Commit | Line | Data |
---|---|---|
2340bcd3 JVH |
1 | /* Based on src/http/ngx_http_parse.c from NGINX copyright Igor Sysoev |
2 | * | |
3 | * Additional changes are licensed under the same terms as NGINX and | |
4 | * copyright Joyent, Inc. and other Node contributors. All rights reserved. | |
5 | * | |
6 | * Permission is hereby granted, free of charge, to any person obtaining a copy | |
7 | * of this software and associated documentation files (the "Software"), to | |
8 | * deal in the Software without restriction, including without limitation the | |
9 | * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or | |
10 | * sell copies of the Software, and to permit persons to whom the Software is | |
11 | * furnished to do so, subject to the following conditions: | |
12 | * | |
13 | * The above copyright notice and this permission notice shall be included in | |
14 | * all copies or substantial portions of the Software. | |
15 | * | |
16 | * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR | |
17 | * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, | |
18 | * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE | |
19 | * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER | |
20 | * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING | |
21 | * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS | |
22 | * IN THE SOFTWARE. | |
23 | */ | |
24 | #include "http_parser.h" | |
25 | #include <assert.h> | |
26 | #include <stddef.h> | |
27 | ||
28 | ||
29 | #ifndef MIN | |
30 | # define MIN(a,b) ((a) < (b) ? (a) : (b)) | |
31 | #endif | |
32 | ||
33 | ||
34 | #if HTTP_PARSER_DEBUG | |
35 | #define SET_ERRNO(e) \ | |
36 | do { \ | |
37 | parser->http_errno = (e); \ | |
38 | parser->error_lineno = __LINE__; \ | |
39 | } while (0) | |
40 | #else | |
41 | #define SET_ERRNO(e) \ | |
42 | do { \ | |
43 | parser->http_errno = (e); \ | |
44 | } while(0) | |
45 | #endif | |
46 | ||
47 | ||
48 | #define CALLBACK2(FOR) \ | |
49 | do { \ | |
50 | if (settings->on_##FOR) { \ | |
51 | if (0 != settings->on_##FOR(parser)) { \ | |
52 | SET_ERRNO(HPE_CB_##FOR); \ | |
53 | return (p - data); \ | |
54 | } \ | |
55 | } \ | |
56 | } while (0) | |
57 | ||
58 | ||
59 | #define MARK(FOR) \ | |
60 | do { \ | |
61 | FOR##_mark = p; \ | |
62 | } while (0) | |
63 | ||
64 | #define CALLBACK(FOR) \ | |
65 | do { \ | |
66 | if (FOR##_mark) { \ | |
67 | if (settings->on_##FOR) { \ | |
68 | if (0 != settings->on_##FOR(parser, \ | |
69 | FOR##_mark, \ | |
70 | p - FOR##_mark)) \ | |
71 | { \ | |
72 | SET_ERRNO(HPE_CB_##FOR); \ | |
73 | return (p - data); \ | |
74 | } \ | |
75 | } \ | |
76 | FOR##_mark = NULL; \ | |
77 | } \ | |
78 | } while (0) | |
79 | ||
80 | ||
81 | #define PROXY_CONNECTION "proxy-connection" | |
82 | #define CONNECTION "connection" | |
83 | #define CONTENT_LENGTH "content-length" | |
84 | #define TRANSFER_ENCODING "transfer-encoding" | |
85 | #define UPGRADE "upgrade" | |
86 | #define CHUNKED "chunked" | |
87 | #define KEEP_ALIVE "keep-alive" | |
88 | #define CLOSE "close" | |
89 | ||
90 | ||
91 | static const char *method_strings[] = | |
92 | { "DELETE" | |
93 | , "GET" | |
94 | , "HEAD" | |
95 | , "POST" | |
96 | , "PUT" | |
97 | , "CONNECT" | |
98 | , "OPTIONS" | |
99 | , "TRACE" | |
100 | , "COPY" | |
101 | , "LOCK" | |
102 | , "MKCOL" | |
103 | , "MOVE" | |
104 | , "PROPFIND" | |
105 | , "PROPPATCH" | |
106 | , "UNLOCK" | |
107 | , "REPORT" | |
108 | , "MKACTIVITY" | |
109 | , "CHECKOUT" | |
110 | , "MERGE" | |
111 | , "M-SEARCH" | |
112 | , "NOTIFY" | |
113 | , "SUBSCRIBE" | |
114 | , "UNSUBSCRIBE" | |
115 | , "PATCH" | |
116 | , "DESCRIBE" | |
117 | , "ANNOUNCE" | |
118 | , "SETUP" | |
119 | , "PLAY" | |
120 | , "PAUSE" | |
121 | , "TEARDOWN" | |
122 | , "GET_PARAMETER" | |
123 | , "SET_PARAMETER" | |
124 | , "REDIRECT" | |
125 | , "RECORD" | |
126 | , "FLUSH" | |
127 | }; | |
128 | ||
129 | ||
130 | /* Tokens as defined by rfc 2616. Also lowercases them. | |
131 | * token = 1*<any CHAR except CTLs or separators> | |
132 | * separators = "(" | ")" | "<" | ">" | "@" | |
133 | * | "," | ";" | ":" | "\" | <"> | |
134 | * | "/" | "[" | "]" | "?" | "=" | |
135 | * | "{" | "}" | SP | HT | |
136 | */ | |
137 | static const char tokens[256] = { | |
138 | /* 0 nul 1 soh 2 stx 3 etx 4 eot 5 enq 6 ack 7 bel */ | |
139 | 0, 0, 0, 0, 0, 0, 0, 0, | |
140 | /* 8 bs 9 ht 10 nl 11 vt 12 np 13 cr 14 so 15 si */ | |
141 | 0, 0, 0, 0, 0, 0, 0, 0, | |
142 | /* 16 dle 17 dc1 18 dc2 19 dc3 20 dc4 21 nak 22 syn 23 etb */ | |
143 | 0, 0, 0, 0, 0, 0, 0, 0, | |
144 | /* 24 can 25 em 26 sub 27 esc 28 fs 29 gs 30 rs 31 us */ | |
145 | 0, 0, 0, 0, 0, 0, 0, 0, | |
146 | /* 32 sp 33 ! 34 " 35 # 36 $ 37 % 38 & 39 ' */ | |
147 | ' ', '!', '"', '#', '$', '%', '&', '\'', | |
148 | /* 40 ( 41 ) 42 * 43 + 44 , 45 - 46 . 47 / */ | |
149 | 0, 0, '*', '+', 0, '-', '.', '/', | |
150 | /* 48 0 49 1 50 2 51 3 52 4 53 5 54 6 55 7 */ | |
151 | '0', '1', '2', '3', '4', '5', '6', '7', | |
152 | /* 56 8 57 9 58 : 59 ; 60 < 61 = 62 > 63 ? */ | |
153 | '8', '9', 0, 0, 0, 0, 0, 0, | |
154 | /* 64 @ 65 A 66 B 67 C 68 D 69 E 70 F 71 G */ | |
155 | 0, 'a', 'b', 'c', 'd', 'e', 'f', 'g', | |
156 | /* 72 H 73 I 74 J 75 K 76 L 77 M 78 N 79 O */ | |
157 | 'h', 'i', 'j', 'k', 'l', 'm', 'n', 'o', | |
158 | /* 80 P 81 Q 82 R 83 S 84 T 85 U 86 V 87 W */ | |
159 | 'p', 'q', 'r', 's', 't', 'u', 'v', 'w', | |
160 | /* 88 X 89 Y 90 Z 91 [ 92 \ 93 ] 94 ^ 95 _ */ | |
161 | 'x', 'y', 'z', 0, 0, 0, '^', '_', | |
162 | /* 96 ` 97 a 98 b 99 c 100 d 101 e 102 f 103 g */ | |
163 | '`', 'a', 'b', 'c', 'd', 'e', 'f', 'g', | |
164 | /* 104 h 105 i 106 j 107 k 108 l 109 m 110 n 111 o */ | |
165 | 'h', 'i', 'j', 'k', 'l', 'm', 'n', 'o', | |
166 | /* 112 p 113 q 114 r 115 s 116 t 117 u 118 v 119 w */ | |
167 | 'p', 'q', 'r', 's', 't', 'u', 'v', 'w', | |
168 | /* 120 x 121 y 122 z 123 { 124 | 125 } 126 ~ 127 del */ | |
169 | 'x', 'y', 'z', 0, '|', '}', '~', 0 }; | |
170 | ||
171 | ||
172 | static const int8_t unhex[256] = | |
173 | {-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1 | |
174 | ,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1 | |
175 | ,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1 | |
176 | , 0, 1, 2, 3, 4, 5, 6, 7, 8, 9,-1,-1,-1,-1,-1,-1 | |
177 | ,-1,10,11,12,13,14,15,-1,-1,-1,-1,-1,-1,-1,-1,-1 | |
178 | ,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1 | |
179 | ,-1,10,11,12,13,14,15,-1,-1,-1,-1,-1,-1,-1,-1,-1 | |
180 | ,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1 | |
181 | }; | |
182 | ||
183 | ||
184 | static const uint8_t normal_url_char[256] = { | |
185 | /* 0 nul 1 soh 2 stx 3 etx 4 eot 5 enq 6 ack 7 bel */ | |
186 | 0, 0, 0, 0, 0, 0, 0, 0, | |
187 | /* 8 bs 9 ht 10 nl 11 vt 12 np 13 cr 14 so 15 si */ | |
188 | 0, 0, 0, 0, 0, 0, 0, 0, | |
189 | /* 16 dle 17 dc1 18 dc2 19 dc3 20 dc4 21 nak 22 syn 23 etb */ | |
190 | 0, 0, 0, 0, 0, 0, 0, 0, | |
191 | /* 24 can 25 em 26 sub 27 esc 28 fs 29 gs 30 rs 31 us */ | |
192 | 0, 0, 0, 0, 0, 0, 0, 0, | |
193 | /* 32 sp 33 ! 34 " 35 # 36 $ 37 % 38 & 39 ' */ | |
194 | 0, 1, 1, 0, 1, 1, 1, 1, | |
195 | /* 40 ( 41 ) 42 * 43 + 44 , 45 - 46 . 47 / */ | |
196 | 1, 1, 1, 1, 1, 1, 1, 1, | |
197 | /* 48 0 49 1 50 2 51 3 52 4 53 5 54 6 55 7 */ | |
198 | 1, 1, 1, 1, 1, 1, 1, 1, | |
199 | /* 56 8 57 9 58 : 59 ; 60 < 61 = 62 > 63 ? */ | |
200 | 1, 1, 1, 1, 1, 1, 1, 0, | |
201 | /* 64 @ 65 A 66 B 67 C 68 D 69 E 70 F 71 G */ | |
202 | 1, 1, 1, 1, 1, 1, 1, 1, | |
203 | /* 72 H 73 I 74 J 75 K 76 L 77 M 78 N 79 O */ | |
204 | 1, 1, 1, 1, 1, 1, 1, 1, | |
205 | /* 80 P 81 Q 82 R 83 S 84 T 85 U 86 V 87 W */ | |
206 | 1, 1, 1, 1, 1, 1, 1, 1, | |
207 | /* 88 X 89 Y 90 Z 91 [ 92 \ 93 ] 94 ^ 95 _ */ | |
208 | 1, 1, 1, 1, 1, 1, 1, 1, | |
209 | /* 96 ` 97 a 98 b 99 c 100 d 101 e 102 f 103 g */ | |
210 | 1, 1, 1, 1, 1, 1, 1, 1, | |
211 | /* 104 h 105 i 106 j 107 k 108 l 109 m 110 n 111 o */ | |
212 | 1, 1, 1, 1, 1, 1, 1, 1, | |
213 | /* 112 p 113 q 114 r 115 s 116 t 117 u 118 v 119 w */ | |
214 | 1, 1, 1, 1, 1, 1, 1, 1, | |
215 | /* 120 x 121 y 122 z 123 { 124 | 125 } 126 ~ 127 del */ | |
216 | 1, 1, 1, 1, 1, 1, 1, 0, }; | |
217 | ||
218 | ||
219 | enum state | |
220 | { s_dead = 1 /* important that this is > 0 */ | |
221 | ||
222 | , s_start_req_or_res | |
223 | , s_res_or_resp_H | |
224 | , s_start_res | |
225 | , s_res_H | |
226 | , s_res_HT | |
227 | , s_res_HTT | |
228 | , s_res_HTTP | |
229 | , s_res_first_http_major | |
230 | , s_res_http_major | |
231 | , s_res_first_http_minor | |
232 | , s_res_http_minor | |
233 | , s_res_first_status_code | |
234 | , s_res_status_code | |
235 | , s_res_status | |
236 | , s_res_line_almost_done | |
237 | ||
238 | , s_start_req | |
239 | ||
240 | , s_req_method | |
241 | , s_req_spaces_before_url | |
242 | , s_req_schema | |
243 | , s_req_schema_slash | |
244 | , s_req_schema_slash_slash | |
245 | , s_req_host | |
246 | , s_req_port | |
247 | , s_req_path | |
248 | , s_req_query_string_start | |
249 | , s_req_query_string | |
250 | , s_req_fragment_start | |
251 | , s_req_fragment | |
252 | , s_req_http_start | |
253 | , s_req_http_H | |
254 | , s_req_http_HT | |
255 | , s_req_http_HTT | |
256 | , s_req_http_HTTP | |
257 | , s_req_first_http_major | |
258 | , s_req_http_major | |
259 | , s_req_first_http_minor | |
260 | , s_req_http_minor | |
261 | , s_req_line_almost_done | |
262 | ||
263 | , s_header_field_start | |
264 | , s_header_field | |
265 | , s_header_value_start | |
266 | , s_header_value | |
267 | , s_header_value_lws | |
268 | ||
269 | , s_header_almost_done | |
270 | ||
271 | , s_chunk_size_start | |
272 | , s_chunk_size | |
273 | , s_chunk_parameters | |
274 | , s_chunk_size_almost_done | |
275 | ||
276 | , s_headers_almost_done | |
277 | /* Important: 's_headers_almost_done' must be the last 'header' state. All | |
278 | * states beyond this must be 'body' states. It is used for overflow | |
279 | * checking. See the PARSING_HEADER() macro. | |
280 | */ | |
281 | ||
282 | , s_chunk_data | |
283 | , s_chunk_data_almost_done | |
284 | , s_chunk_data_done | |
285 | ||
286 | , s_body_identity | |
287 | , s_body_identity_eof | |
288 | }; | |
289 | ||
290 | ||
291 | #define PARSING_HEADER(state) (state <= s_headers_almost_done) | |
292 | ||
293 | ||
294 | enum header_states | |
295 | { h_general = 0 | |
296 | , h_C | |
297 | , h_CO | |
298 | , h_CON | |
299 | ||
300 | , h_matching_connection | |
301 | , h_matching_proxy_connection | |
302 | , h_matching_content_length | |
303 | , h_matching_transfer_encoding | |
304 | , h_matching_upgrade | |
305 | ||
306 | , h_connection | |
307 | , h_content_length | |
308 | , h_transfer_encoding | |
309 | , h_upgrade | |
310 | ||
311 | , h_matching_transfer_encoding_chunked | |
312 | , h_matching_connection_keep_alive | |
313 | , h_matching_connection_close | |
314 | ||
315 | , h_transfer_encoding_chunked | |
316 | , h_connection_keep_alive | |
317 | , h_connection_close | |
318 | }; | |
319 | ||
320 | ||
321 | /* Macros for character classes; depends on strict-mode */ | |
322 | #define CR '\r' | |
323 | #define LF '\n' | |
324 | #define LOWER(c) (unsigned char)(c | 0x20) | |
325 | #define TOKEN(c) (tokens[(unsigned char)c]) | |
326 | #define IS_ALPHA(c) (LOWER(c) >= 'a' && LOWER(c) <= 'z') | |
327 | #define IS_NUM(c) ((c) >= '0' && (c) <= '9') | |
328 | #define IS_ALPHANUM(c) (IS_ALPHA(c) || IS_NUM(c)) | |
329 | #define IS_NUMERIC_CHAR(c) (IS_NUM(c) || (LOWER(c) >= 'a' && LOWER(c) <= 'f') || (c) == '.' || (c) == ':') | |
330 | ||
331 | #if HTTP_PARSER_STRICT | |
332 | #define IS_URL_CHAR(c) (normal_url_char[(unsigned char) (c)]) | |
333 | #define IS_HOST_CHAR(c) (IS_ALPHANUM(c) || (c) == '.' || (c) == '-') | |
334 | #else | |
335 | #define IS_URL_CHAR(c) \ | |
336 | (normal_url_char[(unsigned char) (c)] || ((c) & 0x80)) | |
337 | #define IS_HOST_CHAR(c) \ | |
338 | (IS_ALPHANUM(c) || (c) == '.' || (c) == '-' || (c) == '_') | |
339 | #endif | |
340 | ||
341 | ||
342 | #define start_state (parser->type == HTTP_REQUEST ? s_start_req : s_start_res) | |
343 | ||
344 | ||
345 | #if HTTP_PARSER_STRICT | |
346 | # define STRICT_CHECK(cond) \ | |
347 | do { \ | |
348 | if (cond) { \ | |
349 | SET_ERRNO(HPE_STRICT); \ | |
350 | goto error; \ | |
351 | } \ | |
352 | } while (0) | |
353 | # define NEW_MESSAGE() (http_should_keep_alive(parser) ? start_state : s_dead) | |
354 | #else | |
355 | # define STRICT_CHECK(cond) | |
356 | # define NEW_MESSAGE() start_state | |
357 | #endif | |
358 | ||
359 | ||
360 | /* Map errno values to strings for human-readable output */ | |
361 | #define HTTP_STRERROR_GEN(n, s) { "HPE_" #n, s }, | |
362 | static struct { | |
363 | const char *name; | |
364 | const char *description; | |
365 | } http_strerror_tab[] = { | |
366 | HTTP_ERRNO_MAP(HTTP_STRERROR_GEN) | |
367 | }; | |
368 | #undef HTTP_STRERROR_GEN | |
369 | ||
370 | ||
371 | size_t http_parser_execute (http_parser *parser, | |
372 | const http_parser_settings *settings, | |
373 | const char *data, | |
374 | size_t len) | |
375 | { | |
376 | char c, ch; | |
377 | int8_t unhex_val; | |
378 | const char *p = data, *pe; | |
379 | int64_t to_read; | |
380 | enum state state; | |
381 | enum header_states header_state; | |
382 | uint64_t index = parser->index; | |
383 | uint64_t nread = parser->nread; | |
384 | ||
385 | /* technically we could combine all of these (except for url_mark) into one | |
386 | variable, saving stack space, but it seems more clear to have them | |
387 | separated. */ | |
388 | const char *header_field_mark = 0; | |
389 | const char *header_value_mark = 0; | |
390 | const char *url_mark = 0; | |
391 | ||
392 | /* We're in an error state. Don't bother doing anything. */ | |
393 | if (HTTP_PARSER_ERRNO(parser) != HPE_OK) { | |
394 | return 0; | |
395 | } | |
396 | ||
397 | state = (enum state) parser->state; | |
398 | header_state = (enum header_states) parser->header_state; | |
399 | ||
400 | if (len == 0) { | |
401 | switch (state) { | |
402 | case s_body_identity_eof: | |
403 | CALLBACK2(message_complete); | |
404 | return 0; | |
405 | ||
406 | case s_dead: | |
407 | case s_start_req_or_res: | |
408 | case s_start_res: | |
409 | case s_start_req: | |
410 | return 0; | |
411 | ||
412 | default: | |
413 | SET_ERRNO(HPE_INVALID_EOF_STATE); | |
414 | return 1; | |
415 | } | |
416 | } | |
417 | ||
418 | ||
419 | if (state == s_header_field) | |
420 | header_field_mark = data; | |
421 | if (state == s_header_value) | |
422 | header_value_mark = data; | |
423 | if (state == s_req_path || state == s_req_schema || state == s_req_schema_slash | |
424 | || state == s_req_schema_slash_slash || state == s_req_port | |
425 | || state == s_req_query_string_start || state == s_req_query_string | |
426 | || state == s_req_host | |
427 | || state == s_req_fragment_start || state == s_req_fragment) | |
428 | url_mark = data; | |
429 | ||
430 | for (p=data, pe=data+len; p != pe; p++) { | |
431 | ch = *p; | |
432 | ||
433 | if (PARSING_HEADER(state)) { | |
434 | ++nread; | |
435 | /* Buffer overflow attack */ | |
436 | if (nread > HTTP_MAX_HEADER_SIZE) { | |
437 | SET_ERRNO(HPE_HEADER_OVERFLOW); | |
438 | goto error; | |
439 | } | |
440 | } | |
441 | ||
442 | switch (state) { | |
443 | ||
444 | case s_dead: | |
445 | /* this state is used after a 'Connection: close' message | |
446 | * the parser will error out if it reads another message | |
447 | */ | |
448 | SET_ERRNO(HPE_CLOSED_CONNECTION); | |
449 | goto error; | |
450 | ||
451 | case s_start_req_or_res: | |
452 | { | |
453 | if (ch == CR || ch == LF) | |
454 | break; | |
455 | parser->flags = 0; | |
456 | parser->content_length = -1; | |
457 | ||
458 | CALLBACK2(message_begin); | |
459 | ||
460 | if (ch == 'H') | |
461 | state = s_res_or_resp_H; | |
462 | else { | |
463 | parser->type = HTTP_REQUEST; | |
464 | goto start_req_method_assign; | |
465 | } | |
466 | break; | |
467 | } | |
468 | ||
469 | case s_res_or_resp_H: | |
470 | if (ch == 'T') { | |
471 | parser->type = HTTP_RESPONSE; | |
472 | state = s_res_HT; | |
473 | } else { | |
474 | if (ch != 'E') { | |
475 | SET_ERRNO(HPE_INVALID_CONSTANT); | |
476 | goto error; | |
477 | } | |
478 | ||
479 | parser->type = HTTP_REQUEST; | |
480 | parser->method = HTTP_HEAD; | |
481 | index = 2; | |
482 | state = s_req_method; | |
483 | } | |
484 | break; | |
485 | ||
486 | case s_start_res: | |
487 | { | |
488 | parser->flags = 0; | |
489 | parser->content_length = -1; | |
490 | ||
491 | CALLBACK2(message_begin); | |
492 | ||
493 | switch (ch) { | |
494 | case 'H': | |
495 | state = s_res_H; | |
496 | break; | |
497 | ||
498 | case CR: | |
499 | case LF: | |
500 | break; | |
501 | ||
502 | default: | |
503 | SET_ERRNO(HPE_INVALID_CONSTANT); | |
504 | goto error; | |
505 | } | |
506 | break; | |
507 | } | |
508 | ||
509 | case s_res_H: | |
510 | STRICT_CHECK(ch != 'T'); | |
511 | state = s_res_HT; | |
512 | break; | |
513 | ||
514 | case s_res_HT: | |
515 | STRICT_CHECK(ch != 'T'); | |
516 | state = s_res_HTT; | |
517 | break; | |
518 | ||
519 | case s_res_HTT: | |
520 | STRICT_CHECK(ch != 'P'); | |
521 | state = s_res_HTTP; | |
522 | break; | |
523 | ||
524 | case s_res_HTTP: | |
525 | STRICT_CHECK(ch != '/'); | |
526 | state = s_res_first_http_major; | |
527 | break; | |
528 | ||
529 | case s_res_first_http_major: | |
530 | if (ch < '0' || ch > '9') { | |
531 | SET_ERRNO(HPE_INVALID_VERSION); | |
532 | goto error; | |
533 | } | |
534 | ||
535 | parser->http_major = ch - '0'; | |
536 | state = s_res_http_major; | |
537 | break; | |
538 | ||
539 | /* major HTTP version or dot */ | |
540 | case s_res_http_major: | |
541 | { | |
542 | if (ch == '.') { | |
543 | state = s_res_first_http_minor; | |
544 | break; | |
545 | } | |
546 | ||
547 | if (!IS_NUM(ch)) { | |
548 | SET_ERRNO(HPE_INVALID_VERSION); | |
549 | goto error; | |
550 | } | |
551 | ||
552 | parser->http_major *= 10; | |
553 | parser->http_major += ch - '0'; | |
554 | ||
555 | if (parser->http_major > 999) { | |
556 | SET_ERRNO(HPE_INVALID_VERSION); | |
557 | goto error; | |
558 | } | |
559 | ||
560 | break; | |
561 | } | |
562 | ||
563 | /* first digit of minor HTTP version */ | |
564 | case s_res_first_http_minor: | |
565 | if (!IS_NUM(ch)) { | |
566 | SET_ERRNO(HPE_INVALID_VERSION); | |
567 | goto error; | |
568 | } | |
569 | ||
570 | parser->http_minor = ch - '0'; | |
571 | state = s_res_http_minor; | |
572 | break; | |
573 | ||
574 | /* minor HTTP version or end of request line */ | |
575 | case s_res_http_minor: | |
576 | { | |
577 | if (ch == ' ') { | |
578 | state = s_res_first_status_code; | |
579 | break; | |
580 | } | |
581 | ||
582 | if (!IS_NUM(ch)) { | |
583 | SET_ERRNO(HPE_INVALID_VERSION); | |
584 | goto error; | |
585 | } | |
586 | ||
587 | parser->http_minor *= 10; | |
588 | parser->http_minor += ch - '0'; | |
589 | ||
590 | if (parser->http_minor > 999) { | |
591 | SET_ERRNO(HPE_INVALID_VERSION); | |
592 | goto error; | |
593 | } | |
594 | ||
595 | break; | |
596 | } | |
597 | ||
598 | case s_res_first_status_code: | |
599 | { | |
600 | if (!IS_NUM(ch)) { | |
601 | if (ch == ' ') { | |
602 | break; | |
603 | } | |
604 | ||
605 | SET_ERRNO(HPE_INVALID_STATUS); | |
606 | goto error; | |
607 | } | |
608 | parser->status_code = ch - '0'; | |
609 | state = s_res_status_code; | |
610 | break; | |
611 | } | |
612 | ||
613 | case s_res_status_code: | |
614 | { | |
615 | if (!IS_NUM(ch)) { | |
616 | switch (ch) { | |
617 | case ' ': | |
618 | state = s_res_status; | |
619 | break; | |
620 | case CR: | |
621 | state = s_res_line_almost_done; | |
622 | break; | |
623 | case LF: | |
624 | state = s_header_field_start; | |
625 | break; | |
626 | default: | |
627 | SET_ERRNO(HPE_INVALID_STATUS); | |
628 | goto error; | |
629 | } | |
630 | break; | |
631 | } | |
632 | ||
633 | parser->status_code *= 10; | |
634 | parser->status_code += ch - '0'; | |
635 | ||
636 | if (parser->status_code > 999) { | |
637 | SET_ERRNO(HPE_INVALID_STATUS); | |
638 | goto error; | |
639 | } | |
640 | ||
641 | break; | |
642 | } | |
643 | ||
644 | case s_res_status: | |
645 | /* the human readable status. e.g. "NOT FOUND" | |
646 | * we are not humans so just ignore this */ | |
647 | if (ch == CR) { | |
648 | state = s_res_line_almost_done; | |
649 | break; | |
650 | } | |
651 | ||
652 | if (ch == LF) { | |
653 | state = s_header_field_start; | |
654 | break; | |
655 | } | |
656 | break; | |
657 | ||
658 | case s_res_line_almost_done: | |
659 | STRICT_CHECK(ch != LF); | |
660 | state = s_header_field_start; | |
661 | break; | |
662 | ||
663 | case s_start_req: | |
664 | { | |
665 | if (ch == CR || ch == LF) | |
666 | break; | |
667 | parser->flags = 0; | |
668 | parser->content_length = -1; | |
669 | ||
670 | CALLBACK2(message_begin); | |
671 | ||
672 | if (!IS_ALPHA(ch)) { | |
673 | SET_ERRNO(HPE_INVALID_METHOD); | |
674 | goto error; | |
675 | } | |
676 | ||
677 | start_req_method_assign: | |
678 | parser->method = (enum http_method) 0; | |
679 | index = 1; | |
680 | switch (ch) { | |
681 | case 'A': parser->method = HTTP_ANNOUNCE; break; | |
682 | case 'C': parser->method = HTTP_CONNECT; /* or COPY, CHECKOUT */ break; | |
683 | case 'D': parser->method = HTTP_DELETE; /* or DESCRIBE */ break; | |
684 | case 'F': parser->method = HTTP_FLUSH; break; | |
685 | case 'G': parser->method = HTTP_GET; /* or GET_PARAMETER */ break; | |
686 | case 'H': parser->method = HTTP_HEAD; break; | |
687 | case 'L': parser->method = HTTP_LOCK; break; | |
688 | case 'M': parser->method = HTTP_MKCOL; /* or MOVE, MKACTIVITY, MERGE, M-SEARCH */ break; | |
689 | case 'N': parser->method = HTTP_NOTIFY; break; | |
690 | case 'O': parser->method = HTTP_OPTIONS; break; | |
691 | case 'P': parser->method = HTTP_POST; | |
692 | /* or PROPFIND or PROPPATCH or PUT or PATCH or PLAY or PAUSE */ | |
693 | break; | |
694 | case 'R': parser->method = HTTP_REPORT; /* or REDIRECT, RECORD */ break; | |
695 | case 'S': parser->method = HTTP_SUBSCRIBE; /* or SETUP, SET_PARAMETER */ break; | |
696 | case 'T': parser->method = HTTP_TRACE; /* or TEARDOWN */ break; | |
697 | case 'U': parser->method = HTTP_UNLOCK; /* or UNSUBSCRIBE */ break; | |
698 | default: | |
699 | SET_ERRNO(HPE_INVALID_METHOD); | |
700 | goto error; | |
701 | } | |
702 | state = s_req_method; | |
703 | break; | |
704 | } | |
705 | ||
706 | case s_req_method: | |
707 | { | |
708 | const char *matcher; | |
709 | if (ch == '\0') { | |
710 | SET_ERRNO(HPE_INVALID_METHOD); | |
711 | goto error; | |
712 | } | |
713 | ||
714 | matcher = method_strings[parser->method]; | |
715 | if (ch == ' ' && matcher[index] == '\0') { | |
716 | state = s_req_spaces_before_url; | |
717 | } else if (ch == matcher[index]) { | |
718 | ; /* nada */ | |
719 | } else if (parser->method == HTTP_CONNECT) { | |
720 | if (index == 1 && ch == 'H') { | |
721 | parser->method = HTTP_CHECKOUT; | |
722 | } else if (index == 2 && ch == 'P') { | |
723 | parser->method = HTTP_COPY; | |
724 | } else { | |
725 | goto error; | |
726 | } | |
727 | } else if (index == 2 && parser->method == HTTP_DELETE && ch == 'S') { | |
728 | parser->method = HTTP_DESCRIBE; | |
729 | } else if (index == 3 && parser->method == HTTP_GET && ch == '_') { | |
730 | parser->method = HTTP_GET_PARAMETER; | |
731 | } else if (parser->method == HTTP_MKCOL) { | |
732 | if (index == 1 && ch == 'O') { | |
733 | parser->method = HTTP_MOVE; | |
734 | } else if (index == 1 && ch == 'E') { | |
735 | parser->method = HTTP_MERGE; | |
736 | } else if (index == 1 && ch == '-') { | |
737 | parser->method = HTTP_MSEARCH; | |
738 | } else if (index == 2 && ch == 'A') { | |
739 | parser->method = HTTP_MKACTIVITY; | |
740 | } else { | |
741 | goto error; | |
742 | } | |
743 | } else if (index == 1 && parser->method == HTTP_POST) { | |
744 | if (ch == 'R') { | |
745 | parser->method = HTTP_PROPFIND; /* or HTTP_PROPPATCH */ | |
746 | } else if (ch == 'U') { | |
747 | parser->method = HTTP_PUT; | |
748 | } else if (ch == 'A') { | |
749 | parser->method = HTTP_PATCH; /* or HTTP_PAUSE */ | |
750 | } else if (ch == 'L') { | |
751 | parser->method = HTTP_PLAY; | |
752 | } else { | |
753 | goto error; | |
754 | } | |
755 | } else if (index == 2 && parser->method == HTTP_REPORT) { | |
756 | if (ch == 'D') { | |
757 | parser->method = HTTP_REDIRECT; | |
758 | } else if (ch == 'C') { | |
759 | parser->method = HTTP_RECORD; | |
760 | } else { | |
761 | goto error; | |
762 | } | |
763 | } else if (index == 1 && parser->method == HTTP_SUBSCRIBE && ch == 'E') { | |
764 | parser->method = HTTP_SETUP; /* or HTTP_SET_PARAMETER */ | |
765 | } else if (index == 3 && parser->method == HTTP_SETUP && ch == '_') { | |
766 | parser->method = HTTP_SET_PARAMETER; | |
767 | } else if (index == 1 && parser->method == HTTP_TRACE && ch == 'E') { | |
768 | parser->method = HTTP_TEARDOWN; | |
769 | } else if (index == 2 && parser->method == HTTP_UNLOCK && ch == 'S') { | |
770 | parser->method = HTTP_UNSUBSCRIBE; | |
771 | } else if (index == 4 && parser->method == HTTP_PROPFIND && ch == 'P') { | |
772 | parser->method = HTTP_PROPPATCH; | |
773 | } else if (index == 2 && parser->method == HTTP_PATCH && ch == 'U') { | |
774 | parser->method = HTTP_PAUSE; | |
775 | } else { | |
776 | SET_ERRNO(HPE_INVALID_METHOD); | |
777 | goto error; | |
778 | } | |
779 | ||
780 | ++index; | |
781 | break; | |
782 | } | |
783 | case s_req_spaces_before_url: | |
784 | { | |
785 | if (ch == ' ') break; | |
786 | ||
787 | if (ch == '/' || ch == '*') { | |
788 | MARK(url); | |
789 | state = s_req_path; | |
790 | break; | |
791 | } | |
792 | ||
793 | /* Proxied requests are followed by scheme of an absolute URI (alpha). | |
794 | * CONNECT is followed by a hostname, which begins with alphanum. | |
795 | * All other methods are followed by '/' or '*' (handled above). | |
796 | */ | |
797 | if (IS_ALPHA(ch) || (parser->method == HTTP_CONNECT && IS_NUM(ch))) { | |
798 | MARK(url); | |
799 | state = (parser->method == HTTP_CONNECT) ? s_req_host : s_req_schema; | |
800 | break; | |
801 | } | |
802 | ||
803 | SET_ERRNO(HPE_INVALID_URL); | |
804 | goto error; | |
805 | } | |
806 | ||
807 | case s_req_schema: | |
808 | { | |
809 | if (IS_ALPHA(ch)) break; | |
810 | ||
811 | if (ch == ':') { | |
812 | state = s_req_schema_slash; | |
813 | break; | |
814 | } | |
815 | ||
816 | SET_ERRNO(HPE_INVALID_URL); | |
817 | goto error; | |
818 | } | |
819 | ||
820 | case s_req_schema_slash: | |
821 | STRICT_CHECK(ch != '/'); | |
822 | state = s_req_schema_slash_slash; | |
823 | break; | |
824 | ||
825 | case s_req_schema_slash_slash: | |
826 | STRICT_CHECK(ch != '/'); | |
827 | state = s_req_host; | |
828 | break; | |
829 | ||
830 | case s_req_host: | |
831 | { | |
832 | if (parser->numerichost) { | |
833 | if (IS_NUMERIC_CHAR(ch)) break; | |
834 | } else { | |
835 | if (IS_HOST_CHAR(ch)) break; | |
836 | } | |
837 | switch (ch) { | |
838 | case ':': | |
839 | state = s_req_port; | |
840 | break; | |
841 | case '/': | |
842 | state = s_req_path; | |
843 | break; | |
844 | case ' ': | |
845 | /* The request line looks like: | |
846 | * "GET http://foo.bar.com HTTP/1.1" | |
847 | * That is, there is no path. | |
848 | */ | |
849 | CALLBACK(url); | |
850 | state = s_req_http_start; | |
851 | break; | |
852 | case '?': | |
853 | state = s_req_query_string_start; | |
854 | break; | |
855 | default: | |
856 | SET_ERRNO(HPE_INVALID_HOST); | |
857 | goto error; | |
858 | } | |
859 | break; | |
860 | } | |
861 | ||
862 | case s_req_port: | |
863 | { | |
864 | if (IS_NUM(ch)) break; | |
865 | switch (ch) { | |
866 | case '/': | |
867 | state = s_req_path; | |
868 | break; | |
869 | case ' ': | |
870 | /* The request line looks like: | |
871 | * "GET http://foo.bar.com:1234 HTTP/1.1" | |
872 | * That is, there is no path. | |
873 | */ | |
874 | CALLBACK(url); | |
875 | state = s_req_http_start; | |
876 | break; | |
877 | case '?': | |
878 | state = s_req_query_string_start; | |
879 | break; | |
880 | default: | |
881 | SET_ERRNO(HPE_INVALID_PORT); | |
882 | goto error; | |
883 | } | |
884 | break; | |
885 | } | |
886 | ||
887 | case s_req_path: | |
888 | { | |
889 | if (IS_URL_CHAR(ch)) break; | |
890 | ||
891 | switch (ch) { | |
892 | case ' ': | |
893 | CALLBACK(url); | |
894 | state = s_req_http_start; | |
895 | break; | |
896 | case CR: | |
897 | CALLBACK(url); | |
898 | parser->http_major = 0; | |
899 | parser->http_minor = 9; | |
900 | state = s_req_line_almost_done; | |
901 | break; | |
902 | case LF: | |
903 | CALLBACK(url); | |
904 | parser->http_major = 0; | |
905 | parser->http_minor = 9; | |
906 | state = s_header_field_start; | |
907 | break; | |
908 | case '?': | |
909 | state = s_req_query_string_start; | |
910 | break; | |
911 | case '#': | |
912 | state = s_req_fragment_start; | |
913 | break; | |
914 | default: | |
915 | SET_ERRNO(HPE_INVALID_PATH); | |
916 | goto error; | |
917 | } | |
918 | break; | |
919 | } | |
920 | ||
921 | case s_req_query_string_start: | |
922 | { | |
923 | if (IS_URL_CHAR(ch)) { | |
924 | state = s_req_query_string; | |
925 | break; | |
926 | } | |
927 | ||
928 | switch (ch) { | |
929 | case '?': | |
930 | break; /* XXX ignore extra '?' ... is this right? */ | |
931 | case ' ': | |
932 | CALLBACK(url); | |
933 | state = s_req_http_start; | |
934 | break; | |
935 | case CR: | |
936 | CALLBACK(url); | |
937 | parser->http_major = 0; | |
938 | parser->http_minor = 9; | |
939 | state = s_req_line_almost_done; | |
940 | break; | |
941 | case LF: | |
942 | CALLBACK(url); | |
943 | parser->http_major = 0; | |
944 | parser->http_minor = 9; | |
945 | state = s_header_field_start; | |
946 | break; | |
947 | case '#': | |
948 | state = s_req_fragment_start; | |
949 | break; | |
950 | default: | |
951 | SET_ERRNO(HPE_INVALID_QUERY_STRING); | |
952 | goto error; | |
953 | } | |
954 | break; | |
955 | } | |
956 | ||
957 | case s_req_query_string: | |
958 | { | |
959 | if (IS_URL_CHAR(ch)) break; | |
960 | ||
961 | switch (ch) { | |
962 | case '?': | |
963 | /* allow extra '?' in query string */ | |
964 | break; | |
965 | case ' ': | |
966 | CALLBACK(url); | |
967 | state = s_req_http_start; | |
968 | break; | |
969 | case CR: | |
970 | CALLBACK(url); | |
971 | parser->http_major = 0; | |
972 | parser->http_minor = 9; | |
973 | state = s_req_line_almost_done; | |
974 | break; | |
975 | case LF: | |
976 | CALLBACK(url); | |
977 | parser->http_major = 0; | |
978 | parser->http_minor = 9; | |
979 | state = s_header_field_start; | |
980 | break; | |
981 | case '#': | |
982 | state = s_req_fragment_start; | |
983 | break; | |
984 | default: | |
985 | SET_ERRNO(HPE_INVALID_QUERY_STRING); | |
986 | goto error; | |
987 | } | |
988 | break; | |
989 | } | |
990 | ||
991 | case s_req_fragment_start: | |
992 | { | |
993 | if (IS_URL_CHAR(ch)) { | |
994 | state = s_req_fragment; | |
995 | break; | |
996 | } | |
997 | ||
998 | switch (ch) { | |
999 | case ' ': | |
1000 | CALLBACK(url); | |
1001 | state = s_req_http_start; | |
1002 | break; | |
1003 | case CR: | |
1004 | CALLBACK(url); | |
1005 | parser->http_major = 0; | |
1006 | parser->http_minor = 9; | |
1007 | state = s_req_line_almost_done; | |
1008 | break; | |
1009 | case LF: | |
1010 | CALLBACK(url); | |
1011 | parser->http_major = 0; | |
1012 | parser->http_minor = 9; | |
1013 | state = s_header_field_start; | |
1014 | break; | |
1015 | case '?': | |
1016 | state = s_req_fragment; | |
1017 | break; | |
1018 | case '#': | |
1019 | break; | |
1020 | default: | |
1021 | SET_ERRNO(HPE_INVALID_FRAGMENT); | |
1022 | goto error; | |
1023 | } | |
1024 | break; | |
1025 | } | |
1026 | ||
1027 | case s_req_fragment: | |
1028 | { | |
1029 | if (IS_URL_CHAR(ch)) break; | |
1030 | ||
1031 | switch (ch) { | |
1032 | case ' ': | |
1033 | CALLBACK(url); | |
1034 | state = s_req_http_start; | |
1035 | break; | |
1036 | case CR: | |
1037 | CALLBACK(url); | |
1038 | parser->http_major = 0; | |
1039 | parser->http_minor = 9; | |
1040 | state = s_req_line_almost_done; | |
1041 | break; | |
1042 | case LF: | |
1043 | CALLBACK(url); | |
1044 | parser->http_major = 0; | |
1045 | parser->http_minor = 9; | |
1046 | state = s_header_field_start; | |
1047 | break; | |
1048 | case '?': | |
1049 | case '#': | |
1050 | break; | |
1051 | default: | |
1052 | SET_ERRNO(HPE_INVALID_FRAGMENT); | |
1053 | goto error; | |
1054 | } | |
1055 | break; | |
1056 | } | |
1057 | ||
1058 | case s_req_http_start: | |
1059 | switch (ch) { | |
1060 | case 'H': | |
1061 | case 'R': | |
1062 | state = s_req_http_H; | |
1063 | break; | |
1064 | case ' ': | |
1065 | break; | |
1066 | default: | |
1067 | SET_ERRNO(HPE_INVALID_CONSTANT); | |
1068 | goto error; | |
1069 | } | |
1070 | break; | |
1071 | ||
1072 | case s_req_http_H: | |
1073 | STRICT_CHECK(ch != 'T'); | |
1074 | state = s_req_http_HT; | |
1075 | break; | |
1076 | ||
1077 | case s_req_http_HT: | |
1078 | STRICT_CHECK(ch != 'T'); | |
1079 | state = s_req_http_HTT; | |
1080 | break; | |
1081 | ||
1082 | case s_req_http_HTT: | |
1083 | STRICT_CHECK(ch != 'P'); | |
1084 | state = s_req_http_HTTP; | |
1085 | break; | |
1086 | ||
1087 | case s_req_http_HTTP: | |
1088 | STRICT_CHECK(ch != '/'); | |
1089 | state = s_req_first_http_major; | |
1090 | break; | |
1091 | ||
1092 | /* first digit of major HTTP version */ | |
1093 | case s_req_first_http_major: | |
1094 | if (ch < '1' || ch > '9') { | |
1095 | SET_ERRNO(HPE_INVALID_VERSION); | |
1096 | goto error; | |
1097 | } | |
1098 | ||
1099 | parser->http_major = ch - '0'; | |
1100 | state = s_req_http_major; | |
1101 | break; | |
1102 | ||
1103 | /* major HTTP version or dot */ | |
1104 | case s_req_http_major: | |
1105 | { | |
1106 | if (ch == '.') { | |
1107 | state = s_req_first_http_minor; | |
1108 | break; | |
1109 | } | |
1110 | ||
1111 | if (!IS_NUM(ch)) { | |
1112 | SET_ERRNO(HPE_INVALID_VERSION); | |
1113 | goto error; | |
1114 | } | |
1115 | ||
1116 | parser->http_major *= 10; | |
1117 | parser->http_major += ch - '0'; | |
1118 | ||
1119 | if (parser->http_major > 999) { | |
1120 | SET_ERRNO(HPE_INVALID_VERSION); | |
1121 | goto error; | |
1122 | } | |
1123 | ||
1124 | break; | |
1125 | } | |
1126 | ||
1127 | /* first digit of minor HTTP version */ | |
1128 | case s_req_first_http_minor: | |
1129 | if (!IS_NUM(ch)) { | |
1130 | SET_ERRNO(HPE_INVALID_VERSION); | |
1131 | goto error; | |
1132 | } | |
1133 | ||
1134 | parser->http_minor = ch - '0'; | |
1135 | state = s_req_http_minor; | |
1136 | break; | |
1137 | ||
1138 | /* minor HTTP version or end of request line */ | |
1139 | case s_req_http_minor: | |
1140 | { | |
1141 | if (ch == CR) { | |
1142 | state = s_req_line_almost_done; | |
1143 | break; | |
1144 | } | |
1145 | ||
1146 | if (ch == LF) { | |
1147 | state = s_header_field_start; | |
1148 | break; | |
1149 | } | |
1150 | ||
1151 | /* XXX allow spaces after digit? */ | |
1152 | ||
1153 | if (!IS_NUM(ch)) { | |
1154 | SET_ERRNO(HPE_INVALID_VERSION); | |
1155 | goto error; | |
1156 | } | |
1157 | ||
1158 | parser->http_minor *= 10; | |
1159 | parser->http_minor += ch - '0'; | |
1160 | ||
1161 | if (parser->http_minor > 999) { | |
1162 | SET_ERRNO(HPE_INVALID_VERSION); | |
1163 | goto error; | |
1164 | } | |
1165 | ||
1166 | break; | |
1167 | } | |
1168 | ||
1169 | /* end of request line */ | |
1170 | case s_req_line_almost_done: | |
1171 | { | |
1172 | if (ch != LF) { | |
1173 | SET_ERRNO(HPE_LF_EXPECTED); | |
1174 | goto error; | |
1175 | } | |
1176 | ||
1177 | state = s_header_field_start; | |
1178 | break; | |
1179 | } | |
1180 | ||
1181 | case s_header_field_start: | |
1182 | header_field_start: | |
1183 | { | |
1184 | if (ch == CR) { | |
1185 | state = s_headers_almost_done; | |
1186 | break; | |
1187 | } | |
1188 | ||
1189 | if (ch == LF) { | |
1190 | /* they might be just sending \n instead of \r\n so this would be | |
1191 | * the second \n to denote the end of headers*/ | |
1192 | state = s_headers_almost_done; | |
1193 | goto headers_almost_done; | |
1194 | } | |
1195 | ||
1196 | c = TOKEN(ch); | |
1197 | ||
1198 | if (!c) { | |
1199 | SET_ERRNO(HPE_INVALID_HEADER_TOKEN); | |
1200 | goto error; | |
1201 | } | |
1202 | ||
1203 | MARK(header_field); | |
1204 | ||
1205 | index = 0; | |
1206 | state = s_header_field; | |
1207 | ||
1208 | switch (c) { | |
1209 | case 'c': | |
1210 | header_state = h_C; | |
1211 | break; | |
1212 | ||
1213 | case 'p': | |
1214 | header_state = h_matching_proxy_connection; | |
1215 | break; | |
1216 | ||
1217 | case 't': | |
1218 | header_state = h_matching_transfer_encoding; | |
1219 | break; | |
1220 | ||
1221 | case 'u': | |
1222 | header_state = h_matching_upgrade; | |
1223 | break; | |
1224 | ||
1225 | default: | |
1226 | header_state = h_general; | |
1227 | break; | |
1228 | } | |
1229 | break; | |
1230 | } | |
1231 | ||
1232 | case s_header_field: | |
1233 | { | |
1234 | c = TOKEN(ch); | |
1235 | ||
1236 | if (c) { | |
1237 | switch (header_state) { | |
1238 | case h_general: | |
1239 | break; | |
1240 | ||
1241 | case h_C: | |
1242 | index++; | |
1243 | header_state = (c == 'o' ? h_CO : h_general); | |
1244 | break; | |
1245 | ||
1246 | case h_CO: | |
1247 | index++; | |
1248 | header_state = (c == 'n' ? h_CON : h_general); | |
1249 | break; | |
1250 | ||
1251 | case h_CON: | |
1252 | index++; | |
1253 | switch (c) { | |
1254 | case 'n': | |
1255 | header_state = h_matching_connection; | |
1256 | break; | |
1257 | case 't': | |
1258 | header_state = h_matching_content_length; | |
1259 | break; | |
1260 | default: | |
1261 | header_state = h_general; | |
1262 | break; | |
1263 | } | |
1264 | break; | |
1265 | ||
1266 | /* connection */ | |
1267 | ||
1268 | case h_matching_connection: | |
1269 | index++; | |
1270 | if (index > sizeof(CONNECTION)-1 | |
1271 | || c != CONNECTION[index]) { | |
1272 | header_state = h_general; | |
1273 | } else if (index == sizeof(CONNECTION)-2) { | |
1274 | header_state = h_connection; | |
1275 | } | |
1276 | break; | |
1277 | ||
1278 | /* proxy-connection */ | |
1279 | ||
1280 | case h_matching_proxy_connection: | |
1281 | index++; | |
1282 | if (index > sizeof(PROXY_CONNECTION)-1 | |
1283 | || c != PROXY_CONNECTION[index]) { | |
1284 | header_state = h_general; | |
1285 | } else if (index == sizeof(PROXY_CONNECTION)-2) { | |
1286 | header_state = h_connection; | |
1287 | } | |
1288 | break; | |
1289 | ||
1290 | /* content-length */ | |
1291 | ||
1292 | case h_matching_content_length: | |
1293 | index++; | |
1294 | if (index > sizeof(CONTENT_LENGTH)-1 | |
1295 | || c != CONTENT_LENGTH[index]) { | |
1296 | header_state = h_general; | |
1297 | } else if (index == sizeof(CONTENT_LENGTH)-2) { | |
1298 | header_state = h_content_length; | |
1299 | } | |
1300 | break; | |
1301 | ||
1302 | /* transfer-encoding */ | |
1303 | ||
1304 | case h_matching_transfer_encoding: | |
1305 | index++; | |
1306 | if (index > sizeof(TRANSFER_ENCODING)-1 | |
1307 | || c != TRANSFER_ENCODING[index]) { | |
1308 | header_state = h_general; | |
1309 | } else if (index == sizeof(TRANSFER_ENCODING)-2) { | |
1310 | header_state = h_transfer_encoding; | |
1311 | } | |
1312 | break; | |
1313 | ||
1314 | /* upgrade */ | |
1315 | ||
1316 | case h_matching_upgrade: | |
1317 | index++; | |
1318 | if (index > sizeof(UPGRADE)-1 | |
1319 | || c != UPGRADE[index]) { | |
1320 | header_state = h_general; | |
1321 | } else if (index == sizeof(UPGRADE)-2) { | |
1322 | header_state = h_upgrade; | |
1323 | } | |
1324 | break; | |
1325 | ||
1326 | case h_connection: | |
1327 | case h_content_length: | |
1328 | case h_transfer_encoding: | |
1329 | case h_upgrade: | |
1330 | if (ch != ' ') header_state = h_general; | |
1331 | break; | |
1332 | ||
1333 | default: | |
1334 | assert(0 && "Unknown header_state"); | |
1335 | break; | |
1336 | } | |
1337 | break; | |
1338 | } | |
1339 | ||
1340 | if (ch == ':') { | |
1341 | CALLBACK(header_field); | |
1342 | state = s_header_value_start; | |
1343 | break; | |
1344 | } | |
1345 | ||
1346 | if (ch == CR) { | |
1347 | state = s_header_almost_done; | |
1348 | CALLBACK(header_field); | |
1349 | break; | |
1350 | } | |
1351 | ||
1352 | if (ch == LF) { | |
1353 | CALLBACK(header_field); | |
1354 | state = s_header_field_start; | |
1355 | break; | |
1356 | } | |
1357 | ||
1358 | SET_ERRNO(HPE_INVALID_HEADER_TOKEN); | |
1359 | goto error; | |
1360 | } | |
1361 | ||
1362 | case s_header_value_start: | |
1363 | { | |
1364 | if (ch == ' ' || ch == '\t') break; | |
1365 | ||
1366 | MARK(header_value); | |
1367 | ||
1368 | state = s_header_value; | |
1369 | index = 0; | |
1370 | ||
1371 | if (ch == CR) { | |
1372 | CALLBACK(header_value); | |
1373 | header_state = h_general; | |
1374 | state = s_header_almost_done; | |
1375 | break; | |
1376 | } | |
1377 | ||
1378 | if (ch == LF) { | |
1379 | CALLBACK(header_value); | |
1380 | state = s_header_field_start; | |
1381 | break; | |
1382 | } | |
1383 | ||
1384 | c = LOWER(ch); | |
1385 | ||
1386 | switch (header_state) { | |
1387 | case h_upgrade: | |
1388 | parser->flags |= F_UPGRADE; | |
1389 | header_state = h_general; | |
1390 | break; | |
1391 | ||
1392 | case h_transfer_encoding: | |
1393 | /* looking for 'Transfer-Encoding: chunked' */ | |
1394 | if ('c' == c) { | |
1395 | header_state = h_matching_transfer_encoding_chunked; | |
1396 | } else { | |
1397 | header_state = h_general; | |
1398 | } | |
1399 | break; | |
1400 | ||
1401 | case h_content_length: | |
1402 | if (!IS_NUM(ch)) { | |
1403 | SET_ERRNO(HPE_INVALID_CONTENT_LENGTH); | |
1404 | goto error; | |
1405 | } | |
1406 | ||
1407 | parser->content_length = ch - '0'; | |
1408 | break; | |
1409 | ||
1410 | case h_connection: | |
1411 | /* looking for 'Connection: keep-alive' */ | |
1412 | if (c == 'k') { | |
1413 | header_state = h_matching_connection_keep_alive; | |
1414 | /* looking for 'Connection: close' */ | |
1415 | } else if (c == 'c') { | |
1416 | header_state = h_matching_connection_close; | |
1417 | } else { | |
1418 | header_state = h_general; | |
1419 | } | |
1420 | break; | |
1421 | ||
1422 | default: | |
1423 | header_state = h_general; | |
1424 | break; | |
1425 | } | |
1426 | break; | |
1427 | } | |
1428 | ||
1429 | case s_header_value: | |
1430 | { | |
1431 | ||
1432 | if (ch == CR) { | |
1433 | CALLBACK(header_value); | |
1434 | state = s_header_almost_done; | |
1435 | break; | |
1436 | } | |
1437 | ||
1438 | if (ch == LF) { | |
1439 | CALLBACK(header_value); | |
1440 | goto header_almost_done; | |
1441 | } | |
1442 | ||
1443 | c = LOWER(ch); | |
1444 | ||
1445 | switch (header_state) { | |
1446 | case h_general: | |
1447 | break; | |
1448 | ||
1449 | case h_connection: | |
1450 | case h_transfer_encoding: | |
1451 | assert(0 && "Shouldn't get here."); | |
1452 | break; | |
1453 | ||
1454 | case h_content_length: | |
1455 | if (ch == ' ') break; | |
1456 | if (!IS_NUM(ch)) { | |
1457 | SET_ERRNO(HPE_INVALID_CONTENT_LENGTH); | |
1458 | goto error; | |
1459 | } | |
1460 | ||
1461 | parser->content_length *= 10; | |
1462 | parser->content_length += ch - '0'; | |
1463 | break; | |
1464 | ||
1465 | /* Transfer-Encoding: chunked */ | |
1466 | case h_matching_transfer_encoding_chunked: | |
1467 | index++; | |
1468 | if (index > sizeof(CHUNKED)-1 | |
1469 | || c != CHUNKED[index]) { | |
1470 | header_state = h_general; | |
1471 | } else if (index == sizeof(CHUNKED)-2) { | |
1472 | header_state = h_transfer_encoding_chunked; | |
1473 | } | |
1474 | break; | |
1475 | ||
1476 | /* looking for 'Connection: keep-alive' */ | |
1477 | case h_matching_connection_keep_alive: | |
1478 | index++; | |
1479 | if (index > sizeof(KEEP_ALIVE)-1 | |
1480 | || c != KEEP_ALIVE[index]) { | |
1481 | header_state = h_general; | |
1482 | } else if (index == sizeof(KEEP_ALIVE)-2) { | |
1483 | header_state = h_connection_keep_alive; | |
1484 | } | |
1485 | break; | |
1486 | ||
1487 | /* looking for 'Connection: close' */ | |
1488 | case h_matching_connection_close: | |
1489 | index++; | |
1490 | if (index > sizeof(CLOSE)-1 || c != CLOSE[index]) { | |
1491 | header_state = h_general; | |
1492 | } else if (index == sizeof(CLOSE)-2) { | |
1493 | header_state = h_connection_close; | |
1494 | } | |
1495 | break; | |
1496 | ||
1497 | case h_transfer_encoding_chunked: | |
1498 | case h_connection_keep_alive: | |
1499 | case h_connection_close: | |
1500 | if (ch != ' ') header_state = h_general; | |
1501 | break; | |
1502 | ||
1503 | default: | |
1504 | state = s_header_value; | |
1505 | header_state = h_general; | |
1506 | break; | |
1507 | } | |
1508 | break; | |
1509 | } | |
1510 | ||
1511 | case s_header_almost_done: | |
1512 | header_almost_done: | |
1513 | { | |
1514 | STRICT_CHECK(ch != LF); | |
1515 | ||
1516 | state = s_header_value_lws; | |
1517 | ||
1518 | switch (header_state) { | |
1519 | case h_connection_keep_alive: | |
1520 | parser->flags |= F_CONNECTION_KEEP_ALIVE; | |
1521 | break; | |
1522 | case h_connection_close: | |
1523 | parser->flags |= F_CONNECTION_CLOSE; | |
1524 | break; | |
1525 | case h_transfer_encoding_chunked: | |
1526 | parser->flags |= F_CHUNKED; | |
1527 | break; | |
1528 | default: | |
1529 | break; | |
1530 | } | |
1531 | break; | |
1532 | } | |
1533 | ||
1534 | case s_header_value_lws: | |
1535 | { | |
1536 | if (ch == ' ' || ch == '\t') | |
1537 | state = s_header_value_start; | |
1538 | else | |
1539 | { | |
1540 | state = s_header_field_start; | |
1541 | goto header_field_start; | |
1542 | } | |
1543 | break; | |
1544 | } | |
1545 | ||
1546 | case s_headers_almost_done: | |
1547 | headers_almost_done: | |
1548 | { | |
1549 | STRICT_CHECK(ch != LF); | |
1550 | ||
1551 | if (parser->flags & F_TRAILING) { | |
1552 | /* End of a chunked request */ | |
1553 | CALLBACK2(message_complete); | |
1554 | state = NEW_MESSAGE(); | |
1555 | break; | |
1556 | } | |
1557 | ||
1558 | nread = 0; | |
1559 | ||
1560 | if (parser->flags & F_UPGRADE || parser->method == HTTP_CONNECT) { | |
1561 | parser->upgrade = 1; | |
1562 | } | |
1563 | ||
1564 | /* Here we call the headers_complete callback. This is somewhat | |
1565 | * different than other callbacks because if the user returns 1, we | |
1566 | * will interpret that as saying that this message has no body. This | |
1567 | * is needed for the annoying case of recieving a response to a HEAD | |
1568 | * request. | |
1569 | */ | |
1570 | if (settings->on_headers_complete) { | |
1571 | switch (settings->on_headers_complete(parser)) { | |
1572 | case 0: | |
1573 | break; | |
1574 | ||
1575 | case 1: | |
1576 | parser->flags |= F_SKIPBODY; | |
1577 | break; | |
1578 | ||
1579 | default: | |
1580 | parser->state = state; | |
1581 | SET_ERRNO(HPE_CB_headers_complete); | |
1582 | return p - data; /* Error */ | |
1583 | } | |
1584 | } | |
1585 | ||
1586 | /* Exit, the rest of the connect is in a different protocol. */ | |
1587 | if (parser->upgrade) { | |
1588 | CALLBACK2(message_complete); | |
1589 | return (p - data) + 1; | |
1590 | } | |
1591 | ||
1592 | if (parser->flags & F_SKIPBODY) { | |
1593 | CALLBACK2(message_complete); | |
1594 | state = NEW_MESSAGE(); | |
1595 | } else if (parser->flags & F_CHUNKED) { | |
1596 | /* chunked encoding - ignore Content-Length header */ | |
1597 | state = s_chunk_size_start; | |
1598 | } else { | |
1599 | if (parser->content_length == 0) { | |
1600 | /* Content-Length header given but zero: Content-Length: 0\r\n */ | |
1601 | CALLBACK2(message_complete); | |
1602 | state = NEW_MESSAGE(); | |
1603 | } else if (parser->content_length > 0) { | |
1604 | /* Content-Length header given and non-zero */ | |
1605 | state = s_body_identity; | |
1606 | } else { | |
1607 | if (parser->type == HTTP_REQUEST || http_should_keep_alive(parser)) { | |
1608 | /* Assume content-length 0 - read the next */ | |
1609 | CALLBACK2(message_complete); | |
1610 | state = NEW_MESSAGE(); | |
1611 | } else { | |
1612 | /* Read body until EOF */ | |
1613 | state = s_body_identity_eof; | |
1614 | } | |
1615 | } | |
1616 | } | |
1617 | ||
1618 | break; | |
1619 | } | |
1620 | ||
1621 | case s_body_identity: | |
1622 | to_read = MIN(pe - p, (int64_t)parser->content_length); | |
1623 | if (to_read > 0) { | |
1624 | if (settings->on_body) settings->on_body(parser, p, to_read); | |
1625 | p += to_read - 1; | |
1626 | parser->content_length -= to_read; | |
1627 | if (parser->content_length == 0) { | |
1628 | CALLBACK2(message_complete); | |
1629 | state = NEW_MESSAGE(); | |
1630 | } | |
1631 | } | |
1632 | break; | |
1633 | ||
1634 | /* read until EOF */ | |
1635 | case s_body_identity_eof: | |
1636 | to_read = pe - p; | |
1637 | if (to_read > 0) { | |
1638 | if (settings->on_body) settings->on_body(parser, p, to_read); | |
1639 | p += to_read - 1; | |
1640 | } | |
1641 | break; | |
1642 | ||
1643 | case s_chunk_size_start: | |
1644 | { | |
1645 | assert(nread == 1); | |
1646 | assert(parser->flags & F_CHUNKED); | |
1647 | ||
1648 | unhex_val = unhex[(unsigned char)ch]; | |
1649 | if (unhex_val == -1) { | |
1650 | SET_ERRNO(HPE_INVALID_CHUNK_SIZE); | |
1651 | goto error; | |
1652 | } | |
1653 | ||
1654 | parser->content_length = unhex_val; | |
1655 | state = s_chunk_size; | |
1656 | break; | |
1657 | } | |
1658 | ||
1659 | case s_chunk_size: | |
1660 | { | |
1661 | assert(parser->flags & F_CHUNKED); | |
1662 | ||
1663 | if (ch == CR) { | |
1664 | state = s_chunk_size_almost_done; | |
1665 | break; | |
1666 | } | |
1667 | ||
1668 | unhex_val = unhex[(unsigned char)ch]; | |
1669 | ||
1670 | if (unhex_val == -1) { | |
1671 | if (ch == ';' || ch == ' ') { | |
1672 | state = s_chunk_parameters; | |
1673 | break; | |
1674 | } | |
1675 | ||
1676 | SET_ERRNO(HPE_INVALID_CHUNK_SIZE); | |
1677 | goto error; | |
1678 | } | |
1679 | ||
1680 | parser->content_length *= 16; | |
1681 | parser->content_length += unhex_val; | |
1682 | break; | |
1683 | } | |
1684 | ||
1685 | case s_chunk_parameters: | |
1686 | { | |
1687 | assert(parser->flags & F_CHUNKED); | |
1688 | /* just ignore this shit. TODO check for overflow */ | |
1689 | if (ch == CR) { | |
1690 | state = s_chunk_size_almost_done; | |
1691 | break; | |
1692 | } | |
1693 | break; | |
1694 | } | |
1695 | ||
1696 | case s_chunk_size_almost_done: | |
1697 | { | |
1698 | assert(parser->flags & F_CHUNKED); | |
1699 | STRICT_CHECK(ch != LF); | |
1700 | ||
1701 | nread = 0; | |
1702 | ||
1703 | if (parser->content_length == 0) { | |
1704 | parser->flags |= F_TRAILING; | |
1705 | state = s_header_field_start; | |
1706 | } else { | |
1707 | state = s_chunk_data; | |
1708 | } | |
1709 | break; | |
1710 | } | |
1711 | ||
1712 | case s_chunk_data: | |
1713 | { | |
1714 | assert(parser->flags & F_CHUNKED); | |
1715 | ||
1716 | to_read = MIN(pe - p, (int64_t)(parser->content_length)); | |
1717 | ||
1718 | if (to_read > 0) { | |
1719 | if (settings->on_body) settings->on_body(parser, p, to_read); | |
1720 | p += to_read - 1; | |
1721 | } | |
1722 | ||
1723 | if (to_read == parser->content_length) { | |
1724 | state = s_chunk_data_almost_done; | |
1725 | } | |
1726 | ||
1727 | parser->content_length -= to_read; | |
1728 | break; | |
1729 | } | |
1730 | ||
1731 | case s_chunk_data_almost_done: | |
1732 | assert(parser->flags & F_CHUNKED); | |
1733 | STRICT_CHECK(ch != CR); | |
1734 | state = s_chunk_data_done; | |
1735 | break; | |
1736 | ||
1737 | case s_chunk_data_done: | |
1738 | assert(parser->flags & F_CHUNKED); | |
1739 | STRICT_CHECK(ch != LF); | |
1740 | state = s_chunk_size_start; | |
1741 | break; | |
1742 | ||
1743 | default: | |
1744 | assert(0 && "unhandled state"); | |
1745 | SET_ERRNO(HPE_INVALID_INTERNAL_STATE); | |
1746 | goto error; | |
1747 | } | |
1748 | } | |
1749 | ||
1750 | CALLBACK(header_field); | |
1751 | CALLBACK(header_value); | |
1752 | CALLBACK(url); | |
1753 | ||
1754 | parser->state = state; | |
1755 | parser->header_state = header_state; | |
1756 | parser->index = index; | |
1757 | parser->nread = nread; | |
1758 | ||
1759 | return len; | |
1760 | ||
1761 | error: | |
1762 | if (HTTP_PARSER_ERRNO(parser) == HPE_OK) { | |
1763 | SET_ERRNO(HPE_UNKNOWN); | |
1764 | } | |
1765 | ||
1766 | return (p - data); | |
1767 | } | |
1768 | ||
1769 | ||
1770 | int | |
1771 | http_should_keep_alive (http_parser *parser) | |
1772 | { | |
1773 | if (parser->http_major > 0 && parser->http_minor > 0) { | |
1774 | /* HTTP/1.1 */ | |
1775 | if (parser->flags & F_CONNECTION_CLOSE) { | |
1776 | return 0; | |
1777 | } else { | |
1778 | return 1; | |
1779 | } | |
1780 | } else { | |
1781 | /* HTTP/1.0 or earlier */ | |
1782 | if (parser->flags & F_CONNECTION_KEEP_ALIVE) { | |
1783 | return 1; | |
1784 | } else { | |
1785 | return 0; | |
1786 | } | |
1787 | } | |
1788 | } | |
1789 | ||
1790 | ||
1791 | const char * http_method_str (enum http_method m) | |
1792 | { | |
1793 | return method_strings[m]; | |
1794 | } | |
1795 | ||
1796 | ||
1797 | void | |
1798 | http_parser_init (http_parser *parser, enum http_parser_type t, char numerichost) | |
1799 | { | |
1800 | parser->numerichost = numerichost; | |
1801 | parser->type = t; | |
1802 | parser->state = (t == HTTP_REQUEST ? s_start_req : (t == HTTP_RESPONSE ? s_start_res : s_start_req_or_res)); | |
1803 | parser->nread = 0; | |
1804 | parser->upgrade = 0; | |
1805 | parser->flags = 0; | |
1806 | parser->method = 0; | |
1807 | parser->http_errno = 0; | |
1808 | } | |
1809 | ||
1810 | const char * | |
1811 | http_errno_name(enum http_errno err) { | |
1812 | assert(err < (sizeof(http_strerror_tab)/sizeof(http_strerror_tab[0]))); | |
1813 | return http_strerror_tab[err].name; | |
1814 | } | |
1815 | ||
1816 | const char * | |
1817 | http_errno_description(enum http_errno err) { | |
1818 | assert(err < (sizeof(http_strerror_tab)/sizeof(http_strerror_tab[0]))); | |
1819 | return http_strerror_tab[err].description; | |
1820 | } |