Re: Concerns over VAX Port

> However, one point on which I cannot be so satisfied concerns the
> possible fate of the OpenBSD VAX port in general.  I might be going
> slightly paranoid, or somesuch, but I noticed a message recently about
> a particular port which was going to be removed from the supported
> list--I forget which one.

Your fears are unjustified.

OpenBSD ports are maintained and supported, as long as the platform is
considered relevant to the project, and there are enough developers
willing to maintain it. This also implies that developers have a decent
supply of hardware to do so.

The OpenBSD/cats port, which is being phased out, meets none of this
criteria. These boards are pretty rare, their reliability is
questionable (we have had several failures in less than two years for
boards bought new in box, including one who spontaneously combusted),
and their so-called firmware is atrocious; they have been using as a
fundation and starting point for the successfull OpenBSD/zaurus port,
and the new OpenBSD/armish port. The OpenBSD/cats port was never
intended to be maintained on the long term, although it would probably
have been discontinued later if not for the hardware unreliability.

On the other hand, the OpenBSD/vax port has a much better future. VAXens
are more common accross developers, including ``fast'' models (i.e.
VS4k/9x and VAX 4000-10[56]), and developers without VAXens can use SIMH
on a reasonably fast machine[*]. And due to its unique MMU features, it
is a very good platform to exercise all mmap(2)-based developments, such
as OpenBSD's current malloc(3) implementation, which was refined on a
VAXstation 4000/VLC.

So unless there is a huge shortage of developers in the next years, the
vax port will be maintained and improved.

Miod

[*] And they can also use gxemul to emulate a cats board, but right now
gxemul can not emulate an ethernet interface as well as SIMH can, which
makes it a real PITA to do development within the simulator.