devspec.en_US/project/recutils/lib/stat.c

   1 /* Work around platform bugs in stat.
   2    Copyright (C) 2009-2019 Free Software Foundation, Inc.
   3
   4    This program is free software: you can redistribute it and/or modify
   5    it under the terms of the GNU General Public License as published by
   6    the Free Software Foundation; either version 3 of the License, or
   7    (at your option) any later version.
   8
   9    This program is distributed in the hope that it will be useful,
  10    but WITHOUT ANY WARRANTY; without even the implied warranty of
  11    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  12    GNU General Public License for more details.
  13
  14    You should have received a copy of the GNU General Public License
  15    along with this program.  If not, see <https://www.gnu.org/licenses/>.  */
  16
  17 /* Written by Eric Blake and Bruno Haible.  */
  18
  19 /* If the user's config.h happens to include <sys/stat.h>, let it include only
  20    the system's <sys/stat.h> here, so that orig_stat doesn't recurse to
  21    rpl_stat.  */
  22 #define __need_system_sys_stat_h
  23 #include <config.h>
  24
  25 /* Get the original definition of stat.  It might be defined as a macro.  */
  26 #include <sys/types.h>
  27 #include <sys/stat.h>
  28 #undef __need_system_sys_stat_h
  29
  30 #if defined _WIN32 && ! defined __CYGWIN__
  31 # define WINDOWS_NATIVE
  32 #endif
  33
  34 #if !defined WINDOWS_NATIVE
  35
  36 static int
  37 orig_stat (const char *filename, struct stat *buf)
  38 {
  39   return stat (filename, buf);
  40 }
  41
  42 #endif
  43
  44 /* Specification.  */
  45 /* Write "sys/stat.h" here, not <sys/stat.h>, otherwise OSF/1 5.1 DTK cc
  46    eliminates this include because of the preliminary #include <sys/stat.h>
  47    above.  */
  48 #include "sys/stat.h"
  49
  50 #include "stat-time.h"
  51
  52 #include <errno.h>
  53 #include <limits.h>
  54 #include <stdbool.h>
  55 #include <string.h>
  56 #include "filename.h"
  57 #include "malloca.h"
  58 #include "verify.h"
  59
  60 #ifdef WINDOWS_NATIVE
  61 # define WIN32_LEAN_AND_MEAN
  62 # include <windows.h>
  63 # include "stat-w32.h"
  64 #endif
  65
  66 #ifdef WINDOWS_NATIVE
  67 /* Return TRUE if the given file name denotes an UNC root.  */
  68 static BOOL
  69 is_unc_root (const char *rname)
  70 {
  71   /* Test whether it has the syntax '\\server\share'.  */
  72   if (ISSLASH (rname[0]) && ISSLASH (rname[1]))
  73     {
  74       /* It starts with two slashes.  Find the next slash.  */
  75       const char *p = rname + 2;
  76       const char *q = p;
  77       while (*q != '\0' && !ISSLASH (*q))
  78         q++;
  79       if (q > p && *q != '\0')
  80         {
  81           /* Found the next slash at q.  */
  82           q++;
  83           const char *r = q;
  84           while (*r != '\0' && !ISSLASH (*r))
  85             r++;
  86           if (r > q && *r == '\0')
  87             return TRUE;
  88         }
  89     }
  90   return FALSE;
  91 }
  92 #endif
  93
  94 /* Store information about NAME into ST.  Work around bugs with
  95    trailing slashes.  Mingw has other bugs (such as st_ino always
  96    being 0 on success) which this wrapper does not work around.  But
  97    at least this implementation provides the ability to emulate fchdir
  98    correctly.  */
  99
 100 int
 101 rpl_stat (char const *name, struct stat *buf)
 102 {
 103 #ifdef WINDOWS_NATIVE
 104   /* Fill the fields ourselves, because the original stat function returns
 105      values for st_atime, st_mtime, st_ctime that depend on the current time
 106      zone.  See
 107      <https://lists.gnu.org/r/bug-gnulib/2017-04/msg00134.html>  */
 108   /* XXX Should we convert to wchar_t* and prepend '\\?\', in order to work
 109      around length limitations
 110      <https://msdn.microsoft.com/en-us/library/aa365247.aspx> ?  */
 111
 112   /* POSIX <http://pubs.opengroup.org/onlinepubs/9699919799/basedefs/V1_chap04.html#tag_04_13>
 113      specifies: "More than two leading <slash> characters shall be treated as
 114      a single <slash> character."  */
 115   if (ISSLASH (name[0]) && ISSLASH (name[1]) && ISSLASH (name[2]))
 116     {
 117       name += 2;
 118       while (ISSLASH (name[1]))
 119         name++;
 120     }
 121
 122   size_t len = strlen (name);
 123   size_t drive_prefix_len = (HAS_DEVICE (name) ? 2 : 0);
 124
 125   /* Remove trailing slashes (except the very first one, at position
 126      drive_prefix_len), but remember their presence.  */
 127   size_t rlen;
 128   bool check_dir = false;
 129
 130   rlen = len;
 131   while (rlen > drive_prefix_len && ISSLASH (name[rlen-1]))
 132     {
 133       check_dir = true;
 134       if (rlen == drive_prefix_len + 1)
 135         break;
 136       rlen--;
 137     }
 138
 139   /* Handle '' and 'C:'.  */
 140   if (!check_dir && rlen == drive_prefix_len)
 141     {
 142       errno = ENOENT;
 143       return -1;
 144     }
 145
 146   /* Handle '\\'.  */
 147   if (rlen == 1 && ISSLASH (name[0]) && len >= 2)
 148     {
 149       errno = ENOENT;
 150       return -1;
 151     }
 152
 153   const char *rname;
 154   char *malloca_rname;
 155   if (rlen == len)
 156     {
 157       rname = name;
 158       malloca_rname = NULL;
 159     }
 160   else
 161     {
 162       malloca_rname = malloca (rlen + 1);
 163       if (malloca_rname == NULL)
 164         {
 165           errno = ENOMEM;
 166           return -1;
 167         }
 168       memcpy (malloca_rname, name, rlen);
 169       malloca_rname[rlen] = '\0';
 170       rname = malloca_rname;
 171     }
 172
 173   /* There are two ways to get at the requested information:
 174        - by scanning the parent directory and examining the relevant
 175          directory entry,
 176        - by opening the file directly.
 177      The first approach fails for root directories (e.g. 'C:\') and
 178      UNC root directories (e.g. '\\server\share').
 179      The second approach fails for some system files (e.g. 'C:\pagefile.sys'
 180      and 'C:\hiberfil.sys'): ERROR_SHARING_VIOLATION.
 181      The second approach gives more information (in particular, correct
 182      st_dev, st_ino, st_nlink fields).
 183      So we use the second approach and, as a fallback except for root and
 184      UNC root directories, also the first approach.  */
 185   {
 186     int ret;
 187
 188     {
 189       /* Approach based on the file.  */
 190
 191       /* Open a handle to the file.
 192          CreateFile
 193          <https://msdn.microsoft.com/en-us/library/aa363858.aspx>
 194          <https://msdn.microsoft.com/en-us/library/aa363874.aspx>  */
 195       HANDLE h =
 196         CreateFile (rname,
 197                     FILE_READ_ATTRIBUTES,
 198                     FILE_SHARE_READ | FILE_SHARE_WRITE | FILE_SHARE_DELETE,
 199                     NULL,
 200                     OPEN_EXISTING,
 201                     /* FILE_FLAG_POSIX_SEMANTICS (treat file names that differ only
 202                        in case as different) makes sense only when applied to *all*
 203                        filesystem operations.  */
 204                     FILE_FLAG_BACKUP_SEMANTICS /* | FILE_FLAG_POSIX_SEMANTICS */,
 205                     NULL);
 206       if (h != INVALID_HANDLE_VALUE)
 207         {
 208           ret = _gl_fstat_by_handle (h, rname, buf);
 209           CloseHandle (h);
 210           goto done;
 211         }
 212     }
 213
 214     /* Test for root and UNC root directories.  */
 215     if ((rlen == drive_prefix_len + 1 && ISSLASH (rname[drive_prefix_len]))
 216         || is_unc_root (rname))
 217       goto failed;
 218
 219     /* Fallback.  */
 220     {
 221       /* Approach based on the directory entry.  */
 222
 223       if (strchr (rname, '?') != NULL || strchr (rname, '*') != NULL)
 224         {
 225           /* Other Windows API functions would fail with error
 226              ERROR_INVALID_NAME.  */
 227           if (malloca_rname != NULL)
 228             freea (malloca_rname);
 229           errno = ENOENT;
 230           return -1;
 231         }
 232
 233       /* Get the details about the directory entry.  This can be done through
 234          FindFirstFile
 235          <https://msdn.microsoft.com/en-us/library/aa364418.aspx>
 236          <https://msdn.microsoft.com/en-us/library/aa365740.aspx>
 237          or through
 238          FindFirstFileEx with argument FindExInfoBasic
 239          <https://msdn.microsoft.com/en-us/library/aa364419.aspx>
 240          <https://msdn.microsoft.com/en-us/library/aa364415.aspx>
 241          <https://msdn.microsoft.com/en-us/library/aa365740.aspx>  */
 242       WIN32_FIND_DATA info;
 243       HANDLE h = FindFirstFile (rname, &info);
 244       if (h == INVALID_HANDLE_VALUE)
 245         goto failed;
 246
 247       /* Test for error conditions before starting to fill *buf.  */
 248       if (sizeof (buf->st_size) <= 4 && info.nFileSizeHigh > 0)
 249         {
 250           FindClose (h);
 251           if (malloca_rname != NULL)
 252             freea (malloca_rname);
 253           errno = EOVERFLOW;
 254           return -1;
 255         }
 256
 257 # if _GL_WINDOWS_STAT_INODES
 258       buf->st_dev = 0;
 259 #  if _GL_WINDOWS_STAT_INODES == 2
 260       buf->st_ino._gl_ino[0] = buf->st_ino._gl_ino[1] = 0;
 261 #  else /* _GL_WINDOWS_STAT_INODES == 1 */
 262       buf->st_ino = 0;
 263 #  endif
 264 # else
 265       /* st_ino is not wide enough for identifying a file on a device.
 266          Without st_ino, st_dev is pointless.  */
 267       buf->st_dev = 0;
 268       buf->st_ino = 0;
 269 # endif
 270
 271       /* st_mode.  */
 272       unsigned int mode =
 273         /* XXX How to handle FILE_ATTRIBUTE_REPARSE_POINT ?  */
 274         ((info.dwFileAttributes & FILE_ATTRIBUTE_DIRECTORY) ? _S_IFDIR | S_IEXEC_UGO : _S_IFREG)
 275         | S_IREAD_UGO
 276         | ((info.dwFileAttributes & FILE_ATTRIBUTE_READONLY) ? 0 : S_IWRITE_UGO);
 277       if (!(info.dwFileAttributes & FILE_ATTRIBUTE_DIRECTORY))
 278         {
 279           /* Determine whether the file is executable by looking at the file
 280              name suffix.  */
 281           if (info.nFileSizeHigh > 0 || info.nFileSizeLow > 0)
 282             {
 283               const char *last_dot = NULL;
 284               const char *p;
 285               for (p = info.cFileName; *p != '\0'; p++)
 286                 if (*p == '.')
 287                   last_dot = p;
 288               if (last_dot != NULL)
 289                 {
 290                   const char *suffix = last_dot + 1;
 291                   if (_stricmp (suffix, "exe") == 0
 292                       || _stricmp (suffix, "bat") == 0
 293                       || _stricmp (suffix, "cmd") == 0
 294                       || _stricmp (suffix, "com") == 0)
 295                     mode |= S_IEXEC_UGO;
 296                 }
 297             }
 298         }
 299       buf->st_mode = mode;
 300
 301       /* st_nlink.  Ignore hard links here.  */
 302       buf->st_nlink = 1;
 303
 304       /* There's no easy way to map the Windows SID concept to an integer.  */
 305       buf->st_uid = 0;
 306       buf->st_gid = 0;
 307
 308       /* st_rdev is irrelevant for normal files and directories.  */
 309       buf->st_rdev = 0;
 310
 311       /* st_size.  */
 312       if (sizeof (buf->st_size) <= 4)
 313         /* Range check already done above.  */
 314         buf->st_size = info.nFileSizeLow;
 315       else
 316         buf->st_size = ((long long) info.nFileSizeHigh << 32) | (long long) info.nFileSizeLow;
 317
 318       /* st_atime, st_mtime, st_ctime.  */
 319 # if _GL_WINDOWS_STAT_TIMESPEC
 320       buf->st_atim = _gl_convert_FILETIME_to_timespec (&info.ftLastAccessTime);
 321       buf->st_mtim = _gl_convert_FILETIME_to_timespec (&info.ftLastWriteTime);
 322       buf->st_ctim = _gl_convert_FILETIME_to_timespec (&info.ftCreationTime);
 323 # else
 324       buf->st_atime = _gl_convert_FILETIME_to_POSIX (&info.ftLastAccessTime);
 325       buf->st_mtime = _gl_convert_FILETIME_to_POSIX (&info.ftLastWriteTime);
 326       buf->st_ctime = _gl_convert_FILETIME_to_POSIX (&info.ftCreationTime);
 327 # endif
 328
 329       FindClose (h);
 330
 331       ret = 0;
 332     }
 333
 334    done:
 335     if (ret >= 0 && check_dir && !S_ISDIR (buf->st_mode))
 336       {
 337         errno = ENOTDIR;
 338         ret = -1;
 339       }
 340     if (malloca_rname != NULL)
 341       {
 342         int saved_errno = errno;
 343         freea (malloca_rname);
 344         errno = saved_errno;
 345       }
 346     return ret;
 347   }
 348
 349  failed:
 350   {
 351     DWORD error = GetLastError ();
 352     #if 0
 353     fprintf (stderr, "rpl_stat error 0x%x\n", (unsigned int) error);
 354     #endif
 355
 356     if (malloca_rname != NULL)
 357       freea (malloca_rname);
 358
 359     switch (error)
 360       {
 361       /* Some of these errors probably cannot happen with the specific flags
 362          that we pass to CreateFile.  But who knows...  */
 363       case ERROR_FILE_NOT_FOUND: /* The last component of rname does not exist.  */
 364       case ERROR_PATH_NOT_FOUND: /* Some directory component in rname does not exist.  */
 365       case ERROR_BAD_PATHNAME:   /* rname is such as '\\server'.  */
 366       case ERROR_BAD_NET_NAME:   /* rname is such as '\\server\nonexistentshare'.  */
 367       case ERROR_INVALID_NAME:   /* rname contains wildcards, misplaced colon, etc.  */
 368       case ERROR_DIRECTORY:
 369         errno = ENOENT;
 370         break;
 371
 372       case ERROR_ACCESS_DENIED:  /* rname is such as 'C:\System Volume Information\foo'.  */
 373       case ERROR_SHARING_VIOLATION: /* rname is such as 'C:\pagefile.sys' (second approach only).  */
 374                                     /* XXX map to EACCESS or EPERM? */
 375         errno = EACCES;
 376         break;
 377
 378       case ERROR_OUTOFMEMORY:
 379         errno = ENOMEM;
 380         break;
 381
 382       case ERROR_WRITE_PROTECT:
 383         errno = EROFS;
 384         break;
 385
 386       case ERROR_WRITE_FAULT:
 387       case ERROR_READ_FAULT:
 388       case ERROR_GEN_FAILURE:
 389         errno = EIO;
 390         break;
 391
 392       case ERROR_BUFFER_OVERFLOW:
 393       case ERROR_FILENAME_EXCED_RANGE:
 394         errno = ENAMETOOLONG;
 395         break;
 396
 397       case ERROR_DELETE_PENDING: /* XXX map to EACCESS or EPERM? */
 398         errno = EPERM;
 399         break;
 400
 401       default:
 402         errno = EINVAL;
 403         break;
 404       }
 405
 406     return -1;
 407   }
 408 #else
 409   int result = orig_stat (name, buf);
 410   if (result == 0)
 411     {
 412 # if REPLACE_FUNC_STAT_FILE
 413       /* Solaris 9 mistakenly succeeds when given a non-directory with a
 414          trailing slash.  */
 415       if (!S_ISDIR (buf->st_mode))
 416         {
 417           size_t len = strlen (name);
 418           if (ISSLASH (name[len - 1]))
 419             {
 420               errno = ENOTDIR;
 421               return -1;
 422             }
 423         }
 424 # endif /* REPLACE_FUNC_STAT_FILE */
 425       result = stat_time_normalize (result, buf);
 426     }
 427   return result;
 428 #endif
 429 }