http_parser.c 59.7 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
/* Based on src/http/ngx_http_parse.c from NGINX copyright Igor Sysoev
 *
 * Additional changes are licensed under the same terms as NGINX and
 * copyright Joyent, Inc. and other Node contributors. All rights reserved.
 *
 * Permission is hereby granted, free of charge, to any person obtaining a copy
 * of this software and associated documentation files (the "Software"), to
 * deal in the Software without restriction, including without limitation the
 * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
 * sell copies of the Software, and to permit persons to whom the Software is
 * furnished to do so, subject to the following conditions:
 *
 * The above copyright notice and this permission notice shall be included in
 * all copies or substantial portions of the Software.
 *
 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
 * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
 * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
 * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
 * IN THE SOFTWARE.
 */
24
#include "http_parser.h"
25 26
#include <assert.h>
#include <stddef.h>
27 28 29 30
#include <ctype.h>
#include <stdlib.h>
#include <string.h>
#include <limits.h>
31

32 33 34
#ifndef ULLONG_MAX
# define ULLONG_MAX ((uint64_t) -1) /* 2^64-1 */
#endif
35 36 37 38 39

#ifndef MIN
# define MIN(a,b) ((a) < (b) ? (a) : (b))
#endif

40 41 42 43 44 45 46 47 48 49 50 51 52
#ifndef ARRAY_SIZE
# define ARRAY_SIZE(a) (sizeof(a) / sizeof((a)[0]))
#endif

#ifndef BIT_AT
# define BIT_AT(a, i)                                                \
  (!!((unsigned int) (a)[(unsigned int) (i) >> 3] &                  \
   (1 << ((unsigned int) (i) & 7))))
#endif

#ifndef ELEM_AT
# define ELEM_AT(a, i, v) ((unsigned int) (i) < ARRAY_SIZE(a) ? (a)[(i)] : (v))
#endif
53 54 55 56 57 58 59

#define SET_ERRNO(e)                                                 \
do {                                                                 \
  parser->http_errno = (e);                                          \
} while(0)


60 61
/* Run the notify callback FOR, returning ER if it fails */
#define CALLBACK_NOTIFY_(FOR, ER)                                    \
62
do {                                                                 \
63 64
  assert(HTTP_PARSER_ERRNO(parser) == HPE_OK);                       \
                                                                     \
65 66 67
  if (settings->on_##FOR) {                                          \
    if (0 != settings->on_##FOR(parser)) {                           \
      SET_ERRNO(HPE_CB_##FOR);                                       \
68 69 70 71 72
    }                                                                \
                                                                     \
    /* We either errored above or got paused; get out */             \
    if (HTTP_PARSER_ERRNO(parser) != HPE_OK) {                       \
      return (ER);                                                   \
73 74 75 76
    }                                                                \
  }                                                                  \
} while (0)

77 78
/* Run the notify callback FOR and consume the current byte */
#define CALLBACK_NOTIFY(FOR)            CALLBACK_NOTIFY_(FOR, p - data + 1)
79

80 81
/* Run the notify callback FOR and don't consume the current byte */
#define CALLBACK_NOTIFY_NOADVANCE(FOR)  CALLBACK_NOTIFY_(FOR, p - data)
82

83 84
/* Run data callback FOR with LEN bytes, returning ER if it fails */
#define CALLBACK_DATA_(FOR, LEN, ER)                                 \
85
do {                                                                 \
86 87
  assert(HTTP_PARSER_ERRNO(parser) == HPE_OK);                       \
                                                                     \
88 89
  if (FOR##_mark) {                                                  \
    if (settings->on_##FOR) {                                        \
90
      if (0 != settings->on_##FOR(parser, FOR##_mark, (LEN))) {      \
91
        SET_ERRNO(HPE_CB_##FOR);                                     \
92 93 94 95 96
      }                                                              \
                                                                     \
      /* We either errored above or got paused; get out */           \
      if (HTTP_PARSER_ERRNO(parser) != HPE_OK) {                     \
        return (ER);                                                 \
97 98 99 100 101
      }                                                              \
    }                                                                \
    FOR##_mark = NULL;                                               \
  }                                                                  \
} while (0)
102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117
  
/* Run the data callback FOR and consume the current byte */
#define CALLBACK_DATA(FOR)                                           \
    CALLBACK_DATA_(FOR, p - FOR##_mark, p - data + 1)

/* Run the data callback FOR and don't consume the current byte */
#define CALLBACK_DATA_NOADVANCE(FOR)                                 \
    CALLBACK_DATA_(FOR, p - FOR##_mark, p - data)

/* Set the mark FOR; non-destructive if mark is already set */
#define MARK(FOR)                                                    \
do {                                                                 \
  if (!FOR##_mark) {                                                 \
    FOR##_mark = p;                                                  \
  }                                                                  \
} while (0)
118 119 120 121 122 123 124 125 126 127 128 129 130


#define PROXY_CONNECTION "proxy-connection"
#define CONNECTION "connection"
#define CONTENT_LENGTH "content-length"
#define TRANSFER_ENCODING "transfer-encoding"
#define UPGRADE "upgrade"
#define CHUNKED "chunked"
#define KEEP_ALIVE "keep-alive"
#define CLOSE "close"


static const char *method_strings[] =
131 132 133 134
  {
#define XX(num, name, string) #string,
  HTTP_METHOD_MAP(XX)
#undef XX
135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154
  };


/* Tokens as defined by rfc 2616. Also lowercases them.
 *        token       = 1*<any CHAR except CTLs or separators>
 *     separators     = "(" | ")" | "<" | ">" | "@"
 *                    | "," | ";" | ":" | "\" | <">
 *                    | "/" | "[" | "]" | "?" | "="
 *                    | "{" | "}" | SP | HT
 */
static const char tokens[256] = {
/*   0 nul    1 soh    2 stx    3 etx    4 eot    5 enq    6 ack    7 bel  */
        0,       0,       0,       0,       0,       0,       0,       0,
/*   8 bs     9 ht    10 nl    11 vt    12 np    13 cr    14 so    15 si   */
        0,       0,       0,       0,       0,       0,       0,       0,
/*  16 dle   17 dc1   18 dc2   19 dc3   20 dc4   21 nak   22 syn   23 etb */
        0,       0,       0,       0,       0,       0,       0,       0,
/*  24 can   25 em    26 sub   27 esc   28 fs    29 gs    30 rs    31 us  */
        0,       0,       0,       0,       0,       0,       0,       0,
/*  32 sp    33  !    34  "    35  #    36  $    37  %    38  &    39  '  */
155
        0,      '!',      0,      '#',     '$',     '%',     '&',    '\'',
156
/*  40  (    41  )    42  *    43  +    44  ,    45  -    46  .    47  /  */
157
        0,       0,      '*',     '+',      0,      '-',     '.',      0,
158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176
/*  48  0    49  1    50  2    51  3    52  4    53  5    54  6    55  7  */
       '0',     '1',     '2',     '3',     '4',     '5',     '6',     '7',
/*  56  8    57  9    58  :    59  ;    60  <    61  =    62  >    63  ?  */
       '8',     '9',      0,       0,       0,       0,       0,       0,
/*  64  @    65  A    66  B    67  C    68  D    69  E    70  F    71  G  */
        0,      'a',     'b',     'c',     'd',     'e',     'f',     'g',
/*  72  H    73  I    74  J    75  K    76  L    77  M    78  N    79  O  */
       'h',     'i',     'j',     'k',     'l',     'm',     'n',     'o',
/*  80  P    81  Q    82  R    83  S    84  T    85  U    86  V    87  W  */
       'p',     'q',     'r',     's',     't',     'u',     'v',     'w',
/*  88  X    89  Y    90  Z    91  [    92  \    93  ]    94  ^    95  _  */
       'x',     'y',     'z',      0,       0,       0,      '^',     '_',
/*  96  `    97  a    98  b    99  c   100  d   101  e   102  f   103  g  */
       '`',     'a',     'b',     'c',     'd',     'e',     'f',     'g',
/* 104  h   105  i   106  j   107  k   108  l   109  m   110  n   111  o  */
       'h',     'i',     'j',     'k',     'l',     'm',     'n',     'o',
/* 112  p   113  q   114  r   115  s   116  t   117  u   118  v   119  w  */
       'p',     'q',     'r',     's',     't',     'u',     'v',     'w',
/* 120  x   121  y   122  z   123  {   124  |   125  }   126  ~   127 del */
177
       'x',     'y',     'z',      0,      '|',      0,      '~',       0 };
178 179 180 181 182 183 184 185 186 187 188 189 190 191


static const int8_t unhex[256] =
  {-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1
  ,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1
  ,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1
  , 0, 1, 2, 3, 4, 5, 6, 7, 8, 9,-1,-1,-1,-1,-1,-1
  ,-1,10,11,12,13,14,15,-1,-1,-1,-1,-1,-1,-1,-1,-1
  ,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1
  ,-1,10,11,12,13,14,15,-1,-1,-1,-1,-1,-1,-1,-1,-1
  ,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1
  };


192 193 194 195 196 197 198 199
#if HTTP_PARSER_STRICT
# define T(v) 0
#else
# define T(v) v
#endif


static const uint8_t normal_url_char[32] = {
200
/*   0 nul    1 soh    2 stx    3 etx    4 eot    5 enq    6 ack    7 bel  */
201
        0    |   0    |   0    |   0    |   0    |   0    |   0    |   0,
202
/*   8 bs     9 ht    10 nl    11 vt    12 np    13 cr    14 so    15 si   */
203
        0    | T(2)   |   0    |   0    | T(16)  |   0    |   0    |   0,
204
/*  16 dle   17 dc1   18 dc2   19 dc3   20 dc4   21 nak   22 syn   23 etb */
205
        0    |   0    |   0    |   0    |   0    |   0    |   0    |   0,
206
/*  24 can   25 em    26 sub   27 esc   28 fs    29 gs    30 rs    31 us  */
207
        0    |   0    |   0    |   0    |   0    |   0    |   0    |   0,
208
/*  32 sp    33  !    34  "    35  #    36  $    37  %    38  &    39  '  */
209
        0    |   2    |   4    |   0    |   16   |   32   |   64   |  128,
210
/*  40  (    41  )    42  *    43  +    44  ,    45  -    46  .    47  /  */
211
        1    |   2    |   4    |   8    |   16   |   32   |   64   |  128,
212
/*  48  0    49  1    50  2    51  3    52  4    53  5    54  6    55  7  */
213
        1    |   2    |   4    |   8    |   16   |   32   |   64   |  128,
214
/*  56  8    57  9    58  :    59  ;    60  <    61  =    62  >    63  ?  */
215
        1    |   2    |   4    |   8    |   16   |   32   |   64   |   0,
216
/*  64  @    65  A    66  B    67  C    68  D    69  E    70  F    71  G  */
217
        1    |   2    |   4    |   8    |   16   |   32   |   64   |  128,
218
/*  72  H    73  I    74  J    75  K    76  L    77  M    78  N    79  O  */
219
        1    |   2    |   4    |   8    |   16   |   32   |   64   |  128,
220
/*  80  P    81  Q    82  R    83  S    84  T    85  U    86  V    87  W  */
221
        1    |   2    |   4    |   8    |   16   |   32   |   64   |  128,
222
/*  88  X    89  Y    90  Z    91  [    92  \    93  ]    94  ^    95  _  */
223
        1    |   2    |   4    |   8    |   16   |   32   |   64   |  128,
224
/*  96  `    97  a    98  b    99  c   100  d   101  e   102  f   103  g  */
225
        1    |   2    |   4    |   8    |   16   |   32   |   64   |  128,
226
/* 104  h   105  i   106  j   107  k   108  l   109  m   110  n   111  o  */
227
        1    |   2    |   4    |   8    |   16   |   32   |   64   |  128,
228
/* 112  p   113  q   114  r   115  s   116  t   117  u   118  v   119  w  */
229
        1    |   2    |   4    |   8    |   16   |   32   |   64   |  128,
230
/* 120  x   121  y   122  z   123  {   124  |   125  }   126  ~   127 del */
231
        1    |   2    |   4    |   8    |   16   |   32   |   64   |   0, };
232

233
#undef T
234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260

enum state
  { s_dead = 1 /* important that this is > 0 */

  , s_start_req_or_res
  , s_res_or_resp_H
  , s_start_res
  , s_res_H
  , s_res_HT
  , s_res_HTT
  , s_res_HTTP
  , s_res_first_http_major
  , s_res_http_major
  , s_res_first_http_minor
  , s_res_http_minor
  , s_res_first_status_code
  , s_res_status_code
  , s_res_status
  , s_res_line_almost_done

  , s_start_req

  , s_req_method
  , s_req_spaces_before_url
  , s_req_schema
  , s_req_schema_slash
  , s_req_schema_slash_slash
261 262 263
  , s_req_server_start
  , s_req_server
  , s_req_server_with_at
264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291
  , s_req_path
  , s_req_query_string_start
  , s_req_query_string
  , s_req_fragment_start
  , s_req_fragment
  , s_req_http_start
  , s_req_http_H
  , s_req_http_HT
  , s_req_http_HTT
  , s_req_http_HTTP
  , s_req_first_http_major
  , s_req_http_major
  , s_req_first_http_minor
  , s_req_http_minor
  , s_req_line_almost_done

  , s_header_field_start
  , s_header_field
  , s_header_value_start
  , s_header_value
  , s_header_value_lws

  , s_header_almost_done

  , s_chunk_size_start
  , s_chunk_size
  , s_chunk_parameters
  , s_chunk_size_almost_done
292

293
  , s_headers_almost_done
294 295 296
  , s_headers_done

  /* Important: 's_headers_done' must be the last 'header' state. All
297 298 299 300 301 302 303 304 305 306
   * states beyond this must be 'body' states. It is used for overflow
   * checking. See the PARSING_HEADER() macro.
   */

  , s_chunk_data
  , s_chunk_data_almost_done
  , s_chunk_data_done

  , s_body_identity
  , s_body_identity_eof
307 308

  , s_message_done
309 310 311
  };


312
#define PARSING_HEADER(state) (state <= s_headers_done)
313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340


enum header_states
  { h_general = 0
  , h_C
  , h_CO
  , h_CON

  , h_matching_connection
  , h_matching_proxy_connection
  , h_matching_content_length
  , h_matching_transfer_encoding
  , h_matching_upgrade

  , h_connection
  , h_content_length
  , h_transfer_encoding
  , h_upgrade

  , h_matching_transfer_encoding_chunked
  , h_matching_connection_keep_alive
  , h_matching_connection_close

  , h_transfer_encoding_chunked
  , h_connection_keep_alive
  , h_connection_close
  };

341 342 343 344 345 346 347 348 349 350 351 352 353
enum http_host_state
  {
    s_http_host_dead = 1
  , s_http_userinfo_start
  , s_http_userinfo
  , s_http_host_start
  , s_http_host_v6_start
  , s_http_host
  , s_http_host_v6
  , s_http_host_v6_end
  , s_http_host_port_start
  , s_http_host_port
};
354 355 356 357 358 359 360 361

/* Macros for character classes; depends on strict-mode  */
#define CR                  '\r'
#define LF                  '\n'
#define LOWER(c)            (unsigned char)(c | 0x20)
#define IS_ALPHA(c)         (LOWER(c) >= 'a' && LOWER(c) <= 'z')
#define IS_NUM(c)           ((c) >= '0' && (c) <= '9')
#define IS_ALPHANUM(c)      (IS_ALPHA(c) || IS_NUM(c))
362 363 364 365 366 367 368
#define IS_HEX(c)           (IS_NUM(c) || (LOWER(c) >= 'a' && LOWER(c) <= 'f'))
#define IS_MARK(c)          ((c) == '-' || (c) == '_' || (c) == '.' || \
  (c) == '!' || (c) == '~' || (c) == '*' || (c) == '\'' || (c) == '(' || \
  (c) == ')')
#define IS_USERINFO_CHAR(c) (IS_ALPHANUM(c) || IS_MARK(c) || (c) == '%' || \
  (c) == ';' || (c) == ':' || (c) == '&' || (c) == '=' || (c) == '+' || \
  (c) == '$' || (c) == ',')
369 370

#if HTTP_PARSER_STRICT
371 372
#define TOKEN(c)            (tokens[(unsigned char)c])
#define IS_URL_CHAR(c)      (BIT_AT(normal_url_char, (unsigned char)c))
373 374
#define IS_HOST_CHAR(c)     (IS_ALPHANUM(c) || (c) == '.' || (c) == '-')
#else
375
#define TOKEN(c)            ((c == ' ') ? ' ' : tokens[(unsigned char)c])
376
#define IS_URL_CHAR(c)                                                         \
377
  (BIT_AT(normal_url_char, (unsigned char)c) || ((c) & 0x80))
378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410
#define IS_HOST_CHAR(c)                                                        \
  (IS_ALPHANUM(c) || (c) == '.' || (c) == '-' || (c) == '_')
#endif


#define start_state (parser->type == HTTP_REQUEST ? s_start_req : s_start_res)


#if HTTP_PARSER_STRICT
# define STRICT_CHECK(cond)                                          \
do {                                                                 \
  if (cond) {                                                        \
    SET_ERRNO(HPE_STRICT);                                           \
    goto error;                                                      \
  }                                                                  \
} while (0)
# define NEW_MESSAGE() (http_should_keep_alive(parser) ? start_state : s_dead)
#else
# define STRICT_CHECK(cond)
# define NEW_MESSAGE() start_state
#endif


/* Map errno values to strings for human-readable output */
#define HTTP_STRERROR_GEN(n, s) { "HPE_" #n, s },
static struct {
  const char *name;
  const char *description;
} http_strerror_tab[] = {
  HTTP_ERRNO_MAP(HTTP_STRERROR_GEN)
};
#undef HTTP_STRERROR_GEN

411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570
int http_message_needs_eof(const http_parser *parser);

/* Our URL parser.
 *
 * This is designed to be shared by http_parser_execute() for URL validation,
 * hence it has a state transition + byte-for-byte interface. In addition, it
 * is meant to be embedded in http_parser_parse_url(), which does the dirty
 * work of turning state transitions URL components for its API.
 *
 * This function should only be invoked with non-space characters. It is
 * assumed that the caller cares about (and can detect) the transition between
 * URL and non-URL states by looking for these.
 */
static enum state
parse_url_char(enum state s, const char ch)
{
  if (ch == ' ' || ch == '\r' || ch == '\n') {
    return s_dead;
  }

#if HTTP_PARSER_STRICT
  if (ch == '\t' || ch == '\f') {
    return s_dead;
  }
#endif

  switch (s) {
    case s_req_spaces_before_url:
      /* Proxied requests are followed by scheme of an absolute URI (alpha).
       * All methods except CONNECT are followed by '/' or '*'.
       */

      if (ch == '/' || ch == '*') {
        return s_req_path;
      }

      if (IS_ALPHA(ch)) {
        return s_req_schema;
      }

      break;

    case s_req_schema:
      if (IS_ALPHA(ch)) {
        return s;
      }

      if (ch == ':') {
        return s_req_schema_slash;
      }

      break;

    case s_req_schema_slash:
      if (ch == '/') {
        return s_req_schema_slash_slash;
      }

      break;

    case s_req_schema_slash_slash:
      if (ch == '/') {
        return s_req_server_start;
      }

      break;

    case s_req_server_with_at:
      if (ch == '@') {
        return s_dead;
      }

    /* FALLTHROUGH */
    case s_req_server_start:
    case s_req_server:
      if (ch == '/') {
        return s_req_path;
      }

      if (ch == '?') {
        return s_req_query_string_start;
      }

      if (ch == '@') {
        return s_req_server_with_at;
      }

      if (IS_USERINFO_CHAR(ch) || ch == '[' || ch == ']') {
        return s_req_server;
      }

      break;

    case s_req_path:
      if (IS_URL_CHAR(ch)) {
        return s;
      }

      switch (ch) {
        case '?':
          return s_req_query_string_start;

        case '#':
          return s_req_fragment_start;
      }

      break;

    case s_req_query_string_start:
    case s_req_query_string:
      if (IS_URL_CHAR(ch)) {
        return s_req_query_string;
      }

      switch (ch) {
        case '?':
          /* allow extra '?' in query string */
          return s_req_query_string;

        case '#':
          return s_req_fragment_start;
      }

      break;

    case s_req_fragment_start:
      if (IS_URL_CHAR(ch)) {
        return s_req_fragment;
      }

      switch (ch) {
        case '?':
          return s_req_fragment;

        case '#':
          return s;
      }

      break;

    case s_req_fragment:
      if (IS_URL_CHAR(ch)) {
        return s;
      }

      switch (ch) {
        case '?':
        case '#':
          return s;
      }

      break;

    default:
      break;
  }

  /* We should never fall out of the switch above unless there's an error */
  return s_dead;
}
571 572 573 574 575 576 577 578

size_t http_parser_execute (http_parser *parser,
                            const http_parser_settings *settings,
                            const char *data,
                            size_t len)
{
  char c, ch;
  int8_t unhex_val;
579 580 581 582 583
  const char *p = data;
  const char *header_field_mark = 0;
  const char *header_value_mark = 0;
  const char *url_mark = 0;
  const char *body_mark = 0;
584 585 586 587 588 589 590

  /* We're in an error state. Don't bother doing anything. */
  if (HTTP_PARSER_ERRNO(parser) != HPE_OK) {
    return 0;
  }

  if (len == 0) {
591
    switch (parser->state) {
592
      case s_body_identity_eof:
593 594 595 596
        /* Use of CALLBACK_NOTIFY() here would erroneously return 1 byte read if
         * we got paused.
         */
        CALLBACK_NOTIFY_NOADVANCE(message_complete);
597 598 599 600 601 602 603 604 605 606 607 608 609 610 611
        return 0;

      case s_dead:
      case s_start_req_or_res:
      case s_start_res:
      case s_start_req:
        return 0;

      default:
        SET_ERRNO(HPE_INVALID_EOF_STATE);
        return 1;
    }
  }


612
  if (parser->state == s_header_field)
613
    header_field_mark = data;
614
  if (parser->state == s_header_value)
615
    header_value_mark = data;
616 617 618 619 620 621 622 623 624 625 626 627
  switch (parser->state) {
  case s_req_path:
  case s_req_schema:
  case s_req_schema_slash:
  case s_req_schema_slash_slash:
  case s_req_server_start:
  case s_req_server:
  case s_req_server_with_at:
  case s_req_query_string_start:
  case s_req_query_string:
  case s_req_fragment_start:
  case s_req_fragment:
628
    url_mark = data;
629 630
    break;
  }
631

632
  for (p=data; p != data + len; p++) {
633 634
    ch = *p;

635 636
    if (PARSING_HEADER(parser->state)) {
      ++parser->nread;
637
      /* Buffer overflow attack */
638
      if (parser->nread > HTTP_MAX_HEADER_SIZE) {
639 640 641 642 643
        SET_ERRNO(HPE_HEADER_OVERFLOW);
        goto error;
      }
    }

644 645
    reexecute_byte:
    switch (parser->state) {
646 647 648 649 650

      case s_dead:
        /* this state is used after a 'Connection: close' message
         * the parser will error out if it reads another message
         */
651 652 653
        if (ch == CR || ch == LF)
          break;

654 655 656 657 658 659 660 661
        SET_ERRNO(HPE_CLOSED_CONNECTION);
        goto error;

      case s_start_req_or_res:
      {
        if (ch == CR || ch == LF)
          break;
        parser->flags = 0;
662
        parser->content_length = ULLONG_MAX;
663

664 665
        if (ch == 'H') {
          parser->state = s_res_or_resp_H;
666

667 668
          CALLBACK_NOTIFY(message_begin);
        } else {
669
          parser->type = HTTP_REQUEST;
670 671
          parser->state = s_start_req;
          goto reexecute_byte;
672
        }
673

674 675 676 677 678 679
        break;
      }

      case s_res_or_resp_H:
        if (ch == 'T') {
          parser->type = HTTP_RESPONSE;
680
          parser->state = s_res_HT;
681 682 683 684 685 686 687 688
        } else {
          if (ch != 'E') {
            SET_ERRNO(HPE_INVALID_CONSTANT);
            goto error;
          }

          parser->type = HTTP_REQUEST;
          parser->method = HTTP_HEAD;
689 690
          parser->index = 2;
          parser->state = s_req_method;
691 692 693 694 695 696
        }
        break;

      case s_start_res:
      {
        parser->flags = 0;
697
        parser->content_length = ULLONG_MAX;
698 699 700

        switch (ch) {
          case 'H':
701
            parser->state = s_res_H;
702 703 704 705 706 707 708 709 710 711
            break;

          case CR:
          case LF:
            break;

          default:
            SET_ERRNO(HPE_INVALID_CONSTANT);
            goto error;
        }
712 713

        CALLBACK_NOTIFY(message_begin);
714 715 716 717 718
        break;
      }

      case s_res_H:
        STRICT_CHECK(ch != 'T');
719
        parser->state = s_res_HT;
720 721 722 723
        break;

      case s_res_HT:
        STRICT_CHECK(ch != 'T');
724
        parser->state = s_res_HTT;
725 726 727 728
        break;

      case s_res_HTT:
        STRICT_CHECK(ch != 'P');
729
        parser->state = s_res_HTTP;
730 731 732 733
        break;

      case s_res_HTTP:
        STRICT_CHECK(ch != '/');
734
        parser->state = s_res_first_http_major;
735 736 737
        break;

      case s_res_first_http_major:
738
        if (ch < '0' || ch > '9') {
739 740 741 742 743
          SET_ERRNO(HPE_INVALID_VERSION);
          goto error;
        }

        parser->http_major = ch - '0';
744
        parser->state = s_res_http_major;
745 746 747 748 749 750
        break;

      /* major HTTP version or dot */
      case s_res_http_major:
      {
        if (ch == '.') {
751
          parser->state = s_res_first_http_minor;
752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778
          break;
        }

        if (!IS_NUM(ch)) {
          SET_ERRNO(HPE_INVALID_VERSION);
          goto error;
        }

        parser->http_major *= 10;
        parser->http_major += ch - '0';

        if (parser->http_major > 999) {
          SET_ERRNO(HPE_INVALID_VERSION);
          goto error;
        }

        break;
      }

      /* first digit of minor HTTP version */
      case s_res_first_http_minor:
        if (!IS_NUM(ch)) {
          SET_ERRNO(HPE_INVALID_VERSION);
          goto error;
        }

        parser->http_minor = ch - '0';
779
        parser->state = s_res_http_minor;
780 781 782 783 784 785
        break;

      /* minor HTTP version or end of request line */
      case s_res_http_minor:
      {
        if (ch == ' ') {
786
          parser->state = s_res_first_status_code;
787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816
          break;
        }

        if (!IS_NUM(ch)) {
          SET_ERRNO(HPE_INVALID_VERSION);
          goto error;
        }

        parser->http_minor *= 10;
        parser->http_minor += ch - '0';

        if (parser->http_minor > 999) {
          SET_ERRNO(HPE_INVALID_VERSION);
          goto error;
        }

        break;
      }

      case s_res_first_status_code:
      {
        if (!IS_NUM(ch)) {
          if (ch == ' ') {
            break;
          }

          SET_ERRNO(HPE_INVALID_STATUS);
          goto error;
        }
        parser->status_code = ch - '0';
817
        parser->state = s_res_status_code;
818 819 820 821 822 823 824 825
        break;
      }

      case s_res_status_code:
      {
        if (!IS_NUM(ch)) {
          switch (ch) {
            case ' ':
826
              parser->state = s_res_status;
827 828
              break;
            case CR:
829
              parser->state = s_res_line_almost_done;
830 831
              break;
            case LF:
832
              parser->state = s_header_field_start;
833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855
              break;
            default:
              SET_ERRNO(HPE_INVALID_STATUS);
              goto error;
          }
          break;
        }

        parser->status_code *= 10;
        parser->status_code += ch - '0';

        if (parser->status_code > 999) {
          SET_ERRNO(HPE_INVALID_STATUS);
          goto error;
        }

        break;
      }

      case s_res_status:
        /* the human readable status. e.g. "NOT FOUND"
         * we are not humans so just ignore this */
        if (ch == CR) {
856
          parser->state = s_res_line_almost_done;
857 858 859 860
          break;
        }

        if (ch == LF) {
861
          parser->state = s_header_field_start;
862 863 864 865 866 867
          break;
        }
        break;

      case s_res_line_almost_done:
        STRICT_CHECK(ch != LF);
868
        parser->state = s_header_field_start;
869 870 871 872 873 874 875
        break;

      case s_start_req:
      {
        if (ch == CR || ch == LF)
          break;
        parser->flags = 0;
876
        parser->content_length = ULLONG_MAX;
877 878 879 880 881 882 883

        if (!IS_ALPHA(ch)) {
          SET_ERRNO(HPE_INVALID_METHOD);
          goto error;
        }

        parser->method = (enum http_method) 0;
884
        parser->index = 1;
885 886 887 888 889 890 891 892 893 894
        switch (ch) {
          case 'C': parser->method = HTTP_CONNECT; /* or COPY, CHECKOUT */ break;
          case 'D': parser->method = HTTP_DELETE; break;
          case 'G': parser->method = HTTP_GET; break;
          case 'H': parser->method = HTTP_HEAD; break;
          case 'L': parser->method = HTTP_LOCK; break;
          case 'M': parser->method = HTTP_MKCOL; /* or MOVE, MKACTIVITY, MERGE, M-SEARCH */ break;
          case 'N': parser->method = HTTP_NOTIFY; break;
          case 'O': parser->method = HTTP_OPTIONS; break;
          case 'P': parser->method = HTTP_POST;
895
            /* or PROPFIND|PROPPATCH|PUT|PATCH|PURGE */
896 897
            break;
          case 'R': parser->method = HTTP_REPORT; break;
898
          case 'S': parser->method = HTTP_SUBSCRIBE; /* or SEARCH */ break;
899 900 901 902 903 904
          case 'T': parser->method = HTTP_TRACE; break;
          case 'U': parser->method = HTTP_UNLOCK; /* or UNSUBSCRIBE */ break;
          default:
            SET_ERRNO(HPE_INVALID_METHOD);
            goto error;
        }
905 906 907 908
        parser->state = s_req_method;

        CALLBACK_NOTIFY(message_begin);

909 910 911 912 913
        break;
      }

      case s_req_method:
      {
914
        const char *matcher;
915 916 917 918 919
        if (ch == '\0') {
          SET_ERRNO(HPE_INVALID_METHOD);
          goto error;
        }

920
        matcher = method_strings[parser->method];
921 922 923
        if (ch == ' ' && matcher[parser->index] == '\0') {
          parser->state = s_req_spaces_before_url;
        } else if (ch == matcher[parser->index]) {
924 925
          ; /* nada */
        } else if (parser->method == HTTP_CONNECT) {
926
          if (parser->index == 1 && ch == 'H') {
927
            parser->method = HTTP_CHECKOUT;
928
          } else if (parser->index == 2  && ch == 'P') {
929 930 931 932 933
            parser->method = HTTP_COPY;
          } else {
            goto error;
          }
        } else if (parser->method == HTTP_MKCOL) {
934
          if (parser->index == 1 && ch == 'O') {
935
            parser->method = HTTP_MOVE;
936
          } else if (parser->index == 1 && ch == 'E') {
937
            parser->method = HTTP_MERGE;
938
          } else if (parser->index == 1 && ch == '-') {
939
            parser->method = HTTP_MSEARCH;
940
          } else if (parser->index == 2 && ch == 'A') {
941 942 943 944
            parser->method = HTTP_MKACTIVITY;
          } else {
            goto error;
          }
945 946 947 948 949 950 951
        } else if (parser->method == HTTP_SUBSCRIBE) {
          if (parser->index == 1 && ch == 'E') {
            parser->method = HTTP_SEARCH;
          } else {
            goto error;
          }
        } else if (parser->index == 1 && parser->method == HTTP_POST) {
952 953 954
          if (ch == 'R') {
            parser->method = HTTP_PROPFIND; /* or HTTP_PROPPATCH */
          } else if (ch == 'U') {
955
            parser->method = HTTP_PUT; /* or HTTP_PURGE */
956 957 958 959 960
          } else if (ch == 'A') {
            parser->method = HTTP_PATCH;
          } else {
            goto error;
          }
961 962 963 964 965 966 967
        } else if (parser->index == 2) {
          if (parser->method == HTTP_PUT) {
            if (ch == 'R') parser->method = HTTP_PURGE;
          } else if (parser->method == HTTP_UNLOCK) {
            if (ch == 'S') parser->method = HTTP_UNSUBSCRIBE;
          }
        } else if (parser->index == 4 && parser->method == HTTP_PROPFIND && ch == 'P') {
968 969 970 971 972 973
          parser->method = HTTP_PROPPATCH;
        } else {
          SET_ERRNO(HPE_INVALID_METHOD);
          goto error;
        }

974
        ++parser->index;
975 976
        break;
      }
977

978 979 980 981
      case s_req_spaces_before_url:
      {
        if (ch == ' ') break;

982 983 984
        MARK(url);
        if (parser->method == HTTP_CONNECT) {
          parser->state = s_req_server_start;
985 986
        }

987 988 989 990
        parser->state = parse_url_char((enum state)parser->state, ch);
        if (parser->state == s_dead) {
          SET_ERRNO(HPE_INVALID_URL);
          goto error;
991 992
        }

993
        break;
994 995 996 997 998
      }

      case s_req_schema:
      case s_req_schema_slash:
      case s_req_schema_slash_slash:
999
      case s_req_server_start:
1000 1001
      {
        switch (ch) {
1002
          /* No whitespace allowed here */
1003 1004 1005
          case ' ':
          case CR:
          case LF:
1006
            SET_ERRNO(HPE_INVALID_URL);
1007 1008
            goto error;
          default:
1009 1010 1011 1012 1013
            parser->state = parse_url_char((enum state)parser->state, ch);
            if (parser->state == s_dead) {
              SET_ERRNO(HPE_INVALID_URL);
              goto error;
            }
1014 1015 1016 1017 1018
        }

        break;
      }

1019 1020 1021 1022 1023
      case s_req_server:
      case s_req_server_with_at:
      case s_req_path:
      case s_req_query_string_start:
      case s_req_query_string:
1024 1025 1026 1027 1028
      case s_req_fragment_start:
      case s_req_fragment:
      {
        switch (ch) {
          case ' ':
1029 1030
            parser->state = s_req_http_start;
            CALLBACK_DATA(url);
1031 1032 1033 1034 1035
            break;
          case CR:
          case LF:
            parser->http_major = 0;
            parser->http_minor = 9;
1036 1037 1038 1039
            parser->state = (ch == CR) ?
              s_req_line_almost_done :
              s_header_field_start;
            CALLBACK_DATA(url);
1040 1041
            break;
          default:
1042 1043 1044 1045 1046
            parser->state = parse_url_char((enum state)parser->state, ch);
            if (parser->state == s_dead) {
              SET_ERRNO(HPE_INVALID_URL);
              goto error;
            }
1047 1048 1049 1050 1051 1052 1053
        }
        break;
      }

      case s_req_http_start:
        switch (ch) {
          case 'H':
1054
            parser->state = s_req_http_H;
1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065
            break;
          case ' ':
            break;
          default:
            SET_ERRNO(HPE_INVALID_CONSTANT);
            goto error;
        }
        break;

      case s_req_http_H:
        STRICT_CHECK(ch != 'T');
1066
        parser->state = s_req_http_HT;
1067 1068 1069 1070
        break;

      case s_req_http_HT:
        STRICT_CHECK(ch != 'T');
1071
        parser->state = s_req_http_HTT;
1072 1073 1074 1075
        break;

      case s_req_http_HTT:
        STRICT_CHECK(ch != 'P');
1076
        parser->state = s_req_http_HTTP;
1077 1078 1079 1080
        break;

      case s_req_http_HTTP:
        STRICT_CHECK(ch != '/');
1081
        parser->state = s_req_first_http_major;
1082 1083 1084 1085 1086 1087 1088 1089 1090 1091
        break;

      /* first digit of major HTTP version */
      case s_req_first_http_major:
        if (ch < '1' || ch > '9') {
          SET_ERRNO(HPE_INVALID_VERSION);
          goto error;
        }

        parser->http_major = ch - '0';
1092
        parser->state = s_req_http_major;
1093 1094 1095 1096 1097 1098
        break;

      /* major HTTP version or dot */
      case s_req_http_major:
      {
        if (ch == '.') {
1099
          parser->state = s_req_first_http_minor;
1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126
          break;
        }

        if (!IS_NUM(ch)) {
          SET_ERRNO(HPE_INVALID_VERSION);
          goto error;
        }

        parser->http_major *= 10;
        parser->http_major += ch - '0';

        if (parser->http_major > 999) {
          SET_ERRNO(HPE_INVALID_VERSION);
          goto error;
        }

        break;
      }

      /* first digit of minor HTTP version */
      case s_req_first_http_minor:
        if (!IS_NUM(ch)) {
          SET_ERRNO(HPE_INVALID_VERSION);
          goto error;
        }

        parser->http_minor = ch - '0';
1127
        parser->state = s_req_http_minor;
1128 1129 1130 1131 1132 1133
        break;

      /* minor HTTP version or end of request line */
      case s_req_http_minor:
      {
        if (ch == CR) {
1134
          parser->state = s_req_line_almost_done;
1135 1136 1137 1138
          break;
        }

        if (ch == LF) {
1139
          parser->state = s_header_field_start;
1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168
          break;
        }

        /* XXX allow spaces after digit? */

        if (!IS_NUM(ch)) {
          SET_ERRNO(HPE_INVALID_VERSION);
          goto error;
        }

        parser->http_minor *= 10;
        parser->http_minor += ch - '0';

        if (parser->http_minor > 999) {
          SET_ERRNO(HPE_INVALID_VERSION);
          goto error;
        }

        break;
      }

      /* end of request line */
      case s_req_line_almost_done:
      {
        if (ch != LF) {
          SET_ERRNO(HPE_LF_EXPECTED);
          goto error;
        }

1169
        parser->state = s_header_field_start;
1170 1171 1172 1173 1174 1175
        break;
      }

      case s_header_field_start:
      {
        if (ch == CR) {
1176
          parser->state = s_headers_almost_done;
1177 1178 1179 1180 1181 1182
          break;
        }

        if (ch == LF) {
          /* they might be just sending \n instead of \r\n so this would be
           * the second \n to denote the end of headers*/
1183 1184
          parser->state = s_headers_almost_done;
          goto reexecute_byte;
1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195
        }

        c = TOKEN(ch);

        if (!c) {
          SET_ERRNO(HPE_INVALID_HEADER_TOKEN);
          goto error;
        }

        MARK(header_field);

1196 1197
        parser->index = 0;
        parser->state = s_header_field;
1198 1199 1200

        switch (c) {
          case 'c':
1201
            parser->header_state = h_C;
1202 1203 1204
            break;

          case 'p':
1205
            parser->header_state = h_matching_proxy_connection;
1206 1207 1208
            break;

          case 't':
1209
            parser->header_state = h_matching_transfer_encoding;
1210 1211 1212
            break;

          case 'u':
1213
            parser->header_state = h_matching_upgrade;
1214 1215 1216
            break;

          default:
1217
            parser->header_state = h_general;
1218 1219 1220 1221 1222 1223 1224 1225 1226 1227
            break;
        }
        break;
      }

      case s_header_field:
      {
        c = TOKEN(ch);

        if (c) {
1228
          switch (parser->header_state) {
1229 1230 1231 1232
            case h_general:
              break;

            case h_C:
1233 1234
              parser->index++;
              parser->header_state = (c == 'o' ? h_CO : h_general);
1235 1236 1237
              break;

            case h_CO:
1238 1239
              parser->index++;
              parser->header_state = (c == 'n' ? h_CON : h_general);
1240 1241 1242
              break;

            case h_CON:
1243
              parser->index++;
1244 1245
              switch (c) {
                case 'n':
1246
                  parser->header_state = h_matching_connection;
1247 1248
                  break;
                case 't':
1249
                  parser->header_state = h_matching_content_length;
1250 1251
                  break;
                default:
1252
                  parser->header_state = h_general;
1253 1254 1255 1256 1257 1258 1259
                  break;
              }
              break;

            /* connection */

            case h_matching_connection:
1260 1261 1262 1263 1264 1265
              parser->index++;
              if (parser->index > sizeof(CONNECTION)-1
                  || c != CONNECTION[parser->index]) {
                parser->header_state = h_general;
              } else if (parser->index == sizeof(CONNECTION)-2) {
                parser->header_state = h_connection;
1266 1267 1268 1269 1270 1271
              }
              break;

            /* proxy-connection */

            case h_matching_proxy_connection:
1272 1273 1274 1275 1276 1277
              parser->index++;
              if (parser->index > sizeof(PROXY_CONNECTION)-1
                  || c != PROXY_CONNECTION[parser->index]) {
                parser->header_state = h_general;
              } else if (parser->index == sizeof(PROXY_CONNECTION)-2) {
                parser->header_state = h_connection;
1278 1279 1280 1281 1282 1283
              }
              break;

            /* content-length */

            case h_matching_content_length:
1284 1285 1286 1287 1288 1289
              parser->index++;
              if (parser->index > sizeof(CONTENT_LENGTH)-1
                  || c != CONTENT_LENGTH[parser->index]) {
                parser->header_state = h_general;
              } else if (parser->index == sizeof(CONTENT_LENGTH)-2) {
                parser->header_state = h_content_length;
1290 1291 1292 1293 1294 1295
              }
              break;

            /* transfer-encoding */

            case h_matching_transfer_encoding:
1296 1297 1298 1299 1300 1301
              parser->index++;
              if (parser->index > sizeof(TRANSFER_ENCODING)-1
                  || c != TRANSFER_ENCODING[parser->index]) {
                parser->header_state = h_general;
              } else if (parser->index == sizeof(TRANSFER_ENCODING)-2) {
                parser->header_state = h_transfer_encoding;
1302 1303 1304 1305 1306 1307
              }
              break;

            /* upgrade */

            case h_matching_upgrade:
1308 1309 1310 1311 1312 1313
              parser->index++;
              if (parser->index > sizeof(UPGRADE)-1
                  || c != UPGRADE[parser->index]) {
                parser->header_state = h_general;
              } else if (parser->index == sizeof(UPGRADE)-2) {
                parser->header_state = h_upgrade;
1314 1315 1316 1317 1318 1319 1320
              }
              break;

            case h_connection:
            case h_content_length:
            case h_transfer_encoding:
            case h_upgrade:
1321
              if (ch != ' ') parser->header_state = h_general;
1322 1323 1324 1325 1326 1327 1328 1329 1330 1331
              break;

            default:
              assert(0 && "Unknown header_state");
              break;
          }
          break;
        }

        if (ch == ':') {
1332 1333
          parser->state = s_header_value_start;
          CALLBACK_DATA(header_field);
1334 1335 1336 1337
          break;
        }

        if (ch == CR) {
1338 1339
          parser->state = s_header_almost_done;
          CALLBACK_DATA(header_field);
1340 1341 1342 1343
          break;
        }

        if (ch == LF) {
1344 1345
          parser->state = s_header_field_start;
          CALLBACK_DATA(header_field);
1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358
          break;
        }

        SET_ERRNO(HPE_INVALID_HEADER_TOKEN);
        goto error;
      }

      case s_header_value_start:
      {
        if (ch == ' ' || ch == '\t') break;

        MARK(header_value);

1359 1360
        parser->state = s_header_value;
        parser->index = 0;
1361 1362

        if (ch == CR) {
1363 1364 1365
          parser->header_state = h_general;
          parser->state = s_header_almost_done;
          CALLBACK_DATA(header_value);
1366 1367 1368 1369
          break;
        }

        if (ch == LF) {
1370 1371
          parser->state = s_header_field_start;
          CALLBACK_DATA(header_value);
1372 1373 1374 1375 1376
          break;
        }

        c = LOWER(ch);

1377
        switch (parser->header_state) {
1378 1379
          case h_upgrade:
            parser->flags |= F_UPGRADE;
1380
            parser->header_state = h_general;
1381 1382 1383 1384 1385
            break;

          case h_transfer_encoding:
            /* looking for 'Transfer-Encoding: chunked' */
            if ('c' == c) {
1386
              parser->header_state = h_matching_transfer_encoding_chunked;
1387
            } else {
1388
              parser->header_state = h_general;
1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403
            }
            break;

          case h_content_length:
            if (!IS_NUM(ch)) {
              SET_ERRNO(HPE_INVALID_CONTENT_LENGTH);
              goto error;
            }

            parser->content_length = ch - '0';
            break;

          case h_connection:
            /* looking for 'Connection: keep-alive' */
            if (c == 'k') {
1404
              parser->header_state = h_matching_connection_keep_alive;
1405 1406
            /* looking for 'Connection: close' */
            } else if (c == 'c') {
1407
              parser->header_state = h_matching_connection_close;
1408
            } else {
1409
              parser->header_state = h_general;
1410 1411 1412 1413
            }
            break;

          default:
1414
            parser->header_state = h_general;
1415 1416 1417 1418 1419 1420 1421 1422 1423
            break;
        }
        break;
      }

      case s_header_value:
      {

        if (ch == CR) {
1424 1425
          parser->state = s_header_almost_done;
          CALLBACK_DATA(header_value);
1426 1427 1428 1429
          break;
        }

        if (ch == LF) {
1430 1431 1432
          parser->state = s_header_almost_done;
          CALLBACK_DATA_NOADVANCE(header_value);
          goto reexecute_byte;
1433 1434 1435 1436
        }

        c = LOWER(ch);

1437
        switch (parser->header_state) {
1438 1439 1440 1441 1442 1443 1444 1445 1446
          case h_general:
            break;

          case h_connection:
          case h_transfer_encoding:
            assert(0 && "Shouldn't get here.");
            break;

          case h_content_length:
1447 1448 1449
          {
            uint64_t t;

1450
            if (ch == ' ') break;
1451

1452 1453 1454 1455 1456
            if (!IS_NUM(ch)) {
              SET_ERRNO(HPE_INVALID_CONTENT_LENGTH);
              goto error;
            }

1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467
            t = parser->content_length;
            t *= 10;
            t += ch - '0';

            /* Overflow? */
            if (t < parser->content_length || t == ULLONG_MAX) {
              SET_ERRNO(HPE_INVALID_CONTENT_LENGTH);
              goto error;
            }

            parser->content_length = t;
1468
            break;
1469
          }
1470 1471 1472

          /* Transfer-Encoding: chunked */
          case h_matching_transfer_encoding_chunked:
1473 1474 1475 1476 1477 1478
            parser->index++;
            if (parser->index > sizeof(CHUNKED)-1
                || c != CHUNKED[parser->index]) {
              parser->header_state = h_general;
            } else if (parser->index == sizeof(CHUNKED)-2) {
              parser->header_state = h_transfer_encoding_chunked;
1479 1480 1481 1482 1483
            }
            break;

          /* looking for 'Connection: keep-alive' */
          case h_matching_connection_keep_alive:
1484 1485 1486 1487 1488 1489
            parser->index++;
            if (parser->index > sizeof(KEEP_ALIVE)-1
                || c != KEEP_ALIVE[parser->index]) {
              parser->header_state = h_general;
            } else if (parser->index == sizeof(KEEP_ALIVE)-2) {
              parser->header_state = h_connection_keep_alive;
1490 1491 1492 1493 1494
            }
            break;

          /* looking for 'Connection: close' */
          case h_matching_connection_close:
1495 1496 1497 1498 1499
            parser->index++;
            if (parser->index > sizeof(CLOSE)-1 || c != CLOSE[parser->index]) {
              parser->header_state = h_general;
            } else if (parser->index == sizeof(CLOSE)-2) {
              parser->header_state = h_connection_close;
1500 1501 1502 1503 1504 1505
            }
            break;

          case h_transfer_encoding_chunked:
          case h_connection_keep_alive:
          case h_connection_close:
1506
            if (ch != ' ') parser->header_state = h_general;
1507 1508 1509
            break;

          default:
1510 1511
            parser->state = s_header_value;
            parser->header_state = h_general;
1512 1513 1514 1515 1516 1517 1518 1519 1520
            break;
        }
        break;
      }

      case s_header_almost_done:
      {
        STRICT_CHECK(ch != LF);

1521
        parser->state = s_header_value_lws;
1522

1523
        switch (parser->header_state) {
1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535
          case h_connection_keep_alive:
            parser->flags |= F_CONNECTION_KEEP_ALIVE;
            break;
          case h_connection_close:
            parser->flags |= F_CONNECTION_CLOSE;
            break;
          case h_transfer_encoding_chunked:
            parser->flags |= F_CHUNKED;
            break;
          default:
            break;
        }
1536

1537 1538 1539 1540 1541 1542
        break;
      }

      case s_header_value_lws:
      {
        if (ch == ' ' || ch == '\t')
1543
          parser->state = s_header_value_start;
1544 1545
        else
        {
1546 1547
          parser->state = s_header_field_start;
          goto reexecute_byte;
1548 1549 1550 1551 1552 1553 1554 1555 1556 1557
        }
        break;
      }

      case s_headers_almost_done:
      {
        STRICT_CHECK(ch != LF);

        if (parser->flags & F_TRAILING) {
          /* End of a chunked request */
1558 1559
          parser->state = NEW_MESSAGE();
          CALLBACK_NOTIFY(message_complete);
1560 1561 1562
          break;
        }

1563
        parser->state = s_headers_done;
1564

1565 1566 1567
        /* Set this here so that on_headers_complete() callbacks can see it */
        parser->upgrade =
          (parser->flags & F_UPGRADE || parser->method == HTTP_CONNECT);
1568 1569 1570 1571 1572 1573

        /* Here we call the headers_complete callback. This is somewhat
         * different than other callbacks because if the user returns 1, we
         * will interpret that as saying that this message has no body. This
         * is needed for the annoying case of recieving a response to a HEAD
         * request.
1574 1575 1576
         *
         * We'd like to use CALLBACK_NOTIFY_NOADVANCE() here but we cannot, so
         * we have to simulate it by handling a change in errno below.
1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592
         */
        if (settings->on_headers_complete) {
          switch (settings->on_headers_complete(parser)) {
            case 0:
              break;

            case 1:
              parser->flags |= F_SKIPBODY;
              break;

            default:
              SET_ERRNO(HPE_CB_headers_complete);
              return p - data; /* Error */
          }
        }

1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605
        if (HTTP_PARSER_ERRNO(parser) != HPE_OK) {
          return p - data;
        }

        goto reexecute_byte;
      }

      case s_headers_done:
      {
        STRICT_CHECK(ch != LF);

        parser->nread = 0;

1606 1607
        /* Exit, the rest of the connect is in a different protocol. */
        if (parser->upgrade) {
1608 1609
          parser->state = NEW_MESSAGE();
          CALLBACK_NOTIFY(message_complete);
1610 1611 1612 1613
          return (p - data) + 1;
        }

        if (parser->flags & F_SKIPBODY) {
1614 1615
          parser->state = NEW_MESSAGE();
          CALLBACK_NOTIFY(message_complete);
1616 1617
        } else if (parser->flags & F_CHUNKED) {
          /* chunked encoding - ignore Content-Length header */
1618
          parser->state = s_chunk_size_start;
1619 1620 1621
        } else {
          if (parser->content_length == 0) {
            /* Content-Length header given but zero: Content-Length: 0\r\n */
1622 1623 1624
            parser->state = NEW_MESSAGE();
            CALLBACK_NOTIFY(message_complete);
          } else if (parser->content_length != ULLONG_MAX) {
1625
            /* Content-Length header given and non-zero */
1626
            parser->state = s_body_identity;
1627
          } else {
1628 1629
            if (parser->type == HTTP_REQUEST ||
                !http_message_needs_eof(parser)) {
1630
              /* Assume content-length 0 - read the next */
1631 1632
              parser->state = NEW_MESSAGE();
              CALLBACK_NOTIFY(message_complete);
1633 1634
            } else {
              /* Read body until EOF */
1635
              parser->state = s_body_identity_eof;
1636 1637 1638 1639 1640 1641 1642 1643
            }
          }
        }

        break;
      }

      case s_body_identity:
1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673
      {
        uint64_t to_read = MIN(parser->content_length,
                               (uint64_t) ((data + len) - p));

        assert(parser->content_length != 0
            && parser->content_length != ULLONG_MAX);

        /* The difference between advancing content_length and p is because
         * the latter will automaticaly advance on the next loop iteration.
         * Further, if content_length ends up at 0, we want to see the last
         * byte again for our message complete callback.
         */
        MARK(body);
        parser->content_length -= to_read;
        p += to_read - 1;

        if (parser->content_length == 0) {
          parser->state = s_message_done;

          /* Mimic CALLBACK_DATA_NOADVANCE() but with one extra byte.
           *
           * The alternative to doing this is to wait for the next byte to
           * trigger the data callback, just as in every other case. The
           * problem with this is that this makes it difficult for the test
           * harness to distinguish between complete-on-EOF and
           * complete-on-length. It's not clear that this distinction is
           * important for applications, but let's keep it for now.
           */
          CALLBACK_DATA_(body, p - body_mark + 1, p - data);
          goto reexecute_byte;
1674
        }
1675

1676
        break;
1677
      }
1678 1679 1680

      /* read until EOF */
      case s_body_identity_eof:
1681 1682 1683 1684 1685 1686 1687 1688
        MARK(body);
        p = data + len - 1;

        break;

      case s_message_done:
        parser->state = NEW_MESSAGE();
        CALLBACK_NOTIFY(message_complete);
1689 1690 1691 1692
        break;

      case s_chunk_size_start:
      {
1693
        assert(parser->nread == 1);
1694 1695 1696 1697 1698 1699 1700 1701 1702
        assert(parser->flags & F_CHUNKED);

        unhex_val = unhex[(unsigned char)ch];
        if (unhex_val == -1) {
          SET_ERRNO(HPE_INVALID_CHUNK_SIZE);
          goto error;
        }

        parser->content_length = unhex_val;
1703
        parser->state = s_chunk_size;
1704 1705 1706 1707 1708
        break;
      }

      case s_chunk_size:
      {
1709 1710
        uint64_t t;

1711 1712 1713
        assert(parser->flags & F_CHUNKED);

        if (ch == CR) {
1714
          parser->state = s_chunk_size_almost_done;
1715 1716 1717 1718 1719 1720 1721
          break;
        }

        unhex_val = unhex[(unsigned char)ch];

        if (unhex_val == -1) {
          if (ch == ';' || ch == ' ') {
1722
            parser->state = s_chunk_parameters;
1723 1724 1725 1726 1727 1728 1729
            break;
          }

          SET_ERRNO(HPE_INVALID_CHUNK_SIZE);
          goto error;
        }

1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740
        t = parser->content_length;
        t *= 16;
        t += unhex_val;

        /* Overflow? */
        if (t < parser->content_length || t == ULLONG_MAX) {
          SET_ERRNO(HPE_INVALID_CONTENT_LENGTH);
          goto error;
        }

        parser->content_length = t;
1741 1742 1743 1744 1745 1746
        break;
      }

      case s_chunk_parameters:
      {
        assert(parser->flags & F_CHUNKED);
1747
        /* just ignore this. TODO check for overflow */
1748
        if (ch == CR) {
1749
          parser->state = s_chunk_size_almost_done;
1750 1751 1752 1753 1754 1755 1756 1757 1758 1759
          break;
        }
        break;
      }

      case s_chunk_size_almost_done:
      {
        assert(parser->flags & F_CHUNKED);
        STRICT_CHECK(ch != LF);

1760
        parser->nread = 0;
1761 1762 1763

        if (parser->content_length == 0) {
          parser->flags |= F_TRAILING;
1764
          parser->state = s_header_field_start;
1765
        } else {
1766
          parser->state = s_chunk_data;
1767 1768 1769 1770 1771 1772
        }
        break;
      }

      case s_chunk_data:
      {
1773 1774
        uint64_t to_read = MIN(parser->content_length,
                               (uint64_t) ((data + len) - p));
1775

1776 1777 1778
        assert(parser->flags & F_CHUNKED);
        assert(parser->content_length != 0
            && parser->content_length != ULLONG_MAX);
1779

1780 1781 1782 1783 1784 1785
        /* See the explanation in s_body_identity for why the content
         * length and data pointers are managed this way.
         */
        MARK(body);
        parser->content_length -= to_read;
        p += to_read - 1;
1786

1787 1788
        if (parser->content_length == 0) {
          parser->state = s_chunk_data_almost_done;
1789 1790 1791 1792 1793 1794 1795
        }

        break;
      }

      case s_chunk_data_almost_done:
        assert(parser->flags & F_CHUNKED);
1796
        assert(parser->content_length == 0);
1797
        STRICT_CHECK(ch != CR);
1798 1799
        parser->state = s_chunk_data_done;
        CALLBACK_DATA(body);
1800 1801 1802 1803 1804
        break;

      case s_chunk_data_done:
        assert(parser->flags & F_CHUNKED);
        STRICT_CHECK(ch != LF);
1805 1806
        parser->nread = 0;
        parser->state = s_chunk_size_start;
1807 1808 1809 1810 1811 1812 1813 1814 1815
        break;

      default:
        assert(0 && "unhandled state");
        SET_ERRNO(HPE_INVALID_INTERNAL_STATE);
        goto error;
    }
  }

1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829
  /* Run callbacks for any marks that we have leftover after we ran our of
   * bytes. There should be at most one of these set, so it's OK to invoke
   * them in series (unset marks will not result in callbacks).
   *
   * We use the NOADVANCE() variety of callbacks here because 'p' has already
   * overflowed 'data' and this allows us to correct for the off-by-one that
   * we'd otherwise have (since CALLBACK_DATA() is meant to be run with a 'p'
   * value that's in-bounds).
   */

  assert(((header_field_mark ? 1 : 0) +
          (header_value_mark ? 1 : 0) +
          (url_mark ? 1 : 0)  +
          (body_mark ? 1 : 0)) <= 1);
1830

1831 1832 1833 1834
  CALLBACK_DATA_NOADVANCE(header_field);
  CALLBACK_DATA_NOADVANCE(header_value);
  CALLBACK_DATA_NOADVANCE(url);
  CALLBACK_DATA_NOADVANCE(body);
1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846

  return len;

error:
  if (HTTP_PARSER_ERRNO(parser) == HPE_OK) {
    SET_ERRNO(HPE_UNKNOWN);
  }

  return (p - data);
}


1847
/* Does the parser need to see an EOF to find the end of the message? */
1848
int
1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872
http_message_needs_eof (const http_parser *parser)
{
  if (parser->type == HTTP_REQUEST) {
    return 0;
  }

  /* See RFC 2616 section 4.4 */
  if (parser->status_code / 100 == 1 || /* 1xx e.g. Continue */
      parser->status_code == 204 ||     /* No Content */
      parser->status_code == 304 ||     /* Not Modified */
      parser->flags & F_SKIPBODY) {     /* response to a HEAD request */
    return 0;
  }

  if ((parser->flags & F_CHUNKED) || parser->content_length != ULLONG_MAX) {
    return 0;
  }

  return 1;
}


int
http_should_keep_alive (const http_parser *parser)
1873 1874 1875 1876 1877 1878 1879 1880
{
  if (parser->http_major > 0 && parser->http_minor > 0) {
    /* HTTP/1.1 */
    if (parser->flags & F_CONNECTION_CLOSE) {
      return 0;
    }
  } else {
    /* HTTP/1.0 or earlier */
1881
    if (!(parser->flags & F_CONNECTION_KEEP_ALIVE)) {
1882 1883 1884
      return 0;
    }
  }
1885 1886

  return !http_message_needs_eof(parser);
1887 1888 1889
}


1890 1891
const char *
http_method_str (enum http_method m)
1892
{
1893
  return ELEM_AT(method_strings, m, "<unknown>");
1894 1895 1896 1897 1898 1899
}


void
http_parser_init (http_parser *parser, enum http_parser_type t)
{
1900 1901 1902
  void *data = parser->data; /* preserve application data */
  memset(parser, 0, sizeof(*parser));
  parser->data = data;
1903 1904
  parser->type = t;
  parser->state = (t == HTTP_REQUEST ? s_start_req : (t == HTTP_RESPONSE ? s_start_res : s_start_req_or_res));
1905
  parser->http_errno = HPE_OK;
1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918
}

const char *
http_errno_name(enum http_errno err) {
  assert(err < (sizeof(http_strerror_tab)/sizeof(http_strerror_tab[0])));
  return http_strerror_tab[err].name;
}

const char *
http_errno_description(enum http_errno err) {
  assert(err < (sizeof(http_strerror_tab)/sizeof(http_strerror_tab[0])));
  return http_strerror_tab[err].description;
}
1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 2097 2098 2099 2100 2101 2102 2103 2104 2105 2106 2107 2108 2109 2110 2111 2112 2113 2114 2115 2116 2117 2118 2119 2120 2121 2122 2123 2124 2125 2126 2127 2128 2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140 2141 2142 2143 2144 2145 2146 2147 2148 2149 2150 2151 2152 2153 2154 2155 2156 2157 2158 2159 2160 2161 2162 2163 2164 2165 2166 2167 2168 2169 2170 2171 2172 2173 2174

static enum http_host_state
http_parse_host_char(enum http_host_state s, const char ch) {
  switch(s) {
    case s_http_userinfo:
    case s_http_userinfo_start:
      if (ch == '@') {
        return s_http_host_start;
      }

      if (IS_USERINFO_CHAR(ch)) {
        return s_http_userinfo;
      }
      break;

    case s_http_host_start:
      if (ch == '[') {
        return s_http_host_v6_start;
      }

      if (IS_HOST_CHAR(ch)) {
        return s_http_host;
      }

      break;

    case s_http_host:
      if (IS_HOST_CHAR(ch)) {
        return s_http_host;
      }

    /* FALLTHROUGH */
    case s_http_host_v6_end:
      if (ch == ':') {
        return s_http_host_port_start;
      }

      break;

    case s_http_host_v6:
      if (ch == ']') {
        return s_http_host_v6_end;
      }

    /* FALLTHROUGH */
    case s_http_host_v6_start:
      if (IS_HEX(ch) || ch == ':') {
        return s_http_host_v6;
      }

      break;

    case s_http_host_port:
    case s_http_host_port_start:
      if (IS_NUM(ch)) {
        return s_http_host_port;
      }

      break;

    default:
      break;
  }
  return s_http_host_dead;
}

static int
http_parse_host(const char * buf, struct http_parser_url *u, int found_at) {
  enum http_host_state s;

  const char *p;
  size_t buflen = u->field_data[UF_HOST].off + u->field_data[UF_HOST].len;

  u->field_data[UF_HOST].len = 0;

  s = found_at ? s_http_userinfo_start : s_http_host_start;

  for (p = buf + u->field_data[UF_HOST].off; p < buf + buflen; p++) {
    enum http_host_state new_s = http_parse_host_char(s, *p);

    if (new_s == s_http_host_dead) {
      return 1;
    }

    switch(new_s) {
      case s_http_host:
        if (s != s_http_host) {
          u->field_data[UF_HOST].off = p - buf;
        }
        u->field_data[UF_HOST].len++;
        break;

      case s_http_host_v6:
        if (s != s_http_host_v6) {
          u->field_data[UF_HOST].off = p - buf;
        }
        u->field_data[UF_HOST].len++;
        break;

      case s_http_host_port:
        if (s != s_http_host_port) {
          u->field_data[UF_PORT].off = p - buf;
          u->field_data[UF_PORT].len = 0;
          u->field_set |= (1 << UF_PORT);
        }
        u->field_data[UF_PORT].len++;
        break;

      case s_http_userinfo:
        if (s != s_http_userinfo) {
          u->field_data[UF_USERINFO].off = p - buf ;
          u->field_data[UF_USERINFO].len = 0;
          u->field_set |= (1 << UF_USERINFO);
        }
        u->field_data[UF_USERINFO].len++;
        break;

      default:
        break;
    }
    s = new_s;
  }

  /* Make sure we don't end somewhere unexpected */
  switch (s) {
    case s_http_host_start:
    case s_http_host_v6_start:
    case s_http_host_v6:
    case s_http_host_port_start:
    case s_http_userinfo:
    case s_http_userinfo_start:
      return 1;
    default:
      break;
  }

  return 0;
}

int
http_parser_parse_url(const char *buf, size_t buflen, int is_connect,
                      struct http_parser_url *u)
{
  enum state s;
  const char *p;
  enum http_parser_url_fields uf, old_uf;
  int found_at = 0;

  u->port = u->field_set = 0;
  s = is_connect ? s_req_server_start : s_req_spaces_before_url;
  uf = old_uf = UF_MAX;

  for (p = buf; p < buf + buflen; p++) {
    s = parse_url_char(s, *p);

    /* Figure out the next field that we're operating on */
    switch (s) {
      case s_dead:
        return 1;

      /* Skip delimeters */
      case s_req_schema_slash:
      case s_req_schema_slash_slash:
      case s_req_server_start:
      case s_req_query_string_start:
      case s_req_fragment_start:
        continue;

      case s_req_schema:
        uf = UF_SCHEMA;
        break;

      case s_req_server_with_at:
        found_at = 1;

      /* FALLTROUGH */
      case s_req_server:
        uf = UF_HOST;
        break;

      case s_req_path:
        uf = UF_PATH;
        break;

      case s_req_query_string:
        uf = UF_QUERY;
        break;

      case s_req_fragment:
        uf = UF_FRAGMENT;
        break;

      default:
        assert(!"Unexpected state");
        return 1;
    }

    /* Nothing's changed; soldier on */
    if (uf == old_uf) {
      u->field_data[uf].len++;
      continue;
    }

    u->field_data[uf].off = p - buf;
    u->field_data[uf].len = 1;

    u->field_set |= (1 << uf);
    old_uf = uf;
  }

  /* host must be present if there is a schema */
  /* parsing http:///toto will fail */
  if ((u->field_set & ((1 << UF_SCHEMA) | (1 << UF_HOST))) != 0) {
    if (http_parse_host(buf, u, found_at) != 0) {
      return 1;
    }
  }

  /* CONNECT requests can only contain "hostname:port" */
  if (is_connect && u->field_set != ((1 << UF_HOST)|(1 << UF_PORT))) {
    return 1;
  }

  if (u->field_set & (1 << UF_PORT)) {
    /* Don't bother with endp; we've already validated the string */
    unsigned long v = strtoul(buf + u->field_data[UF_PORT].off, NULL, 10);

    /* Ports have a max value of 2^16 */
    if (v > 0xffff) {
      return 1;
    }

    u->port = (uint16_t) v;
  }

  return 0;
}

void
http_parser_pause(http_parser *parser, int paused) {
  /* Users should only be pausing/unpausing a parser that is not in an error
   * state. In non-debug builds, there's not much that we can do about this
   * other than ignore it.
   */
  if (HTTP_PARSER_ERRNO(parser) == HPE_OK ||
      HTTP_PARSER_ERRNO(parser) == HPE_PAUSED) {
    SET_ERRNO((paused) ? HPE_PAUSED : HPE_OK);
  } else {
    assert(0 && "Attempting to pause parser in error state");
  }
}

int
http_body_is_final(const struct http_parser *parser) {
    return parser->state == s_message_done;
}