mmio_decode()

mmio_decode()

[Jan Beulich]

Could anyone enlighten me why this function decodes arbitrary opcodes?
I.e. 0x20...0x23 are all getting decoded, but only 0x00 (and not
0x01...0x03) and only 0x2B (and not 0x2A). I guess this is based on
observed instructions used for mmio, but - this doesn't make sense! That
way, we'll never get to a point where arbitrary OSes (or other
environments) can get emulated.

(To make things odd, I'm complaining about this while the absence of
a decode for 0x00 actually helped pointing out another problem that
I'm in the process of trying to address: Newer isolinux runs partly in
non-flat protected mode, which apparently isn't generally being
considered a supported environment).

Jan

Re: mmio_decode()

[Keir Fraser]

On 21/11/06 12:42, "Jan Beulich" <jbeulich@novell.com> wrote:

> Could anyone enlighten me why this function decodes arbitrary opcodes?
> I.e. 0x20...0x23 are all getting decoded, but only 0x00 (and not
> 0x01...0x03) and only 0x2B (and not 0x2A). I guess this is based on
> observed instructions used for mmio, but - this doesn't make sense! That
> way, we'll never get to a point where arbitrary OSes (or other
> environments) can get emulated.

Opcodes get faulted in as needed. Pretty stupid I agree.

 -- Keir