Simple Detached pThread does not cancel ->cout blocks and interleaves even if mutexed

I have a hard problem here, which I can not solve and do not find the right answer on the net:
I have created a detached thread with a clean up routing, the problem is that on my Imac and Ubuntu 9.1 (Dual Core). I am not able to correctly cancel the detached thread in the fallowing code:

Code:

#include <iostream>
#include <pthread.h>
#include <sched.h>
#include <signal.h>

#include <time.h>

pthread_mutex_t mutex_t;

using namespace std;

static void cleanup(void *arg){
    pthread_mutex_lock(&mutex_t);
    cout << " doing clean up"<<endl;
    pthread_mutex_unlock(&mutex_t);
}


static void *thread(void *aArgument)
{

    pthread_setcancelstate(PTHREAD_CANCEL_ENABLE,NULL);
    pthread_setcanceltype(PTHREAD_CANCEL_DEFERRED,NULL);

    pthread_cleanup_push(&cleanup,NULL);
    int n=0;
    while(1){
        pthread_testcancel();
        sched_yield();
        n++;

        pthread_mutex_lock(&mutex_t);
        cout << " Thread 2: "<< n<<endl;
        pthread_mutex_unlock(&mutex_t);

    }
    pthread_cleanup_pop(0);

    return NULL;
}


int main()
{

    pthread_t thread_id;

    pthread_attr_t attr;
    pthread_attr_init(&attr);
    pthread_attr_setdetachstate(&attr,PTHREAD_CREATE_DETACHED);

    int error;

    if (pthread_mutex_init(&mutex_t,NULL) != 0) return 1;

    if (pthread_create(&thread_id, &attr, &(thread) , NULL) != 0) return 1;

    pthread_mutex_lock(&mutex_t);
    cout << "waiting 1s for thread...\n" <<endl;
    pthread_mutex_unlock(&mutex_t);

    int n =0;

    while(n<1E3){
        pthread_testcancel();
        sched_yield();
        n++;

        pthread_mutex_lock(&mutex_t);
        cout << " Thread 1: "<< n<<endl;
        pthread_mutex_unlock(&mutex_t);
    }

    pthread_mutex_lock(&mutex_t);
    cout << "canceling thread...\n" <<endl;
    pthread_mutex_unlock(&mutex_t);

    if (pthread_cancel(thread_id) == 0)
    {
        //This doesn't wait for the thread to exit
        pthread_mutex_lock(&mutex_t);
        cout << "detaching thread...\n"<<endl;
        pthread_mutex_unlock(&mutex_t);

        pthread_detach(thread_id);

        while (pthread_kill(thread_id,0)==0)
        {
                sched_yield();
        }

        pthread_mutex_lock(&mutex_t);
        cout << "thread is canceled";
        pthread_mutex_unlock(&mutex_t);

    }

    pthread_mutex_lock(&mutex_t);
    cout << "exit"<<endl;
    pthread_mutex_unlock(&mutex_t);

    return 0;
}

When I replace the Cout with printf() i workes to the end "exit" , but with the cout (even locked) the executable hangs after outputting "detaching thread...
It would be very cool to know from a Pro, what the problem here is?. Why does this not work even when cout is locked by a mutex!?
Thanks a lot for your support!!