ext/unicorn/http11/http11.c

   1 /**
   2  * Copyright (c) 2009 Eric Wong (all bugs are Eric's fault)
   3  * Copyright (c) 2005 Zed A. Shaw
   4  * You can redistribute it and/or modify it under the same terms as Ruby.
   5  */
   6 #include "ruby.h"
   7 #include "ext_help.h"
   8 #include <assert.h>
   9 #include <string.h>
  10 #include "http11_parser.h"
  11
  12 #ifndef RSTRING_PTR
  13 #define RSTRING_PTR(s) (RSTRING(s)->ptr)
  14 #endif
  15 #ifndef RSTRING_LEN
  16 #define RSTRING_LEN(s) (RSTRING(s)->len)
  17 #endif
  18
  19 static VALUE mUnicorn;
  20 static VALUE cHttpParser;
  21 static VALUE eHttpParserError;
  22 static VALUE sym_http_body;
  23
  24 #define HTTP_PREFIX "HTTP_"
  25 #define HTTP_PREFIX_LEN (sizeof(HTTP_PREFIX) - 1)
  26
  27 static VALUE global_rack_url_scheme;
  28 static VALUE global_request_method;
  29 static VALUE global_request_uri;
  30 static VALUE global_fragment;
  31 static VALUE global_query_string;
  32 static VALUE global_http_version;
  33 static VALUE global_request_path;
  34 static VALUE global_path_info;
  35 static VALUE global_server_name;
  36 static VALUE global_server_port;
  37 static VALUE global_server_protocol;
  38 static VALUE global_server_protocol_value;
  39 static VALUE global_http_host;
  40 static VALUE global_http_x_forwarded_proto;
  41 static VALUE global_port_80;
  42 static VALUE global_port_443;
  43 static VALUE global_localhost;
  44 static VALUE global_http;
  45
  46 /** Defines common length and error messages for input length validation. */
  47 #define DEF_MAX_LENGTH(N,length) const size_t MAX_##N##_LENGTH = length; const char *MAX_##N##_LENGTH_ERR = "HTTP element " # N  " is longer than the " # length " allowed length."
  48
  49 /** Validates the max length of given input and throws an HttpParserError exception if over. */
  50 #define VALIDATE_MAX_LENGTH(len, N) if(len > MAX_##N##_LENGTH) { rb_raise(eHttpParserError, MAX_##N##_LENGTH_ERR); }
  51
  52 /** Defines global strings in the init method. */
  53 #define DEF_GLOBAL(N, val)   global_##N = rb_obj_freeze(rb_str_new2(val)); rb_global_variable(&global_##N)
  54
  55
  56 /* Defines the maximum allowed lengths for various input elements.*/
  57 DEF_MAX_LENGTH(FIELD_NAME, 256);
  58 DEF_MAX_LENGTH(FIELD_VALUE, 80 * 1024);
  59 DEF_MAX_LENGTH(REQUEST_URI, 1024 * 12);
  60 DEF_MAX_LENGTH(FRAGMENT, 1024); /* Don't know if this length is specified somewhere or not */
  61 DEF_MAX_LENGTH(REQUEST_PATH, 1024);
  62 DEF_MAX_LENGTH(QUERY_STRING, (1024 * 10));
  63 DEF_MAX_LENGTH(HEADER, (1024 * (80 + 32)));
  64
  65 struct common_field {
  66         const signed long len;
  67         const char *name;
  68         VALUE value;
  69 };
  70
  71 /*
  72  * A list of common HTTP headers we expect to receive.
  73  * This allows us to avoid repeatedly creating identical string
  74  * objects to be used with rb_hash_aset().
  75  */
  76 static struct common_field common_http_fields[] = {
  77 # define f(N) { (sizeof(N) - 1), N, Qnil }
  78         f("ACCEPT"),
  79         f("ACCEPT_CHARSET"),
  80         f("ACCEPT_ENCODING"),
  81         f("ACCEPT_LANGUAGE"),
  82         f("ALLOW"),
  83         f("AUTHORIZATION"),
  84         f("CACHE_CONTROL"),
  85         f("CONNECTION"),
  86         f("CONTENT_ENCODING"),
  87         f("CONTENT_LENGTH"),
  88         f("CONTENT_TYPE"),
  89         f("COOKIE"),
  90         f("DATE"),
  91         f("EXPECT"),
  92         f("FROM"),
  93         f("HOST"),
  94         f("IF_MATCH"),
  95         f("IF_MODIFIED_SINCE"),
  96         f("IF_NONE_MATCH"),
  97         f("IF_RANGE"),
  98         f("IF_UNMODIFIED_SINCE"),
  99         f("KEEP_ALIVE"), /* Firefox sends this */
 100         f("MAX_FORWARDS"),
 101         f("PRAGMA"),
 102         f("PROXY_AUTHORIZATION"),
 103         f("RANGE"),
 104         f("REFERER"),
 105         f("TE"),
 106         f("TRAILER"),
 107         f("TRANSFER_ENCODING"),
 108         f("UPGRADE"),
 109         f("USER_AGENT"),
 110         f("VIA"),
 111         f("X_FORWARDED_FOR"), /* common for proxies */
 112         f("X_FORWARDED_PROTO"), /* common for proxies */
 113         f("X_REAL_IP"), /* common for proxies */
 114         f("WARNING")
 115 # undef f
 116 };
 117
 118 /* this function is not performance-critical */
 119 static void init_common_fields(void)
 120 {
 121   int i;
 122   struct common_field *cf = common_http_fields;
 123   char tmp[256]; /* MAX_FIELD_NAME_LENGTH */
 124   memcpy(tmp, HTTP_PREFIX, HTTP_PREFIX_LEN);
 125
 126   for(i = 0; i < ARRAY_SIZE(common_http_fields); cf++, i++) {
 127     /* Rack doesn't like certain headers prefixed with "HTTP_" */
 128     if (!strcmp("CONTENT_LENGTH", cf->name) ||
 129         !strcmp("CONTENT_TYPE", cf->name)) {
 130       cf->value = rb_str_new(cf->name, cf->len);
 131     } else {
 132       memcpy(tmp + HTTP_PREFIX_LEN, cf->name, cf->len + 1);
 133       cf->value = rb_str_new(tmp, HTTP_PREFIX_LEN + cf->len);
 134     }
 135     cf->value = rb_obj_freeze(cf->value);
 136     rb_global_variable(&cf->value);
 137   }
 138 }
 139
 140 static VALUE find_common_field_value(const char *field, size_t flen)
 141 {
 142   int i;
 143   struct common_field *cf = common_http_fields;
 144   for(i = 0; i < ARRAY_SIZE(common_http_fields); i++, cf++) {
 145     if (cf->len == flen && !memcmp(cf->name, field, flen))
 146       return cf->value;
 147   }
 148   return Qnil;
 149 }
 150
 151 static void http_field(void *data, const char *field,
 152                        size_t flen, const char *value, size_t vlen)
 153 {
 154   VALUE req = (VALUE)data;
 155   VALUE v = Qnil;
 156   VALUE f = Qnil;
 157
 158   VALIDATE_MAX_LENGTH(flen, FIELD_NAME);
 159   VALIDATE_MAX_LENGTH(vlen, FIELD_VALUE);
 160
 161   v = rb_str_new(value, vlen);
 162
 163   f = find_common_field_value(field, flen);
 164
 165   if (f == Qnil) {
 166     /*
 167      * We got a strange header that we don't have a memoized value for.
 168      * Fallback to creating a new string to use as a hash key.
 169      *
 170      * using rb_str_new(NULL, len) here is faster than rb_str_buf_new(len)
 171      * in my testing, because: there's no minimum allocation length (and
 172      * no check for it, either), RSTRING_LEN(f) does not need to be
 173      * written twice, and and RSTRING_PTR(f) will already be
 174      * null-terminated for us.
 175      */
 176     f = rb_str_new(NULL, HTTP_PREFIX_LEN + flen);
 177     memcpy(RSTRING_PTR(f), HTTP_PREFIX, HTTP_PREFIX_LEN);
 178     memcpy(RSTRING_PTR(f) + HTTP_PREFIX_LEN, field, flen);
 179     assert(*(RSTRING_PTR(f) + RSTRING_LEN(f)) == '\0'); /* paranoia */
 180     /* fprintf(stderr, "UNKNOWN HEADER <%s>\n", RSTRING_PTR(f)); */
 181   }
 182
 183   rb_hash_aset(req, f, v);
 184 }
 185
 186 static void request_method(void *data, const char *at, size_t length)
 187 {
 188   VALUE req = (VALUE)data;
 189   VALUE val = Qnil;
 190
 191   val = rb_str_new(at, length);
 192   rb_hash_aset(req, global_request_method, val);
 193 }
 194
 195 static void request_uri(void *data, const char *at, size_t length)
 196 {
 197   VALUE req = (VALUE)data;
 198   VALUE val = Qnil;
 199
 200   VALIDATE_MAX_LENGTH(length, REQUEST_URI);
 201
 202   val = rb_str_new(at, length);
 203   rb_hash_aset(req, global_request_uri, val);
 204
 205   /* "OPTIONS * HTTP/1.1\r\n" is a valid request */
 206   if (length == 1 && *at == '*') {
 207     val = rb_str_new(NULL, 0);
 208     rb_hash_aset(req, global_request_path, val);
 209     rb_hash_aset(req, global_path_info, val);
 210   }
 211 }
 212
 213 static void fragment(void *data, const char *at, size_t length)
 214 {
 215   VALUE req = (VALUE)data;
 216   VALUE val = Qnil;
 217
 218   VALIDATE_MAX_LENGTH(length, FRAGMENT);
 219
 220   val = rb_str_new(at, length);
 221   rb_hash_aset(req, global_fragment, val);
 222 }
 223
 224 static void request_path(void *data, const char *at, size_t length)
 225 {
 226   VALUE req = (VALUE)data;
 227   VALUE val = Qnil;
 228
 229   VALIDATE_MAX_LENGTH(length, REQUEST_PATH);
 230
 231   val = rb_str_new(at, length);
 232   rb_hash_aset(req, global_request_path, val);
 233
 234   /* rack says PATH_INFO must start with "/" or be empty */
 235   if (!(length == 1 && *at == '*'))
 236     rb_hash_aset(req, global_path_info, val);
 237 }
 238
 239 static void query_string(void *data, const char *at, size_t length)
 240 {
 241   VALUE req = (VALUE)data;
 242   VALUE val = Qnil;
 243
 244   VALIDATE_MAX_LENGTH(length, QUERY_STRING);
 245
 246   val = rb_str_new(at, length);
 247   rb_hash_aset(req, global_query_string, val);
 248 }
 249
 250 static void http_version(void *data, const char *at, size_t length)
 251 {
 252   VALUE req = (VALUE)data;
 253   VALUE val = rb_str_new(at, length);
 254   rb_hash_aset(req, global_http_version, val);
 255 }
 256
 257 /** Finalizes the request header to have a bunch of stuff that's needed. */
 258 static void header_done(void *data, const char *at, size_t length)
 259 {
 260   VALUE req = (VALUE)data;
 261   VALUE server_name = global_localhost;
 262   VALUE server_port = global_port_80;
 263   VALUE temp;
 264
 265   /* rack requires QUERY_STRING */
 266   if (rb_hash_aref(req, global_query_string) == Qnil)
 267     rb_hash_aset(req, global_query_string, rb_str_new(NULL, 0));
 268
 269   /* set rack.url_scheme to "https" or "http", no others are allowed by Rack */
 270   if ((temp = rb_hash_aref(req, global_http_x_forwarded_proto)) != Qnil &&
 271       RSTRING_LEN(temp) == 5 &&
 272       !memcmp("https", RSTRING_PTR(temp), 5))
 273     server_port = global_port_443;
 274   else
 275     temp = global_http;
 276   rb_hash_aset(req, global_rack_url_scheme, temp);
 277
 278   /* parse and set the SERVER_NAME and SERVER_PORT variables */
 279   if ((temp = rb_hash_aref(req, global_http_host)) != Qnil) {
 280     char *colon = memchr(RSTRING_PTR(temp), ':', RSTRING_LEN(temp));
 281     if (colon) {
 282       server_name = rb_str_substr(temp, 0, colon - RSTRING_PTR(temp));
 283       server_port = rb_str_substr(temp, colon - RSTRING_PTR(temp)+1,
 284                                   RSTRING_LEN(temp));
 285     } else {
 286       server_name = temp;
 287     }
 288   }
 289   rb_hash_aset(req, global_server_name, server_name);
 290   rb_hash_aset(req, global_server_port, server_port);
 291
 292   /* grab the initial body and stuff it into the hash */
 293   rb_hash_aset(req, sym_http_body, rb_str_new(at, length));
 294   rb_hash_aset(req, global_server_protocol, global_server_protocol_value);
 295 }
 296
 297 static void HttpParser_free(void *data) {
 298   TRACE();
 299
 300   if(data) {
 301     free(data);
 302   }
 303 }
 304
 305
 306 static VALUE HttpParser_alloc(VALUE klass)
 307 {
 308   VALUE obj;
 309   http_parser *hp = ALLOC_N(http_parser, 1);
 310   TRACE();
 311   hp->http_field = http_field;
 312   hp->request_method = request_method;
 313   hp->request_uri = request_uri;
 314   hp->fragment = fragment;
 315   hp->request_path = request_path;
 316   hp->query_string = query_string;
 317   hp->http_version = http_version;
 318   hp->header_done = header_done;
 319   http_parser_init(hp);
 320
 321   obj = Data_Wrap_Struct(klass, NULL, HttpParser_free, hp);
 322
 323   return obj;
 324 }
 325
 326
 327 /**
 328  * call-seq:
 329  *    parser.new -> parser
 330  *
 331  * Creates a new parser.
 332  */
 333 static VALUE HttpParser_init(VALUE self)
 334 {
 335   http_parser *http = NULL;
 336   DATA_GET(self, http_parser, http);
 337   http_parser_init(http);
 338
 339   return self;
 340 }
 341
 342
 343 /**
 344  * call-seq:
 345  *    parser.reset -> nil
 346  *
 347  * Resets the parser to it's initial state so that you can reuse it
 348  * rather than making new ones.
 349  */
 350 static VALUE HttpParser_reset(VALUE self)
 351 {
 352   http_parser *http = NULL;
 353   DATA_GET(self, http_parser, http);
 354   http_parser_init(http);
 355
 356   return Qnil;
 357 }
 358
 359
 360 /**
 361  * call-seq:
 362  *    parser.execute(req_hash, data) -> true/false
 363  *
 364  * Takes a Hash and a String of data, parses the String of data filling
 365  * in the Hash returning a boolean to indicate whether or not parsing
 366  * is finished.
 367  *
 368  * This function now throws an exception when there is a parsing error.
 369  * This makes the logic for working with the parser much easier.  You
 370  * will need to wrap the parser with an exception handling block.
 371  */
 372
 373 static VALUE HttpParser_execute(VALUE self, VALUE req_hash, VALUE data)
 374 {
 375   http_parser *http;
 376   char *dptr = RSTRING_PTR(data);
 377   long dlen = RSTRING_LEN(data);
 378
 379   DATA_GET(self, http_parser, http);
 380
 381   if (http->nread < dlen) {
 382     http->data = (void *)req_hash;
 383     http_parser_execute(http, dptr, dlen);
 384
 385     VALIDATE_MAX_LENGTH(http->nread, HEADER);
 386
 387     if (!http_parser_has_error(http))
 388       return http_parser_is_finished(http) ? Qtrue : Qfalse;
 389
 390     rb_raise(eHttpParserError, "Invalid HTTP format, parsing fails.");
 391   }
 392   rb_raise(eHttpParserError, "Requested start is after data buffer end.");
 393 }
 394
 395 void Init_http11()
 396 {
 397
 398   mUnicorn = rb_define_module("Unicorn");
 399
 400   DEF_GLOBAL(rack_url_scheme, "rack.url_scheme");
 401   DEF_GLOBAL(request_method, "REQUEST_METHOD");
 402   DEF_GLOBAL(request_uri, "REQUEST_URI");
 403   DEF_GLOBAL(fragment, "FRAGMENT");
 404   DEF_GLOBAL(query_string, "QUERY_STRING");
 405   DEF_GLOBAL(http_version, "HTTP_VERSION");
 406   DEF_GLOBAL(request_path, "REQUEST_PATH");
 407   DEF_GLOBAL(path_info, "PATH_INFO");
 408   DEF_GLOBAL(server_name, "SERVER_NAME");
 409   DEF_GLOBAL(server_port, "SERVER_PORT");
 410   DEF_GLOBAL(server_protocol, "SERVER_PROTOCOL");
 411   DEF_GLOBAL(server_protocol_value, "HTTP/1.1");
 412   DEF_GLOBAL(http_host, "HTTP_HOST");
 413   DEF_GLOBAL(http_x_forwarded_proto, "HTTP_X_FORWARDED_PROTO");
 414   DEF_GLOBAL(port_80, "80");
 415   DEF_GLOBAL(port_443, "443");
 416   DEF_GLOBAL(localhost, "localhost");
 417   DEF_GLOBAL(http, "http");
 418
 419   eHttpParserError = rb_define_class_under(mUnicorn, "HttpParserError", rb_eIOError);
 420
 421   cHttpParser = rb_define_class_under(mUnicorn, "HttpParser", rb_cObject);
 422   rb_define_alloc_func(cHttpParser, HttpParser_alloc);
 423   rb_define_method(cHttpParser, "initialize", HttpParser_init,0);
 424   rb_define_method(cHttpParser, "reset", HttpParser_reset,0);
 425   rb_define_method(cHttpParser, "execute", HttpParser_execute,2);
 426   sym_http_body = ID2SYM(rb_intern("http_body"));
 427   init_common_fields();
 428 }