On 06/28/2015 03:38 AM, Florian Fainelli wrote:

Le 06/27/15 15:08, Rafa? Mi?ecki a ?crit :

quoted

On 24 June 2015 at 01:51, Florian Fainelli [off-list ref] wrote:

quoted

Enable the use of UART0 by overriding its default status property.

Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
---
 arch/arm/boot/dts/bcm4708-smartrg-sr400ac.dts | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/arch/arm/boot/dts/bcm4708-smartrg-sr400ac.dts b/arch/arm/boot/dts/bcm4708-smartrg-sr400ac.dts
index d6a033b97c70..64a5e8ab65e0 100644
--- a/arch/arm/boot/dts/bcm4708-smartrg-sr400ac.dts
+++ b/arch/arm/boot/dts/bcm4708-smartrg-sr400ac.dts

@@ -118,3 +118,7 @@
                };
        };
 };
+
+&uart0 {
+       status = "okay";
+};

We have many more changes like this in OpenWrt, I just didn't upstream
them because of current chipcommonA state.

It was added before we got "brcm,bus-axi" and I believe Hauke wanted
move it to the "correct" place at some point. Since UART is part of
ChipCommon device and ChipCommon is part of "brcm,bus-axi",.I guess we
should add UARTs as a ChipCommon device subnodes. We already have
chipcommon: chipcommon at 0 {
        reg = <0x00000000 0x1000>;

        gpio-controller;
        #gpio-cells = <2>;
};
, is it possible to move UARTs there?

I added it at the beginning of the port so I have serial, even without
adding support for bcma. I haven't look at that part later, but it
should go into the chipcommon part. Thats for the patches.

quoted

I'm not sure if this UART cleanup should block your change. I guess it
depends on the way it'll finally look like.

I do not think it will, see below:

quoted

I'm also wondering: is there any preference between overwriting status
with something like
&uart0 { };
and
chipcommonA {
    uart0: serial at 0300 { };
};
? I don't know DT that well, just asking.

An alias can be located pretty much anywhere in the DTS as long as the
name is unique, which is why I took this approach, because I do not have
to do something like this and need to know the full depth of the tree:

foo0 {
	bar0 {
		baz0 {
			status = "okay";
		};
	};
};

This is both more compact, and more robust to re-parenting the UART0
node in case you ever wanted to do it in the future.

I will happily fix the other DTSes to use that convention if we agree
this is the route to take.

I haven't used that earlier because I did not know of this feature. I
think Rafa? already did or is currently doing this steps.

Hauke