diff mbox

[RFC,v2,12/13] ARM: mach-vt8500: cpus/cpu nodes dts updates

Message ID 1366644455-16550-13-git-send-email-lorenzo.pieralisi@arm.com (mailing list archive)
State New, archived
Headers show

Commit Message

Lorenzo Pieralisi April 22, 2013, 3:27 p.m. UTC
This patch updates the in-kernel dts files according to the latest cpus
and cpu bindings updates for ARM.

Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
---
 arch/arm/boot/dts/wm8505.dtsi | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

Comments

Tony Prisk April 23, 2013, 2:13 a.m. UTC | #1
On 23/04/13 03:27, Lorenzo Pieralisi wrote:
> This patch updates the in-kernel dts files according to the latest cpus
> and cpu bindings updates for ARM.
>
> Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
> ---
>   arch/arm/boot/dts/wm8505.dtsi | 2 +-
>   1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/arch/arm/boot/dts/wm8505.dtsi b/arch/arm/boot/dts/wm8505.dtsi
> index e74a1c0..a470808 100644
> --- a/arch/arm/boot/dts/wm8505.dtsi
> +++ b/arch/arm/boot/dts/wm8505.dtsi
> @@ -13,7 +13,7 @@
>   
>   	cpus {
>   		cpu@0 {
> -			compatible = "arm,arm926ejs";
> +			compatible = "arm,arm926";
>   		};
>   	};
>   
As the only author for that file, and the arch-vt8500 maintainer, a CC 
would have been nice :)

This binding is still incomplete.device_type, compatible and reg are all 
required properties.
Also, you seem to have only fixed the wm8505.dtsi, while vt8500.dtsi, 
wm8650.dtsi and wm8850.dtsi are all left incorrect.

I am already carrying a patch to fix all this properly so could you drop 
the arch-vt8500 related patch and I will send in a more complete version 
for 3.11

Regards
Tony Prisk
Tony Prisk April 23, 2013, 2:20 a.m. UTC | #2
On 23/04/13 14:13, Tony Prisk wrote:
> On 23/04/13 03:27, Lorenzo Pieralisi wrote:
>> This patch updates the in-kernel dts files according to the latest cpus
>> and cpu bindings updates for ARM.
>>
>> Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
>> ---
>>   arch/arm/boot/dts/wm8505.dtsi | 2 +-
>>   1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/arch/arm/boot/dts/wm8505.dtsi 
>> b/arch/arm/boot/dts/wm8505.dtsi
>> index e74a1c0..a470808 100644
>> --- a/arch/arm/boot/dts/wm8505.dtsi
>> +++ b/arch/arm/boot/dts/wm8505.dtsi
>> @@ -13,7 +13,7 @@
>>         cpus {
>>           cpu@0 {
>> -            compatible = "arm,arm926ejs";
>> +            compatible = "arm,arm926";
>>           };
>>       };
> As the only author for that file, and the arch-vt8500 maintainer, a CC 
> would have been nice :)
>
> This binding is still incomplete.device_type, compatible and reg are 
> all required properties.
> Also, you seem to have only fixed the wm8505.dtsi, while vt8500.dtsi, 
> wm8650.dtsi and wm8850.dtsi are all left incorrect.
>
> I am already carrying a patch to fix all this properly so could you 
> drop the arch-vt8500 related patch and I will send in a more complete 
> version for 3.11
>
> Regards
> Tony Prisk

I didn't notice the rules had been changed in the last patch - shouldn't 
you be changing the rules (read: binding) before changing the devicetree 
files?
Request still stands - please drop this patch and I will submit a more 
complete one with the other SoCs updated as well.

Tony P
Tony Prisk April 23, 2013, 2:43 a.m. UTC | #3
On 23/04/13 03:27, Lorenzo Pieralisi wrote:
> This patch updates the in-kernel dts files according to the latest cpus
> and cpu bindings updates for ARM.
>
> Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
> ---
>   arch/arm/boot/dts/wm8505.dtsi | 2 +-
>   1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/arch/arm/boot/dts/wm8505.dtsi b/arch/arm/boot/dts/wm8505.dtsi
> index e74a1c0..a470808 100644
> --- a/arch/arm/boot/dts/wm8505.dtsi
> +++ b/arch/arm/boot/dts/wm8505.dtsi
> @@ -13,7 +13,7 @@
>   
>   	cpus {
>   		cpu@0 {
> -			compatible = "arm,arm926ejs";
> +			compatible = "arm,arm926";
>   		};
>   	};
>   
The more I look at this, the more wrong it is :/

 From the new binding documentation,

+	A cpus node must define the following properties:
+
+	- #address-cells
+		Usage: required
+		Value type: <u32>
+		Definition: must be set to 1 for 32-bit systems and 2 for
+			    64-bit systems
+	- #size-cells
+		Usage: required
+		Value type: <u32>
+		Definition: must be set to 0

...

+- cpu node
+
+	Description: Describes a CPU in an ARM based system
+
+	PROPERTIES
+
+	- device_type
+		Usage: required
+		Value type: <string>
+		Definition: must be "cpu"

Three required properties that aren't present in the patch.

cpus {
     #size-cells = <0>;
     #address-cells = <1>;

     cpu {
         device_type = "cpu"
         compatible = "arm,arm926";
     };
};

Regards
Tony P
Lorenzo Pieralisi April 23, 2013, 9:15 a.m. UTC | #4
On Tue, Apr 23, 2013 at 03:13:18AM +0100, Tony Prisk wrote:
> On 23/04/13 03:27, Lorenzo Pieralisi wrote:
> > This patch updates the in-kernel dts files according to the latest cpus
> > and cpu bindings updates for ARM.
> >
> > Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
> > ---
> >   arch/arm/boot/dts/wm8505.dtsi | 2 +-
> >   1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/arch/arm/boot/dts/wm8505.dtsi b/arch/arm/boot/dts/wm8505.dtsi
> > index e74a1c0..a470808 100644
> > --- a/arch/arm/boot/dts/wm8505.dtsi
> > +++ b/arch/arm/boot/dts/wm8505.dtsi
> > @@ -13,7 +13,7 @@
> >   
> >   	cpus {
> >   		cpu@0 {
> > -			compatible = "arm,arm926ejs";
> > +			compatible = "arm,arm926";
> >   		};
> >   	};
> >   
> As the only author for that file, and the arch-vt8500 maintainer, a CC 
> would have been nice :)

Apologies.

> This binding is still incomplete.device_type, compatible and reg are all 
> required properties.

device_type should already be there and I should not be in charge of
adding it; reg is not required for that processor.

> Also, you seem to have only fixed the wm8505.dtsi, while vt8500.dtsi, 
> wm8650.dtsi and wm8850.dtsi are all left incorrect.

There is no cpus node in there. I can't add cpus node in all dts files
in the kernel where it is missing (why have those dts files been
accepted in the first place, with no cpus node ? why ? cpus node is mandatory
since device tree was introduced, I really do not understand why those dts
landed in the kernel as they are), I would kindly ask maintainers to do
that, I just can not know what cpus are there for every given SoC.

> I am already carrying a patch to fix all this properly so could you drop 
> the arch-vt8500 related patch and I will send in a more complete version 
> for 3.11

Good, thanks.
Lorenzo
Lorenzo Pieralisi April 23, 2013, 9:16 a.m. UTC | #5
On Tue, Apr 23, 2013 at 03:20:46AM +0100, Tony Prisk wrote:
> On 23/04/13 14:13, Tony Prisk wrote:
> > On 23/04/13 03:27, Lorenzo Pieralisi wrote:
> >> This patch updates the in-kernel dts files according to the latest cpus
> >> and cpu bindings updates for ARM.
> >>
> >> Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
> >> ---
> >>   arch/arm/boot/dts/wm8505.dtsi | 2 +-
> >>   1 file changed, 1 insertion(+), 1 deletion(-)
> >>
> >> diff --git a/arch/arm/boot/dts/wm8505.dtsi 
> >> b/arch/arm/boot/dts/wm8505.dtsi
> >> index e74a1c0..a470808 100644
> >> --- a/arch/arm/boot/dts/wm8505.dtsi
> >> +++ b/arch/arm/boot/dts/wm8505.dtsi
> >> @@ -13,7 +13,7 @@
> >>         cpus {
> >>           cpu@0 {
> >> -            compatible = "arm,arm926ejs";
> >> +            compatible = "arm,arm926";
> >>           };
> >>       };
> > As the only author for that file, and the arch-vt8500 maintainer, a CC 
> > would have been nice :)
> >
> > This binding is still incomplete.device_type, compatible and reg are 
> > all required properties.
> > Also, you seem to have only fixed the wm8505.dtsi, while vt8500.dtsi, 
> > wm8650.dtsi and wm8850.dtsi are all left incorrect.
> >
> > I am already carrying a patch to fix all this properly so could you 
> > drop the arch-vt8500 related patch and I will send in a more complete 
> > version for 3.11
> >
> > Regards
> > Tony Prisk
> 
> I didn't notice the rules had been changed in the last patch - shouldn't 
> you be changing the rules (read: binding) before changing the devicetree 
> files?

Fair enough, point taken.

> Request still stands - please drop this patch and I will submit a more 
> complete one with the other SoCs updated as well.

Great, thanks,
Lorenzo
Lorenzo Pieralisi April 23, 2013, 9:26 a.m. UTC | #6
On Tue, Apr 23, 2013 at 03:43:48AM +0100, Tony Prisk wrote:
> On 23/04/13 03:27, Lorenzo Pieralisi wrote:
> > This patch updates the in-kernel dts files according to the latest cpus
> > and cpu bindings updates for ARM.
> >
> > Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
> > ---
> >   arch/arm/boot/dts/wm8505.dtsi | 2 +-
> >   1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/arch/arm/boot/dts/wm8505.dtsi b/arch/arm/boot/dts/wm8505.dtsi
> > index e74a1c0..a470808 100644
> > --- a/arch/arm/boot/dts/wm8505.dtsi
> > +++ b/arch/arm/boot/dts/wm8505.dtsi
> > @@ -13,7 +13,7 @@
> >   
> >   	cpus {
> >   		cpu@0 {
> > -			compatible = "arm,arm926ejs";
> > +			compatible = "arm,arm926";
> >   		};
> >   	};
> >   
> The more I look at this, the more wrong it is :/
> 
>  From the new binding documentation,
> 
> +	A cpus node must define the following properties:
> +
> +	- #address-cells
> +		Usage: required
> +		Value type: <u32>
> +		Definition: must be set to 1 for 32-bit systems and 2 for
> +			    64-bit systems
> +	- #size-cells
> +		Usage: required
> +		Value type: <u32>
> +		Definition: must be set to 0
> 
> ...
> 
> +- cpu node
> +
> +	Description: Describes a CPU in an ARM based system
> +
> +	PROPERTIES
> +
> +	- device_type
> +		Usage: required
> +		Value type: <string>
> +		Definition: must be "cpu"
> 

This property is required since DT was invented, again, I should not be
in charge of adding it.

> Three required properties that aren't present in the patch.
> 
> cpus {
>      #size-cells = <0>;
>      #address-cells = <1>;
> 
>      cpu {
>          device_type = "cpu"
>          compatible = "arm,arm926";
>      };
> };

I am of two minds about this. That processor does not have an MPIDR
equivalent so in theory #address-cells and #size-cells could be omitted
because the reg property is meaningless. I do not know to be honest the
best way to handle this.

Thoughts ?

Thanks for the review,
Lorenzo
diff mbox

Patch

diff --git a/arch/arm/boot/dts/wm8505.dtsi b/arch/arm/boot/dts/wm8505.dtsi
index e74a1c0..a470808 100644
--- a/arch/arm/boot/dts/wm8505.dtsi
+++ b/arch/arm/boot/dts/wm8505.dtsi
@@ -13,7 +13,7 @@ 
 
 	cpus {
 		cpu@0 {
-			compatible = "arm,arm926ejs";
+			compatible = "arm,arm926";
 		};
 	};