Re: [PATCH] Input: atmel_mxt_ts: Add of node type to the i2c table

[Javier Martinez Canillas <javier.martinez@collabora.co.uk>]

Hello Nick,

On 09/11/2014 11:19 AM, Nick Dyer wrote:
> 
> Thanks for the clear explanation.
> 
> The i2c aliases are a bit confusing. The original device the driver was
> written for was called qt602240, which was renamed by Atmel to mXT224 when
> the chip series was called "maXTouch". The driver now actually supports
> many other chips which aren't listed (more than 20 devices that I've
> personally tested). I could add them all, but it would be an extremely long
> list. It may be preferable to use the generic name maXTouch.
> 
> So I think the sensible thing to do here would be to add "maxtouch" to the
> i2c list to fix the module autoload issue.
> 

While this will actually fix the module auto-load issue on the atmel driver,
I'm concerned that the I2C core is not reporting the correct module
'of:N*T*Catmel,maxtouch*' alias when probing using DT.

Since as Lee said on his cover letter for the mentioned series [0], an I2C ID
table shouldn't be mandatory for drivers that only support DT based platforms
(e.g: a driver that depends on OF) but in that case I2C module auto-loading
would not work AFAICT.

Best regards,
Javier

[0]: https://lkml.org/lkml/2014/6/20/199