ngx_int_t ngx_http_parse_request_line(ngx_http_request_t *r, ngx_buf_t *b) { u_char c, ch, *p, *m; enum { sw_start = 0, sw_method, sw_spaces_before_uri, sw_schema, sw_schema_slash, sw_schema_slash_slash, sw_host, sw_port, sw_after_slash_in_uri, sw_check_uri, sw_uri, sw_http_09, sw_http_H, sw_http_HT, sw_http_HTT, sw_http_HTTP, sw_first_major_digit, sw_major_digit, sw_first_minor_digit, sw_minor_digit, sw_spaces_after_digit, sw_almost_done } state; state = r->state; for (p = b->pos; p < b->last; p++) { ch = *p; switch (state) { /* HTTP methods: GET, HEAD, POST */ case sw_start: r->request_start = p; if (ch == CR || ch == LF) { break; } if (ch < 'A' || ch > 'Z') { return NGX_HTTP_PARSE_INVALID_METHOD; } state = sw_method; break; case sw_method: if (ch == ' ') { r->method_end = p - 1; m = r->request_start; switch (p - m) { case 3: if (ngx_str3_cmp(m, 'G', 'E', 'T', ' ')) { r->method = NGX_HTTP_GET; break; } if (ngx_str3_cmp(m, 'P', 'U', 'T', ' ')) { r->method = NGX_HTTP_PUT; break; } break; case 4: if (m[1] == 'O') { if (ngx_str3Ocmp(m, 'P', 'O', 'S', 'T')) { r->method = NGX_HTTP_POST; break; } if (ngx_str3Ocmp(m, 'C', 'O', 'P', 'Y')) { r->method = NGX_HTTP_COPY; break; } if (ngx_str3Ocmp(m, 'M', 'O', 'V', 'E')) { r->method = NGX_HTTP_MOVE; break; } if (ngx_str3Ocmp(m, 'L', 'O', 'C', 'K')) { r->method = NGX_HTTP_LOCK; break; } } else { if (ngx_str4cmp(m, 'H', 'E', 'A', 'D')) { r->method = NGX_HTTP_HEAD; break; } } break; case 5: if (ngx_str5cmp(m, 'M', 'K', 'C', 'O', 'L')) { r->method = NGX_HTTP_MKCOL; } if (ngx_str5cmp(m, 'T', 'R', 'A', 'C', 'E')) { r->method = NGX_HTTP_TRACE; } break; case 6: if (ngx_str6cmp(m, 'D', 'E', 'L', 'E', 'T', 'E')) { r->method = NGX_HTTP_DELETE; break; } if (ngx_str6cmp(m, 'U', 'N', 'L', 'O', 'C', 'K')) { r->method = NGX_HTTP_UNLOCK; break; } break; case 7: if (ngx_str7_cmp(m, 'O', 'P', 'T', 'I', 'O', 'N', 'S', ' ')) { r->method = NGX_HTTP_OPTIONS; } break; case 8: if (ngx_str8cmp(m, 'P', 'R', 'O', 'P', 'F', 'I', 'N', 'D')) { r->method = NGX_HTTP_PROPFIND; } break; case 9: if (ngx_str9cmp(m, 'P', 'R', 'O', 'P', 'P', 'A', 'T', 'C', 'H')) { r->method = NGX_HTTP_PROPPATCH; } break; } state = sw_spaces_before_uri; break; } if (ch < 'A' || ch > 'Z') { return NGX_HTTP_PARSE_INVALID_METHOD; } break; /* space* before URI */ case sw_spaces_before_uri: if (ch == '/' ){ r->uri_start = p; state = sw_after_slash_in_uri; break; } c = (u_char) (ch | 0x20); if (c >= 'a' && c <= 'z') { r->schema_start = p; state = sw_schema; break; } switch (ch) { case ' ': break; default: return NGX_HTTP_PARSE_INVALID_REQUEST; } break; case sw_schema: c = (u_char) (ch | 0x20); if (c >= 'a' && c <= 'z') { break; } switch (ch) { case ':': r->schema_end = p; state = sw_schema_slash; break; default: return NGX_HTTP_PARSE_INVALID_REQUEST; } break; case sw_schema_slash: switch (ch) { case '/': state = sw_schema_slash_slash; break; default: return NGX_HTTP_PARSE_INVALID_REQUEST; } break; case sw_schema_slash_slash: switch (ch) { case '/': r->host_start = p + 1; state = sw_host; break; default: return NGX_HTTP_PARSE_INVALID_REQUEST; } break; case sw_host: c = (u_char) (ch | 0x20); if (c >= 'a' && c <= 'z') { break; } if ((ch >= '0' && ch <= '9') || ch == '.' || ch == '-') { break; } r->host_end = p; switch (ch) { case ':': state = sw_port; break; case '/': r->uri_start = p; state = sw_after_slash_in_uri; break; case ' ': /* * use single "/" from request line to preserve pointers, * if request line will be copied to large client buffer */ r->uri_start = r->schema_end + 1; r->uri_end = r->schema_end + 2; state = sw_http_09; break; default: return NGX_HTTP_PARSE_INVALID_REQUEST; } break; case sw_port: if (ch >= '0' && ch <= '9') { break; } switch (ch) { case '/': r->port_end = p; r->uri_start = p; state = sw_after_slash_in_uri; break; case ' ': r->port_end = p; /* * use single "/" from request line to preserve pointers, * if request line will be copied to large client buffer */ r->uri_start = r->schema_end + 1; r->uri_end = r->schema_end + 2; state = sw_http_09; break; default: return NGX_HTTP_PARSE_INVALID_REQUEST; } break; /* check "/.", "//", "%", and "\" (Win32) in URI */ case sw_after_slash_in_uri: if (usual[ch >> 5] & (1 << (ch & 0x1f))) { state = sw_check_uri; break; } switch (ch) { case ' ': r->uri_end = p; state = sw_http_09; break; case CR: r->uri_end = p; r->http_minor = 9; state = sw_almost_done; break; case LF: r->uri_end = p; r->http_minor = 9; goto done; case '.': r->complex_uri = 1; state = sw_uri; break; case '%': r->quoted_uri = 1; state = sw_uri; break; case '/': r->complex_uri = 1; state = sw_uri; break; #if (NGX_WIN32) case '\\': r->complex_uri = 1; state = sw_uri; break; #endif case '?': r->args_start = p + 1; state = sw_uri; break; case '#': r->complex_uri = 1; state = sw_uri; break; case '+': r->plus_in_uri = 1; break; case '\0': r->zero_in_uri = 1; break; default: state = sw_check_uri; break; } break; /* check "/", "%" and "\" (Win32) in URI */ case sw_check_uri: if (usual[ch >> 5] & (1 << (ch & 0x1f))) { break; } switch (ch) { case '/': r->uri_ext = NULL; state = sw_after_slash_in_uri; break; case '.': r->uri_ext = p + 1; break; case ' ': r->uri_end = p; state = sw_http_09; break; case CR: r->uri_end = p; r->http_minor = 9; state = sw_almost_done; break; case LF: r->uri_end = p; r->http_minor = 9; goto done; #if (NGX_WIN32) case '\\': r->complex_uri = 1; state = sw_after_slash_in_uri; break; #endif case '%': r->quoted_uri = 1; state = sw_uri; break; case '?': r->args_start = p + 1; state = sw_uri; break; case '#': r->complex_uri = 1; state = sw_uri; break; case '+': r->plus_in_uri = 1; break; case '\0': r->zero_in_uri = 1; break; } break; /* URI */ case sw_uri: if (usual[ch >> 5] & (1 << (ch & 0x1f))) { break; } switch (ch) { case ' ': r->uri_end = p; state = sw_http_09; break; case CR: r->uri_end = p; r->http_minor = 9; state = sw_almost_done; break; case LF: r->uri_end = p; r->http_minor = 9; goto done; case '#': r->complex_uri = 1; break; case '\0': r->zero_in_uri = 1; break; } break; /* space+ after URI */ case sw_http_09: switch (ch) { case ' ': break; case CR: r->http_minor = 9; state = sw_almost_done; break; case LF: r->http_minor = 9; goto done; case 'H': r->http_protocol.data = p; state = sw_http_H; break; default: return NGX_HTTP_PARSE_INVALID_REQUEST; } break; case sw_http_H: switch (ch) { case 'T': state = sw_http_HT; break; default: return NGX_HTTP_PARSE_INVALID_REQUEST; } break; case sw_http_HT: switch (ch) { case 'T': state = sw_http_HTT; break; default: return NGX_HTTP_PARSE_INVALID_REQUEST; } break; case sw_http_HTT: switch (ch) { case 'P': state = sw_http_HTTP; break; default: return NGX_HTTP_PARSE_INVALID_REQUEST; } break; case sw_http_HTTP: switch (ch) { case '/': state = sw_first_major_digit; break; default: return NGX_HTTP_PARSE_INVALID_REQUEST; } break; /* first digit of major HTTP version */ case sw_first_major_digit: if (ch < '1' || ch > '9') { return NGX_HTTP_PARSE_INVALID_REQUEST; } r->http_major = ch - '0'; state = sw_major_digit; break; /* major HTTP version or dot */ case sw_major_digit: if (ch == '.') { state = sw_first_minor_digit; break; } if (ch < '0' || ch > '9') { return NGX_HTTP_PARSE_INVALID_REQUEST; } r->http_major = r->http_major * 10 + ch - '0'; break; /* first digit of minor HTTP version */ case sw_first_minor_digit: if (ch < '0' || ch > '9') { return NGX_HTTP_PARSE_INVALID_REQUEST; } r->http_minor = ch - '0'; state = sw_minor_digit; break; /* minor HTTP version or end of request line */ case sw_minor_digit: if (ch == CR) { state = sw_almost_done; break; } if (ch == LF) { goto done; } if (ch == ' ') { state = sw_spaces_after_digit; break; } if (ch < '0' || ch > '9') { return NGX_HTTP_PARSE_INVALID_REQUEST; } r->http_minor = r->http_minor * 10 + ch - '0'; break; case sw_spaces_after_digit: switch (ch) { case ' ': break; case CR: state = sw_almost_done; break; case LF: goto done; default: return NGX_HTTP_PARSE_INVALID_REQUEST; } break; /* end of request line */ case sw_almost_done: r->request_end = p - 1; switch (ch) { case LF: goto done; default: return NGX_HTTP_PARSE_INVALID_REQUEST; } } } b->pos = p; r->state = state; return NGX_AGAIN; done: b->pos = p + 1; if (r->request_end == NULL) { r->request_end = p; } r->http_version = r->http_major * 1000 + r->http_minor; r->state = sw_start; if (r->http_version == 9 && r->method != NGX_HTTP_GET) { return NGX_HTTP_PARSE_INVALID_09_METHOD; } return NGX_OK; }
char * ngx_http_upstream_drizzle_keepalive(ngx_conf_t *cf, ngx_command_t *cmd, void *conf) { ngx_http_upstream_drizzle_srv_conf_t *dscf = conf; ngx_str_t *value; ngx_uint_t i; ngx_int_t n; u_char *data; ngx_uint_t len; if (dscf->max_cached) { return "is duplicate"; } value = cf->args->elts; for (i = 1; i < cf->args->nelts; i++) { if (ngx_strncmp(value[i].data, "max=", sizeof("max=") - 1) == 0) { len = value[i].len - (sizeof("max=") - 1); data = &value[i].data[sizeof("max=") - 1]; n = ngx_atoi(data, len); if (n == NGX_ERROR || n < 0) { ngx_conf_log_error(NGX_LOG_EMERG, cf, 0, "invalid \"max\" value \"%V\" " "in \"%V\" directive", &value[i], &cmd->name); return NGX_CONF_ERROR; } dscf->max_cached = n; continue; } if (ngx_strncmp(value[i].data, "mode=", sizeof("mode=") - 1) == 0) { len = value[i].len - (sizeof("mode=") - 1); data = &value[i].data[sizeof("mode=") - 1]; switch (len) { case 6: if (ngx_str6cmp(data, 's', 'i', 'n', 'g', 'l', 'e')) { dscf->single = 1; } break; case 5: if (ngx_str5cmp(data, 'm', 'u', 'l', 't', 'i')) { dscf->single = 0; } break; default: ngx_conf_log_error(NGX_LOG_EMERG, cf, 0, "drizzle: invalid \"mode\" value \"%V\" " "in \"%V\" directive", &value[i], &cmd->name); return NGX_CONF_ERROR; } continue; } if (ngx_strncmp(value[i].data, "overflow=", sizeof("overflow=") - 1) == 0) { len = value[i].len - (sizeof("overflow=") - 1); data = &value[i].data[sizeof("overflow=") - 1]; switch (len) { case 6: if (ngx_str6cmp(data, 'r', 'e', 'j', 'e', 'c', 't')) { dscf->overflow = drizzle_keepalive_overflow_reject; } else if (ngx_str6cmp(data, 'i', 'g', 'n', 'o', 'r', 'e')) { dscf->overflow = drizzle_keepalive_overflow_ignore; } break; default: ngx_conf_log_error(NGX_LOG_EMERG, cf, 0, "drizzle: invalid \"overflow\" value \"%V\" " "in \"%V\" directive", &value[i], &cmd->name); return NGX_CONF_ERROR; } continue; } ngx_conf_log_error(NGX_LOG_EMERG, cf, 0, "drizzle: invalid parameter \"%V\" in" " \"%V\" directive", &value[i], &cmd->name); return NGX_CONF_ERROR; } return NGX_CONF_OK; }
static ngx_int_t ngx_http_echo_parse_method_name(ngx_str_t **method_name_ptr) { const ngx_str_t* method_name = *method_name_ptr; switch (method_name->len) { case 3: if (ngx_str3cmp(method_name->data, 'G', 'E', 'T')) { *method_name_ptr = &ngx_http_echo_get_method; return NGX_HTTP_GET; break; } if (ngx_str3cmp(method_name->data, 'P', 'U', 'T')) { *method_name_ptr = &ngx_http_echo_put_method; return NGX_HTTP_PUT; break; } return NGX_HTTP_UNKNOWN; break; case 4: if (ngx_str4cmp(method_name->data, 'P', 'O', 'S', 'T')) { *method_name_ptr = &ngx_http_echo_post_method; return NGX_HTTP_POST; break; } if (ngx_str4cmp(method_name->data, 'H', 'E', 'A', 'D')) { *method_name_ptr = &ngx_http_echo_head_method; return NGX_HTTP_HEAD; break; } if (ngx_str4cmp(method_name->data, 'C', 'O', 'P', 'Y')) { *method_name_ptr = &ngx_http_echo_copy_method; return NGX_HTTP_COPY; break; } if (ngx_str4cmp(method_name->data, 'M', 'O', 'V', 'E')) { *method_name_ptr = &ngx_http_echo_move_method; return NGX_HTTP_MOVE; break; } if (ngx_str4cmp(method_name->data, 'L', 'O', 'C', 'K')) { *method_name_ptr = &ngx_http_echo_lock_method; return NGX_HTTP_LOCK; break; } return NGX_HTTP_UNKNOWN; break; case 5: if (ngx_strncmp("MKCOL", method_name->data, method_name->len) == 0) { *method_name_ptr = &ngx_http_echo_mkcol_method; return NGX_HTTP_MKCOL; break; } if (ngx_strncmp("TRACE", method_name->data, method_name->len) == 0) { *method_name_ptr = &ngx_http_echo_trace_method; return NGX_HTTP_TRACE; break; } return NGX_HTTP_UNKNOWN; break; case 6: if (ngx_str6cmp(method_name->data, 'D', 'E', 'L', 'E', 'T', 'E')) { *method_name_ptr = &ngx_http_echo_delete_method; return NGX_HTTP_DELETE; break; } if (ngx_str6cmp(method_name->data, 'U', 'N', 'L', 'O', 'C', 'K')) { *method_name_ptr = &ngx_http_echo_unlock_method; return NGX_HTTP_UNLOCK; break; } return NGX_HTTP_UNKNOWN; break; case 7: if (ngx_strncmp("OPTIONS", method_name->data, method_name->len) == 0) { *method_name_ptr = &ngx_http_echo_options_method; return NGX_HTTP_OPTIONS; break; } return NGX_HTTP_UNKNOWN; break; case 8: if (ngx_strncmp("PROPFIND", method_name->data, method_name->len) == 0) { *method_name_ptr = &ngx_http_echo_propfind_method; return NGX_HTTP_PROPFIND; break; } return NGX_HTTP_UNKNOWN; break; case 9: if (ngx_strncmp("PROPPATCH", method_name->data, method_name->len) == 0) { *method_name_ptr = &ngx_http_echo_proppatch_method; return NGX_HTTP_PROPPATCH; break; } return NGX_HTTP_UNKNOWN; break; default: return NGX_HTTP_UNKNOWN; break; } return NGX_HTTP_UNKNOWN; }
size_t http_parser_execute (http_parser *parser, const http_parser_settings *settings, const char *data, size_t len) { char c, ch; const char *p = data, *pe; ssize_t to_read; enum state state = (enum state) parser->state; enum header_states header_state = (enum header_states) parser->header_state; size_t index = parser->index; size_t nread = parser->nread; if (len == 0) { if (state == s_body_identity_eof) { CALLBACK2(message_complete); } return 0; } if (parser->header_field_mark) parser->header_field_mark = data; if (parser->header_value_mark) parser->header_value_mark = data; if (parser->fragment_mark) parser->fragment_mark = data; if (parser->query_string_mark) parser->query_string_mark = data; if (parser->path_mark) parser->path_mark = data; if (parser->url_mark) parser->url_mark = data; for (p=data, pe=data+len; p != pe; p++) { ch = *p; if (++nread > HTTP_MAX_HEADER_SIZE && PARSING_HEADER(state)) { /* Buffer overflow attack */ goto error; } switch (state) { case s_dead: /* this state is used after a 'Connection: close' message * the parser will error out if it reads another message */ goto error; case s_start_res: { parser->flags = 0; parser->content_length = -1; CALLBACK2(message_begin); switch (ch) { case 'H': state = s_res_H; break; case CR: case LF: break; default: goto error; } break; } case s_res_H: STRICT_CHECK(ch != 'T'); state = s_res_HT; break; case s_res_HT: STRICT_CHECK(ch != 'T'); state = s_res_HTT; break; case s_res_HTT: STRICT_CHECK(ch != 'P'); state = s_res_HTTP; break; case s_res_HTTP: STRICT_CHECK(ch != '/'); state = s_res_first_http_major; break; case s_res_first_http_major: if (ch < '1' || ch > '9') goto error; parser->http_major = ch - '0'; state = s_res_http_major; break; /* major HTTP version or dot */ case s_res_http_major: { if (ch == '.') { state = s_res_first_http_minor; break; } if (ch < '0' || ch > '9') goto error; parser->http_major *= 10; parser->http_major += ch - '0'; if (parser->http_major > 999) goto error; break; } /* first digit of minor HTTP version */ case s_res_first_http_minor: if (ch < '0' || ch > '9') goto error; parser->http_minor = ch - '0'; state = s_res_http_minor; break; /* minor HTTP version or end of request line */ case s_res_http_minor: { if (ch == ' ') { state = s_res_first_status_code; break; } if (ch < '0' || ch > '9') goto error; parser->http_minor *= 10; parser->http_minor += ch - '0'; if (parser->http_minor > 999) goto error; break; } case s_res_first_status_code: { if (ch < '0' || ch > '9') { if (ch == ' ') { break; } goto error; } parser->status_code = ch - '0'; state = s_res_status_code; break; } case s_res_status_code: { if (ch < '0' || ch > '9') { switch (ch) { case ' ': state = s_res_status; break; case CR: state = s_res_line_almost_done; break; case LF: state = s_header_field_start; break; default: goto error; } break; } parser->status_code *= 10; parser->status_code += ch - '0'; if (parser->status_code > 999) 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) { state = s_res_line_almost_done; break; } if (ch == LF) { state = s_header_field_start; break; } break; case s_res_line_almost_done: STRICT_CHECK(ch != LF); state = s_header_field_start; break; case s_start_req: { if (ch == CR || ch == LF) break; parser->flags = 0; parser->content_length = -1; CALLBACK2(message_begin); if (ch < 'A' || 'Z' < ch) goto error; parser->method = (enum http_method) 0; index = 0; parser->buffer[0] = ch; state = s_req_method; break; } case s_req_method: if (ch == ' ') { assert(index+1 < HTTP_PARSER_MAX_METHOD_LEN); parser->buffer[index+1] = '\0'; switch (index+1) { case 3: if (ngx_str3_cmp(parser->buffer, 'G', 'E', 'T')) { parser->method = HTTP_GET; break; } if (ngx_str3_cmp(parser->buffer, 'P', 'U', 'T')) { parser->method = HTTP_PUT; break; } break; case 4: if (ngx_str4cmp(parser->buffer, 'P', 'O', 'S', 'T')) { parser->method = HTTP_POST; break; } if (ngx_str4cmp(parser->buffer, 'H', 'E', 'A', 'D')) { parser->method = HTTP_HEAD; break; } if (ngx_str4cmp(parser->buffer, 'C', 'O', 'P', 'Y')) { parser->method = HTTP_COPY; break; } if (ngx_str4cmp(parser->buffer, 'M', 'O', 'V', 'E')) { parser->method = HTTP_MOVE; break; } break; case 5: if (ngx_str5cmp(parser->buffer, 'M', 'K', 'C', 'O', 'L')) { parser->method = HTTP_MKCOL; break; } if (ngx_str5cmp(parser->buffer, 'T', 'R', 'A', 'C', 'E')) { parser->method = HTTP_TRACE; break; } break; case 6: if (ngx_str6cmp(parser->buffer, 'D', 'E', 'L', 'E', 'T', 'E')) { parser->method = HTTP_DELETE; break; } if (ngx_str6cmp(parser->buffer, 'U', 'N', 'L', 'O', 'C', 'K')) { parser->method = HTTP_UNLOCK; break; } break; case 7: if (ngx_str7_cmp(parser->buffer, 'O', 'P', 'T', 'I', 'O', 'N', 'S', '\0')) { parser->method = HTTP_OPTIONS; break; } if (ngx_str7_cmp(parser->buffer, 'C', 'O', 'N', 'N', 'E', 'C', 'T', '\0')) { parser->method = HTTP_CONNECT; break; } break; case 8: if (ngx_str8cmp(parser->buffer, 'P', 'R', 'O', 'P', 'F', 'I', 'N', 'D')) { parser->method = HTTP_PROPFIND; break; } break; case 9: if (ngx_str9cmp(parser->buffer, 'P', 'R', 'O', 'P', 'P', 'A', 'T', 'C', 'H')) { parser->method = HTTP_PROPPATCH; break; } break; } state = s_req_spaces_before_url; break; } if (ch < 'A' || 'Z' < ch) goto error; if (++index >= HTTP_PARSER_MAX_METHOD_LEN - 1) { goto error; } parser->buffer[index] = ch; break; case s_req_spaces_before_url: { if (ch == ' ') break; if (ch == '/') { MARK(url); MARK(path); state = s_req_path; break; } c = LOWER(ch); if (c >= 'a' && c <= 'z') { MARK(url); state = s_req_schema; break; } goto error; } case s_req_schema: { c = LOWER(ch); if (c >= 'a' && c <= 'z') break; if (ch == ':') { state = s_req_schema_slash; break; } goto error; } case s_req_schema_slash: STRICT_CHECK(ch != '/'); state = s_req_schema_slash_slash; break; case s_req_schema_slash_slash: STRICT_CHECK(ch != '/'); state = s_req_host; break; case s_req_host: { c = LOWER(ch); if (c >= 'a' && c <= 'z') break; if ((ch >= '0' && ch <= '9') || ch == '.' || ch == '-') break; switch (ch) { case ':': state = s_req_port; break; case '/': MARK(path); state = s_req_path; break; case ' ': /* The request line looks like: * "GET http://foo.bar.com HTTP/1.1" * That is, there is no path. */ CALLBACK(url); state = s_req_http_start; break; default: goto error; } break; } case s_req_port: { if (ch >= '0' && ch <= '9') break; switch (ch) { case '/': MARK(path); state = s_req_path; break; case ' ': /* The request line looks like: * "GET http://foo.bar.com:1234 HTTP/1.1" * That is, there is no path. */ CALLBACK(url); state = s_req_http_start; break; default: goto error; } break; } case s_req_path: { if (USUAL(ch)) break; switch (ch) { case ' ': CALLBACK(url); CALLBACK(path); state = s_req_http_start; break; case CR: CALLBACK(url); CALLBACK(path); parser->http_minor = 9; state = s_req_line_almost_done; break; case LF: CALLBACK(url); CALLBACK(path); parser->http_minor = 9; state = s_header_field_start; break; case '?': CALLBACK(path); state = s_req_query_string_start; break; case '#': CALLBACK(path); state = s_req_fragment_start; break; default: goto error; } break; } case s_req_query_string_start: { if (USUAL(ch)) { MARK(query_string); state = s_req_query_string; break; } switch (ch) { case '?': break; // XXX ignore extra '?' ... is this right? case ' ': CALLBACK(url); state = s_req_http_start; break; case CR: CALLBACK(url); parser->http_minor = 9; state = s_req_line_almost_done; break; case LF: CALLBACK(url); parser->http_minor = 9; state = s_header_field_start; break; case '#': state = s_req_fragment_start; break; default: goto error; } break; } case s_req_query_string: { if (USUAL(ch)) break; switch (ch) { case '?': // allow extra '?' in query string break; case ' ': CALLBACK(url); CALLBACK(query_string); state = s_req_http_start; break; case CR: CALLBACK(url); CALLBACK(query_string); parser->http_minor = 9; state = s_req_line_almost_done; break; case LF: CALLBACK(url); CALLBACK(query_string); parser->http_minor = 9; state = s_header_field_start; break; case '#': CALLBACK(query_string); state = s_req_fragment_start; break; default: goto error; } break; } case s_req_fragment_start: { if (USUAL(ch)) { MARK(fragment); state = s_req_fragment; break; } switch (ch) { case ' ': CALLBACK(url); state = s_req_http_start; break; case CR: CALLBACK(url); parser->http_minor = 9; state = s_req_line_almost_done; break; case LF: CALLBACK(url); parser->http_minor = 9; state = s_header_field_start; break; case '?': MARK(fragment); state = s_req_fragment; break; case '#': break; default: goto error; } break; } case s_req_fragment: { if (USUAL(ch)) break; switch (ch) { case ' ': CALLBACK(url); CALLBACK(fragment); state = s_req_http_start; break; case CR: CALLBACK(url); CALLBACK(fragment); parser->http_minor = 9; state = s_req_line_almost_done; break; case LF: CALLBACK(url); CALLBACK(fragment); parser->http_minor = 9; state = s_header_field_start; break; case '?': case '#': break; default: goto error; } break; } case s_req_http_start: switch (ch) { case 'H': state = s_req_http_H; break; case ' ': break; default: goto error; } break; case s_req_http_H: STRICT_CHECK(ch != 'T'); state = s_req_http_HT; break; case s_req_http_HT: STRICT_CHECK(ch != 'T'); state = s_req_http_HTT; break; case s_req_http_HTT: STRICT_CHECK(ch != 'P'); state = s_req_http_HTTP; break; case s_req_http_HTTP: STRICT_CHECK(ch != '/'); state = s_req_first_http_major; break; /* first digit of major HTTP version */ case s_req_first_http_major: if (ch < '1' || ch > '9') goto error; parser->http_major = ch - '0'; state = s_req_http_major; break; /* major HTTP version or dot */ case s_req_http_major: { if (ch == '.') { state = s_req_first_http_minor; break; } if (ch < '0' || ch > '9') goto error; parser->http_major *= 10; parser->http_major += ch - '0'; if (parser->http_major > 999) goto error; break; } /* first digit of minor HTTP version */ case s_req_first_http_minor: if (ch < '0' || ch > '9') goto error; parser->http_minor = ch - '0'; state = s_req_http_minor; break; /* minor HTTP version or end of request line */ case s_req_http_minor: { if (ch == CR) { state = s_req_line_almost_done; break; } if (ch == LF) { state = s_header_field_start; break; } /* XXX allow spaces after digit? */ if (ch < '0' || ch > '9') goto error; parser->http_minor *= 10; parser->http_minor += ch - '0'; if (parser->http_minor > 999) goto error; break; } /* end of request line */ case s_req_line_almost_done: { if (ch != LF) goto error; state = s_header_field_start; break; } case s_header_field_start: { if (ch == CR) { 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*/ state = s_headers_almost_done; goto headers_almost_done; } c = LOWER(ch); if (c < 'a' || 'z' < c) goto error; MARK(header_field); index = 0; state = s_header_field; switch (c) { case 'c': header_state = h_C; break; case 'p': header_state = h_matching_proxy_connection; break; case 't': header_state = h_matching_transfer_encoding; break; case 'u': header_state = h_matching_upgrade; break; default: header_state = h_general; break; } break; } case s_header_field: { c = lowcase[(int)ch]; if (c) { switch (header_state) { case h_general: break; case h_C: index++; header_state = (c == 'o' ? h_CO : h_general); break; case h_CO: index++; header_state = (c == 'n' ? h_CON : h_general); break; case h_CON: index++; switch (c) { case 'n': header_state = h_matching_connection; break; case 't': header_state = h_matching_content_length; break; default: header_state = h_general; break; } break; /* connection */ case h_matching_connection: index++; if (index > sizeof(CONNECTION)-1 || c != CONNECTION[index]) { header_state = h_general; } else if (index == sizeof(CONNECTION)-2) { header_state = h_connection; } break; /* proxy-connection */ case h_matching_proxy_connection: index++; if (index > sizeof(PROXY_CONNECTION)-1 || c != PROXY_CONNECTION[index]) { header_state = h_general; } else if (index == sizeof(PROXY_CONNECTION)-2) { header_state = h_connection; } break; /* content-length */ case h_matching_content_length: index++; if (index > sizeof(CONTENT_LENGTH)-1 || c != CONTENT_LENGTH[index]) { header_state = h_general; } else if (index == sizeof(CONTENT_LENGTH)-2) { header_state = h_content_length; } break; /* transfer-encoding */ case h_matching_transfer_encoding: index++; if (index > sizeof(TRANSFER_ENCODING)-1 || c != TRANSFER_ENCODING[index]) { header_state = h_general; } else if (index == sizeof(TRANSFER_ENCODING)-2) { header_state = h_transfer_encoding; } break; /* upgrade */ case h_matching_upgrade: index++; if (index > sizeof(UPGRADE)-1 || c != UPGRADE[index]) { header_state = h_general; } else if (index == sizeof(UPGRADE)-2) { header_state = h_upgrade; } break; case h_connection: case h_content_length: case h_transfer_encoding: case h_upgrade: if (ch != ' ') header_state = h_general; break; default: assert(0 && "Unknown header_state"); break; } break; } if (ch == ':') { CALLBACK(header_field); state = s_header_value_start; break; } if (ch == CR) { state = s_header_almost_done; CALLBACK(header_field); break; } if (ch == LF) { CALLBACK(header_field); state = s_header_field_start; break; } goto error; } case s_header_value_start: { if (ch == ' ') break; MARK(header_value); state = s_header_value; index = 0; c = lowcase[(int)ch]; if (!c) { if (ch == CR) { header_state = h_general; state = s_header_almost_done; break; } if (ch == LF) { state = s_header_field_start; break; } header_state = h_general; break; } switch (header_state) { case h_upgrade: parser->flags |= F_UPGRADE; header_state = h_general; break; case h_transfer_encoding: /* looking for 'Transfer-Encoding: chunked' */ if ('c' == c) { header_state = h_matching_transfer_encoding_chunked; } else { header_state = h_general; } break; case h_content_length: if (ch < '0' || ch > '9') goto error; parser->content_length = ch - '0'; break; case h_connection: /* looking for 'Connection: keep-alive' */ if (c == 'k') { header_state = h_matching_connection_keep_alive; /* looking for 'Connection: close' */ } else if (c == 'c') { header_state = h_matching_connection_close; } else { header_state = h_general; } break; default: header_state = h_general; break; } break; } case s_header_value: { c = lowcase[(int)ch]; if (!c) { if (ch == CR) { CALLBACK(header_value); state = s_header_almost_done; break; } if (ch == LF) { CALLBACK(header_value); goto header_almost_done; } break; } switch (header_state) { case h_general: break; case h_connection: case h_transfer_encoding: assert(0 && "Shouldn't get here."); break; case h_content_length: if (ch < '0' || ch > '9') goto error; parser->content_length *= 10; parser->content_length += ch - '0'; break; /* Transfer-Encoding: chunked */ case h_matching_transfer_encoding_chunked: index++; if (index > sizeof(CHUNKED)-1 || c != CHUNKED[index]) { header_state = h_general; } else if (index == sizeof(CHUNKED)-2) { header_state = h_transfer_encoding_chunked; } break; /* looking for 'Connection: keep-alive' */ case h_matching_connection_keep_alive: index++; if (index > sizeof(KEEP_ALIVE)-1 || c != KEEP_ALIVE[index]) { header_state = h_general; } else if (index == sizeof(KEEP_ALIVE)-2) { header_state = h_connection_keep_alive; } break; /* looking for 'Connection: close' */ case h_matching_connection_close: index++; if (index > sizeof(CLOSE)-1 || c != CLOSE[index]) { header_state = h_general; } else if (index == sizeof(CLOSE)-2) { header_state = h_connection_close; } break; case h_transfer_encoding_chunked: case h_connection_keep_alive: case h_connection_close: if (ch != ' ') header_state = h_general; break; default: state = s_header_value; header_state = h_general; break; } break; } case s_header_almost_done: header_almost_done: { STRICT_CHECK(ch != LF); state = s_header_field_start; switch (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_headers_almost_done: headers_almost_done: { STRICT_CHECK(ch != LF); if (parser->flags & F_TRAILING) { /* End of a chunked request */ CALLBACK2(message_complete); state = NEW_MESSAGE(); break; } parser->body_read = 0; nread = 0; if (parser->flags & F_UPGRADE) parser->upgrade = 1; CALLBACK2(headers_complete); // Exit, the rest of the connect is in a different protocol. if (parser->flags & F_UPGRADE) { CALLBACK2(message_complete); return (p - data); } if (parser->flags & F_CHUNKED) { /* chunked encoding - ignore Content-Length header */ state = s_chunk_size_start; } else { if (parser->content_length == 0) { /* Content-Length header given but zero: Content-Length: 0\r\n */ CALLBACK2(message_complete); state = NEW_MESSAGE(); } else if (parser->content_length > 0) { /* Content-Length header given and non-zero */ state = s_body_identity; } else { if (parser->type == HTTP_REQUEST || http_should_keep_alive(parser)) { /* Assume content-length 0 - read the next */ CALLBACK2(message_complete); state = NEW_MESSAGE(); } else { /* Read body until EOF */ state = s_body_identity_eof; } } } break; } case s_body_identity: to_read = MIN(pe - p, (ssize_t)(parser->content_length - parser->body_read)); if (to_read > 0) { if (settings->on_body) settings->on_body(parser, p, to_read); p += to_read - 1; parser->body_read += to_read; if (parser->body_read == parser->content_length) { CALLBACK2(message_complete); state = NEW_MESSAGE(); } } break; /* read until EOF */ case s_body_identity_eof: to_read = pe - p; if (to_read > 0) { if (settings->on_body) settings->on_body(parser, p, to_read); p += to_read - 1; parser->body_read += to_read; } break; case s_chunk_size_start: { assert(parser->flags & F_CHUNKED); c = unhex[(int)ch]; if (c == -1) goto error; parser->content_length = c; state = s_chunk_size; break; } case s_chunk_size: { assert(parser->flags & F_CHUNKED); if (ch == CR) { state = s_chunk_size_almost_done; break; } c = unhex[(int)ch]; if (c == -1) { if (ch == ';' || ch == ' ') { state = s_chunk_parameters; break; } goto error; } parser->content_length *= 16; parser->content_length += c; break; } case s_chunk_parameters: { assert(parser->flags & F_CHUNKED); /* just ignore this shit. TODO check for overflow */ if (ch == CR) { state = s_chunk_size_almost_done; break; } break; } case s_chunk_size_almost_done: { assert(parser->flags & F_CHUNKED); STRICT_CHECK(ch != LF); if (parser->content_length == 0) { parser->flags |= F_TRAILING; state = s_header_field_start; } else { state = s_chunk_data; } break; } case s_chunk_data: { assert(parser->flags & F_CHUNKED); to_read = MIN(pe - p, (ssize_t)(parser->content_length)); if (to_read > 0) { if (settings->on_body) settings->on_body(parser, p, to_read); p += to_read - 1; } if (to_read == parser->content_length) { state = s_chunk_data_almost_done; } parser->content_length -= to_read; break; } case s_chunk_data_almost_done: assert(parser->flags & F_CHUNKED); STRICT_CHECK(ch != CR); state = s_chunk_data_done; break; case s_chunk_data_done: assert(parser->flags & F_CHUNKED); STRICT_CHECK(ch != LF); state = s_chunk_size_start; break; default: assert(0 && "unhandled state"); goto error; } } CALLBACK_NOCLEAR(header_field); CALLBACK_NOCLEAR(header_value); CALLBACK_NOCLEAR(fragment); CALLBACK_NOCLEAR(query_string); CALLBACK_NOCLEAR(path); CALLBACK_NOCLEAR(url); parser->state = state; parser->header_state = header_state; parser->index = index; parser->nread = nread; return len; error: return (p - data); }