Experimental Physics and
Industrial Control System

Benjamin Franksen <[email protected]> · Tue, 6 Oct 2009 02:09:09 +0200

On Montag, 5. Oktober 2009, Zelazny, Michael Stanley wrote:
> With no delay's or event flags, isn't your SNL below compute bound?

My mistake. This state change

>     when () {
>     } state done

from state idle will loop. You can leave it out, or, if you want selective 
responsiveness, it needs a condition

  when (cmd != 0) {
  } state done

Thus state idle is enetered only when cmd==0, and is left only if someone 
else sets it to something different to 0. The program will block and not 
consume cpu until cmd changes to something different from 0.

Cheers
Ben

> On Montag, 5. Oktober 2009, Benjamin Franksen wrote:
> On Montag, 5. Oktober 2009, Andrew Johnson wrote:
> > I suspect it would be a good idea to try and get the the SNL compiler
> > to issue a warning whenever it sees someone calling pvPut() on a
> > variable which has a monitor on it, since this is IMHO an unsafe
> > practice.
>
> It is not if you do it right. I have never observed any problems with it.
> You just have to make sure that the value your SNL program writes is
> different from what the outside writes. Typically I reserve some value
> (usually 0) for 'done' (a.k.a. 'idle'). The outside is not allowed to
> write this value (this can be enforced using yet another record). Then I
> have some SNL code like
>
> int cmd;
> assign cmd to "whatever";
> monitor cmd;
>
> state idle {
>     when (cmd==DO_SOMETHING) {
>         setup_execution();
>     } state first_intermediate_state
>     ...
>     when () {
>     } state done
> }
> ...intermediate states...
> state done {
>     when ... {
>         cmd = 0;
>         pvPut(cmd);
>     } state idle
> }
>
> The advantage is that I don't need event flags which I always found to be
> unreliable.

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System