Experimental Physics and
Industrial Control System

Andrew Johnson <[email protected]> · Mon, 16 Jul 2001 14:54:09 -0500

Geoff Savage wrote:
> 
> We see an INVALID alarm when writing to fields in a Fermilab specific
> record that are not the VAL field using a longout record and Soft
> Channel device support.  I think the INVALID alarm is correct based on
> the code but I don't know why it should be an error.  In dbPut
<snip> 
> Is this how it should work?

I'm not sure about your description of the Fermilab specific record stuff,
but hopefully the following description will help.  The UDF field tells
you whether the contents of the VAL field have ever been set by somebody
(it's like gcc's warning that an uninitialized variable is being used). 
If nothing ever sets a record's .VAL field the record should remain in the
UDF_ALARM state.  If your longout record is in a UDF state, is processed
and its OUT link has the MS (Maximize Severity) flag set, then the link
destination record will also be put into LINK_ALARM state with an INVALID
severity.  I suspect the latter is why your fermilab record type is in
INVALID alarm severity.  Check the value of the STAT as well as the SEVR
fields, they provide different information.

If you want some default initial value in the longout record, you should
be able to just set the longout's DOL field to that value, but I forget
whether you need to set OMSL to closed_loop or not, probably not.

- Andrew
-- 
The world is such a cheerful place when viewed from upside-down
It makes a rise of every fall, a smile of every frown

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System