- Boost-users - lists.preview.boost.org

boost::asio blocking socket read with timeout with multiple threads
by tom＠quarendon.net 24 Mar '18

24 Mar '18

I've got a boost::asio based server, and generally it's working well. The server runs multiple threads, each simply running the io_service::run method on the same io_service object. It accepts multiple connections and processes multiple concurrent sockets. It uses a strand per connection to ensure that the data for each connection is processed strictly sequentially. For various reasons it uses a mixture of async and sync reads. The trouble I'm having is implementing a timeout on the synchronous (blocking) reads. I want to protect the server in the case of a client sending an erroneous or malicious packet for example, and make sure a read doesn't block forever. There was a post back in 2007 that gave a technique (https://lists.boost.org/Archives/boost/2007/04/120339.php) There's a stack overflow question that references that (https://stackoverflow.com/questions/13126776/asioread-with-timeout) There are some examples in the docs that perform a similar kind of technique. In the examples, the server is all asynchronous. The client is synchronous, but only uses one thread. The example from 2007 also only uses one thread. The examples all revolve around the technique of starting an *async* read, then performing a nested loop of io_service::run_one. However, so far I've been unable to find a form of code that works reliably in a multithreaded environment. The basis for my experimentation is here: https://gist.github.com/anonymous/1160c11f8ed9c29b9184325191a3a63b It starts a server thread, then starts a client that makes a connection and then writes nothing, to simulate a "bad" client and to provoke a "timeout" condition. What I'm finding is this: With one server thread I can make it work OK. The nested loop of "run_one()" operates correctly. With multiple server threads though, I can't make it work reliably. The basis of the loop is: bool timedOut = false; bool readComplete = false; timer.async_wait(strand_.wrap(boost::bind(&handleReadTimeout, &timedOut, &socket_, _1))); async_read(socket_, buffer(b), strand_.wrap(boost::bind(handleReadComplete, &ec, &bytesRead, &readComplete, &timer, _1, _2))); while (socket_.get_io_service().run_one()) { if (timedOut && readComplete) { break; } } handleReadTimeout sets the timedOut flag to true and calls "cancel" on the socket. The handleReadComplete callback sets readComplete to true and calls cancel on the timer. With multiple threads, the handleReadTimeout/handleReadComplete callbacks are run on other threads. So the while loop here just blocks, as there is never anything to run. That's my surmise of what's going on anyway. I've experimented with strands to try and force it all onto the same thread, but so far failed (if the above code is called in the context of the same strand, it just seems to block the handleReadTimeout and handleReadComplete callbacks from being called). Note that if I expend my example to include other clients talking to a simple "echo" service to simulate other traffic, then the while loop above *does* wake, as that other traffic gives it something to do. But that's fragile. The alternative formulation, the one I started with actually, is to do an async wait on the timer, and a normal synchronous read on the socket. The timer callback performs a cancel (or close -- I tried both) on the socket hoping to cause the socket read to error. This is the kind of thing you'd do if you were programming raw sockets. That works fine on Windows, but won't work on Linux. The documentation for cancel does say it cancels any outstanding *async* calls, but I'm surprised calling close doesn't work and cause the read to wake. So I'm stuck. I was hoping someone here could point me in the right direction, as it seems an obvious thing (to me) to want try and do. Is there a form of code that allows me to perform a synchronous read from a socket and time it out, which will work in a multi threaded/single io service environment? This is all with boost 1.64 by the way. Thanks for any help in advance.

6 27

Re: [Boost-users] boost::asio blocking socket read with timeout with multiple threads
by tom＠quarendon.net 23 Mar '18

23 Mar '18

> You're right. That poll_read loop consumes the alarm :/ > > So then i think the fallback has to be a raw Berkley sockets read against > the native_handle(), with a signal to cause the EINTR. Yes, and I've tried basically that method earlier. It works fine, as long as you don't want to use the SSL wrapper, which is what I ultimately want to do. Of course there, you don't have the control to do that. Well, I could write my own socket class variant and wrap that in the SSL wrapper, but that's daft. You can't "test" the socket by using a poll with a timeout, because there may be unprocessed data that the SSL code has read off the socket that it hasn't returned to you yet, so you'd block when actually there was data to return (been there). There is, as far as I can see, no sensible solution to this. The best I've seen involves mutexes and condition variables, but that's just horrible for what should just consist of passing a timeout value to the poll call. It's just that from the socket class you don't have that control. You DO have that control if you use the basic_socket_streambuf class and wrap the socket in an iostream. That's the odd thing. The functionality IS available in some form, so it's clearly recognised as useful at some level, it's just never been pushed down into the socket itself.

1 0

[boost.test] floating point comparison with user defined type
by Georgios Sermaidis 22 Mar '18

22 Mar '18

Hello, I have created a class to represent the euro currency and I am trying to use boost.test to check if two euro amounts are sufficiently close. I was expecting the following to pass the test given the tolerance level I have set, but instead it fails. What is the proper way to compare two user-defined decimals? Thank you for your time. #define BOOST_TEST_MODULE My Test #include <boost/test/included/unit_test.hpp> class eur { long double m_amount; public: eur(long double amount) : m_amount(amount){} bool operator ==(const eur& rhs) const { return m_amount == rhs.m_amount; } bool operator !=(const eur& rhs) const { return m_amount != rhs.m_amount; } friend std::ostream& operator <<(std::ostream& os, const eur& rhs); }; std::ostream& operator <<(std::ostream& os, const eur& rhs) { os << "EUR " << rhs.m_amount; return os; } BOOST_TEST_DECORATOR(* boost::unit_test::tolerance(eur(0.1))) BOOST_AUTO_TEST_CASE(first_test) { eur x(1.1); eur y(1.2); BOOST_TEST(x == y); }

2 1

boost::units::information::byte_base_unit off by a factor of 8
by Simon Doppler 21 Mar '18

21 Mar '18

I am trying to build a program that will display file sizes in human readable form. The file size is in bytes and I am using boost::units to convert the bytes to a binary prefixed unit. Using the example code found here[1], I build a small program to test the library and the output is off by a factor of 8: boost::units::quantity<boost::units::information::byte_base_unit::unit_type> b = 2048. * boost::units::information::byte_base_unit::unit_type(); std::cout << boost::units::engineering_prefix << b << std::endl; // expected 2.048 kb std::cout << boost::units::symbol_format << boost::units::binary_prefix << b << std::endl; // "expected 2 Kib" but the output on stdout is: 16.384 kb 16 Kib which is 8 times the expected result. The whole example code can be found here[2]. I have tested this code both on Archlinux (boost 1.66.0) as well as Ubuntu Xenial Server (libboost-dev 1.58.0.1ubuntu1). Do you have any idea on how to get the correct output (other than divide the number of bytes by 8 on the input) ? Best regards Simon Doppler [1] http://www.boost.org/doc/libs/1_65_0/doc/html/boost_units/Examples.html#boo… [2] https://dopsi.ch/tmp/main.cpp

2 1

[boost.test] floating point comparison with user defined type
by Georgios Sermaidis 20 Mar '18

20 Mar '18

Apologies, I just found the relevant documentation: http://www.boost.org/doc/libs/1_66_0/libs/test/doc/html/boost_test/testing_… <http://www.boost.org/doc/libs/1_66_0/libs/test/doc/html/boost_test/testing_…>

1 0

[Asio] How to implement a server and client that replies to all clients but one
by Gonzalo Garramuño 20 Mar '18

20 Mar '18

Hi. I want to have a variation on the chat server/client model in that there's a language with instructions being sent. Both the server and the client need to parse this language. When the server changes something, it sends the change in the language to all clients. When the client changes something it sends the change to the server which in turn resends it to all clients but, here's the catch, not to the one that sent the change in the first place. I managed the removal of the client by doing: ParserList::const_iterator i = clients.begin(); ParserList::const_iterator e = clients.end(); for ( ; i != e; ++i ) { try { if ( !(*i)->socket_.is_open() ) continue; std::string p = boost::lexical_cast<std::string>( (*i)->socket_.remote_endpoint() ); if ( p == id ) continue; (*i)->deliver( cmd ); } catch( const std::exception& e ) { LOG_CONN( "Parser::write " << e.what() ); } catch( ... ) { LOG_CONN( "Parser::write unhandled exception" ); } } } I managed to create a working version of the above also by modifying the asio async example using multiple inheritance for the parser (which both the client and the tcp_session inherit). However I am unhappy with this. I ask here to see if someone can suggest a more simple approach to this. -- Gonzalo Garramuño

2 1

[asio] Boost-1.65.0 TlsSlot leak on win32
by Klebsch, Mario 20 Mar '18

20 Mar '18

Hi, I am using boost version 1.65.0 in a 32 bit DLL. I use VS2005 When using ApplicationVerifier, I got application verifier stops due to leaked TLS slots. The TLS slots are allocated in win_tss_ptr_create(). Each slot is wrapped into RAII class win_tss_ptr. The constructor allocates the TLS slot and the destructor de-allocates it. I put printf-Statements in both methods and I see, that the constructor is called twice, while the destructor is not called at all. If I put a breakpoint into the destructor, it never triggers. The printf() satemets show, that the following instances have been constructed: boost::asio::detail::win_tss_ptr<class boost::asio::detail::call_stack<class boost::asio::detail::win_iocp_io_service,struct boost::asio::detail::win_iocp_thread_info>::context>::win_tss_ptr boost::asio::detail::win_tss_ptr<class boost::asio::detail::call_stack<class boost::asio::detail::strand_service::strand_impl,unsigned char>::context>::win_tss_ptr When I put a breakpoint into the constructor, I see, that both instances seem to be global variables > xxx.dll!boost::asio::detail::win_tss_ptr_create() Line 40 C++ xxx.dll!boost::asio::detail::win_tss_ptr<boost::asio::detail::call_stack<boost::asio::detail::win_iocp_io_service,boost::asio::detail::win_iocp_thread_info>::context>::win_tss_ptr<boost::asio::detail::call_stack<boost::asio::detail::win_iocp_io_service,boost::asio::detail::win_iocp_thread_info>::context>() Line 42 + 0x3a bytes C++ xxx.dll!boost::asio::detail::tss_ptr<boost::asio::detail::call_stack<boost::asio::detail::win_iocp_io_service,boost::asio::detail::win_iocp_thread_info>::context>::tss_ptr<boost::asio::detail::call_stack<boost::asio::detail::win_iocp_io_service,boost::asio::detail::win_iocp_thread_info>::context>() + 0xf bytes C++ xxx.dll!`dynamic initializer for 'boost::asio::detail::call_stack<boost::asio::detail::win_iocp_io_service,boost::asio::detail::win_iocp_thread_info>::top_''() Line 165 + 0xd bytes C++ msvcr80d.dll!_initterm(void (void)* * pfbegin=0x101c43a0, void (void)* * pfend=0x101ce9c0) Line 855 C xxx.dll!_CRT_INIT(void * hDllHandle=0x10000000, unsigned long dwReason=0x00000001, void * lpreserved=0x00000000) Line 316 + 0xf bytes C xxx.dll!__DllMainCRTStartup(void * hDllHandle=0x10000000, unsigned long dwReason=0x00000001, void * lpreserved=0x00000000) Line 492 + 0x11 bytes C xxx.dll!_DllMainCRTStartup(void * hDllHandle=0x10000000, unsigned long dwReason=0x00000001, void * lpreserved=0x00000000) Line 462 + 0x11 bytes C Does anyone on this list has an idea, why the destructors aren't called? Thanks in advance, Mario Klebsch -- Mario Klebsch Actia I+ME GmbH Mario.klebsch(a)ime-actia.de<mailto:Mario.klebsch@ime-actia.de> Dresdenstrasse 17/18 Fon: +49 531 38 701 716 38124 Braunschweig Fax: +49 531 38 701 88 Germany

2 2

C++Now 2018 Program Published
by Jon Kalb 19 Mar '18

19 Mar '18

Boosters, The program has (mostly) been published for C++Now 2018: http://cppnow.org/announcements/2018/03/2018-program-announced/ It promises to be an exciting event. It isn’t too late to decide to join us in Aspen this may: http://cppnow.org/registration/ Jon

1 0

singleton.hpp, leak
by Matt Vinson 19 Mar '18

19 Mar '18

Hello,I am using 64-bit, Windows 7, Visual Studio 2017, 15.5.7 on a MFC project. Since updating to 1.66, I think I have leaks traced to singleton.hpp. I put a _crtBreakAlloc = Allocation number and it stops near singleton.hpp.Is it possible this class now has leaks?It true, is there a work around? Any help is appreciated,

2 1

[Asio] How to assign raw sockets (descriptors) to boost::asio::ssl::stream ?
by Dmitriy V 17 Mar '18

17 Mar '18

Hi guys ! I have client server app written on boost::asio. I want to implement updating the server through `execve`. But I don't know how to assign integer socket descriptor (given by streamssl_.lowest_layer().native_handle()) to newly initialized server object after the updating (descriptors strored/restored in shared memory ...) Any help, please ! -- the best regards

2 2