Experimental Physics and
Industrial Control System

[email protected] · Tue, 10 Oct 95 10:19:00 -0500

>Thanks for your reply concerning the use of vxworks and the mvme162.
>...
>So if you could, send me instructions on how I can obtain a copy
>of at least your modified version, and hopefully the unmodified as
>well.

The 'config' tree of the vxWorks 5.2 distribution with my EPICS
driver mods in it can be found:

ftp://epics.aps.anl.gov/vxWorks/README
ftp://epics.aps.anl.gov/vxWorks/config.tar.gz

To access the file you must use the 'beta' account on epics.aps.anl.gov. 
You may contact [email protected] if you don't know the password.

Read the README file first or you will void your warranty ;-)  A copy of
it is included here:

-----------------------------------------------------------------------

This is a modified config subtree for the vxWorks distribution of release 5.2.
Note that this is NOT a BSP!!!!!!!!!  This is simply some tailoring!  This 
contains an improved memory setup as well as some additional functions that 
were added to support some of the EPICS drivers that require the mapping of 
memory into the A24 VME region.

To use it you must first install vxWorks on your system (including the BSPs.)  
When that is complete, you will have two directories, vw and vxgccV2.2.3.1.  
To install these BSP mods, you should do the following:

cd vw
mv config config.DISTRIB
tar xvf <whereever>/config.tar.gz
cd config/mv167
sh bld
cd ../mv162
sh build

The 'bld' scripts may have to be tailored for your site.  As-is, they should
be OK if you are on a sun4 and install vxworks as sugested in the distribution 
notes.  However since they uses relative paths for references to other 
directories, they must be run from the config/<arch> directories.

You will find both the 167 and 162 versions of the BSP mods here.  

If you get errors about missing routines with names like 'muldi3' or other
things that have 'mul' or 'div' in them, you are building your code for 
soft-float support and have forgotten to load the software floating point 
support libraries.  You may either figure out how to load the soft-float
library or rebuild your code without the soft-float flag set on the compiler.

I do not use soft-float, you may send email to [email protected] to ask
around about it.  If there is a FAQ posted, I will add it to this README file.
However, I suspect that the vxWorks distribution discusses it somewhere.

For those of you that are ${PC_COMMENT}, this soft-float stuff is related to 
the version of the 162 that you buy.  Some of the lesser priced boards don't 
have floating point support in the hardware, thus it must be taken care of in 
software.  These 'mul' and 'div' functions are that support.

Have fun!

--John Winans
[email protected]

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System