfile, w_curs, cur, next - cur); + return; + } + + /* ...otherwise we have no fixup, and can write it verbatim */ + } + + copy_pack_data(out, reuse_packfile, w_curs, offset, next - offset); +} + +static size_t write_reused_pack_verbatim(struct hashfile *out, + struct pack_window **w_curs) +{ + size_t pos = 0; + + while (pos < reuse_packfile_bitmap->word_alloc && + reuse_packfile_bitmap->words[pos] == (eword_t)~0) + pos++; + + if (pos) { + off_t to_write; + + written = (pos * BITS_IN_EWORD); + to_write = reuse_packfile->revindex[written].offset + - sizeof(struct pack_header); + + /* We're recording one chunk, not one object. */ + record_reused_object(sizeof(struct pack_header), 0); + hashflush(out); + copy_pack_data(out, reuse_packfile, w_curs, + sizeof(struct pack_header), to_write); - /* - * We don't know the actual number of objects written, - * only how many bytes written, how many bytes total, and - * how many objects total. So we can fake it by pretending all - * objects we are writing are the same size. This gives us a - * smooth progress meter, and at the end it matches the true - * answer. - */ - written = reuse_packfile_objects * - (((double)(total - to_write)) / total); display_progress(progress_state, written); } + return pos; +} + +static void write_reused_pack(struct hashfile *f) +{ + size_t i = 0; + uint32_t offset; + struct pack_window *w_curs = NULL; + + if (allow_ofs_delta) + i = write_reused_pack_verbatim(f, &w_curs); + + for (; i < reuse_packfile_bitmap->word_alloc; ++i) { + eword_t word = reuse_packfile_bitmap->words[i]; + size_t pos = (i * BITS_IN_EWORD); + + for (offset = 0; offset < BITS_IN_EWORD; ++offset) { + if ((word >> offset) == 0) + break; + + offset += ewah_bit_ctz64(word >> offset); + write_reused_pack_one(pos + offset, f, &w_curs); + display_progress(progress_state, ++written); + } + } - close(fd); - written = reuse_packfile_objects; - display_progress(progress_state, written); - return reuse_packfile_offset - sizeof(struct pack_header); + unuse_pack(&w_curs); } static const char no_split_warning[] = N_( @@ -863,11 +998,9 @@ static void write_pack_file(void) offset = write_pack_header(f, nr_remaining); if (reuse_packfile) { - off_t packfile_size; assert(pack_to_stdout); - - packfile_size = write_reused_pack(f); - offset += packfile_size; + write_reused_pack(f); + offset = hashfile_total(f); } nr_written = 0; @@ -892,7 +1025,8 @@ static void write_pack_file(void) nr_written, oid.hash, offset); close(fd); if (write_bitmap_index) { - warning(_(no_split_warning)); + if (write_bitmap_index != WRITE_BITMAP_QUIET) + warning(_(no_split_warning)); write_bitmap_index = 0; } } @@ -991,12 +1125,15 @@ static int no_try_delta(const char *path) * few lines later when we want to add the new entry. */ static int have_duplicate_entry(const struct object_id *oid, - int exclude, - uint32_t *index_pos) + int exclude) { struct object_entry *entry; - entry = packlist_find(&to_pack, oid->hash, index_pos); + if (reuse_packfile_bitmap && + bitmap_walk_contains(bitmap_git, reuse_packfile_bitmap, oid)) + return 1; + + entry = packlist_find(&to_pack, oid); if (!entry) return 0; @@ -1136,13 +1273,12 @@ static void create_object_entry(const struct object_id *oid, uint32_t hash, int exclude, int no_try_delta, - uint32_t index_pos, struct packed_git *found_pack, off_t found_offset) { struct object_entry *entry; - entry = packlist_alloc(&to_pack, oid->hash, index_pos); + entry = packlist_alloc(&to_pack, oid); entry->hash = hash; oe_set_type(entry, type); if (exclude) @@ -1166,17 +1302,17 @@ static int add_object_entry(const struct object_id *oid, enum object_type type, { struct packed_git *found_pack = NULL; off_t found_offset = 0; - uint32_t index_pos; display_progress(progress_state, ++nr_seen); - if (have_duplicate_entry(oid, exclude, &index_pos)) + if (have_duplicate_entry(oid, exclude)) return 0; if (!want_object_in_pack(oid, exclude, &found_pack, &found_offset)) { /* The pack is missing an object, so it will not have closure */ if (write_bitmap_index) { - warning(_(no_closure_warning)); + if (write_bitmap_index != WRITE_BITMAP_QUIET) + warning(_(no_closure_warning)); write_bitmap_index = 0; } return 0; @@ -1184,7 +1320,7 @@ static int add_object_entry(const struct object_id *oid, enum object_type type, create_object_entry(oid, type, pack_name_hash(name), exclude, name && no_try_delta(name), - index_pos, found_pack, found_offset); + found_pack, found_offset); return 1; } @@ -1193,17 +1329,15 @@ static int add_object_entry_from_bitmap(const struct object_id *oid, int flags, uint32_t name_hash, struct packed_git *pack, off_t offset) { - uint32_t index_pos; - display_progress(progress_state, ++nr_seen); - if (have_duplicate_entry(oid, 0, &index_pos)) + if (have_duplicate_entry(oid, 0)) return 0; if (!want_object_in_pack(oid, 0, &pack, &offset)) return 0; - create_object_entry(oid, type, name_hash, 0, 0, index_pos, pack, offset); + create_object_entry(oid, type, name_hash, 0, 0, pack, offset); return 1; } @@ -1428,7 +1562,8 @@ static void add_preferred_base(struct object_id *oid) if (window <= num_preferred_base++) return; - data = read_object_with_reference(oid, tree_type, &size, &tree_oid); + data = read_object_with_reference(the_repository, oid, + tree_type, &size, &tree_oid); if (!data) return; @@ -1484,21 +1619,17 @@ static void cleanup_preferred_base(void) * deltify other objects against, in order to avoid * circular deltas. */ -static int can_reuse_delta(const unsigned char *base_sha1, +static int can_reuse_delta(const struct object_id *base_oid, struct object_entry *delta, struct object_entry **base_out) { struct object_entry *base; - struct object_id base_oid; - - if (!base_sha1) - return 0; /* * First see if we're already sending the base (or it's explicitly in * our "excluded" list). */ - base = packlist_find(&to_pack, base_sha1, NULL); + base = packlist_find(&to_pack, base_oid); if (base) { if (!in_same_island(&delta->idx.oid, &base->idx.oid)) return 0; @@ -1511,10 +1642,9 @@ static int can_reuse_delta(const unsigned char *base_sha1, * even if it was buried too deep in history to make it into the * packing list. */ - oidread(&base_oid, base_sha1); - if (thin && bitmap_has_oid_in_uninteresting(bitmap_git, &base_oid)) { + if (thin && bitmap_has_oid_in_uninteresting(bitmap_git, base_oid)) { if (use_delta_islands) { - if (!in_same_island(&delta->idx.oid, &base_oid)) + if (!in_same_island(&delta->idx.oid, base_oid)) return 0; } *base_out = NULL; @@ -1531,7 +1661,8 @@ static void check_object(struct object_entry *entry) if (IN_PACK(entry)) { struct packed_git *p = IN_PACK(entry); struct pack_window *w_curs = NULL; - const unsigned char *base_ref = NULL; + int have_base = 0; + struct object_id base_ref; struct object_entry *base_entry; unsigned long used, used_0; unsigned long avail; @@ -1572,9 +1703,13 @@ static void check_object(struct object_entry *entry) unuse_pack(&w_curs); return; case OBJ_REF_DELTA: - if (reuse_delta && !entry->preferred_base) - base_ref = use_pack(p, &w_curs, - entry->in_pack_offset + used, NULL); + if (reuse_delta && !entry->preferred_base) { + oidread(&base_ref, + use_pack(p, &w_curs, + entry->in_pack_offset + used, + NULL)); + have_base = 1; + } entry->in_pack_header_size = used + the_hash_algo->rawsz; break; case OBJ_OFS_DELTA: @@ -1604,13 +1739,15 @@ static void check_object(struct object_entry *entry) revidx = find_pack_revindex(p, ofs); if (!revidx) goto give_up; - base_ref = nth_packed_object_sha1(p, revidx->nr); + if (!nth_packed_object_id(&base_ref, p, revidx->nr)) + have_base = 1; } entry->in_pack_header_size = used + used_0; break; } - if (can_reuse_delta(base_ref, entry, &base_entry)) { + if (have_base && + can_reuse_delta(&base_ref, entry, &base_entry)) { oe_set_type(entry, entry->in_pack_type); SET_SIZE(entry, in_pack_size); /* delta size */ SET_DELTA_SIZE(entry, in_pack_size); @@ -1620,7 +1757,7 @@ static void check_object(struct object_entry *entry) entry->delta_sibling_idx = base_entry->delta_child_idx; SET_DELTA_CHILD(base_entry, entry); } else { - SET_DELTA_EXT(entry, base_ref); + SET_DELTA_EXT(entry, &base_ref); } unuse_pack(&w_curs); @@ -2334,15 +2471,6 @@ static void find_deltas(struct object_entry **list, unsigned *list_size, free(array); } -static void try_to_free_from_threads(size_t size) -{ - packing_data_lock(&to_pack); - release_pack_memory(size); - packing_data_unlock(&to_pack); -} - -static try_to_free_t old_try_to_free_routine; - /* * The main object list is split into smaller lists, each is handed to * one worker. @@ -2383,12 +2511,10 @@ static void init_threaded_search(void) pthread_mutex_init(&cache_mutex, NULL); pthread_mutex_init(&progress_mutex, NULL); pthread_cond_init(&progress_cond, NULL); - old_try_to_free_routine = set_try_to_free_routine(try_to_free_from_threads); } static void cleanup_threaded_search(void) { - set_try_to_free_routine(old_try_to_free_routine); pthread_cond_destroy(&progress_cond); pthread_mutex_destroy(&cache_mutex); pthread_mutex_destroy(&progress_mutex); @@ -2560,6 +2686,13 @@ static void ll_find_deltas(struct object_entry **list, unsigned list_size, free(p); } +static int obj_is_packed(const struct object_id *oid) +{ + return packlist_find(&to_pack, oid) || + (reuse_packfile_bitmap && + bitmap_walk_contains(bitmap_git, reuse_packfile_bitmap, oid)); +} + static void add_tag_chain(const struct object_id *oid) { struct tag *tag; @@ -2571,7 +2704,7 @@ static void add_tag_chain(const struct object_id *oid) * it was included via bitmaps, we would not have parsed it * previously). */ - if (packlist_find(&to_pack, oid->hash, NULL)) + if (obj_is_packed(oid)) return; tag = lookup_tag(the_repository, oid); @@ -2595,7 +2728,7 @@ static int add_ref_tag(const char *path, const struct object_id *oid, int flag, if (starts_with(path, "refs/tags/") && /* is a tag? */ !peel_ref(path, &peeled) && /* peelable? */ - packlist_find(&to_pack, peeled.hash, NULL)) /* object packed? */ + obj_is_packed(&peeled)) /* object packed? */ add_tag_chain(oid); return 0; } @@ -2663,6 +2796,7 @@ static void prepare_pack(int window, int depth) if (nr_deltas && n > 1) { unsigned nr_done = 0; + if (progress) progress_state = start_progress(_("Compressing objects"), nr_deltas); @@ -2707,8 +2841,8 @@ static int git_pack_config(const char *k, const char *v, void *cb) use_bitmap_index_default = git_config_bool(k, v); return 0; } - if (!strcmp(k, "pack.usesparse")) { - sparse = git_config_bool(k, v); + if (!strcmp(k, "pack.allowpackreuse")) { + allow_pack_reuse = git_config_bool(k, v); return 0; } if (!strcmp(k, "pack.threads")) { @@ -2795,7 +2929,7 @@ static void show_object(struct object *obj, const char *name, void *data) for (p = strchr(name, '/'); p; p = strchr(p + 1, '/')) depth++; - ent = packlist_find(&to_pack, obj->oid.hash, NULL); + ent = packlist_find(&to_pack, &obj->oid); if (ent && depth > oe_tree_depth(&to_pack, ent)) oe_set_tree_depth(&to_pack, ent, depth); } @@ -2899,7 +3033,7 @@ static int ofscmp(const void *a_, const void *b_) return oidcmp(&a->object->oid, &b->object->oid); } -static void add_objects_in_unpacked_packs(struct rev_info *revs) +static void add_objects_in_unpacked_packs(void) { struct packed_git *p; struct in_pack in_pack; @@ -2921,8 +3055,8 @@ static void add_objects_in_unpacked_packs(struct rev_info *revs) in_pack.alloc); for (i = 0; i < p->num_objects; i++) { - nth_packed_object_oid(&oid, p, i); - o = lookup_unknown_object(oid.hash); + nth_packed_object_id(&oid, p, i); + o = lookup_unknown_object(&oid); if (!(o->flags & OBJECT_ADDED)) mark_in_pack_object(o, p, &in_pack); o->flags |= OBJECT_ADDED; @@ -3011,7 +3145,7 @@ static int loosened_object_can_be_discarded(const struct object_id *oid, return 1; } -static void loosen_unused_packed_objects(struct rev_info *revs) +static void loosen_unused_packed_objects(void) { struct packed_git *p; uint32_t i; @@ -3025,8 +3159,8 @@ static void loosen_unused_packed_objects(struct rev_info *revs) die(_("cannot open pack index")); for (i = 0; i < p->num_objects; i++) { - nth_packed_object_oid(&oid, p, i); - if (!packlist_find(&to_pack, oid.hash, NULL) && + nth_packed_object_id(&oid, p, i); + if (!packlist_find(&to_pack, &oid) && !has_sha1_pack_kept_or_nonlocal(&oid) && !loosened_object_can_be_discarded(&oid, p->mtime)) if (force_object_loose(&oid, p->mtime)) @@ -3042,8 +3176,8 @@ static void loosen_unused_packed_objects(struct rev_info *revs) */ static int pack_options_allow_reuse(void) { - return pack_to_stdout && - allow_ofs_delta && + return allow_pack_reuse && + pack_to_stdout && !ignore_packed_keep_on_disk && !ignore_packed_keep_in_core && (!local || !have_non_local_packs) && @@ -3052,7 +3186,7 @@ static int pack_options_allow_reuse(void) static int get_object_list_from_bitmap(struct rev_info *revs) { - if (!(bitmap_git = prepare_bitmap_walk(revs))) + if (!(bitmap_git = prepare_bitmap_walk(revs, &filter_options))) return -1; if (pack_options_allow_reuse() && @@ -3060,13 +3194,14 @@ static int get_object_list_from_bitmap(struct rev_info *revs) bitmap_git, &reuse_packfile, &reuse_packfile_objects, - &reuse_packfile_offset)) { + &reuse_packfile_bitmap)) { assert(reuse_packfile_objects); nr_result += reuse_packfile_objects; display_progress(progress_state, nr_result); } - traverse_bitmap_commit_list(bitmap_git, &add_object_entry_from_bitmap); + traverse_bitmap_commit_list(bitmap_git, revs, + &add_object_entry_from_bitmap); return 0; } @@ -3158,11 +3293,11 @@ static void get_object_list(int ac, const char **av) } if (keep_unreachable) - add_objects_in_unpacked_packs(&revs); + add_objects_in_unpacked_packs(); if (pack_loose_unreachable) add_unreachable_loose_objects(); if (unpack_unreachable) - loosen_unused_packed_objects(&revs); + loosen_unused_packed_objects(); oid_array_clear(&recent_objects); } @@ -3311,8 +3446,13 @@ int cmd_pack_objects(int argc, const char **argv, const char *prefix) N_("do not hide commits by grafts"), 0), OPT_BOOL(0, "use-bitmap-index", &use_bitmap_index, N_("use a bitmap index if available to speed up counting objects")), - OPT_BOOL(0, "write-bitmap-index", &write_bitmap_index, - N_("write a bitmap index together with the pack index")), + OPT_SET_INT(0, "write-bitmap-index", &write_bitmap_index, + N_("write a bitmap index together with the pack index"), + WRITE_BITMAP_TRUE), + OPT_SET_INT_F(0, "write-bitmap-index-quiet", + &write_bitmap_index, + N_("write a bitmap index if possible"), + WRITE_BITMAP_QUIET, PARSE_OPT_HIDDEN), OPT_PARSE_LIST_OBJECTS_FILTER(&filter_options), { OPTION_CALLBACK, 0, "missing", NULL, N_("action"), N_("handling for missing objects"), PARSE_OPT_NONEG, @@ -3330,6 +3470,10 @@ int cmd_pack_objects(int argc, const char **argv, const char *prefix) read_replace_refs = 0; sparse = git_env_bool("GIT_TEST_PACK_SPARSE", 0); + prepare_repo_settings(the_repository); + if (!sparse && the_repository->settings.pack_use_sparse != -1) + sparse = the_repository->settings.pack_use_sparse; + reset_pack_idx_option(&pack_idx_opts); git_config(git_pack_config, NULL); @@ -3421,7 +3565,6 @@ int cmd_pack_objects(int argc, const char **argv, const char *prefix) if (filter_options.choice) { if (!pack_to_stdout) die(_("cannot use --filter without --stdout")); - use_bitmap_index = 0; } /* @@ -3512,7 +3655,9 @@ int cmd_pack_objects(int argc, const char **argv, const char *prefix) if (progress) fprintf_ln(stderr, _("Total %"PRIu32" (delta %"PRIu32")," - " reused %"PRIu32" (delta %"PRIu32")"), - written, written_delta, reused, reused_delta); + " reused %"PRIu32" (delta %"PRIu32")," + " pack-reused %"PRIu32), + written, written_delta, reused, reused_delta, + reuse_packfile_objects); return 0; } |