Experimental Physics and
Industrial Control System

<[email protected]> · Wed, 23 Oct 2013 09:14:43 +0000

Hi Andrew,

Sorry to resurrect an old thread, but I think I have an issue with the 0.1s penalty for over-runs of periodic tasks. This means that if the 0.1s scan task overruns, it will delay by 0.1s, which causes it to overrun the next time, and for every scan forever. I've noticed this behaviour when we had a misbehaving IOC running on a badly configured realtime linux system. All scan threads would overrun once, but the 0.1s then overran every time after that.

Reducing the delay to a bit less than the fastest scan period fixes this. I set it to epicsThreadSleepQuantum(), but that is probably a bit drastic...

Thanks,
Tom Cobb

> -----Original Message-----
> From: [email protected] [mailto:tech-talk-
> [email protected]] On Behalf Of Andrew Johnson
> Sent: 09 January 2013 19:46
> To: [email protected]
> Cc: Eric Norum
> Subject: Re: time drift in camonitor timestamps
> 
> Hi Eric,
> 
> FYI Jack wasn't the reporter.
> 
> Your fixed version works fine for me.  I'm adding some code to print
> warnings
> about continuous overruns (at increasing periods, up to hourly) before
> I
> commit it.
> 
> I'm also inclined to keep the original delay = 0.1 penalty for an over-
> run
> instead of dropping it to 0.02 seconds, but I will consider arguments
> against
> that if there are objections.
> 
> - Andrew
> 
> On 2013-01-09 Eric Norum wrote:
> > Sorry.
> > Previous message had wrong periodicTask code.  Here's what I meant to
> send.
> >
> > Jack, can you try this and see if it fixes the problem that you were
> >  seeing/
> >
> >
> > static void periodicTask(void *arg)
> > {
> >     periodic_scan_list *ppsl = (periodic_scan_list *)arg;
> >
> >     epicsTimeStamp nextScan, now;
> >     double delay;
> >
> >     taskwdInsert(0, NULL, NULL);
> >     epicsEventSignal(startStopEvent);
> >
> >     epicsTimeGetCurrent(&nextScan);
> >     while (ppsl->scanCtl != ctlExit) {
> >         if (ppsl->scanCtl == ctlRun) scanList(&ppsl->scan_list);
> >         epicsTimeAddSeconds(&nextScan, ppsl->period);
> >         epicsTimeGetCurrent(&now);
> >         delay = epicsTimeDiffInSeconds(&nextScan, &now);
> >         if (delay <= 0.0) {
> >             ppsl->overrunCount++;
> >             delay = 0.02;
> >             epicsTimeGetCurrent(&nextScan);
> >         }
> >         epicsEventWaitWithTimeout(ppsl->loopEvent, delay);
> >     }
> >
> >     taskwdRemove(0);
> >     epicsEventSignal(startStopEvent);
> > }
> 
> --
> There is no such thing as a free lunch.  When invited for lunch,
> it is best to check if you are there to eat, or to be eaten.
> -- Clive Robinson

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System