RE: DB Question

["Szalata, Zenon M." <zms@slac.stanford.edu>]

It is an asynPortDriver implementation.

I have a skeleton readInt32 routine, which I intend to flesh out as needed.  It was returning asynSuccess, which was indeed the source of my problem.  Well, I learned something useful from this.

Thanks Mark,

Zen

 

From: Mark Rivers [mailto:rivers@cars.uchicago.edu]
Sent: Friday, November 16, 2012 9:25 AM
To: Mark Rivers; Szalata, Zenon M.; tech-talk@aps.anl.gov
Subject: RE: DB Question

 

One further note:

 

If you driver is based on asynPortDriver then you will often not be implementing the readInt32() method.  The base class readInt32() method will return asynSuccess if and only if your driver has ever called setIntegerParam() for that parameter.   So if you don’t want output records to initialize then don’t call setIntegerParam() for that parameter in your constructor.

 

Mark

 

 

From: tech-talk-bounces@aps.anl.gov [mailto:tech-talk-bounces@aps.anl.gov] On Behalf Of Mark Rivers
Sent: Friday, November 16, 2012 10:58 AM
To: 'Szalata, Zenon M.'; tech-talk@aps.anl.gov
Subject: RE: DB Question

 

Hi Zen,

 

Remember that standard asyn device support for output records does the following at iocInit:

 

1)      Attempts to read the current value from the asyn port driver

2)      If that read returns asynSuccess then the VAL field in the record is replaced by the one read from the driver

3)      If that return does not return asynSuccess then the VAL field is not modified

 

This behavior is designed to support “bumpless reboots”.

 

I suspect your driver is returning asynSuccess on a call to asynInt32->read() at iocInit.  If you want to preserve the value of the VAL field from the database file then your driver must not return asynSuccess.

 

Mark

 

 

From: tech-talk-bounces@aps.anl.gov [mailto:tech-talk-bounces@aps.anl.gov] On Behalf Of Szalata, Zenon M.
Sent: Friday, November 16, 2012 10:32 AM
To: tech-talk@aps.anl.gov
Subject: DB Question

 

Hi,

I am a bit confused by the following behavior.

I am using EPICS R3.14.12.2 and asyn R4.20.

I have a simple longout record:

 

record( longout, "$(P):$(N):LO:BTOI:NCOEF"){

  field( DESC, "N Coefs:")

  field( DTYP, "asynInt32")

  field( OUT,  "@asyn($(PORT),$(N),1)LO_NCOEF")

 field( VAL,  "2")

  field( PINI, "1")

}

 

After this record is done initializing, caget on it gives me 0.

 

I have introduced another record as follows:

 

record( longout, "$(P):$(N):LO:NCOEF"){

  field( DESC, "N Coefs:")

  field( VAL,  "2")

  field( OUT,  "$(P):$(N):LO:BTOI:NCOEF PP")

  field( PINI, "1")

}

record( longout, "$(P):$(N):LO:BTOI:NCOEF"){

  field( DESC, "N Coefs:")

  field( DTYP, "asynInt32")

  field( OUT,  "@asyn($(PORT),$(N),1)LO_NCOEF")

# field( VAL,  "2")

#  field( PINI, "1")

}

 

With these two records the second one now has the correct value.

Why does the single record solution lose its VAL value?

 

Thanks in advance,

Zen