pax/ftree.c

   1 /*      $OpenBSD: ftree.c,v 1.25 2004/04/16 22:50:23 deraadt Exp $      */
   2 /*      $NetBSD: ftree.c,v 1.4 1995/03/21 09:07:21 cgd Exp $    */
   3
   4 /*-
   5  * Copyright (c) 1992 Keith Muller.
   6  * Copyright (c) 1992, 1993
   7  *      The Regents of the University of California.  All rights reserved.
   8  *
   9  * This code is derived from software contributed to Berkeley by
  10  * Keith Muller of the University of California, San Diego.
  11  *
  12  * Redistribution and use in source and binary forms, with or without
  13  * modification, are permitted provided that the following conditions
  14  * are met:
  15  * 1. Redistributions of source code must retain the above copyright
  16  *    notice, this list of conditions and the following disclaimer.
  17  * 2. Redistributions in binary form must reproduce the above copyright
  18  *    notice, this list of conditions and the following disclaimer in the
  19  *    documentation and/or other materials provided with the distribution.
  20  * 3. Neither the name of the University nor the names of its contributors
  21  *    may be used to endorse or promote products derived from this software
  22  *    without specific prior written permission.
  23  *
  24  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
  25  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
  26  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
  27  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
  28  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
  29  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
  30  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
  31  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
  32  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
  33  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
  34  * SUCH DAMAGE.
  35  */
  36
  37 #ifndef lint
  38 #if 0
  39 static const char sccsid[] = "@(#)ftree.c       8.2 (Berkeley) 4/18/94";
  40 #else
  41 static const char rcsid[] __attribute__((__unused__)) = "$OpenBSD: ftree.c,v 1.25 2004/04/16 22:50:23 deraadt Exp $";
  42 #endif
  43 #endif /* not lint */
  44
  45 #include <sys/types.h>
  46 #include <sys/time.h>
  47 #include <sys/stat.h>
  48 #include <sys/param.h>
  49 #include <unistd.h>
  50 #include <string.h>
  51 #include <stdio.h>
  52 #include <errno.h>
  53 #include <stdlib.h>
  54 #include <fts.h>
  55 #include "pax.h"
  56 #include "ftree.h"
  57 #include "extern.h"
  58
  59 /*
  60  * routines to interface with the fts library function.
  61  *
  62  * file args supplied to pax are stored on a single linked list (of type FTREE)
  63  * and given to fts to be processed one at a time. pax "selects" files from
  64  * the expansion of each arg into the corresponding file tree (if the arg is a
  65  * directory, otherwise the node itself is just passed to pax). The selection
  66  * is modified by the -n and -u flags. The user is informed when a specific
  67  * file arg does not generate any selected files. -n keeps expanding the file
  68  * tree arg until one of its files is selected, then skips to the next file
  69  * arg. when the user does not supply the file trees as command line args to
  70  * pax, they are read from stdin
  71  */
  72
  73 static FTS *ftsp = NULL;                /* current FTS handle */
  74 static int ftsopts;                     /* options to be used on fts_open */
  75 static char *farray[2];                 /* array for passing each arg to fts */
  76 static FTREE *fthead = NULL;            /* head of linked list of file args */
  77 static FTREE *fttail = NULL;            /* tail of linked list of file args */
  78 static FTREE *ftcur = NULL;             /* current file arg being processed */
  79 static FTSENT *ftent = NULL;            /* current file tree entry */
  80 static int ftree_skip;                  /* when set skip to next file arg */
  81
  82 static int ftree_arg(void);
  83 static char *getpathname(char *, int);
  84
  85 /*
  86  * ftree_start()
  87  *      initialize the options passed to fts_open() during this run of pax
  88  *      options are based on the selection of pax options by the user
  89  *      fts_start() also calls fts_arg() to open the first valid file arg. We
  90  *      also attempt to reset directory access times when -t (tflag) is set.
  91  * Return:
  92  *      0 if there is at least one valid file arg to process, -1 otherwise
  93  */
  94
  95 int
  96 ftree_start(void)
  97 {
  98         /*
  99          * set up the operation mode of fts, open the first file arg. We must
 100          * use FTS_NOCHDIR, as the user may have to open multiple archives and
 101          * if fts did a chdir off into the boondocks, we may create an archive
 102          * volume in an place where the user did not expect to.
 103          */
 104         ftsopts = FTS_NOCHDIR;
 105
 106         /*
 107          * optional user flags that effect file traversal
 108          * -H command line symlink follow only (half follow)
 109          * -L follow sylinks (logical)
 110          * -P do not follow sylinks (physical). This is the default.
 111          * -X do not cross over mount points
 112          * -t preserve access times on files read.
 113          * -n select only the first member of a file tree when a match is found
 114          * -d do not extract subtrees rooted at a directory arg.
 115          */
 116         if (Lflag)
 117                 ftsopts |= FTS_LOGICAL;
 118         else
 119                 ftsopts |= FTS_PHYSICAL;
 120         if (Hflag)
 121                 ftsopts |= FTS_COMFOLLOW;
 122         if (Xflag)
 123                 ftsopts |= FTS_XDEV;
 124
 125         if ((fthead == NULL) && ((farray[0] = malloc(PAXPATHLEN+2)) == NULL)) {
 126                 paxwarn(1, "Unable to allocate memory for file name buffer");
 127                 return(-1);
 128         }
 129
 130         if (ftree_arg() < 0)
 131                 return(-1);
 132         if (tflag && (atdir_start() < 0))
 133                 return(-1);
 134         return(0);
 135 }
 136
 137 /*
 138  * ftree_add()
 139  *      add the arg to the linked list of files to process. Each will be
 140  *      processed by fts one at a time
 141  * Return:
 142  *      0 if added to the linked list, -1 if failed
 143  */
 144
 145 int
 146 ftree_add(char *str, int chflg)
 147 {
 148         FTREE *ft;
 149         int len;
 150
 151         /*
 152          * simple check for bad args
 153          */
 154         if ((str == NULL) || (*str == '\0')) {
 155                 paxwarn(0, "Invalid file name argument");
 156                 return(-1);
 157         }
 158
 159         /*
 160          * allocate FTREE node and add to the end of the linked list (args are
 161          * processed in the same order they were passed to pax). Get rid of any
 162          * trailing / the user may pass us. (watch out for / by itself).
 163          */
 164         if ((ft = (FTREE *)malloc(sizeof(FTREE))) == NULL) {
 165                 paxwarn(0, "Unable to allocate memory for filename");
 166                 return(-1);
 167         }
 168
 169         if (((len = strlen(str) - 1) > 0) && (str[len] == '/'))
 170                 str[len] = '\0';
 171         ft->fname = str;
 172         ft->refcnt = 0;
 173         ft->chflg = chflg;
 174         ft->fow = NULL;
 175         if (fthead == NULL) {
 176                 fttail = fthead = ft;
 177                 return(0);
 178         }
 179         fttail->fow = ft;
 180         fttail = ft;
 181         return(0);
 182 }
 183
 184 /*
 185  * ftree_sel()
 186  *      this entry has been selected by pax. bump up reference count and handle
 187  *      -n and -d processing.
 188  */
 189
 190 void
 191 ftree_sel(ARCHD *arcn)
 192 {
 193         /*
 194          * set reference bit for this pattern. This linked list is only used
 195          * when file trees are supplied pax as args. The list is not used when
 196          * the trees are read from stdin.
 197          */
 198         if (ftcur != NULL)
 199                 ftcur->refcnt = 1;
 200
 201         /*
 202          * if -n we are done with this arg, force a skip to the next arg when
 203          * pax asks for the next file in next_file().
 204          * if -d we tell fts only to match the directory (if the arg is a dir)
 205          * and not the entire file tree rooted at that point.
 206          */
 207         if (nflag)
 208                 ftree_skip = 1;
 209
 210         if (!dflag || (arcn->type != PAX_DIR))
 211                 return;
 212
 213         if (ftent != NULL)
 214                 (void)fts_set(ftsp, ftent, FTS_SKIP);
 215 }
 216
 217 /*
 218  * ftree_chk()
 219  *      called at end on pax execution. Prints all those file args that did not
 220  *      have a selected member (reference count still 0)
 221  */
 222
 223 void
 224 ftree_chk(void)
 225 {
 226         FTREE *ft;
 227         int wban = 0;
 228
 229         /*
 230          * make sure all dir access times were reset.
 231          */
 232         if (tflag)
 233                 atdir_end();
 234
 235         /*
 236          * walk down list and check reference count. Print out those members
 237          * that never had a match
 238          */
 239         for (ft = fthead; ft != NULL; ft = ft->fow) {
 240                 if ((ft->refcnt > 0) || ft->chflg)
 241                         continue;
 242                 if (wban == 0) {
 243                         paxwarn(1,"WARNING! These file names were not selected:");
 244                         ++wban;
 245                 }
 246                 (void)fprintf(stderr, "%s\n", ft->fname);
 247         }
 248 }
 249
 250 /*
 251  * ftree_arg()
 252  *      Get the next file arg for fts to process. Can be from either the linked
 253  *      list or read from stdin when the user did not them as args to pax. Each
 254  *      arg is processed until the first successful fts_open().
 255  * Return:
 256  *      0 when the next arg is ready to go, -1 if out of file args (or EOF on
 257  *      stdin).
 258  */
 259
 260 static int
 261 ftree_arg(void)
 262 {
 263
 264         /*
 265          * close off the current file tree
 266          */
 267         if (ftsp != NULL) {
 268                 (void)fts_close(ftsp);
 269                 ftsp = NULL;
 270         }
 271
 272         /*
 273          * keep looping until we get a valid file tree to process. Stop when we
 274          * reach the end of the list (or get an eof on stdin)
 275          */
 276         for (;;) {
 277                 if (fthead == NULL) {
 278                         /*
 279                          * the user didn't supply any args, get the file trees
 280                          * to process from stdin;
 281                          */
 282                         if (getpathname(farray[0], PAXPATHLEN+1) == NULL)
 283                                 return(-1);
 284                 } else {
 285                         /*
 286                          * the user supplied the file args as arguments to pax
 287                          */
 288                         if (ftcur == NULL)
 289                                 ftcur = fthead;
 290                         else if ((ftcur = ftcur->fow) == NULL)
 291                                 return(-1);
 292                         if (ftcur->chflg) {
 293                                 /* First fchdir() back... */
 294                                 if (fchdir(cwdfd) < 0) {
 295                                         syswarn(1, errno,
 296                                           "Can't fchdir to starting directory");
 297                                         return(-1);
 298                                 }
 299                                 if (chdir(ftcur->fname) < 0) {
 300                                         syswarn(1, errno, "Can't chdir to %s",
 301                                             ftcur->fname);
 302                                         return(-1);
 303                                 }
 304                                 continue;
 305                         } else
 306                                 farray[0] = ftcur->fname;
 307                 }
 308
 309                 /*
 310                  * watch it, fts wants the file arg stored in a array of char
 311                  * ptrs, with the last one a null. we use a two element array
 312                  * and set farray[0] to point at the buffer with the file name
 313                  * in it. We cannot pass all the file args to fts at one shot
 314                  * as we need to keep a handle on which file arg generates what
 315                  * files (the -n and -d flags need this). If the open is
 316                  * successful, return a 0.
 317                  */
 318                 if ((ftsp = fts_open(farray, ftsopts, NULL)) != NULL)
 319                         break;
 320         }
 321         return(0);
 322 }
 323
 324 /*
 325  * next_file()
 326  *      supplies the next file to process in the supplied archd structure.
 327  * Return:
 328  *      0 when contents of arcn have been set with the next file, -1 when done.
 329  */
 330
 331 int
 332 next_file(ARCHD *arcn)
 333 {
 334         int cnt;
 335         time_t atime;
 336         time_t mtime;
 337
 338         /*
 339          * ftree_sel() might have set the ftree_skip flag if the user has the
 340          * -n option and a file was selected from this file arg tree. (-n says
 341          * only one member is matched for each pattern) ftree_skip being 1
 342          * forces us to go to the next arg now.
 343          */
 344         if (ftree_skip) {
 345                 /*
 346                  * clear and go to next arg
 347                  */
 348                 ftree_skip = 0;
 349                 if (ftree_arg() < 0)
 350                         return(-1);
 351         }
 352
 353         /*
 354          * loop until we get a valid file to process
 355          */
 356         for (;;) {
 357                 if ((ftent = fts_read(ftsp)) == NULL) {
 358                         /*
 359                          * out of files in this tree, go to next arg, if none
 360                          * we are done
 361                          */
 362                         if (ftree_arg() < 0)
 363                                 return(-1);
 364                         continue;
 365                 }
 366
 367                 /*
 368                  * handle each type of fts_read() flag
 369                  */
 370                 switch (ftent->fts_info) {
 371                 case FTS_D:
 372                 case FTS_DEFAULT:
 373                 case FTS_F:
 374                 case FTS_SL:
 375                         /*
 376                          * these are all ok
 377                          */
 378                         break;
 379                 case FTS_SLNONE:        /* was same as above cases except Unix
 380                                            conformance requires this error check */
 381                         if (Hflag || Lflag) {       /* -H or -L was specified */
 382                                 if (ftent->fts_errno)
 383                                         paxwarn(1, "%s: %s",
 384                                                 ftent->fts_name, strerror(ftent->fts_errno));
 385                         }
 386                         break;
 387                 case FTS_DP:
 388                         /*
 389                          * already saw this directory. If the user wants file
 390                          * access times reset, we use this to restore the
 391                          * access time for this directory since this is the
 392                          * last time we will see it in this file subtree
 393                          * remember to force the time (this is -t on a read
 394                          * directory, not a created directory).
 395                          */
 396                         if (!tflag || (get_atdir(ftent->fts_statp->st_dev,
 397                             ftent->fts_statp->st_ino, &mtime, &atime) < 0))
 398                                 continue;
 399                         set_ftime(ftent->fts_path, mtime, atime, 1);
 400                         continue;
 401                 case FTS_DC:
 402                         /*
 403                          * fts claims a file system cycle
 404                          */
 405                         paxwarn(1,"File system cycle found at %s",ftent->fts_path);
 406                         continue;
 407                 case FTS_DNR:
 408                         syswarn(1, ftent->fts_errno,
 409                             "Unable to read directory %s", ftent->fts_path);
 410                         continue;
 411                 case FTS_ERR:
 412                         syswarn(1, ftent->fts_errno,
 413                             "File system traversal error");
 414                         continue;
 415                 case FTS_NS:
 416                 case FTS_NSOK:
 417                         syswarn(1, ftent->fts_errno,
 418                             "Unable to access %s", ftent->fts_path);
 419                         continue;
 420                 }
 421
 422                 /*
 423                  * ok got a file tree node to process. copy info into arcn
 424                  * structure (initialize as required)
 425                  */
 426                 arcn->skip = 0;
 427                 arcn->pad = 0;
 428                 arcn->ln_nlen = 0;
 429                 arcn->ln_name[0] = '\0';
 430                 memcpy(&arcn->sb, ftent->fts_statp, sizeof(arcn->sb));
 431
 432                 /*
 433                  * file type based set up and copy into the arcn struct
 434                  * SIDE NOTE:
 435                  * we try to reset the access time on all files and directories
 436                  * we may read when the -t flag is specified. files are reset
 437                  * when we close them after copying. we reset the directories
 438                  * when we are done with their file tree (we also clean up at
 439                  * end in case we cut short a file tree traversal). However
 440                  * there is no way to reset access times on symlinks.
 441                  */
 442                 switch (S_IFMT & arcn->sb.st_mode) {
 443                 case S_IFDIR:
 444                         arcn->type = PAX_DIR;
 445                         if (!tflag)
 446                                 break;
 447                         add_atdir(ftent->fts_path, arcn->sb.st_dev,
 448                             arcn->sb.st_ino, arcn->sb.st_mtime,
 449                             arcn->sb.st_atime);
 450                         break;
 451                 case S_IFCHR:
 452                         arcn->type = PAX_CHR;
 453                         break;
 454                 case S_IFBLK:
 455                         arcn->type = PAX_BLK;
 456                         break;
 457                 case S_IFREG:
 458                         /*
 459                          * only regular files with have data to store on the
 460                          * archive. all others will store a zero length skip.
 461                          * the skip field is used by pax for actual data it has
 462                          * to read (or skip over).
 463                          */
 464                         arcn->type = PAX_REG;
 465                         arcn->skip = arcn->sb.st_size;
 466                         break;
 467                 case S_IFLNK:
 468                         arcn->type = PAX_SLK;
 469                         /*
 470                          * have to read the symlink path from the file
 471                          */
 472                         if ((cnt = readlink(ftent->fts_path, arcn->ln_name,
 473                             PAXPATHLEN)) < 0) {
 474                                 syswarn(1, errno, "Unable to read symlink %s",
 475                                     ftent->fts_path);
 476                                 continue;
 477                         }
 478                         /*
 479                          * set link name length, watch out readlink does not
 480                          * always NUL terminate the link path
 481                          */
 482                         arcn->ln_name[cnt] = '\0';
 483                         arcn->ln_nlen = cnt;
 484                         break;
 485                 case S_IFSOCK:
 486                         /*
 487                          * under BSD storing a socket is senseless but we will
 488                          * let the format specific write function make the
 489                          * decision of what to do with it.
 490                          */
 491                         arcn->type = PAX_SCK;
 492                         break;
 493                 case S_IFIFO:
 494                         arcn->type = PAX_FIF;
 495                         break;
 496                 }
 497                 break;
 498         }
 499
 500         /*
 501          * copy file name, set file name length
 502          */
 503         arcn->nlen = strlcpy(arcn->name, ftent->fts_path, sizeof(arcn->name));
 504         arcn->org_name = ftent->fts_path;
 505         return(0);
 506 }
 507
 508 /*
 509  * getpathname()
 510  *      Reads a pathname from stdin, handling NUL- or newline-termination.
 511  * Return:
 512  *      NULL at end of file, otherwise the NUL-terminated buffer.
 513  */
 514
 515 static char *
 516 getpathname(char *buf, int buflen)
 517 {
 518         char *bp, *ep;
 519         int ch, term;
 520
 521         if (zeroflag) {
 522                 /*
 523                  * Read a NUL-terminated pathname, being especially
 524                  * paranoid about proper termination and pathname length.
 525                  */
 526                 for (bp = buf, ep = buf + buflen; bp < ep; bp++) {
 527                         if ((ch = getchar()) == EOF) {
 528                                 if (bp != buf)
 529                                         paxwarn(1, "Ignoring unterminated "
 530                                             "pathname at EOF");
 531                                 return(NULL);
 532                         }
 533                         if ((*bp = ch) == '\0')
 534                                 return(buf);
 535                 }
 536                 /* Too long - skip this path */
 537                 *--bp = '\0';
 538                 term = '\0';
 539         } else {
 540                 if (fgets(buf, buflen, stdin) == NULL)
 541                         return(NULL);
 542                 if ((bp = strchr(buf, '\n')) != NULL || feof(stdin)) {
 543                         if (bp != NULL)
 544                                 *bp = '\0';
 545                         return(buf);
 546                 }
 547                 /* Too long - skip this path */
 548                 term = '\n';
 549         }
 550         while ((ch = getchar()) != term && ch != EOF)
 551                 ;
 552         paxwarn(1, "Ignoring too-long pathname: %s", buf);
 553         return(NULL);
 554 }