[LTP] [PATCH] pthread_cancel_3-1: rewrite

[Li Wang]

Jan Stancek <jstancek@redhat.com> wrote:

>
> > ​Maybe add:
> >     SAFE_PFUNC(pthread_mutex_unlock(&mutex));
> > here too?​
>
> There are no users at this point. Is there an issue having it locked at
> exit?
>
>
​Not clear too. But after thinking over, maybe it's not a good idea to use
pthread_mutex_lock() in this situation.

The reasons are:
1. to avoid dead lock, it's not suggest to lock a mutex twice in one thread
2. it's not suggest to lock mutex in thread_A but do unlock in thread_B

​So what about using semaphore:
-----------------------------

static sem_t sem_syn;

static void *thread_func(void *unused)
{
...
       sem_post(&sem_syn);

       while (waited_for_cancel_ms < TIMEOUT_MS) {
               usleep(SLEEP_MS*1000);
               waited_for_cancel_ms += SLEEP_MS;
       }
...
}

int main(void)
{
       sem_init(&sem_syn, 0, 0);

       SAFE_PFUNC(pthread_create(&th, NULL, thread_func, NULL));

       /* wait for thread to start */
       sem_wait(&sem_syn);
       SAFE_PFUNC(pthread_cancel(th));
...
}


-- 
Regards,
Li Wang
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.linux.it/pipermail/ltp/attachments/20180604/918a3542/attachment.html>