bin/pax/ftree.c

   1 /*      $OpenBSD: ftree.c,v 1.40 2016/08/26 04:17:48 guenther 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/types.h>
  38 #include <sys/stat.h>
  39 #include <errno.h>
  40 #include <fts.h>
  41 #include <stdio.h>
  42 #include <stdlib.h>
  43 #include <string.h>
  44 #include <unistd.h>
  45
  46 #include "pax.h"
  47 #include "extern.h"
  48
  49 /*
  50  * Data structure used to store the file args to be handed to fts().
  51  * It keeps track of which args generated a "selected" member.
  52  */
  53 typedef struct ftree {
  54         char            *fname;         /* file tree name */
  55         int             refcnt;         /* has tree had a selected file? */
  56         int             newercnt;       /* skipped due to -u/-D */
  57         int             chflg;          /* change directory flag */
  58         struct ftree    *fow;           /* pointer to next entry on list */
  59 } FTREE;
  60
  61
  62 /*
  63  * routines to interface with the fts library function.
  64  *
  65  * file args supplied to pax are stored on a single linked list (of type FTREE)
  66  * and given to fts to be processed one at a time. pax "selects" files from
  67  * the expansion of each arg into the corresponding file tree (if the arg is a
  68  * directory, otherwise the node itself is just passed to pax). The selection
  69  * is modified by the -n and -u flags. The user is informed when a specific
  70  * file arg does not generate any selected files. -n keeps expanding the file
  71  * tree arg until one of its files is selected, then skips to the next file
  72  * arg. when the user does not supply the file trees as command line args to
  73  * pax, they are read from stdin
  74  */
  75
  76 static FTS *ftsp = NULL;                /* current FTS handle */
  77 static int ftsopts;                     /* options to be used on fts_open */
  78 static char *farray[2];                 /* array for passing each arg to fts */
  79 static FTREE *fthead = NULL;            /* head of linked list of file args */
  80 static FTREE *fttail = NULL;            /* tail of linked list of file args */
  81 static FTREE *ftcur = NULL;             /* current file arg being processed */
  82 static FTSENT *ftent = NULL;            /* current file tree entry */
  83 static int ftree_skip;                  /* when set skip to next file arg */
  84
  85 static int ftree_arg(void);
  86 static char *getpathname(char *, int);
  87
  88 /*
  89  * ftree_start()
  90  *      initialize the options passed to fts_open() during this run of pax
  91  *      options are based on the selection of pax options by the user
  92  *      fts_start() also calls fts_arg() to open the first valid file arg. We
  93  *      also attempt to reset directory access times when -t (tflag) is set.
  94  * Return:
  95  *      0 if there is at least one valid file arg to process, -1 otherwise
  96  */
  97
  98 int
  99 ftree_start(void)
 100 {
 101         /*
 102          * set up the operation mode of fts, open the first file arg. We must
 103          * use FTS_NOCHDIR, as the user may have to open multiple archives and
 104          * if fts did a chdir off into the boondocks, we may create an archive
 105          * volume in an place where the user did not expect to.
 106          */
 107         ftsopts = FTS_NOCHDIR;
 108
 109         /*
 110          * optional user flags that effect file traversal
 111          * -H command line symlink follow only (half follow)
 112          * -L follow sylinks (logical)
 113          * -P do not follow sylinks (physical). This is the default.
 114          * -X do not cross over mount points
 115          * -t preserve access times on files read.
 116          * -n select only the first member of a file tree when a match is found
 117          * -d do not extract subtrees rooted at a directory arg.
 118          */
 119         if (Lflag)
 120                 ftsopts |= FTS_LOGICAL;
 121         else
 122                 ftsopts |= FTS_PHYSICAL;
 123         if (Hflag)
 124                 ftsopts |= FTS_COMFOLLOW;
 125         if (Xflag)
 126                 ftsopts |= FTS_XDEV;
 127
 128         if ((fthead == NULL) && ((farray[0] = malloc(PAXPATHLEN+2)) == NULL)) {
 129                 paxwarn(1, "Unable to allocate memory for file name buffer");
 130                 return(-1);
 131         }
 132
 133         if (ftree_arg() < 0)
 134                 return(-1);
 135         if (tflag && (atdir_start() < 0))
 136                 return(-1);
 137         return(0);
 138 }
 139
 140 /*
 141  * ftree_add()
 142  *      add the arg to the linked list of files to process. Each will be
 143  *      processed by fts one at a time
 144  * Return:
 145  *      0 if added to the linked list, -1 if failed
 146  */
 147
 148 int
 149 ftree_add(char *str, int chflg)
 150 {
 151         FTREE *ft;
 152         int len;
 153
 154         /*
 155          * simple check for bad args
 156          */
 157         if ((str == NULL) || (*str == '\0')) {
 158                 paxwarn(0, "Invalid file name argument");
 159                 return(-1);
 160         }
 161
 162         /*
 163          * allocate FTREE node and add to the end of the linked list (args are
 164          * processed in the same order they were passed to pax). Get rid of any
 165          * trailing / the user may pass us. (watch out for / by itself).
 166          */
 167         if ((ft = malloc(sizeof(FTREE))) == NULL) {
 168                 paxwarn(0, "Unable to allocate memory for filename");
 169                 return(-1);
 170         }
 171
 172         if (((len = strlen(str) - 1) > 0) && (str[len] == '/'))
 173                 str[len] = '\0';
 174         ft->fname = str;
 175         ft->refcnt = 0;
 176         ft->newercnt = 0;
 177         ft->chflg = chflg;
 178         ft->fow = NULL;
 179         if (fthead == NULL) {
 180                 fttail = fthead = ft;
 181                 return(0);
 182         }
 183         fttail->fow = ft;
 184         fttail = ft;
 185         return(0);
 186 }
 187
 188 /*
 189  * ftree_sel()
 190  *      this entry has been selected by pax. bump up reference count and handle
 191  *      -n and -d processing.
 192  */
 193
 194 void
 195 ftree_sel(ARCHD *arcn)
 196 {
 197         /*
 198          * set reference bit for this pattern. This linked list is only used
 199          * when file trees are supplied pax as args. The list is not used when
 200          * the trees are read from stdin.
 201          */
 202         if (ftcur != NULL)
 203                 ftcur->refcnt = 1;
 204
 205         /*
 206          * if -n we are done with this arg, force a skip to the next arg when
 207          * pax asks for the next file in next_file().
 208          * if -d we tell fts only to match the directory (if the arg is a dir)
 209          * and not the entire file tree rooted at that point.
 210          */
 211         if (nflag)
 212                 ftree_skip = 1;
 213
 214         if (!dflag || (arcn->type != PAX_DIR))
 215                 return;
 216
 217         if (ftent != NULL)
 218                 (void)fts_set(ftsp, ftent, FTS_SKIP);
 219 }
 220
 221 /*
 222  * ftree_skipped_newer()
 223  *      file has been skipped because a newer file exists and -u/-D given
 224  */
 225
 226 void
 227 ftree_skipped_newer(ARCHD *arcn)
 228 {
 229         /* skipped due to -u/-D, mark accordingly */
 230         if (ftcur != NULL)
 231                 ftcur->newercnt = 1;
 232 }
 233
 234 /*
 235  * ftree_chk()
 236  *      called at end on pax execution. Prints all those file args that did not
 237  *      have a selected member (reference count still 0)
 238  */
 239
 240 void
 241 ftree_chk(void)
 242 {
 243         FTREE *ft;
 244         int wban = 0;
 245
 246         /*
 247          * make sure all dir access times were reset.
 248          */
 249         if (tflag)
 250                 atdir_end();
 251
 252         /*
 253          * walk down list and check reference count. Print out those members
 254          * that never had a match
 255          */
 256         for (ft = fthead; ft != NULL; ft = ft->fow) {
 257                 if ((ft->refcnt > 0) || ft->newercnt > 0 || ft->chflg)
 258                         continue;
 259                 if (wban == 0) {
 260                         paxwarn(1,"WARNING! These file names were not selected:");
 261                         ++wban;
 262                 }
 263                 (void)fprintf(stderr, "%s\n", ft->fname);
 264         }
 265 }
 266
 267 /*
 268  * ftree_arg()
 269  *      Get the next file arg for fts to process. Can be from either the linked
 270  *      list or read from stdin when the user did not them as args to pax. Each
 271  *      arg is processed until the first successful fts_open().
 272  * Return:
 273  *      0 when the next arg is ready to go, -1 if out of file args (or EOF on
 274  *      stdin).
 275  */
 276
 277 static int
 278 ftree_arg(void)
 279 {
 280
 281         /*
 282          * close off the current file tree
 283          */
 284         if (ftsp != NULL) {
 285                 (void)fts_close(ftsp);
 286                 ftsp = NULL;
 287         }
 288
 289         /*
 290          * keep looping until we get a valid file tree to process. Stop when we
 291          * reach the end of the list (or get an eof on stdin)
 292          */
 293         for (;;) {
 294                 if (fthead == NULL) {
 295                         /*
 296                          * the user didn't supply any args, get the file trees
 297                          * to process from stdin;
 298                          */
 299                         if (getpathname(farray[0], PAXPATHLEN+1) == NULL)
 300                                 return(-1);
 301                 } else {
 302                         /*
 303                          * the user supplied the file args as arguments to pax
 304                          */
 305                         if (ftcur == NULL)
 306                                 ftcur = fthead;
 307                         else if ((ftcur = ftcur->fow) == NULL)
 308                                 return(-1);
 309                         if (ftcur->chflg) {
 310                                 /* First fchdir() back... */
 311                                 if (fchdir(cwdfd) < 0) {
 312                                         syswarn(1, errno,
 313                                           "Can't fchdir to starting directory");
 314                                         return(-1);
 315                                 }
 316                                 if (chdir(ftcur->fname) < 0) {
 317                                         syswarn(1, errno, "Can't chdir to %s",
 318                                             ftcur->fname);
 319                                         return(-1);
 320                                 }
 321                                 continue;
 322                         } else
 323                                 farray[0] = ftcur->fname;
 324                 }
 325
 326                 /*
 327                  * watch it, fts wants the file arg stored in a array of char
 328                  * ptrs, with the last one a null. we use a two element array
 329                  * and set farray[0] to point at the buffer with the file name
 330                  * in it. We cannot pass all the file args to fts at one shot
 331                  * as we need to keep a handle on which file arg generates what
 332                  * files (the -n and -d flags need this). If the open is
 333                  * successful, return a 0.
 334                  */
 335                 if ((ftsp = fts_open(farray, ftsopts, NULL)) != NULL)
 336                         break;
 337         }
 338         return(0);
 339 }
 340
 341 /*
 342  * next_file()
 343  *      supplies the next file to process in the supplied archd structure.
 344  * Return:
 345  *      0 when contents of arcn have been set with the next file, -1 when done.
 346  */
 347
 348 int
 349 next_file(ARCHD *arcn)
 350 {
 351         int cnt;
 352
 353         /*
 354          * ftree_sel() might have set the ftree_skip flag if the user has the
 355          * -n option and a file was selected from this file arg tree. (-n says
 356          * only one member is matched for each pattern) ftree_skip being 1
 357          * forces us to go to the next arg now.
 358          */
 359         if (ftree_skip) {
 360                 /*
 361                  * clear and go to next arg
 362                  */
 363                 ftree_skip = 0;
 364                 if (ftree_arg() < 0)
 365                         return(-1);
 366         }
 367
 368         /*
 369          * loop until we get a valid file to process
 370          */
 371         for (;;) {
 372                 if ((ftent = fts_read(ftsp)) == NULL) {
 373                         if (errno)
 374                                 syswarn(1, errno, "next_file");
 375                         /*
 376                          * out of files in this tree, go to next arg, if none
 377                          * we are done
 378                          */
 379                         if (ftree_arg() < 0)
 380                                 return(-1);
 381                         continue;
 382                 }
 383
 384                 /*
 385                  * handle each type of fts_read() flag
 386                  */
 387                 switch (ftent->fts_info) {
 388                 case FTS_D:
 389                 case FTS_DEFAULT:
 390                 case FTS_F:
 391                 case FTS_SL:
 392                 case FTS_SLNONE:
 393                         /*
 394                          * these are all ok
 395                          */
 396                         break;
 397                 case FTS_DP:
 398                         /*
 399                          * already saw this directory. If the user wants file
 400                          * access times reset, we use this to restore the
 401                          * access time for this directory since this is the
 402                          * last time we will see it in this file subtree
 403                          * remember to force the time (this is -t on a read
 404                          * directory, not a created directory).
 405                          */
 406                         if (!tflag)
 407                                 continue;
 408                         do_atdir(ftent->fts_path, ftent->fts_statp->st_dev,
 409                             ftent->fts_statp->st_ino);
 410                         continue;
 411                 case FTS_DC:
 412                         /*
 413                          * fts claims a file system cycle
 414                          */
 415                         paxwarn(1,"File system cycle found at %s",ftent->fts_path);
 416                         continue;
 417                 case FTS_DNR:
 418                         syswarn(1, ftent->fts_errno,
 419                             "Unable to read directory %s", ftent->fts_path);
 420                         continue;
 421                 case FTS_ERR:
 422                         syswarn(1, ftent->fts_errno,
 423                             "File system traversal error");
 424                         continue;
 425                 case FTS_NS:
 426                 case FTS_NSOK:
 427                         syswarn(1, ftent->fts_errno,
 428                             "Unable to access %s", ftent->fts_path);
 429                         continue;
 430                 }
 431
 432                 /*
 433                  * ok got a file tree node to process. copy info into arcn
 434                  * structure (initialize as required)
 435                  */
 436                 arcn->skip = 0;
 437                 arcn->pad = 0;
 438                 arcn->ln_nlen = 0;
 439                 arcn->ln_name[0] = '\0';
 440                 memcpy(&arcn->sb, ftent->fts_statp, sizeof(arcn->sb));
 441
 442                 /*
 443                  * file type based set up and copy into the arcn struct
 444                  * SIDE NOTE:
 445                  * we try to reset the access time on all files and directories
 446                  * we may read when the -t flag is specified. files are reset
 447                  * when we close them after copying. we reset the directories
 448                  * when we are done with their file tree (we also clean up at
 449                  * end in case we cut short a file tree traversal). However
 450                  * there is no way to reset access times on symlinks.
 451                  */
 452                 switch (S_IFMT & arcn->sb.st_mode) {
 453                 case S_IFDIR:
 454                         arcn->type = PAX_DIR;
 455                         if (!tflag)
 456                                 break;
 457                         add_atdir(ftent->fts_path, arcn->sb.st_dev,
 458                             arcn->sb.st_ino, &arcn->sb.st_mtim,
 459                             &arcn->sb.st_atim);
 460                         break;
 461                 case S_IFCHR:
 462                         arcn->type = PAX_CHR;
 463                         break;
 464                 case S_IFBLK:
 465                         arcn->type = PAX_BLK;
 466                         break;
 467                 case S_IFREG:
 468                         /*
 469                          * only regular files with have data to store on the
 470                          * archive. all others will store a zero length skip.
 471                          * the skip field is used by pax for actual data it has
 472                          * to read (or skip over).
 473                          */
 474                         arcn->type = PAX_REG;
 475                         arcn->skip = arcn->sb.st_size;
 476                         break;
 477                 case S_IFLNK:
 478                         arcn->type = PAX_SLK;
 479                         /*
 480                          * have to read the symlink path from the file
 481                          */
 482                         if ((cnt = readlink(ftent->fts_path, arcn->ln_name,
 483                             PAXPATHLEN)) < 0) {
 484                                 syswarn(1, errno, "Unable to read symlink %s",
 485                                     ftent->fts_path);
 486                                 continue;
 487                         }
 488                         /*
 489                          * set link name length, watch out readlink does not
 490                          * NUL terminate the link path
 491                          */
 492                         arcn->ln_name[cnt] = '\0';
 493                         arcn->ln_nlen = cnt;
 494                         break;
 495                 case S_IFSOCK:
 496                         /*
 497                          * under BSD storing a socket is senseless but we will
 498                          * let the format specific write function make the
 499                          * decision of what to do with it.
 500                          */
 501                         arcn->type = PAX_SCK;
 502                         break;
 503                 case S_IFIFO:
 504                         arcn->type = PAX_FIF;
 505                         break;
 506                 }
 507                 break;
 508         }
 509
 510         /*
 511          * copy file name, set file name length
 512          */
 513         arcn->nlen = strlcpy(arcn->name, ftent->fts_path, sizeof(arcn->name));
 514         if (arcn->nlen >= sizeof(arcn->name))
 515                 arcn->nlen = sizeof(arcn->name) - 1; /* XXX truncate? */
 516         arcn->org_name = ftent->fts_path;
 517         return(0);
 518 }
 519
 520 /*
 521  * getpathname()
 522  *      Reads a pathname from stdin, handling NUL- or newline-termination.
 523  * Return:
 524  *      NULL at end of file, otherwise the NUL-terminated buffer.
 525  */
 526
 527 static char *
 528 getpathname(char *buf, int buflen)
 529 {
 530         char *bp, *ep;
 531         int ch, term;
 532
 533         if (zeroflag) {
 534                 /*
 535                  * Read a NUL-terminated pathname, being especially
 536                  * paranoid about proper termination and pathname length.
 537                  */
 538                 for (bp = buf, ep = buf + buflen; bp < ep; bp++) {
 539                         if ((ch = getchar()) == EOF) {
 540                                 if (bp != buf)
 541                                         paxwarn(1, "Ignoring unterminated "
 542                                             "pathname at EOF");
 543                                 return(NULL);
 544                         }
 545                         if ((*bp = ch) == '\0')
 546                                 return(buf);
 547                 }
 548                 /* Too long - skip this path */
 549                 *--bp = '\0';
 550                 term = '\0';
 551         } else {
 552                 if (fgets(buf, buflen, stdin) == NULL)
 553                         return(NULL);
 554                 if ((bp = strchr(buf, '\n')) != NULL || feof(stdin)) {
 555                         if (bp != NULL)
 556                                 *bp = '\0';
 557                         return(buf);
 558                 }
 559                 /* Too long - skip this path */
 560                 term = '\n';
 561         }
 562         while ((ch = getchar()) != term && ch != EOF)
 563                 continue;
 564         paxwarn(1, "Ignoring too-long pathname: %s", buf);
 565         return(NULL);
 566 }