Experimental Physics and
Industrial Control System

Andrew Johnson <[email protected]> · Tue, 23 Mar 1999 11:00:52 -0600

Till Straumann wrote:
> 
> I completely agree. That's why I suggested to just disable interrupts while copying
> small multiple word data types (such as scalar doubles). This should be quite
> efficient
> and not affecting interrupt latency very much.
> AFAIK, this could be done at a small number of places, namely in the
> putXXXDouble routines.

The taskLock() (NB an intLock() is unnecessarily strong) would also need
to be in any record or device code which updates a multi-word field of its
records.  Given the large number of doubles which occur in record support
modules this won't be quite as easy as you suggest.  However I suspect
that only the cheaper MVME162 CPUs will need this for doubles as they use
software floating-point.  Other CPUs will get the FP co-processor to write
doubles, and those operations are probably not interruptable.  I've not
checked the compiler output though, I'm just guessing.

- Andrew
-- 
New country, job, address, car and signature; same old Andrew Johnson

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System