So, patch time again. With this patch we are able to use
easy-to-assemble kernels from OSKit. So we have all the
power from OSKit to write our test kernels, and I think
that's quite comfortable. Also Fiasco does a bit more
before it stops...
To get this working I had to change something with segment
virtualization and to add ljmp virtualization.
There was a little discussion about it some weeks ago. I see
my solution not as the best and I'm sure we have to change
it, especially when our page tables work fine, but it works
for now and we can do further experiments.
(Here comes one point from yesterday again: how to detect
the cpl of the protection faulting app?)
jens
patch_fiasco2.diff
