tcs@usna.UUCP (Terry Slattery <tcs@usna>) (02/11/85)
> /**** waltz:net.lan / sabre!martin / 5:24 pm Dec 26, 1984 ****/ > /*vaxbusters*/ > > I'm looking for people who have tried to bring up the Excelan Unibus > board under Vax 5.0, i about to try but i can't quite see what's the > best machine model to start from. > > I'm also looking for poeple who have tried to use the boards with a > standalone system, ie without unix. > > martin levy. > bellcore. > /* ---------- */ There probably isn't a "best model" for a vax because none of the machines that they have ported the stuff to have anything like a unibus map or extended address bits (b_xmem in buf.h). You'll have to pick a model that has the correct byte swapping for a Vax and then look through the code for places where it wants to look at the memory and bracket them with mapalloc()'s to map real memory into the unibus. You may also have to diddle with their exvtop() macro and all the calls to it to allow the use of b_xmem. I have just gotten their stuff running on 2.9 (only test routines though). It was a bitch because of the pdp11 architecture (read: limitations) which require mapped out disk buffers and the use of b_xmem. You might call them and find out if someone else has done this work and will release the fixes. I can't comment on how their stuff runs in production; maybe in another week or so. I have some questions about some of the things in the code that I have asked Excelan about but havn't gotten an answer yet (I only asked last week so I'm expecting an answer soon.) I can't help with the standalone use. -tcs Terry Slattery U.S. Naval Academy 301-267-4413 ARPA: tcs@brl-bmd UUCP: decvax!brl-bmd!usna!tcs