Re: [PATCH 2/7] ARM: i.MX53: Add sata to devicetree

Alexander Shiyan shc_work at mail.ru
Mon Feb 10 04:22:43 EST 2014


Понедельник, 10 февраля 2014, 10:15 +01:00 от Sascha Hauer <s.hauer at pengutronix.de>:
> On Mon, Feb 10, 2014 at 10:07:35AM +0100, Sascha Hauer wrote:
> > On Sat, Feb 08, 2014 at 06:26:41PM +0400, Alexander Shiyan wrote:
> > > Суббота,  8 февраля 2014, 15:16 +01:00 от Eric Bénard <eric at eukrea.com>:
> > > > Signed-off-by: Eric Bénard <eric at eukrea.com>
> > > > ---
> > > > arch/arm/dts/imx53.dtsi | 9 +++++++++
> > > > 1 file changed, 9 insertions(+)
> > > > 
> > > > diff --git a/arch/arm/dts/imx53.dtsi b/arch/arm/dts/imx53.dtsi
> > > > index 678547c..91659b8 100644
> > > > --- a/arch/arm/dts/imx53.dtsi
> > > > +++ b/arch/arm/dts/imx53.dtsi
> > > > @@ -75,6 +75,15 @@
> > > > interrupt-parent = <&tzic>;
> > > > ranges;
> > > > 
> > > > +		sata: sata at 10000000 {
> > > > +			compatible = "fsl,imx53-ahci";
> > > > +			reg = <0x10000000 0x1000>;
> > > > +			interrupts = <28>;
> > > > +			clocks = <&clks 173>, <&clks 5>, <&clks 188>;
> > > > +			clock-names = "sata", "ahb", "per";
> > > > +			status = "disabled";
> > > > +		};
> > > 
> > > Sascha, we stop changing DTS of the kernel or not?
> > > What is the ultimate solution?
> > 
> > I don't want to give out the rule "Only use unmodified dts in barebox"
> > and then use dts files from Linux-next. If we only want to use
> > unmodified dts files we have to use the ones from Linux mainline.
> > Otherwise we are doomed when a pull request doesn't make it in the
> > Kernel and then the dts files are changed significantly before the
> > next merge window, as just happens with Shawns dts changes for 3.14.
> 
> So what I suggest for now is:
> 
> - I welcome all patches moving the clearly barebox specific stuff to
> separate files, like Sebastian already started.
> - Use plain Linux Kernel dts files where possible. I will try and deny
> all changes to these files except "sync to Linux-x.y"
> - Leave the i.MX dts files how they are until the changes currently
> queued in Linux-next hit mainline.

Nice. 
Something I already updated back into the kernel (imx27-phycard series).

---


More information about the barebox mailing list