On Thu, 7 Nov 2013 06:57:00 Benjamin Herrenschmidt wrote:
> On Wed, 2013-11-06 at 18:39 +1100, Alistair Popple wrote:
[snip]
>
> I would go even further and add the 44x workarounds to the normal
> platform device, with a compatible check in there. That isn't the
> first time we add quirks to an
On Wed, 6 Nov 2013 11:14:44 Alan Stern wrote:
> On Wed, 6 Nov 2013, Alistair Popple wrote:
[snip]
> > + /* Initialise platform data from device tree if available. */
> > + if (!dn) {
>
> Shouldn't this be "if (dn)"?
Yep. Thanks.
> > + if (of_get_property(dn, "big-endian", NULL))
On Wed, 2013-11-06 at 18:39 +1100, Alistair Popple wrote:
> diff --git a/arch/powerpc/boot/dts/sequoia.dts
> b/arch/powerpc/boot/dts/sequoia.dts
> index b1d3292..e28371e 100644
> --- a/arch/powerpc/boot/dts/sequoia.dts
> +++ b/arch/powerpc/boot/dts/sequoia.dts
> @@ -153,7 +153,7 @@
>
On Wed, 6 Nov 2013, Alistair Popple wrote:
> Currently the ppc-of driver uses the compatibility string
> "usb-ehci". This means platforms that use device-tree and implement an
> EHCI compatible interface have to either use the ppc-of driver or add
> a compatible line to the ehci-platform driver. I
Currently the ppc-of driver uses the compatibility string
"usb-ehci". This means platforms that use device-tree and implement an
EHCI compatible interface have to either use the ppc-of driver or add
a compatible line to the ehci-platform driver. It would be more
appropriate for the platform driver
