> I must admit I do not know what all that values mean, but the book
> "Designing cards and drivers for Macintosh ..." discusses for some
> length the video controller. Although the internal video may be
> different from NuBus card video, I hope the logic is close. If you do
> not have access to this book, I have copied few pages from it.

I own Cards 'n' Drivers. :) It's kind of funny, actually, I'm writing a
video driver as though I built the internal video. I had to consult it
many times so that I could meet your grfioctl interface requirements :)

> As far as I can tell, videorowbytes is the offset to the next row, not
> necessarily screen size. Some portion of video buffer is not mapped to
> screen. There is also an offset to the next page hidden somewhere in
> data structures.

I've worked with some low level frame buffers before. It seems like the Mac
puts it's other pages *horizontally* on the same line. The buffers I worked
with put each page *following* each other. Interesting.

It looks like I can use the value I calculated for fboffset instead of
rowbytes.

Now things make some sense. I find this very interesting and I wonder if
there is a hardware mode that supports pointing the controller at the next
page offset. Hmmm....more research!

The data you gave was most helpful.

Thanks to everyone who answered this!