1/*************************************************************************** 2 * _ _ ____ _ 3 * Project ___| | | | _ \| | 4 * / __| | | | |_) | | 5 * | (__| |_| | _ <| |___ 6 * \___|\___/|_| \_\_____| 7 * 8 * Copyright (C) Daniel Stenberg, <daniel@haxx.se>, et al. 9 * 10 * This software is licensed as described in the file COPYING, which 11 * you should have received as part of this distribution. The terms 12 * are also available at https://curl.se/docs/copyright.html. 13 * 14 * You may opt to use, copy, modify, merge, publish, distribute and/or sell 15 * copies of the Software, and permit persons to whom the Software is 16 * furnished to do so, under the terms of the COPYING file. 17 * 18 * This software is distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY 19 * KIND, either express or implied. 20 * 21 * SPDX-License-Identifier: curl 22 * 23 ***************************************************************************/ 24 25#include "curl_setup.h" 26 27#ifndef CURL_DISABLE_HTTP 28 29#include "urldata.h" 30#include <curl/curl.h> 31#include "http.h" 32#include "http1.h" 33#include "urlapi-int.h" 34 35/* The last 3 #include files should be in this order */ 36#include "curl_printf.h" 37#include "curl_memory.h" 38#include "memdebug.h" 39 40 41#define H1_MAX_URL_LEN (8*1024) 42 43void Curl_h1_req_parse_init(struct h1_req_parser *parser, size_t max_line_len) 44{ 45 memset(parser, 0, sizeof(*parser)); 46 parser->max_line_len = max_line_len; 47 Curl_dyn_init(&parser->scratch, max_line_len); 48} 49 50void Curl_h1_req_parse_free(struct h1_req_parser *parser) 51{ 52 if(parser) { 53 Curl_http_req_free(parser->req); 54 Curl_dyn_free(&parser->scratch); 55 parser->req = NULL; 56 parser->done = FALSE; 57 } 58} 59 60static CURLcode trim_line(struct h1_req_parser *parser, int options) 61{ 62 DEBUGASSERT(parser->line); 63 if(parser->line_len) { 64 if(parser->line[parser->line_len - 1] == '\n') 65 --parser->line_len; 66 if(parser->line_len) { 67 if(parser->line[parser->line_len - 1] == '\r') 68 --parser->line_len; 69 else if(options & H1_PARSE_OPT_STRICT) 70 return CURLE_URL_MALFORMAT; 71 } 72 else if(options & H1_PARSE_OPT_STRICT) 73 return CURLE_URL_MALFORMAT; 74 } 75 else if(options & H1_PARSE_OPT_STRICT) 76 return CURLE_URL_MALFORMAT; 77 78 if(parser->line_len > parser->max_line_len) { 79 return CURLE_URL_MALFORMAT; 80 } 81 return CURLE_OK; 82} 83 84static ssize_t detect_line(struct h1_req_parser *parser, 85 const char *buf, const size_t buflen, 86 CURLcode *err) 87{ 88 const char *line_end; 89 90 DEBUGASSERT(!parser->line); 91 line_end = memchr(buf, '\n', buflen); 92 if(!line_end) { 93 *err = CURLE_AGAIN; 94 return -1; 95 } 96 parser->line = buf; 97 parser->line_len = line_end - buf + 1; 98 *err = CURLE_OK; 99 return (ssize_t)parser->line_len; 100} 101 102static ssize_t next_line(struct h1_req_parser *parser, 103 const char *buf, const size_t buflen, int options, 104 CURLcode *err) 105{ 106 ssize_t nread = 0; 107 108 if(parser->line) { 109 parser->line = NULL; 110 parser->line_len = 0; 111 Curl_dyn_reset(&parser->scratch); 112 } 113 114 nread = detect_line(parser, buf, buflen, err); 115 if(nread >= 0) { 116 if(Curl_dyn_len(&parser->scratch)) { 117 /* append detected line to scratch to have the complete line */ 118 *err = Curl_dyn_addn(&parser->scratch, parser->line, parser->line_len); 119 if(*err) 120 return -1; 121 parser->line = Curl_dyn_ptr(&parser->scratch); 122 parser->line_len = Curl_dyn_len(&parser->scratch); 123 } 124 *err = trim_line(parser, options); 125 if(*err) 126 return -1; 127 } 128 else if(*err == CURLE_AGAIN) { 129 /* no line end in `buf`, add it to our scratch */ 130 *err = Curl_dyn_addn(&parser->scratch, (const unsigned char *)buf, buflen); 131 nread = (*err)? -1 : (ssize_t)buflen; 132 } 133 return nread; 134} 135 136static CURLcode start_req(struct h1_req_parser *parser, 137 const char *scheme_default, int options) 138{ 139 const char *p, *m, *target, *hv, *scheme, *authority, *path; 140 size_t m_len, target_len, hv_len, scheme_len, authority_len, path_len; 141 size_t i; 142 CURLU *url = NULL; 143 CURLcode result = CURLE_URL_MALFORMAT; /* Use this as default fail */ 144 145 DEBUGASSERT(!parser->req); 146 /* line must match: "METHOD TARGET HTTP_VERSION" */ 147 p = memchr(parser->line, ' ', parser->line_len); 148 if(!p || p == parser->line) 149 goto out; 150 151 m = parser->line; 152 m_len = p - parser->line; 153 target = p + 1; 154 target_len = hv_len = 0; 155 hv = NULL; 156 157 /* URL may contain spaces so scan backwards */ 158 for(i = parser->line_len; i > m_len; --i) { 159 if(parser->line[i] == ' ') { 160 hv = &parser->line[i + 1]; 161 hv_len = parser->line_len - i; 162 target_len = (hv - target) - 1; 163 break; 164 } 165 } 166 /* no SPACE found or empty TARGET or empty HTTP_VERSION */ 167 if(!target_len || !hv_len) 168 goto out; 169 170 /* TODO: we do not check HTTP_VERSION for conformity, should 171 + do that when STRICT option is supplied. */ 172 (void)hv; 173 174 /* The TARGET can be (rfc 9112, ch. 3.2): 175 * origin-form: path + optional query 176 * absolute-form: absolute URI 177 * authority-form: host+port for CONNECT 178 * asterisk-form: '*' for OPTIONS 179 * 180 * from TARGET, we derive `scheme` `authority` `path` 181 * origin-form -- -- TARGET 182 * absolute-form URL* URL* URL* 183 * authority-form -- TARGET -- 184 * asterisk-form -- -- TARGET 185 */ 186 scheme = authority = path = NULL; 187 scheme_len = authority_len = path_len = 0; 188 189 if(target_len == 1 && target[0] == '*') { 190 /* asterisk-form */ 191 path = target; 192 path_len = target_len; 193 } 194 else if(!strncmp("CONNECT", m, m_len)) { 195 /* authority-form */ 196 authority = target; 197 authority_len = target_len; 198 } 199 else if(target[0] == '/') { 200 /* origin-form */ 201 path = target; 202 path_len = target_len; 203 } 204 else { 205 /* origin-form OR absolute-form */ 206 CURLUcode uc; 207 char tmp[H1_MAX_URL_LEN]; 208 209 /* default, unless we see an absolute URL */ 210 path = target; 211 path_len = target_len; 212 213 /* URL parser wants 0-termination */ 214 if(target_len >= sizeof(tmp)) 215 goto out; 216 memcpy(tmp, target, target_len); 217 tmp[target_len] = '\0'; 218 /* See if treating TARGET as an absolute URL makes sense */ 219 if(Curl_is_absolute_url(tmp, NULL, 0, FALSE)) { 220 int url_options; 221 222 url = curl_url(); 223 if(!url) { 224 result = CURLE_OUT_OF_MEMORY; 225 goto out; 226 } 227 url_options = (CURLU_NON_SUPPORT_SCHEME| 228 CURLU_PATH_AS_IS| 229 CURLU_NO_DEFAULT_PORT); 230 if(!(options & H1_PARSE_OPT_STRICT)) 231 url_options |= CURLU_ALLOW_SPACE; 232 uc = curl_url_set(url, CURLUPART_URL, tmp, url_options); 233 if(uc) { 234 goto out; 235 } 236 } 237 238 if(!url && (options & H1_PARSE_OPT_STRICT)) { 239 /* we should have an absolute URL or have seen `/` earlier */ 240 goto out; 241 } 242 } 243 244 if(url) { 245 result = Curl_http_req_make2(&parser->req, m, m_len, url, scheme_default); 246 } 247 else { 248 if(!scheme && scheme_default) { 249 scheme = scheme_default; 250 scheme_len = strlen(scheme_default); 251 } 252 result = Curl_http_req_make(&parser->req, m, m_len, scheme, scheme_len, 253 authority, authority_len, path, path_len); 254 } 255 256out: 257 curl_url_cleanup(url); 258 return result; 259} 260 261ssize_t Curl_h1_req_parse_read(struct h1_req_parser *parser, 262 const char *buf, size_t buflen, 263 const char *scheme_default, int options, 264 CURLcode *err) 265{ 266 ssize_t nread = 0, n; 267 268 *err = CURLE_OK; 269 while(!parser->done) { 270 n = next_line(parser, buf, buflen, options, err); 271 if(n < 0) { 272 if(*err != CURLE_AGAIN) { 273 nread = -1; 274 } 275 *err = CURLE_OK; 276 goto out; 277 } 278 279 /* Consume this line */ 280 nread += (size_t)n; 281 buf += (size_t)n; 282 buflen -= (size_t)n; 283 284 if(!parser->line) { 285 /* consumed bytes, but line not complete */ 286 if(!buflen) 287 goto out; 288 } 289 else if(!parser->req) { 290 *err = start_req(parser, scheme_default, options); 291 if(*err) { 292 nread = -1; 293 goto out; 294 } 295 } 296 else if(parser->line_len == 0) { 297 /* last, empty line, we are finished */ 298 if(!parser->req) { 299 *err = CURLE_URL_MALFORMAT; 300 nread = -1; 301 goto out; 302 } 303 parser->done = TRUE; 304 Curl_dyn_reset(&parser->scratch); 305 /* last chance adjustments */ 306 } 307 else { 308 *err = Curl_dynhds_h1_add_line(&parser->req->headers, 309 parser->line, parser->line_len); 310 if(*err) { 311 nread = -1; 312 goto out; 313 } 314 } 315 } 316 317out: 318 return nread; 319} 320 321CURLcode Curl_h1_req_write_head(struct httpreq *req, int http_minor, 322 struct dynbuf *dbuf) 323{ 324 CURLcode result; 325 326 result = Curl_dyn_addf(dbuf, "%s %s%s%s%s HTTP/1.%d\r\n", 327 req->method, 328 req->scheme? req->scheme : "", 329 req->scheme? "://" : "", 330 req->authority? req->authority : "", 331 req->path? req->path : "", 332 http_minor); 333 if(result) 334 goto out; 335 336 result = Curl_dynhds_h1_dprint(&req->headers, dbuf); 337 if(result) 338 goto out; 339 340 result = Curl_dyn_addn(dbuf, STRCONST("\r\n")); 341 342out: 343 return result; 344} 345 346#endif /* !CURL_DISABLE_HTTP */ 347