> On Mon, Jun 17, 2013 at 9:42 AM, Bryan Wu <cooloney@xxxxxxxxx> wrote:
> > On Fri, Jun 14, 2013 at 8:58 PM, Alexander Shiyan <shc_work@xxxxxxx> wrote:
> >> Patch adds a stubs for of_get_next_child and of_get_next_available_child
> >> non-OF builds.
> >>
> >
> > So how other drivers deal with this non-OF builds if they're using
> > of_get_next_child()?
> >
> > And this patch is just for Device Tree core stuff, please resend it to
> > Device Tree mail list.
> >
> > Thanks,
> > -Bryan
> >
>
> I guess you probably need to fix it in the driver firstly like others doing:
>
> #ifdef CONFIG_OF
> /* DT related probing code, like of_get_next_child() */
> ...
> #else
> /* non-DT probing code without of_get_next_child() */
> #endif
This is incorrect since I can have 2 board version in the kernel.
DT and non-DT, same as I doing now for testing. So we should use proper variant at runtime.
> After we fix this, we can try to push your OF core patch to Device Tree.
I'll do it.
---
��.n��������+%������w��{.n�����{��W����ܨ}���Ơz�j:+v�����w����ޙ��&�)ߡ�a����z�ޗ���ݢj��w�f
