Experimental Physics and
Industrial Control System

Hello,

I wanted to follow up on this message from March where Tim suggests using the SDIS link for propagating an invalid severity in a record that scans at "I/O Intr".  In order to get the "I/O Intr" record to process (when the task that does the scanIoRequest doesn't do it since it doesn't detect the communication problem for whatever reason), the SDIS link must have "CP" set in addition to "MS" and the link itself should only post a monitor update when the severity changes.  It does mean double-processing when the severity goes good and the scanIoRequests start happening again.  Better to correct the logic of the task doing the scanIoRequest but this isn't always a simple option.

Stephanie Allison

> -----Original Message-----
> From: [email protected] 
> [mailto:[email protected]] On Behalf Of Tim Mooney
> Sent: Wednesday, March 19, 2008 6:52 AM
> To: Schuh, Stephen
> Cc: [email protected]
> Subject: Re: Severity of I/O interrupt records
> 
> Stephen,
> If the I/O-Interrupt-scanned records have input links to the record
> that sends the query, and if those links have the attribute "MS",
> then the I/O-Interrupt-scanned records should get a link alarm from
> the querying record.  The SDIS input link seems a likely candidate for
> the job.  I think you want the records to process, so they can post
> their discontent.  This means the SDIS links should go to some field
> whose value won't match the linker's DISV value.  If this works, you'd
> better comment the strategy in the database, because nobody will guess
> that the SDIS links are only being used for alarm propagation.
> 
> -- 
> Tim Mooney ([email protected]) (630)252-5417
> Beamline Controls & Data Acquisition Group
> Advanced Photon Source, Argonne National Lab.
> 
> Schuh, Stephen wrote:
> > I am using streamdevice to communicate with a VME crate controller.  One
> > record sends a query string to the crate and receives a string as a reply.
> > The reply string contains a lot of information and needs to be mapped to
> > several records.  The records that receive this information are all
> > streamdevice records with I/O interrupt processing.  This all works.
> > 
> > My problem is that I want the severity of all the I/O interrupt records to be
> > set to invalid if there is a problem with the communication to the crate.
> > The record that sends the query string has invalid severity in this case.
> > Can anyone suggest a nice way to propagate that invalid severity to all the
> > I/O interrupt records?  I thought about disabling the records, but I think I
> > would need to process the records to make this work, and I don't know how to
> > cause I/O interrupt records to process.  Is there an asyn option that will
> > cause I/O interrupt records to process after a connection problem?
> > 
> > thank you, Stephen
> > 
> 

Navigate by Date:

Prev: Re: camonitor prints CA errors to stdout rather than stderr J. Lewis Muir

Next: Another question about mbboDirect record Carl Lionberger

Index: 1994  1995  1996  1997  1998  1999  2000  2001  2002  2003  2004  2005  2006  2007  <2008>  2009  2010  2011  2012  2013  2014  2015  2016  2017  2018  2019  2020  2021  2022  2023  2024 

Navigate by Thread:

Prev: RE: Severity of I/O interrupt records Mark Rivers

Next: sysAtReboot and epicsExit Chris Slominski

Index: 1994  1995  1996  1997  1998  1999  2000  2001  2002  2003  2004  2005  2006  2007  <2008>  2009  2010  2011  2012  2013  2014  2015  2016  2017  2018  2019  2020  2021  2022  2023  2024