/* 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. */ #include "ryah_http_parser.h" #include #include #include #include #include #include #ifndef ULLONG_MAX # define ULLONG_MAX ((uint64_t) -1) /* 2^64-1 */ #endif #ifndef MIN # define MIN(a,b) ((a) < (b) ? (a) : (b)) #endif #if HTTP_PARSER_DEBUG #define SET_ERRNO(e) \ do { \ parser->http_errno = (e); \ parser->error_lineno = __LINE__; \ } while (0) #else #define SET_ERRNO(e) \ do { \ parser->http_errno = (e); \ } while(0) #endif /* Run the notify callback FOR, returning ER if it fails */ #define CALLBACK_NOTIFY_(FOR, ER) \ do { \ assert(HTTP_PARSER_ERRNO(parser) == HPE_OK); \ \ if (settings->on_##FOR) { \ if (0 != settings->on_##FOR(parser)) { \ SET_ERRNO(HPE_CB_##FOR); \ } \ \ /* We either errored above or got paused; get out */ \ if (HTTP_PARSER_ERRNO(parser) != HPE_OK) { \ return (ER); \ } \ } \ } while (0) /* Run the notify callback FOR and consume the current byte */ #define CALLBACK_NOTIFY(FOR) CALLBACK_NOTIFY_(FOR, p - data + 1) /* Run the notify callback FOR and don't consume the current byte */ #define CALLBACK_NOTIFY_NOADVANCE(FOR) CALLBACK_NOTIFY_(FOR, p - data) /* Run data callback FOR with LEN bytes, returning ER if it fails */ #define CALLBACK_DATA_(FOR, LEN, ER) \ do { \ assert(HTTP_PARSER_ERRNO(parser) == HPE_OK); \ \ if (FOR##_mark) { \ if (settings->on_##FOR) { \ if (0 != settings->on_##FOR(parser, FOR##_mark, (LEN))) { \ SET_ERRNO(HPE_CB_##FOR); \ } \ \ /* We either errored above or got paused; get out */ \ if (HTTP_PARSER_ERRNO(parser) != HPE_OK) { \ return (ER); \ } \ } \ FOR##_mark = NULL; \ } \ } while (0) /* 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) #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[] = { "DELETE" , "GET" , "HEAD" , "POST" , "PUT" , "CONNECT" , "OPTIONS" , "TRACE" , "COPY" , "LOCK" , "MKCOL" , "MOVE" , "PROPFIND" , "PROPPATCH" , "UNLOCK" , "REPORT" , "MKACTIVITY" , "CHECKOUT" , "MERGE" , "M-SEARCH" , "NOTIFY" , "SUBSCRIBE" , "UNSUBSCRIBE" , "PATCH" , "PURGE" }; /* Tokens as defined by rfc 2616. Also lowercases them. * token = 1* * 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 ' */ 0, '!', 0, '#', '$', '%', '&', '\'', /* 40 ( 41 ) 42 * 43 + 44 , 45 - 46 . 47 / */ 0, 0, '*', '+', 0, '-', '.', 0, /* 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 */ 'x', 'y', 'z', 0, '|', 0, '~', 0 }; 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 }; static const uint8_t normal_url_char[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 ' */ 0, 1, 1, 0, 1, 1, 1, 1, /* 40 ( 41 ) 42 * 43 + 44 , 45 - 46 . 47 / */ 1, 1, 1, 1, 1, 1, 1, 1, /* 48 0 49 1 50 2 51 3 52 4 53 5 54 6 55 7 */ 1, 1, 1, 1, 1, 1, 1, 1, /* 56 8 57 9 58 : 59 ; 60 < 61 = 62 > 63 ? */ 1, 1, 1, 1, 1, 1, 1, 0, /* 64 @ 65 A 66 B 67 C 68 D 69 E 70 F 71 G */ 1, 1, 1, 1, 1, 1, 1, 1, /* 72 H 73 I 74 J 75 K 76 L 77 M 78 N 79 O */ 1, 1, 1, 1, 1, 1, 1, 1, /* 80 P 81 Q 82 R 83 S 84 T 85 U 86 V 87 W */ 1, 1, 1, 1, 1, 1, 1, 1, /* 88 X 89 Y 90 Z 91 [ 92 \ 93 ] 94 ^ 95 _ */ 1, 1, 1, 1, 1, 1, 1, 1, /* 96 ` 97 a 98 b 99 c 100 d 101 e 102 f 103 g */ 1, 1, 1, 1, 1, 1, 1, 1, /* 104 h 105 i 106 j 107 k 108 l 109 m 110 n 111 o */ 1, 1, 1, 1, 1, 1, 1, 1, /* 112 p 113 q 114 r 115 s 116 t 117 u 118 v 119 w */ 1, 1, 1, 1, 1, 1, 1, 1, /* 120 x 121 y 122 z 123 { 124 | 125 } 126 ~ 127 del */ 1, 1, 1, 1, 1, 1, 1, 0, }; 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 , s_req_host_start , s_req_host_v6_start , s_req_host_v6 , s_req_host_v6_end , s_req_host , s_req_port_start , s_req_port , 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 , s_headers_almost_done , s_headers_done /* Important: 's_headers_done' must be the last 'header' state. All * 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 , s_message_done }; #define PARSING_HEADER(state) (state <= s_headers_done) 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 }; /* 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)) #define IS_HEX(c) (IS_NUM(c) || (LOWER(c) >= 'a' && LOWER(c) <= 'f')) #if HTTP_PARSER_STRICT #define TOKEN(c) (tokens[(unsigned char)c]) #define IS_URL_CHAR(c) (normal_url_char[(unsigned char) (c)]) #define IS_HOST_CHAR(c) (IS_ALPHANUM(c) || (c) == '.' || (c) == '-') #else #define TOKEN(c) ((c == ' ') ? ' ' : tokens[(unsigned char)c]) #define IS_URL_CHAR(c) \ (normal_url_char[(unsigned char) (c)] || ((c) & 0x80)) #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 int http_message_needs_eof(ryah_http_parser *parser); /* Our URL parser. * * This is designed to be shared by ryah_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 ryah_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) { assert(!isspace(ch)); 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_host_start; } break; case s_req_host_start: if (ch == '[') { return s_req_host_v6_start; } if (IS_HOST_CHAR(ch)) { return s_req_host; } break; case s_req_host: if (IS_HOST_CHAR(ch)) { return s_req_host; } /* FALLTHROUGH */ case s_req_host_v6_end: switch (ch) { case ':': return s_req_port_start; case '/': return s_req_path; case '?': return s_req_query_string_start; } break; case s_req_host_v6: if (ch == ']') { return s_req_host_v6_end; } /* FALLTHROUGH */ case s_req_host_v6_start: if (IS_HEX(ch) || ch == ':') { return s_req_host_v6; } break; case s_req_port: switch (ch) { case '/': return s_req_path; case '?': return s_req_query_string_start; } /* FALLTHROUGH */ case s_req_port_start: if (IS_NUM(ch)) { return s_req_port; } 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; } size_t ryah_http_parser_execute (ryah_http_parser *parser, const ryah_http_parser_settings *settings, const char *data, size_t len) { char c, ch; int8_t unhex_val; 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; /* We're in an error state. Don't bother doing anything. */ if (HTTP_PARSER_ERRNO(parser) != HPE_OK) { return 0; } if (len == 0) { switch (parser->state) { case s_body_identity_eof: /* Use of CALLBACK_NOTIFY() here would erroneously return 1 byte read if * we got paused. */ CALLBACK_NOTIFY_NOADVANCE(message_complete); 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; } } if (parser->state == s_header_field) header_field_mark = data; if (parser->state == s_header_value) header_value_mark = data; 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_host_start: case s_req_host_v6_start: case s_req_host_v6: case s_req_host_v6_end: case s_req_host: case s_req_port_start: case s_req_port: case s_req_query_string_start: case s_req_query_string: case s_req_fragment_start: case s_req_fragment: url_mark = data; break; } for (p=data; p != data + len; p++) { ch = *p; if (PARSING_HEADER(parser->state)) { ++parser->nread; /* Buffer overflow attack */ if (parser->nread > HTTP_MAX_HEADER_SIZE) { SET_ERRNO(HPE_HEADER_OVERFLOW); goto error; } } reexecute_byte: switch (parser->state) { case s_dead: /* this state is used after a 'Connection: close' message * the parser will error out if it reads another message */ if (ch == CR || ch == LF) break; SET_ERRNO(HPE_CLOSED_CONNECTION); goto error; case s_start_req_or_res: { if (ch == CR || ch == LF) break; parser->flags = 0; parser->content_length = ULLONG_MAX; if (ch == 'H') { parser->state = s_res_or_resp_H; CALLBACK_NOTIFY(message_begin); } else { parser->type = HTTP_REQUEST; parser->state = s_start_req; goto reexecute_byte; } break; } case s_res_or_resp_H: if (ch == 'T') { parser->type = HTTP_RESPONSE; parser->state = s_res_HT; } else { if (ch != 'E') { SET_ERRNO(HPE_INVALID_CONSTANT); goto error; } parser->type = HTTP_REQUEST; parser->method = HTTP_HEAD; parser->index = 2; parser->state = s_req_method; } break; case s_start_res: { parser->flags = 0; parser->content_length = ULLONG_MAX; switch (ch) { case 'H': parser->state = s_res_H; break; case CR: case LF: break; default: SET_ERRNO(HPE_INVALID_CONSTANT); goto error; } CALLBACK_NOTIFY(message_begin); break; } case s_res_H: STRICT_CHECK(ch != 'T'); parser->state = s_res_HT; break; case s_res_HT: STRICT_CHECK(ch != 'T'); parser->state = s_res_HTT; break; case s_res_HTT: STRICT_CHECK(ch != 'P'); parser->state = s_res_HTTP; break; case s_res_HTTP: STRICT_CHECK(ch != '/'); parser->state = s_res_first_http_major; break; case s_res_first_http_major: if (ch < '0' || ch > '9') { SET_ERRNO(HPE_INVALID_VERSION); goto error; } parser->http_major = ch - '0'; parser->state = s_res_http_major; break; /* major HTTP version or dot */ case s_res_http_major: { if (ch == '.') { parser->state = s_res_first_http_minor; 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'; parser->state = s_res_http_minor; break; /* minor HTTP version or end of request line */ case s_res_http_minor: { if (ch == ' ') { parser->state = s_res_first_status_code; 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'; parser->state = s_res_status_code; break; } case s_res_status_code: { if (!IS_NUM(ch)) { switch (ch) { case ' ': parser->state = s_res_status; break; case CR: parser->state = s_res_line_almost_done; break; case LF: parser->state = s_header_field_start; 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) { parser->state = s_res_line_almost_done; break; } if (ch == LF) { parser->state = s_header_field_start; break; } break; case s_res_line_almost_done: STRICT_CHECK(ch != LF); parser->state = s_header_field_start; break; case s_start_req: { if (ch == CR || ch == LF) break; parser->flags = 0; parser->content_length = ULLONG_MAX; if (!IS_ALPHA(ch)) { SET_ERRNO(HPE_INVALID_METHOD); goto error; } parser->method = (enum http_method) 0; parser->index = 1; 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; /* or PROPFIND|PROPPATCH|PUT|PATCH|PURGE */ break; case 'R': parser->method = HTTP_REPORT; break; case 'S': parser->method = HTTP_SUBSCRIBE; break; case 'T': parser->method = HTTP_TRACE; break; case 'U': parser->method = HTTP_UNLOCK; /* or UNSUBSCRIBE */ break; default: SET_ERRNO(HPE_INVALID_METHOD); goto error; } parser->state = s_req_method; CALLBACK_NOTIFY(message_begin); break; } case s_req_method: { const char *matcher; if (ch == '\0') { SET_ERRNO(HPE_INVALID_METHOD); goto error; } matcher = method_strings[parser->method]; if (ch == ' ' && matcher[parser->index] == '\0') { parser->state = s_req_spaces_before_url; } else if (ch == matcher[parser->index]) { ; /* nada */ } else if (parser->method == HTTP_CONNECT) { if (parser->index == 1 && ch == 'H') { parser->method = HTTP_CHECKOUT; } else if (parser->index == 2 && ch == 'P') { parser->method = HTTP_COPY; } else { goto error; } } else if (parser->method == HTTP_MKCOL) { if (parser->index == 1 && ch == 'O') { parser->method = HTTP_MOVE; } else if (parser->index == 1 && ch == 'E') { parser->method = HTTP_MERGE; } else if (parser->index == 1 && ch == '-') { parser->method = HTTP_MSEARCH; } else if (parser->index == 2 && ch == 'A') { parser->method = HTTP_MKACTIVITY; } else { goto error; } } else if (parser->index == 1 && parser->method == HTTP_POST) { if (ch == 'R') { parser->method = HTTP_PROPFIND; /* or HTTP_PROPPATCH */ } else if (ch == 'U') { parser->method = HTTP_PUT; /* or HTTP_PURGE */ } else if (ch == 'A') { parser->method = HTTP_PATCH; } else { goto error; } } 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') { parser->method = HTTP_PROPPATCH; } else { SET_ERRNO(HPE_INVALID_METHOD); goto error; } ++parser->index; break; } case s_req_spaces_before_url: { if (ch == ' ') break; MARK(url); if (parser->method == HTTP_CONNECT) { parser->state = s_req_host_start; } parser->state = parse_url_char((enum state)parser->state, ch); if (parser->state == s_dead) { SET_ERRNO(HPE_INVALID_URL); goto error; } break; } case s_req_schema: case s_req_schema_slash: case s_req_schema_slash_slash: case s_req_host_start: case s_req_host_v6_start: case s_req_host_v6: case s_req_port_start: { switch (ch) { /* No whitespace allowed here */ case ' ': case CR: case LF: SET_ERRNO(HPE_INVALID_URL); goto error; default: parser->state = parse_url_char((enum state)parser->state, ch); if (parser->state == s_dead) { SET_ERRNO(HPE_INVALID_URL); goto error; } } break; } case s_req_host: case s_req_host_v6_end: case s_req_port: case s_req_path: case s_req_query_string_start: case s_req_query_string: case s_req_fragment_start: case s_req_fragment: { switch (ch) { case ' ': parser->state = s_req_http_start; CALLBACK_DATA(url); break; case CR: case LF: parser->http_major = 0; parser->http_minor = 9; parser->state = (ch == CR) ? s_req_line_almost_done : s_header_field_start; CALLBACK_DATA(url); break; default: parser->state = parse_url_char((enum state)parser->state, ch); if (parser->state == s_dead) { SET_ERRNO(HPE_INVALID_URL); goto error; } } break; } case s_req_http_start: switch (ch) { case 'H': parser->state = s_req_http_H; break; case ' ': break; default: SET_ERRNO(HPE_INVALID_CONSTANT); goto error; } break; case s_req_http_H: STRICT_CHECK(ch != 'T'); parser->state = s_req_http_HT; break; case s_req_http_HT: STRICT_CHECK(ch != 'T'); parser->state = s_req_http_HTT; break; case s_req_http_HTT: STRICT_CHECK(ch != 'P'); parser->state = s_req_http_HTTP; break; case s_req_http_HTTP: STRICT_CHECK(ch != '/'); parser->state = s_req_first_http_major; 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'; parser->state = s_req_http_major; break; /* major HTTP version or dot */ case s_req_http_major: { if (ch == '.') { parser->state = s_req_first_http_minor; 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'; parser->state = s_req_http_minor; break; /* minor HTTP version or end of request line */ case s_req_http_minor: { if (ch == CR) { parser->state = s_req_line_almost_done; break; } if (ch == LF) { parser->state = s_header_field_start; 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; } parser->state = s_header_field_start; break; } case s_header_field_start: { if (ch == CR) { parser->state = s_headers_almost_done; 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*/ parser->state = s_headers_almost_done; goto reexecute_byte; } c = TOKEN(ch); if (!c) { SET_ERRNO(HPE_INVALID_HEADER_TOKEN); goto error; } MARK(header_field); parser->index = 0; parser->state = s_header_field; switch (c) { case 'c': parser->header_state = h_C; break; case 'p': parser->header_state = h_matching_proxy_connection; break; case 't': parser->header_state = h_matching_transfer_encoding; break; case 'u': parser->header_state = h_matching_upgrade; break; default: parser->header_state = h_general; break; } break; } case s_header_field: { c = TOKEN(ch); if (c) { switch (parser->header_state) { case h_general: break; case h_C: parser->index++; parser->header_state = (c == 'o' ? h_CO : h_general); break; case h_CO: parser->index++; parser->header_state = (c == 'n' ? h_CON : h_general); break; case h_CON: parser->index++; switch (c) { case 'n': parser->header_state = h_matching_connection; break; case 't': parser->header_state = h_matching_content_length; break; default: parser->header_state = h_general; break; } break; /* connection */ case h_matching_connection: 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; } break; /* proxy-connection */ case h_matching_proxy_connection: 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; } break; /* content-length */ case h_matching_content_length: 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; } break; /* transfer-encoding */ case h_matching_transfer_encoding: 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; } break; /* upgrade */ case h_matching_upgrade: 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; } break; case h_connection: case h_content_length: case h_transfer_encoding: case h_upgrade: if (ch != ' ') parser->header_state = h_general; break; default: assert(0 && "Unknown header_state"); break; } break; } if (ch == ':') { parser->state = s_header_value_start; CALLBACK_DATA(header_field); break; } if (ch == CR) { parser->state = s_header_almost_done; CALLBACK_DATA(header_field); break; } if (ch == LF) { parser->state = s_header_field_start; CALLBACK_DATA(header_field); break; } SET_ERRNO(HPE_INVALID_HEADER_TOKEN); goto error; } case s_header_value_start: { if (ch == ' ' || ch == '\t') break; MARK(header_value); parser->state = s_header_value; parser->index = 0; if (ch == CR) { parser->header_state = h_general; parser->state = s_header_almost_done; CALLBACK_DATA(header_value); break; } if (ch == LF) { parser->state = s_header_field_start; CALLBACK_DATA(header_value); break; } c = LOWER(ch); switch (parser->header_state) { case h_upgrade: parser->flags |= F_UPGRADE; parser->header_state = h_general; break; case h_transfer_encoding: /* looking for 'Transfer-Encoding: chunked' */ if ('c' == c) { parser->header_state = h_matching_transfer_encoding_chunked; } else { parser->header_state = h_general; } 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') { parser->header_state = h_matching_connection_keep_alive; /* looking for 'Connection: close' */ } else if (c == 'c') { parser->header_state = h_matching_connection_close; } else { parser->header_state = h_general; } break; default: parser->header_state = h_general; break; } break; } case s_header_value: { if (ch == CR) { parser->state = s_header_almost_done; CALLBACK_DATA(header_value); break; } if (ch == LF) { parser->state = s_header_almost_done; CALLBACK_DATA_NOADVANCE(header_value); goto reexecute_byte; } c = LOWER(ch); switch (parser->header_state) { case h_general: break; case h_connection: case h_transfer_encoding: assert(0 && "Shouldn't get here."); break; case h_content_length: { uint64_t t; if (ch == ' ') break; if (!IS_NUM(ch)) { SET_ERRNO(HPE_INVALID_CONTENT_LENGTH); goto error; } 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; break; } /* Transfer-Encoding: chunked */ case h_matching_transfer_encoding_chunked: 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; } break; /* looking for 'Connection: keep-alive' */ case h_matching_connection_keep_alive: 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; } break; /* looking for 'Connection: close' */ case h_matching_connection_close: 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; } break; case h_transfer_encoding_chunked: case h_connection_keep_alive: case h_connection_close: if (ch != ' ') parser->header_state = h_general; break; default: parser->state = s_header_value; parser->header_state = h_general; break; } break; } case s_header_almost_done: { STRICT_CHECK(ch != LF); parser->state = s_header_value_lws; switch (parser->header_state) { 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; } break; } case s_header_value_lws: { if (ch == ' ' || ch == '\t') parser->state = s_header_value_start; else { parser->state = s_header_field_start; goto reexecute_byte; } break; } case s_headers_almost_done: { STRICT_CHECK(ch != LF); if (parser->flags & F_TRAILING) { /* End of a chunked request */ parser->state = NEW_MESSAGE(); CALLBACK_NOTIFY(message_complete); break; } parser->state = s_headers_done; /* Set this here so that on_headers_complete() callbacks can see it */ parser->upgrade = (parser->flags & F_UPGRADE || parser->method == HTTP_CONNECT); /* 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. * * 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. */ 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 */ } } if (HTTP_PARSER_ERRNO(parser) != HPE_OK) { return p - data; } goto reexecute_byte; } case s_headers_done: { STRICT_CHECK(ch != LF); parser->nread = 0; /* Exit, the rest of the connect is in a different protocol. */ if (parser->upgrade) { parser->state = NEW_MESSAGE(); CALLBACK_NOTIFY(message_complete); return (p - data) + 1; } if (parser->flags & F_SKIPBODY) { parser->state = NEW_MESSAGE(); CALLBACK_NOTIFY(message_complete); } else if (parser->flags & F_CHUNKED) { /* chunked encoding - ignore Content-Length header */ parser->state = s_chunk_size_start; } else { if (parser->content_length == 0) { /* Content-Length header given but zero: Content-Length: 0\r\n */ parser->state = NEW_MESSAGE(); CALLBACK_NOTIFY(message_complete); } else if (parser->content_length != ULLONG_MAX) { /* Content-Length header given and non-zero */ parser->state = s_body_identity; } else { if (parser->type == HTTP_REQUEST || !http_message_needs_eof(parser)) { /* Assume content-length 0 - read the next */ parser->state = NEW_MESSAGE(); CALLBACK_NOTIFY(message_complete); } else { /* Read body until EOF */ parser->state = s_body_identity_eof; } } } break; } case s_body_identity: { 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; } break; } /* read until EOF */ case s_body_identity_eof: MARK(body); p = data + len - 1; break; case s_message_done: parser->state = NEW_MESSAGE(); CALLBACK_NOTIFY(message_complete); break; case s_chunk_size_start: { assert(parser->nread == 1); 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; parser->state = s_chunk_size; break; } case s_chunk_size: { uint64_t t; assert(parser->flags & F_CHUNKED); if (ch == CR) { parser->state = s_chunk_size_almost_done; break; } unhex_val = unhex[(unsigned char)ch]; if (unhex_val == -1) { if (ch == ';' || ch == ' ') { parser->state = s_chunk_parameters; break; } SET_ERRNO(HPE_INVALID_CHUNK_SIZE); goto error; } 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; break; } case s_chunk_parameters: { assert(parser->flags & F_CHUNKED); /* just ignore this shit. TODO check for overflow */ if (ch == CR) { parser->state = s_chunk_size_almost_done; break; } break; } case s_chunk_size_almost_done: { assert(parser->flags & F_CHUNKED); STRICT_CHECK(ch != LF); parser->nread = 0; if (parser->content_length == 0) { parser->flags |= F_TRAILING; parser->state = s_header_field_start; } else { parser->state = s_chunk_data; } break; } case s_chunk_data: { uint64_t to_read = MIN(parser->content_length, (uint64_t) ((data + len) - p)); assert(parser->flags & F_CHUNKED); assert(parser->content_length != 0 && parser->content_length != ULLONG_MAX); /* 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; if (parser->content_length == 0) { parser->state = s_chunk_data_almost_done; } break; } case s_chunk_data_almost_done: assert(parser->flags & F_CHUNKED); assert(parser->content_length == 0); STRICT_CHECK(ch != CR); parser->state = s_chunk_data_done; CALLBACK_DATA(body); break; case s_chunk_data_done: assert(parser->flags & F_CHUNKED); STRICT_CHECK(ch != LF); parser->nread = 0; parser->state = s_chunk_size_start; break; default: assert(0 && "unhandled state"); SET_ERRNO(HPE_INVALID_INTERNAL_STATE); goto error; } } /* 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); CALLBACK_DATA_NOADVANCE(header_field); CALLBACK_DATA_NOADVANCE(header_value); CALLBACK_DATA_NOADVANCE(url); CALLBACK_DATA_NOADVANCE(body); return len; error: if (HTTP_PARSER_ERRNO(parser) == HPE_OK) { SET_ERRNO(HPE_UNKNOWN); } return (p - data); } /* Does the parser need to see an EOF to find the end of the message? */ int http_message_needs_eof (ryah_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 (ryah_http_parser *parser) { 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 */ if (!(parser->flags & F_CONNECTION_KEEP_ALIVE)) { return 0; } } return !http_message_needs_eof(parser); } const char * http_method_str (enum http_method m) { return method_strings[m]; } void ryah_http_parser_init (ryah_http_parser *parser, enum ryah_http_parser_type t) { void *data = parser->data; /* preserve application data */ memset(parser, 0, sizeof(*parser)); parser->data = data; parser->type = t; parser->state = (t == HTTP_REQUEST ? s_start_req : (t == HTTP_RESPONSE ? s_start_res : s_start_req_or_res)); parser->http_errno = HPE_OK; } 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; } int ryah_http_parser_parse_url(const char *buf, size_t buflen, int is_connect, struct ryah_http_parser_url *u) { enum state s; const char *p; enum ryah_http_parser_url_fields uf, old_uf; u->port = u->field_set = 0; s = is_connect ? s_req_host_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_host_start: case s_req_host_v6_start: case s_req_host_v6_end: case s_req_port_start: case s_req_query_string_start: case s_req_fragment_start: continue; case s_req_schema: uf = UF_SCHEMA; break; case s_req_host: case s_req_host_v6: uf = UF_HOST; break; case s_req_port: uf = UF_PORT; 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; } /* CONNECT requests can only contain "hostname:port" */ if (is_connect && u->field_set != ((1 << UF_HOST)|(1 << UF_PORT))) { return 1; } /* Make sure we don't end somewhere unexpected */ switch (s) { case s_req_host_v6_start: case s_req_host_v6: case s_req_host_v6_end: case s_req_host: case s_req_port_start: return 1; default: break; } 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 ryah_http_parser_pause(ryah_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"); } }