src/rangecgi.c

   1 /*
   2
   3 rangecgi.c -- rangecgi utility to serve multiple files as one with range support
   4 Copyright (C) 2014,2015,2016,2019,2020 Kyle J. McKay
   5 All rights reserved
   6
   7 This program is free software; you can redistribute it and/or
   8 modify it under the terms of the GNU General Public License
   9 as published by the Free Software Foundation; either version 2
  10 of the License, or (at your option) any later version.
  11
  12 This program is distributed in the hope that it will be useful,
  13 but WITHOUT ANY WARRANTY; without even the implied warranty of
  14 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  15 GNU General Public License for more details.
  16
  17 You should have received a copy of the GNU General Public License
  18 along with this program; if not, write to the Free Software
  19 Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301, USA.
  20
  21 */
  22
  23 /*
  24   This utility serves multiple files (currently exactly two) as though they
  25   were one file and allows a continuation download using the "Range:" header.
  26
  27   Only GET and HEAD requests are supported with either no "Range:" header or
  28   a "Range:" header with exactly one range.
  29
  30   USAGE:
  31     rangecgi ([--etag] | [-c <content-type>] [-f <filename>] [-e <days>]) [-m 0|1|2] file1 file2
  32
  33   If --etag is given then all environment variables are ignored and the
  34   computed ETag value (with the "", but without the "ETag:" prefix part) is
  35   output to standard output on success.  Otherwise there is no output and the
  36   exit code will be non-zero.
  37
  38   If --etag is given then no other options except -m are allowed.  If
  39   -c <content-type> is given then the specified content type will be used as-is
  40   for the returned item.  If -e <days> is given then a cache-control and expires
  41   header will be output with the expiration set that many days into the future.
  42   If -f <filename> is given, then a "Content-Disposition: attachment;
  43   filename=<filename>" header will be emitted.  If no -f <filename> option is
  44   given, then no "Content-Disposition:" header will be emitted!
  45
  46   The default is "-m 0" which means use the latest mtime of the given files
  47   when computing the ETag value.  With "-m 1" always use the mtime from the
  48   first file and with "-m 2" always use the mtime from the second file.
  49
  50   Other CGI parameters MUST be passed as environment variables in particular
  51   REQUEST_METHOD MUST be set and to request a range, HTTP_RANGE MUST be set.
  52   HTTP_IF_RANGE MAY be set.  No other environment variables are examined.
  53
  54   Exit code 0 for CGI success (Status: header etc. output)
  55   Exit code 1 for no REQUEST_METHOD.
  56   Exit code 2 for file1 and/or file2 not given or wrong type or bad option.
  57   Exit code 3 for --etag mode and file1 and/or file2 could not be opened.
  58
  59   If file1 and/or file2 is not found a 404 status will be output.
  60
  61   Normally a front end script will begin processing the initial CGI request
  62   from the web server and then pass it on to this utility as appropriate.
  63
  64   If a "Range:" header is present and a non-empty "If-Range:" header is also
  65   present then if the value in the "If-Range:" header does not exactly match
  66   the computed "ETag:" value then the "Range:" header will be silently ignored.
  67 */
  68
  69 #undef _FILE_OFFSET_BITS
  70 #define _FILE_OFFSET_BITS 64
  71 #include <sys/types.h>
  72 #include <sys/uio.h>
  73 #include <errno.h>
  74 #include <fcntl.h>
  75 #include <stdio.h>
  76 #include <stdlib.h>
  77 #include <string.h>
  78 #include <sys/stat.h>
  79 #include <time.h>
  80 #include <unistd.h>
  81 #ifdef __APPLE__
  82 #include <AvailabilityMacros.h>
  83 #ifndef MAC_OS_X_VERSION_10_5
  84 #define MAC_OS_X_VERSION_10_5 1050
  85 #endif
  86 #if MAC_OS_X_VERSION_MIN_REQUIRED < MAC_OS_X_VERSION_10_5
  87 typedef struct stat statrec;
  88 #define fstatfunc(p,b) fstat(p,b)
  89 #else
  90 typedef struct stat64 statrec;
  91 #define fstatfunc(p,b) fstat64(p,b)
  92 #endif
  93 #else
  94 typedef struct stat statrec;
  95 #define fstatfunc(p,b) fstat(p,b)
  96 #endif
  97 typedef unsigned long long bignum;
  98
  99 static void errorfail_(unsigned code, const char *status, const char *extrahdr)
 100 {
 101   (void)code;
 102   (void)status;
 103   (void)extrahdr;
 104   exit(3);
 105 }
 106
 107 static void errorexit_(unsigned code, const char *status, const char *extrahdr)
 108 {
 109   printf("Status: %u %s\r\n", code, status);
 110   printf("%s\r\n", "Expires: Fri, 01 Jan 1980 00:00:00 GMT");
 111   printf("%s\r\n", "Pragma: no-cache");
 112   printf("%s\r\n", "Cache-Control: no-cache,max-age=0,must-revalidate");
 113   printf("%s\r\n", "Accept-Ranges: bytes");
 114   if (extrahdr)
 115     printf("%s\r\n", extrahdr);
 116   printf("%s\r\n", "Content-Type: text/plain; charset=utf-8; format=fixed");
 117   printf("%s\r\n", "");
 118   printf("%s\n", status);
 119   fflush(stdout);
 120   exit(0);
 121 }
 122
 123 static void emithdrs(const char *ct, const char *fn, int exp, time_t lm,
 124                      const char *etag, bignum tl, int isr, bignum r1, bignum r2)
 125 {
 126   struct tm gt;
 127   char dtstr[32];
 128   const char *xtra = "";
 129
 130   if (isr)
 131     if (isr > 0)
 132       printf("Status: %u %s\r\n", 206, "Partial Content");
 133     else
 134       printf("Status: %u %s\r\n", 416, "Requested Range Not Satisfiable");
 135   else
 136     printf("Status: %u %s\r\n", 200, "OK");
 137   if (exp > 0) {
 138     time_t epsecs = time(NULL);
 139     long esecs = 86400 * exp;
 140     gt = *gmtime(&epsecs);
 141     strftime(dtstr, sizeof(dtstr), "%a, %d %b %Y %H:%M:%S GMT", &gt);
 142     printf("Date: %s\r\n", dtstr);
 143     epsecs += esecs;
 144     gt = *gmtime(&epsecs);
 145     strftime(dtstr, sizeof(dtstr), "%a, %d %b %Y %H:%M:%S GMT", &gt);
 146     printf("Expires: %s\r\n", dtstr);
 147     printf("Cache-Control: public,max-age=%ld\r\n", esecs);
 148   } else if (!exp) {
 149     printf("%s\r\n", "Expires: Fri, 01 Jan 1980 00:00:00 GMT");
 150     printf("%s\r\n", "Pragma: no-cache");
 151     printf("%s\r\n", "Cache-Control: no-cache,max-age=0,must-revalidate");
 152   }
 153   printf("%s\r\n", "Accept-Ranges: bytes");
 154   gt = *gmtime(&lm);
 155   strftime(dtstr, sizeof(dtstr), "%a, %d %b %Y %H:%M:%S GMT", &gt);
 156   printf("Last-Modified: %s\r\n", dtstr);
 157   if (etag)
 158     printf("ETag: %s\r\n", etag);
 159   if (!isr) {
 160     printf("Content-Length: %llu\r\n", tl);
 161   } else if (isr > 0) {
 162     printf("Content-Length: %llu\r\n", r2 - r1 + 1);
 163     printf("Content-Range: bytes %llu-%llu/%llu\r\n", r1, r2, tl);
 164   } else {
 165     printf("Content-Range: bytes */%llu\r\n", tl);
 166   }
 167   if (isr >= 0) {
 168     if (!ct || !*ct)
 169       ct = "application/octet-stream";
 170     printf("Content-Type: %s\r\n", ct);
 171     if (fn && *fn)
 172       printf("Content-Disposition: attachment; filename=\"%s\"\r\n", fn);
 173     printf("%s\r\n", "Vary: Accept-Encoding");
 174   } else {
 175     printf("%s\r\n", "Content-Type: text/plain; charset=utf-8; format=fixed");
 176     xtra = "Requested Range Not Satisfiable\n";
 177   }
 178   printf("\r\n%s", xtra);
 179 }
 180
 181 static void error416(time_t lm, const char *etag, bignum tl)
 182 {
 183   emithdrs(NULL, NULL, -1, lm, etag, tl, -1, 0, 0);
 184   fflush(stdout);
 185   exit(0);
 186 }
 187
 188 static void die(const char *msg)
 189 {
 190   fprintf(stderr, "%s\n", msg);
 191   fflush(stderr);
 192   exit(2);
 193 }
 194
 195 static void readx(int fd, void *buf, size_t count)
 196 {
 197   char *buff = (char *)buf;
 198   int err = 0;
 199   while (count && (!err || err == EINTR || err == EAGAIN || err == EWOULDBLOCK)) {
 200     ssize_t amt = read(fd, buff, count);
 201     if (amt == -1) {
 202       err = errno;
 203       continue;
 204     }
 205     err = 0;
 206     if (!amt)
 207       break;
 208     buff += (size_t)amt;
 209     count -= (size_t)amt;
 210   }
 211   if (count) {
 212     if (err)
 213       die("failed reading file (error)");
 214     else
 215       die("failed reading file (EOF)");
 216   }
 217 }
 218
 219 static void writex(int fd, const void *buf, size_t count)
 220 {
 221   const char *buff = (const char *)buf;
 222   int err = 0;
 223   while (count && (!err || err == EINTR || err == EAGAIN || err == EWOULDBLOCK)) {
 224     ssize_t amt = write(fd, buff, count);
 225     if (amt == -1) {
 226       err = errno;
 227       continue;
 228     }
 229     err = 0;
 230     if (!amt)
 231       break;
 232     buff += (size_t)amt;
 233     count -= (size_t)amt;
 234   }
 235   if (count) {
 236     if (err)
 237       die("failed writing file (error)");
 238     else
 239       die("failed writing file (EOF)");
 240   }
 241 }
 242
 243 #define SIZEPWR 15
 244 static char dumpbuff[1U << SIZEPWR];
 245 void dumpfile(int fd, bignum start, bignum len)
 246 {
 247   off_t loc = lseek(fd, (off_t)start, SEEK_SET);
 248   size_t maxread;
 249   if (loc != (off_t)start)
 250     die("lseek failed");
 251   if (start & ((bignum)(sizeof(dumpbuff) - 1)))
 252     maxread = sizeof(dumpbuff) - (size_t)(start & ((bignum)(sizeof(dumpbuff) - 1)));
 253   else
 254     maxread = sizeof(dumpbuff);
 255   while (len) {
 256     size_t cnt = len > (bignum)maxread ? maxread : (size_t)len;
 257     readx(fd, dumpbuff, cnt);
 258     writex(STDOUT_FILENO, dumpbuff, cnt);
 259     len -= (bignum)cnt;
 260     maxread = sizeof(dumpbuff);
 261   }
 262 }
 263 #undef SIZEPWR
 264
 265 int main(int argc, char *argv[])
 266 {
 267   int isetag = 0;
 268   void (*errorexit)(unsigned,const char *,const char *);
 269   statrec f1, f2;
 270   int e1, e2, i;
 271   bignum l1, l2, tl;
 272   bignum r1=0, r2=0;
 273   bignum start, length;
 274   time_t lm;
 275   const char *rm = NULL;
 276   const char *hr = NULL;
 277   const char *hir = NULL;
 278   const char *ct = NULL;
 279   const char *fn = NULL;
 280   int expdays = -1, opt_e = 0;
 281   /* "inode_inode-size-time_t_micros" each in hex up to 8 bytes gives */
 282   /* "16bytes_16bytes-16bytes-16bytes" plus NUL = 70 bytes (including "") */
 283   char etag[70];
 284   int fd1 = -1, fd2 = -1;
 285   int mno = 0;
 286
 287   opterr = 0;
 288   for (;;) {
 289     int ch;
 290     if (optind < argc && !strcmp(argv[optind], "--etag")) {
 291       ch = -2;
 292       ++optind;
 293     } else {
 294       ch = getopt(argc, argv, "c:e:f:m:");
 295     }
 296     if (ch == -1)
 297       break;
 298     switch (ch) {
 299       case -2:
 300         isetag = 1;
 301         break;
 302       case 'c':
 303         ct = optarg;
 304         break;
 305       case 'f':
 306         fn = optarg;
 307         break;
 308       case 'e':
 309       {
 310         int v, n;
 311         if (sscanf(optarg, "%i%n", &v, &n) != 1 || n != (int)strlen(optarg))
 312           exit(2);
 313         expdays = v;
 314         opt_e = 1;
 315         break;
 316       }
 317       case 'm':
 318         if (!optarg[0] || optarg[1])
 319           exit(2);
 320         if (optarg[0] != '0' && optarg[0] != '1' && optarg[0] != '2')
 321           exit(2);
 322         mno = optarg[0] - '0';
 323         break;
 324       default:
 325         exit(2);
 326     }
 327   }
 328   if (argc - optind != 2)
 329     exit(2);
 330   i = optind;
 331
 332   if (isetag) {
 333     if (ct || fn || opt_e)
 334       exit(2);
 335     errorexit = errorfail_;
 336   } else {
 337     rm = getenv("REQUEST_METHOD");
 338     if (!rm || !*rm)
 339       exit(1);
 340     hr = getenv("HTTP_RANGE");
 341     if (hr)
 342       hir = getenv("HTTP_IF_RANGE");
 343     errorexit = errorexit_;
 344     if (strcmp(rm, "GET") && strcmp(rm, "HEAD"))
 345       errorexit(405, "Method Not Allowed", "Allow: GET,HEAD");
 346   }
 347
 348   fd1 = open(argv[i], O_RDONLY);
 349   e1 = fd1 >= 0 ? 0 : errno;
 350   fd2 = open(argv[i+1], O_RDONLY);
 351   e2 = fd2 >= 0 ? 0 : errno;
 352   if (e1 == EACCES || e2 == EACCES)
 353     errorexit(403, "Forbidden", NULL);
 354   if (e1 == ENOENT || e1 == ENOTDIR || e2 == ENOENT || e2 == ENOTDIR)
 355     errorexit(404, "Not Found", NULL);
 356   e1 = fstatfunc(fd1, &f1) ? errno : 0;
 357   e2 = fstatfunc(fd2, &f2) ? errno : 0;
 358   if (e1 || e2)
 359     errorexit(500, "Internal Server Error", NULL);
 360   if (!S_ISREG(f1.st_mode) || !S_ISREG(f2.st_mode))
 361     errorexit(500, "Internal Server Error", NULL);
 362   if (mno == 1)
 363     lm = f1.st_mtime;
 364   else if (mno == 2)
 365     lm = f2.st_mtime;
 366   else if (f1.st_mtime >= f2.st_mtime)
 367     lm = f1.st_mtime;
 368   else
 369     lm = f2.st_mtime;
 370   l1 = f1.st_size;
 371   l2 = f2.st_size;
 372   tl = l1 + l2;
 373   sprintf(etag, "\"%llx_%llx-%llx-%llx\"", (unsigned long long)f1.st_ino,
 374     (unsigned long long)f2.st_ino, tl, (unsigned long long)lm * 1000000U);
 375
 376   if (isetag) {
 377     close(fd2);
 378     close(fd1);
 379     printf("%s\n", etag);
 380     exit(0);
 381   }
 382
 383   if (hir && *hir && strcmp(etag, hir))
 384     hr = NULL;
 385
 386   if (hr && !tl)
 387     error416(lm, etag, tl);  /* Range: not allowed on zero length content */
 388
 389   if (hr) {
 390     /* Only one range may be specified and it must be bytes */
 391     /* with a 2^64 value we could have "Range: bytes = 20-digit - 20-digit" */
 392     int pos = -1;
 393     int s = sscanf(hr, " %*[Bb]%*[Yy]%*[Tt]%*[Ee]%*[Ss] = %n", &pos);
 394     if (s != 0 || pos < 6 || strchr(hr, ','))
 395       errorexit(400, "Bad Request", NULL);
 396     hr += pos;
 397     if (*hr == '-') {
 398       bignum trail;
 399       ++hr;
 400       /* It's a request for the trailing part */
 401       if (strchr(hr, '-'))
 402         errorexit(400, "Bad Request", NULL);
 403       pos = -1;
 404       s = sscanf(hr, " %llu%n", &trail, &pos);
 405       if (s != 1 || pos < 1 || hr[pos])
 406         errorexit(400, "Bad Request", NULL);
 407       if (!trail || trail > tl)
 408         error416(lm, etag, tl);
 409       r1 = tl - trail;
 410       r2 = tl - 1;
 411     } else {
 412       pos = -1;
 413       s = sscanf(hr, "%llu - %n", &r1, &pos);
 414       if (s != 1 || pos < 2)
 415         errorexit(400, "Bad Request", NULL);
 416       hr += pos;
 417       if (*hr) {
 418         if (*hr == '-')
 419           errorexit(400, "Bad Request", NULL);
 420         pos = -1;
 421         s = sscanf(hr, "%llu %n", &r2, &pos);
 422         if (s != 1 || pos < 1 || hr[pos])
 423           errorexit(400, "Bad Request", NULL);
 424       } else {
 425         r2 = tl - 1;
 426       }
 427       if (r1 > r2 || r2 >= tl)
 428         error416(lm, etag, tl);
 429     }
 430     start = r1;
 431     length = r2 - r1 + 1;
 432   } else {
 433     start = 0;
 434     length = tl;
 435   }
 436
 437   emithdrs(ct, fn, expdays, lm, etag, tl, hr?1:0, r1, r2);
 438   fflush(stdout);
 439
 440   if (strcmp(rm, "HEAD")) {
 441     if (start < l1) {
 442       bignum dl = l1 - start;
 443       if (dl > length) dl = length;
 444       dumpfile(fd1, start, dl);
 445       start += dl;
 446       length -= dl;
 447     }
 448     if (length && start >= l1) {
 449       bignum dl;
 450       start -= l1;
 451       dl = l2 - start;
 452       if (dl > length) dl = length;
 453       dumpfile(fd2, start, dl);
 454     }
 455   }
 456
 457   close(fd2);
 458   close(fd1);
 459   return 0;
 460 }