Re: [PATCH 2/3] ARM: sunxi: Add an ahci-platform compatible AHCI driver for the Allwinner SUNXi series of SoCs

[Olliver Schinagl <oliver@schinagl.nl>]

Hey all,

On 04-12-13 14:23, Tejun Heo wrote:
> Hello,
>
> (cc'ing Richard and Shawn, hi!)
>
> On Wed, Dec 04, 2013 at 02:16:49PM +0100, Olliver Schinagl wrote:
>> On 04-12-13 14:14, Tejun Heo wrote:
>>> Hello,
>>>
>>> On Wed, Dec 04, 2013 at 01:56:23PM +0100, Oliver Schinagl wrote:
>>>> I took the imx driver as example, as I wasn't sure on where to
>>>> start. But I don't think it's possible yet without improving
>>>> ahci_platform as I suggested in the cover letter. So if
>>>> ahci_platform needs to be improved, I guess a separate patch series
>>>> would be more appropriate?
>>>>
>>>> So would it be acceptable to have this as the 2nd (and last?)
>>>> ahci_platform driver and go from there? Or do you want to block new
>>>> ahci_XXX drivers until ahci_platform has been improved?
>>> I don't want to block new drivers unconditionally but at least I want
>>> to know which direction we're headed in the longer term.  Right now it
>>> feels like we could be at the beginning of an uncoordinated explosion
>>> of these drivers which will take a hell lot mpore effort to clean up
>>> after the fact.  I could be wrong and these could actually be
>>> different enough to justify separate drivers and there isn't gonna be
>>> an avalanche of these but again I at least want to know the general
>>> direction things are headed before making any decisions.
>> I'd be happy to pour it in any form that's needed. I even do the
>> modification/rewrite of ahci_platform if I get enough help as it
>> might be a little over my head initially ;)
>>
>> That said, I don't think it's much different at all and I do think
>> it could be much simpler. In my mind, the sunxi_ahci driver wouldn't
>> need to be much bigger then a few lines that are specific to the SoC
>> (hardware init) and registerd to the ahci_platform framework via
>> platform_ahci_register() instead of platform_device_register().
>>
>> But again, point me (for dummies ;) in the right direction and I'll
>> work on it with some help.
> Richard and Shawn recently worked on ahci_imx.  Can you guys please
> talk with each other and figure out what can be done to share as much
> as possible among these new platform-specific drivers?  I'd really
> like to see the common things factored out as much as possible with
> only the actual hardware differences described for each device.
Working on this and studying the existing ahci_platform/shci_platform 
drivers the last few days and was figuring out why ahci_platform only 
supports 1 clock. IMX handles this by having 3 clocks defined in the DT, 
the first one gets enabled by default via ahci_platform, the other 2 get 
enabled in IMX's probe function.

Is it an idea to extend this to support all clocks that would be 
required (via a callback)? Or do we prefer having the clocks separated 
for other technical reasons? Or do we want to handle the clocks via the 
ahci_platform framework and extend hpriv->clk to an array of clocks?

Oliver

>
> Thanks a lot!
>