[PATCH 6/8] arm: mach-armada: add support for Armada XP board with device tree

[robherring2@gmail.com (Rob Herring)]

On 05/15/2012 10:46 AM, Thomas Petazzoni wrote:
> Le Tue, 15 May 2012 15:32:20 +0000,
> Arnd Bergmann <arnd@arndb.de> a ?crit :
> 
>>>> Agreed. When looking at the code, I wasn't sure why PCI mappings
>>>> were needed so early. Surely, the minimal mappings to get an UART
>>>> are quite useful to have very early, but why PCI?
>>>
>>> I think it was because that PCI was initialised early and requires a
>>> large mapping space. It is very likely to be legacy kernel stuff
>>> that people have kept copying through.
>>
>> No, as Rob pointed out, we might not need them early, but we
>> definitely want a fixed address for the PIO window, which ioremap
>> does not provide.
> 
> A fixed address for the PIO window of the PCI registers? Or a fixed
> address for the PIO window used to access UART registers?
> 
> I would understand why a fixed address is needed for the latter, but
> not really for the former.
> 

I'm not sure which PCI registers you mean, but we're talking about PCI
I/O space. The i/o space should be a fixed virtual address so that the
__io() is a constant base for ALL arm platforms. This isn't in place
yet, but that is the plan. Although current macros are constant base.

Rob