diff mbox series

arm64: dts: rockchip: Correct clock-output-names for rk817

Message ID 20220216165837.4721-1-macroalpha82@gmail.com (mailing list archive)
State New, archived
Headers show
Series arm64: dts: rockchip: Correct clock-output-names for rk817 | expand

Commit Message

Chris Morgan Feb. 16, 2022, 4:58 p.m. UTC
From: Chris Morgan <macromorgan@hotmail.com>

According to the datasheet for the rk817 there is only a single output
clock for the PMIC. I don't believe this is causing any harm though
and testing this change doesn't appear to have any noticeable effects.

Signed-off-by: Chris Morgan <macromorgan@hotmail.com>
---
 arch/arm64/boot/dts/rockchip/rk3326-odroid-go2.dts | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

Comments

Heiko Stuebner Feb. 16, 2022, 8:42 p.m. UTC | #1
Hi,

Am Mittwoch, 16. Februar 2022, 17:58:37 CET schrieb Chris Morgan:
> From: Chris Morgan <macromorgan@hotmail.com>
> 
> According to the datasheet for the rk817 there is only a single output
> clock for the PMIC. I don't believe this is causing any harm though
> and testing this change doesn't appear to have any noticeable effects.

hmm, looking at the go2's schematics, I'd disagree:

- yes, the rk817 has only one clock output
- but no, it is used for the clkout_32k signal, which gets piped back into
  the SoC. As that clock dynamically appears at some unspecified point
  when the rk817 probes, the clock controller expects it to be named
  xin32k to bind into its clock-tree.

The driver (drivers/clk/clk-rk808.c) really puts that rk817-clk32k at
position 2 but I'm not sure if that simply papers over some implementation


Heiko

> Signed-off-by: Chris Morgan <macromorgan@hotmail.com>
> ---
>  arch/arm64/boot/dts/rockchip/rk3326-odroid-go2.dts | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/arch/arm64/boot/dts/rockchip/rk3326-odroid-go2.dts b/arch/arm64/boot/dts/rockchip/rk3326-odroid-go2.dts
> index ea0695b51ecd..5b23e607876e 100644
> --- a/arch/arm64/boot/dts/rockchip/rk3326-odroid-go2.dts
> +++ b/arch/arm64/boot/dts/rockchip/rk3326-odroid-go2.dts
> @@ -296,7 +296,7 @@ rk817: pmic@20 {
>  		reg = <0x20>;
>  		interrupt-parent = <&gpio0>;
>  		interrupts = <RK_PB2 IRQ_TYPE_LEVEL_LOW>;
> -		clock-output-names = "rk808-clkout1", "xin32k";
> +		clock-output-names = "rk808-clkout1";
>  		clock-names = "mclk";
>  		clocks = <&cru SCLK_I2S1_OUT>;
>  		pinctrl-names = "default";
>
Robin Murphy Feb. 16, 2022, 9:34 p.m. UTC | #2
On 2022-02-16 20:42, Heiko Stübner wrote:
> Hi,
> 
> Am Mittwoch, 16. Februar 2022, 17:58:37 CET schrieb Chris Morgan:
>> From: Chris Morgan <macromorgan@hotmail.com>
>>
>> According to the datasheet for the rk817 there is only a single output
>> clock for the PMIC. I don't believe this is causing any harm though
>> and testing this change doesn't appear to have any noticeable effects.
> 
> hmm, looking at the go2's schematics, I'd disagree:
> 
> - yes, the rk817 has only one clock output
> - but no, it is used for the clkout_32k signal, which gets piped back into
>    the SoC. As that clock dynamically appears at some unspecified point
>    when the rk817 probes, the clock controller expects it to be named
>    xin32k to bind into its clock-tree.
> 
> The driver (drivers/clk/clk-rk808.c) really puts that rk817-clk32k at
> position 2 but I'm not sure if that simply papers over some implementation

Indeed it appears to be a hangover from RK808 where CLK32KOUT1 is 
always-on and CLK32KOUT2 is the switchable one. It's even more obvious 
on RK805 which still shares the original control register layout.

Robin.

> 
> 
> Heiko
> 
>> Signed-off-by: Chris Morgan <macromorgan@hotmail.com>
>> ---
>>   arch/arm64/boot/dts/rockchip/rk3326-odroid-go2.dts | 2 +-
>>   1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/arch/arm64/boot/dts/rockchip/rk3326-odroid-go2.dts b/arch/arm64/boot/dts/rockchip/rk3326-odroid-go2.dts
>> index ea0695b51ecd..5b23e607876e 100644
>> --- a/arch/arm64/boot/dts/rockchip/rk3326-odroid-go2.dts
>> +++ b/arch/arm64/boot/dts/rockchip/rk3326-odroid-go2.dts
>> @@ -296,7 +296,7 @@ rk817: pmic@20 {
>>   		reg = <0x20>;
>>   		interrupt-parent = <&gpio0>;
>>   		interrupts = <RK_PB2 IRQ_TYPE_LEVEL_LOW>;
>> -		clock-output-names = "rk808-clkout1", "xin32k";
>> +		clock-output-names = "rk808-clkout1";
>>   		clock-names = "mclk";
>>   		clocks = <&cru SCLK_I2S1_OUT>;
>>   		pinctrl-names = "default";
>>
> 
> 
> 
> 
> 
> _______________________________________________
> Linux-rockchip mailing list
> Linux-rockchip@lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/linux-rockchip
Chris Morgan Feb. 22, 2022, 7:14 p.m. UTC | #3
On Wed, Feb 16, 2022 at 09:34:52PM +0000, Robin Murphy wrote:
> On 2022-02-16 20:42, Heiko Stübner wrote:
> > Hi,
> > 
> > Am Mittwoch, 16. Februar 2022, 17:58:37 CET schrieb Chris Morgan:
> > > From: Chris Morgan <macromorgan@hotmail.com>
> > > 
> > > According to the datasheet for the rk817 there is only a single output
> > > clock for the PMIC. I don't believe this is causing any harm though
> > > and testing this change doesn't appear to have any noticeable effects.
> > 
> > hmm, looking at the go2's schematics, I'd disagree:
> > 
> > - yes, the rk817 has only one clock output
> > - but no, it is used for the clkout_32k signal, which gets piped back into
> >    the SoC. As that clock dynamically appears at some unspecified point
> >    when the rk817 probes, the clock controller expects it to be named
> >    xin32k to bind into its clock-tree.
> > 
> > The driver (drivers/clk/clk-rk808.c) really puts that rk817-clk32k at
> > position 2 but I'm not sure if that simply papers over some implementation
> 
> Indeed it appears to be a hangover from RK808 where CLK32KOUT1 is always-on
> and CLK32KOUT2 is the switchable one. It's even more obvious on RK805 which
> still shares the original control register layout.
> 
> Robin.

Okay, so I'm thinking for now we cancel this patch, note in the yaml updates
patch series that this issue is present, update the driver, and then notify
the board dts maintainers to switch over and test. Does that sound good?

Thank you.

> 
> > 
> > 
> > Heiko
> > 
> > > Signed-off-by: Chris Morgan <macromorgan@hotmail.com>
> > > ---
> > >   arch/arm64/boot/dts/rockchip/rk3326-odroid-go2.dts | 2 +-
> > >   1 file changed, 1 insertion(+), 1 deletion(-)
> > > 
> > > diff --git a/arch/arm64/boot/dts/rockchip/rk3326-odroid-go2.dts b/arch/arm64/boot/dts/rockchip/rk3326-odroid-go2.dts
> > > index ea0695b51ecd..5b23e607876e 100644
> > > --- a/arch/arm64/boot/dts/rockchip/rk3326-odroid-go2.dts
> > > +++ b/arch/arm64/boot/dts/rockchip/rk3326-odroid-go2.dts
> > > @@ -296,7 +296,7 @@ rk817: pmic@20 {
> > >   		reg = <0x20>;
> > >   		interrupt-parent = <&gpio0>;
> > >   		interrupts = <RK_PB2 IRQ_TYPE_LEVEL_LOW>;
> > > -		clock-output-names = "rk808-clkout1", "xin32k";
> > > +		clock-output-names = "rk808-clkout1";
> > >   		clock-names = "mclk";
> > >   		clocks = <&cru SCLK_I2S1_OUT>;
> > >   		pinctrl-names = "default";
> > > 
> > 
> > 
> > 
> > 
> > 
> > _______________________________________________
> > Linux-rockchip mailing list
> > Linux-rockchip@lists.infradead.org
> > http://lists.infradead.org/mailman/listinfo/linux-rockchip
diff mbox series

Patch

diff --git a/arch/arm64/boot/dts/rockchip/rk3326-odroid-go2.dts b/arch/arm64/boot/dts/rockchip/rk3326-odroid-go2.dts
index ea0695b51ecd..5b23e607876e 100644
--- a/arch/arm64/boot/dts/rockchip/rk3326-odroid-go2.dts
+++ b/arch/arm64/boot/dts/rockchip/rk3326-odroid-go2.dts
@@ -296,7 +296,7 @@  rk817: pmic@20 {
 		reg = <0x20>;
 		interrupt-parent = <&gpio0>;
 		interrupts = <RK_PB2 IRQ_TYPE_LEVEL_LOW>;
-		clock-output-names = "rk808-clkout1", "xin32k";
+		clock-output-names = "rk808-clkout1";
 		clock-names = "mclk";
 		clocks = <&cru SCLK_I2S1_OUT>;
 		pinctrl-names = "default";