diff mbox

[v6,03/10] Documentation: devicetree: Update Samsung FIMC DT binding

Message ID 1394122819-9582-4-git-send-email-s.nawrocki@samsung.com (mailing list archive)
State New, archived
Headers show

Commit Message

This patch documents following updates of the Exynos4 SoC camera subsystem
devicetree binding:

 - addition of #clock-cells property to 'camera' node - the #clock-cells
   property is needed when the sensor sub-devices use clock provided by
   the camera host interface,
 - addition of an optional clock-output-names property,
 - change of the clock-frequency at image sensor node from mandatory to
   an optional property - the sensor devices can now control their clock
   themselves and there should be no need to require this property by the
   camera host device binding, a default frequency value can often be used.

Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
Acked-by: Kyungmin Park <kyungmin.park@samsung.com>
---
Changes since v5:
 - none.

Changes since v4:
 - dropped a requirement of specific order of values in clocks/
   clock-names properties (Mark) and reference to clock-names in
   clock-output-names property description (Mark).
---
 .../devicetree/bindings/media/samsung-fimc.txt     |   34 +++++++++++++++-----
 1 file changed, 26 insertions(+), 8 deletions(-)

Comments

Laurent Pinchart March 11, 2014, 12:30 p.m. UTC | #1
Hi Sylwester,

Thank you for the patch.

On Thursday 06 March 2014 17:20:12 Sylwester Nawrocki wrote:
> This patch documents following updates of the Exynos4 SoC camera subsystem
> devicetree binding:
> 
>  - addition of #clock-cells property to 'camera' node - the #clock-cells
>    property is needed when the sensor sub-devices use clock provided by
>    the camera host interface,
>  - addition of an optional clock-output-names property,
>  - change of the clock-frequency at image sensor node from mandatory to
>    an optional property - the sensor devices can now control their clock
>    themselves and there should be no need to require this property by the
>    camera host device binding, a default frequency value can often be used.
> 
> Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
> Acked-by: Kyungmin Park <kyungmin.park@samsung.com>
> ---
> Changes since v5:
>  - none.
> 
> Changes since v4:
>  - dropped a requirement of specific order of values in clocks/
>    clock-names properties (Mark) and reference to clock-names in
>    clock-output-names property description (Mark).
> ---
>  .../devicetree/bindings/media/samsung-fimc.txt     |   34 ++++++++++++-----
>  1 file changed, 26 insertions(+), 8 deletions(-)
> 
> diff --git a/Documentation/devicetree/bindings/media/samsung-fimc.txt
> b/Documentation/devicetree/bindings/media/samsung-fimc.txt index
> 96312f6..dbd4020 100644
> --- a/Documentation/devicetree/bindings/media/samsung-fimc.txt
> +++ b/Documentation/devicetree/bindings/media/samsung-fimc.txt
> @@ -32,6 +32,21 @@ way around.
> 
>  The 'camera' node must include at least one 'fimc' child node.
> 
> +Optional properties:
> +
> +- #clock-cells: from the common clock bindings
> (../clock/clock-bindings.txt),
> +  must be 1. A clock provider is associated with the 'camera' node and it
> should
> +  be referenced by external sensors that use clocks provided by the SoC on
> +  CAM_*_CLKOUT pins. The clock specifier cell stores an index of a clock.
> +  The indices are 0, 1 for CAM_A_CLKOUT, CAM_B_CLKOUT clocks respectively.
> +
> +- clock-output-names: from the common clock bindings, should contain names
> of
> +  clocks registered by the camera subsystem corresponding to CAM_A_CLKOUT,
> +  CAM_B_CLKOUT output clocks respectively.

Wouldn't it be better to document the "cam_mclk_a" and "cam_mclk_b" names 
explicitly ? Or do you expect different names to be used in different DT files 
? And as they correspond to the CAM_A_CLKOUT and CAM_B_CLKOUT pins, shouldn't 
they be named "cam_a_clkout" and "cam_b_clkout" ?

> +
> +Note: #clock-cells and clock-output-names are mandatory properties if
> external
> +image sensor devices reference 'camera' device node as a clock provider.
> +

What's the reason not to make them always mandatory ? Backward compatibility 
only ? If so wouldn't it make sense to document the properties as mandatory 
from now on, and treating them as optional in the driver for backward 
compatibility ?

>  'fimc' device nodes
>  -------------------
> 
> @@ -97,8 +112,8 @@ Image sensor nodes
>  The sensor device nodes should be added to their control bus controller
> (e.g. I2C0) nodes and linked to a port node in the csis or the
> parallel-ports node, using the common video interfaces bindings, defined in
> video-interfaces.txt.
> -The implementation of this bindings requires clock-frequency property to be
> -present in the sensor device nodes.
> +An optional clock-frequency property needs to be present in the sensor
> device
> +nodes. Default value when this property is not present is 24 MHz.

This bothers me. Having the FIMC driver read the clock-frequence property from 
the sensor DT nodes feels like a layering violation. Shouldn't the sensor 
drivers call clk_set_rate() explicitly instead ?

>  Example:
> 
> @@ -114,7 +129,7 @@ Example:
>  			vddio-supply = <...>;
> 
>  			clock-frequency = <24000000>;
> -			clocks = <...>;
> +			clocks = <&camera 1>;
>  			clock-names = "mclk";
> 
>  			port {
> @@ -135,7 +150,7 @@ Example:
>  			vddio-supply = <...>;
> 
>  			clock-frequency = <24000000>;
> -			clocks = <...>;
> +			clocks = <&camera 0>;
>  			clock-names = "mclk";
> 
>  			port {
> @@ -149,12 +164,15 @@ Example:
> 
>  	camera {
>  		compatible = "samsung,fimc", "simple-bus";
> -		#address-cells = <1>;
> -		#size-cells = <1>;
> -		status = "okay";
> -
> +		clocks = <&clock 132>, <&clock 133>;
> +		clock-names = "sclk_cam0", "sclk_cam1";

The documentation mentions that clock-names must contain "sclk_cam0", 
"sclk_cam1", "pxl_async0", "pxl_async1". Are the last two optional ? If so I 
think you should clarify the description of the clock-names property. This can 
be done in a separate patch.

> +		#clock-cells = <1>;
> +		clock-output-names = "cam_mclk_a", "cam_mclk_b";
>  		pinctrl-names = "default";
>  		pinctrl-0 = <&cam_port_a_clk_active>;
> +		status = "okay";
> +		#address-cells = <1>;
> +		#size-cells = <1>;
> 
>  		/* parallel camera ports */
>  		parallel-ports {
Hi Laurent,

Thanks for your review.

On 11/03/14 13:30, Laurent Pinchart wrote:
[...]
>> ---
>>  .../devicetree/bindings/media/samsung-fimc.txt     |   34 ++++++++++++-----
>>  1 file changed, 26 insertions(+), 8 deletions(-)
>>
>> diff --git a/Documentation/devicetree/bindings/media/samsung-fimc.txt
>> b/Documentation/devicetree/bindings/media/samsung-fimc.txt index
>> 96312f6..dbd4020 100644
>> --- a/Documentation/devicetree/bindings/media/samsung-fimc.txt
>> +++ b/Documentation/devicetree/bindings/media/samsung-fimc.txt
>> @@ -32,6 +32,21 @@ way around.
>>
>>  The 'camera' node must include at least one 'fimc' child node.
>>
>> +Optional properties:
>> +
>> +- #clock-cells: from the common clock bindings
>> (../clock/clock-bindings.txt),
>> +  must be 1. A clock provider is associated with the 'camera' node and it
>> should
>> +  be referenced by external sensors that use clocks provided by the SoC on
>> +  CAM_*_CLKOUT pins. The clock specifier cell stores an index of a clock.
>> +  The indices are 0, 1 for CAM_A_CLKOUT, CAM_B_CLKOUT clocks respectively.
>> +
>> +- clock-output-names: from the common clock bindings, should contain names
>> of
>> +  clocks registered by the camera subsystem corresponding to CAM_A_CLKOUT,
>> +  CAM_B_CLKOUT output clocks respectively.
> 
> Wouldn't it be better to document the "cam_mclk_a" and "cam_mclk_b" names 
> explicitly ? Or do you expect different names to be used in different DT files 
> ? And as they correspond to the CAM_A_CLKOUT and CAM_B_CLKOUT pins, shouldn't 
> they be named "cam_a_clkout" and "cam_b_clkout" ?

Basically I could use fixed names for these clocks, I just wanted to keep
a possibility to override them in dts to avoid any possible clock name
collisions, rather than keep a list of different names per SoC in the driver. 
Right now fixed names could also be used for all SoCs I'm aware of, 
nevertheless I would prefer to keep the clock-output-names property.
"cam_a_clkout", "cam_b_clkout" may be indeed better names, I'll change
that.

>> +Note: #clock-cells and clock-output-names are mandatory properties if
>> external
>> +image sensor devices reference 'camera' device node as a clock provider.
>> +
> 
> What's the reason not to make them always mandatory ? Backward compatibility 
> only ? If so wouldn't it make sense to document the properties as mandatory 
> from now on, and treating them as optional in the driver for backward 
> compatibility ?

Yes, it's for backwards compatibility only. It may be a good idea to just 
document them as required, since this is how the device is expected to be 
described in DT from now. I'll just make these a required properties, 
the driver already handles them as optional.

>>  'fimc' device nodes
>>  -------------------
>>
>> @@ -97,8 +112,8 @@ Image sensor nodes
>>  The sensor device nodes should be added to their control bus controller
>> (e.g. I2C0) nodes and linked to a port node in the csis or the
>> parallel-ports node, using the common video interfaces bindings, defined in
>> video-interfaces.txt.
>> -The implementation of this bindings requires clock-frequency property to be
>> -present in the sensor device nodes.
>> +An optional clock-frequency property needs to be present in the sensor
>> device
>> +nodes. Default value when this property is not present is 24 MHz.
> 
> This bothers me. Having the FIMC driver read the clock-frequence property from 
> the sensor DT nodes feels like a layering violation. Shouldn't the sensor 
> drivers call clk_set_rate() explicitly instead ?

It is supposed to do so, after this whole patch series. So the camera
controller driver will not need such properties. What do you think about
removing this sentence altogether ?

>>  Example:
>>
>> @@ -114,7 +129,7 @@ Example:
>>  			vddio-supply = <...>;
>>
>>  			clock-frequency = <24000000>;
>> -			clocks = <...>;
>> +			clocks = <&camera 1>;
>>  			clock-names = "mclk";
>>
>>  			port {
>> @@ -135,7 +150,7 @@ Example:
>>  			vddio-supply = <...>;
>>
>>  			clock-frequency = <24000000>;
>> -			clocks = <...>;
>> +			clocks = <&camera 0>;
>>  			clock-names = "mclk";
>>
>>  			port {
>> @@ -149,12 +164,15 @@ Example:
>>
>>  	camera {
>>  		compatible = "samsung,fimc", "simple-bus";
>> -		#address-cells = <1>;
>> -		#size-cells = <1>;
>> -		status = "okay";
>> -
>> +		clocks = <&clock 132>, <&clock 133>;
>> +		clock-names = "sclk_cam0", "sclk_cam1";
> 
> The documentation mentions that clock-names must contain "sclk_cam0", 
> "sclk_cam1", "pxl_async0", "pxl_async1". Are the last two optional ? If so I 
> think you should clarify the description of the clock-names property. This can 
> be done in a separate patch.

"pxl_async0", "pxl_async1" are mandatory, I'll add them also into
this example dts.

>> +		#clock-cells = <1>;
>> +		clock-output-names = "cam_mclk_a", "cam_mclk_b";
>>  		pinctrl-names = "default";
>>  		pinctrl-0 = <&cam_port_a_clk_active>;
>> +		status = "okay";
>> +		#address-cells = <1>;
>> +		#size-cells = <1>;
>>
>>  		/* parallel camera ports */
>>  		parallel-ports {

--
Regards,
Sylwester
--
To unsubscribe from this list: send the line "unsubscribe linux-samsung-soc" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Laurent Pinchart March 11, 2014, 2:58 p.m. UTC | #3
Hi Sylwester,

On Tuesday 11 March 2014 14:38:37 Sylwester Nawrocki wrote:
> Hi Laurent,
> 
> Thanks for your review.

You're welcome.

> On 11/03/14 13:30, Laurent Pinchart wrote:
> [...]
> 
> >> ---
> >> 
> >>  .../devicetree/bindings/media/samsung-fimc.txt     |   34 +++++++++-----
> >>  1 file changed, 26 insertions(+), 8 deletions(-)
> >> 
> >> diff --git a/Documentation/devicetree/bindings/media/samsung-fimc.txt
> >> b/Documentation/devicetree/bindings/media/samsung-fimc.txt index
> >> 96312f6..dbd4020 100644
> >> --- a/Documentation/devicetree/bindings/media/samsung-fimc.txt
> >> +++ b/Documentation/devicetree/bindings/media/samsung-fimc.txt
> >> @@ -32,6 +32,21 @@ way around.
> >> 
> >>  The 'camera' node must include at least one 'fimc' child node.
> >> 
> >> +Optional properties:
> >> +
> >> +- #clock-cells: from the common clock bindings
> >> (../clock/clock-bindings.txt),
> >> +  must be 1. A clock provider is associated with the 'camera' node and
> >> it should
> >> +  be referenced by external sensors that use clocks provided by the SoC
> >> on
> >> +  CAM_*_CLKOUT pins. The clock specifier cell stores an index of a
> >> clock.
> >> +  The indices are 0, 1 for CAM_A_CLKOUT, CAM_B_CLKOUT clocks
> >> respectively.
> >> +
> >> +- clock-output-names: from the common clock bindings, should contain
> >> names of
> >> +  clocks registered by the camera subsystem corresponding to
> >> CAM_A_CLKOUT,
> >> +  CAM_B_CLKOUT output clocks respectively.
> > 
> > Wouldn't it be better to document the "cam_mclk_a" and "cam_mclk_b" names
> > explicitly ? Or do you expect different names to be used in different DT
> > files ? And as they correspond to the CAM_A_CLKOUT and CAM_B_CLKOUT pins,
> > shouldn't they be named "cam_a_clkout" and "cam_b_clkout" ?
> 
> Basically I could use fixed names for these clocks, I just wanted to keep
> a possibility to override them in dts to avoid any possible clock name
> collisions, rather than keep a list of different names per SoC in the
> driver. Right now fixed names could also be used for all SoCs I'm aware of,
> nevertheless I would prefer to keep the clock-output-names property.
> "cam_a_clkout", "cam_b_clkout" may be indeed better names, I'll change
> that.

OK, variable names are fine with me.

> >> +Note: #clock-cells and clock-output-names are mandatory properties if
> >> external
> >> +image sensor devices reference 'camera' device node as a clock provider.
> >> +
> > 
> > What's the reason not to make them always mandatory ? Backward
> > compatibility only ? If so wouldn't it make sense to document the
> > properties as mandatory from now on, and treating them as optional in the
> > driver for backward compatibility ?
> 
> Yes, it's for backwards compatibility only. It may be a good idea to just
> document them as required, since this is how the device is expected to be
> described in DT from now. I'll just make these a required properties,
> the driver already handles them as optional.

OK.

> >>  'fimc' device nodes
> >>  -------------------
> >> 
> >> @@ -97,8 +112,8 @@ Image sensor nodes
> >> 
> >>  The sensor device nodes should be added to their control bus controller
> >> 
> >> (e.g. I2C0) nodes and linked to a port node in the csis or the
> >> parallel-ports node, using the common video interfaces bindings, defined
> >> in video-interfaces.txt.
> >> -The implementation of this bindings requires clock-frequency property to
> >> be
> >> -present in the sensor device nodes.
> >> +An optional clock-frequency property needs to be present in the sensor
> >> device
> >> +nodes. Default value when this property is not present is 24 MHz.
> > 
> > This bothers me. Having the FIMC driver read the clock-frequence property
> > from the sensor DT nodes feels like a layering violation. Shouldn't the
> > sensor drivers call clk_set_rate() explicitly instead ?
> 
> It is supposed to do so, after this whole patch series. So the camera
> controller driver will not need such properties. What do you think about
> removing this sentence altogether ?

Sure. As the FIMC won't access the clock-frequency property of the sensor 
anymore after this patch series, let's just drop the mention of clock-
frequency.

> >>  Example:
> >> @@ -114,7 +129,7 @@ Example:
> >>  			vddio-supply = <...>;
> >>  			
> >>  			clock-frequency = <24000000>;
> >> 
> >> -			clocks = <...>;
> >> +			clocks = <&camera 1>;
> >> 
> >>  			clock-names = "mclk";
> >>  			
> >>  			port {
> >> 
> >> @@ -135,7 +150,7 @@ Example:
> >>  			vddio-supply = <...>;
> >>  			
> >>  			clock-frequency = <24000000>;
> >> 
> >> -			clocks = <...>;
> >> +			clocks = <&camera 0>;
> >> 
> >>  			clock-names = "mclk";
> >>  			
> >>  			port {
> >> 
> >> @@ -149,12 +164,15 @@ Example:
> >>  	camera {
> >>  	
> >>  		compatible = "samsung,fimc", "simple-bus";
> >> 
> >> -		#address-cells = <1>;
> >> -		#size-cells = <1>;
> >> -		status = "okay";
> >> -
> >> +		clocks = <&clock 132>, <&clock 133>;
> >> +		clock-names = "sclk_cam0", "sclk_cam1";
> > 
> > The documentation mentions that clock-names must contain "sclk_cam0",
> > "sclk_cam1", "pxl_async0", "pxl_async1". Are the last two optional ? If so
> > I think you should clarify the description of the clock-names property.
> > This can be done in a separate patch.
> 
> "pxl_async0", "pxl_async1" are mandatory, I'll add them also into
> this example dts.

OK.

> >> +		#clock-cells = <1>;
> >> +		clock-output-names = "cam_mclk_a", "cam_mclk_b";
> >> 
> >>  		pinctrl-names = "default";
> >>  		pinctrl-0 = <&cam_port_a_clk_active>;
> >> 
> >> +		status = "okay";
> >> +		#address-cells = <1>;
> >> +		#size-cells = <1>;
> >> 
> >>  		/* parallel camera ports */
> >>  		parallel-ports {
diff mbox

Patch

diff --git a/Documentation/devicetree/bindings/media/samsung-fimc.txt b/Documentation/devicetree/bindings/media/samsung-fimc.txt
index 96312f6..dbd4020 100644
--- a/Documentation/devicetree/bindings/media/samsung-fimc.txt
+++ b/Documentation/devicetree/bindings/media/samsung-fimc.txt
@@ -32,6 +32,21 @@  way around.
 
 The 'camera' node must include at least one 'fimc' child node.
 
+Optional properties:
+
+- #clock-cells: from the common clock bindings (../clock/clock-bindings.txt),
+  must be 1. A clock provider is associated with the 'camera' node and it should
+  be referenced by external sensors that use clocks provided by the SoC on
+  CAM_*_CLKOUT pins. The clock specifier cell stores an index of a clock.
+  The indices are 0, 1 for CAM_A_CLKOUT, CAM_B_CLKOUT clocks respectively.
+
+- clock-output-names: from the common clock bindings, should contain names of
+  clocks registered by the camera subsystem corresponding to CAM_A_CLKOUT,
+  CAM_B_CLKOUT output clocks respectively.
+
+Note: #clock-cells and clock-output-names are mandatory properties if external
+image sensor devices reference 'camera' device node as a clock provider.
+
 'fimc' device nodes
 -------------------
 
@@ -97,8 +112,8 @@  Image sensor nodes
 The sensor device nodes should be added to their control bus controller (e.g.
 I2C0) nodes and linked to a port node in the csis or the parallel-ports node,
 using the common video interfaces bindings, defined in video-interfaces.txt.
-The implementation of this bindings requires clock-frequency property to be
-present in the sensor device nodes.
+An optional clock-frequency property needs to be present in the sensor device
+nodes. Default value when this property is not present is 24 MHz.
 
 Example:
 
@@ -114,7 +129,7 @@  Example:
 			vddio-supply = <...>;
 
 			clock-frequency = <24000000>;
-			clocks = <...>;
+			clocks = <&camera 1>;
 			clock-names = "mclk";
 
 			port {
@@ -135,7 +150,7 @@  Example:
 			vddio-supply = <...>;
 
 			clock-frequency = <24000000>;
-			clocks = <...>;
+			clocks = <&camera 0>;
 			clock-names = "mclk";
 
 			port {
@@ -149,12 +164,15 @@  Example:
 
 	camera {
 		compatible = "samsung,fimc", "simple-bus";
-		#address-cells = <1>;
-		#size-cells = <1>;
-		status = "okay";
-
+		clocks = <&clock 132>, <&clock 133>;
+		clock-names = "sclk_cam0", "sclk_cam1";
+		#clock-cells = <1>;
+		clock-output-names = "cam_mclk_a", "cam_mclk_b";
 		pinctrl-names = "default";
 		pinctrl-0 = <&cam_port_a_clk_active>;
+		status = "okay";
+		#address-cells = <1>;
+		#size-cells = <1>;
 
 		/* parallel camera ports */
 		parallel-ports {