src/tar.h

   1 /* GNU tar Archive Format description.
   2
   3    Copyright (C) 1988, 1989, 1991, 1992, 1993, 1994, 1995, 1996, 1997,
   4    2000, 2001, 2003, 2004, 2005, 2006, 2007, 2012
   5    Free Software Foundation, Inc.
   6
   7    This program is free software; you can redistribute it and/or modify it
   8    under the terms of the GNU General Public License as published by the
   9    Free Software Foundation; either version 3, or (at your option) any later
  10    version.
  11
  12    This program is distributed in the hope that it will be useful, but
  13    WITHOUT ANY WARRANTY; without even the implied warranty of
  14    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General
  15    Public License for more details.
  16
  17    You should have received a copy of the GNU General Public License along
  18    with this program; if not, write to the Free Software Foundation, Inc.,
  19    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.  */
  20
  21 /* tar Header Block, from POSIX 1003.1-1990.  */
  22
  23 /* POSIX header.  */
  24
  25 struct posix_header
  26 {                               /* byte offset */
  27   char name[100];               /*   0 */
  28   char mode[8];                 /* 100 */
  29   char uid[8];                  /* 108 */
  30   char gid[8];                  /* 116 */
  31   char size[12];                /* 124 */
  32   char mtime[12];               /* 136 */
  33   char chksum[8];               /* 148 */
  34   char typeflag;                /* 156 */
  35   char linkname[100];           /* 157 */
  36   char magic[6];                /* 257 */
  37   char version[2];              /* 263 */
  38   char uname[32];               /* 265 */
  39   char gname[32];               /* 297 */
  40   char devmajor[8];             /* 329 */
  41   char devminor[8];             /* 337 */
  42   char prefix[155];             /* 345 */
  43                                 /* 500 */
  44 };
  45
  46 #define TMAGIC   "ustar"        /* ustar and a null */
  47 #define TMAGLEN  6
  48 #define TVERSION "00"           /* 00 and no null */
  49 #define TVERSLEN 2
  50
  51 /* Values used in typeflag field.  */
  52 #define REGTYPE  '0'            /* regular file */
  53 #define AREGTYPE '\0'           /* regular file */
  54 #define LNKTYPE  '1'            /* link */
  55 #define SYMTYPE  '2'            /* reserved */
  56 #define CHRTYPE  '3'            /* character special */
  57 #define BLKTYPE  '4'            /* block special */
  58 #define DIRTYPE  '5'            /* directory */
  59 #define FIFOTYPE '6'            /* FIFO special */
  60 #define CONTTYPE '7'            /* reserved */
  61
  62 #define XHDTYPE  'x'            /* Extended header referring to the
  63                                    next file in the archive */
  64 #define XGLTYPE  'g'            /* Global extended header */
  65
  66 /* Bits used in the mode field, values in octal.  */
  67 #define TSUID    04000          /* set UID on execution */
  68 #define TSGID    02000          /* set GID on execution */
  69 #define TSVTX    01000          /* reserved */
  70                                 /* file permissions */
  71 #define TUREAD   00400          /* read by owner */
  72 #define TUWRITE  00200          /* write by owner */
  73 #define TUEXEC   00100          /* execute/search by owner */
  74 #define TGREAD   00040          /* read by group */
  75 #define TGWRITE  00020          /* write by group */
  76 #define TGEXEC   00010          /* execute/search by group */
  77 #define TOREAD   00004          /* read by other */
  78 #define TOWRITE  00002          /* write by other */
  79 #define TOEXEC   00001          /* execute/search by other */
  80
  81 /* tar Header Block, GNU extensions.  */
  82
  83 /* In GNU tar, SYMTYPE is for to symbolic links, and CONTTYPE is for
  84    contiguous files, so maybe disobeying the "reserved" comment in POSIX
  85    header description.  I suspect these were meant to be used this way, and
  86    should not have really been "reserved" in the published standards.  */
  87
  88 /* *BEWARE* *BEWARE* *BEWARE* that the following information is still
  89    boiling, and may change.  Even if the OLDGNU format description should be
  90    accurate, the so-called GNU format is not yet fully decided.  It is
  91    surely meant to use only extensions allowed by POSIX, but the sketch
  92    below repeats some ugliness from the OLDGNU format, which should rather
  93    go away.  Sparse files should be saved in such a way that they do *not*
  94    require two passes at archive creation time.  Huge files get some POSIX
  95    fields to overflow, alternate solutions have to be sought for this.  */
  96
  97 /* Descriptor for a single file hole.  */
  98
  99 struct sparse
 100 {                               /* byte offset */
 101   char offset[12];              /*   0 */
 102   char numbytes[12];            /*  12 */
 103                                 /*  24 */
 104 };
 105
 106 /* Sparse files are not supported in POSIX ustar format.  For sparse files
 107    with a POSIX header, a GNU extra header is provided which holds overall
 108    sparse information and a few sparse descriptors.  When an old GNU header
 109    replaces both the POSIX header and the GNU extra header, it holds some
 110    sparse descriptors too.  Whether POSIX or not, if more sparse descriptors
 111    are still needed, they are put into as many successive sparse headers as
 112    necessary.  The following constants tell how many sparse descriptors fit
 113    in each kind of header able to hold them.  */
 114
 115 #define SPARSES_IN_EXTRA_HEADER  16
 116 #define SPARSES_IN_OLDGNU_HEADER 4
 117 #define SPARSES_IN_SPARSE_HEADER 21
 118
 119 /* Extension header for sparse files, used immediately after the GNU extra
 120    header, and used only if all sparse information cannot fit into that
 121    extra header.  There might even be many such extension headers, one after
 122    the other, until all sparse information has been recorded.  */
 123
 124 struct sparse_header
 125 {                               /* byte offset */
 126   struct sparse sp[SPARSES_IN_SPARSE_HEADER];
 127                                 /*   0 */
 128   char isextended;              /* 504 */
 129                                 /* 505 */
 130 };
 131
 132 /* The old GNU format header conflicts with POSIX format in such a way that
 133    POSIX archives may fool old GNU tar's, and POSIX tar's might well be
 134    fooled by old GNU tar archives.  An old GNU format header uses the space
 135    used by the prefix field in a POSIX header, and cumulates information
 136    normally found in a GNU extra header.  With an old GNU tar header, we
 137    never see any POSIX header nor GNU extra header.  Supplementary sparse
 138    headers are allowed, however.  */
 139
 140 struct oldgnu_header
 141 {                               /* byte offset */
 142   char unused_pad1[345];        /*   0 */
 143   char atime[12];               /* 345 Incr. archive: atime of the file */
 144   char ctime[12];               /* 357 Incr. archive: ctime of the file */
 145   char offset[12];              /* 369 Multivolume archive: the offset of
 146                                    the start of this volume */
 147   char longnames[4];            /* 381 Not used */
 148   char unused_pad2;             /* 385 */
 149   struct sparse sp[SPARSES_IN_OLDGNU_HEADER];
 150                                 /* 386 */
 151   char isextended;              /* 482 Sparse file: Extension sparse header
 152                                    follows */
 153   char realsize[12];            /* 483 Sparse file: Real size*/
 154                                 /* 495 */
 155 };
 156
 157 /* OLDGNU_MAGIC uses both magic and version fields, which are contiguous.
 158    Found in an archive, it indicates an old GNU header format, which will be
 159    hopefully become obsolescent.  With OLDGNU_MAGIC, uname and gname are
 160    valid, though the header is not truly POSIX conforming.  */
 161 #define OLDGNU_MAGIC "ustar  "  /* 7 chars and a null */
 162
 163 /* The standards committee allows only capital A through capital Z for
 164    user-defined expansion.  Other letters in use include:
 165
 166    'A' Solaris Access Control List
 167    'E' Solaris Extended Attribute File
 168    'I' Inode only, as in 'star'
 169    'N' Obsolete GNU tar, for file names that do not fit into the main header.
 170    'X' POSIX 1003.1-2001 eXtended (VU version)  */
 171
 172 /* This is a dir entry that contains the names of files that were in the
 173    dir at the time the dump was made.  */
 174 #define GNUTYPE_DUMPDIR 'D'
 175
 176 /* Identifies the *next* file on the tape as having a long linkname.  */
 177 #define GNUTYPE_LONGLINK 'K'
 178
 179 /* Identifies the *next* file on the tape as having a long name.  */
 180 #define GNUTYPE_LONGNAME 'L'
 181
 182 /* This is the continuation of a file that began on another volume.  */
 183 #define GNUTYPE_MULTIVOL 'M'
 184
 185 /* This is for sparse files.  */
 186 #define GNUTYPE_SPARSE 'S'
 187
 188 /* This file is a tape/volume header.  Ignore it on extraction.  */
 189 #define GNUTYPE_VOLHDR 'V'
 190
 191 /* Solaris extended header */
 192 #define SOLARIS_XHDTYPE 'X'
 193
 194 /* J@"org Schilling star header */
 195
 196 struct star_header
 197 {                               /* byte offset */
 198   char name[100];               /*   0 */
 199   char mode[8];                 /* 100 */
 200   char uid[8];                  /* 108 */
 201   char gid[8];                  /* 116 */
 202   char size[12];                /* 124 */
 203   char mtime[12];               /* 136 */
 204   char chksum[8];               /* 148 */
 205   char typeflag;                /* 156 */
 206   char linkname[100];           /* 157 */
 207   char magic[6];                /* 257 */
 208   char version[2];              /* 263 */
 209   char uname[32];               /* 265 */
 210   char gname[32];               /* 297 */
 211   char devmajor[8];             /* 329 */
 212   char devminor[8];             /* 337 */
 213   char prefix[131];             /* 345 */
 214   char atime[12];               /* 476 */
 215   char ctime[12];               /* 488 */
 216                                 /* 500 */
 217 };
 218
 219 #define SPARSES_IN_STAR_HEADER      4
 220 #define SPARSES_IN_STAR_EXT_HEADER  21
 221
 222 struct star_in_header
 223 {
 224   char fill[345];       /*   0  Everything that is before t_prefix */
 225   char prefix[1];       /* 345  t_name prefix */
 226   char fill2;           /* 346  */
 227   char fill3[8];        /* 347  */
 228   char isextended;      /* 355  */
 229   struct sparse sp[SPARSES_IN_STAR_HEADER]; /* 356  */
 230   char realsize[12];    /* 452  Actual size of the file */
 231   char offset[12];      /* 464  Offset of multivolume contents */
 232   char atime[12];       /* 476  */
 233   char ctime[12];       /* 488  */
 234   char mfill[8];        /* 500  */
 235   char xmagic[4];       /* 508  "tar" */
 236 };
 237
 238 struct star_ext_header
 239 {
 240   struct sparse sp[SPARSES_IN_STAR_EXT_HEADER];
 241   char isextended;
 242 };
 243
 244 /* END */
 245 \f
 246
 247 /* tar Header Block, overall structure.  */
 248
 249 /* tar files are made in basic blocks of this size.  */
 250 #define BLOCKSIZE 512
 251
 252 enum archive_format
 253 {
 254   DEFAULT_FORMAT,               /* format to be decided later */
 255   V7_FORMAT,                    /* old V7 tar format */
 256   OLDGNU_FORMAT,                /* GNU format as per before tar 1.12 */
 257   USTAR_FORMAT,                 /* POSIX.1-1988 (ustar) format */
 258   POSIX_FORMAT,                 /* POSIX.1-2001 format */
 259   STAR_FORMAT,                  /* Star format defined in 1994 */
 260   GNU_FORMAT                    /* Same as OLDGNU_FORMAT with one exception:
 261                                    see FIXME note for to_chars() function
 262                                    (create.c:189) */
 263 };
 264
 265 /* Information about a sparse file.  */
 266 struct sp_array
 267 {
 268   off_t offset;
 269   off_t numbytes;
 270 };
 271
 272 struct xheader
 273 {
 274   struct obstack *stk;
 275   size_t size;
 276   char *buffer;
 277   uintmax_t string_length;
 278 };
 279
 280 /* Information about xattrs for a file.  */
 281 struct xattr_array
 282   {
 283     char *xkey;
 284     char *xval_ptr;
 285     size_t xval_len;
 286   };
 287
 288 struct tar_stat_info
 289 {
 290   char *orig_file_name;     /* name of file read from the archive header */
 291   char *file_name;          /* name of file for the current archive entry
 292                                after being normalized.  */
 293   bool had_trailing_slash;  /* true if the current archive entry had a
 294                                trailing slash before it was normalized. */
 295   char *link_name;          /* name of link for the current archive entry.  */
 296
 297   char          *uname;     /* user name of owner */
 298   char          *gname;     /* group name of owner */
 299
 300   char *cntx_name;          /* SELinux context for the current archive entry. */
 301
 302   char *acls_a_ptr;         /* Access ACLs for the current archive entry. */
 303   size_t acls_a_len;        /* Access ACLs for the current archive entry. */
 304
 305   char *acls_d_ptr;         /* Default ACLs for the current archive entry. */
 306   size_t acls_d_len;        /* Default ACLs for the current archive entry. */
 307
 308   struct stat   stat;       /* regular filesystem stat */
 309
 310   /* STAT doesn't always have access, data modification, and status
 311      change times in a convenient form, so store them separately.  */
 312   struct timespec atime;
 313   struct timespec mtime;
 314   struct timespec ctime;
 315
 316   off_t archive_file_size;  /* Size of file as stored in the archive.
 317                                Equals stat.st_size for non-sparse files */
 318
 319   bool   is_sparse;         /* Is the file sparse */
 320
 321   /* For sparse files: */
 322   unsigned sparse_major;
 323   unsigned sparse_minor;
 324   size_t sparse_map_avail;  /* Index to the first unused element in
 325                                sparse_map array. Zero if the file is
 326                                not sparse */
 327   size_t sparse_map_size;   /* Size of the sparse map */
 328   struct sp_array *sparse_map;
 329
 330   size_t xattr_map_size;   /* Size of the xattr map */
 331   struct xattr_array *xattr_map;
 332
 333   /* Extended headers */
 334   struct xheader xhdr;
 335
 336   /* For dumpdirs */
 337   bool is_dumpdir;          /* Is the member a dumpdir? */
 338   bool skipped;             /* The member contents is already read
 339                                (for GNUTYPE_DUMPDIR) */
 340   char *dumpdir;            /* Contents of the dump directory */
 341
 342   /* Parent directory, if creating an archive.  This is null if the
 343      file is at the top level.  */
 344   struct tar_stat_info *parent;
 345
 346   /* Directory stream.  If this is not null, it is in control of FD,
 347      and should be closed instead of FD.  */
 348   DIR *dirstream;
 349
 350   /* File descriptor, if creating an archive, and if a directory or a
 351      regular file or a contiguous file.
 352
 353      It is zero if no file descriptor is available, either because it
 354      was never needed or because it was open and then closed to
 355      conserve on file descriptors.  (Standard input is never used
 356      here, so zero cannot be a valid file descriptor.)
 357
 358      It is negative if it could not be reopened after it was closed.
 359      Negate it to find out what errno was when the reopen failed.  */
 360   int fd;
 361 };
 362
 363 union block
 364 {
 365   char buffer[BLOCKSIZE];
 366   struct posix_header header;
 367   struct star_header star_header;
 368   struct oldgnu_header oldgnu_header;
 369   struct sparse_header sparse_header;
 370   struct star_in_header star_in_header;
 371   struct star_ext_header star_ext_header;
 372 };