Hi Aitor,
Aitor Santamaría Merino escreveu:
This gets complicated when Kernel is 386 optimized. MS-DOS never
modified this because it is 16-bit. It could be argued that if a
kernel is to be compatible it cannot modify them either.
Now the other way round: can the kernel be sensistive to 386 registers
being modified by drivers? IMHO not, because the original one is not
and hence some drivers will stop working.
Then I believe that it is the kernel resposibility to preserve it's
registers.
Well, I disagree. As it was posted, drivers are to preserve registers.
This is the problem: Old programs assume that the kernel is 16bit. As
such no 386 register will be ever modified because this is plainly
impossible. So if a kernel is to be *compatible* it should not do
something that the original does not.
Notice further that modern 31 but programs that use protected mode
managers like WDOSX are quite imune to that. *but* not long ago I had a
very big head-ache (not completely solved) because I instaled a TSR in
INT-28h that did that and a lot of programs crashed. Most notably Borland's.
So I am *sure* that if any 386 register is modified by the kernel a lot
of programs will not work.
BUT, Eric guarantied me that this does not happen, everyting is properly
saved/restored at the entrypoints. I hope this is 100% true.
Anyway, can't this be the cause why EMM386 + 386-optimised kernel are
not working together?
According do Eric, yes. He convinced me, but I am not sure that Michael
is so sure of it...
Alain
-------------------------------------------------------
SF.Net email is sponsored by: Discover Easy Linux Migration Strategies
from IBM. Find simple to follow Roadmaps, straightforward articles,
informative Webcasts and more! Get everything you need to get up to
speed, fast. http://ads.osdn.com/?ad_id=7477&alloc_id=16492&op=click
_______________________________________________
Freedos-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/freedos-devel
