#include <pthread.h>
#include <iostream>
using namespace std;
void OnCreateThread()
{
cout << "Create a thread." << endl;
}
void OnExitThread()
{
cout << "Exit a thread." << endl;
}
void f(void*) {}
int main()
{
//
// What to do here ???
//
pthread_t dummy;
pthread_create(&dummy, 0, f, 0);
pthread_create(&dummy, 0, f, 0);
while (true);
}
The code creates two native threads, other than std::thread, and I want it to output as follows:
Create a thread.
Create a thread.
Exit a thread.
Exit a thread.
It can be done under Windows by using FlsXXX functions.
However, I don't know whether it can also be done under Linux.
Is there a standard way under Linux?
How to call a function on a thread's creation and exit?
Pthreads API does not provide callbacks for thread creation (nor does std::thread API).
Solution is pretty simple however: Call the functions at the beginning and end of start_routine callback.
void* f(void*) {
OnCreateThread();
OnExitThread();
return nullptr;
}
In case you might want OnExitThread to be called even when the thread has been terminated prematurely, you might want to use pthread_cleanup_push to register it as a callback.
PS. The start_routine callback must return void*.
There exists at least a pthread_cleanup_push function that lets you add a function that will be called just after thread termination. Never heard about the same for creation, but some API may have such.
Related
Is there a better way to use CreateThread than creating a free function each time for the sole purpose of casting lpParameter?
Are there any modern alternatives to CreateThread for creating persistent threads?
Edit: Perhaps you should just use std::async(lambda). I imagine that it's just implemented with CreateThread. Maybe the answer to this question is looking up how std::async is implemented (assuming it's a library feature).
DWORD WINAPI MyThreadFunction(
_In_ LPVOID lpParameter
)
{
((MyClass*)lpParameter)->RunLoop();
}
void MyClass::LaunchThread()
{
CreateThread(
NULL, // default security attributes
0, // use default stack size
MyThreadFunction, // thread function name
this, // argument to thread function
0, // use default creation flags
NULL); // returns the thread identifier
}
There are several mechanisms for achieving parallelism (std::async etc. as mentioned above).
But the modern one which is most similar to your original code with CreateThread is std::thread. It can be constructed with a global function, a lambda, or a class method (which seems the best fit for you):
m_thread = std::thread([this](){ RunLoop(); }); // pass a lambda
or
m_thread = std::thread(&MyClass::RunLoop, this); // pass a method
Note that a std::thread starts to run (potentially) when constructed. Also note that, std::async does not guarantee that it will run on a separate thread and even if it does run on a thread, it could be a thread from a pool. The behaviour might not be the same as with your original CreateThread.
Here's a complete example of using std::thread (including cancellation):
#include <thread>
#include <chrono>
#include <atomic>
#include <iostream>
class MyClass
{
public:
MyClass() {}
~MyClass() { EndThread(); }
void LaunchThread()
{
EndThread(); // in case it was already running
m_bThreadShouldExit = false;
// Start the thread with a class method:
m_thread = std::thread(&MyClass::RunLoop, this);
}
void EndThread()
{
// Singal the thread to exit, and wait for it:
m_bThreadShouldExit = true;
if (m_thread.joinable())
{
m_thread.join();
}
}
void RunLoop()
{
std::cout << "RunLoop started" << std::endl;
while (!m_bThreadShouldExit)
{
std::cout << "RunLoop doing something ..." << std::endl;
std::this_thread::sleep_for(std::chrono::milliseconds(1000));
}
std::cout << "RunLoop ended" << std::endl;
}
private:
std::thread m_thread;
std::atomic_bool m_bThreadShouldExit{ false };
};
int main()
{
MyClass m;
m.LaunchThread();
std::this_thread::sleep_for(std::chrono::milliseconds(5000));
m.EndThread();
}
Possible output:
RunLoop started
RunLoop doing something ...
RunLoop doing something ...
RunLoop doing something ...
RunLoop doing something ...
RunLoop doing something ...
RunLoop ended
std::async() and std::thread(, <args...>) are most likely internally implemented as you just did, the only exception is that lambdas without captures can be implicitly converted to function pointers, which pretty much can be passed straight away to CreateThread function with nullptr lpParameter.
Lambdas with capture list are pretty much syntactic sugar but internally they translate to sth like this (very simplified):
struct <internal_lambda_name>
{
<capture list...> fields...;
void operator()(<arguments...>){<code...>;}
};
So they pretty much translate to objects of struct type thus they need some way to store all those captures and in order to be executed on other thread with CreateThread function they need some way of ensuring that the capture list data stored in them will be available during their execution.
I looked in to MSVC implementation of std::async and they implemented it using ::Concurrency::create_task which straight forwardly accepts a callable object.
https://learn.microsoft.com/en-us/cpp/parallel/concrt/task-parallelism-concurrency-runtime
I also looked into their implementation of create_task
template<typename _Ty>
__declspec(noinline) // Ask for no inlining so that the _CAPTURE_CALLSTACK gives us the expected result
explicit task(_Ty _Param)
{
task_options _TaskOptions;
details::_ValidateTaskConstructorArgs<_ReturnType,_Ty>(_Param);
_CreateImpl(_TaskOptions.get_cancellation_token()._GetImplValue(), _TaskOptions.get_scheduler());
// Do not move the next line out of this function. It is important that _CAPTURE_CALLSTACK() evaluates to the call site of the task constructor.
_SetTaskCreationCallstack(_CAPTURE_CALLSTACK());
_TaskInitMaybeFunctor(_Param, decltype(details::_IsCallable(_Param,0))());
}
and so it turns out that launching a lambda on a new thread is quite difficult and beyond the scope of this question.
I am doing a kind of shell: depending of the user's entry, I must call some function. I cannot modify the content of those called functions since my program is only a client and has no visibility of them.
But I want the possibility for the user to kill the call using CTRL+C. Here is the minimal code:
#include <csignal>
#include <iostream>
#include <unistd.h>
void do_thing(void)
{
std::cout << "entering in do_thing()\n";
while(42)
::sleep(1);
}
extern "C" {
void signal_handler(int);
}
class Shell
{
friend void signal_handler(int);
public:
static Shell & Instance(void)
{
static Shell instance;
return instance;
}
int run(void)
{
std::string buff;
while ((std::cin >> buff))
{
if (buff == "foo")
do_thing(); // this must be terminable
else
std::cout << "(nothing)\n";
}
return 0;
}
private:
Shell(void)
{
::signal(SIGINT, signal_handler);
}
void signal(int sig)
{
if (sig == SIGINT)
;// must terminal the function call
}
};
extern "C" {
void signal_handler(int sig)
{
Shell::Instance().signal(sig);
}
}
int main(void)
{
return Shell::Instance().run();
}
I considered three possibilities:
I tried to create a thread class derived from std::thread, with a kill() method that throws an exception. The function call is in a try-catch block. It works, but this is a bad solution since the destructor cannot be called, and the resource is never freed.
I considered using fork, but I think it is an overkill to just get the possibility of interrupt a function call.
I tried to throw an exception from the signal handler, but I saw that this is a bad idea since this is very compiler/OS dependent code.
How could you do the thing? What is the better solution?
Note: I deleted the old post because it was close requested, and took into consideration the C/C++ tags.
Essentially, no, there is no standard why to interrupt a thread in C++. Threads run co-operatively and as such, they need to "give up" control.
If the code for do_thing were modifiable, then you can create a flag (atomic) to signal that the thread should give up and exit. This can be periodically checked by the thread and complete as required.
Given the code for do_thing is not modifiable, there is a small window of opportunity that can be used to "kill" or "cancel" the thread (albeit it won't be "standard" and support will be limited to targeted platforms).
std::thread offers a function to retrieve a native_handle() that is implementation defined. Once obtained (and converted), it can be used to kill or cancel the thread.
If pthreads are being used, see pthread_kill (or pthread_cancel if supported by the target thread).
On windows, see the TerminateThread function.
Be warned; aside from the platform specific code required, the thread terminations generally leave the objects on that thread in "limbo" and with them, the resources they control.
I have a function that I want to run whenever my program exits:
void foo() {
std::cout<< "Exiting" << std::endl;
}
How do I register it to be run whenever the program exists, regardless of when and why it exits - due to signal, exit() call, etc?
You can use the aptly named std::atexit function in the cstdlib header:
#include <cstdlib>
void exiting() {
std::cout << "Exiting";
}
int main() {
std::atexit(exiting);
}
The system will maintain a stack of functions registered with atexit and call them each in the reverse order of their registration when either the exit function is called, or the program returns from main. You can register at least 32 functions this way.
I am answering as a Linux user, but all of this should apply to windows.
I had this similar question, so hopefully I can sum up previous answers and add my two cents.
Signals and abort(): ^C and ^Z can be "intercepted" to call your function before exiting, presumably with exit(). Signals SIGQUIT AKA ^\ and SIGKILL which has no key stroke cannot be intercepted. Here's an example for using the csignal header and a C++ lambda.
#include <iostream>
#include <csignal>
#include <cstdlib>
using namespace std;
int main()
{
//signal requires lam take an int parameter
//this parameter is equal to the signals value
auto lam =
[] (int i) { cout << "aborting" << endl; exit(0); };
//^C
signal(SIGINT, lam);
//abort()
signal(SIGABRT, lam);
//sent by "kill" command
signal(SIGTERM, lam);
//^Z
signal(SIGTSTP, lam);
while(1)
{
}
return 0;
}
Exit: Since I used exit() in my examples above, care must be taken here. If the function being run is a clean-up function that only needs to run once, perhaps a static variable has_run could be used. Or in the example above, raise() a signal that you can't intercept. But those tend to come with core dumps which just feels dirty. Your choice, here. An example follows
#include <cstdlib>
#include <iostream>
using namespace std;
int main()
{
//called with no parameters
auto lam = [] () { cout << "at exit"; };
atexit(lam);
return 0;
}
Take note that c++11 added a quick_exit which has an accompanying at_quick_exit which act the same as above. But with quick_exit no clean up tasks are performed. In contrast, with exit object destructors are called and C streams are closed, with only automatic storage variables not getting cleaned up.
You could put it in the destructor of a class with a global instance.
class SomeGlobalStuff {
~SomeGlobalStuff() {
foo();
}
static SomeGlobalStuff instance;
};
// putting this in a single compilation unit.
SomeGlobalStuff SomeGlobalStuff::instance instance;
But like any other method, you have to remember that you cannot use any data if you cannot garantee that it still exists. Deallocation of global objects is done in a arbitrary order, so basically, you cannot use std::cout in the foo() function. atexit() is worse in this regard, because whether it executes before or after destruction of global objects depends on the compiler and compiler options.
And anyway, you still have to handle signals correctly. You have to choose which signals to handle and which to not handle (you most likely don't want to handle SIGSEGV). You cannot escape signal handling. And remember that signals may interrupt your program at any time (unless masked) so your data structures might be in a arbitrary state, in the middle of an update.
The only way (in Unix and Unix-like operating systems) to regain control after a process exits is to wait(2) for it. Short of a powerfail, kernel panic, or forced reboot, this should work:
#include <sys/types.h>
#include <sys/wait.h>
#include <iostream>
int AtExit() {
pid_t pid = fork();
if(pid < 0) return pid;
if(pid == 0) return pid;
pid = waitpid(pid, 0, 0);
return pid;
}
int main () {
if(AtExit()) {
std::cout << "Exiting\n";
return 0;
}
std::cout << 7 << "\n";
}
I want to call a function which will be asynchronous (I will give a callback when this task is done).
I want to do this in single thread.
This can be done portably with modern C++ or even with old C++ and some boost. Both boost and C++11 include sophisticated facilities to obtain asynchronous values from threads, but if all you want is a callback, just launch a thread and call it.
1998 C++/boost approach:
#include <iostream>
#include <string>
#include <boost/thread.hpp>
void callback(const std::string& data)
{
std::cout << "Callback called because: " << data << '\n';
}
void task(int time)
{
boost::this_thread::sleep(boost::posix_time::seconds(time));
callback("async task done");
}
int main()
{
boost::thread bt(task, 1);
std::cout << "async task launched\n";
boost::this_thread::sleep(boost::posix_time::seconds(5));
std::cout << "main done\n";
bt.join();
}
2011 C++ approach (using gcc 4.5.2, which needs this #define)
#define _GLIBCXX_USE_NANOSLEEP
#include <iostream>
#include <string>
#include <thread>
void callback(const std::string& data)
{
std::cout << "Callback called because: " << data << '\n';
}
void task(int time)
{
std::this_thread::sleep_for(std::chrono::seconds(time));
callback("async task done");
}
int main()
{
std::thread bt(task, 1);
std::cout << "async task launched\n";
std::this_thread::sleep_for(std::chrono::seconds(5));
std::cout << "main done\n";
bt.join();
}
As of C++11, plain c++ does have a concept of threads, but the most concise way to call a function asynchronously is to use the C++11 async command along with futures. This ends up looking a lot like the way you'd do the same thing in pthreads, but it's 100% portable to all OSes and platforms:
Say your function has a return value... int = MyFunc(int x, int y)
#include <future>
Just do:
// This function is called asynchronously
std::future<int> EventualValue = std::async(std::launch::async, MyFunc, x, y);
Catch? How do you know when it's done? (The barrier.)
Eventually, do:
int MyReturnValue = EventualValue.get(); // block until MyFunc is done
Note it's easy to do a parallel for loop this way - just create an array of futures.
You can't in plain C++. You'll need to use an OS-specific mechanism, and you need a point where execution is suspended in a way that allows the OS to execute the callback. E.g. for Windows, QueueUserAPC - the callback will be executed when you e.g. SleepEx or WaitForSingleObjectEx
The long answer involves implementing your own task scheduler and wrapping your "function" up into one or more tasks. I'm not sure you want the long answer. It certainly doesn't allow you to call something, completely forget about it, and then be notified when that thing is done; however if you are feeling ambitious, it will allow you to simulate coroutines on some level without reaching outside of standard C++.
The short answer is that this isn't possible. Use multiple threads or multiple processes. I can give you more specific information if you divulge what OS/platform you're developing for.
There are two bits to doing this.
Firstly, packing up the function call so that it can be executed later.
Secondly, scheduling it.
It is the scheduling which depends on other aspects of the implementation. If you know "when this task is done", then that's all you need - to go back and retrieve the "function call" and call it. So I am not sure this is necessarily a big problem.
The first part is then really about function objects, or even function pointers. The latter are the traditional callback mechanism from C.
For a FO, you might have:
class Callback
{
public:
virtual void callMe() = 0;
};
You derive from this and implement that as you see fit for your specific problem. The asyncronous event queue is then nothing more than a list<> of callbacks:
std::list<Callback*> asyncQ; // Or shared_ptr or whatever.
I'm not sure I understand what you want, but if it's how to make use of a callback: It works by defining a function pointer, like this (untested):
// Define callback signature.
typedef void (*DoneCallback) (int reason, char *explanation);
// A method that takes a callback as argument.
void doSomeWorkWithCallback(DoneCallback done)
{
...
if (done) {
done(1, "Finished");
}
}
//////
// A callback
void myCallback(int reason, char *explanation)
{
printf("Callback called with reason %d: %s", reason, explanation);
}
/////
// Put them together
doSomeWortkWithCallback(myCallback);
As others have said, you technically can't in plain C++.
However, you can create a manager that takes your task and does time-slicing or time scheduling; with each function call, the manager uses a timer to measure the amount of time the process took; if the process took less time than scheduled, and it thinks it can finish another call and use up the remaining time without going over, it can call it again; if the function does go over the alloted time, it means the function has less time next update to run. So, this will involve creating a somewhat complex system to handle it for you.
Or, if you have a specific platform in mind, you could use threading, or create another process to handle the work.
This question's answers are a community effort. Edit existing answers to improve this post. It is not currently accepting new answers or interactions.
Can someone post a simple example of starting two (Object Oriented) threads in C++.
I'm looking for actual C++ thread objects that I can extend run methods on (or something similar) as opposed to calling a C-style thread library.
I left out any OS specific requests in the hopes that whoever replied would reply with cross platform libraries to use. I'm just making that explicit now.
Create a function that you want the thread to execute, for example:
void task1(std::string msg)
{
std::cout << "task1 says: " << msg;
}
Now create the thread object that will ultimately invoke the function above like so:
std::thread t1(task1, "Hello");
(You need to #include <thread> to access the std::thread class.)
The constructor's first argument is the function the thread will execute, followed by the function's parameters. The thread is automatically started upon construction.
If later on you want to wait for the thread to be done executing the function, call:
t1.join();
(Joining means that the thread who invoked the new thread will wait for the new thread to finish execution, before it will continue its own execution.)
The Code
#include <string>
#include <iostream>
#include <thread>
using namespace std;
// The function we want to execute on the new thread.
void task1(string msg)
{
cout << "task1 says: " << msg;
}
int main()
{
// Constructs the new thread and runs it. Does not block execution.
thread t1(task1, "Hello");
// Do other things...
// Makes the main thread wait for the new thread to finish execution, therefore blocks its own execution.
t1.join();
}
More information about std::thread here
On GCC, compile with -std=c++0x -pthread.
This should work for any operating-system, granted your compiler supports this (C++11) feature.
Well, technically any such object will wind up being built over a C-style thread library because C++ only just specified a stock std::thread model in C++0x, which was just nailed down and hasn't yet been implemented.
The problem is somewhat systemic. Technically the existing C++ memory model isn't strict enough to allow for well-defined semantics for all of the 'happens before' cases. Hans Boehm wrote an paper on the topic a while back and was instrumental in hammering out the C++0x standard on the topic.
Threads Cannot be Implemented as a Library
That said, there are several cross-platform thread C++ libraries that work just fine in practice. The Intel thread building blocks contains a tbb::thread object that closely approximates the C++0x standard and Boost has a boost::thread library that does the same.
oneAPI Threading Building Blocks
Chapter 19. Thread (Boost documentation)
Using boost::thread, you'd get something like:
#include <boost/thread.hpp>
void task1() {
// do stuff
}
void task2() {
// do stuff
}
int main (int argc, char ** argv) {
using namespace boost;
thread thread_1 = thread(task1);
thread thread_2 = thread(task2);
// do other stuff
thread_2.join();
thread_1.join();
return 0;
}
#include <thread>
#include <iostream>
#include <vector>
using namespace std;
void doSomething(int id) {
cout << id << "\n";
}
/**
* Spawns n threads
*/
void spawnThreads(int n)
{
std::vector<thread> threads(n);
// spawn n threads:
for (int i = 0; i < n; i++) {
threads[i] = thread(doSomething, i + 1);
}
for (auto& th : threads) {
th.join();
}
}
int main()
{
spawnThreads(10);
}
There is also a POSIX library for POSIX operating systems.
Check for compatibility:
#include <stdio.h>
#include <stdlib.h>
#include <pthread.h>
#include <iostream>
void *task(void *argument){
char* msg;
msg = (char*)argument;
std::cout << msg << std::endl;
}
int main(){
pthread_t thread1, thread2;
int i1, i2;
i1 = pthread_create(&thread1, NULL, task, (void*) "thread 1");
i2 = pthread_create(&thread2, NULL, task, (void*) "thread 2");
pthread_join(thread1, NULL);
pthread_join(thread2, NULL);
return 0;
}
Compile with -lpthread.
POSIX Threads
When searching for an example of a C++ class that calls one of its own instance methods in a new thread, this question comes up, but we were not able to use any of these answers that way. Here's an example that does that:
Class.h
class DataManager
{
public:
bool hasData;
void getData();
bool dataAvailable();
};
Class.cpp
#include "DataManager.h"
void DataManager::getData()
{
// perform background data munging
hasData = true;
// be sure to notify on the main thread
}
bool DataManager::dataAvailable()
{
if (hasData)
{
return true;
}
else
{
std::thread t(&DataManager::getData, this);
t.detach(); // as opposed to .join, which runs on the current thread
}
}
Note that this example doesn't get into mutex or locking.
Unless one wants a separate function in the global namespace, we can use lambda functions for creating threads.
One of the major advantage of creating a thread using lambda is that we don't need to pass local parameters as an argument list. We can use the capture list for the same and the closure property of lambda will take care of the lifecycle.
Here is sample code:
int main() {
int localVariable = 100;
thread th { [=]() {
cout << "The value of local variable => " << localVariable << endl;
}};
th.join();
return 0;
}
By far, I've found C++ lambdas to be the best way of creating threads especially for simpler thread functions.
It largely depends on the library you decide to use. For instance, if you use the wxWidgets library, the creation of a thread would look like this:
class RThread : public wxThread {
public:
RThread()
: wxThread(wxTHREAD_JOINABLE){
}
private:
RThread(const RThread ©);
public:
void *Entry(void){
//Do...
return 0;
}
};
wxThread *CreateThread() {
//Create thread
wxThread *_hThread = new RThread();
//Start thread
_hThread->Create();
_hThread->Run();
return _hThread;
}
If your main thread calls the CreateThread method, you'll create a new thread that will start executing the code in your "Entry" method. You'll have to keep a reference to the thread in most cases to join or stop it.
More information is in the wxThread documentation.