spi/rockchip: Add dt-binding for Rockchip rk3399 spi
diff mbox

Message ID 1455614538-4415-1-git-send-email-jay.xu@rock-chips.com
State New
Headers show

Commit Message

Jianqun Xu Feb. 16, 2016, 9:22 a.m. UTC
From: Xu Jianqun <jay.xu@rock-chips.com>

Add compatible support for Rockchip rk3399 spi.

Signed-off-by: Xu Jianqun <jay.xu@rock-chips.com>
---
 Documentation/devicetree/bindings/spi/spi-rockchip.txt | 1 +
 1 file changed, 1 insertion(+)

Comments

Mark Brown Feb. 16, 2016, 1:22 p.m. UTC | #1
On Tue, Feb 16, 2016 at 05:22:18PM +0800, jianqun.xu wrote:

>  Documentation/devicetree/bindings/spi/spi-rockchip.txt | 1 +
>  1 file changed, 1 insertion(+)

I'd expect this to be added to both the driver and the binding document
not just the document.
Heiko Stuebner Feb. 16, 2016, 2:37 p.m. UTC | #2
Am Dienstag, 16. Februar 2016, 13:22:23 schrieb Mark Brown:
> On Tue, Feb 16, 2016 at 05:22:18PM +0800, jianqun.xu wrote:
> >  Documentation/devicetree/bindings/spi/spi-rockchip.txt | 1 +
> >  1 file changed, 1 insertion(+)
> 
> I'd expect this to be added to both the driver and the binding document
> not just the document.

It seems the driver can still use the rk3066-variant - aka nobody has found 
specific issues in the rk3399 IP implementation.

Having the more specific IP names already in the dts before actually using 
them was suggested by devicetree people to not have to exchange everything 
once specific issues were found. 

So right now it's using the fallback-mechanism of binding to the rk3066 
name, but if some obstacle appears it will only take adding the binding in 
the driver to make it work (even with old devicetrees).
Jianqun Xu Feb. 17, 2016, 12:21 a.m. UTC | #3
Hi Heiko

? 16/02/2016 22:37, Heiko Stuebner ??:
> Am Dienstag, 16. Februar 2016, 13:22:23 schrieb Mark Brown:
>> On Tue, Feb 16, 2016 at 05:22:18PM +0800, jianqun.xu wrote:
>>>   Documentation/devicetree/bindings/spi/spi-rockchip.txt | 1 +
>>>   1 file changed, 1 insertion(+)
>>
>> I'd expect this to be added to both the driver and the binding document
>> not just the document.
>
> It seems the driver can still use the rk3066-variant - aka nobody has found
> specific issues in the rk3399 IP implementation.
>
> Having the more specific IP names already in the dts before actually using
> them was suggested by devicetree people to not have to exchange everything
> once specific issues were found.
>
> So right now it's using the fallback-mechanism of binding to the rk3066
> name, but if some obstacle appears it will only take adding the binding in
> the driver to make it work (even with old devicetrees).
>
>
Thank you, got it, and I will re-send this patch with the driver patch.
>
Heiko Stuebner Feb. 17, 2016, 6:45 a.m. UTC | #4
Hi Jianqun,

Am Mittwoch, 17. Februar 2016, 08:21:58 schrieb Jianqun Xu:
> ? 16/02/2016 22:37, Heiko Stuebner ??:
> > Am Dienstag, 16. Februar 2016, 13:22:23 schrieb Mark Brown:
> >> On Tue, Feb 16, 2016 at 05:22:18PM +0800, jianqun.xu wrote:
> >>>   Documentation/devicetree/bindings/spi/spi-rockchip.txt | 1 +
> >>>   1 file changed, 1 insertion(+)
> >> 
> >> I'd expect this to be added to both the driver and the binding document
> >> not just the document.
> > 
> > It seems the driver can still use the rk3066-variant - aka nobody has
> > found specific issues in the rk3399 IP implementation.
> > 
> > Having the more specific IP names already in the dts before actually
> > using them was suggested by devicetree people to not have to exchange
> > everything once specific issues were found.
> > 
> > So right now it's using the fallback-mechanism of binding to the rk3066
> > name, but if some obstacle appears it will only take adding the binding
> > in the driver to make it work (even with old devicetrees).
> 
> Thank you, got it, and I will re-send this patch with the driver patch.

you have me confused now :-) .

That was meant as a reply to Mark's comment, explaining why the double 
compatible and the driver only binding to the rk3066-one right now.


Heiko
Mark Brown Feb. 17, 2016, 10:53 a.m. UTC | #5
On Wed, Feb 17, 2016 at 07:45:09AM +0100, Heiko Stuebner wrote:

> That was meant as a reply to Mark's comment, explaining why the double 
> compatible and the driver only binding to the rk3066-one right now.

It's still better to have the compatible explicitly listed in the
driver even if it doesn't change the configuration at all at this time.
That way we've at least got a record that the handling is currently
identical to whatever other devices it is the same as.

Patch
diff mbox

diff --git a/Documentation/devicetree/bindings/spi/spi-rockchip.txt b/Documentation/devicetree/bindings/spi/spi-rockchip.txt
index 0c491bd..1b14d69 100644
--- a/Documentation/devicetree/bindings/spi/spi-rockchip.txt
+++ b/Documentation/devicetree/bindings/spi/spi-rockchip.txt
@@ -9,6 +9,7 @@  Required Properties:
     "rockchip,rk3066-spi" for rk3066.
     "rockchip,rk3188-spi", "rockchip,rk3066-spi" for rk3188.
     "rockchip,rk3288-spi", "rockchip,rk3066-spi" for rk3288.
+    "rockchip,rk3399-spi", "rockchip,rk3066-spi" for rk3399.
 - reg: physical base address of the controller and length of memory mapped
        region.
 - interrupts: The interrupt number to the cpu. The interrupt specifier format