Experimental Physics and
Industrial Control System

Benjamin Franksen <[email protected]> · Mon, 30 Apr 2007 10:35:07 +0200

On Friday 27 April 2007 21:37, Mark Rivers wrote:
> epics> threadTest 1
> Calling pthread_create
> pthread_create status=0
> Sleeping 5 seconds
> epics>
> epics>
> epics> threadTest 0
> Calling pthread_create
> pthread_create status=11
> pthread_create error Resource temporarily unavailable
> Segmentation fault (core dumped)
>
> So pthread_create is returning status=11 (EAGAIN) when things mess up.
> Any idea what this means and how to fix it?  And why is the error
> causing a seg fault, since it did simply return an error.

Hi Mark,

you probably know this but just in case: EAGAIN is the hallmark of 
cheap-and-dirty system design. In Unix, when a system call gets interrupted 
by a signal and must be restarted, the system doesn't automatically restart 
the call. Instead the system call returns the error code EAGAIN. Thus, in 
principle, every such call must be wrapped (by the user!) in a loop like 
this

do {
  ret = syscall(...);
} while (ret != EAGAIN);

How this relates to the concrete problem you are having I can't say, other 
than that it seems the reason in this case is not a signal but rather a 
'temporarily unavailable resource'.

BTW, since the status you get is not EPERM, the crash must happen some where 
inside the rest of the wrapper routine (the part which you did not quote), 
which is

static void * start_routine(void *arg)
{

	/*...error not in this part...*/

    if(status) {
        free_threadInfo(pthreadInfo);
        return 0;
    }
    status = pthread_sigmask(SIG_SETMASK,&oldSig,NULL);
    checkStatusOnce(status,"pthread_sigmask");
    return(pthreadInfo);
}

that is, either in free_threadInfo() or pthread_sigmask(). (I am assuming 
the crash really happens inside epicsThreadCreate().)

HTH
Ben

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System