Ulrich Weigand wrote:
>
> Kevin Lawton wrote:
>
> > OK, I resync'd with Ulrich's changes to the CVS tree.
> > Here's my diffs to the current CVS tree to add my
> > newest changes.
>
> OK, I've added this patch to CVS.
>
> > The 'paging' guest now works!
>
> Unfortunately, I can't confirm this; I keep getting aborts
> due to monPageFault: MapLinPPageOOB and the like ...
>
> Furthermore, the exact symptoms appear to depend on the
> Linux kernel version. I'll be looking into this problem.
Oops, I tested that version only on Linux=2.0.36,gcc=2.7.2.3,
where it seems to work fine for all 3 guests.
On a Linux=2.2.15,gcc=2.95.3 system, I get weirdness.
First I had to implement emulation of another instruction.
This happens depending on the version of the compile generating
code for any of the guests. But after that, the preemptive
kernel works fine, the cooperative one only displays one
running thread (not very cooperative), and the paging one
just sits there. It doesn't segfault for me, and I can quit
it, so it's not hung.
I wonder if these are weird side effects of holding up
the host OS for so long. I'm working on adding support
to make the monitor interruptible. We need that anyways,
and I'd like to see what effect is has.
Otherwise, I'm not sure what the bug is yet. It would figure
mine wouldn't segfault.
-Kevin
