On 5 September 2013 19:55, Andreas Färber <afaer...@suse.de> wrote:
> Am 05.09.2013 18:43, schrieb Peter Maydell:
>> /* Indexed by pl110_version */
>> static const unsigned char *idregs[] = {
>> pl110_id,
>> - pl110_versatile_id,
>> + /* The ARM documentation (DDI0224C) says the CLCDC on the Versatile
>> board
>> + * has a different ID (0x93, 0x10, 0x04, 0x00, ...). However the
>> hardware
>> + * itself has the same ID values as a stock PL110, and guests (in
>> + * particular Linux) rely on this. We emulate what the hardware does,
>> + * rather than what the docs claim it ought to do.
>> + */
>> + pl110_id,
>> pl111_id
>> };
>>
>
> I vaguely remember us having a conversation that we might store these in
> the class, but me not wanting to refactor that in my 1.6 candidate
> patchset, right?
I don't feel very strongly about it one way or the other for this
sort of "these two things are almost exactly the same but
not quite" device; we could do it the way we do now, or with
something in the class, I guess. The pl11x is not something
I'm likely to do any new work on and the current code's not
unacceptably ugly so unless you have a strong opinion I'd
just leave it the way it is.
(Possibly I leant the other way last time the issue came up:
if so, that's a measure of my lack-of-strong-opinion :-))
-- PMM
