[PATCH 1/4] [POWERPC] cpm2: Updates for CPM2 pic

Kumar Gala galak at kernel.crashing.org
Wed Dec 13 19:47:36 EST 2006


On Dec 13, 2006, at 2:35 AM, Vitaly Bordug wrote:

> On Tue, 12 Dec 2006 13:58:44 -0600
> Kumar Gala <galak at kernel.crashing.org> wrote:
>
>>
>> On Dec 12, 2006, at 12:55 PM, Vitaly Bordug wrote:
>>
>>> On Mon, 11 Dec 2006 20:21:52 -0600
>>> Kumar Gala <galak at kernel.crashing.org> wrote:
>>>
>>>>
>>>> On Dec 11, 2006, at 4:36 PM, Vitaly Bordug wrote:
>>>>
>>>>>
>>>>> This contains important fixes for the CPM2 PIC code. Eliminated
>>>>> CPM_IRQ_OFFSET, pulling the respective interrupt numbers from the
>>>>> interrupt
>>>>> mapping. Updated devicetree files to reflect that. Changed direct
>>>>> IC-related IO accesses to the IO accessors.
>>>>
>>>> Since your changing the interrupt numbers, do you mind changing the
>>>> interrupt sense to match ipic?  I was hoping to introduce a generic
>>>> host_xlate function to cover the various embedded pic's.
>>> Umm, can you pls elaborate what do you want to achieve?
>>
>> Its a minor thing, but the values we picked for ipic to encode sense/
>> polarity match the linux definitions.  The idea is that  if we use
>> the same encoding on multiple pic's we can sure the host_xlate code.
>>
> I started it but faced with the problem:
>
> in 8560, there are both mpic and cpm2 pic. To keep consistency (so  
> that cpm irqs encode sence one way, and generic - mpic -  in  
> another), we'll have to modify mpic too, and all the 85xx dts's as  
> well.

Don't worry about mpic, just the cpm2.  The idea was for pic's that  
didn't of existing OF specs like mpic/openpic/i8259, etc. to use the  
same encoding.  Consider mpic grandfathered in and don't worry about it.

- k

> Yet doable, it's pretty big change, and I think it should be put to  
> the patch, independent of small fixes I have introduced, and that  
> should come separately (or else we'll have hard time git-bisecting  
> for troubles).
>
> If you envision this way as a proper one for mpic too, I'll head  
> for it though.
>
> -- 
> Sincerely,
> Vitaly




More information about the Linuxppc-dev mailing list