diff options
Diffstat (limited to 'builtin/pack-objects.c')
-rw-r--r-- | builtin/pack-objects.c | 289 |
1 files changed, 209 insertions, 80 deletions
diff --git a/builtin/pack-objects.c b/builtin/pack-objects.c index 76ce906946..b1998202fb 100644 --- a/builtin/pack-objects.c +++ b/builtin/pack-objects.c @@ -92,10 +92,11 @@ 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 allow_pack_reuse = 1; static enum { WRITE_BITMAP_FALSE = 0, WRITE_BITMAP_QUIET, @@ -163,7 +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. */ @@ -610,12 +611,12 @@ 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, 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, NULL); + entry = packlist_find(&to_pack, &peeled); if (entry) entry->tagged = 1; } @@ -784,57 +785,185 @@ 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) { - unsigned char buffer[8192]; - off_t to_write, total; - int fd; + if (reused_chunks_nr && reused_chunks[reused_chunks_nr-1].difference == offset) + return; - if (!is_pack_valid(reuse_packfile)) - die(_("packfile is invalid: %s"), reuse_packfile->pack_name); + 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++; +} - fd = git_open(reuse_packfile->pack_name); - if (fd < 0) - die_errno(_("unable to open packfile for reuse: %s"), - reuse_packfile->pack_name); +/* + * 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; + } - if (lseek(fd, sizeof(struct pack_header), SEEK_SET) == -1) - die_errno(_("unable to seek in reused packfile")); + /* + * The first chunk starts at zero, so we can't have gone below + * there. + */ + assert(lo); + return reused_chunks[lo-1].difference; +} - if (reuse_packfile_offset < 0) - reuse_packfile_offset = reuse_packfile->pack_size - the_hash_algo->rawsz; +static void write_reused_pack_one(size_t pos, struct hashfile *out, + struct pack_window **w_curs) +{ + off_t offset, next, cur; + enum object_type type; + unsigned long size; - total = to_write = reuse_packfile_offset - sizeof(struct pack_header); + offset = reuse_packfile->revindex[pos].offset; + next = reuse_packfile->revindex[pos + 1].offset; - while (to_write) { - int read_pack = xread(fd, buffer, sizeof(buffer)); + record_reused_object(offset, offset - hashfile_total(out)); - if (read_pack <= 0) - die_errno(_("unable to read from reused packfile")); + cur = offset; + type = unpack_object_header(reuse_packfile, w_curs, &cur, &size); + assert(type >= 0); - if (read_pack > to_write) - read_pack = to_write; + if (type == OBJ_OFS_DELTA) { + off_t base_offset; + off_t fixup; + + unsigned char header[MAX_PACK_OBJECT_HEADER]; + unsigned len; + + base_offset = get_delta_base(reuse_packfile, w_curs, &cur, type, offset); + assert(base_offset != 0); + + /* Convert to REF_DELTA if we must... */ + if (!allow_ofs_delta) { + int base_pos = find_revindex_position(reuse_packfile, base_offset); + const unsigned char *base_sha1 = + nth_packed_object_sha1(reuse_packfile, + reuse_packfile->revindex[base_pos].nr); + + len = encode_in_pack_object_header(header, sizeof(header), + OBJ_REF_DELTA, size); + hashwrite(out, header, len); + hashwrite(out, base_sha1, 20); + copy_pack_data(out, reuse_packfile, w_curs, cur, next - cur); + return; + } - hashwrite(f, buffer, read_pack); - to_write -= read_pack; + /* 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_( @@ -867,11 +996,9 @@ 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; @@ -996,12 +1123,15 @@ 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, 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; @@ -1141,13 +1271,12 @@ 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) @@ -1171,11 +1300,10 @@ 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)) { @@ -1190,7 +1318,7 @@ 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; } @@ -1199,17 +1327,15 @@ 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; } @@ -1507,7 +1633,7 @@ static int can_reuse_delta(const unsigned char *base_sha1, * First see if we're already sending the base (or it's explicitly in * our "excluded" list). */ - base = packlist_find(&to_pack, &base_oid, NULL); + base = packlist_find(&to_pack, &base_oid); if (base) { if (!in_same_island(&delta->idx.oid, &base->idx.oid)) return 0; @@ -2342,15 +2468,6 @@ 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. @@ -2391,12 +2508,10 @@ 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); @@ -2568,6 +2683,13 @@ 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; @@ -2579,7 +2701,7 @@ 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, NULL)) + if (obj_is_packed(oid)) return; tag = lookup_tag(the_repository, oid); @@ -2603,7 +2725,7 @@ 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, NULL)) /* object packed? */ + obj_is_packed(&peeled)) /* object packed? */ add_tag_chain(oid); return 0; } @@ -2671,6 +2793,7 @@ 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); @@ -2715,8 +2838,8 @@ 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")) { @@ -2803,7 +2926,7 @@ 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, 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); } @@ -3034,7 +3157,7 @@ static void loosen_unused_packed_objects(void) for (i = 0; i < p->num_objects; i++) { nth_packed_object_oid(&oid, p, i); - if (!packlist_find(&to_pack, &oid, NULL) && + 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)) @@ -3050,8 +3173,8 @@ static void loosen_unused_packed_objects(void) */ 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) && @@ -3068,7 +3191,7 @@ 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); @@ -3343,6 +3466,10 @@ 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); @@ -3525,7 +3652,9 @@ 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; } |