Re: [PATCH 1/2] ARM: shmobile: R8A7779: add Ether support

[Simon Horman <horms@verge.net.au>]

On Thu, Apr 04, 2013 at 03:34:10PM +0900, Simon Horman wrote:
> On Wed, Apr 03, 2013 at 05:22:22PM +0400, Sergei Shtylyov wrote:
> > Hello.
> > 
> > On 03-04-2013 4:08, Kuninori Morimoto wrote:
> > 
> > >>>Then, this (void *pdata) should be
> > >>>(struct sh_eth_plat_data *pdata) IMO
> > 
> > >>     ether_device.dev.platform_data is 'void *'. I didn't want to bring in
> > >>extra header for the little use.
> > 
> > >Not enough reason for me.
> > 
> > >"void *" means there is no pointer check,
> > >and extra header is just 1 line. No ?
> > 
> >    There's no pointer check either if we just initialize the 'platform_data'
> > member as part of the platfrom device initializer, so we can
> > actually stuff pointer to any nonsense there. Why make this case
> > different?
> > 
> > >If you want to use this style,
> > >then, additional extra header is fate, IMO
> > 
> >    We can agree to disagree here. :-)
> 
> I would like void * changed to struct sh_eth_plat_data * so
> that callers of r8a7779_add_ether_device() will get a warning
> if they pass an argument of the wrong type.
> 
> Other than that I believe that I am happy with this patch.
>
> Ditto for the similar patch for the R8A7778.

As mentioned in a subsequent email in another subthread,
I would also like you to update the patches to not use
platform_register_device().