On 10/26/07, Enache Adrian <[EMAIL PROTECTED]> wrote:
> > - boots install42.iso
> > without kqemu, i can do a full install. not as fast as real metal but
> > acceptably quick. after install, openssl speed runs 12x slower than on
> > real metal. with kqemu, kernel crashes right after "root on rd0a swap
> > on rd0b dump on rd0b"
> > ====================================
> > uvm_fault(0xd069f620, 0x0, 0, 3) -> e
>
> While I think I have fixed that particular problem (it was a mistake
> I've taken over from kqemu-win32.c), neither openbsd nor netbsd will
> work inside kqemu (should it be with linux, windows or openbsd as a host
> - with full virtualization they crash, with user mode virtualization
> they hang).

What are you mean by "user mode virtualization" - just the pure
software emulation?... oh, nevermind... "-kernel-kqemu", "" and
"-no-kqemu" do different things.

kqemu causes win98se to fails with a "windows protection error", full
software emulation still works.

winxp pro works with kernel-kqemu

> I do hope to come with a solution for that (at least for the second case;
> running OpenBSD with '-kernel-kqemu' may need modifications to the OpenBSD
> kernel itself - or maybe not).

my knee-jerk reaction is that kqemu should probably be fixed if it's
presenting an incorrect virtual machine.

CK

-- 
GDB has a 'break' feature; why doesn't it have 'fix' too?

Reply via email to