On 24 August 2013 20:02, Tomasz Figa [off-list ref] wrote:

On Saturday 24 of August 2013 12:05:36 Rahul Sharma wrote:

quoted

quoted

Hi Rahul,

Please see my comments inline.

On Friday 23 of August 2013 15:54:18 Rahul Sharma wrote:

quoted

Add hdmi, mixer, ddc device tree nodes for Exynos 5420 SoC.

Signed-off-by: Rahul Sharma <redacted>
---

 arch/arm/boot/dts/exynos5420-smdk5420.dts |   19 +++++++++++++++++++
 arch/arm/boot/dts/exynos5420.dtsi         |   14 ++++++++++++++
 2 files changed, 33 insertions(+)

diff --git a/arch/arm/boot/dts/exynos5420-smdk5420.dts

b/arch/arm/boot/dts/exynos5420-smdk5420.dts index bafba25..7dbf1fa
100644

--- a/arch/arm/boot/dts/exynos5420-smdk5420.dts
+++ b/arch/arm/boot/dts/exynos5420-smdk5420.dts

@@ -61,4 +61,23 @@

              };

      };

+     hdmi@14530000 {
+             status = "okay";
+             hpd-gpio = <&gpx3 7 0>;
+     };
+
+     mixer@14450000 {
+             status = "okay";
+     };
+
+     i2c_2: i2c@12C80000 {
+             samsung,i2c-sda-delay = <100>;
+             samsung,i2c-max-bus-freq = <66000>;
+             status = "okay";
+
+             hdmiddc@50 {
+                     compatible = "samsung,exynos4210-hdmiddc";
+                     reg = <0x50>;
+             };
+     };

 };

You should separate above board-specific changes from this patch and
squash it with patch 6/7, adjusting patch subject and description
appropriately.

quoted

diff --git a/arch/arm/boot/dts/exynos5420.dtsi

b/arch/arm/boot/dts/exynos5420.dtsi index 05dee86..7c6214d 100644

--- a/arch/arm/boot/dts/exynos5420.dtsi
+++ b/arch/arm/boot/dts/exynos5420.dtsi

@@ -274,4 +274,18 @@

              pinctrl-0 = <&i2c3_bus>;
              status = "disabled";

      };

+
+     hdmi@14530000 {
+             compatible = "samsung,exynos4212-hdmi";
+             reg = <0x14530000 0x70000>;
+             interrupts = <0 95 0>;
+             status = "disabled";
+     };
+
+     mixer@14450000 {
+             compatible = "samsung,exynos5420-mixer";
+             reg = <0x14450000 0x10000>;
+             interrupts = <0 94 0>;
+             status = "disabled";

This node looks like a candidate to be enabled here on SoC level,
since it does not require any board specific information.

Hi Tomasz,

Both hdmi and mixer should be enabled at same time. Mixer don't
but hdmi needs hpd gpio information which is board dependent. Thats
why I kept them disabled in SoC file. What you say?

Hmm, is it a hardware restriction? I mean, if HDMI is not enabled, is it
impossible to enable mixer and use it for some purposes? I don't know much
about the Exynos HDMI subsystem, so sorry for the noise if this is
obvious.

Hi Tomasz,

No, It is not a hardware restriction. Mixer can be probed independently
but there is NO usecase where mixer can be used without hdmi
enabled. Mixer output is hardwired to hdmi block in SoC.

What I have in mind is that device tree should describe the hardware,
regardless of use case. This is, all the hardware that can possibly
operate should be listed as operational.

I understand your point. We can keep mixer enabled in DTSI.

regards,
Rahul Sharma.

Best regards,
Tomasz