Re: Port of EPICS 3.14.9 to ETRAX CRIS architecture - Strange Data

[Andrew Johnson <anj@aps.anl.gov>]

Hi Peter,

Peter Zumbruch wrote:
>    - The file "$EPICS-BASE/src/libCom/osi/default/osdWireFormat.h"
>    has been extended by those two lines to comply for the 
> little-endianness :
>
>     #elif defined (_cris_)
>     #     define EPICS_LITTLE_ENDIAN

Are you sure that this architecture is little-endian?  It looks to me 
from your results below like you should try using EPICS_BIG_ENDIAN. 
Here's why:

> name:   zumbruchHost:aiExample2
> DBR_STRING      3
> DBR_SHORT       768
> DBR_LONG        50331648
>
> and what it should look like:
>
> name:   zumbruch2Host:aiExample2
> DBR_STRING      3
> DBR_SHORT       3
> DBR_LONG        3

If I convert the short result 768 to hex, the result is 0x0300 instead 
of 0x0003, i.e. the high and low bytes of the 16 bit result have been 
swapped from what they should contain.  Similarly converting the long 
result to hex gives 0x03000000 which implies the wrong endianness.

Try making that change in osdWireFormat.h and recompiling and I suspect 
this problem will disappear.

Once this is resolved I'll be happy to add this new architecture to the 
next official release.

- Andrew
--
When a distinguished but elderly scientist states that something is
possible, he is almost certainly right.  When he states that something
is impossible, he is very probably wrong.  -- Arthur C. Clarke