[RFC] i2c: Providing hooks for i2c multimaster bus arbitration.

[Simon Glass]

Hi,

On Tue, Feb 5, 2013 at 9:33 PM, Yuvaraj Kumar <yuvaraj.cd at gmail.com> wrote:
> On Wed, Feb 6, 2013 at 10:58 AM, Yuvaraj Kumar <yuvaraj.cd at gmail.com> wrote:
>> On Wed, Feb 6, 2013 at 5:48 AM, Doug Anderson <dianders at chromium.org> wrote:
>>> Yavaraj,
>>>
>>> On Mon, Feb 4, 2013 at 1:03 AM, Yuvaraj Kumar C D <yuvaraj.cd at gmail.com>
>>> wrote:
>>>>
>>>> This RFC patch is w.r.t multimaster bus arbitration which is already
>>>> being discussing in the mainline.
>>>> This patch provides hooks for the i2c multimaster bus arbitration
>>>> and to have the arbitration parameters.
>>>
>>>
>>> I may have missed something in all the threads, but my recollection was that
>>> the request was that this should be implemented without touching the i2c
>>> core code.  Your patch modifies the i2c core code.
>
>    Different i2c devices will directly talk to i2c core using i2c_transfer().
>     If we put this out of the i2c core,the i2c devices like tpsxxxx and etc
>     should be registered through the bridge/mux .
>
>>>
>>> My impression was that the best solution was to use the infrastructure in
>>> place for i2c multiplexing.  ...but in your case you would only multiplex
>>> one thing.  This was suggested by Grant Likely and seems the cleanest...
>>
>      I agree that Grants idea was the cleanest but we may end up in change in
>       i2c device probe to register through mux/bridge.(for the
> device's connected on that
>      particular bus which requires arbitration).

Does the device tree not handle that case automatically? How are
muxes/bridges done in that case? It's not problem to adjust the device
tree file accordingly if Grant's approach is the cleanest method.

Regards,
Simon

>>
>>>
>>> -Doug