我基于 boost coroutine echo server example 制作了我的服务器，只是接收和写回一些数据。它在向客户端写入数据时崩溃，更奇怪的是，它只在使用多核时崩溃。
这是服务器，它读取 4 个字节并在 1 秒内写回“OK”作为超时：
#include <winsock2.h>
#include <windows.h>
#include <iostream>
using namespace std;
#include <boost/thread/thread.hpp>
#include <boost/asio.hpp>
#include <boost/asio/spawn.hpp>
using namespace boost;
using namespace boost::asio;
using namespace boost::asio::ip;
#define SERVER_PORT 1234
#define DATA_LEN_4 4
#define TIMEOUT_LIMIT 1 // second
struct session : public std::enable_shared_from_this<session>
{
tcp::socket socket_;
boost::asio::steady_timer timer_;
boost::asio::strand<boost::asio::io_context::executor_type> strand_;
explicit session(boost::asio::io_context& io_context, tcp::socket socket)
: socket_(std::move(socket)),
timer_(io_context),
strand_(io_context.get_executor())
{ }
void go()
{
auto self(shared_from_this());
boost::asio::spawn(strand_, [this, self](boost::asio::yield_context yield)
{
try
{
timer_.expires_from_now(std::chrono::seconds(TIMEOUT_LIMIT));
// recv data
string packet;
packet.resize(DATA_LEN_4); // alloc memory
size_t received_len = 0;
// read data
{
size_t rs;
while(received_len < DATA_LEN_4) { // recv 4 bytes
boost::system::error_code ec;
rs = socket_.async_read_some(
boost::asio::buffer((char*)(packet.c_str()+received_len), DATA_LEN_4-received_len), yield[ec]);
if(ec==boost::asio::error::eof)
break; //connection closed cleanly by peer
else if(ec) {
throw "read_fail";
}
received_len += rs;
}
}
if(received_len < DATA_LEN_4) {
throw "recv too short, maybe timeout";
}
// write back "OK"
{
boost::system::error_code ecw;
boost::asio::async_write(socket_, boost::asio::buffer(string("OK")), yield[ecw]);
if(ecw==boost::asio::error::eof)
return; //connection closed cleanly by peer
else if(ecw)
throw "write_fail"; // some other error
}
}
catch (const char* reason)
{
printf("exception reason: %s\n", reason);
boost::system::error_code ecw;
/*
* Question 1: why this 'async_write' line causes crash?
*/
// write the error reason to client
boost::asio::async_write(socket_, boost::asio::buffer(string(reason)), yield[ecw]);
socket_.close();
timer_.cancel();
}
catch (...)
{
printf("unknown exception\n");
socket_.close();
timer_.cancel();
}
});
boost::asio::spawn(strand_, [this, self](boost::asio::yield_context yield)
{
while (socket_.is_open())
{
boost::system::error_code ignored_ec;
timer_.async_wait(yield[ignored_ec]);
if (timer_.expires_from_now() <= std::chrono::seconds(0))
socket_.close();
}
});
}
};
int main() {
boost::asio::io_context io_context;
boost::asio::spawn(io_context, [&](boost::asio::yield_context yield)
{
tcp::acceptor acceptor(io_context,
tcp::endpoint(tcp::v4(), SERVER_PORT));
for (;;)
{
boost::system::error_code ec;
tcp::socket socket(io_context);
acceptor.async_accept(socket, yield[ec]);
if (!ec)
std::make_shared<session>(io_context, std::move(socket))->go();
}
});
/*
* When run on 1 CPU, it runs fine, no Crash
*/
// io_context.run();
/*
* Question 2:
* But when run on multiple CPUs, it Crashes !!!
* Why?
*/
auto thread_count = std::thread::hardware_concurrency();
boost::thread_group tgroup;
for (auto i = 0; i < thread_count; ++i)
tgroup.create_thread(boost::bind(&boost::asio::io_context::run, &io_context));
tgroup.join_all();
}
请注意，4 字节数据包和1 秒超时只是为了说明问题，真实服务器使用大数据包可能会在网络状况不佳时导致超时。为了模拟这一点，客户端每秒写入 1 个字节以触发服务器上的读取超时。
客户端：
#include <iostream>
#include <boost/asio.hpp>
using namespace std;
using boost::asio::ip::tcp;
#define SERVER "127.0.0.1"
#define PORT "1234"
int main() {
boost::asio::io_context io_context;
unsigned i = 1;
while(1) {
try {
tcp::socket s(io_context);
tcp::resolver resolver(io_context);
boost::asio::connect(s, resolver.resolve(SERVER, PORT));
// to simulate the bad network condition,
// write 4 bytes in 4 seconds to trigger the receive timeout on server, which is 1 second
for(int i=0; i<4; i++) {
boost::asio::write(s, boost::asio::buffer(string("A")));
std::this_thread::sleep_for(std::chrono::seconds(1)); // sleep 1 second
}
// read echo
char x[64] = {0};
s.read_some(boost::asio::buffer(x, sizeof(x)));
cout << i++ << ". received: " << x << endl;
} catch (...) {
cout << i++ << " exception" << endl;
}
}
return 0;
}
问题一：
为什么这条线会导致崩溃？
boost::asio::async_write(socket_, boost::asio::buffer(string(reason)), yield[ecw]);
问题 2：
为什么服务器在 1 cpu 上运行时不会崩溃：io_context.run();？
并在多个 CPU 上使用thread_group?
我的环境：Win10-64bit、boost-1.71.0-64bit、VisualStudio-2017-Community