At 9:19 AM -0800 4/30/98, Jonathan S. Shapiro wrote:
The KeyKOS system for the Omron 88000 had a special range which allocated the pages of the frame buffer. Otherwise they checkpointed normally. It was quite impressive to see the screen quickly come back upon restart. It was so impressive that people tended not to notice that is was another 10-15 seconds before it reacted to user input.
> [Reactions from the KeyKOS crowd? Is this necessary? Is it okay
> for the page fault handler to convert to DK(0) and yet special case
> this? One does not seem intrinsically more offensive than another
> to me.]
Since the allocator is in a position to know what kind of page these pages were, there is no information hiding principle that would make it undesirable to tell the allocator directly. The problem comes when you want to virtualize these pages and use normal pages to replace them. You need an architecture which will allow you the rescind the normal page, but give the expected error code. You might be able to use single page kept segments to achieve this.
Bill Frantz | If hate must be my prison | Periwinkle -- Consulting (408)356-8506 | lock, then love must be | 16345 Englewood Ave. frantz@netcom.com | the key. - Phil Ochs | Los Gatos, CA 95032, USA