Experimental Physics and
Industrial Control System

Tony Cox - (415)926-3105 <[email protected]> · Thu, 13 Apr 1995 15:22:03 PST

Bob writes:-

>Also - what is the state of the Alpha port?

The Alpha/OSF-1 port is proceeding. Here are the major milestones so far.

	1) DEC VxWorks builds the IOC components without problems. There are,
	   however, some symbols which need to be changed from the defaults
	   when you build the vxWorks image so that all the routines needed
	   to support IOC code are resident. I sent (someone?) a short
	   README on this. Is it in the distribution??

	2) iocCore & all the rest load and appear to run just fine. I have
	   not extensively tested them, but there is no reason to believe
	   anything extraordinary will happen in production. We aim to go
	   'live' - controlling hardware on our beamline - starting May/June,
	   so stay tuned!

	3) Channel access now works. Jeff & I found & removed the last known
	   bug (missing endian-swap) some 3 weeks ago. This would have
	   manifested itself in MEDM on any little-endian system.

	4) cau/caget/caput check out just fine. (there were some minor changes
	   needed). Note that for all the `host side' software, I have
	   tested & debugged everything in full 64 bit mode (there is a 
	   compiler switch which will force pointers into 32 bits & a linker
	   flag which will constrain all addresses to the lower 32 bits of
	   memory. I've avoided this `easy way out', since there are long
	   term benefits in doing the port _right_.)

	5) MEDM works `as advertized', but without XrtGraph (since I don't 
	   have it). There is a dummy XrtGraph.h file to satisfy the
	   compiler and linker; use in place of the official XrtGraph.h file.
	   With fvwm there is a problem with pop-up menus - the window 
	   pops up but then immediately disappears from view under the main
	   window. This is most likely a problem which isn't unique to Alphas,
	   and could be fvwm related.

	6) Alpha building of default.dctsdr (and the sum file) for downline
	   loading into the IOC are bust.
	   This is because of endian problems & a host of SDR definitions which
	   need to be patched up. My solution is to steal a default.dctsdr
	   file from someone who has already built it on a Sun. Long term
	   solution to this was discussed in Tucson earlier this year.

	7) I have a working version of DCT, but this isn't yet in the kit.
	   (Anyone interested can contact me for it). It needs the
 	   default.dctsdr file which the kit _does_ build (see above).

	8) I'm working on free_gdct. My hangup at the moment is getting 
	   Interviews working with g++ (see various threads on
	   comp.windows.interviews for latest). Perhaps I'll fold & buy
	   DEC's C++ compiler, which is rumoured to be very good & for
	   which the standard interviews.stanford.edu kit is built to use.

Thats about it. Can't be at Dallas meeting, since I'm on-line at SSRL in
early May. Chug a beer there for me, folks!

Tony

--------------------------------------------------------------------------------
Dr Anthony D Cox
Computer Systems Specialist
Stanford Synchrotron Radiation Laboratory
Stanford Linear Accelerator Center
MS 69, Box 4349
Stanford CA 94305
[email protected]
--------------------------------------------------------------------------------

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System