On 17/12/2014 22:10, Laszlo Ersek wrote:
> Addressing review comments received for v4 (with many thanks).
> 
> Deeper changes in v5 include (see also the notes per patch):
> 
> - The I/O port mapping and the MMIO mapping have been split into
>   separate QOM subclasses. This is new territory for me (what is not,
>   heh), so bear with me.
> 
>   I've done this because a wide data I/O port didn't seem overly useful,
>   but marrying it with the overlapped selector/data case was certainly a
>   huge mess. This is just so much cleaner. (Assuming I got it right.)

Pretty much.

> - The memory mapped selector and data registers are explicitly big
>   endian now.
> 
> - The memory mapped data register performs its own value
>   (de)composition, the memory subsystem is left out from that role.

Patches 1-7

Reviewed-by: Paolo Bonzini <pbonz...@redhat.com>

Paolo

Reply via email to