Re: [PATCH 5/9] ARM: dts: Add GPMC node for OMAP2, OMAP4 and OMAP5

[Benoit Cousson <b-cousson@ti.com>]

On 03/14/2013 05:00 PM, Jon Hunter wrote:
> 
> 
> On 03/14/2013 10:58 AM, Benoit Cousson wrote:
>> On 03/14/2013 04:50 PM, Jon Hunter wrote:
>>>
>>> On 03/14/2013 10:45 AM, Benoit Cousson wrote:
>>>> On 03/11/2013 06:56 PM, Jon Hunter wrote:
>>>>>
>>>>> On 03/09/2013 06:42 AM, Ezequiel Garcia wrote:
>>>>>> On Fri, Mar 8, 2013 at 10:25 PM, Javier Martinez Canillas
>>>>>> <javier@dowhile0.org> wrote:
>>>>>>> On Fri, Mar 8, 2013 at 10:41 PM, Jon Hunter <jon-hunter@ti.com> wrote:
>>>>>>>>
>>>>>>>> Yes you are correct. In general, I have been trying to stay some-what
>>>>>>>> consistent with what hwmod was doing as this was being auto-generated by
>>>>>>>> some hardware design specs and I believe they wanted to eventually get
>>>>>>>> to the point where DT files would be auto-generated too for OMAP.
>>>>>>>> Furthermore my understanding is that the smallest page that can be
>>>>>>>> mapped by the kernel for ARM is 4kB. So if you declare it as 0x2d0 or
>>>>>>>> 0x1000 it will map a 4kB page (I could be wrong here).
>>>>>>>>
>>>>>>>> I don't have any strong feelings here but will do what the consensus
>>>>>>>> prefers.
>>>>>>>>
>>>>>>>
>>>>>>> Yes, you are right here.
>>>>>>>
>>>>>>> I forget that ioremap() does a page-aligned mapping and since the
>>>>>>> minimum page size for ARM is 4KB as you said, there is no difference
>>>>>>> between using 0x2d0 and 0x1000. Sorry for the noise.
>>>>>>>
>>>>>>
>>>>>> Certainly, I don't have strong feelings about this.
>>>>>> FWIW, mvebu maintainers imposes a "minimal" address space request
>>>>>> policy.
>>>>>>
>>>>>> On the other side, it seems to me we shouldn't look at internal kernel
>>>>>> implementation (i.e. ioremap page-alignment) to make this decision.
>>>>>
>>>>> I agree with that. I am not sure if Tony/Benoit have any comments on
>>>>> what they would like to do here to be consistent for the omap bindings.
>>>>
>>>> Yes, I full agree with that as well. The size should be purely HW
>>>> related. So we should not take any assumption about the page size /
>>>> alignment.
>>>
>>> Ok, what is best to use? The size from hwmod structures or the size from
>>> the documentation?
>>
>> Well, in theory both are supposed to be identical :-)
>> I'm just applying a rounding to the closet power of two, that's why it
>> cannot be 0x2d0.
> 
> Ok I understand. However, still not clear what you want me to use :-(

That's on purpose :-)

Take 0x2d0, we could always remove the rounding in the generation part
to stick to the HW documentation.

Regards
Benoit