Experimental Physics and
Industrial Control System

Till Straumann <[email protected]> · Tue, 23 Mar 1999 15:37:35 +0100

Marty Kraimer wrote:

> If each get/put is locked in addition to locking the lock set then database
> access must also lock each get/put. On an mv167 it taks about 50 microseconds to
> process a soft ai records which fetches a value from another record via a
> DB_LINK. It takes several microseconds (I cant remember the number but I think
> it was something like 5) to do a semTake,semGive. Thus it is a significent cpu
> overhead to lock each get/put.
>

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.

Regarding string fields, it's probably the easiest  to drop the NPP NMS exception for
such fields.

The `pointer alternative' i.e. adding an extra level of indirection to database links
and atomically switching a pointer before and after copying fields of more complex
data in order to always refer to valid data, would probably imply major changes.
The overhead however would not be outragous (one extra pointer deref)
- considering the fact that in many  cases the `old' or `last' value of a field is
already buffered.

Well, I don't want to get on your nerves too much with this issue ;-) As I said, I
have no
problem with the current behavior - I was just wondering.

Best regards

Till Straumann.

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System