hlink.c

   1 /*
   2  * Routines to support hard-linking.
   3  *
   4  * Copyright (C) 1996 Andrew Tridgell
   5  * Copyright (C) 1996 Paul Mackerras
   6  * Copyright (C) 2002 Martin Pool <mbp@samba.org>
   7  * Copyright (C) 2004-2008 Wayne Davison
   8  *
   9  * This program is free software; you can redistribute it and/or modify
  10  * it under the terms of the GNU General Public License as published by
  11  * the Free Software Foundation; either version 3 of the License, or
  12  * (at your option) any later version.
  13  *
  14  * This program is distributed in the hope that it will be useful,
  15  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  16  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  17  * GNU General Public License for more details.
  18  *
  19  * You should have received a copy of the GNU General Public License along
  20  * with this program; if not, visit the http://fsf.org website.
  21  */
  22
  23 #include "rsync.h"
  24
  25 extern int verbose;
  26 extern int dry_run;
  27 extern int list_only;
  28 extern int am_sender;
  29 extern int inc_recurse;
  30 extern int do_xfers;
  31 extern int link_dest;
  32 extern int preserve_acls;
  33 extern int make_backups;
  34 extern int protocol_version;
  35 extern int remove_source_files;
  36 extern int stdout_format_has_i;
  37 extern int maybe_ATTRS_REPORT;
  38 extern int unsort_ndx;
  39 extern char *basis_dir[];
  40 extern struct file_list *cur_flist, *first_flist;
  41
  42 #ifdef SUPPORT_HARD_LINKS
  43
  44 /* Starting with protocol 30, we use a simple hashtable on the sending side
  45  * for hashing the st_dev and st_ino info.  The receiving side gets told
  46  * (via flags and a "group index") which items are hard-linked together, so
  47  * we can avoid the pool of dev+inode data.  For incremental recursion mode,
  48  * the receiver will use a ndx hash to remember old pathnames. */
  49
  50 static struct hashtable *dev_tbl;
  51
  52 static struct hashtable *prior_hlinks;
  53
  54 static struct file_list *hlink_flist;
  55
  56 void init_hard_links(void)
  57 {
  58         if (am_sender || protocol_version < 30)
  59                 dev_tbl = hashtable_create(16, SIZEOF_INT64 == 8);
  60         else if (inc_recurse)
  61                 prior_hlinks = hashtable_create(1024, 0);
  62 }
  63
  64 struct ht_int64_node *idev_find(int64 dev, int64 ino)
  65 {
  66         static struct ht_int64_node *dev_node = NULL;
  67         struct hashtable *tbl;
  68
  69         if (!dev_node || dev_node->key != dev) {
  70                 /* We keep a separate hash table of inodes for every device. */
  71                 dev_node = hashtable_find(dev_tbl, dev, 1);
  72                 if (!(tbl = dev_node->data))
  73                         tbl = dev_node->data = hashtable_create(512, SIZEOF_INT64 == 8);
  74         } else
  75                 tbl = dev_node->data;
  76
  77         return hashtable_find(tbl, ino, 1);
  78 }
  79
  80 void idev_destroy(void)
  81 {
  82         int i;
  83
  84         for (i = 0; i < dev_tbl->size; i++) {
  85                 struct ht_int32_node *node = HT_NODE(dev_tbl, dev_tbl->nodes, i);
  86                 if (node->data)
  87                         hashtable_destroy(node->data);
  88         }
  89
  90         hashtable_destroy(dev_tbl);
  91 }
  92
  93 static int hlink_compare_gnum(int *int1, int *int2)
  94 {
  95         struct file_struct *f1 = hlink_flist->sorted[*int1];
  96         struct file_struct *f2 = hlink_flist->sorted[*int2];
  97         int32 gnum1 = F_HL_GNUM(f1);
  98         int32 gnum2 = F_HL_GNUM(f2);
  99
 100         if (gnum1 != gnum2)
 101                 return gnum1 > gnum2 ? 1 : -1;
 102
 103         return *int1 > *int2 ? 1 : -1;
 104 }
 105
 106 static void match_gnums(int32 *ndx_list, int ndx_count)
 107 {
 108         int32 from, prev;
 109         struct file_struct *file, *file_next;
 110         struct ht_int32_node *node = NULL;
 111         int32 gnum, gnum_next;
 112
 113         qsort(ndx_list, ndx_count, sizeof ndx_list[0],
 114              (int (*)()) hlink_compare_gnum);
 115
 116         for (from = 0; from < ndx_count; from++) {
 117                 file = hlink_flist->sorted[ndx_list[from]];
 118                 gnum = F_HL_GNUM(file);
 119                 if (inc_recurse) {
 120                         node = hashtable_find(prior_hlinks, gnum, 1);
 121                         if (!node->data) {
 122                                 node->data = new_array0(char, 5);
 123                                 assert(gnum >= hlink_flist->ndx_start);
 124                                 file->flags |= FLAG_HLINK_FIRST;
 125                                 prev = -1;
 126                         } else if (CVAL(node->data, 0) == 0) {
 127                                 struct file_list *flist;
 128                                 prev = IVAL(node->data, 1);
 129                                 flist = flist_for_ndx(prev);
 130                                 if (flist)
 131                                         flist->files[prev - flist->ndx_start]->flags &= ~FLAG_HLINK_LAST;
 132                                 else {
 133                                         /* We skipped all prior files in this
 134                                          * group, so mark this as a "first". */
 135                                         file->flags |= FLAG_HLINK_FIRST;
 136                                         prev = -1;
 137                                 }
 138                         } else
 139                                 prev = -1;
 140                 } else {
 141                         file->flags |= FLAG_HLINK_FIRST;
 142                         prev = -1;
 143                 }
 144                 for ( ; from < ndx_count-1; file = file_next, gnum = gnum_next, from++) { /*SHARED ITERATOR*/
 145                         file_next = hlink_flist->sorted[ndx_list[from+1]];
 146                         gnum_next = F_HL_GNUM(file_next);
 147                         if (gnum != gnum_next)
 148                                 break;
 149                         F_HL_PREV(file) = prev;
 150                         /* The linked list uses over-the-wire ndx values. */
 151                         if (unsort_ndx)
 152                                 prev = F_NDX(file);
 153                         else
 154                                 prev = ndx_list[from] + hlink_flist->ndx_start;
 155                 }
 156                 if (prev < 0 && !inc_recurse) {
 157                         /* Disable hard-link bit and set DONE so that
 158                          * HLINK_BUMP()-dependent values are unaffected. */
 159                         file->flags &= ~(FLAG_HLINKED | FLAG_HLINK_FIRST);
 160                         file->flags |= FLAG_HLINK_DONE;
 161                         continue;
 162                 }
 163
 164                 file->flags |= FLAG_HLINK_LAST;
 165                 F_HL_PREV(file) = prev;
 166                 if (inc_recurse && CVAL(node->data, 0) == 0) {
 167                         if (unsort_ndx)
 168                                 prev = F_NDX(file);
 169                         else
 170                                 prev = ndx_list[from] + hlink_flist->ndx_start;
 171                         SIVAL(node->data, 1, prev);
 172                 }
 173         }
 174 }
 175
 176 /* Analyze the hard-links in the file-list by creating a list of all the
 177  * items that have hlink data, sorting them, and matching up identical
 178  * values into clusters.  These will be a single linked list from last
 179  * to first when we're done. */
 180 void match_hard_links(struct file_list *flist)
 181 {
 182         if (!list_only) {
 183                 int i, ndx_count = 0;
 184                 int32 *ndx_list;
 185
 186                 if (!(ndx_list = new_array(int32, flist->used)))
 187                         out_of_memory("match_hard_links");
 188
 189                 for (i = 0; i < flist->used; i++) {
 190                         if (F_IS_HLINKED(flist->sorted[i]))
 191                                 ndx_list[ndx_count++] = i;
 192                 }
 193
 194                 hlink_flist = flist;
 195
 196                 if (ndx_count)
 197                         match_gnums(ndx_list, ndx_count);
 198
 199                 free(ndx_list);
 200         }
 201         if (protocol_version < 30)
 202                 idev_destroy();
 203 }
 204
 205 static int maybe_hard_link(struct file_struct *file, int ndx,
 206                            const char *fname, int statret, stat_x *sxp,
 207                            const char *oldname, STRUCT_STAT *old_stp,
 208                            const char *realname, int itemizing, enum logcode code)
 209 {
 210         if (statret == 0) {
 211                 if (sxp->st.st_dev == old_stp->st_dev
 212                  && sxp->st.st_ino == old_stp->st_ino) {
 213                         if (itemizing) {
 214                                 itemize(fname, file, ndx, statret, sxp,
 215                                         ITEM_LOCAL_CHANGE | ITEM_XNAME_FOLLOWS,
 216                                         0, "");
 217                         }
 218                         if (verbose > 1 && maybe_ATTRS_REPORT)
 219                                 rprintf(FCLIENT, "%s is uptodate\n", fname);
 220                         file->flags |= FLAG_HLINK_DONE;
 221                         return 0;
 222                 }
 223                 if (make_backups > 0) {
 224                         if (!make_backup(fname))
 225                                 return -1;
 226                 } else if (robust_unlink(fname)) {
 227                         rsyserr(FERROR_XFER, errno, "unlink %s failed",
 228                                 full_fname(fname));
 229                         return -1;
 230                 }
 231         }
 232
 233         if (hard_link_one(file, fname, oldname, 0)) {
 234                 if (itemizing) {
 235                         itemize(fname, file, ndx, statret, sxp,
 236                                 ITEM_LOCAL_CHANGE | ITEM_XNAME_FOLLOWS, 0,
 237                                 realname);
 238                 }
 239                 if (code != FNONE && verbose)
 240                         rprintf(code, "%s => %s\n", fname, realname);
 241                 return 0;
 242         }
 243         return -1;
 244 }
 245
 246 /* Figure out if a prior entry is still there or if we just have a
 247  * cached name for it. */
 248 static char *check_prior(struct file_struct *file, int gnum,
 249                          int *prev_ndx_p, struct file_list **flist_p)
 250 {
 251         struct file_struct *fp;
 252         struct ht_int32_node *node;
 253         int prev_ndx = F_HL_PREV(file);
 254
 255         while (1) {
 256                 struct file_list *flist;
 257                 if (prev_ndx < 0
 258                  || (flist = flist_for_ndx(prev_ndx)) == NULL)
 259                         break;
 260                 fp = flist->files[prev_ndx - flist->ndx_start];
 261                 if (!(fp->flags & FLAG_SKIP_HLINK)) {
 262                         *prev_ndx_p = prev_ndx;
 263                         *flist_p = flist;
 264                         return NULL;
 265                 }
 266                 F_HL_PREV(file) = prev_ndx = F_HL_PREV(fp);
 267         }
 268
 269         if (inc_recurse
 270          && (node = hashtable_find(prior_hlinks, gnum, 0)) != NULL) {
 271                 assert(node->data != NULL);
 272                 if (CVAL(node->data, 0) != 0) {
 273                         *prev_ndx_p = -1;
 274                         *flist_p = NULL;
 275                         return node->data;
 276                 }
 277                 /* The prior file must have been skipped. */
 278                 F_HL_PREV(file) = -1;
 279         }
 280
 281         *prev_ndx_p = -1;
 282         *flist_p = NULL;
 283         return NULL;
 284 }
 285
 286 /* Only called if FLAG_HLINKED is set and FLAG_HLINK_FIRST is not.  Returns:
 287  * 0 = process the file, 1 = skip the file, -1 = error occurred. */
 288 int hard_link_check(struct file_struct *file, int ndx, const char *fname,
 289                     int statret, stat_x *sxp, int itemizing,
 290                     enum logcode code)
 291 {
 292         STRUCT_STAT prev_st;
 293         char namebuf[MAXPATHLEN], altbuf[MAXPATHLEN];
 294         char *realname, *prev_name;
 295         struct file_list *flist;
 296         int gnum = inc_recurse ? F_HL_GNUM(file) : -1;
 297         int prev_ndx;
 298
 299         prev_name = realname = check_prior(file, gnum, &prev_ndx, &flist);
 300
 301         if (!prev_name) {
 302                 struct file_struct *prev_file;
 303
 304                 if (!flist) {
 305                         /* The previous file was skipped, so this one is
 306                          * treated as if it were the first in its group. */
 307                         return 0;
 308                 }
 309
 310                 prev_file = flist->files[prev_ndx - flist->ndx_start];
 311
 312                 /* Is the previous link not complete yet? */
 313                 if (!(prev_file->flags & FLAG_HLINK_DONE)) {
 314                         /* Is the previous link being transferred? */
 315                         if (prev_file->flags & FLAG_FILE_SENT) {
 316                                 /* Add ourselves to the list of files that will
 317                                  * be updated when the transfer completes, and
 318                                  * mark ourself as waiting for the transfer. */
 319                                 F_HL_PREV(file) = F_HL_PREV(prev_file);
 320                                 F_HL_PREV(prev_file) = ndx;
 321                                 file->flags |= FLAG_FILE_SENT;
 322                                 cur_flist->in_progress++;
 323                                 return 1;
 324                         }
 325                         return 0;
 326                 }
 327
 328                 /* There is a finished file to link with! */
 329                 if (!(prev_file->flags & FLAG_HLINK_FIRST)) {
 330                         /* The previous previous is FIRST when prev is not. */
 331                         prev_name = realname = check_prior(prev_file, gnum, &prev_ndx, &flist);
 332                         assert(prev_name != NULL || flist != NULL);
 333                         /* Update our previous pointer to point to the FIRST. */
 334                         F_HL_PREV(file) = prev_ndx;
 335                 }
 336
 337                 if (!prev_name) {
 338                         int alt_dest;
 339
 340                         prev_file = flist->files[prev_ndx - flist->ndx_start];
 341                         /* F_HL_PREV() is alt_dest value when DONE && FIRST. */
 342                         alt_dest = F_HL_PREV(prev_file);
 343
 344                         if (alt_dest >= 0 && dry_run) {
 345                                 pathjoin(namebuf, MAXPATHLEN, basis_dir[alt_dest],
 346                                          f_name(prev_file, NULL));
 347                                 prev_name = namebuf;
 348                                 realname = f_name(prev_file, altbuf);
 349                         } else {
 350                                 prev_name = f_name(prev_file, namebuf);
 351                                 realname = prev_name;
 352                         }
 353                 }
 354         }
 355
 356         if (link_stat(prev_name, &prev_st, 0) < 0) {
 357                 rsyserr(FERROR_XFER, errno, "stat %s failed",
 358                         full_fname(prev_name));
 359                 return -1;
 360         }
 361
 362         if (statret < 0 && basis_dir[0] != NULL) {
 363                 /* If we match an alt-dest item, we don't output this as a change. */
 364                 char cmpbuf[MAXPATHLEN];
 365                 stat_x alt_sx;
 366                 int j = 0;
 367 #ifdef SUPPORT_ACLS
 368                 alt_sx.acc_acl = alt_sx.def_acl = NULL;
 369 #endif
 370                 do {
 371                         pathjoin(cmpbuf, MAXPATHLEN, basis_dir[j], fname);
 372                         if (link_stat(cmpbuf, &alt_sx.st, 0) < 0)
 373                                 continue;
 374                         if (link_dest) {
 375                                 if (prev_st.st_dev != alt_sx.st.st_dev
 376                                  || prev_st.st_ino != alt_sx.st.st_ino)
 377                                         continue;
 378                                 statret = 1;
 379                                 if (stdout_format_has_i == 0
 380                                  || (verbose < 2 && stdout_format_has_i < 2)) {
 381                                         itemizing = 0;
 382                                         code = FNONE;
 383                                         if (verbose > 1 && maybe_ATTRS_REPORT)
 384                                                 rprintf(FCLIENT, "%s is uptodate\n", fname);
 385                                 }
 386                                 break;
 387                         }
 388                         if (!unchanged_file(cmpbuf, file, &alt_sx.st))
 389                                 continue;
 390                         statret = 1;
 391                         if (unchanged_attrs(cmpbuf, file, &alt_sx))
 392                                 break;
 393                 } while (basis_dir[++j] != NULL);
 394                 if (statret == 1) {
 395                         sxp->st = alt_sx.st;
 396 #ifdef SUPPORT_ACLS
 397                         if (preserve_acls && !S_ISLNK(file->mode)) {
 398                                 if (!ACL_READY(*sxp))
 399                                         get_acl(cmpbuf, sxp);
 400                                 else {
 401                                         sxp->acc_acl = alt_sx.acc_acl;
 402                                         sxp->def_acl = alt_sx.def_acl;
 403                                 }
 404                         }
 405 #endif
 406                 }
 407 #ifdef SUPPORT_ACLS
 408                 else if (preserve_acls)
 409                         free_acl(&alt_sx);
 410 #endif
 411         }
 412
 413         if (maybe_hard_link(file, ndx, fname, statret, sxp, prev_name, &prev_st,
 414                             realname, itemizing, code) < 0)
 415                 return -1;
 416
 417         if (remove_source_files == 1 && do_xfers)
 418                 send_msg_int(MSG_SUCCESS, ndx);
 419
 420         return 1;
 421 }
 422
 423 int hard_link_one(struct file_struct *file, const char *fname,
 424                   const char *oldname, int terse)
 425 {
 426         if (do_link(oldname, fname) < 0) {
 427                 enum logcode code;
 428                 if (terse) {
 429                         if (!verbose)
 430                                 return 0;
 431                         code = FINFO;
 432                 } else
 433                         code = FERROR_XFER;
 434                 rsyserr(code, errno, "link %s => %s failed",
 435                         full_fname(fname), oldname);
 436                 return 0;
 437         }
 438
 439         file->flags |= FLAG_HLINK_DONE;
 440
 441         return 1;
 442 }
 443
 444 void finish_hard_link(struct file_struct *file, const char *fname, int fin_ndx,
 445                       STRUCT_STAT *stp, int itemizing, enum logcode code,
 446                       int alt_dest)
 447 {
 448         stat_x prev_sx;
 449         STRUCT_STAT st;
 450         char prev_name[MAXPATHLEN], alt_name[MAXPATHLEN];
 451         const char *our_name;
 452         struct file_list *flist;
 453         int prev_statret, ndx, prev_ndx = F_HL_PREV(file);
 454
 455         if (stp == NULL && prev_ndx >= 0) {
 456                 if (link_stat(fname, &st, 0) < 0) {
 457                         rsyserr(FERROR_XFER, errno, "stat %s failed",
 458                                 full_fname(fname));
 459                         return;
 460                 }
 461                 stp = &st;
 462         }
 463
 464         /* FIRST combined with DONE means we were the first to get done. */
 465         file->flags |= FLAG_HLINK_FIRST | FLAG_HLINK_DONE;
 466         F_HL_PREV(file) = alt_dest;
 467         if (alt_dest >= 0 && dry_run) {
 468                 pathjoin(alt_name, MAXPATHLEN, basis_dir[alt_dest],
 469                          f_name(file, NULL));
 470                 our_name = alt_name;
 471         } else
 472                 our_name = fname;
 473
 474 #ifdef SUPPORT_ACLS
 475         prev_sx.acc_acl = prev_sx.def_acl = NULL;
 476 #endif
 477
 478         while ((ndx = prev_ndx) >= 0) {
 479                 int val;
 480                 flist = flist_for_ndx(ndx);
 481                 if (flist == NULL) {
 482                         int start1 = first_flist ? first_flist->ndx_start : 0;
 483                         int start2 = first_flist ? first_flist->prev->ndx_start : 0;
 484                         int used = first_flist ? first_flist->prev->used : 0;
 485                         rprintf(FERROR,
 486                                 "File index not found: %d (%d - %d)\n",
 487                                 ndx, start1 - 1, start2 + used - 1);
 488                         exit_cleanup(RERR_PROTOCOL);
 489                 }
 490                 file = flist->files[ndx - flist->ndx_start];
 491                 file->flags = (file->flags & ~FLAG_HLINK_FIRST) | FLAG_HLINK_DONE;
 492                 prev_ndx = F_HL_PREV(file);
 493                 F_HL_PREV(file) = fin_ndx;
 494                 prev_statret = link_stat(f_name(file, prev_name), &prev_sx.st, 0);
 495                 val = maybe_hard_link(file, ndx, prev_name, prev_statret, &prev_sx,
 496                                       our_name, stp, fname, itemizing, code);
 497                 flist->in_progress--;
 498 #ifdef SUPPORT_ACLS
 499                 if (preserve_acls)
 500                         free_acl(&prev_sx);
 501 #endif
 502                 if (val < 0)
 503                         continue;
 504                 if (remove_source_files == 1 && do_xfers)
 505                         send_msg_int(MSG_SUCCESS, ndx);
 506         }
 507
 508         if (inc_recurse) {
 509                 int gnum = F_HL_GNUM(file);
 510                 struct ht_int32_node *node = hashtable_find(prior_hlinks, gnum, 0);
 511                 assert(node != NULL && node->data != NULL);
 512                 assert(CVAL(node->data, 0) == 0);
 513                 free(node->data);
 514                 if (!(node->data = strdup(our_name)))
 515                         out_of_memory("finish_hard_link");
 516         }
 517 }
 518
 519 int skip_hard_link(struct file_struct *file, struct file_list **flist_p)
 520 {
 521         struct file_list *flist;
 522         int prev_ndx;
 523
 524         file->flags |= FLAG_SKIP_HLINK;
 525         if (!(file->flags & FLAG_HLINK_LAST))
 526                 return -1;
 527
 528         check_prior(file, F_HL_GNUM(file), &prev_ndx, &flist);
 529         if (prev_ndx >= 0) {
 530                 file = flist->files[prev_ndx - flist->ndx_start];
 531                 if (file->flags & (FLAG_HLINK_DONE|FLAG_FILE_SENT))
 532                         return -1;
 533                 file->flags |= FLAG_HLINK_LAST;
 534                 *flist_p = flist;
 535         }
 536
 537         return prev_ndx;
 538 }
 539 #endif