Experimental Physics and
Industrial Control System

Andrew Johnson <[email protected]> · Fri, 4 Dec 2009 17:55:01 -0600

Hi Stephanie,

On Friday 04 December 2009 17:16:34 Allison, Stephanie wrote:
>
> I notice that camonitor and caget do not honor the "-0x" option (option
> to printout as hex) for PVs that are unsigned long's.  Such PVs are sent
> over CA as double's and -0x is not available for doubles, per design.
> Are there other options that I could use instead?    Should I look into
> changing the tools to allow for -0x for unsigned long's?

CA can't transport unsigned types as such, although the IOC database supports 
them.  If you do a cainfo of one of your ULONG PVs you'll see that its native 
type actually gets reported as a DBF_DOUBLE, which is the smallest type that 
CA can guarantee will pass any ULONG value correctly.  Unfortunately CA gives 
no indication of what the real database type is, so I don't think the tool 
could properly detect that the field is really a ULONG.

However if you ask caget or camonitor to use DBR_LONG when fetching the value 
(add the switch -d LONG) then it does honor the -0x switch, but the display 
format is different.  Also if your client is a 64 bit machine you might get 
the sign bit extended for display.  This is the output I get on linux-x86_64 
when the top bit is set (note that the DBF_LONG and DBR_LONG types are only 32 
bits wide even on 64-bit architectures):

tux% caget -d LONG -0x anjHost:compressExample.N
anjHost:compressExample.N
    Native data type: DBF_DOUBLE
    Request type:     DBR_LONG
    Element count:    1
    Value:            0xFFFFFFFF80000000

HTH, HANW,

- Andrew
-- 
The best FOSS code is written to be read by other humans -- Harald Welte

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System