Experimental Physics and
Industrial Control System

Eric Norum <[email protected]> · Mon, 31 Jan 2011 10:48:48 -0800

Is EINTR possible if there's some third-party code that is catching signals?
Also -- what changes are making?  Are you switching  vxWorks and RTEMS epicsEvent and epicsMessageQueue to queue-by-priority?  Are you taking the wait queue out of the POSIX epicsMessageQueue?  

On Jan 31, 2011, at 10:41 AM, Andrew Johnson wrote:

> Hi Till,
> 
> Thanks for the explanation, that helped a lot, I'll make the changes.
> 
> Has anybody ever seen any of these error messages generated:
>    pthread_mutex_lock returned EINTR. Violates SUSv3
>    pthread_cond_timedwait returned EINTR. Violates SUSv3
>    pthread_cond_wait returned EINTR. Violates SUSv3
> 
> I'm considering removing the internal routines that check for and print those 
> in the 3.15 series since I've never seen them myself or any reports of them on 
> tech-talk, and I suspect none of our current OSs would return EINTR from those 
> routines.
> 

-- 
Eric Norum
[email protected]

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System