Experimental Physics and
Industrial Control System

Andrew Johnson <[email protected]> · Tue, 24 Jan 2012 11:34:10 -0600

Hi Vesna,

On 2012-01-24 Vesna Samardzic-Boban wrote:
> 
> A Trouble Report has been submitted to our local (Australian) support and
>  then to WindRiver. In the meantime, I have a fix for the .h files that
>  enabled me to build the base - if anyone is interested to see/check, let
>  me know and I can send it.

Interesting; this may be something to do with the _VSB_CONFIG_FILE definitions 
that we have in a few places, e.g. see src/libCom/osi/os/vxWorks/osdSock.c but 
I'm not sure about that.

I would strongly recommend that you run the libCom test suite from your 
resulting binaries, which should tell you about any major problems.  Load the 
vxTestHarness.munch binary from the appropriate base/bin/vxWorks-<...> 
directory and then run the vxWorks shell command epicsRunLibComTests, then 
wait for it to finish (there's one test that takes 5 minues to run, so don't 
be too impatient with it).  Fix any failures (post the output from the failing 
tests here for help) before going any further.

> Is there any reason to not use vxWorks 6.9  for epics 3.14.12.1 or .2
>  and/or mvme 6100 or 5100 ?

We don't yet have a copy of 6.9 here at APS, but I believe that Wind River 
removed some of the APIs that EPICS uses between vxWorks 6.8 (which we use) 
and 6.9.  Since no other site has contributed suitable code yet there are no 
released versions of Base that will work on 6.9 (as you discovered). One 
reason for this is that they introduced the ability to run vxWorks on SMP 
systems, which are incompatible with one or more of the APIs that we use 
(taskVarLib for example). To support SMP properly we're going to have to 
introduce a new set of target architectures for vxWorks, and there's quite a 
bit of rewriting needed for some of the internals, but I'm not sure what if 
anything we have to do for UP support on vxWorks 6.9.

- Andrew
-- 
Optimization is the process of taking something that works and
replacing it with something that almost works, but costs less.
-- Roger Needham

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System