| Jacob H. Suter wrote:
| > 
| > I'd assume it'd be kinda easy to
| > copy the contents of the ROM into some memory segment as the memory
| > "managment" was initialized in the Kernel.
| > 
|
|   Ragge has mentioned that, at least on the MicroVAX II, the ROM character
| output routines only work with the MMU turned off.  Someone (Bertram?) talked
| about a hack to disable the MMU, print a character, and turn it back on again.
| I don't know how far that got.
|
| -- 
| David Evans          (NeXTMail/MIME OK)             dfevans@bbcr.uwaterloo.ca

???  Does this mean that a MicroVAX II, running Ultrix or VMS, cannot
operate in text mode on the console after the OS has started?  Seems
crazy, but in all honesty, I've only seen MicroVAXen running in graphics
mode (X11) after bootup (my own have no monitors, serial terminals only).
Surely, *if* the ROM contents were mirrored to RAM, which would seem to be
the way to do it anyway (don't want a sluggish console like Suns have,
yuck!), access to the particular fixed addresses necessary to hit the console
display could be mediated by a device driver, in the usual Unix way?
Perhaps the MMU would have to be turned off once at startup, to read the
ROMs?

What, exactly, do the ROM routines do, anyway?  If one could just access
the raw bitmap of the display, perhaps one could generate ones own
characters without using the ROM routines, mirrored or not?  This type
of access will be needed for the eventual X server anyway...

--ben