gl/stat.c

   1 /* Work around platform bugs in stat.
   2    Copyright (C) 2009-2024 Free Software Foundation, Inc.
   3
   4    This file is free software: you can redistribute it and/or modify
   5    it under the terms of the GNU Lesser General Public License as
   6    published by the Free Software Foundation; either version 2.1 of the
   7    License, or (at your option) any later version.
   8
   9    This file 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 Lesser General Public License for more details.
  13
  14    You should have received a copy of the GNU Lesser 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 #ifdef __osf__
  46 /* Write "sys/stat.h" here, not <sys/stat.h>, otherwise OSF/1 5.1 DTK cc
  47    eliminates this include because of the preliminary #include <sys/stat.h>
  48    above.  */
  49 # include "sys/stat.h"
  50 #else
  51 # include <sys/stat.h>
  52 #endif
  53
  54 #include "stat-time.h"
  55
  56 #include <errno.h>
  57 #include <limits.h>
  58 #include <string.h>
  59 #include "filename.h"
  60 #include "malloca.h"
  61
  62 #ifdef WINDOWS_NATIVE
  63 # define WIN32_LEAN_AND_MEAN
  64 # include <windows.h>
  65 # include "stat-w32.h"
  66 /* Don't assume that UNICODE is not defined.  */
  67 # undef WIN32_FIND_DATA
  68 # define WIN32_FIND_DATA WIN32_FIND_DATAA
  69 # undef CreateFile
  70 # define CreateFile CreateFileA
  71 # undef FindFirstFile
  72 # define FindFirstFile FindFirstFileA
  73 #endif
  74
  75 #ifdef WINDOWS_NATIVE
  76 /* Return TRUE if the given file name denotes an UNC root.  */
  77 static BOOL
  78 is_unc_root (const char *rname)
  79 {
  80   /* Test whether it has the syntax '\\server\share'.  */
  81   if (ISSLASH (rname[0]) && ISSLASH (rname[1]))
  82     {
  83       /* It starts with two slashes.  Find the next slash.  */
  84       const char *p = rname + 2;
  85       const char *q = p;
  86       while (*q != '\0' && !ISSLASH (*q))
  87         q++;
  88       if (q > p && *q != '\0')
  89         {
  90           /* Found the next slash at q.  */
  91           q++;
  92           const char *r = q;
  93           while (*r != '\0' && !ISSLASH (*r))
  94             r++;
  95           if (r > q && *r == '\0')
  96             return TRUE;
  97         }
  98     }
  99   return FALSE;
 100 }
 101 #endif
 102
 103 /* Store information about NAME into ST.  Work around bugs with
 104    trailing slashes.  Mingw has other bugs (such as st_ino always
 105    being 0 on success) which this wrapper does not work around.  But
 106    at least this implementation provides the ability to emulate fchdir
 107    correctly.  */
 108
 109 int
 110 rpl_stat (char const *name, struct stat *buf)
 111 {
 112 #ifdef WINDOWS_NATIVE
 113   /* Fill the fields ourselves, because the original stat function returns
 114      values for st_atime, st_mtime, st_ctime that depend on the current time
 115      zone.  See
 116      <https://lists.gnu.org/r/bug-gnulib/2017-04/msg00134.html>  */
 117   /* XXX Should we convert to wchar_t* and prepend '\\?\', in order to work
 118      around length limitations
 119      <https://docs.microsoft.com/en-us/windows/desktop/FileIO/naming-a-file> ?  */
 120
 121   /* POSIX <https://pubs.opengroup.org/onlinepubs/9699919799/basedefs/V1_chap04.html#tag_04_13>
 122      specifies: "More than two leading <slash> characters shall be treated as
 123      a single <slash> character."  */
 124   if (ISSLASH (name[0]) && ISSLASH (name[1]) && ISSLASH (name[2]))
 125     {
 126       name += 2;
 127       while (ISSLASH (name[1]))
 128         name++;
 129     }
 130
 131   size_t len = strlen (name);
 132   size_t drive_prefix_len = (HAS_DEVICE (name) ? 2 : 0);
 133
 134   /* Remove trailing slashes (except the very first one, at position
 135      drive_prefix_len), but remember their presence.  */
 136   size_t rlen;
 137   bool check_dir = false;
 138
 139   rlen = len;
 140   while (rlen > drive_prefix_len && ISSLASH (name[rlen-1]))
 141     {
 142       check_dir = true;
 143       if (rlen == drive_prefix_len + 1)
 144         break;
 145       rlen--;
 146     }
 147
 148   /* Handle '' and 'C:'.  */
 149   if (!check_dir && rlen == drive_prefix_len)
 150     {
 151       errno = ENOENT;
 152       return -1;
 153     }
 154
 155   /* Handle '\\'.  */
 156   if (rlen == 1 && ISSLASH (name[0]) && len >= 2)
 157     {
 158       errno = ENOENT;
 159       return -1;
 160     }
 161
 162   const char *rname;
 163   char *malloca_rname;
 164   if (rlen == len)
 165     {
 166       rname = name;
 167       malloca_rname = NULL;
 168     }
 169   else
 170     {
 171       malloca_rname = malloca (rlen + 1);
 172       if (malloca_rname == NULL)
 173         {
 174           errno = ENOMEM;
 175           return -1;
 176         }
 177       memcpy (malloca_rname, name, rlen);
 178       malloca_rname[rlen] = '\0';
 179       rname = malloca_rname;
 180     }
 181
 182   /* There are two ways to get at the requested information:
 183        - by scanning the parent directory and examining the relevant
 184          directory entry,
 185        - by opening the file directly.
 186      The first approach fails for root directories (e.g. 'C:\') and
 187      UNC root directories (e.g. '\\server\share').
 188      The second approach fails for some system files (e.g. 'C:\pagefile.sys'
 189      and 'C:\hiberfil.sys'): ERROR_SHARING_VIOLATION.
 190      The second approach gives more information (in particular, correct
 191      st_dev, st_ino, st_nlink fields).
 192      So we use the second approach and, as a fallback except for root and
 193      UNC root directories, also the first approach.  */
 194   {
 195     int ret;
 196
 197     {
 198       /* Approach based on the file.  */
 199
 200       /* Open a handle to the file.
 201          CreateFile
 202          <https://docs.microsoft.com/en-us/windows/desktop/api/fileapi/nf-fileapi-createfilea>
 203          <https://docs.microsoft.com/en-us/windows/desktop/FileIO/creating-and-opening-files>  */
 204       HANDLE h =
 205         CreateFile (rname,
 206                     FILE_READ_ATTRIBUTES,
 207                     FILE_SHARE_READ | FILE_SHARE_WRITE | FILE_SHARE_DELETE,
 208                     NULL,
 209                     OPEN_EXISTING,
 210                     /* FILE_FLAG_POSIX_SEMANTICS (treat file names that differ only
 211                        in case as different) makes sense only when applied to *all*
 212                        filesystem operations.  */
 213                     FILE_FLAG_BACKUP_SEMANTICS /* | FILE_FLAG_POSIX_SEMANTICS */,
 214                     NULL);
 215       if (h != INVALID_HANDLE_VALUE)
 216         {
 217           ret = _gl_fstat_by_handle (h, rname, buf);
 218           CloseHandle (h);
 219           goto done;
 220         }
 221     }
 222
 223     /* Test for root and UNC root directories.  */
 224     if ((rlen == drive_prefix_len + 1 && ISSLASH (rname[drive_prefix_len]))
 225         || is_unc_root (rname))
 226       goto failed;
 227
 228     /* Fallback.  */
 229     {
 230       /* Approach based on the directory entry.  */
 231
 232       if (strchr (rname, '?') != NULL || strchr (rname, '*') != NULL)
 233         {
 234           /* Other Windows API functions would fail with error
 235              ERROR_INVALID_NAME.  */
 236           if (malloca_rname != NULL)
 237             freea (malloca_rname);
 238           errno = ENOENT;
 239           return -1;
 240         }
 241
 242       /* Get the details about the directory entry.  This can be done through
 243          FindFirstFile
 244          <https://docs.microsoft.com/en-us/windows/desktop/api/fileapi/nf-fileapi-findfirstfilea>
 245          <https://docs.microsoft.com/en-us/windows/desktop/api/minwinbase/ns-minwinbase-_win32_find_dataa>
 246          or through
 247          FindFirstFileEx with argument FindExInfoBasic
 248          <https://docs.microsoft.com/en-us/windows/desktop/api/fileapi/nf-fileapi-findfirstfileexa>
 249          <https://docs.microsoft.com/en-us/windows/desktop/api/minwinbase/ne-minwinbase-findex_info_levels>
 250          <https://docs.microsoft.com/en-us/windows/desktop/api/minwinbase/ns-minwinbase-_win32_find_dataa>  */
 251       WIN32_FIND_DATA info;
 252       HANDLE h = FindFirstFile (rname, &info);
 253       if (h == INVALID_HANDLE_VALUE)
 254         goto failed;
 255
 256       /* Test for error conditions before starting to fill *buf.  */
 257       if (sizeof (buf->st_size) <= 4 && info.nFileSizeHigh > 0)
 258         {
 259           FindClose (h);
 260           if (malloca_rname != NULL)
 261             freea (malloca_rname);
 262           errno = EOVERFLOW;
 263           return -1;
 264         }
 265
 266 # if _GL_WINDOWS_STAT_INODES
 267       buf->st_dev = 0;
 268 #  if _GL_WINDOWS_STAT_INODES == 2
 269       buf->st_ino._gl_ino[0] = buf->st_ino._gl_ino[1] = 0;
 270 #  else /* _GL_WINDOWS_STAT_INODES == 1 */
 271       buf->st_ino = 0;
 272 #  endif
 273 # else
 274       /* st_ino is not wide enough for identifying a file on a device.
 275          Without st_ino, st_dev is pointless.  */
 276       buf->st_dev = 0;
 277       buf->st_ino = 0;
 278 # endif
 279
 280       /* st_mode.  */
 281       unsigned int mode =
 282         /* XXX How to handle FILE_ATTRIBUTE_REPARSE_POINT ?  */
 283         ((info.dwFileAttributes & FILE_ATTRIBUTE_DIRECTORY) ? _S_IFDIR | S_IEXEC_UGO : _S_IFREG)
 284         | S_IREAD_UGO
 285         | ((info.dwFileAttributes & FILE_ATTRIBUTE_READONLY) ? 0 : S_IWRITE_UGO);
 286       if (!(info.dwFileAttributes & FILE_ATTRIBUTE_DIRECTORY))
 287         {
 288           /* Determine whether the file is executable by looking at the file
 289              name suffix.  */
 290           if (info.nFileSizeHigh > 0 || info.nFileSizeLow > 0)
 291             {
 292               const char *last_dot = NULL;
 293               const char *p;
 294               for (p = info.cFileName; *p != '\0'; p++)
 295                 if (*p == '.')
 296                   last_dot = p;
 297               if (last_dot != NULL)
 298                 {
 299                   const char *suffix = last_dot + 1;
 300                   if (_stricmp (suffix, "exe") == 0
 301                       || _stricmp (suffix, "bat") == 0
 302                       || _stricmp (suffix, "cmd") == 0
 303                       || _stricmp (suffix, "com") == 0)
 304                     mode |= S_IEXEC_UGO;
 305                 }
 306             }
 307         }
 308       buf->st_mode = mode;
 309
 310       /* st_nlink.  Ignore hard links here.  */
 311       buf->st_nlink = 1;
 312
 313       /* There's no easy way to map the Windows SID concept to an integer.  */
 314       buf->st_uid = 0;
 315       buf->st_gid = 0;
 316
 317       /* st_rdev is irrelevant for normal files and directories.  */
 318       buf->st_rdev = 0;
 319
 320       /* st_size.  */
 321       if (sizeof (buf->st_size) <= 4)
 322         /* Range check already done above.  */
 323         buf->st_size = info.nFileSizeLow;
 324       else
 325         buf->st_size = ((long long) info.nFileSizeHigh << 32) | (long long) info.nFileSizeLow;
 326
 327       /* st_atime, st_mtime, st_ctime.  */
 328 # if _GL_WINDOWS_STAT_TIMESPEC
 329       buf->st_atim = _gl_convert_FILETIME_to_timespec (&info.ftLastAccessTime);
 330       buf->st_mtim = _gl_convert_FILETIME_to_timespec (&info.ftLastWriteTime);
 331       buf->st_ctim = _gl_convert_FILETIME_to_timespec (&info.ftCreationTime);
 332 # else
 333       buf->st_atime = _gl_convert_FILETIME_to_POSIX (&info.ftLastAccessTime);
 334       buf->st_mtime = _gl_convert_FILETIME_to_POSIX (&info.ftLastWriteTime);
 335       buf->st_ctime = _gl_convert_FILETIME_to_POSIX (&info.ftCreationTime);
 336 # endif
 337
 338       FindClose (h);
 339
 340       ret = 0;
 341     }
 342
 343    done:
 344     if (ret >= 0 && check_dir && !S_ISDIR (buf->st_mode))
 345       {
 346         errno = ENOTDIR;
 347         ret = -1;
 348       }
 349     if (malloca_rname != NULL)
 350       {
 351         int saved_errno = errno;
 352         freea (malloca_rname);
 353         errno = saved_errno;
 354       }
 355     return ret;
 356   }
 357
 358  failed:
 359   {
 360     DWORD error = GetLastError ();
 361     #if 0
 362     fprintf (stderr, "rpl_stat error 0x%x\n", (unsigned int) error);
 363     #endif
 364
 365     if (malloca_rname != NULL)
 366       freea (malloca_rname);
 367
 368     switch (error)
 369       {
 370       /* Some of these errors probably cannot happen with the specific flags
 371          that we pass to CreateFile.  But who knows...  */
 372       case ERROR_FILE_NOT_FOUND: /* The last component of rname does not exist.  */
 373       case ERROR_PATH_NOT_FOUND: /* Some directory component in rname does not exist.  */
 374       case ERROR_BAD_PATHNAME:   /* rname is such as '\\server'.  */
 375       case ERROR_BAD_NET_NAME:   /* rname is such as '\\server\nonexistentshare'.  */
 376       case ERROR_INVALID_NAME:   /* rname contains wildcards, misplaced colon, etc.  */
 377       case ERROR_DIRECTORY:
 378         errno = ENOENT;
 379         break;
 380
 381       case ERROR_ACCESS_DENIED:  /* rname is such as 'C:\System Volume Information\foo'.  */
 382       case ERROR_SHARING_VIOLATION: /* rname is such as 'C:\pagefile.sys' (second approach only).  */
 383                                     /* XXX map to EACCES or EPERM? */
 384         errno = EACCES;
 385         break;
 386
 387       case ERROR_OUTOFMEMORY:
 388         errno = ENOMEM;
 389         break;
 390
 391       case ERROR_WRITE_PROTECT:
 392         errno = EROFS;
 393         break;
 394
 395       case ERROR_WRITE_FAULT:
 396       case ERROR_READ_FAULT:
 397       case ERROR_GEN_FAILURE:
 398         errno = EIO;
 399         break;
 400
 401       case ERROR_BUFFER_OVERFLOW:
 402       case ERROR_FILENAME_EXCED_RANGE:
 403         errno = ENAMETOOLONG;
 404         break;
 405
 406       case ERROR_DELETE_PENDING: /* XXX map to EACCES or EPERM? */
 407         errno = EPERM;
 408         break;
 409
 410       default:
 411         errno = EINVAL;
 412         break;
 413       }
 414
 415     return -1;
 416   }
 417 #else
 418   int result = orig_stat (name, buf);
 419   if (result == 0)
 420     {
 421 # if REPLACE_FUNC_STAT_FILE
 422       /* Solaris 9 mistakenly succeeds when given a non-directory with a
 423          trailing slash.  */
 424       if (!S_ISDIR (buf->st_mode))
 425         {
 426           size_t len = strlen (name);
 427           if (ISSLASH (name[len - 1]))
 428             {
 429               errno = ENOTDIR;
 430               return -1;
 431             }
 432         }
 433 # endif /* REPLACE_FUNC_STAT_FILE */
 434       result = stat_time_normalize (result, buf);
 435     }
 436   return result;
 437 #endif
 438 }