tar2qfile.c 24 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522523524525526527528529530531532533534535536537538539540541542543544545546547548549550551552553554555556557558559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672673674675676677678679680681682683684685686687688689690691692693694695696697698699700701702703704705706707708709710711712713714715716717718719720721722723724725726727728729730731732733734735736737738739740741742743744745746747748749750751752753754755756757758759760761762763764765766767768769770771772773774775776777778779780781782783784785786787788789790791792793794795796797798799800801802803804805806807808809810811812813814815816817818819820821822823824825826827828829830831832833834835836837838839840841842843844845846847848849850851852853854855856857858859860861862863864865866867868869870871872873874875876877878879880881882883884885886887888889
  1. /* $OpenBSD: tar.h,v 1.7 2003/06/02 23:32:09 millert Exp $ */
  2. /* $NetBSD: tar.h,v 1.3 1995/03/21 09:07:51 cgd Exp $ */
  3. /*-
  4. * Copyright (c) 1992 Keith Muller.
  5. * Copyright (c) 1992, 1993
  6. * The Regents of the University of California. All rights reserved.
  7. *
  8. * This code is derived from software contributed to Berkeley by
  9. * Keith Muller of the University of California, San Diego.
  10. *
  11. * Redistribution and use in source and binary forms, with or without
  12. * modification, are permitted provided that the following conditions
  13. * are met:
  14. * 1. Redistributions of source code must retain the above copyright
  15. * notice, this list of conditions and the following disclaimer.
  16. * 2. Redistributions in binary form must reproduce the above copyright
  17. * notice, this list of conditions and the following disclaimer in the
  18. * documentation and/or other materials provided with the distribution.
  19. * 3. Neither the name of the University nor the names of its contributors
  20. * may be used to endorse or promote products derived from this software
  21. * without specific prior written permission.
  22. *
  23. * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
  24. * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
  25. * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
  26. * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
  27. * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
  28. * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
  29. * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
  30. * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
  31. * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
  32. * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
  33. * SUCH DAMAGE.
  34. *
  35. * @(#)tar.h 8.2 (Berkeley) 4/18/94
  36. */
  37. #define _GNU_SOURCE /* For O_NOFOLLOW. */
  38. #include <errno.h>
  39. #include <ioall.h>
  40. #include <fcntl.h>
  41. #include <sys/time.h>
  42. #include <sys/stat.h>
  43. #include <stdlib.h>
  44. #include <unistd.h>
  45. #include <stdio.h>
  46. #include <string.h>
  47. #include <qfile-utils.h>
  48. /***************************************************
  49. * Most routines extracted from the PAX project (tar.c...) *
  50. ***************************************************/
  51. /*
  52. * BSD PAX global data structures and constants.
  53. */
  54. #define MAXBLK 64512 /* MAX blocksize supported (posix SPEC) */
  55. /* WARNING: increasing MAXBLK past 32256 */
  56. /* will violate posix spec. */
  57. #define MAXBLK_POSIX 32256 /* MAX blocksize supported as per POSIX */
  58. #define BLKMULT 512 /* blocksize must be even mult of 512 bytes */
  59. /* Don't even think of changing this */
  60. #define DEVBLK 8192 /* default read blksize for devices */
  61. #define FILEBLK 10240 /* default read blksize for files */
  62. #define PAXPATHLEN 3072 /* maximium path length for pax. MUST be */
  63. /*
  64. * defines and data structures common to all tar formats
  65. */
  66. #define CHK_LEN 8 /* length of checksum field */
  67. #define TNMSZ 100 /* size of name field */
  68. #define NULLCNT 2 /* number of null blocks in trailer */
  69. #define CHK_OFFSET 148 /* start of chksum field */
  70. #define BLNKSUM 256L /* sum of checksum field using ' ' */
  71. /*
  72. * General Defines
  73. */
  74. #define HEX 16
  75. #define OCT 8
  76. #define _PAX_ 1
  77. #define _TFILE_BASE "paxXXXXXXXXXX"
  78. /*
  79. * General Macros
  80. */
  81. #ifndef MIN
  82. #define MIN(a,b) (((a)<(b))?(a):(b))
  83. #endif
  84. #ifndef MAX
  85. #define MAX(a,b) (((a)>(b))?(a):(b))
  86. #endif
  87. #define MAJOR(x) major(x)
  88. #define MINOR(x) minor(x)
  89. #define TODEV(x, y) makedev((x), (y))
  90. /*
  91. * Values used in typeflag field in all tar formats
  92. * (only REGTYPE, LNKTYPE and SYMTYPE are used in old bsd tar headers)
  93. */
  94. #define REGTYPE '0' /* Regular File */
  95. #define AREGTYPE '\0' /* Regular File */
  96. #define LNKTYPE '1' /* Link */
  97. #define SYMTYPE '2' /* Symlink */
  98. #define CHRTYPE '3' /* Character Special File */
  99. #define BLKTYPE '4' /* Block Special File */
  100. #define DIRTYPE '5' /* Directory */
  101. #define FIFOTYPE '6' /* FIFO */
  102. #define CONTTYPE '7' /* high perf file */
  103. /*
  104. * GNU tar compatibility;
  105. */
  106. #define LONGLINKTYPE 'K' /* Long Symlink */
  107. #define LONGNAMETYPE 'L' /* Long File */
  108. #define EXTHEADERTYPE 'x' /* Extended header */
  109. /*
  110. * Pad with a bit mask, much faster than doing a mod but only works on powers
  111. * of 2. Macro below is for block of 512 bytes.
  112. */
  113. #define TAR_PAD(x) ((512 - ((x) & 511)) & 511)
  114. /*
  115. * Data Interchange Format - Extended tar header format - POSIX 1003.1-1990
  116. */
  117. #define TPFSZ 155
  118. #define TMAGIC "ustar" /* ustar and a null */
  119. #define TMAGLEN 6
  120. #define TVERSION "00" /* 00 and no null */
  121. #define TVERSLEN 2
  122. typedef struct {
  123. char name[TNMSZ]; /* name of entry */
  124. char mode[8]; /* mode */
  125. char uid[8]; /* uid */
  126. char gid[8]; /* gid */
  127. char size[12]; /* size */
  128. char mtime[12]; /* modification time */
  129. char chksum[CHK_LEN]; /* checksum */
  130. char typeflag; /* type of file. */
  131. char linkname[TNMSZ]; /* linked to name */
  132. char magic[TMAGLEN]; /* magic cookie */
  133. char version[TVERSLEN]; /* version */
  134. char uname[32]; /* ascii owner name */
  135. char gname[32]; /* ascii group name */
  136. char devmajor[8]; /* major device number */
  137. char devminor[8]; /* minor device number */
  138. char prefix[TPFSZ]; /* linked to name */
  139. } HD_USTAR;
  140. /*
  141. * Routines for manipulating headers, trailers:
  142. * asc_ul()
  143. * tar_trail()
  144. * tar_chksm()
  145. * ustar_id()
  146. */
  147. static unsigned long tar_chksm (char *, int);
  148. char *gnu_hack_string; /* GNU ././@LongLink hackery */
  149. char untrusted_namebuf[MAX_PATH_LENGTH];
  150. extern int ignore_quota_error;
  151. /*
  152. * asc_ul()
  153. * convert hex/octal character string into a u_long. We do not have to
  154. * check for overflow! (the headers in all supported formats are not large
  155. * enough to create an overflow).
  156. * NOTE: strings passed to us are NOT TERMINATED.
  157. * Return:
  158. * unsigned long value
  159. */
  160. u_long
  161. asc_ul (char *str, int len, int base)
  162. {
  163. char *stop;
  164. u_long tval = 0;
  165. stop = str + len;
  166. /*
  167. * skip over leading blanks and zeros
  168. */
  169. while ((str < stop) && ((*str == ' ') || (*str == '0')))
  170. ++str;
  171. /*
  172. * for each valid digit, shift running value (tval) over to next digit
  173. * and add next digit
  174. */
  175. if (base == HEX)
  176. {
  177. while (str < stop)
  178. {
  179. if ((*str >= '0') && (*str <= '9'))
  180. tval = (tval << 4) + (*str++ - '0');
  181. else if ((*str >= 'A') && (*str <= 'F'))
  182. tval = (tval << 4) + 10 + (*str++ - 'A');
  183. else if ((*str >= 'a') && (*str <= 'f'))
  184. tval = (tval << 4) + 10 + (*str++ - 'a');
  185. else
  186. break;
  187. }
  188. }
  189. else
  190. {
  191. while ((str < stop) && (*str >= '0') && (*str <= '7'))
  192. tval = (tval << 3) + (*str++ - '0');
  193. }
  194. return (tval);
  195. }
  196. /*
  197. * tar_trail()
  198. * Called to determine if a header block is a valid trailer. We are passed
  199. * the block, the in_sync flag (which tells us we are in resync mode;
  200. * looking for a valid header), and cnt (which starts at zero) which is
  201. * used to count the number of empty blocks we have seen so far.
  202. * Return:
  203. * 0 if a valid trailer, -1 if not a valid trailer, or 1 if the block
  204. * could never contain a header.
  205. */
  206. int
  207. tar_trail (char *buf,
  208. int in_resync, int *cnt)
  209. {
  210. register int i;
  211. /*
  212. * look for all zero, trailer is two consecutive blocks of zero
  213. */
  214. for (i = 0; i < BLKMULT; ++i)
  215. {
  216. if (buf[i] != '\0')
  217. break;
  218. }
  219. /*
  220. * if not all zero it is not a trailer, but MIGHT be a header.
  221. */
  222. if (i != BLKMULT)
  223. return (-1);
  224. /*
  225. * When given a zero block, we must be careful!
  226. * If we are not in resync mode, check for the trailer. Have to watch
  227. * out that we do not mis-identify file data as the trailer, so we do
  228. * NOT try to id a trailer during resync mode. During resync mode we
  229. * might as well throw this block out since a valid header can NEVER be
  230. * a block of all 0 (we must have a valid file name).
  231. */
  232. if (!in_resync && (++*cnt >= NULLCNT))
  233. return (0);
  234. return (1);
  235. }
  236. /*
  237. * tar_chksm()
  238. * calculate the checksum for a tar block counting the checksum field as
  239. * all blanks (BLNKSUM is that value pre-calculated, the sum of 8 blanks).
  240. * NOTE: we use len to short circuit summing 0's on write since we ALWAYS
  241. * pad headers with 0.
  242. * Return:
  243. * unsigned long checksum
  244. */
  245. static unsigned long
  246. tar_chksm (char *blk, int len)
  247. {
  248. char *stop;
  249. char *pt;
  250. unsigned int chksm = BLNKSUM; /* initial value is checksum field sum */
  251. /*
  252. * add the part of the block before the checksum field
  253. */
  254. pt = blk;
  255. stop = blk + CHK_OFFSET;
  256. while (pt < stop)
  257. chksm += (*pt++ & 0xff);
  258. /*
  259. * move past the checksum field and keep going, spec counts the
  260. * checksum field as the sum of 8 blanks (which is pre-computed as
  261. * BLNKSUM).
  262. * ASSUMED: len is greater than CHK_OFFSET. (len is where our 0 padding
  263. * starts, no point in summing zero's)
  264. */
  265. pt += CHK_LEN;
  266. stop = blk + len;
  267. while (pt < stop)
  268. chksm += (*pt++ & 0xff);
  269. return chksm;
  270. }
  271. /*
  272. * ustar_id()
  273. * determine if a block given to us is a valid ustar header. We have to
  274. * be on the lookout for those pesky blocks of all zero's
  275. * Return:
  276. * 0 if a ustar header, -1 otherwise
  277. */
  278. int
  279. ustar_id (char *blk, size_t size)
  280. {
  281. HD_USTAR *hd;
  282. if (size < BLKMULT)
  283. return (-1);
  284. hd = (HD_USTAR *) blk;
  285. /*
  286. * check for block of zero's first, a simple and fast test then check
  287. * ustar magic cookie. We should use TMAGLEN, but some USTAR archive
  288. * programs are fouled up and create archives missing the \0. Last we
  289. * check the checksum. If ok we have to assume it is a valid header.
  290. */
  291. if (hd->name[0] == '\0')
  292. return (-1);
  293. if (strncmp (hd->magic, TMAGIC, TMAGLEN - 1) != 0)
  294. return (-1);
  295. if (asc_ul (hd->chksum, sizeof (hd->chksum), OCT) !=
  296. tar_chksm (blk, BLKMULT))
  297. return (-1);
  298. return (0);
  299. }
  300. /*
  301. * Routines for reading tar files
  302. // Source: http://www.mkssoftware.com/docs/man4/pax.4.asp
  303. struct file_header { // PAX header is similar as file_header and can be completely ignored
  304. unsigned char[100] name;
  305. unsigned char[8] mode;
  306. unsigned char[8] uid; // unused
  307. unsigned char[8] gid; // unused
  308. unsigned char[12] size; // 0 if file is a link
  309. unsigned char[12] mtime;
  310. unsigned char[8] chksum;
  311. unsigned char[1] typeflag;
  312. unsigned char[100] linkname;
  313. unsigned char[6] magic; //ustar
  314. unsigned char[2] version; // 00
  315. unsigned char[32] uname; // unused
  316. unsigned char[32] gname; // unused
  317. unsigned char[8] devmajor; // unused ?
  318. unsigned char[8] devminor; // unused ?
  319. unsigned char[155] prefix; // only used for files > 100 characters. could be unused ?
  320. };
  321. enum {
  322. TYPE_REGULAR, //0
  323. TYPE_ARCHIVE_LINK, //1
  324. TYPE_SYMLINK, //2
  325. TYPE_CHARACTER_DEVICE, //3
  326. TYPE_BLOCK_DEVICE, //4
  327. TYPE_DIRECTORY, //5
  328. TYPE_FIFO, //6
  329. // Other types:
  330. TYPE_EXTENDED_USAGE, //xxxxx
  331. // A-Z are available for custom usage
  332. };
  333. // Extended attribute:
  334. // length keyword=value
  335. // atime, charset, comment, gname, linkpath, mtime, path, size, uname
  336. */
  337. enum {
  338. NEED_NOTHING,
  339. NEED_SKIP,
  340. NEED_READ,
  341. NEED_SYNC_TRAIL,
  342. INVALID_HEADER,
  343. MEMORY_ALLOC_FAILED,
  344. };
  345. /*
  346. * ustar_rd()
  347. * extract the values out of block already determined to be a ustar header.
  348. * store the values in the ARCHD parameter.
  349. * Return:
  350. * 0
  351. */
  352. int n_dirs = 0;
  353. char ** dirs_headers_sent = NULL;
  354. int
  355. ustar_rd (int fd, struct file_header * untrusted_hdr, char *buf, struct stat * sb, int filter_count, char **filter)
  356. {
  357. register HD_USTAR *hd;
  358. register char *dest;
  359. register int cnt = 0;
  360. int ret;
  361. int i;
  362. int should_extract;
  363. /* DISABLED: unused
  364. dev_t devmajor;
  365. dev_t devminor;
  366. */
  367. /*
  368. * we only get proper sized buffers
  369. */
  370. fprintf(stderr,"Checking if valid header\n");
  371. if (ustar_id (buf, BLKMULT) < 0) {
  372. fprintf (stderr,"Invalid header\n");
  373. return INVALID_HEADER;
  374. }
  375. fprintf(stderr,"Valid header!\n");
  376. /* DISABLED: Internal to PAX
  377. arcn->org_name = arcn->name;
  378. arcn->sb.st_nlink = 1;
  379. arcn->pat = NULL;
  380. arcn->nlen = 0;
  381. */
  382. untrusted_hdr->namelen = 0;
  383. hd = (HD_USTAR *) buf;
  384. /*
  385. * see if the filename is split into two parts. if, so joint the parts.
  386. * we copy the prefix first and add a / between the prefix and name.
  387. */
  388. dest = untrusted_namebuf;
  389. if (*(hd->prefix) != '\0')
  390. {
  391. cnt = strlen(strncpy (dest, hd->prefix,
  392. MIN(sizeof (untrusted_namebuf) - 1,TPFSZ+1)));
  393. dest += cnt;
  394. *dest++ = '/';
  395. cnt++;
  396. }
  397. if (gnu_hack_string)
  398. {
  399. untrusted_hdr->namelen = cnt + strlen(strncpy (dest, gnu_hack_string,
  400. MIN(TNMSZ+1, sizeof (untrusted_namebuf) - cnt)));
  401. free(gnu_hack_string);
  402. gnu_hack_string = NULL;
  403. } else
  404. untrusted_hdr->namelen = cnt + strlen(strncpy (dest, hd->name,
  405. MIN(TNMSZ+1, sizeof (untrusted_namebuf) - cnt)));
  406. // qfile count the \0 in the namelen
  407. untrusted_hdr->namelen += 1;
  408. fprintf(stderr,"Retrieved name len: %d\n",untrusted_hdr->namelen);
  409. fprintf(stderr,"Retrieved name: %s\n",untrusted_namebuf);
  410. /*
  411. * follow the spec to the letter. we should only have mode bits, strip
  412. * off all other crud we may be passed.
  413. */
  414. sb->st_mode = (mode_t) (asc_ul (hd->mode, sizeof (hd->mode), OCT) &
  415. 0xfff);
  416. untrusted_hdr->mode = sb->st_mode;
  417. #if defined(_FILE_OFFSET_BITS) && _FILE_OFFSET_BITS == 64
  418. sb->st_size = (off_t) asc_uqd (hd->size, sizeof (hd->size), OCT);
  419. #else
  420. sb->st_size = (off_t) asc_ul (hd->size, sizeof (hd->size), OCT);
  421. #endif
  422. untrusted_hdr->filelen = sb->st_size;
  423. untrusted_hdr->atime = (time_t) asc_ul (hd->mtime, sizeof (hd->mtime), OCT);
  424. untrusted_hdr->mtime = untrusted_hdr->atime;
  425. untrusted_hdr->atime_nsec = untrusted_hdr->mtime_nsec = 0;
  426. sb->st_mtime = (time_t) asc_ul (hd->mtime, sizeof (hd->mtime), OCT);
  427. sb->st_ctime = sb->st_atime = sb->st_mtime;
  428. /*
  429. * If we can find the ascii names for gname and uname in the password
  430. * and group files we will use the uid's and gid they bind. Otherwise
  431. * we use the uid and gid values stored in the header. (This is what
  432. * the posix spec wants).
  433. */
  434. /* DISABLED: unused
  435. hd->gname[sizeof (hd->gname) - 1] = '\0';
  436. if (gid_name (hd->gname, &(arcn->sb.st_gid)) < 0)
  437. arcn->sb.st_gid = (gid_t) asc_ul (hd->gid, sizeof (hd->gid), OCT);
  438. hd->uname[sizeof (hd->uname) - 1] = '\0';
  439. if (uid_name (hd->uname, &(arcn->sb.st_uid)) < 0)
  440. arcn->sb.st_uid = (uid_t) asc_ul (hd->uid, sizeof (hd->uid), OCT);
  441. */
  442. /*
  443. * set the defaults, these may be changed depending on the file type
  444. */
  445. /* Disabled: pax specific
  446. arcn->ln_name[0] = '\0';
  447. arcn->ln_nlen = 0;
  448. arcn->pad = 0;
  449. arcn->skip = 0;
  450. arcn->sb.st_rdev = (dev_t) 0;
  451. */
  452. /*
  453. * set the mode and PAX type according to the typeflag in the header
  454. */
  455. switch (hd->typeflag)
  456. {
  457. case FIFOTYPE:
  458. fprintf(stderr,"File is FIFOTYPE\n");
  459. /* DISABLED: unused
  460. arcn->type = PAX_FIF;
  461. arcn->sb.st_mode |= S_IFIFO;
  462. */
  463. break;
  464. case DIRTYPE:
  465. fprintf(stderr,"File is DIRTYPE\n");
  466. /* DISABLED: unused
  467. arcn->type = PAX_DIR;
  468. arcn->sb.st_mode |= S_IFDIR;
  469. arcn->sb.st_nlink = 2;
  470. */
  471. /*
  472. * Some programs that create ustar archives append a '/'
  473. * to the pathname for directories. This clearly violates
  474. * ustar specs, but we will silently strip it off anyway.
  475. */
  476. /*
  477. if (arcn->name[arcn->nlen - 1] == '/')
  478. arcn->name[--arcn->nlen] = '\0';
  479. */
  480. break;
  481. case BLKTYPE:
  482. fprintf(stderr,"File is BLKTYPE\n");
  483. break;
  484. case CHRTYPE:
  485. fprintf(stderr,"File is CHRTYPE\n");
  486. /*
  487. * this type requires the rdev field to be set.
  488. */
  489. if (hd->typeflag == BLKTYPE)
  490. {
  491. /*
  492. arcn->type = PAX_BLK;
  493. arcn->sb.st_mode |= S_IFBLK;
  494. */
  495. }
  496. else
  497. {
  498. /*
  499. arcn->type = PAX_CHR;
  500. arcn->sb.st_mode |= S_IFCHR;
  501. */
  502. }
  503. /* DISABLED: unused
  504. devmajor = (dev_t) asc_ul (hd->devmajor, sizeof (hd->devmajor), OCT);
  505. devminor = (dev_t) asc_ul (hd->devminor, sizeof (hd->devminor), OCT);
  506. */
  507. // arcn->sb.st_rdev = TODEV (devmajor, devminor);
  508. break;
  509. case SYMTYPE:
  510. fprintf(stderr,"File is SYMTYPE\n");
  511. break;
  512. case LNKTYPE:
  513. fprintf(stderr,"File is LNKTYPE\n");
  514. if (hd->typeflag == SYMTYPE)
  515. {
  516. // arcn->type = PAX_SLK;
  517. // arcn->sb.st_mode |= S_IFLNK;
  518. }
  519. else
  520. {
  521. // arcn->type = PAX_HLK;
  522. /*
  523. * so printing looks better
  524. */
  525. // arcn->sb.st_mode |= S_IFREG;
  526. // arcn->sb.st_nlink = 2;
  527. }
  528. /*
  529. * copy the link name
  530. */
  531. // arcn->ln_nlen = strlcpy (arcn->ln_name, hd->linkname,
  532. // MIN(TNMSZ+1,sizeof (arcn->ln_name)));
  533. break;
  534. case LONGLINKTYPE:
  535. fprintf(stderr,"File is LONGLINKTYPE\n");
  536. break;
  537. case LONGNAMETYPE:
  538. fprintf(stderr,"File is LONGNAMETYPE\n");
  539. /*
  540. * GNU long link/file; we tag these here and let the
  541. * pax internals deal with it -- too ugly otherwise.
  542. */
  543. // arcn->type =
  544. // hd->typeflag == LONGLINKTYPE ? PAX_GLL : PAX_GLF;
  545. // arcn->pad = TAR_PAD(arcn->sb.st_size);
  546. // arcn->skip = arcn->sb.st_size;
  547. // arcn->ln_name[0] = '\0';
  548. // arcn->ln_nlen = 0;
  549. break;
  550. case CONTTYPE:
  551. fprintf(stderr,"File is CONTTYPE\n");
  552. break;
  553. case AREGTYPE:
  554. fprintf(stderr,"File is AREGTYPE\n");
  555. break;
  556. case REGTYPE:
  557. fprintf(stderr,"File is REGTYPE of size %d\n",sb->st_size);
  558. // Check if user want to extract this file
  559. should_extract = 1;
  560. for (i=1; i < filter_count; i++) {
  561. should_extract = 0;
  562. fprintf(stderr, "Comparing with filter %s\n", filter[i]);
  563. if (strstr(untrusted_namebuf, filter[i]) == untrusted_namebuf) {
  564. fprintf(stderr, "Match\n");
  565. should_extract = 1;
  566. break;
  567. }
  568. }
  569. if (should_extract != 1) {
  570. fprintf(stderr, "File should be filtered.. Skipping\n");
  571. return NEED_SKIP;
  572. }
  573. // Create a copy of untrusted_namebuf to be used for strtok
  574. char * dirbuf;
  575. dirbuf = malloc(sizeof (char) * (untrusted_hdr->namelen));
  576. if (dirbuf == NULL)
  577. return MEMORY_ALLOC_FAILED;
  578. dirbuf = strncpy(dirbuf, untrusted_namebuf, untrusted_hdr->namelen);
  579. int i = 0;
  580. int dir_found = 0;
  581. size_t pathsize = 0;
  582. char * path = NULL;
  583. struct file_header dir_header;
  584. // Split the path in directories and recompose it incrementally
  585. char * last_token = strtok(dirbuf,"/");
  586. char * token = strtok(NULL, "/");
  587. while (token != NULL) {
  588. fprintf(stderr,"Found directory %s (last:%s)\n",token,last_token);
  589. // Recompose the path based on last discovered directory
  590. if (path == NULL) {
  591. path = malloc(sizeof (char) * (strlen(last_token)+1));
  592. if (path == NULL)
  593. return MEMORY_ALLOC_FAILED;
  594. path = strncpy(path, last_token, strlen(last_token));
  595. path[strlen(last_token)] = '\0';
  596. } else {
  597. pathsize = strlen(path);
  598. path = realloc(path, sizeof (char) * (strlen(path)+1+strlen(last_token)+1));
  599. if (path == NULL)
  600. return MEMORY_ALLOC_FAILED;
  601. path[pathsize] = '/';
  602. strncpy(path+pathsize+1, last_token, strlen(last_token));
  603. path[pathsize+strlen(last_token)+1] = '\0';
  604. }
  605. fprintf(stderr,"Path is %s\n",path);
  606. fprintf(stderr,"Checking from i=0 i<%d\n",n_dirs);
  607. // Verify if qfile headers for the current path have already been sent based on the dirs_headers_sent table
  608. dir_found = 0;
  609. for (i = 0; i < n_dirs; ++i) {
  610. fprintf(stderr,"Comparing with %d %d %s %s\n",i,n_dirs,dirs_headers_sent[i],path);
  611. if (strcmp(dirs_headers_sent[i],path)==0) {
  612. fprintf(stderr,"Directory headers already sent\n");
  613. dir_found=1;
  614. }
  615. }
  616. if (dir_found == 0) {
  617. // Register the current path as being sent in the dirs_headers_sent table
  618. fprintf(stderr,"Inserting %s into register\n",path);
  619. dirs_headers_sent = realloc(dirs_headers_sent, sizeof (char*) * n_dirs++);
  620. if (dirs_headers_sent == NULL)
  621. return MEMORY_ALLOC_FAILED;
  622. dirs_headers_sent[n_dirs-1] = malloc(sizeof (char) * (strlen(path)+1));
  623. if (dirs_headers_sent[n_dirs-1] == NULL)
  624. return MEMORY_ALLOC_FAILED;
  625. strncpy(dirs_headers_sent[n_dirs-1], path, strlen(path)+1);
  626. // Initialize the qfile headers for the current directory path
  627. dir_header.namelen = strlen(path)+1;
  628. dir_header.atime = untrusted_hdr->atime;
  629. dir_header.atime_nsec = untrusted_hdr->atime_nsec;
  630. dir_header.mtime = untrusted_hdr->mtime;
  631. dir_header.mtime_nsec = untrusted_hdr->mtime_nsec;
  632. dir_header.mode = untrusted_hdr->mode | S_IFDIR;
  633. dir_header.filelen = 0;
  634. fprintf(stderr,"Sending directory headers for %s\n",path);
  635. // Send the qfile headers for the current directory path
  636. write_headers(&dir_header, path);
  637. }
  638. last_token = token;
  639. token = strtok(NULL, "/");
  640. }
  641. free(path);
  642. free(dirbuf);
  643. fprintf(stderr,"End of directory checks\n");
  644. // Restore POSIX stat file mode (because PAX format use its own file type)
  645. untrusted_hdr->mode |= S_IFREG;
  646. fprintf(stderr,"Writing file header\n");
  647. // Send header and file content
  648. write_headers(untrusted_hdr, untrusted_namebuf);
  649. fprintf(stderr,"Writing file content\n");
  650. ret = copy_file(1, fd, untrusted_hdr->filelen, &crc32_sum);
  651. fprintf(stderr,"Copyfile returned with error %d\n",ret);
  652. if (ret != COPY_FILE_OK) {
  653. if (ret != COPY_FILE_WRITE_ERROR)
  654. gui_fatal("Copying file %s: %s", untrusted_namebuf,
  655. copy_file_status_to_str(ret));
  656. else {
  657. fprintf(stderr,"UNKNOWN ERROR RETURN STATUS:%d\n.. Waiting...\n",ret);
  658. set_block(0);
  659. wait_for_result();
  660. exit(1);
  661. }
  662. }
  663. // Extract extra padding
  664. fprintf(stderr,"Need to remove pad:%ld %ld\n",sb->st_size,BLKMULT-(sb->st_size%BLKMULT));
  665. ret = read(fd, buf, BLKMULT-(sb->st_size%BLKMULT));
  666. fprintf(stderr,"Removed %d bytes of padding\n",ret);
  667. // Resync trailing headers in order to find next file chunck in the tar file
  668. return NEED_SYNC_TRAIL;
  669. break;
  670. case EXTHEADERTYPE:
  671. fprintf(stderr,"Extended HEADER encountered\n");
  672. return NEED_SKIP;
  673. break;
  674. default:
  675. fprintf(stderr,"Default type detected:%c\n",hd->typeflag);
  676. return NEED_SKIP;
  677. /*
  678. * these types have file data that follows. Set the skip and
  679. * pad fields.
  680. */
  681. // arcn->type = PAX_REG;
  682. // arcn->pad = TAR_PAD (arcn->sb.st_size);
  683. // arcn->skip = arcn->sb.st_size;
  684. // arcn->sb.st_mode |= S_IFREG;
  685. break;
  686. }
  687. return NEED_SKIP;
  688. }
  689. int tar_file_processor(int fd, int filter_count, char **filter)
  690. {
  691. int ret;
  692. int i;
  693. int current;
  694. struct file_header hdr;
  695. struct stat sb; /* stat buffer see stat(2) */
  696. char buf[BLKMULT+1];
  697. size_t size;
  698. i=0;
  699. current = NEED_READ;
  700. size_t to_skip = 0;
  701. int sync_count = 0;
  702. while ((size = read(fd, &buf, BLKMULT))) {
  703. if (size != -1) {
  704. fprintf(stderr,"Read %ld bytes\n",size);
  705. ret = 0;
  706. if (current==NEED_SYNC_TRAIL) {
  707. ret = tar_trail (buf, 1, &sync_count);
  708. fprintf(stderr,"Synchronizing trail: %d %d\n",ret,sync_count);
  709. if (ret != 1) {
  710. current = NEED_READ;
  711. }
  712. }
  713. if (current==NEED_READ) {
  714. current = ustar_rd(fd, &hdr, buf, &sb);
  715. fprintf(stderr,"Return %d\n",ret);
  716. }
  717. if (current==NEED_SKIP) {
  718. fprintf(stderr,"Need to skip %ld bytes\n",sb.st_size);
  719. to_skip = sb.st_size;
  720. while (to_skip > 0) {
  721. to_skip -= read(fd, &buf, MIN(to_skip,BLKMULT));
  722. }
  723. // Extract extra padding
  724. fprintf(stderr,"Need to remove pad:%ld %ld %ld\n",to_skip,sb.st_size,BLKMULT-(sb.st_size%BLKMULT));
  725. ret = read(fd, &buf, BLKMULT-(sb.st_size%BLKMULT));
  726. fprintf(stderr,"Removed %d bytes of padding\n",ret);
  727. current = NEED_SYNC_TRAIL;
  728. }
  729. i++;
  730. }
  731. //if (i >= 10)
  732. // exit(0);
  733. }
  734. }
  735. int main(int argc, char **argv)
  736. {
  737. int i;
  738. char *entry;
  739. char *cwd;
  740. char *sep;
  741. int fd;
  742. int use_stdin = 0;
  743. signal(SIGPIPE, SIG_IGN);
  744. // this will allow checking for possible feedback packet in the middle of transfer
  745. // if disabled, the copy_file process could hang
  746. notify_progress(0, PROGRESS_FLAG_INIT);
  747. //set_size_limit(1500000000, 2048);
  748. crc32_sum = 0;
  749. /* when extracting backup header, dom0 will terminate the transfer with
  750. * EDQUOT just after getting qubes.xml */
  751. ignore_quota_error = 1;
  752. for (i = 1; i < argc; i++) {
  753. set_nonblock(0);
  754. if (strcmp(argv[i], "--ignore-symlinks")==0) {
  755. ignore_symlinks = 1;
  756. continue;
  757. } else if (strcmp(argv[i], "-")==0) {
  758. use_stdin = 1;
  759. break;
  760. } else {
  761. // Parse tar file
  762. entry = argv[i];
  763. fprintf(stderr,"Parsing file %s\n",entry);
  764. fd = open(entry, O_RDONLY);
  765. if (fd < 0)
  766. fprintf(stderr,"Error opening file %s\n",entry);
  767. // At least two arguments can be found in the command line
  768. // (process name and the file to extract)
  769. tar_file_processor(fd, argc-2, argv[2]);
  770. }
  771. }
  772. if (i <= 1 || use_stdin == 1) {
  773. // No argument specified. Use STDIN
  774. fprintf(stderr,"Using STDIN\n");
  775. set_block(0);
  776. // If at least one argument has been found ( process name and - )
  777. if (use_stdin)
  778. tar_file_processor(fileno(stdin), argc-2, argv[2]);
  779. else
  780. tar_file_processor(fileno(stdin), argc-1, argv[1]);
  781. }
  782. //notify_end_and_wait_for_result();
  783. notify_progress(0, PROGRESS_FLAG_DONE);
  784. return 0;
  785. }