Skip to content

Commit

Permalink
rpcdaemon: add picohttpparser module (#1322)
Browse files Browse the repository at this point in the history
  • Loading branch information
@lupin012
lupin012 authored Jul 16, 2023
1 parent 3cbfd0e commit 7512054
Show file tree
Hide file tree
Showing 7 changed files with 74 additions and 343 deletions.
3 changes: 3 additions & 0 deletions .gitmodules
View file
Original file line number Diff line number Diff line change
Expand Up @@ -45,3 +45,6 @@
[submodule "third_party/stun-msg"]
path = third_party/stun-msg
url = https://github.com/battlmonstr/stun-msg.git
[submodule "third_party/picohttpparser"]
path = third_party/picohttpparser
url = https://github.com/h2o/picohttpparser
6 changes: 6 additions & 0 deletions silkworm/silkrpc/CMakeLists.txt
View file
Original file line number Diff line number Diff line change
Expand Up @@ -35,6 +35,10 @@ file(
)
list(FILTER SILKRPC_SRC EXCLUDE REGEX "main\\.cpp$|_test\\.cpp$|\\.pb\\.cc|\\.pb\\.h")

# pico_http_parser
add_library(pico_http_parser "${SILKWORM_MAIN_DIR}/third_party/picohttpparser/picohttpparser.c")
target_compile_options(pico_http_parser PRIVATE -Wno-error)

set(SILKRPC_PUBLIC_LIBRARIES
silkworm_node
silkworm_core
Expand All @@ -50,6 +54,7 @@ set(SILKRPC_PUBLIC_LIBRARIES
Boost::thread
protobuf::libprotobuf
intx::intx
pico_http_parser
)

set(SILKRPC_PRIVATE_LIBRARIES evmc::instructions roaring::roaring)
Expand All @@ -66,6 +71,7 @@ endif()
target_include_directories(
silkrpc
PUBLIC "${SILKWORM_MAIN_DIR}"
"${SILKWORM_MAIN_DIR}/third_party/picohttpparser"
"${SILKWORM_MAIN_DIR}/third_party/cbor-cpp"
"${SILKWORM_MAIN_DIR}/third_party/glaze/include"
"${SILKWORM_MAIN_DIR}/third_party/evmone/lib"
Expand Down
1 change: 0 additions & 1 deletion silkworm/silkrpc/http/connection.cpp
View file
Original file line number Diff line number Diff line change
Expand Up @@ -105,7 +105,6 @@ boost::asio::awaitable<void> Connection::do_write() {

void Connection::clean() {
request_.reset();
request_parser_.reset();
reply_.reset();
}

Expand Down
313 changes: 61 additions & 252 deletions silkworm/silkrpc/http/request_parser.cpp
View file
Original file line number Diff line number Diff line change
Expand Up @@ -22,272 +22,81 @@

#include "request_parser.hpp"

#include <picohttpparser.h>

#include <algorithm>
#include <cstdlib>
#include <cstring>

namespace silkworm::rpc::http {

RequestParser::RequestParser() : state_(method_start) {
}
RequestParser::ResultType RequestParser::parse(Request& req, const char* begin, const char* end) {
const char* method_name; // uninitialised here because phr_parse_request initialises it
size_t method_len; // uninitialised here because phr_parse_request initialises it
const char* path; // uninitialised here because phr_parse_request initialises it
size_t path_len; // uninitialised here because phr_parse_request initialises it
int minor_version; // uninitialised here because phr_parse_request initialises it
struct phr_header headers[100];
size_t num_headers = sizeof(headers) / sizeof(headers[0]);
size_t last_len = 0;

void RequestParser::reset() {
state_ = method_start;
}
size_t slen = static_cast<size_t>(end - begin);

RequestParser::ResultType RequestParser::consume(Request& req, char input) {
switch (state_) {
case method_start:
if (!is_char(input) || is_ctl(input) || is_tspecial(input)) {
return ResultType::bad;
} else {
state_ = method;
req.method.push_back(input);
return ResultType::indeterminate;
}
case method:
if (input == ' ') {
state_ = uri;
return ResultType::indeterminate;
} else if (!is_char(input) || is_ctl(input) || is_tspecial(input)) {
return ResultType::bad;
} else {
req.method.push_back(input);
return ResultType::indeterminate;
}
case uri:
if (input == ' ') {
state_ = http_version_h;
return ResultType::indeterminate;
} else if (is_ctl(input)) {
return ResultType::bad;
} else {
req.uri.push_back(input);
return ResultType::indeterminate;
}
case http_version_h:
if (input == 'H') {
state_ = http_version_t_1;
return ResultType::indeterminate;
} else {
return ResultType::bad;
}
case http_version_t_1:
if (input == 'T') {
state_ = http_version_t_2;
return ResultType::indeterminate;
} else {
return ResultType::bad;
}
case http_version_t_2:
if (input == 'T') {
state_ = http_version_p;
return ResultType::indeterminate;
} else {
return ResultType::bad;
}
case http_version_p:
if (input == 'P') {
state_ = http_version_slash;
return ResultType::indeterminate;
} else {
return ResultType::bad;
}
case http_version_slash:
if (input == '/') {
req.http_version_major = 0;
req.http_version_minor = 0;
state_ = http_version_major_start;
return ResultType::indeterminate;
} else {
return ResultType::bad;
}
case http_version_major_start:
if (is_digit(input)) {
req.http_version_major = req.http_version_major * 10 + input - '0';
state_ = http_version_major;
return ResultType::indeterminate;
} else {
return ResultType::bad;
}
case http_version_major:
if (input == '.') {
state_ = http_version_minor_start;
return ResultType::indeterminate;
} else if (is_digit(input)) {
req.http_version_major = req.http_version_major * 10 + input - '0';
return ResultType::indeterminate;
} else {
return ResultType::bad;
}
case http_version_minor_start:
if (is_digit(input)) {
req.http_version_minor = req.http_version_minor * 10 + input - '0';
state_ = http_version_minor;
return ResultType::indeterminate;
} else {
return ResultType::bad;
}
case http_version_minor:
if (input == '\r') {
state_ = expecting_newline_1;
return ResultType::indeterminate;
} else if (is_digit(input)) {
req.http_version_minor = req.http_version_minor * 10 + input - '0';
return ResultType::indeterminate;
} else {
return ResultType::bad;
}
case expecting_newline_1:
if (input == '\n') {
state_ = header_line_start;
return ResultType::indeterminate;
} else {
return ResultType::bad;
}
case header_line_start:
if (input == '\r') {
state_ = expecting_newline_3;
return ResultType::indeterminate;
} else if (!req.headers.empty() && (input == ' ' || input == '\t')) {
state_ = header_lws;
return ResultType::indeterminate;
} else if (!is_char(input) || is_ctl(input) || is_tspecial(input)) {
return ResultType::bad;
} else {
req.headers.emplace_back();
req.headers.back().name.push_back(input);
state_ = header_name;
return ResultType::indeterminate;
}
case header_lws:
if (input == '\r') {
state_ = expecting_newline_2;
return ResultType::indeterminate;
} else if (input == ' ' || input == '\t') {
return ResultType::indeterminate;
} else if (is_ctl(input)) {
return ResultType::bad;
} else {
state_ = header_value;
req.headers.back().value.push_back(input);
return ResultType::indeterminate;
}
case header_name:
if (input == ':') {
state_ = space_before_header_value;
return ResultType::indeterminate;
} else if (!is_char(input) || is_ctl(input) || is_tspecial(input)) {
return ResultType::bad;
} else {
req.headers.back().name.push_back(input);
return ResultType::indeterminate;
}
case space_before_header_value:
if (input == ' ') {
state_ = header_value;
return ResultType::indeterminate;
} else {
return ResultType::bad;
}
case header_value:
if (input == '\r') {
state_ = expecting_newline_2;
return ResultType::indeterminate;
} else if (is_ctl(input)) {
return ResultType::bad;
} else {
req.headers.back().value.push_back(input);
return ResultType::indeterminate;
}
case expecting_newline_2:
if (input == '\n') {
state_ = header_line_start;
return ResultType::indeterminate;
} else {
return ResultType::bad;
}
case expecting_newline_3:
if (input == '\n') {
state_ = content_start;
// Look for Content-Length header to get content size
if (req.content_length == 0) {
const auto it = std::find_if(req.headers.begin(), req.headers.end(), [&](const Header& h) {
return h.name == "Content-Length";
});
if (it == req.headers.end()) {
return ResultType::bad;
}
const char* str{it->value.c_str()};
char* end_ptr{nullptr};
errno = 0;
long long len{std::strtoll(str, &end_ptr, 0)};
if (errno == ERANGE || end_ptr == str || len < 0 || len > UINT32_MAX) {
return ResultType::bad;
}
req.content_length = static_cast<uint32_t>(len);
}
if (req.content_length == 0) {
return ResultType::good;
}
// Look for Expect header to handle continuation request
const auto it = std::find_if(req.headers.begin(), req.headers.end(), [&](const Header& h) {
return h == kExpectRequestHeader;
});
if (it != req.headers.end()) {
return ResultType::processing_continue;
}
return ResultType::indeterminate;
} else {
return ResultType::bad;
if (req.content_length != 0 && req.content.length() < req.content_length) {
for (size_t ii = 0; ii < slen; ii++) {
req.content.push_back(begin[ii]);
}
if (req.content.length() < req.content_length)
return ResultType::indeterminate;
else
return ResultType::good;
}

const auto res = phr_parse_request(begin, slen, &method_name, &method_len, &path, &path_len, &minor_version, headers, &num_headers, last_len);
if (res < 0) {
return ResultType::bad;
}

bool expect_request = false;
bool content_length_present = false;

for (size_t ii = 0; ii < num_headers; ii++) {
if (std::memcmp(headers[ii].name, "Content-Length", headers[ii].name_len) == 0) {
req.content_length = static_cast<uint32_t>(atoi(headers[ii].value));
content_length_present = true;
}

else if (std::memcmp(headers[ii].name, "Expect", headers[ii].name_len) == 0) {
expect_request = true;
}

else if (std::memcmp(headers[ii].name, "Authorization", headers[ii].name_len) == 0) {
req.headers.emplace_back();
for (size_t index = 0; index < static_cast<size_t>(headers[ii].name_len); index++) {
req.headers.back().name.push_back(headers[ii].name[index]);
}
case content_start:
req.content.push_back(input);
if (req.content.length() < req.content_length) {
return ResultType::indeterminate;
} else {
return ResultType::good;
for (size_t index = 0; index < static_cast<size_t>(headers[ii].value_len); index++) {
req.headers.back().value.push_back(headers[ii].value[index]);
}
}
}
return ResultType::bad;
}

inline bool RequestParser::is_char(int c) {
return c >= 0 && c <= 127;
}

inline bool RequestParser::is_ctl(int c) {
return (c >= 0 && c <= 31) || (c == 127);
}
if (content_length_present == false) {
return ResultType::bad;
}

inline bool RequestParser::is_tspecial(int c) {
switch (c) {
case '(':
case ')':
case '<':
case '>':
case '@':
case ',':
case ';':
case ':':
case '\\':
case '"':
case '/':
case '[':
case ']':
case '?':
case '=':
case '{':
case '}':
case ' ':
case '\t':
return true;
default:
return false;
if (req.content_length == 0) {
return ResultType::good;
}
}

inline bool RequestParser::is_digit(int c) {
return c >= '0' && c <= '9';
req.content.resize(slen - static_cast<size_t>(res));
std::memcpy(req.content.data(), &begin[res], slen - static_cast<size_t>(res));
if (expect_request == true)
return ResultType::processing_continue;
else if (req.content.length() < req.content_length)
return ResultType::indeterminate;
else
return ResultType::good;
}

} // namespace silkworm::rpc::http
Loading

0 comments on commit 7512054

Please sign in to comment.