Thread Exiting Problem

[Kohinoor24]

I have an application which have to be run under Windows as well as Linux.It is a multithreaded application also.

Iam having problems(Run Time Errror) Exiting with the threads in Linux when it comes to the pthread_join statement.
The Windows side runs well.I have implemented it under windows & Linux as follows.
Could anyone tell me what should be done under Linux Side for safely exiting the threads

Thanks in advance....


For Windows:
------------

thread_write->Start();
thread_read->Start();
// Getting Thread Handles.
HANDLE hHandles[2] = {thread_write->GetHandle(),thread_read->GetHandle()};
int NumThreads = 2;
do
{
switch(WaitForMultipleObjects(NumThreads,hHandles,false,INFINITE))
{
case WAIT_OBJECT_0 +0:
if(hHandles[0]==thread_write->GetHandle())
m_Readaccess->SetExitFlag();// Write thread terminated - so allow reader to exit gracefully

// Put last handle in where this one completed and
// reduce number of handles in array
hHandles[0]= hHandles[--NumThreads];
break;

default:
hHandles[0]= hHandles[--NumThreads];
break;
}

}
while(NumThreads)






For Linux:
----------

typedef WriterThread<LinuxThread> WriterThread;
typedef ReaderThread<LinuxThread> ReaderThread;
m_pCycleBuffer = new T_BUFFERTYPE(BUF_SIZE);
m_BasePtr = new LinuxThread();
m_Readaccess = new C_BufferReadAccess(*m_pCycleBuffer,* m_BasePtr);
WriterThread *thread_write = new WriterThread(*m_pCycleBuffer,*m_pdma,* m_RdiFilename,*m_pmta);
ReaderThread *thread_read = new ReaderThread(*m_pCycleBuffer,*m_pida,*m_Readaccess);
thread_write->Start();
thread_read->Start();
void *status = NULL;
pthread_join(thread_write->GetHandle(),&status);
pthread_join(thread_read->GetHandle(),&status);

[cup]

Doesn't look like anything wrong with the code: are they compiled for multi-threaded operation?

If you want to force termination, you could always use _exit but that is just like a power cut: nothing exits cleanly.

[Kohinoor24]

Yes,They are compiled for Multithreaded operation.


What shold be given as the 2nd paramter to pthread_join.Iam giving it as follows now.(Null).Could that be the problem..
I want to make the writer thread exit first & then the reader Thread.
void *status = NULL;
pthread_join(thread_write->GetHandle(),&status);

[cup]

It doesn't have to be anything as it is overwritten when the thread exits. The threads have to be created as void* (void*).

Does the following example work? I don't have the pthread libraries on my Linux system at the moment so I cannot try it out.

#include <pthread.h>
#include <iostream>
#include <stdlib.h>
using namespace std;
bool running = true;
void* One (void*)
{
while (running)
cout << "One" << endl;

return (void*) 0;
}

void* Two (void*)
{
while (running)
cout << "Two" << endl;

return (void*) 0;
}

main ()
{
pthread_t one, two;
void* status;
pthread_create (&one, 0, One, (void*) 0);
pthread_create (&two, 0, Two, (void*) 0);

sleep (60);
running = false;
pthread_join (one, &status);
pthread_join (two, &status);
return 0;
}

Another possibility is to use pthread_exit ((void*) 0) instead of return (void*) 0.

[Kohinoor24]

I made a project with ur file.I produced the .o object.No problem.I commented out the sleep(), call as it was showing undeclared identifier.
I got an *.exe under linux.
Iam new to Linux.I dont know how to run the exe from konsole

[cup]

Sorry - I forgot: sleep is in unistd.h. If you add unistd.h to the bunch of includes, it should work with sleep in. Just type the program name with a ./ in front. For instance, if it is called a.out

./a.out

Minor note: if you are eventually going to transfer this to Solaris, sleep might not work as both pthreads and sleep use SIGALRM. If you are staying with Linux, it is not a problem.

[PaulWendt]

This kind of indirectly answers your question. If you used a thread
library that already supports multiple platforms, you have one
client-code base. In addition, if you use one that's already out
there, it'll probably have underwent a lot of testing.

So, you should check out the boost threading stuff. I believe they
implement all the thread stuf natively, but it's all abstracted so
that you don't HAVE to know the details.

It's just a suggestion.

--Paul

[Kohinoor24]

When Iam Debugging I have another problem ,Iam getting a Deadlock

I have implemented the LinuxMutex class as follows:

LinMutex::LinMutex()
{
pthread_mutex_init(&m_hMutex,(pthread_mutexattr_t*)NULL);

}
void LinMutex::AcquireLock()
{
pthread_mutex_lock(&m_hMutex);
}

void LinMutex::ReleaseLock()
{
pthread_mutex_unlock(&m_hMutex);
}

LinMutex::~LinMutex()
{
pthread_mutex_destroy(&m_hMutex);

}

Is it alright.I heard from a colleuge,that If a thread which already owns a Mutex,Try to access,it again there will be a deadlock under Linux.

In windows,I have implemented the lock() as follows to remove the DeadLock as said in Msdn:
// from Msdn
The thread that owns a mutex can specify the same mutex in repeated wait function calls without blocking its execution. Typically, you would not wait repeatedly for the same mutex, but this mechanism prevents a thread from deadlocking itself while waiting for a mutex that it already owns. However, to release its ownership, the thread must call ReleaseMutex once for each time that the mutex satisfied a wait.

void WinMutex::AcquireLock()
{
while (WaitForSingleObject(m_hMutex,1000L)!= WAIT_OBJECT_0);
}


What should I do to remove the Deadlock.What should be done in The AcquireLock() of LinMutex.
Could u help me on this regard.....