pax/ftree.c

   1 /*      $OpenBSD: ftree.c,v 1.28 2008/05/06 06:54:28 henning 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 #include <sys/cdefs.h>
  38 #ifndef lint
  39 #if 0
  40 static const char sccsid[] = "@(#)ftree.c       8.2 (Berkeley) 4/18/94";
  41 #else
  42 __used static const char rcsid[] = "$OpenBSD: ftree.c,v 1.28 2008/05/06 06:54:28 henning Exp $";
  43 #endif
  44 #endif /* not lint */
  45
  46 #include <sys/types.h>
  47 #include <sys/time.h>
  48 #include <sys/stat.h>
  49 #include <sys/param.h>
  50 #include <unistd.h>
  51 #include <string.h>
  52 #include <stdio.h>
  53 #include <errno.h>
  54 #include <stdlib.h>
  55 #include <fts.h>
  56 #include "pax.h"
  57 #include "ftree.h"
  58 #include "extern.h"
  59
  60 /*
  61  * routines to interface with the fts library function.
  62  *
  63  * file args supplied to pax are stored on a single linked list (of type FTREE)
  64  * and given to fts to be processed one at a time. pax "selects" files from
  65  * the expansion of each arg into the corresponding file tree (if the arg is a
  66  * directory, otherwise the node itself is just passed to pax). The selection
  67  * is modified by the -n and -u flags. The user is informed when a specific
  68  * file arg does not generate any selected files. -n keeps expanding the file
  69  * tree arg until one of its files is selected, then skips to the next file
  70  * arg. when the user does not supply the file trees as command line args to
  71  * pax, they are read from stdin
  72  */
  73
  74 static FTS *ftsp = NULL;                /* current FTS handle */
  75 static int ftsopts;                     /* options to be used on fts_open */
  76 static char *farray[2];                 /* array for passing each arg to fts */
  77 static FTREE *fthead = NULL;            /* head of linked list of file args */
  78 static FTREE *fttail = NULL;            /* tail of linked list of file args */
  79 static FTREE *ftcur = NULL;             /* current file arg being processed */
  80 static FTSENT *ftent = NULL;            /* current file tree entry */
  81 static int ftree_skip;                  /* when set skip to next file arg */
  82
  83 static int ftree_arg(void);
  84 static char *getpathname(char *, int);
  85
  86 /*
  87  * ftree_start()
  88  *      initialize the options passed to fts_open() during this run of pax
  89  *      options are based on the selection of pax options by the user
  90  *      fts_start() also calls fts_arg() to open the first valid file arg. We
  91  *      also attempt to reset directory access times when -t (tflag) is set.
  92  * Return:
  93  *      0 if there is at least one valid file arg to process, -1 otherwise
  94  */
  95
  96 int
  97 ftree_start(void)
  98 {
  99         /*
 100          * set up the operation mode of fts, open the first file arg. We must
 101          * use FTS_NOCHDIR, as the user may have to open multiple archives and
 102          * if fts did a chdir off into the boondocks, we may create an archive
 103          * volume in an place where the user did not expect to.
 104          */
 105         ftsopts = FTS_NOCHDIR;
 106
 107         /*
 108          * optional user flags that effect file traversal
 109          * -H command line symlink follow only (half follow)
 110          * -L follow sylinks (logical)
 111          * -P do not follow sylinks (physical). This is the default.
 112          * -X do not cross over mount points
 113          * -t preserve access times on files read.
 114          * -n select only the first member of a file tree when a match is found
 115          * -d do not extract subtrees rooted at a directory arg.
 116          */
 117         if (Lflag)
 118                 ftsopts |= FTS_LOGICAL;
 119         else
 120                 ftsopts |= FTS_PHYSICAL;
 121         if (Hflag)
 122                 ftsopts |= FTS_COMFOLLOW;
 123         if (Xflag)
 124                 ftsopts |= FTS_XDEV;
 125
 126         if ((fthead == NULL) && ((farray[0] = malloc(PAXPATHLEN+2)) == NULL)) {
 127                 paxwarn(1, "Unable to allocate memory for file name buffer");
 128                 return(-1);
 129         }
 130
 131         if (ftree_arg() < 0)
 132                 return(-1);
 133         if (tflag && (atdir_start() < 0))
 134                 return(-1);
 135         return(0);
 136 }
 137
 138 /*
 139  * ftree_add()
 140  *      add the arg to the linked list of files to process. Each will be
 141  *      processed by fts one at a time
 142  * Return:
 143  *      0 if added to the linked list, -1 if failed
 144  */
 145
 146 int
 147 ftree_add(char *str, int chflg)
 148 {
 149         FTREE *ft;
 150         int len;
 151
 152         /*
 153          * simple check for bad args
 154          */
 155         if ((str == NULL) || (*str == '\0')) {
 156                 paxwarn(0, "Invalid file name argument");
 157                 return(-1);
 158         }
 159
 160         /*
 161          * allocate FTREE node and add to the end of the linked list (args are
 162          * processed in the same order they were passed to pax). Get rid of any
 163          * trailing / the user may pass us. (watch out for / by itself).
 164          */
 165         if ((ft = (FTREE *)malloc(sizeof(FTREE))) == NULL) {
 166                 paxwarn(0, "Unable to allocate memory for filename");
 167                 return(-1);
 168         }
 169
 170         if (((len = strlen(str) - 1) > 0) && (str[len] == '/'))
 171                 str[len] = '\0';
 172         ft->fname = str;
 173         ft->refcnt = 0;
 174         ft->newercnt = 0;
 175         ft->chflg = chflg;
 176         ft->fow = NULL;
 177         if (fthead == NULL) {
 178                 fttail = fthead = ft;
 179                 return(0);
 180         }
 181         fttail->fow = ft;
 182         fttail = ft;
 183         return(0);
 184 }
 185
 186 /*
 187  * ftree_sel()
 188  *      this entry has been selected by pax. bump up reference count and handle
 189  *      -n and -d processing.
 190  */
 191
 192 void
 193 ftree_sel(ARCHD *arcn)
 194 {
 195         /*
 196          * set reference bit for this pattern. This linked list is only used
 197          * when file trees are supplied pax as args. The list is not used when
 198          * the trees are read from stdin.
 199          */
 200         if (ftcur != NULL)
 201                 ftcur->refcnt = 1;
 202
 203         /*
 204          * if -n we are done with this arg, force a skip to the next arg when
 205          * pax asks for the next file in next_file().
 206          * if -d we tell fts only to match the directory (if the arg is a dir)
 207          * and not the entire file tree rooted at that point.
 208          */
 209         if (nflag)
 210                 ftree_skip = 1;
 211
 212         if (!dflag || (arcn->type != PAX_DIR))
 213                 return;
 214
 215         if (ftent != NULL)
 216                 (void)fts_set(ftsp, ftent, FTS_SKIP);
 217 }
 218
 219 /*
 220  * ftree_notsel()
 221  *      this entry has not been selected by pax.
 222  */
 223
 224 void
 225 ftree_notsel()
 226 {
 227         if (ftent != NULL)
 228                 (void)fts_set(ftsp, ftent, FTS_SKIP);
 229 }
 230
 231 /*
 232  * ftree_skipped_newer()
 233  *      file has been skipped because a newer file exists and -u/-D given
 234  */
 235
 236 void
 237 ftree_skipped_newer(ARCHD *arcn)
 238 {
 239         /* skipped due to -u/-D, mark accordingly */
 240         if (ftcur != NULL)
 241                 ftcur->newercnt = 1;
 242 }
 243
 244 /*
 245  * ftree_chk()
 246  *      called at end on pax execution. Prints all those file args that did not
 247  *      have a selected member (reference count still 0)
 248  */
 249
 250 void
 251 ftree_chk(void)
 252 {
 253         FTREE *ft;
 254         int wban = 0;
 255
 256         /*
 257          * make sure all dir access times were reset.
 258          */
 259         if (tflag)
 260                 atdir_end();
 261
 262         /*
 263          * walk down list and check reference count. Print out those members
 264          * that never had a match
 265          */
 266         for (ft = fthead; ft != NULL; ft = ft->fow) {
 267                 if ((ft->refcnt > 0) || ft->newercnt > 0 || ft->chflg)
 268                         continue;
 269                 if (wban == 0) {
 270                         paxwarn(1,"WARNING! These file names were not selected:");
 271                         ++wban;
 272                 }
 273                 (void)fprintf(stderr, "%s\n", ft->fname);
 274         }
 275 }
 276
 277 /*
 278  * ftree_arg()
 279  *      Get the next file arg for fts to process. Can be from either the linked
 280  *      list or read from stdin when the user did not them as args to pax. Each
 281  *      arg is processed until the first successful fts_open().
 282  * Return:
 283  *      0 when the next arg is ready to go, -1 if out of file args (or EOF on
 284  *      stdin).
 285  */
 286
 287 static int
 288 ftree_arg(void)
 289 {
 290
 291         /*
 292          * close off the current file tree
 293          */
 294         if (ftsp != NULL) {
 295                 (void)fts_close(ftsp);
 296                 ftsp = NULL;
 297         }
 298
 299         /*
 300          * keep looping until we get a valid file tree to process. Stop when we
 301          * reach the end of the list (or get an eof on stdin)
 302          */
 303         for (;;) {
 304                 if (fthead == NULL) {
 305                         /*
 306                          * the user didn't supply any args, get the file trees
 307                          * to process from stdin;
 308                          */
 309                         if (getpathname(farray[0], PAXPATHLEN+1) == NULL)
 310                                 return(-1);
 311                 } else {
 312                         /*
 313                          * the user supplied the file args as arguments to pax
 314                          */
 315                         if (ftcur == NULL)
 316                                 ftcur = fthead;
 317                         else if ((ftcur = ftcur->fow) == NULL)
 318                                 return(-1);
 319                         if (ftcur->chflg) {
 320                                 /* First fchdir() back... */
 321                                 if (fdochdir(cwdfd) < 0) {
 322                                         syswarn(1, errno,
 323                                           "Can't fchdir to starting directory");
 324                                         return(-1);
 325                                 }
 326                                 if (dochdir(ftcur->fname) < 0) {
 327                                         syswarn(1, errno, "Can't chdir to %s",
 328                                             ftcur->fname);
 329                                         return(-1);
 330                                 }
 331                                 continue;
 332                         } else
 333                                 farray[0] = ftcur->fname;
 334                 }
 335
 336                 /*
 337                  * watch it, fts wants the file arg stored in a array of char
 338                  * ptrs, with the last one a null. we use a two element array
 339                  * and set farray[0] to point at the buffer with the file name
 340                  * in it. We cannot pass all the file args to fts at one shot
 341                  * as we need to keep a handle on which file arg generates what
 342                  * files (the -n and -d flags need this). If the open is
 343                  * successful, return a 0.
 344                  */
 345                 if ((ftsp = fts_open(farray, ftsopts, NULL)) != NULL)
 346                         break;
 347         }
 348         return(0);
 349 }
 350
 351 /*
 352  * next_file()
 353  *      supplies the next file to process in the supplied archd structure.
 354  * Return:
 355  *      0 when contents of arcn have been set with the next file, -1 when done.
 356  */
 357
 358 int
 359 next_file(ARCHD *arcn)
 360 {
 361         int cnt;
 362         time_t atime;
 363         time_t mtime;
 364
 365         /*
 366          * ftree_sel() might have set the ftree_skip flag if the user has the
 367          * -n option and a file was selected from this file arg tree. (-n says
 368          * only one member is matched for each pattern) ftree_skip being 1
 369          * forces us to go to the next arg now.
 370          */
 371         if (ftree_skip) {
 372                 /*
 373                  * clear and go to next arg
 374                  */
 375                 ftree_skip = 0;
 376                 if (ftree_arg() < 0)
 377                         return(-1);
 378         }
 379
 380         /*
 381          * loop until we get a valid file to process
 382          */
 383         for (;;) {
 384                 if ((ftent = fts_read(ftsp)) == NULL) {
 385                         if (errno)
 386                                 syswarn(1, errno, "next_file");
 387                         /*
 388                          * out of files in this tree, go to next arg, if none
 389                          * we are done
 390                          */
 391                         if (ftree_arg() < 0)
 392                                 return(-1);
 393                         continue;
 394                 }
 395
 396                 /*
 397                  * handle each type of fts_read() flag
 398                  */
 399                 switch (ftent->fts_info) {
 400                 case FTS_D:
 401                 case FTS_DEFAULT:
 402                 case FTS_F:
 403                 case FTS_SL:
 404                         /*
 405                          * these are all ok
 406                          */
 407                         break;
 408                 case FTS_SLNONE:        /* was same as above cases except Unix
 409                                            conformance requires this error check */
 410                         if (Hflag || Lflag) {       /* -H or -L was specified */
 411                                 if (ftent->fts_errno)
 412                                         paxwarn(1, "%s: %s",
 413                                                 ftent->fts_name, strerror(ftent->fts_errno));
 414                         }
 415                         break;
 416                 case FTS_DP:
 417                         /*
 418                          * already saw this directory. If the user wants file
 419                          * access times reset, we use this to restore the
 420                          * access time for this directory since this is the
 421                          * last time we will see it in this file subtree
 422                          * remember to force the time (this is -t on a read
 423                          * directory, not a created directory).
 424                          */
 425                         if (!tflag || (get_atdir(ftent->fts_statp->st_dev,
 426                             ftent->fts_statp->st_ino, &mtime, &atime) < 0))
 427                                 continue;
 428                         set_ftime(ftent->fts_path, mtime, atime, 1);
 429                         continue;
 430                 case FTS_DC:
 431                         /*
 432                          * fts claims a file system cycle
 433                          */
 434                         paxwarn(1,"File system cycle found at %s",ftent->fts_path);
 435                         continue;
 436                 case FTS_DNR:
 437                         syswarn(1, ftent->fts_errno,
 438                             "Unable to read directory %s", ftent->fts_path);
 439                         continue;
 440                 case FTS_ERR:
 441                         syswarn(1, ftent->fts_errno,
 442                             "File system traversal error");
 443                         continue;
 444                 case FTS_NS:
 445                 case FTS_NSOK:
 446                         syswarn(1, ftent->fts_errno,
 447                             "Unable to access %s", ftent->fts_path);
 448                         continue;
 449                 }
 450
 451                 /*
 452                  * ok got a file tree node to process. copy info into arcn
 453                  * structure (initialize as required)
 454                  */
 455                 arcn->skip = 0;
 456                 arcn->pad = 0;
 457                 arcn->ln_nlen = 0;
 458                 arcn->ln_name[0] = '\0';
 459                 memcpy(&arcn->sb, ftent->fts_statp, sizeof(arcn->sb));
 460
 461                 /*
 462                  * file type based set up and copy into the arcn struct
 463                  * SIDE NOTE:
 464                  * we try to reset the access time on all files and directories
 465                  * we may read when the -t flag is specified. files are reset
 466                  * when we close them after copying. we reset the directories
 467                  * when we are done with their file tree (we also clean up at
 468                  * end in case we cut short a file tree traversal). However
 469                  * there is no way to reset access times on symlinks.
 470                  */
 471                 switch (S_IFMT & arcn->sb.st_mode) {
 472                 case S_IFDIR:
 473                         arcn->type = PAX_DIR;
 474                         if (!tflag)
 475                                 break;
 476                         add_atdir(ftent->fts_path, arcn->sb.st_dev,
 477                             arcn->sb.st_ino, arcn->sb.st_mtime,
 478                             arcn->sb.st_atime);
 479                         break;
 480                 case S_IFCHR:
 481                         arcn->type = PAX_CHR;
 482                         break;
 483                 case S_IFBLK:
 484                         arcn->type = PAX_BLK;
 485                         break;
 486                 case S_IFREG:
 487                         /*
 488                          * only regular files with have data to store on the
 489                          * archive. all others will store a zero length skip.
 490                          * the skip field is used by pax for actual data it has
 491                          * to read (or skip over).
 492                          */
 493                         arcn->type = PAX_REG;
 494                         arcn->skip = arcn->sb.st_size;
 495                         break;
 496                 case S_IFLNK:
 497                         arcn->type = PAX_SLK;
 498                         /*
 499                          * have to read the symlink path from the file
 500                          */
 501                         if ((cnt = readlink(ftent->fts_path, arcn->ln_name,
 502                             PAXPATHLEN)) < 0) {
 503                                 syswarn(1, errno, "Unable to read symlink %s",
 504                                     ftent->fts_path);
 505                                 continue;
 506                         }
 507                         /*
 508                          * set link name length, watch out readlink does not
 509                          * always NUL terminate the link path
 510                          */
 511                         arcn->ln_name[cnt] = '\0';
 512                         arcn->ln_nlen = cnt;
 513                         break;
 514                 case S_IFSOCK:
 515                         /*
 516                          * under BSD storing a socket is senseless but we will
 517                          * let the format specific write function make the
 518                          * decision of what to do with it.
 519                          */
 520                         arcn->type = PAX_SCK;
 521                         break;
 522                 case S_IFIFO:
 523                         arcn->type = PAX_FIF;
 524                         break;
 525                 }
 526                 break;
 527         }
 528
 529         /*
 530          * copy file name, set file name length
 531          */
 532         arcn->nlen = strlcpy(arcn->name, ftent->fts_path, sizeof(arcn->name));
 533         if (arcn->nlen >= sizeof(arcn->name))
 534                 arcn->nlen = sizeof(arcn->name) - 1; /* XXX truncate? */
 535         arcn->org_name = ftent->fts_path;
 536         return(0);
 537 }
 538
 539 /*
 540  * getpathname()
 541  *      Reads a pathname from stdin, handling NUL- or newline-termination.
 542  * Return:
 543  *      NULL at end of file, otherwise the NUL-terminated buffer.
 544  */
 545
 546 static char *
 547 getpathname(char *buf, int buflen)
 548 {
 549         char *bp, *ep;
 550         int ch, term;
 551
 552         if (zeroflag) {
 553                 /*
 554                  * Read a NUL-terminated pathname, being especially
 555                  * paranoid about proper termination and pathname length.
 556                  */
 557                 for (bp = buf, ep = buf + buflen; bp < ep; bp++) {
 558                         if ((ch = getchar()) == EOF) {
 559                                 if (bp != buf)
 560                                         paxwarn(1, "Ignoring unterminated "
 561                                             "pathname at EOF");
 562                                 return(NULL);
 563                         }
 564                         if ((*bp = ch) == '\0')
 565                                 return(buf);
 566                 }
 567                 /* Too long - skip this path */
 568                 *--bp = '\0';
 569                 term = '\0';
 570         } else {
 571                 if (fgets(buf, buflen, stdin) == NULL)
 572                         return(NULL);
 573                 if ((bp = strchr(buf, '\n')) != NULL || feof(stdin)) {
 574                         if (bp != NULL)
 575                                 *bp = '\0';
 576                         return(buf);
 577                 }
 578                 /* Too long - skip this path */
 579                 term = '\n';
 580         }
 581         while ((ch = getchar()) != term && ch != EOF)
 582                 ;
 583         paxwarn(1, "Ignoring too-long pathname: %s", buf);
 584         return(NULL);
 585 }