Subject: None
To: None <port-pmax@NetBSD.ORG>
From: Jonathan Stone <jonathan@DSG.Stanford.EDU>
List: port-pmax
Date: 02/24/1997 16:56:11
>To: port-pmax@NetBSD.ORG
>Subject: Graphic Adapter
>Sender: port-pmax-owner@NetBSD.ORG
>Precedence: list
>Delivered-To: port-pmax@NetBSD.ORG
>Hardware:  DECstation 5000/200

There are two problems here.  The first is that NetBSD/pmax doesn't
support the PMAG-C, PMAG-D, or PMAG-F i860-based accelerated
framebuffers -- not up to 1.2C (today's current), inclusive.

The second is 1.0A is a truly ancient kernel. I'm no longer sure what
what 1.0A is doing internally after it prints the CPU description
information.  It's quite possible that the kernel is enabling
interrupts, and is getting vertical-retrace interrupts from the
accelerated framebuffer. 

The OS doesn't know how to handle the framebuffer, so the interrupt
handler ignores it, returns from the interrupt hanlder while the
interrupt is still asserted.  The interrupt is taken again, and

You can find out what the framebuffer is by typing
	>> conf

at the PROM prompt.  If you have a PMAG-C, PMAG-D, PMAG-E, or PMAG-F,
then NetBSD won't support it.  It's  possible that upgrading to 1.2 or
to -current will give you a kernel that works with a serial console.

The problem with these graphics adaptors is that it's impossible to
get documentation on them.  There may be some useful information in
the Mach microkernel code (Mach mk83 or thereabouts) but the copyright
status of that code is unclear....