Re: [MiNT] XaAES / GEM memory issues

[Sven Karlsson <svenka@it.kth.se>]

Frank,

> Oh, just a question: Why not move all hardware stuff into the AES? This is
> the fastest way at all! Why not move all hardware stuff into the
> applications? Then they are fast as demos ...

Please, I never said that and you know it.

> Btw. to follow your way a VDI/AES is strictly required to made a runnable
> system. Also you have crossdependencies as you can't get the underlying
> GEMDOS running without it. Such crossdependencies are really bad and
> always an indicator for a wrong design.

There is no cross dependence You need one central mouse/keyboard event
handler. That is all. In fact, if you draw a process graph you see that
it is strictly acyclic.

> Btw. you can simply avoid the context switches if you integerate the AES
> into the kern itself (or at least the syscall handler and the event
> manager that dispatch between the AES tasks and the AES server).

Now that was what I wrote in the last mail, wasn't it? And if you do
that (which you get if you implement parts of the AES as a device
driver) you would never eliminate all context switches as you need to
switch to the end AES application.

Most modern OSes such as NeXTSTEP, Plan9, BeOS and even NT pay special
attention to how the mouse and keyboard is handled and associates it
with the windowmanager.

Best regards
 Sven