]>
Commit | Line | Data |
---|---|---|
8b0eca7c DH |
1 | #include "cache.h" |
2 | #include "pack.h" | |
aa7e44bf GB |
3 | #include "csum-file.h" |
4 | ||
ebcfb379 JH |
5 | void reset_pack_idx_option(struct pack_idx_option *opts) |
6 | { | |
7 | memset(opts, 0, sizeof(*opts)); | |
8 | opts->version = 2; | |
9 | opts->off32_limit = 0x7fffffff; | |
10 | } | |
aa7e44bf GB |
11 | |
12 | static int sha1_compare(const void *_a, const void *_b) | |
13 | { | |
14 | struct pack_idx_entry *a = *(struct pack_idx_entry **)_a; | |
15 | struct pack_idx_entry *b = *(struct pack_idx_entry **)_b; | |
e6a492b7 | 16 | return oidcmp(&a->oid, &b->oid); |
aa7e44bf GB |
17 | } |
18 | ||
3c9fc074 JH |
19 | static int cmp_uint32(const void *a_, const void *b_) |
20 | { | |
21 | uint32_t a = *((uint32_t *)a_); | |
22 | uint32_t b = *((uint32_t *)b_); | |
23 | ||
24 | return (a < b) ? -1 : (a != b); | |
25 | } | |
26 | ||
fb956c1f JH |
27 | static int need_large_offset(off_t offset, const struct pack_idx_option *opts) |
28 | { | |
3c9fc074 JH |
29 | uint32_t ofsval; |
30 | ||
31 | if ((offset >> 31) || (opts->off32_limit < offset)) | |
32 | return 1; | |
33 | if (!opts->anomaly_nr) | |
34 | return 0; | |
35 | ofsval = offset; | |
36 | return !!bsearch(&ofsval, opts->anomaly, opts->anomaly_nr, | |
37 | sizeof(ofsval), cmp_uint32); | |
fb956c1f JH |
38 | } |
39 | ||
aa7e44bf | 40 | /* |
e2bfa50a JB |
41 | * The *sha1 contains the pack content SHA1 hash. |
42 | * The objects array passed in will be sorted by SHA1 on exit. | |
aa7e44bf | 43 | */ |
3bb72562 | 44 | const char *write_idx_file(const char *index_name, struct pack_idx_entry **objects, |
ebcfb379 | 45 | int nr_objects, const struct pack_idx_option *opts, |
1190a1ac | 46 | const unsigned char *sha1) |
aa7e44bf | 47 | { |
98a3beab | 48 | struct hashfile *f; |
aa7e44bf GB |
49 | struct pack_idx_entry **sorted_by_sha, **list, **last; |
50 | off_t last_obj_offset = 0; | |
51 | uint32_t array[256]; | |
52 | int i, fd; | |
aa7e44bf GB |
53 | uint32_t index_version; |
54 | ||
55 | if (nr_objects) { | |
56 | sorted_by_sha = objects; | |
57 | list = sorted_by_sha; | |
58 | last = sorted_by_sha + nr_objects; | |
59 | for (i = 0; i < nr_objects; ++i) { | |
60 | if (objects[i]->offset > last_obj_offset) | |
61 | last_obj_offset = objects[i]->offset; | |
62 | } | |
9ed0d8d6 | 63 | QSORT(sorted_by_sha, nr_objects, sha1_compare); |
aa7e44bf GB |
64 | } |
65 | else | |
66 | sorted_by_sha = list = last = NULL; | |
67 | ||
e337a04d JH |
68 | if (opts->flags & WRITE_IDX_VERIFY) { |
69 | assert(index_name); | |
98a3beab | 70 | f = hashfd_check(index_name); |
aa7e44bf | 71 | } else { |
e337a04d | 72 | if (!index_name) { |
594fa999 JK |
73 | struct strbuf tmp_file = STRBUF_INIT; |
74 | fd = odb_mkstemp(&tmp_file, "pack/tmp_idx_XXXXXX"); | |
75 | index_name = strbuf_detach(&tmp_file, NULL); | |
e337a04d JH |
76 | } else { |
77 | unlink(index_name); | |
78 | fd = open(index_name, O_CREAT|O_EXCL|O_WRONLY, 0600); | |
892e723a JK |
79 | if (fd < 0) |
80 | die_errno("unable to create '%s'", index_name); | |
e337a04d | 81 | } |
98a3beab | 82 | f = hashfd(fd, index_name); |
aa7e44bf | 83 | } |
aa7e44bf GB |
84 | |
85 | /* if last object's offset is >= 2^31 we should use index V2 */ | |
fb956c1f | 86 | index_version = need_large_offset(last_obj_offset, opts) ? 2 : opts->version; |
aa7e44bf GB |
87 | |
88 | /* index versions 2 and above need a header */ | |
89 | if (index_version >= 2) { | |
90 | struct pack_idx_header hdr; | |
91 | hdr.idx_signature = htonl(PACK_IDX_SIGNATURE); | |
92 | hdr.idx_version = htonl(index_version); | |
98a3beab | 93 | hashwrite(f, &hdr, sizeof(hdr)); |
aa7e44bf GB |
94 | } |
95 | ||
96 | /* | |
97 | * Write the first-level table (the list is sorted, | |
98 | * but we use a 256-entry lookup to be able to avoid | |
99 | * having to do eight extra binary search iterations). | |
100 | */ | |
101 | for (i = 0; i < 256; i++) { | |
102 | struct pack_idx_entry **next = list; | |
103 | while (next < last) { | |
104 | struct pack_idx_entry *obj = *next; | |
e6a492b7 | 105 | if (obj->oid.hash[0] != i) |
aa7e44bf GB |
106 | break; |
107 | next++; | |
108 | } | |
109 | array[i] = htonl(next - sorted_by_sha); | |
110 | list = next; | |
111 | } | |
98a3beab | 112 | hashwrite(f, array, 256 * 4); |
aa7e44bf | 113 | |
aa7e44bf GB |
114 | /* |
115 | * Write the actual SHA1 entries.. | |
116 | */ | |
117 | list = sorted_by_sha; | |
118 | for (i = 0; i < nr_objects; i++) { | |
119 | struct pack_idx_entry *obj = *list++; | |
120 | if (index_version < 2) { | |
121 | uint32_t offset = htonl(obj->offset); | |
98a3beab | 122 | hashwrite(f, &offset, 4); |
aa7e44bf | 123 | } |
98a3beab | 124 | hashwrite(f, obj->oid.hash, the_hash_algo->rawsz); |
68be2fea | 125 | if ((opts->flags & WRITE_IDX_STRICT) && |
4a7e27e9 | 126 | (i && oideq(&list[-2]->oid, &obj->oid))) |
68be2fea | 127 | die("The same object %s appears twice in the pack", |
e6a492b7 | 128 | oid_to_hex(&obj->oid)); |
aa7e44bf GB |
129 | } |
130 | ||
131 | if (index_version >= 2) { | |
132 | unsigned int nr_large_offset = 0; | |
133 | ||
134 | /* write the crc32 table */ | |
135 | list = sorted_by_sha; | |
136 | for (i = 0; i < nr_objects; i++) { | |
137 | struct pack_idx_entry *obj = *list++; | |
138 | uint32_t crc32_val = htonl(obj->crc32); | |
98a3beab | 139 | hashwrite(f, &crc32_val, 4); |
aa7e44bf GB |
140 | } |
141 | ||
142 | /* write the 32-bit offset table */ | |
143 | list = sorted_by_sha; | |
144 | for (i = 0; i < nr_objects; i++) { | |
145 | struct pack_idx_entry *obj = *list++; | |
fb956c1f JH |
146 | uint32_t offset; |
147 | ||
148 | offset = (need_large_offset(obj->offset, opts) | |
149 | ? (0x80000000 | nr_large_offset++) | |
150 | : obj->offset); | |
aa7e44bf | 151 | offset = htonl(offset); |
98a3beab | 152 | hashwrite(f, &offset, 4); |
aa7e44bf GB |
153 | } |
154 | ||
155 | /* write the large offset table */ | |
156 | list = sorted_by_sha; | |
157 | while (nr_large_offset) { | |
158 | struct pack_idx_entry *obj = *list++; | |
159 | uint64_t offset = obj->offset; | |
fb956c1f JH |
160 | uint32_t split[2]; |
161 | ||
162 | if (!need_large_offset(offset, opts)) | |
163 | continue; | |
164 | split[0] = htonl(offset >> 32); | |
165 | split[1] = htonl(offset & 0xffffffff); | |
98a3beab | 166 | hashwrite(f, split, 8); |
fb956c1f | 167 | nr_large_offset--; |
aa7e44bf GB |
168 | } |
169 | } | |
170 | ||
98a3beab | 171 | hashwrite(f, sha1, the_hash_algo->rawsz); |
cfe83216 DS |
172 | finalize_hashfile(f, NULL, CSUM_HASH_IN_STREAM | CSUM_CLOSE | |
173 | ((opts->flags & WRITE_IDX_VERIFY) | |
174 | ? 0 : CSUM_FSYNC)); | |
aa7e44bf GB |
175 | return index_name; |
176 | } | |
8b0eca7c | 177 | |
98a3beab | 178 | off_t write_pack_header(struct hashfile *f, uint32_t nr_entries) |
c0ad4657 JH |
179 | { |
180 | struct pack_header hdr; | |
181 | ||
182 | hdr.hdr_signature = htonl(PACK_SIGNATURE); | |
183 | hdr.hdr_version = htonl(PACK_VERSION); | |
184 | hdr.hdr_entries = htonl(nr_entries); | |
98a3beab | 185 | hashwrite(f, &hdr, sizeof(hdr)); |
c0ad4657 JH |
186 | return sizeof(hdr); |
187 | } | |
188 | ||
abeb40e5 NP |
189 | /* |
190 | * Update pack header with object_count and compute new SHA1 for pack data | |
191 | * associated to pack_fd, and write that SHA1 at the end. That new SHA1 | |
192 | * is also returned in new_pack_sha1. | |
193 | * | |
194 | * If partial_pack_sha1 is non null, then the SHA1 of the existing pack | |
195 | * (without the header update) is computed and validated against the | |
196 | * one provided in partial_pack_sha1. The validation is performed at | |
197 | * partial_pack_offset bytes in the pack file. The SHA1 of the remaining | |
198 | * data (i.e. from partial_pack_offset to the end) is then computed and | |
199 | * returned in partial_pack_sha1. | |
200 | * | |
201 | * Note that new_pack_sha1 is updated last, so both new_pack_sha1 and | |
202 | * partial_pack_sha1 can refer to the same buffer if the caller is not | |
203 | * interested in the resulting SHA1 of pack data above partial_pack_offset. | |
204 | */ | |
8b0eca7c | 205 | void fixup_pack_header_footer(int pack_fd, |
81c58cd4 | 206 | unsigned char *new_pack_hash, |
8b0eca7c | 207 | const char *pack_name, |
abeb40e5 | 208 | uint32_t object_count, |
81c58cd4 | 209 | unsigned char *partial_pack_hash, |
abeb40e5 | 210 | off_t partial_pack_offset) |
8b0eca7c | 211 | { |
d35825da | 212 | int aligned_sz, buf_sz = 8 * 1024; |
81c58cd4 | 213 | git_hash_ctx old_hash_ctx, new_hash_ctx; |
8b0eca7c DH |
214 | struct pack_header hdr; |
215 | char *buf; | |
90dca671 | 216 | ssize_t read_result; |
8b0eca7c | 217 | |
81c58cd4 | 218 | the_hash_algo->init_fn(&old_hash_ctx); |
219 | the_hash_algo->init_fn(&new_hash_ctx); | |
abeb40e5 | 220 | |
8b0eca7c | 221 | if (lseek(pack_fd, 0, SEEK_SET) != 0) |
d824cbba | 222 | die_errno("Failed seeking to start of '%s'", pack_name); |
90dca671 JK |
223 | read_result = read_in_full(pack_fd, &hdr, sizeof(hdr)); |
224 | if (read_result < 0) | |
d824cbba | 225 | die_errno("Unable to reread header of '%s'", pack_name); |
90dca671 JK |
226 | else if (read_result != sizeof(hdr)) |
227 | die_errno("Unexpected short read for header of '%s'", | |
228 | pack_name); | |
8b0eca7c | 229 | if (lseek(pack_fd, 0, SEEK_SET) != 0) |
d824cbba | 230 | die_errno("Failed seeking to start of '%s'", pack_name); |
81c58cd4 | 231 | the_hash_algo->update_fn(&old_hash_ctx, &hdr, sizeof(hdr)); |
8b0eca7c | 232 | hdr.hdr_entries = htonl(object_count); |
81c58cd4 | 233 | the_hash_algo->update_fn(&new_hash_ctx, &hdr, sizeof(hdr)); |
8b0eca7c | 234 | write_or_die(pack_fd, &hdr, sizeof(hdr)); |
abeb40e5 | 235 | partial_pack_offset -= sizeof(hdr); |
8b0eca7c DH |
236 | |
237 | buf = xmalloc(buf_sz); | |
d35825da | 238 | aligned_sz = buf_sz - sizeof(hdr); |
8b0eca7c | 239 | for (;;) { |
abeb40e5 | 240 | ssize_t m, n; |
81c58cd4 | 241 | m = (partial_pack_hash && partial_pack_offset < aligned_sz) ? |
d35825da | 242 | partial_pack_offset : aligned_sz; |
abeb40e5 | 243 | n = xread(pack_fd, buf, m); |
8b0eca7c DH |
244 | if (!n) |
245 | break; | |
246 | if (n < 0) | |
d824cbba | 247 | die_errno("Failed to checksum '%s'", pack_name); |
81c58cd4 | 248 | the_hash_algo->update_fn(&new_hash_ctx, buf, n); |
abeb40e5 | 249 | |
d35825da NP |
250 | aligned_sz -= n; |
251 | if (!aligned_sz) | |
252 | aligned_sz = buf_sz; | |
253 | ||
81c58cd4 | 254 | if (!partial_pack_hash) |
abeb40e5 NP |
255 | continue; |
256 | ||
81c58cd4 | 257 | the_hash_algo->update_fn(&old_hash_ctx, buf, n); |
abeb40e5 NP |
258 | partial_pack_offset -= n; |
259 | if (partial_pack_offset == 0) { | |
81c58cd4 | 260 | unsigned char hash[GIT_MAX_RAWSZ]; |
261 | the_hash_algo->final_fn(hash, &old_hash_ctx); | |
67947c34 | 262 | if (!hasheq(hash, partial_pack_hash)) |
abeb40e5 NP |
263 | die("Unexpected checksum for %s " |
264 | "(disk corruption?)", pack_name); | |
265 | /* | |
266 | * Now let's compute the SHA1 of the remainder of the | |
267 | * pack, which also means making partial_pack_offset | |
268 | * big enough not to matter anymore. | |
269 | */ | |
81c58cd4 | 270 | the_hash_algo->init_fn(&old_hash_ctx); |
abeb40e5 NP |
271 | partial_pack_offset = ~partial_pack_offset; |
272 | partial_pack_offset -= MSB(partial_pack_offset, 1); | |
273 | } | |
8b0eca7c DH |
274 | } |
275 | free(buf); | |
276 | ||
81c58cd4 | 277 | if (partial_pack_hash) |
278 | the_hash_algo->final_fn(partial_pack_hash, &old_hash_ctx); | |
279 | the_hash_algo->final_fn(new_pack_hash, &new_hash_ctx); | |
280 | write_or_die(pack_fd, new_pack_hash, the_hash_algo->rawsz); | |
0c68d386 | 281 | fsync_or_die(pack_fd, pack_name); |
8b0eca7c | 282 | } |
106764e6 SP |
283 | |
284 | char *index_pack_lockfile(int ip_out) | |
285 | { | |
81c58cd4 | 286 | char packname[GIT_MAX_HEXSZ + 6]; |
287 | const int len = the_hash_algo->hexsz + 6; | |
106764e6 SP |
288 | |
289 | /* | |
6e180cdc | 290 | * The first thing we expect from index-pack's output |
106764e6 SP |
291 | * is "pack\t%40s\n" or "keep\t%40s\n" (46 bytes) where |
292 | * %40s is the newly created pack SHA1 name. In the "keep" | |
293 | * case, we need it to remove the corresponding .keep file | |
294 | * later on. If we don't get that then tough luck with it. | |
295 | */ | |
81c58cd4 | 296 | if (read_in_full(ip_out, packname, len) == len && packname[len-1] == '\n') { |
d7731444 | 297 | const char *name; |
81c58cd4 | 298 | packname[len-1] = 0; |
d7731444 RS |
299 | if (skip_prefix(packname, "keep\t", &name)) |
300 | return xstrfmt("%s/pack/pack-%s.keep", | |
301 | get_object_directory(), name); | |
106764e6 SP |
302 | } |
303 | return NULL; | |
304 | } | |
f965c525 NP |
305 | |
306 | /* | |
307 | * The per-object header is a pretty dense thing, which is | |
308 | * - first byte: low four bits are "size", then three bits of "type", | |
309 | * and the high bit is "size continues". | |
310 | * - each byte afterwards: low seven bits are size continuation, | |
311 | * with the high bit being "size continues" | |
312 | */ | |
7202a6fa JK |
313 | int encode_in_pack_object_header(unsigned char *hdr, int hdr_len, |
314 | enum object_type type, uintmax_t size) | |
f965c525 NP |
315 | { |
316 | int n = 1; | |
317 | unsigned char c; | |
318 | ||
319 | if (type < OBJ_COMMIT || type > OBJ_REF_DELTA) | |
320 | die("bad type %d", type); | |
321 | ||
322 | c = (type << 4) | (size & 15); | |
323 | size >>= 4; | |
324 | while (size) { | |
7202a6fa JK |
325 | if (n == hdr_len) |
326 | die("object size is too enormous to format"); | |
f965c525 NP |
327 | *hdr++ = c | 0x80; |
328 | c = size & 0x7f; | |
329 | size >>= 7; | |
330 | n++; | |
331 | } | |
332 | *hdr = c; | |
333 | return n; | |
334 | } | |
cdf9db3c | 335 | |
98a3beab | 336 | struct hashfile *create_tmp_packfile(char **pack_tmp_name) |
cdf9db3c | 337 | { |
594fa999 | 338 | struct strbuf tmpname = STRBUF_INIT; |
cdf9db3c JH |
339 | int fd; |
340 | ||
594fa999 JK |
341 | fd = odb_mkstemp(&tmpname, "pack/tmp_pack_XXXXXX"); |
342 | *pack_tmp_name = strbuf_detach(&tmpname, NULL); | |
98a3beab | 343 | return hashfd(fd, *pack_tmp_name); |
cdf9db3c | 344 | } |
0e990530 | 345 | |
58892711 | 346 | void finish_tmp_packfile(struct strbuf *name_buffer, |
0e990530 JH |
347 | const char *pack_tmp_name, |
348 | struct pack_idx_entry **written_list, | |
349 | uint32_t nr_written, | |
350 | struct pack_idx_option *pack_idx_opts, | |
894c0f66 | 351 | unsigned char hash[]) |
0e990530 JH |
352 | { |
353 | const char *idx_tmp_name; | |
58892711 | 354 | int basename_len = name_buffer->len; |
0e990530 JH |
355 | |
356 | if (adjust_shared_perm(pack_tmp_name)) | |
357 | die_errno("unable to make temporary pack file readable"); | |
358 | ||
359 | idx_tmp_name = write_idx_file(NULL, written_list, nr_written, | |
894c0f66 | 360 | pack_idx_opts, hash); |
0e990530 JH |
361 | if (adjust_shared_perm(idx_tmp_name)) |
362 | die_errno("unable to make temporary index file readable"); | |
363 | ||
894c0f66 | 364 | strbuf_addf(name_buffer, "%s.pack", hash_to_hex(hash)); |
0e990530 | 365 | |
58892711 | 366 | if (rename(pack_tmp_name, name_buffer->buf)) |
0e990530 JH |
367 | die_errno("unable to rename temporary pack file"); |
368 | ||
58892711 SH |
369 | strbuf_setlen(name_buffer, basename_len); |
370 | ||
894c0f66 | 371 | strbuf_addf(name_buffer, "%s.idx", hash_to_hex(hash)); |
58892711 | 372 | if (rename(idx_tmp_name, name_buffer->buf)) |
0e990530 JH |
373 | die_errno("unable to rename temporary index file"); |
374 | ||
58892711 | 375 | strbuf_setlen(name_buffer, basename_len); |
7cc8f971 | 376 | |
0e990530 JH |
377 | free((void *)idx_tmp_name); |
378 | } |