> Object frames cannot be paged out by the kernel, because the
> kernel does not implement persistence.
KeyKOS used the same mechanism for persistance as it did to support
objects (etc.) larger than the real memory of the machine.
Yes. Since NewSys will not implement persistence, this is not an option.
If the kernel can "call" the memory object manager, it could ask for the optional storage map. Then the kernel could throw out the information whenever it seemed reasonable and reconstruct it by asking the memory object manager again. Or am I missing something?
This is vaguely what I have in mind. There remains the challenge of designing a clean protocol for doing this... More on this in a moment.
Jonathan