1
0
Fork 0
mirror of https://github.com/ossrs/srs.git synced 2025-02-15 04:42:04 +00:00
srs/trunk/src/protocol/srs_http_stack.cpp

3437 lines
99 KiB
C++
Raw Normal View History

2017-03-25 09:21:39 +00:00
/**
* The MIT License (MIT)
*
2019-01-01 13:37:28 +00:00
* Copyright (c) 2013-2019 Winlin
2017-03-25 09:21:39 +00:00
*
* 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 <srs_http_stack.hpp>
#if !defined(SRS_EXPORT_LIBRTMP)
2015-05-23 07:22:20 +00:00
#include <stdlib.h>
#include <sstream>
2015-05-23 11:32:27 +00:00
#include <algorithm>
using namespace std;
#include <srs_kernel_error.hpp>
#include <srs_kernel_log.hpp>
#include <srs_kernel_utility.hpp>
#include <srs_kernel_file.hpp>
2015-08-21 08:20:19 +00:00
#include <srs_protocol_json.hpp>
#define SRS_HTTP_DEFAULT_PAGE "index.html"
// get the status text of code.
string srs_generate_http_status_text(int status)
{
static std::map<int, std::string> _status_map;
if (_status_map.empty()) {
_status_map[SRS_CONSTS_HTTP_Continue] = SRS_CONSTS_HTTP_Continue_str;
_status_map[SRS_CONSTS_HTTP_SwitchingProtocols] = SRS_CONSTS_HTTP_SwitchingProtocols_str;
_status_map[SRS_CONSTS_HTTP_OK] = SRS_CONSTS_HTTP_OK_str;
_status_map[SRS_CONSTS_HTTP_Created] = SRS_CONSTS_HTTP_Created_str;
_status_map[SRS_CONSTS_HTTP_Accepted] = SRS_CONSTS_HTTP_Accepted_str;
_status_map[SRS_CONSTS_HTTP_NonAuthoritativeInformation] = SRS_CONSTS_HTTP_NonAuthoritativeInformation_str;
_status_map[SRS_CONSTS_HTTP_NoContent] = SRS_CONSTS_HTTP_NoContent_str;
_status_map[SRS_CONSTS_HTTP_ResetContent] = SRS_CONSTS_HTTP_ResetContent_str;
_status_map[SRS_CONSTS_HTTP_PartialContent] = SRS_CONSTS_HTTP_PartialContent_str;
_status_map[SRS_CONSTS_HTTP_MultipleChoices] = SRS_CONSTS_HTTP_MultipleChoices_str;
_status_map[SRS_CONSTS_HTTP_MovedPermanently] = SRS_CONSTS_HTTP_MovedPermanently_str;
_status_map[SRS_CONSTS_HTTP_Found] = SRS_CONSTS_HTTP_Found_str;
_status_map[SRS_CONSTS_HTTP_SeeOther] = SRS_CONSTS_HTTP_SeeOther_str;
_status_map[SRS_CONSTS_HTTP_NotModified] = SRS_CONSTS_HTTP_NotModified_str;
_status_map[SRS_CONSTS_HTTP_UseProxy] = SRS_CONSTS_HTTP_UseProxy_str;
_status_map[SRS_CONSTS_HTTP_TemporaryRedirect] = SRS_CONSTS_HTTP_TemporaryRedirect_str;
_status_map[SRS_CONSTS_HTTP_BadRequest] = SRS_CONSTS_HTTP_BadRequest_str;
_status_map[SRS_CONSTS_HTTP_Unauthorized] = SRS_CONSTS_HTTP_Unauthorized_str;
_status_map[SRS_CONSTS_HTTP_PaymentRequired] = SRS_CONSTS_HTTP_PaymentRequired_str;
_status_map[SRS_CONSTS_HTTP_Forbidden] = SRS_CONSTS_HTTP_Forbidden_str;
_status_map[SRS_CONSTS_HTTP_NotFound] = SRS_CONSTS_HTTP_NotFound_str;
_status_map[SRS_CONSTS_HTTP_MethodNotAllowed] = SRS_CONSTS_HTTP_MethodNotAllowed_str;
_status_map[SRS_CONSTS_HTTP_NotAcceptable] = SRS_CONSTS_HTTP_NotAcceptable_str;
_status_map[SRS_CONSTS_HTTP_ProxyAuthenticationRequired] = SRS_CONSTS_HTTP_ProxyAuthenticationRequired_str;
_status_map[SRS_CONSTS_HTTP_RequestTimeout] = SRS_CONSTS_HTTP_RequestTimeout_str;
_status_map[SRS_CONSTS_HTTP_Conflict] = SRS_CONSTS_HTTP_Conflict_str;
_status_map[SRS_CONSTS_HTTP_Gone] = SRS_CONSTS_HTTP_Gone_str;
_status_map[SRS_CONSTS_HTTP_LengthRequired] = SRS_CONSTS_HTTP_LengthRequired_str;
_status_map[SRS_CONSTS_HTTP_PreconditionFailed] = SRS_CONSTS_HTTP_PreconditionFailed_str;
_status_map[SRS_CONSTS_HTTP_RequestEntityTooLarge] = SRS_CONSTS_HTTP_RequestEntityTooLarge_str;
_status_map[SRS_CONSTS_HTTP_RequestURITooLarge] = SRS_CONSTS_HTTP_RequestURITooLarge_str;
_status_map[SRS_CONSTS_HTTP_UnsupportedMediaType] = SRS_CONSTS_HTTP_UnsupportedMediaType_str;
_status_map[SRS_CONSTS_HTTP_RequestedRangeNotSatisfiable] = SRS_CONSTS_HTTP_RequestedRangeNotSatisfiable_str;
_status_map[SRS_CONSTS_HTTP_ExpectationFailed] = SRS_CONSTS_HTTP_ExpectationFailed_str;
_status_map[SRS_CONSTS_HTTP_InternalServerError] = SRS_CONSTS_HTTP_InternalServerError_str;
_status_map[SRS_CONSTS_HTTP_NotImplemented] = SRS_CONSTS_HTTP_NotImplemented_str;
_status_map[SRS_CONSTS_HTTP_BadGateway] = SRS_CONSTS_HTTP_BadGateway_str;
_status_map[SRS_CONSTS_HTTP_ServiceUnavailable] = SRS_CONSTS_HTTP_ServiceUnavailable_str;
_status_map[SRS_CONSTS_HTTP_GatewayTimeout] = SRS_CONSTS_HTTP_GatewayTimeout_str;
_status_map[SRS_CONSTS_HTTP_HTTPVersionNotSupported] = SRS_CONSTS_HTTP_HTTPVersionNotSupported_str;
}
std::string status_text;
if (_status_map.find(status) == _status_map.end()) {
status_text = "Status Unknown";
} else {
status_text = _status_map[status];
}
return status_text;
}
// bodyAllowedForStatus reports whether a given response status code
// permits a body. See RFC2616, section 4.4.
bool srs_go_http_body_allowd(int status)
{
2019-12-14 02:12:25 +00:00
if (status >= SRS_CONSTS_HTTP_Continue && status < SRS_CONSTS_HTTP_OK) {
return false;
2019-12-14 02:12:25 +00:00
} else if (status == SRS_CONSTS_HTTP_NoContent || status == SRS_CONSTS_HTTP_NotModified) {
return false;
}
return true;
}
// DetectContentType implements the algorithm described
// at http://mimesniff.spec.whatwg.org/ to determine the
// Content-Type of the given data. It considers at most the
// first 512 bytes of data. DetectContentType always returns
// a valid MIME type: if it cannot determine a more specific one, it
// returns "application/octet-stream".
string srs_go_http_detect(char* data, int size)
{
2019-12-14 02:12:25 +00:00
// TODO: Implement the content detecting.
return "application/octet-stream"; // fallback
}
2017-07-29 13:39:57 +00:00
srs_error_t srs_go_http_error(ISrsHttpResponseWriter* w, int code)
2015-08-22 05:36:15 +00:00
{
return srs_go_http_error(w, code, srs_generate_http_status_text(code));
}
2017-07-29 13:39:57 +00:00
srs_error_t srs_go_http_error(ISrsHttpResponseWriter* w, int code, string error)
{
srs_error_t err = srs_success;
w->header()->set_content_type("text/plain; charset=utf-8");
w->header()->set_content_length(error.length());
w->write_header(code);
if ((err = w->write((char*)error.data(), (int)error.length())) != srs_success) {
return srs_error_wrap(err, "http write");
2017-07-29 13:39:57 +00:00
}
return err;
}
SrsHttpHeader::SrsHttpHeader()
{
}
SrsHttpHeader::~SrsHttpHeader()
{
}
void SrsHttpHeader::set(string key, string value)
{
headers[key] = value;
}
string SrsHttpHeader::get(string key)
{
std::string v;
if (headers.find(key) != headers.end()) {
v = headers[key];
}
return v;
}
int64_t SrsHttpHeader::content_length()
{
std::string cl = get("Content-Length");
if (cl.empty()) {
return -1;
}
return (int64_t)::atof(cl.c_str());
}
void SrsHttpHeader::set_content_length(int64_t size)
{
set("Content-Length", srs_int2str(size));
}
string SrsHttpHeader::content_type()
{
return get("Content-Type");
}
void SrsHttpHeader::set_content_type(string ct)
{
set("Content-Type", ct);
}
void SrsHttpHeader::write(stringstream& ss)
{
std::map<std::string, std::string>::iterator it;
for (it = headers.begin(); it != headers.end(); ++it) {
ss << it->first << ": " << it->second << SRS_HTTP_CRLF;
}
}
ISrsHttpResponseWriter::ISrsHttpResponseWriter()
{
}
ISrsHttpResponseWriter::~ISrsHttpResponseWriter()
{
}
ISrsHttpResponseReader::ISrsHttpResponseReader()
{
}
ISrsHttpResponseReader::~ISrsHttpResponseReader()
{
}
ISrsHttpHandler::ISrsHttpHandler()
{
entry = NULL;
}
ISrsHttpHandler::~ISrsHttpHandler()
{
}
bool ISrsHttpHandler::is_not_found()
{
return false;
}
SrsHttpRedirectHandler::SrsHttpRedirectHandler(string u, int c)
{
url = u;
code = c;
}
SrsHttpRedirectHandler::~SrsHttpRedirectHandler()
{
}
2017-07-29 13:39:57 +00:00
srs_error_t SrsHttpRedirectHandler::serve_http(ISrsHttpResponseWriter* w, ISrsHttpMessage* r)
{
string location = url;
if (!r->query().empty()) {
location += "?" + r->query();
}
string msg = "Redirect to" + location;
2017-03-25 09:21:39 +00:00
2015-08-11 07:23:46 +00:00
w->header()->set_content_type("text/plain; charset=utf-8");
w->header()->set_content_length(msg.length());
w->header()->set("Location", location);
2015-08-11 07:23:46 +00:00
w->write_header(code);
2017-03-25 09:21:39 +00:00
2015-08-11 07:23:46 +00:00
w->write((char*)msg.data(), (int)msg.length());
w->final_request();
2017-03-25 09:21:39 +00:00
srs_info("redirect to %s.", location.c_str());
2017-07-29 13:39:57 +00:00
return srs_success;
}
SrsHttpNotFoundHandler::SrsHttpNotFoundHandler()
{
}
SrsHttpNotFoundHandler::~SrsHttpNotFoundHandler()
{
}
bool SrsHttpNotFoundHandler::is_not_found()
{
return true;
}
2017-07-29 13:39:57 +00:00
srs_error_t SrsHttpNotFoundHandler::serve_http(ISrsHttpResponseWriter* w, ISrsHttpMessage* r)
{
2015-08-22 05:36:15 +00:00
return srs_go_http_error(w, SRS_CONSTS_HTTP_NotFound);
}
SrsHttpFileServer::SrsHttpFileServer(string root_dir)
{
dir = root_dir;
}
SrsHttpFileServer::~SrsHttpFileServer()
{
}
2017-07-29 13:39:57 +00:00
srs_error_t SrsHttpFileServer::serve_http(ISrsHttpResponseWriter* w, ISrsHttpMessage* r)
{
string upath = r->path();
// add default pages.
if (srs_string_ends_with(upath, "/")) {
upath += SRS_HTTP_DEFAULT_PAGE;
}
string fullpath = dir + "/";
// remove the virtual directory.
srs_assert(entry);
size_t pos = entry->pattern.find("/");
if (upath.length() > entry->pattern.length() && pos != string::npos) {
fullpath += upath.substr(entry->pattern.length() - pos);
} else {
fullpath += upath;
}
// stat current dir, if exists, return error.
if (!srs_path_exists(fullpath)) {
srs_warn("http miss file=%s, pattern=%s, upath=%s",
fullpath.c_str(), entry->pattern.c_str(), upath.c_str());
return SrsHttpNotFoundHandler().serve_http(w, r);
}
srs_trace("http match file=%s, pattern=%s, upath=%s",
fullpath.c_str(), entry->pattern.c_str(), upath.c_str());
// handle file according to its extension.
// use vod stream for .flv/.fhv
if (srs_string_ends_with(fullpath, ".flv") || srs_string_ends_with(fullpath, ".fhv")) {
return serve_flv_file(w, r, fullpath);
} else if (srs_string_ends_with(fullpath, ".mp4")) {
return serve_mp4_file(w, r, fullpath);
}
// serve common static file.
return serve_file(w, r, fullpath);
}
2017-07-29 13:39:57 +00:00
srs_error_t SrsHttpFileServer::serve_file(ISrsHttpResponseWriter* w, ISrsHttpMessage* r, string fullpath)
{
2017-07-29 13:39:57 +00:00
srs_error_t err = srs_success;
// open the target file.
SrsFileReader fs;
if ((err = fs.open(fullpath)) != srs_success) {
return srs_error_wrap(err, "open file %s", fullpath.c_str());
}
int64_t length = fs.filesize();
// unset the content length to encode in chunked encoding.
w->header()->set_content_length(length);
static std::map<std::string, std::string> _mime;
if (_mime.empty()) {
_mime[".ts"] = "video/MP2T";
_mime[".flv"] = "video/x-flv";
_mime[".m4v"] = "video/x-m4v";
_mime[".3gpp"] = "video/3gpp";
_mime[".3gp"] = "video/3gpp";
_mime[".mp4"] = "video/mp4";
_mime[".aac"] = "audio/x-aac";
_mime[".mp3"] = "audio/mpeg";
_mime[".m4a"] = "audio/x-m4a";
_mime[".ogg"] = "audio/ogg";
// @see hls-m3u8-draft-pantos-http-live-streaming-12.pdf, page 5.
_mime[".m3u8"] = "application/vnd.apple.mpegurl"; // application/x-mpegURL
_mime[".rss"] = "application/rss+xml";
_mime[".json"] = "application/json";
_mime[".swf"] = "application/x-shockwave-flash";
_mime[".doc"] = "application/msword";
_mime[".zip"] = "application/zip";
_mime[".rar"] = "application/x-rar-compressed";
_mime[".xml"] = "text/xml";
_mime[".html"] = "text/html";
_mime[".js"] = "text/javascript";
_mime[".css"] = "text/css";
_mime[".ico"] = "image/x-icon";
_mime[".png"] = "image/png";
_mime[".jpeg"] = "image/jpeg";
_mime[".jpg"] = "image/jpeg";
_mime[".gif"] = "image/gif";
2017-02-26 12:40:33 +00:00
// For MPEG-DASH.
//_mime[".mpd"] = "application/dash+xml";
_mime[".mpd"] = "text/xml";
_mime[".m4s"] = "video/iso.segment";
_mime[".mp4v"] = "video/mp4";
}
if (true) {
2015-10-13 08:06:37 +00:00
std::string ext = srs_path_filext(fullpath);
if (_mime.find(ext) == _mime.end()) {
w->header()->set_content_type("application/octet-stream");
} else {
w->header()->set_content_type(_mime[ext]);
}
}
// write body.
int64_t left = length;
2017-07-29 13:39:57 +00:00
if ((err = copy(w, &fs, r, (int)left)) != srs_success) {
return srs_error_wrap(err, "copy file=%s size=%d", fullpath.c_str(), left);
}
if ((err = w->final_request()) != srs_success) {
return srs_error_wrap(err, "final request");
2017-07-29 13:39:57 +00:00
}
return err;
}
2017-07-29 13:39:57 +00:00
srs_error_t SrsHttpFileServer::serve_flv_file(ISrsHttpResponseWriter* w, ISrsHttpMessage* r, string fullpath)
{
std::string start = r->query_get("start");
if (start.empty()) {
return serve_file(w, r, fullpath);
}
int offset = ::atoi(start.c_str());
if (offset <= 0) {
return serve_file(w, r, fullpath);
}
return serve_flv_stream(w, r, fullpath, offset);
}
2017-07-29 13:39:57 +00:00
srs_error_t SrsHttpFileServer::serve_mp4_file(ISrsHttpResponseWriter* w, ISrsHttpMessage* r, string fullpath)
{
// for flash to request mp4 range in query string.
// for example, http://digitalprimates.net/dash/DashTest.html?url=http://dashdemo.edgesuite.net/digitalprimates/nexus/oops-20120802-manifest.mpd
std::string range = r->query_get("range");
// or, use bytes to request range,
// for example, http://dashas.castlabs.com/demo/try.html
if (range.empty()) {
range = r->query_get("bytes");
}
// rollback to serve whole file.
size_t pos = string::npos;
if (range.empty() || (pos = range.find("-")) == string::npos) {
return serve_file(w, r, fullpath);
}
// parse the start in query string
int start = 0;
if (pos > 0) {
start = ::atoi(range.substr(0, pos).c_str());
}
// parse end in query string.
int end = -1;
if (pos < range.length() - 1) {
end = ::atoi(range.substr(pos + 1).c_str());
}
// invalid param, serve as whole mp4 file.
if (start < 0 || (end != -1 && start > end)) {
return serve_file(w, r, fullpath);
}
return serve_mp4_stream(w, r, fullpath, start, end);
}
2017-07-29 13:39:57 +00:00
srs_error_t SrsHttpFileServer::serve_flv_stream(ISrsHttpResponseWriter* w, ISrsHttpMessage* r, string fullpath, int offset)
{
return serve_file(w, r, fullpath);
}
2017-07-29 13:39:57 +00:00
srs_error_t SrsHttpFileServer::serve_mp4_stream(ISrsHttpResponseWriter* w, ISrsHttpMessage* r, string fullpath, int start, int end)
{
return serve_file(w, r, fullpath);
}
2017-07-29 13:39:57 +00:00
srs_error_t SrsHttpFileServer::copy(ISrsHttpResponseWriter* w, SrsFileReader* fs, ISrsHttpMessage* r, int size)
{
srs_error_t err = srs_success;
int left = size;
char* buf = r->http_ts_send_buffer();
while (left > 0) {
ssize_t nread = -1;
int max_read = srs_min(left, SRS_HTTP_TS_SEND_BUFFER_SIZE);
if ((err = fs->read(buf, max_read, &nread)) != srs_success) {
break;
}
left -= nread;
if ((err = w->write(buf, (int)nread)) != srs_success) {
break;
}
}
if (err != srs_success) {
return srs_error_wrap(err, "copy");
2017-07-29 13:39:57 +00:00
}
return err;
}
SrsHttpMuxEntry::SrsHttpMuxEntry()
{
enabled = true;
explicit_match = false;
handler = NULL;
}
SrsHttpMuxEntry::~SrsHttpMuxEntry()
{
srs_freep(handler);
}
ISrsHttpMatchHijacker::ISrsHttpMatchHijacker()
{
}
ISrsHttpMatchHijacker::~ISrsHttpMatchHijacker()
{
}
ISrsHttpServeMux::ISrsHttpServeMux()
{
}
ISrsHttpServeMux::~ISrsHttpServeMux()
{
}
SrsHttpServeMux::SrsHttpServeMux()
{
}
SrsHttpServeMux::~SrsHttpServeMux()
{
std::map<std::string, SrsHttpMuxEntry*>::iterator it;
for (it = entries.begin(); it != entries.end(); ++it) {
SrsHttpMuxEntry* entry = it->second;
srs_freep(entry);
}
entries.clear();
vhosts.clear();
hijackers.clear();
}
2017-06-09 05:29:23 +00:00
srs_error_t SrsHttpServeMux::initialize()
{
2017-06-09 05:29:23 +00:00
srs_error_t err = srs_success;
// TODO: FIXME: Implements it.
return err;
}
void SrsHttpServeMux::hijack(ISrsHttpMatchHijacker* h)
{
std::vector<ISrsHttpMatchHijacker*>::iterator it = ::find(hijackers.begin(), hijackers.end(), h);
if (it != hijackers.end()) {
return;
}
hijackers.push_back(h);
}
void SrsHttpServeMux::unhijack(ISrsHttpMatchHijacker* h)
{
std::vector<ISrsHttpMatchHijacker*>::iterator it = ::find(hijackers.begin(), hijackers.end(), h);
if (it == hijackers.end()) {
return;
}
hijackers.erase(it);
}
2017-06-10 07:20:48 +00:00
srs_error_t SrsHttpServeMux::handle(std::string pattern, ISrsHttpHandler* handler)
{
srs_assert(handler);
if (pattern.empty()) {
2017-06-10 07:20:48 +00:00
return srs_error_new(ERROR_HTTP_PATTERN_EMPTY, "empty pattern");
}
if (entries.find(pattern) != entries.end()) {
SrsHttpMuxEntry* exists = entries[pattern];
if (exists->explicit_match) {
2017-06-10 07:20:48 +00:00
return srs_error_new(ERROR_HTTP_PATTERN_DUPLICATED, "pattern=%s exists", pattern.c_str());
}
}
std::string vhost = pattern;
if (pattern.at(0) != '/') {
if (pattern.find("/") != string::npos) {
vhost = pattern.substr(0, pattern.find("/"));
}
vhosts[vhost] = handler;
}
if (true) {
SrsHttpMuxEntry* entry = new SrsHttpMuxEntry();
entry->explicit_match = true;
entry->handler = handler;
entry->pattern = pattern;
entry->handler->entry = entry;
if (entries.find(pattern) != entries.end()) {
SrsHttpMuxEntry* exists = entries[pattern];
srs_freep(exists);
}
entries[pattern] = entry;
}
// Helpful behavior:
// If pattern is /tree/, insert an implicit permanent redirect for /tree.
// It can be overridden by an explicit registration.
if (pattern != "/" && !pattern.empty() && pattern.at(pattern.length() - 1) == '/') {
std::string rpattern = pattern.substr(0, pattern.length() - 1);
SrsHttpMuxEntry* entry = NULL;
// free the exists not explicit entry
if (entries.find(rpattern) != entries.end()) {
SrsHttpMuxEntry* exists = entries[rpattern];
if (!exists->explicit_match) {
entry = exists;
}
}
// create implicit redirect.
if (!entry || entry->explicit_match) {
srs_freep(entry);
entry = new SrsHttpMuxEntry();
entry->explicit_match = false;
entry->handler = new SrsHttpRedirectHandler(pattern, SRS_CONSTS_HTTP_Found);
entry->pattern = pattern;
entry->handler->entry = entry;
entries[rpattern] = entry;
}
}
2017-06-10 07:20:48 +00:00
return srs_success;
}
2017-07-29 13:39:57 +00:00
srs_error_t SrsHttpServeMux::serve_http(ISrsHttpResponseWriter* w, ISrsHttpMessage* r)
{
2017-07-29 13:39:57 +00:00
srs_error_t err = srs_success;
ISrsHttpHandler* h = NULL;
2017-07-29 13:39:57 +00:00
if ((err = find_handler(r, &h)) != srs_success) {
return srs_error_wrap(err, "find handler");
}
srs_assert(h);
2017-07-29 13:39:57 +00:00
if ((err = h->serve_http(w, r)) != srs_success) {
return srs_error_wrap(err, "serve http");
}
2017-07-29 13:39:57 +00:00
return err;
}
2017-07-29 13:39:57 +00:00
srs_error_t SrsHttpServeMux::find_handler(ISrsHttpMessage* r, ISrsHttpHandler** ph)
{
2017-07-29 13:39:57 +00:00
srs_error_t err = srs_success;
// TODO: FIXME: support the path . and ..
if (r->url().find("..") != std::string::npos) {
2017-07-29 13:39:57 +00:00
return srs_error_new(ERROR_HTTP_URL_NOT_CLEAN, "url %s not canonical", r->url().c_str());
}
2017-07-29 13:39:57 +00:00
if ((err = match(r, ph)) != srs_success) {
return srs_error_wrap(err, "http match");
}
// always hijack.
if (!hijackers.empty()) {
// notice all hijacker the match failed.
std::vector<ISrsHttpMatchHijacker*>::iterator it;
for (it = hijackers.begin(); it != hijackers.end(); ++it) {
ISrsHttpMatchHijacker* hijacker = *it;
2017-07-29 13:39:57 +00:00
if ((err = hijacker->hijack(r, ph)) != srs_success) {
return srs_error_wrap(err, "http hijack");
}
}
}
static ISrsHttpHandler* h404 = new SrsHttpNotFoundHandler();
if (*ph == NULL) {
*ph = h404;
}
2017-07-29 13:39:57 +00:00
return err;
}
2017-07-29 13:39:57 +00:00
srs_error_t SrsHttpServeMux::match(ISrsHttpMessage* r, ISrsHttpHandler** ph)
{
std::string path = r->path();
// Host-specific pattern takes precedence over generic ones
if (!vhosts.empty() && vhosts.find(r->host()) != vhosts.end()) {
path = r->host() + path;
}
int nb_matched = 0;
ISrsHttpHandler* h = NULL;
std::map<std::string, SrsHttpMuxEntry*>::iterator it;
for (it = entries.begin(); it != entries.end(); ++it) {
std::string pattern = it->first;
SrsHttpMuxEntry* entry = it->second;
if (!entry->enabled) {
continue;
}
if (!path_match(pattern, path)) {
continue;
}
if (!h || (int)pattern.length() > nb_matched) {
nb_matched = (int)pattern.length();
h = entry->handler;
}
}
*ph = h;
2017-07-29 13:39:57 +00:00
return srs_success;
}
bool SrsHttpServeMux::path_match(string pattern, string path)
{
if (pattern.empty()) {
return false;
}
int n = (int)pattern.length();
// not endswith '/', exactly match.
if (pattern.at(n - 1) != '/') {
return pattern == path;
}
// endswith '/', match any,
// for example, '/api/' match '/api/[N]'
if ((int)path.length() >= n) {
if (memcmp(pattern.data(), path.data(), n) == 0) {
return true;
}
}
return false;
}
2016-12-15 07:00:08 +00:00
SrsHttpCorsMux::SrsHttpCorsMux()
2016-12-15 06:10:02 +00:00
{
next = NULL;
enabled = false;
required = false;
}
2016-12-15 07:00:08 +00:00
SrsHttpCorsMux::~SrsHttpCorsMux()
2016-12-15 06:10:02 +00:00
{
}
2017-07-29 13:39:57 +00:00
srs_error_t SrsHttpCorsMux::initialize(ISrsHttpServeMux* worker, bool cros_enabled)
2016-12-15 06:10:02 +00:00
{
next = worker;
enabled = cros_enabled;
2017-07-29 13:39:57 +00:00
return srs_success;
2016-12-15 06:10:02 +00:00
}
2017-07-29 13:39:57 +00:00
srs_error_t SrsHttpCorsMux::serve_http(ISrsHttpResponseWriter* w, ISrsHttpMessage* r)
2016-12-15 06:10:02 +00:00
{
srs_error_t err = srs_success;
2017-07-29 13:39:57 +00:00
2017-02-26 13:45:08 +00:00
// If CORS enabled, and there is a "Origin" header, it's CORS.
if (enabled) {
for (int i = 0; i < r->request_header_count(); i++) {
string k = r->request_header_key_at(i);
if (k == "Origin" || k == "origin") {
required = true;
break;
}
}
2016-12-15 06:10:02 +00:00
}
2017-02-26 13:45:08 +00:00
// When CORS required, set the CORS headers.
2016-12-15 06:10:02 +00:00
if (required) {
2017-02-26 13:45:08 +00:00
SrsHttpHeader* h = w->header();
h->set("Access-Control-Allow-Origin", "*");
h->set("Access-Control-Allow-Methods", "GET, POST, HEAD, PUT, DELETE, OPTIONS");
h->set("Access-Control-Expose-Headers", "Server,range,Content-Length,Content-Range");
h->set("Access-Control-Allow-Headers", "origin,range,accept-encoding,referer,Cache-Control,X-Proxy-Authorization,X-Requested-With,Content-Type");
2016-12-15 06:10:02 +00:00
}
// handle the http options.
if (r->is_http_options()) {
w->header()->set_content_length(0);
if (enabled) {
w->write_header(SRS_CONSTS_HTTP_OK);
} else {
w->write_header(SRS_CONSTS_HTTP_MethodNotAllowed);
}
if ((err = w->final_request()) != srs_success) {
return srs_error_wrap(err, "final request");
2017-07-29 13:39:57 +00:00
}
2016-12-15 06:10:02 +00:00
}
srs_assert(next);
return next->serve_http(w, r);
}
ISrsHttpMessage::ISrsHttpMessage()
{
_http_ts_send_buffer = new char[SRS_HTTP_TS_SEND_BUFFER_SIZE];
}
ISrsHttpMessage::~ISrsHttpMessage()
{
2015-11-02 03:05:39 +00:00
srs_freepa(_http_ts_send_buffer);
}
char* ISrsHttpMessage::http_ts_send_buffer()
{
return _http_ts_send_buffer;
}
2019-04-23 00:26:15 +00:00
SrsHttpUri::SrsHttpUri()
{
port = SRS_DEFAULT_HTTP_PORT;
}
SrsHttpUri::~SrsHttpUri()
{
}
srs_error_t SrsHttpUri::initialize(string _url)
{
srs_error_t err = srs_success;
schema = host = path = query = "";
url = _url;
const char* purl = url.c_str();
http_parser_url hp_u;
int r0;
if((r0 = http_parser_parse_url(purl, url.length(), 0, &hp_u)) != 0){
return srs_error_new(ERROR_HTTP_PARSE_URI, "parse url %s failed, code=%d", purl, r0);
}
std::string field = get_uri_field(url, &hp_u, UF_SCHEMA);
if(!field.empty()){
schema = field;
}
host = get_uri_field(url, &hp_u, UF_HOST);
field = get_uri_field(url, &hp_u, UF_PORT);
if(!field.empty()){
port = atoi(field.c_str());
}
if(port<=0){
port = SRS_DEFAULT_HTTP_PORT;
}
path = get_uri_field(url, &hp_u, UF_PATH);
query = get_uri_field(url, &hp_u, UF_QUERY);
return err;
}
string SrsHttpUri::get_url()
{
return url;
}
string SrsHttpUri::get_schema()
{
return schema;
}
string SrsHttpUri::get_host()
{
return host;
}
int SrsHttpUri::get_port()
{
return port;
}
string SrsHttpUri::get_path()
{
return path;
}
string SrsHttpUri::get_query()
{
return query;
}
string SrsHttpUri::get_uri_field(string uri, void* php_u, int ifield)
{
http_parser_url* hp_u = (http_parser_url*)php_u;
http_parser_url_fields field = (http_parser_url_fields)ifield;
if((hp_u->field_set & (1 << field)) == 0){
return "";
}
int offset = hp_u->field_data[field].off;
int len = hp_u->field_data[field].len;
return uri.substr(offset, len);
}
2019-04-23 00:31:21 +00:00
// For #if !defined(SRS_EXPORT_LIBRTMP)
#endif
2019-04-23 00:31:21 +00:00
/////////////////////////////////////////////////////////////////////////////////////////
/////////////////////////////////////////////////////////////////////////////////////////
/////////////////////////////////////////////////////////////////////////////////////////
/////////////////////////////////////////////////////////////////////////////////////////
/////////////////////////////////////////////////////////////////////////////////////////
/////////////////////////////////////////////////////////////////////////////////////////
/////////////////////////////////////////////////////////////////////////////////////////
/////////////////////////////////////////////////////////////////////////////////////////
/////////////////////////////////////////////////////////////////////////////////////////
/////////////////////////////////////////////////////////////////////////////////////////
/////////////////////////////////////////////////////////////////////////////////////////
/////////////////////////////////////////////////////////////////////////////////////////
/////////////////////////////////////////////////////////////////////////////////////////
/////////////////////////////////////////////////////////////////////////////////////////
/////////////////////////////////////////////////////////////////////////////////////////
/////////////////////////////////////////////////////////////////////////////////////////
/////////////////////////////////////////////////////////////////////////////////////////
/////////////////////////////////////////////////////////////////////////////////////////
/////////////////////////////////////////////////////////////////////////////////////////
/////////////////////////////////////////////////////////////////////////////////////////
/////////////////////////////////////////////////////////////////////////////////////////
/////////////////////////////////////////////////////////////////////////////////////////
/////////////////////////////////////////////////////////////////////////////////////////
/////////////////////////////////////////////////////////////////////////////////////////
2019-04-24 01:37:25 +00:00
/* Copyright Joyent, Inc. and other Node contributors.
2017-03-25 09:21:39 +00:00
*
* 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 "http_parser.h"
#include <assert.h>
#include <stddef.h>
#include <ctype.h>
#include <string.h>
#include <limits.h>
2019-04-24 01:37:25 +00:00
static uint32_t max_header_size = HTTP_MAX_HEADER_SIZE;
#ifndef ULLONG_MAX
# define ULLONG_MAX ((uint64_t) -1) /* 2^64-1 */
#endif
#ifndef MIN
# define MIN(a,b) ((a) < (b) ? (a) : (b))
#endif
#ifndef ARRAY_SIZE
# define ARRAY_SIZE(a) (sizeof(a) / sizeof((a)[0]))
#endif
#ifndef BIT_AT
# define BIT_AT(a, i) \
2019-04-23 00:31:21 +00:00
(!!((unsigned int) (a)[(unsigned int) (i) >> 3] & \
(1 << ((unsigned int) (i) & 7))))
#endif
#ifndef ELEM_AT
# define ELEM_AT(a, i, v) ((unsigned int) (i) < ARRAY_SIZE(a) ? (a)[(i)] : (v))
#endif
#define SET_ERRNO(e) \
do { \
2019-04-24 01:37:25 +00:00
parser->nread = nread; \
2019-04-23 00:31:21 +00:00
parser->http_errno = (e); \
} while(0)
2019-04-24 01:37:25 +00:00
#define CURRENT_STATE() p_state
#define UPDATE_STATE(V) p_state = (enum state) (V);
#define RETURN(V) \
do { \
parser->nread = nread; \
parser->state = CURRENT_STATE(); \
return (V); \
} while (0);
#define REEXECUTE() \
goto reexecute; \
#ifdef __GNUC__
# define LIKELY(X) __builtin_expect(!!(X), 1)
# define UNLIKELY(X) __builtin_expect(!!(X), 0)
#else
# define LIKELY(X) (X)
# define UNLIKELY(X) (X)
#endif
/* Run the notify callback FOR, returning ER if it fails */
#define CALLBACK_NOTIFY_(FOR, ER) \
do { \
2019-04-23 00:31:21 +00:00
assert(HTTP_PARSER_ERRNO(parser) == HPE_OK); \
\
2019-04-24 01:37:25 +00:00
if (LIKELY(settings->on_##FOR)) { \
parser->state = CURRENT_STATE(); \
if (UNLIKELY(0 != settings->on_##FOR(parser))) { \
2019-04-23 00:31:21 +00:00
SET_ERRNO(HPE_CB_##FOR); \
} \
2019-04-24 01:37:25 +00:00
UPDATE_STATE(parser->state); \
2019-04-23 00:31:21 +00:00
\
/* We either errored above or got paused; get out */ \
2019-04-24 01:37:25 +00:00
if (UNLIKELY(HTTP_PARSER_ERRNO(parser) != HPE_OK)) { \
2019-04-23 00:31:21 +00:00
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 { \
2019-04-23 00:31:21 +00:00
assert(HTTP_PARSER_ERRNO(parser) == HPE_OK); \
\
if (FOR##_mark) { \
2019-04-24 01:37:25 +00:00
if (LIKELY(settings->on_##FOR)) { \
parser->state = CURRENT_STATE(); \
if (UNLIKELY(0 != \
settings->on_##FOR(parser, FOR##_mark, (LEN)))) { \
2019-04-23 00:31:21 +00:00
SET_ERRNO(HPE_CB_##FOR); \
} \
2019-04-24 01:37:25 +00:00
UPDATE_STATE(parser->state); \
2019-04-23 00:31:21 +00:00
\
/* We either errored above or got paused; get out */ \
2019-04-24 01:37:25 +00:00
if (UNLIKELY(HTTP_PARSER_ERRNO(parser) != HPE_OK)) { \
2019-04-23 00:31:21 +00:00
return (ER); \
} \
} \
FOR##_mark = NULL; \
} \
} while (0)
/* Run the data callback FOR and consume the current byte */
#define CALLBACK_DATA(FOR) \
2019-04-23 00:31:21 +00:00
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) \
2019-04-23 00:31:21 +00:00
CALLBACK_DATA_(FOR, p - FOR##_mark, p - data)
/* Set the mark FOR; non-destructive if mark is already set */
#define MARK(FOR) \
do { \
2019-04-23 00:31:21 +00:00
if (!FOR##_mark) { \
FOR##_mark = p; \
} \
} while (0)
2019-04-24 01:37:25 +00:00
/* Don't allow the total size of the HTTP headers (including the status
* line) to exceed max_header_size. This check is here to protect
* embedders against denial-of-service attacks where the attacker feeds
* us a never-ending header that the embedder keeps buffering.
*
* This check is arguably the responsibility of embedders but we're doing
* it on the embedder's behalf because most won't bother and this way we
* make the web a little safer. max_header_size is still far bigger
* than any reasonable request or response so this should never affect
* day-to-day operation.
*/
#define COUNT_HEADER_SIZE(V) \
do { \
nread += (uint32_t)(V); \
if (UNLIKELY(nread > max_header_size)) { \
SET_ERRNO(HPE_HEADER_OVERFLOW); \
goto error; \
} \
} 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[] =
2019-04-23 00:31:21 +00:00
{
#define XX(num, name, string) #string,
2019-04-23 00:31:21 +00:00
HTTP_METHOD_MAP(XX)
#undef XX
2019-04-23 00:31:21 +00:00
};
/* Tokens as defined by rfc 2616. Also lowercases them.
* token = 1*<any CHAR except CTLs or separators>
* separators = "(" | ")" | "<" | ">" | "@"
* | "," | ";" | ":" | "\" | <">
* | "/" | "[" | "]" | "?" | "="
* | "{" | "}" | SP | HT
*/
static const char tokens[256] = {
2019-04-23 00:31:21 +00:00
/* 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 ' */
2019-04-24 01:37:25 +00:00
' ', '!', 0, '#', '$', '%', '&', '\'',
2019-04-23 00:31:21 +00:00
/* 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] =
2019-04-23 00:31:21 +00:00
{-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
};
#if HTTP_PARSER_STRICT
# define T(v) 0
#else
# define T(v) v
#endif
static const uint8_t normal_url_char[32] = {
2019-04-23 00:31:21 +00:00
/* 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 | T(2) | 0 | 0 | T(16) | 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 | 2 | 4 | 0 | 16 | 32 | 64 | 128,
/* 40 ( 41 ) 42 * 43 + 44 , 45 - 46 . 47 / */
1 | 2 | 4 | 8 | 16 | 32 | 64 | 128,
/* 48 0 49 1 50 2 51 3 52 4 53 5 54 6 55 7 */
1 | 2 | 4 | 8 | 16 | 32 | 64 | 128,
/* 56 8 57 9 58 : 59 ; 60 < 61 = 62 > 63 ? */
1 | 2 | 4 | 8 | 16 | 32 | 64 | 0,
/* 64 @ 65 A 66 B 67 C 68 D 69 E 70 F 71 G */
1 | 2 | 4 | 8 | 16 | 32 | 64 | 128,
/* 72 H 73 I 74 J 75 K 76 L 77 M 78 N 79 O */
1 | 2 | 4 | 8 | 16 | 32 | 64 | 128,
/* 80 P 81 Q 82 R 83 S 84 T 85 U 86 V 87 W */
1 | 2 | 4 | 8 | 16 | 32 | 64 | 128,
/* 88 X 89 Y 90 Z 91 [ 92 \ 93 ] 94 ^ 95 _ */
1 | 2 | 4 | 8 | 16 | 32 | 64 | 128,
/* 96 ` 97 a 98 b 99 c 100 d 101 e 102 f 103 g */
1 | 2 | 4 | 8 | 16 | 32 | 64 | 128,
/* 104 h 105 i 106 j 107 k 108 l 109 m 110 n 111 o */
1 | 2 | 4 | 8 | 16 | 32 | 64 | 128,
/* 112 p 113 q 114 r 115 s 116 t 117 u 118 v 119 w */
1 | 2 | 4 | 8 | 16 | 32 | 64 | 128,
/* 120 x 121 y 122 z 123 { 124 | 125 } 126 ~ 127 del */
1 | 2 | 4 | 8 | 16 | 32 | 64 | 0, };
#undef T
enum state
2019-04-23 00:31:21 +00:00
{ 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_http_major
2019-04-24 01:37:25 +00:00
, s_res_http_dot
2019-04-23 00:31:21 +00:00
, s_res_http_minor
2019-04-24 01:37:25 +00:00
, s_res_http_end
2019-04-23 00:31:21 +00:00
, s_res_first_status_code
, s_res_status_code
2019-04-24 01:37:25 +00:00
, s_res_status_start
2019-04-23 00:31:21 +00:00
, 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_server_start
, s_req_server
, s_req_server_with_at
, 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
2019-04-24 01:37:25 +00:00
, s_req_http_I
, s_req_http_IC
2019-04-23 00:31:21 +00:00
, s_req_http_major
2019-04-24 01:37:25 +00:00
, s_req_http_dot
2019-04-23 00:31:21 +00:00
, s_req_http_minor
2019-04-24 01:37:25 +00:00
, s_req_http_end
2019-04-23 00:31:21 +00:00
, s_req_line_almost_done
, s_header_field_start
, s_header_field
2019-04-24 01:37:25 +00:00
, s_header_value_discard_ws
, s_header_value_discard_ws_almost_done
, s_header_value_discard_lws
2019-04-23 00:31:21 +00:00
, 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
2019-04-23 00:31:21 +00:00
{ 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
2019-04-24 01:37:25 +00:00
, h_content_length_num
, h_content_length_ws
2019-04-23 00:31:21 +00:00
, h_transfer_encoding
, h_upgrade
, h_matching_transfer_encoding_chunked
2019-04-24 01:37:25 +00:00
, h_matching_connection_token_start
2019-04-23 00:31:21 +00:00
, h_matching_connection_keep_alive
, h_matching_connection_close
2019-04-24 01:37:25 +00:00
, h_matching_connection_upgrade
, h_matching_connection_token
2019-04-23 00:31:21 +00:00
, h_transfer_encoding_chunked
, h_connection_keep_alive
, h_connection_close
2019-04-24 01:37:25 +00:00
, h_connection_upgrade
2019-04-23 00:31:21 +00:00
};
enum http_host_state
2019-04-23 00:31:21 +00:00
{
s_http_host_dead = 1
2019-04-23 00:31:21 +00:00
, s_http_userinfo_start
, s_http_userinfo
, s_http_host_start
, s_http_host_v6_start
, s_http_host
, s_http_host_v6
, s_http_host_v6_end
2019-04-24 01:37:25 +00:00
, s_http_host_v6_zone_start
, s_http_host_v6_zone
2019-04-23 00:31:21 +00:00
, s_http_host_port_start
, s_http_host_port
};
/* 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'))
#define IS_MARK(c) ((c) == '-' || (c) == '_' || (c) == '.' || \
2019-04-23 00:31:21 +00:00
(c) == '!' || (c) == '~' || (c) == '*' || (c) == '\'' || (c) == '(' || \
(c) == ')')
#define IS_USERINFO_CHAR(c) (IS_ALPHANUM(c) || IS_MARK(c) || (c) == '%' || \
2019-04-23 00:31:21 +00:00
(c) == ';' || (c) == ':' || (c) == '&' || (c) == '=' || (c) == '+' || \
(c) == '$' || (c) == ',')
2019-04-24 01:37:25 +00:00
#define STRICT_TOKEN(c) ((c == ' ') ? 0 : tokens[(unsigned char)c])
#if HTTP_PARSER_STRICT
2019-04-24 01:37:25 +00:00
#define TOKEN(c) STRICT_TOKEN(c)
#define IS_URL_CHAR(c) (BIT_AT(normal_url_char, (unsigned char)c))
#define IS_HOST_CHAR(c) (IS_ALPHANUM(c) || (c) == '.' || (c) == '-')
#else
2019-04-24 01:37:25 +00:00
#define TOKEN(c) tokens[(unsigned char)c]
#define IS_URL_CHAR(c) \
2019-04-23 00:31:21 +00:00
(BIT_AT(normal_url_char, (unsigned char)c) || ((c) & 0x80))
#define IS_HOST_CHAR(c) \
2019-04-23 00:31:21 +00:00
(IS_ALPHANUM(c) || (c) == '.' || (c) == '-' || (c) == '_')
#endif
2019-04-24 01:37:25 +00:00
/**
* Verify that a char is a valid visible (printable) US-ASCII
* character or %x80-FF
**/
#define IS_HEADER_CHAR(ch) \
(ch == CR || ch == LF || ch == 9 || ((unsigned char)ch > 31 && ch != 127))
#define start_state (parser->type == HTTP_REQUEST ? s_start_req : s_start_res)
#if HTTP_PARSER_STRICT
# define STRICT_CHECK(cond) \
do { \
2019-04-23 00:31:21 +00:00
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 {
2019-04-23 00:31:21 +00:00
const char *name;
const char *description;
} http_strerror_tab[] = {
2019-04-23 00:31:21 +00:00
HTTP_ERRNO_MAP(HTTP_STRERROR_GEN)
};
#undef HTTP_STRERROR_GEN
int http_message_needs_eof(const http_parser *parser);
/* Our URL parser.
*
* This is designed to be shared by http_parser_execute() for URL validation,
* hence it has a state transition + byte-for-byte interface. In addition, it
* is meant to be embedded in http_parser_parse_url(), which does the dirty
* work of turning state transitions URL components for its API.
*
* This function should only be invoked with non-space characters. It is
* assumed that the caller cares about (and can detect) the transition between
* URL and non-URL states by looking for these.
*/
static enum state
parse_url_char(enum state s, const char ch)
{
2019-04-23 00:31:21 +00:00
if (ch == ' ' || ch == '\r' || ch == '\n') {
return s_dead;
}
#if HTTP_PARSER_STRICT
2019-04-23 00:31:21 +00:00
if (ch == '\t' || ch == '\f') {
return s_dead;
2019-04-23 00:31:21 +00:00
}
#endif
switch (s) {
case s_req_spaces_before_url:
/* Proxied requests are followed by scheme of an absolute URI (alpha).
* All methods except CONNECT are followed by '/' or '*'.
*/
if (ch == '/' || ch == '*') {
return s_req_path;
}
if (IS_ALPHA(ch)) {
return s_req_schema;
}
break;
case s_req_schema:
if (IS_ALPHA(ch)) {
return s;
}
if (ch == ':') {
return s_req_schema_slash;
}
break;
case s_req_schema_slash:
if (ch == '/') {
return s_req_schema_slash_slash;
}
break;
case s_req_schema_slash_slash:
if (ch == '/') {
return s_req_server_start;
}
break;
case s_req_server_with_at:
if (ch == '@') {
return s_dead;
}
2019-04-24 01:37:25 +00:00
/* fall through */
2019-04-23 00:31:21 +00:00
case s_req_server_start:
case s_req_server:
if (ch == '/') {
return s_req_path;
}
if (ch == '?') {
return s_req_query_string_start;
}
if (ch == '@') {
return s_req_server_with_at;
}
if (IS_USERINFO_CHAR(ch) || ch == '[' || ch == ']') {
return s_req_server;
}
break;
case s_req_path:
if (IS_URL_CHAR(ch)) {
return s;
}
switch (ch) {
case '?':
return s_req_query_string_start;
case '#':
return s_req_fragment_start;
}
break;
case s_req_query_string_start:
case s_req_query_string:
if (IS_URL_CHAR(ch)) {
return s_req_query_string;
}
switch (ch) {
case '?':
/* allow extra '?' in query string */
return s_req_query_string;
case '#':
return s_req_fragment_start;
}
break;
case s_req_fragment_start:
if (IS_URL_CHAR(ch)) {
return s_req_fragment;
}
switch (ch) {
case '?':
return s_req_fragment;
case '#':
return s;
}
break;
case s_req_fragment:
if (IS_URL_CHAR(ch)) {
return s;
}
switch (ch) {
case '?':
case '#':
return s;
}
break;
default:
break;
}
/* We should never fall out of the switch above unless there's an error */
return s_dead;
}
size_t http_parser_execute (http_parser *parser,
const http_parser_settings *settings,
const char *data,
size_t len)
{
2019-04-23 00:31:21 +00:00
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;
2019-04-24 01:37:25 +00:00
const char *status_mark = 0;
enum state p_state = (enum state) parser->state;
const unsigned int lenient = parser->lenient_http_headers;
uint32_t nread = parser->nread;
2019-04-23 00:31:21 +00:00
/* We're in an error state. Don't bother doing anything. */
if (HTTP_PARSER_ERRNO(parser) != HPE_OK) {
return 0;
}
2019-04-23 00:31:21 +00:00
if (len == 0) {
2019-04-24 01:37:25 +00:00
switch (CURRENT_STATE()) {
2019-04-23 00:31:21 +00:00
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;
2019-04-23 00:31:21 +00:00
case s_dead:
case s_start_req_or_res:
case s_start_res:
case s_start_req:
return 0;
2019-04-23 00:31:21 +00:00
default:
SET_ERRNO(HPE_INVALID_EOF_STATE);
return 1;
}
2019-04-23 00:31:21 +00:00
}
2019-04-24 01:37:25 +00:00
if (CURRENT_STATE() == s_header_field)
2019-04-23 00:31:21 +00:00
header_field_mark = data;
2019-04-24 01:37:25 +00:00
if (CURRENT_STATE() == s_header_value)
2019-04-23 00:31:21 +00:00
header_value_mark = data;
2019-04-24 01:37:25 +00:00
switch (CURRENT_STATE()) {
2019-04-23 00:31:21 +00:00
case s_req_path:
case s_req_schema:
case s_req_schema_slash:
case s_req_schema_slash_slash:
case s_req_server_start:
case s_req_server:
case s_req_server_with_at:
case s_req_query_string_start:
case s_req_query_string:
case s_req_fragment_start:
case s_req_fragment:
url_mark = data;
break;
2019-04-24 01:37:25 +00:00
case s_res_status:
status_mark = data;
break;
default:
break;
2019-04-23 00:31:21 +00:00
}
for (p=data; p != data + len; p++) {
ch = *p;
2019-04-24 01:37:25 +00:00
if (PARSING_HEADER(CURRENT_STATE()))
COUNT_HEADER_SIZE(1);
2019-04-24 01:37:25 +00:00
reexecute:
switch (CURRENT_STATE()) {
2019-04-23 00:31:21 +00:00
case s_dead:
/* this state is used after a 'Connection: close' message
* the parser will error out if it reads another message
*/
2019-04-24 01:37:25 +00:00
if (LIKELY(ch == CR || ch == LF))
2019-04-23 00:31:21 +00:00
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') {
2019-04-24 01:37:25 +00:00
UPDATE_STATE(s_res_or_resp_H);
2019-04-23 00:31:21 +00:00
CALLBACK_NOTIFY(message_begin);
} else {
parser->type = HTTP_REQUEST;
2019-04-24 01:37:25 +00:00
UPDATE_STATE(s_start_req);
REEXECUTE();
}
2019-04-23 00:31:21 +00:00
break;
}
case s_res_or_resp_H:
if (ch == 'T') {
parser->type = HTTP_RESPONSE;
2019-04-24 01:37:25 +00:00
UPDATE_STATE(s_res_HT);
2019-04-23 00:31:21 +00:00
} else {
2019-04-24 01:37:25 +00:00
if (UNLIKELY(ch != 'E')) {
2019-04-23 00:31:21 +00:00
SET_ERRNO(HPE_INVALID_CONSTANT);
goto error;
}
parser->type = HTTP_REQUEST;
parser->method = HTTP_HEAD;
parser->index = 2;
2019-04-24 01:37:25 +00:00
UPDATE_STATE(s_req_method);
}
2019-04-23 00:31:21 +00:00
break;
2019-04-23 00:31:21 +00:00
case s_start_res:
{
2019-04-24 01:37:25 +00:00
if (ch == CR || ch == LF)
break;
2019-04-23 00:31:21 +00:00
parser->flags = 0;
parser->content_length = ULLONG_MAX;
2019-04-24 01:37:25 +00:00
if (ch == 'H') {
UPDATE_STATE(s_res_H);
} else {
SET_ERRNO(HPE_INVALID_CONSTANT);
goto error;
2019-04-23 00:31:21 +00:00
}
2019-04-23 00:31:21 +00:00
CALLBACK_NOTIFY(message_begin);
break;
}
case s_res_H:
STRICT_CHECK(ch != 'T');
2019-04-24 01:37:25 +00:00
UPDATE_STATE(s_res_HT);
2019-04-23 00:31:21 +00:00
break;
case s_res_HT:
STRICT_CHECK(ch != 'T');
2019-04-24 01:37:25 +00:00
UPDATE_STATE(s_res_HTT);
2019-04-23 00:31:21 +00:00
break;
case s_res_HTT:
STRICT_CHECK(ch != 'P');
2019-04-24 01:37:25 +00:00
UPDATE_STATE(s_res_HTTP);
2019-04-23 00:31:21 +00:00
break;
case s_res_HTTP:
STRICT_CHECK(ch != '/');
2019-04-24 01:37:25 +00:00
UPDATE_STATE(s_res_http_major);
2019-04-23 00:31:21 +00:00
break;
2019-04-24 01:37:25 +00:00
case s_res_http_major:
if (UNLIKELY(!IS_NUM(ch))) {
2019-04-23 00:31:21 +00:00
SET_ERRNO(HPE_INVALID_VERSION);
goto error;
}
2019-04-23 00:31:21 +00:00
parser->http_major = ch - '0';
2019-04-24 01:37:25 +00:00
UPDATE_STATE(s_res_http_dot);
2019-04-23 00:31:21 +00:00
break;
2019-04-24 01:37:25 +00:00
case s_res_http_dot:
2019-04-23 00:31:21 +00:00
{
2019-04-24 01:37:25 +00:00
if (UNLIKELY(ch != '.')) {
2019-04-23 00:31:21 +00:00
SET_ERRNO(HPE_INVALID_VERSION);
goto error;
}
2019-04-24 01:37:25 +00:00
UPDATE_STATE(s_res_http_minor);
2019-04-23 00:31:21 +00:00
break;
}
2019-04-24 01:37:25 +00:00
case s_res_http_minor:
if (UNLIKELY(!IS_NUM(ch))) {
2019-04-23 00:31:21 +00:00
SET_ERRNO(HPE_INVALID_VERSION);
goto error;
}
parser->http_minor = ch - '0';
2019-04-24 01:37:25 +00:00
UPDATE_STATE(s_res_http_end);
2019-04-23 00:31:21 +00:00
break;
2019-04-24 01:37:25 +00:00
case s_res_http_end:
2019-04-23 00:31:21 +00:00
{
2019-04-24 01:37:25 +00:00
if (UNLIKELY(ch != ' ')) {
2019-04-23 00:31:21 +00:00
SET_ERRNO(HPE_INVALID_VERSION);
goto error;
}
2019-04-24 01:37:25 +00:00
UPDATE_STATE(s_res_first_status_code);
2019-04-23 00:31:21 +00:00
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';
2019-04-24 01:37:25 +00:00
UPDATE_STATE(s_res_status_code);
2019-04-23 00:31:21 +00:00
break;
}
case s_res_status_code:
{
if (!IS_NUM(ch)) {
switch (ch) {
case ' ':
2019-04-24 01:37:25 +00:00
UPDATE_STATE(s_res_status_start);
2019-04-23 00:31:21 +00:00
break;
case CR:
case LF:
2019-04-24 01:37:25 +00:00
UPDATE_STATE(s_res_status_start);
REEXECUTE();
2019-04-23 00:31:21 +00:00
break;
default:
SET_ERRNO(HPE_INVALID_STATUS);
goto error;
}
break;
}
parser->status_code *= 10;
parser->status_code += ch - '0';
2019-04-24 01:37:25 +00:00
if (UNLIKELY(parser->status_code > 999)) {
2019-04-23 00:31:21 +00:00
SET_ERRNO(HPE_INVALID_STATUS);
goto error;
}
break;
}
2019-04-24 01:37:25 +00:00
case s_res_status_start:
{
MARK(status);
UPDATE_STATE(s_res_status);
parser->index = 0;
if (ch == CR || ch == LF)
REEXECUTE();
break;
}
2019-04-23 00:31:21 +00:00
case s_res_status:
if (ch == CR) {
2019-04-24 01:37:25 +00:00
UPDATE_STATE(s_res_line_almost_done);
CALLBACK_DATA(status);
2019-04-23 00:31:21 +00:00
break;
}
if (ch == LF) {
2019-04-24 01:37:25 +00:00
UPDATE_STATE(s_header_field_start);
CALLBACK_DATA(status);
2019-04-23 00:31:21 +00:00
break;
}
2019-04-24 01:37:25 +00:00
2019-04-23 00:31:21 +00:00
break;
case s_res_line_almost_done:
STRICT_CHECK(ch != LF);
2019-04-24 01:37:25 +00:00
UPDATE_STATE(s_header_field_start);
2019-04-23 00:31:21 +00:00
break;
case s_start_req:
{
if (ch == CR || ch == LF)
break;
parser->flags = 0;
parser->content_length = ULLONG_MAX;
2019-04-24 01:37:25 +00:00
if (UNLIKELY(!IS_ALPHA(ch))) {
2019-04-23 00:31:21 +00:00
SET_ERRNO(HPE_INVALID_METHOD);
goto error;
}
parser->method = (enum http_method) 0;
parser->index = 1;
switch (ch) {
2019-04-24 01:37:25 +00:00
case 'A': parser->method = HTTP_ACL; break;
case 'B': parser->method = HTTP_BIND; break;
2019-04-23 00:31:21 +00:00
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;
2019-04-24 01:37:25 +00:00
case 'L': parser->method = HTTP_LOCK; /* or LINK */ break;
case 'M': parser->method = HTTP_MKCOL; /* or MOVE, MKACTIVITY, MERGE, M-SEARCH, MKCALENDAR */ break;
2019-04-23 00:31:21 +00:00
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;
2019-04-24 01:37:25 +00:00
case 'R': parser->method = HTTP_REPORT; /* or REBIND */ break;
case 'S': parser->method = HTTP_SUBSCRIBE; /* or SEARCH, SOURCE */ break;
2019-04-23 00:31:21 +00:00
case 'T': parser->method = HTTP_TRACE; break;
2019-04-24 01:37:25 +00:00
case 'U': parser->method = HTTP_UNLOCK; /* or UNSUBSCRIBE, UNBIND, UNLINK */ break;
2019-04-23 00:31:21 +00:00
default:
SET_ERRNO(HPE_INVALID_METHOD);
goto error;
}
2019-04-24 01:37:25 +00:00
UPDATE_STATE(s_req_method);
2019-04-23 00:31:21 +00:00
CALLBACK_NOTIFY(message_begin);
break;
}
case s_req_method:
{
const char *matcher;
2019-04-24 01:37:25 +00:00
if (UNLIKELY(ch == '\0')) {
2019-04-23 00:31:21 +00:00
SET_ERRNO(HPE_INVALID_METHOD);
goto error;
}
matcher = method_strings[parser->method];
if (ch == ' ' && matcher[parser->index] == '\0') {
2019-04-24 01:37:25 +00:00
UPDATE_STATE(s_req_spaces_before_url);
2019-04-23 00:31:21 +00:00
} else if (ch == matcher[parser->index]) {
; /* nada */
2019-04-24 01:37:25 +00:00
} else if ((ch >= 'A' && ch <= 'Z') || ch == '-') {
switch (parser->method << 16 | parser->index << 8 | ch) {
#define XX(meth, pos, ch, new_meth) \
case (HTTP_##meth << 16 | pos << 8 | ch): \
parser->method = HTTP_##new_meth; break;
XX(POST, 1, 'U', PUT)
XX(POST, 1, 'A', PATCH)
XX(POST, 1, 'R', PROPFIND)
XX(PUT, 2, 'R', PURGE)
XX(CONNECT, 1, 'H', CHECKOUT)
XX(CONNECT, 2, 'P', COPY)
XX(MKCOL, 1, 'O', MOVE)
XX(MKCOL, 1, 'E', MERGE)
XX(MKCOL, 1, '-', MSEARCH)
XX(MKCOL, 2, 'A', MKACTIVITY)
XX(MKCOL, 3, 'A', MKCALENDAR)
XX(SUBSCRIBE, 1, 'E', SEARCH)
XX(SUBSCRIBE, 1, 'O', SOURCE)
XX(REPORT, 2, 'B', REBIND)
XX(PROPFIND, 4, 'P', PROPPATCH)
XX(LOCK, 1, 'I', LINK)
XX(UNLOCK, 2, 'S', UNSUBSCRIBE)
XX(UNLOCK, 2, 'B', UNBIND)
XX(UNLOCK, 3, 'I', UNLINK)
#undef XX
default:
SET_ERRNO(HPE_INVALID_METHOD);
goto error;
2019-04-23 00:31:21 +00:00
}
} 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) {
2019-04-24 01:37:25 +00:00
UPDATE_STATE(s_req_server_start);
2019-04-23 00:31:21 +00:00
}
2019-04-24 01:37:25 +00:00
UPDATE_STATE(parse_url_char(CURRENT_STATE(), ch));
if (UNLIKELY(CURRENT_STATE() == s_dead)) {
2019-04-23 00:31:21 +00:00
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_server_start:
{
switch (ch) {
/* No whitespace allowed here */
case ' ':
case CR:
case LF:
SET_ERRNO(HPE_INVALID_URL);
goto error;
default:
2019-04-24 01:37:25 +00:00
UPDATE_STATE(parse_url_char(CURRENT_STATE(), ch));
if (UNLIKELY(CURRENT_STATE() == s_dead)) {
2019-04-23 00:31:21 +00:00
SET_ERRNO(HPE_INVALID_URL);
goto error;
}
2019-04-23 00:31:21 +00:00
}
break;
}
case s_req_server:
case s_req_server_with_at:
case s_req_path:
case s_req_query_string_start:
case s_req_query_string:
case s_req_fragment_start:
case s_req_fragment:
{
switch (ch) {
case ' ':
2019-04-24 01:37:25 +00:00
UPDATE_STATE(s_req_http_start);
2019-04-23 00:31:21 +00:00
CALLBACK_DATA(url);
break;
case CR:
case LF:
parser->http_major = 0;
parser->http_minor = 9;
2019-04-24 01:37:25 +00:00
UPDATE_STATE((ch == CR) ?
2019-04-23 00:31:21 +00:00
s_req_line_almost_done :
2019-04-24 01:37:25 +00:00
s_header_field_start);
2019-04-23 00:31:21 +00:00
CALLBACK_DATA(url);
break;
default:
2019-04-24 01:37:25 +00:00
UPDATE_STATE(parse_url_char(CURRENT_STATE(), ch));
if (UNLIKELY(CURRENT_STATE() == s_dead)) {
2019-04-23 00:31:21 +00:00
SET_ERRNO(HPE_INVALID_URL);
goto error;
}
2019-04-23 00:31:21 +00:00
}
break;
}
case s_req_http_start:
switch (ch) {
case ' ':
break;
2019-04-24 01:37:25 +00:00
case 'H':
UPDATE_STATE(s_req_http_H);
break;
case 'I':
if (parser->method == HTTP_SOURCE) {
UPDATE_STATE(s_req_http_I);
break;
}
/* fall through */
2019-04-23 00:31:21 +00:00
default:
SET_ERRNO(HPE_INVALID_CONSTANT);
goto error;
}
break;
case s_req_http_H:
STRICT_CHECK(ch != 'T');
2019-04-24 01:37:25 +00:00
UPDATE_STATE(s_req_http_HT);
2019-04-23 00:31:21 +00:00
break;
case s_req_http_HT:
STRICT_CHECK(ch != 'T');
2019-04-24 01:37:25 +00:00
UPDATE_STATE(s_req_http_HTT);
2019-04-23 00:31:21 +00:00
break;
case s_req_http_HTT:
STRICT_CHECK(ch != 'P');
2019-04-24 01:37:25 +00:00
UPDATE_STATE(s_req_http_HTTP);
break;
case s_req_http_I:
STRICT_CHECK(ch != 'C');
UPDATE_STATE(s_req_http_IC);
break;
case s_req_http_IC:
STRICT_CHECK(ch != 'E');
UPDATE_STATE(s_req_http_HTTP); /* Treat "ICE" as "HTTP". */
2019-04-23 00:31:21 +00:00
break;
case s_req_http_HTTP:
STRICT_CHECK(ch != '/');
2019-04-24 01:37:25 +00:00
UPDATE_STATE(s_req_http_major);
2019-04-23 00:31:21 +00:00
break;
2019-04-24 01:37:25 +00:00
case s_req_http_major:
if (UNLIKELY(!IS_NUM(ch))) {
2019-04-23 00:31:21 +00:00
SET_ERRNO(HPE_INVALID_VERSION);
goto error;
}
parser->http_major = ch - '0';
2019-04-24 01:37:25 +00:00
UPDATE_STATE(s_req_http_dot);
2019-04-23 00:31:21 +00:00
break;
2019-04-24 01:37:25 +00:00
case s_req_http_dot:
2019-04-23 00:31:21 +00:00
{
2019-04-24 01:37:25 +00:00
if (UNLIKELY(ch != '.')) {
2019-04-23 00:31:21 +00:00
SET_ERRNO(HPE_INVALID_VERSION);
goto error;
}
2019-04-24 01:37:25 +00:00
UPDATE_STATE(s_req_http_minor);
2019-04-23 00:31:21 +00:00
break;
}
2019-04-24 01:37:25 +00:00
case s_req_http_minor:
if (UNLIKELY(!IS_NUM(ch))) {
2019-04-23 00:31:21 +00:00
SET_ERRNO(HPE_INVALID_VERSION);
goto error;
}
parser->http_minor = ch - '0';
2019-04-24 01:37:25 +00:00
UPDATE_STATE(s_req_http_end);
2019-04-23 00:31:21 +00:00
break;
2019-04-24 01:37:25 +00:00
case s_req_http_end:
2019-04-23 00:31:21 +00:00
{
if (ch == CR) {
2019-04-24 01:37:25 +00:00
UPDATE_STATE(s_req_line_almost_done);
2019-04-23 00:31:21 +00:00
break;
}
if (ch == LF) {
2019-04-24 01:37:25 +00:00
UPDATE_STATE(s_header_field_start);
2019-04-23 00:31:21 +00:00
break;
}
2019-04-24 01:37:25 +00:00
SET_ERRNO(HPE_INVALID_VERSION);
goto error;
2019-04-23 00:31:21 +00:00
break;
}
/* end of request line */
case s_req_line_almost_done:
{
2019-04-24 01:37:25 +00:00
if (UNLIKELY(ch != LF)) {
2019-04-23 00:31:21 +00:00
SET_ERRNO(HPE_LF_EXPECTED);
goto error;
}
2019-04-24 01:37:25 +00:00
UPDATE_STATE(s_header_field_start);
2019-04-23 00:31:21 +00:00
break;
}
case s_header_field_start:
{
if (ch == CR) {
2019-04-24 01:37:25 +00:00
UPDATE_STATE(s_headers_almost_done);
2019-04-23 00:31:21 +00:00
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*/
2019-04-24 01:37:25 +00:00
UPDATE_STATE(s_headers_almost_done);
REEXECUTE();
2019-04-23 00:31:21 +00:00
}
c = TOKEN(ch);
2019-04-24 01:37:25 +00:00
if (UNLIKELY(!c)) {
2019-04-23 00:31:21 +00:00
SET_ERRNO(HPE_INVALID_HEADER_TOKEN);
goto error;
}
MARK(header_field);
parser->index = 0;
2019-04-24 01:37:25 +00:00
UPDATE_STATE(s_header_field);
2019-04-23 00:31:21 +00:00
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:
{
2019-04-24 01:37:25 +00:00
const char* start = p;
for (; p != data + len; p++) {
ch = *p;
c = TOKEN(ch);
if (!c)
break;
2019-04-23 00:31:21 +00:00
switch (parser->header_state) {
2019-04-24 01:37:25 +00:00
case h_general: {
size_t left = data + len - p;
const char* pe = p + MIN(left, max_header_size);
while (p+1 < pe && TOKEN(p[1])) {
p++;
}
2019-04-23 00:31:21 +00:00
break;
2019-04-24 01:37:25 +00:00
}
2019-04-23 00:31:21 +00:00
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;
}
}
2019-04-24 01:37:25 +00:00
if (p == data + len) {
--p;
COUNT_HEADER_SIZE(p - start);
2019-04-23 00:31:21 +00:00
break;
}
2019-04-24 01:37:25 +00:00
COUNT_HEADER_SIZE(p - start);
2019-04-23 00:31:21 +00:00
2019-04-24 01:37:25 +00:00
if (ch == ':') {
UPDATE_STATE(s_header_value_discard_ws);
2019-04-23 00:31:21 +00:00
CALLBACK_DATA(header_field);
break;
}
SET_ERRNO(HPE_INVALID_HEADER_TOKEN);
goto error;
}
2019-04-24 01:37:25 +00:00
case s_header_value_discard_ws:
2019-04-23 00:31:21 +00:00
if (ch == ' ' || ch == '\t') break;
if (ch == CR) {
2019-04-24 01:37:25 +00:00
UPDATE_STATE(s_header_value_discard_ws_almost_done);
2019-04-23 00:31:21 +00:00
break;
}
if (ch == LF) {
2019-04-24 01:37:25 +00:00
UPDATE_STATE(s_header_value_discard_lws);
2019-04-23 00:31:21 +00:00
break;
}
2019-04-24 01:37:25 +00:00
/* fall through */
case s_header_value_start:
{
MARK(header_value);
UPDATE_STATE(s_header_value);
parser->index = 0;
2019-04-23 00:31:21 +00:00
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;
}
2019-04-23 00:31:21 +00:00
break;
case h_content_length:
2019-04-24 01:37:25 +00:00
if (UNLIKELY(!IS_NUM(ch))) {
2019-04-23 00:31:21 +00:00
SET_ERRNO(HPE_INVALID_CONTENT_LENGTH);
goto error;
}
2019-04-23 00:31:21 +00:00
2019-04-24 01:37:25 +00:00
if (parser->flags & F_CONTENTLENGTH) {
SET_ERRNO(HPE_UNEXPECTED_CONTENT_LENGTH);
goto error;
}
parser->flags |= F_CONTENTLENGTH;
2019-04-23 00:31:21 +00:00
parser->content_length = ch - '0';
2019-04-24 01:37:25 +00:00
parser->header_state = h_content_length_num;
break;
/* when obsolete line folding is encountered for content length
* continue to the s_header_value state */
case h_content_length_ws:
break;
2019-04-23 00:31:21 +00:00
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;
2019-04-24 01:37:25 +00:00
} else if (c == 'u') {
parser->header_state = h_matching_connection_upgrade;
2019-04-23 00:31:21 +00:00
} else {
2019-04-24 01:37:25 +00:00
parser->header_state = h_matching_connection_token;
}
2019-04-23 00:31:21 +00:00
break;
2019-04-24 01:37:25 +00:00
/* Multi-value `Connection` header */
case h_matching_connection_token_start:
break;
2019-04-23 00:31:21 +00:00
default:
parser->header_state = h_general;
break;
}
break;
}
case s_header_value:
{
2019-04-24 01:37:25 +00:00
const char* start = p;
enum header_states h_state = (enum header_states) parser->header_state;
for (; p != data + len; p++) {
ch = *p;
if (ch == CR) {
UPDATE_STATE(s_header_almost_done);
parser->header_state = h_state;
CALLBACK_DATA(header_value);
break;
}
2019-04-23 00:31:21 +00:00
2019-04-24 01:37:25 +00:00
if (ch == LF) {
UPDATE_STATE(s_header_almost_done);
COUNT_HEADER_SIZE(p - start);
parser->header_state = h_state;
CALLBACK_DATA_NOADVANCE(header_value);
REEXECUTE();
}
2019-04-23 00:31:21 +00:00
2019-04-24 01:37:25 +00:00
if (!lenient && !IS_HEADER_CHAR(ch)) {
SET_ERRNO(HPE_INVALID_HEADER_TOKEN);
goto error;
}
2019-04-23 00:31:21 +00:00
2019-04-24 01:37:25 +00:00
c = LOWER(ch);
2019-04-23 00:31:21 +00:00
2019-04-24 01:37:25 +00:00
switch (h_state) {
case h_general:
{
size_t left = data + len - p;
const char* pe = p + MIN(left, max_header_size);
for (; p != pe; p++) {
ch = *p;
if (ch == CR || ch == LF) {
--p;
break;
}
if (!lenient && !IS_HEADER_CHAR(ch)) {
SET_ERRNO(HPE_INVALID_HEADER_TOKEN);
goto error;
}
}
if (p == data + len)
--p;
break;
}
2019-04-23 00:31:21 +00:00
2019-04-24 01:37:25 +00:00
case h_connection:
case h_transfer_encoding:
assert(0 && "Shouldn't get here.");
break;
2019-04-23 00:31:21 +00:00
2019-04-24 01:37:25 +00:00
case h_content_length:
if (ch == ' ') break;
h_state = h_content_length_num;
/* fall through */
2019-04-23 00:31:21 +00:00
2019-04-24 01:37:25 +00:00
case h_content_length_num:
{
uint64_t t;
2019-04-23 00:31:21 +00:00
2019-04-24 01:37:25 +00:00
if (ch == ' ') {
h_state = h_content_length_ws;
break;
}
if (UNLIKELY(!IS_NUM(ch))) {
SET_ERRNO(HPE_INVALID_CONTENT_LENGTH);
parser->header_state = h_state;
goto error;
}
t = parser->content_length;
t *= 10;
t += ch - '0';
2019-04-23 00:31:21 +00:00
2019-04-24 01:37:25 +00:00
/* Overflow? Test against a conservative limit for simplicity. */
if (UNLIKELY((ULLONG_MAX - 10) / 10 < parser->content_length)) {
SET_ERRNO(HPE_INVALID_CONTENT_LENGTH);
parser->header_state = h_state;
goto error;
}
2019-04-23 00:31:21 +00:00
2019-04-24 01:37:25 +00:00
parser->content_length = t;
break;
}
case h_content_length_ws:
if (ch == ' ') break;
2019-04-23 00:31:21 +00:00
SET_ERRNO(HPE_INVALID_CONTENT_LENGTH);
2019-04-24 01:37:25 +00:00
parser->header_state = h_state;
2019-04-23 00:31:21 +00:00
goto error;
2019-04-24 01:37:25 +00:00
/* Transfer-Encoding: chunked */
case h_matching_transfer_encoding_chunked:
parser->index++;
if (parser->index > sizeof(CHUNKED)-1
|| c != CHUNKED[parser->index]) {
h_state = h_general;
} else if (parser->index == sizeof(CHUNKED)-2) {
h_state = h_transfer_encoding_chunked;
}
break;
2019-04-23 00:31:21 +00:00
2019-04-24 01:37:25 +00:00
case h_matching_connection_token_start:
/* looking for 'Connection: keep-alive' */
if (c == 'k') {
h_state = h_matching_connection_keep_alive;
/* looking for 'Connection: close' */
} else if (c == 'c') {
h_state = h_matching_connection_close;
} else if (c == 'u') {
h_state = h_matching_connection_upgrade;
} else if (STRICT_TOKEN(c)) {
h_state = h_matching_connection_token;
} else if (c == ' ' || c == '\t') {
/* Skip lws */
} else {
h_state = h_general;
}
break;
2019-04-23 00:31:21 +00:00
2019-04-24 01:37:25 +00:00
/* 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]) {
h_state = h_matching_connection_token;
} else if (parser->index == sizeof(KEEP_ALIVE)-2) {
h_state = h_connection_keep_alive;
}
break;
2019-04-23 00:31:21 +00:00
2019-04-24 01:37:25 +00:00
/* looking for 'Connection: close' */
case h_matching_connection_close:
parser->index++;
if (parser->index > sizeof(CLOSE)-1 || c != CLOSE[parser->index]) {
h_state = h_matching_connection_token;
} else if (parser->index == sizeof(CLOSE)-2) {
h_state = h_connection_close;
}
break;
2019-04-23 00:31:21 +00:00
2019-04-24 01:37:25 +00:00
/* looking for 'Connection: upgrade' */
case h_matching_connection_upgrade:
parser->index++;
if (parser->index > sizeof(UPGRADE) - 1 ||
c != UPGRADE[parser->index]) {
h_state = h_matching_connection_token;
} else if (parser->index == sizeof(UPGRADE)-2) {
h_state = h_connection_upgrade;
}
break;
2019-04-23 00:31:21 +00:00
2019-04-24 01:37:25 +00:00
case h_matching_connection_token:
if (ch == ',') {
h_state = h_matching_connection_token_start;
parser->index = 0;
}
break;
case h_transfer_encoding_chunked:
if (ch != ' ') h_state = h_general;
break;
case h_connection_keep_alive:
case h_connection_close:
case h_connection_upgrade:
if (ch == ',') {
if (h_state == h_connection_keep_alive) {
parser->flags |= F_CONNECTION_KEEP_ALIVE;
} else if (h_state == h_connection_close) {
parser->flags |= F_CONNECTION_CLOSE;
} else if (h_state == h_connection_upgrade) {
parser->flags |= F_CONNECTION_UPGRADE;
}
h_state = h_matching_connection_token_start;
parser->index = 0;
} else if (ch != ' ') {
h_state = h_matching_connection_token;
}
break;
default:
UPDATE_STATE(s_header_value);
h_state = h_general;
break;
}
2019-04-23 00:31:21 +00:00
}
2019-04-24 01:37:25 +00:00
parser->header_state = h_state;
if (p == data + len)
--p;
COUNT_HEADER_SIZE(p - start);
2019-04-23 00:31:21 +00:00
break;
}
case s_header_almost_done:
{
2019-04-24 01:37:25 +00:00
if (UNLIKELY(ch != LF)) {
SET_ERRNO(HPE_LF_EXPECTED);
goto error;
}
2019-04-23 00:31:21 +00:00
2019-04-24 01:37:25 +00:00
UPDATE_STATE(s_header_value_lws);
break;
}
case s_header_value_lws:
{
if (ch == ' ' || ch == '\t') {
if (parser->header_state == h_content_length_num) {
/* treat obsolete line folding as space */
parser->header_state = h_content_length_ws;
}
UPDATE_STATE(s_header_value_start);
REEXECUTE();
}
2019-04-23 00:31:21 +00:00
2019-04-24 01:37:25 +00:00
/* finished the header */
2019-04-23 00:31:21 +00:00
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;
2019-04-23 00:31:21 +00:00
case h_transfer_encoding_chunked:
parser->flags |= F_CHUNKED;
break;
2019-04-24 01:37:25 +00:00
case h_connection_upgrade:
parser->flags |= F_CONNECTION_UPGRADE;
break;
2019-04-23 00:31:21 +00:00
default:
break;
}
2019-04-24 01:37:25 +00:00
UPDATE_STATE(s_header_field_start);
REEXECUTE();
}
case s_header_value_discard_ws_almost_done:
{
STRICT_CHECK(ch != LF);
UPDATE_STATE(s_header_value_discard_lws);
2019-04-23 00:31:21 +00:00
break;
}
2019-04-24 01:37:25 +00:00
case s_header_value_discard_lws:
2019-04-23 00:31:21 +00:00
{
2019-04-24 01:37:25 +00:00
if (ch == ' ' || ch == '\t') {
UPDATE_STATE(s_header_value_discard_ws);
break;
} else {
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_connection_upgrade:
parser->flags |= F_CONNECTION_UPGRADE;
break;
case h_transfer_encoding_chunked:
parser->flags |= F_CHUNKED;
break;
case h_content_length:
/* do not allow empty content length */
SET_ERRNO(HPE_INVALID_CONTENT_LENGTH);
goto error;
break;
default:
break;
}
/* header value was empty */
MARK(header_value);
UPDATE_STATE(s_header_field_start);
CALLBACK_DATA_NOADVANCE(header_value);
REEXECUTE();
2019-04-23 00:31:21 +00:00
}
}
case s_headers_almost_done:
{
STRICT_CHECK(ch != LF);
if (parser->flags & F_TRAILING) {
/* End of a chunked request */
2019-04-24 01:37:25 +00:00
UPDATE_STATE(s_message_done);
CALLBACK_NOTIFY_NOADVANCE(chunk_complete);
REEXECUTE();
2019-04-23 00:31:21 +00:00
}
2019-04-24 01:37:25 +00:00
/* Cannot use chunked encoding and a content-length header together
per the HTTP specification. */
if ((parser->flags & F_CHUNKED) &&
(parser->flags & F_CONTENTLENGTH)) {
SET_ERRNO(HPE_UNEXPECTED_CONTENT_LENGTH);
goto error;
}
UPDATE_STATE(s_headers_done);
2019-04-23 00:31:21 +00:00
/* Set this here so that on_headers_complete() callbacks can see it */
2019-04-24 01:37:25 +00:00
if ((parser->flags & F_UPGRADE) &&
(parser->flags & F_CONNECTION_UPGRADE)) {
/* For responses, "Upgrade: foo" and "Connection: upgrade" are
* mandatory only when it is a 101 Switching Protocols response,
* otherwise it is purely informational, to announce support.
*/
parser->upgrade =
(parser->type == HTTP_REQUEST || parser->status_code == 101);
} else {
parser->upgrade = (parser->method == HTTP_CONNECT);
}
2019-04-23 00:31:21 +00:00
/* 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;
2019-04-24 01:37:25 +00:00
case 2:
parser->upgrade = 1;
/* fall through */
2019-04-23 00:31:21 +00:00
case 1:
parser->flags |= F_SKIPBODY;
break;
default:
SET_ERRNO(HPE_CB_headers_complete);
2019-04-24 01:37:25 +00:00
RETURN(p - data); /* Error */
2019-04-23 00:31:21 +00:00
}
}
if (HTTP_PARSER_ERRNO(parser) != HPE_OK) {
2019-04-24 01:37:25 +00:00
RETURN(p - data);
2019-04-23 00:31:21 +00:00
}
2019-04-24 01:37:25 +00:00
REEXECUTE();
2019-04-23 00:31:21 +00:00
}
case s_headers_done:
{
2019-04-24 01:37:25 +00:00
int hasBody;
2019-04-23 00:31:21 +00:00
STRICT_CHECK(ch != LF);
parser->nread = 0;
2019-04-24 01:37:25 +00:00
nread = 0;
hasBody = parser->flags & F_CHUNKED ||
(parser->content_length > 0 && parser->content_length != ULLONG_MAX);
if (parser->upgrade && (parser->method == HTTP_CONNECT ||
(parser->flags & F_SKIPBODY) || !hasBody)) {
/* Exit, the rest of the message is in a different protocol. */
UPDATE_STATE(NEW_MESSAGE());
2019-04-23 00:31:21 +00:00
CALLBACK_NOTIFY(message_complete);
2019-04-24 01:37:25 +00:00
RETURN((p - data) + 1);
2019-04-23 00:31:21 +00:00
}
if (parser->flags & F_SKIPBODY) {
2019-04-24 01:37:25 +00:00
UPDATE_STATE(NEW_MESSAGE());
2019-04-23 00:31:21 +00:00
CALLBACK_NOTIFY(message_complete);
} else if (parser->flags & F_CHUNKED) {
/* chunked encoding - ignore Content-Length header */
2019-04-24 01:37:25 +00:00
UPDATE_STATE(s_chunk_size_start);
2019-04-23 00:31:21 +00:00
} else {
if (parser->content_length == 0) {
/* Content-Length header given but zero: Content-Length: 0\r\n */
2019-04-24 01:37:25 +00:00
UPDATE_STATE(NEW_MESSAGE());
2019-04-23 00:31:21 +00:00
CALLBACK_NOTIFY(message_complete);
} else if (parser->content_length != ULLONG_MAX) {
/* Content-Length header given and non-zero */
2019-04-24 01:37:25 +00:00
UPDATE_STATE(s_body_identity);
2019-04-23 00:31:21 +00:00
} else {
2019-04-24 01:37:25 +00:00
if (!http_message_needs_eof(parser)) {
2019-04-23 00:31:21 +00:00
/* Assume content-length 0 - read the next */
2019-04-24 01:37:25 +00:00
UPDATE_STATE(NEW_MESSAGE());
2019-04-23 00:31:21 +00:00
CALLBACK_NOTIFY(message_complete);
} else {
/* Read body until EOF */
2019-04-24 01:37:25 +00:00
UPDATE_STATE(s_body_identity_eof);
2019-04-23 00:31:21 +00:00
}
}
}
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) {
2019-04-24 01:37:25 +00:00
UPDATE_STATE(s_message_done);
2019-04-23 00:31:21 +00:00
/* 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);
2019-04-24 01:37:25 +00:00
REEXECUTE();
2019-04-23 00:31:21 +00:00
}
break;
}
/* read until EOF */
case s_body_identity_eof:
MARK(body);
p = data + len - 1;
break;
case s_message_done:
2019-04-24 01:37:25 +00:00
UPDATE_STATE(NEW_MESSAGE());
2019-04-23 00:31:21 +00:00
CALLBACK_NOTIFY(message_complete);
2019-04-24 01:37:25 +00:00
if (parser->upgrade) {
/* Exit, the rest of the message is in a different protocol. */
RETURN((p - data) + 1);
}
2019-04-23 00:31:21 +00:00
break;
case s_chunk_size_start:
{
2019-04-24 01:37:25 +00:00
assert(nread == 1);
2019-04-23 00:31:21 +00:00
assert(parser->flags & F_CHUNKED);
unhex_val = unhex[(unsigned char)ch];
2019-04-24 01:37:25 +00:00
if (UNLIKELY(unhex_val == -1)) {
2019-04-23 00:31:21 +00:00
SET_ERRNO(HPE_INVALID_CHUNK_SIZE);
goto error;
}
parser->content_length = unhex_val;
2019-04-24 01:37:25 +00:00
UPDATE_STATE(s_chunk_size);
2019-04-23 00:31:21 +00:00
break;
}
case s_chunk_size:
{
uint64_t t;
assert(parser->flags & F_CHUNKED);
if (ch == CR) {
2019-04-24 01:37:25 +00:00
UPDATE_STATE(s_chunk_size_almost_done);
2019-04-23 00:31:21 +00:00
break;
}
unhex_val = unhex[(unsigned char)ch];
if (unhex_val == -1) {
if (ch == ';' || ch == ' ') {
2019-04-24 01:37:25 +00:00
UPDATE_STATE(s_chunk_parameters);
2019-04-23 00:31:21 +00:00
break;
}
SET_ERRNO(HPE_INVALID_CHUNK_SIZE);
goto error;
}
t = parser->content_length;
t *= 16;
t += unhex_val;
2019-04-24 01:37:25 +00:00
/* Overflow? Test against a conservative limit for simplicity. */
if (UNLIKELY((ULLONG_MAX - 16) / 16 < parser->content_length)) {
2019-04-23 00:31:21 +00:00
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) {
2019-04-24 01:37:25 +00:00
UPDATE_STATE(s_chunk_size_almost_done);
2019-04-23 00:31:21 +00:00
break;
}
break;
}
case s_chunk_size_almost_done:
{
assert(parser->flags & F_CHUNKED);
STRICT_CHECK(ch != LF);
parser->nread = 0;
2019-04-24 01:37:25 +00:00
nread = 0;
2019-04-23 00:31:21 +00:00
if (parser->content_length == 0) {
parser->flags |= F_TRAILING;
2019-04-24 01:37:25 +00:00
UPDATE_STATE(s_header_field_start);
2019-04-23 00:31:21 +00:00
} else {
2019-04-24 01:37:25 +00:00
UPDATE_STATE(s_chunk_data);
2019-04-23 00:31:21 +00:00
}
2019-04-24 01:37:25 +00:00
CALLBACK_NOTIFY(chunk_header);
2019-04-23 00:31:21 +00:00
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) {
2019-04-24 01:37:25 +00:00
UPDATE_STATE(s_chunk_data_almost_done);
2019-04-23 00:31:21 +00:00
}
break;
}
case s_chunk_data_almost_done:
assert(parser->flags & F_CHUNKED);
assert(parser->content_length == 0);
STRICT_CHECK(ch != CR);
2019-04-24 01:37:25 +00:00
UPDATE_STATE(s_chunk_data_done);
2019-04-23 00:31:21 +00:00
CALLBACK_DATA(body);
break;
case s_chunk_data_done:
assert(parser->flags & F_CHUNKED);
STRICT_CHECK(ch != LF);
parser->nread = 0;
2019-04-24 01:37:25 +00:00
nread = 0;
UPDATE_STATE(s_chunk_size_start);
CALLBACK_NOTIFY(chunk_complete);
2019-04-23 00:31:21 +00:00
break;
default:
assert(0 && "unhandled state");
SET_ERRNO(HPE_INVALID_INTERNAL_STATE);
goto error;
}
}
2019-04-24 01:37:25 +00:00
/* Run callbacks for any marks that we have leftover after we ran out of
2019-04-23 00:31:21 +00:00
* 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) +
2019-04-24 01:37:25 +00:00
(body_mark ? 1 : 0) +
(status_mark ? 1 : 0)) <= 1);
2019-04-23 00:31:21 +00:00
CALLBACK_DATA_NOADVANCE(header_field);
CALLBACK_DATA_NOADVANCE(header_value);
CALLBACK_DATA_NOADVANCE(url);
CALLBACK_DATA_NOADVANCE(body);
2019-04-24 01:37:25 +00:00
CALLBACK_DATA_NOADVANCE(status);
2019-04-23 00:31:21 +00:00
2019-04-24 01:37:25 +00:00
RETURN(len);
2019-04-23 00:31:21 +00:00
error:
if (HTTP_PARSER_ERRNO(parser) == HPE_OK) {
SET_ERRNO(HPE_UNKNOWN);
}
2019-04-24 01:37:25 +00:00
RETURN(p - data);
2019-04-23 00:31:21 +00:00
}
/* Does the parser need to see an EOF to find the end of the message? */
int
http_message_needs_eof (const http_parser *parser)
{
if (parser->type == HTTP_REQUEST) {
return 0;
}
/* See RFC 2616 section 4.4 */
if (parser->status_code / 100 == 1 || /* 1xx e.g. Continue */
parser->status_code == 204 || /* No Content */
parser->status_code == 304 || /* Not Modified */
parser->flags & F_SKIPBODY) { /* response to a HEAD request */
return 0;
}
if ((parser->flags & F_CHUNKED) || parser->content_length != ULLONG_MAX) {
return 0;
}
return 1;
}
int
http_should_keep_alive (const http_parser *parser)
{
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;
}
2019-04-23 00:31:21 +00:00
}
return !http_message_needs_eof(parser);
}
const char *
http_method_str (enum http_method m)
{
return ELEM_AT(method_strings, m, "<unknown>");
}
2019-04-24 01:37:25 +00:00
const char *
http_status_str (enum http_status s)
{
switch (s) {
#define XX(num, name, string) case HTTP_STATUS_##name: return #string;
HTTP_STATUS_MAP(XX)
#undef XX
default: return "<unknown>";
}
}
2019-04-23 00:31:21 +00:00
void
http_parser_init (http_parser *parser, enum 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;
}
2019-04-24 01:37:25 +00:00
void
http_parser_settings_init(http_parser_settings *settings)
{
memset(settings, 0, sizeof(*settings));
}
2019-04-23 00:31:21 +00:00
const char *
http_errno_name(enum http_errno err) {
2019-04-24 01:37:25 +00:00
assert(((size_t) err) < ARRAY_SIZE(http_strerror_tab));
2019-04-23 00:31:21 +00:00
return http_strerror_tab[err].name;
}
const char *
http_errno_description(enum http_errno err) {
2019-04-24 01:37:25 +00:00
assert(((size_t) err) < ARRAY_SIZE(http_strerror_tab));
2019-04-23 00:31:21 +00:00
return http_strerror_tab[err].description;
}
static enum http_host_state
http_parse_host_char(enum http_host_state s, const char ch) {
switch(s) {
case s_http_userinfo:
case s_http_userinfo_start:
if (ch == '@') {
return s_http_host_start;
}
if (IS_USERINFO_CHAR(ch)) {
return s_http_userinfo;
}
break;
case s_http_host_start:
if (ch == '[') {
return s_http_host_v6_start;
}
if (IS_HOST_CHAR(ch)) {
return s_http_host;
}
break;
case s_http_host:
if (IS_HOST_CHAR(ch)) {
return s_http_host;
}
2019-04-24 01:37:25 +00:00
/* fall through */
2019-04-23 00:31:21 +00:00
case s_http_host_v6_end:
if (ch == ':') {
return s_http_host_port_start;
}
break;
case s_http_host_v6:
if (ch == ']') {
return s_http_host_v6_end;
}
2019-04-24 01:37:25 +00:00
/* fall through */
2019-04-23 00:31:21 +00:00
case s_http_host_v6_start:
if (IS_HEX(ch) || ch == ':' || ch == '.') {
return s_http_host_v6;
}
2019-04-24 01:37:25 +00:00
if (s == s_http_host_v6 && ch == '%') {
return s_http_host_v6_zone_start;
}
break;
case s_http_host_v6_zone:
if (ch == ']') {
return s_http_host_v6_end;
}
/* fall through */
case s_http_host_v6_zone_start:
/* RFC 6874 Zone ID consists of 1*( unreserved / pct-encoded) */
if (IS_ALPHANUM(ch) || ch == '%' || ch == '.' || ch == '-' || ch == '_' ||
ch == '~') {
return s_http_host_v6_zone;
}
2019-04-23 00:31:21 +00:00
break;
case s_http_host_port:
case s_http_host_port_start:
if (IS_NUM(ch)) {
return s_http_host_port;
}
break;
default:
break;
}
return s_http_host_dead;
}
static int
http_parse_host(const char * buf, struct http_parser_url *u, int found_at) {
2019-04-23 00:31:21 +00:00
enum http_host_state s;
const char *p;
size_t buflen = u->field_data[UF_HOST].off + u->field_data[UF_HOST].len;
2019-04-24 01:37:25 +00:00
assert(u->field_set & (1 << UF_HOST));
2019-04-23 00:31:21 +00:00
u->field_data[UF_HOST].len = 0;
s = found_at ? s_http_userinfo_start : s_http_host_start;
for (p = buf + u->field_data[UF_HOST].off; p < buf + buflen; p++) {
enum http_host_state new_s = http_parse_host_char(s, *p);
if (new_s == s_http_host_dead) {
return 1;
}
switch(new_s) {
case s_http_host:
if (s != s_http_host) {
2019-04-24 01:37:25 +00:00
u->field_data[UF_HOST].off = (uint16_t)(p - buf);
}
2019-04-23 00:31:21 +00:00
u->field_data[UF_HOST].len++;
break;
case s_http_host_v6:
if (s != s_http_host_v6) {
2019-04-24 01:37:25 +00:00
u->field_data[UF_HOST].off = (uint16_t)(p - buf);
}
2019-04-23 00:31:21 +00:00
u->field_data[UF_HOST].len++;
break;
2019-04-24 01:37:25 +00:00
case s_http_host_v6_zone_start:
case s_http_host_v6_zone:
u->field_data[UF_HOST].len++;
break;
2019-04-23 00:31:21 +00:00
case s_http_host_port:
if (s != s_http_host_port) {
2019-04-24 01:37:25 +00:00
u->field_data[UF_PORT].off = (uint16_t)(p - buf);
2019-04-23 00:31:21 +00:00
u->field_data[UF_PORT].len = 0;
u->field_set |= (1 << UF_PORT);
}
u->field_data[UF_PORT].len++;
break;
case s_http_userinfo:
if (s != s_http_userinfo) {
2019-04-24 01:37:25 +00:00
u->field_data[UF_USERINFO].off = (uint16_t)(p - buf);
2019-04-23 00:31:21 +00:00
u->field_data[UF_USERINFO].len = 0;
u->field_set |= (1 << UF_USERINFO);
}
u->field_data[UF_USERINFO].len++;
break;
default:
break;
}
2019-04-23 00:31:21 +00:00
s = new_s;
}
/* Make sure we don't end somewhere unexpected */
switch (s) {
case s_http_host_start:
case s_http_host_v6_start:
case s_http_host_v6:
2019-04-24 01:37:25 +00:00
case s_http_host_v6_zone_start:
case s_http_host_v6_zone:
2019-04-23 00:31:21 +00:00
case s_http_host_port_start:
case s_http_userinfo:
case s_http_userinfo_start:
return 1;
default:
break;
}
return 0;
}
2019-04-24 01:37:25 +00:00
void
http_parser_url_init(struct http_parser_url *u) {
memset(u, 0, sizeof(*u));
}
int
http_parser_parse_url(const char *buf, size_t buflen, int is_connect,
struct http_parser_url *u)
{
2019-04-23 00:31:21 +00:00
enum state s;
const char *p;
enum http_parser_url_fields uf, old_uf;
int found_at = 0;
2019-04-24 01:37:25 +00:00
if (buflen == 0) {
return 1;
}
2019-04-23 00:31:21 +00:00
u->port = u->field_set = 0;
s = is_connect ? s_req_server_start : s_req_spaces_before_url;
2019-04-24 01:37:25 +00:00
old_uf = UF_MAX;
2019-04-23 00:31:21 +00:00
for (p = buf; p < buf + buflen; p++) {
s = parse_url_char(s, *p);
/* Figure out the next field that we're operating on */
switch (s) {
case s_dead:
return 1;
/* Skip delimeters */
case s_req_schema_slash:
case s_req_schema_slash_slash:
case s_req_server_start:
case s_req_query_string_start:
case s_req_fragment_start:
continue;
case s_req_schema:
uf = UF_SCHEMA;
break;
case s_req_server_with_at:
found_at = 1;
2019-04-24 01:37:25 +00:00
/* fall through */
2019-04-23 00:31:21 +00:00
case s_req_server:
uf = UF_HOST;
break;
case s_req_path:
uf = UF_PATH;
break;
case s_req_query_string:
uf = UF_QUERY;
break;
case s_req_fragment:
uf = UF_FRAGMENT;
break;
default:
assert(!"Unexpected state");
return 1;
}
2019-04-23 00:31:21 +00:00
/* Nothing's changed; soldier on */
if (uf == old_uf) {
u->field_data[uf].len++;
continue;
}
2019-04-23 00:31:21 +00:00
2019-04-24 01:37:25 +00:00
u->field_data[uf].off = (uint16_t)(p - buf);
2019-04-23 00:31:21 +00:00
u->field_data[uf].len = 1;
u->field_set |= (1 << uf);
old_uf = uf;
}
/* host must be present if there is a schema */
/* parsing http:///toto will fail */
2019-04-24 01:37:25 +00:00
if ((u->field_set & (1 << UF_SCHEMA)) &&
(u->field_set & (1 << UF_HOST)) == 0) {
return 1;
}
if (u->field_set & (1 << UF_HOST)) {
2019-04-23 00:31:21 +00:00
if (http_parse_host(buf, u, found_at) != 0) {
return 1;
}
2019-04-23 00:31:21 +00:00
}
/* CONNECT requests can only contain "hostname:port" */
if (is_connect && u->field_set != ((1 << UF_HOST)|(1 << UF_PORT))) {
return 1;
}
if (u->field_set & (1 << UF_PORT)) {
2019-04-24 01:37:25 +00:00
uint16_t off;
uint16_t len;
const char* p;
const char* end;
unsigned long v;
off = u->field_data[UF_PORT].off;
len = u->field_data[UF_PORT].len;
end = buf + off + len;
/* NOTE: The characters are already validated and are in the [0-9] range */
assert(off + len <= buflen && "Port number overflow");
v = 0;
for (p = buf + off; p < end; p++) {
v *= 10;
v += *p - '0';
/* Ports have a max value of 2^16 */
if (v > 0xffff) {
return 1;
}
}
2019-04-23 00:31:21 +00:00
u->port = (uint16_t) v;
}
return 0;
}
void
http_parser_pause(http_parser *parser, int paused) {
2019-04-23 00:31:21 +00:00
/* 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) {
2019-04-24 01:37:25 +00:00
uint32_t nread = parser->nread; /* used by the SET_ERRNO macro */
2019-04-23 00:31:21 +00:00
SET_ERRNO((paused) ? HPE_PAUSED : HPE_OK);
} else {
assert(0 && "Attempting to pause parser in error state");
}
}
int
http_body_is_final(const struct http_parser *parser) {
return parser->state == s_message_done;
}
2019-04-24 01:37:25 +00:00
unsigned long
http_parser_version(void) {
return HTTP_PARSER_VERSION_MAJOR * 0x10000 |
HTTP_PARSER_VERSION_MINOR * 0x00100 |
HTTP_PARSER_VERSION_PATCH * 0x00001;
}
void
http_parser_set_max_header_size(uint32_t size) {
max_header_size = size;
}