> Okey :) But I thought about REAL swapping (page based) because this kind
> of swapping require to swap out the WHOLE program. On heavy memory load
> this would be terrible (swapping in/out a whole program almost on any
> context switching ?)
Only if you have more programs actively running thann fit, as opposed to
programs asleep. It is true that swap based systems collapse more spectacularly
than page based ones but that can be offset by scheduler sensibility.
> 286. (I might have posted it too the list some months ago I don't remember
> exactly). Set segment size to for example 1. If the code (or data)
> try to read/write offset greater then this we got an exception.
> But I don't know how we can get the offset (because we must to load
> from hdd data between the setted segment size and the offset casued the
> exception). Maybe we must write an instruction decoder or seomthing ?!
> It looks bad. But it can be considered ...
Xenix 286 did some sort of paging, I think the compiler but big arrays,
each object file and other related stuff in their own segment- so it can be
done and it is truely horrible.
