src/rangecgi.c

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