[PATCH v1 0/3] MIPS: dts: ralink: mt7628a: Tweak for GARDENA smart Gateway

[PATCH v1 0/3] MIPS: dts: ralink: mt7628a: Tweak for GARDENA smart Gateway

[Ezra Buehler]

From: Ezra Buehler <ezra.buehler@husqvarnagroup.com>

Various devicetree changes needed for the MT7688-based GARDENA smart
Gateway.

Ezra Buehler (3):
  MIPS: dts: ralink: mt7628a: Fix sysc's compatible property for MT7688
  MIPS: dts: ralink: mt7628a: Update watchdog node according to bindings
  MIPS: dts: ralink: gardena_smart_gateway_mt7688: Fix power LED

 .../boot/dts/ralink/gardena_smart_gateway_mt7688.dts  |  2 +-
 arch/mips/boot/dts/ralink/mt7628a.dtsi                | 11 +++--------
 2 files changed, 4 insertions(+), 9 deletions(-)

--
2.43.0

[PATCH v1 1/3] MIPS: dts: ralink: mt7628a: Fix sysc's compatible property for MT7688

[Ezra Buehler]

From: Ezra Buehler <ezra.buehler@husqvarnagroup.com>

Otherwise, the MT7688-based GARDENA smart Gateway will fail to boot
printing "Kernel panic - not syncing: unable to get CPU clock, err=-2".

Signed-off-by: Ezra Buehler <ezra.buehler@husqvarnagroup.com>
---
 arch/mips/boot/dts/ralink/mt7628a.dtsi | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/arch/mips/boot/dts/ralink/mt7628a.dtsi b/arch/mips/boot/dts/ralink/mt7628a.dtsi
index 0212700c4fb4..10221a41f02a 100644
--- a/arch/mips/boot/dts/ralink/mt7628a.dtsi
+++ b/arch/mips/boot/dts/ralink/mt7628a.dtsi
@@ -33,7 +33,7 @@ palmbus@10000000 {
 		#size-cells = <1>;
 
 		sysc: syscon@0 {
-			compatible = "ralink,mt7628-sysc", "syscon";
+			compatible = "ralink,mt7628-sysc", "ralink,mt7688-sysc", "syscon";
 			reg = <0x0 0x60>;
 			#clock-cells = <1>;
 			#reset-cells = <1>;
-- 
2.43.0

[PATCH v1 2/3] MIPS: dts: ralink: mt7628a: Update watchdog node according to bindings

[Ezra Buehler]

From: Ezra Buehler <ezra.buehler@husqvarnagroup.com>

Most notably, add the mediatek,sysctl phandle and remove the redundant
reset/interrupt-related properties from the watchdog node. This is in
line with the corresponding devicetree (mt7628an.dtsi) used by the
OpenWrt project.

This has been tested on the MT7688-based GARDENA smart Gateway.

Signed-off-by: Ezra Buehler <ezra.buehler@husqvarnagroup.com>
---
 arch/mips/boot/dts/ralink/mt7628a.dtsi | 9 ++-------
 1 file changed, 2 insertions(+), 7 deletions(-)

diff --git a/arch/mips/boot/dts/ralink/mt7628a.dtsi b/arch/mips/boot/dts/ralink/mt7628a.dtsi
index 10221a41f02a..5d7a6cfa9e2b 100644
--- a/arch/mips/boot/dts/ralink/mt7628a.dtsi
+++ b/arch/mips/boot/dts/ralink/mt7628a.dtsi
@@ -134,13 +134,8 @@ pinmux_p4led_an_gpio: p4led-an-gpio-pins {
 
 		watchdog: watchdog@100 {
 			compatible = "mediatek,mt7621-wdt";
-			reg = <0x100 0x30>;
-
-			resets = <&sysc 8>;
-			reset-names = "wdt";
-
-			interrupt-parent = <&intc>;
-			interrupts = <24>;
+			reg = <0x100 0x100>;
+			mediatek,sysctl = <&sysc>;
 
 			status = "disabled";
 		};
-- 
2.43.0

[PATCH v1 3/3] MIPS: dts: ralink: gardena_smart_gateway_mt7688: Fix power LED

[Ezra Buehler]

From: Ezra Buehler <ezra.buehler@husqvarnagroup.com>

When starting up, the GARDENA smart Gateway's power LED should be
flashing green. It is unclear why it was initially set to "off".

The LED frequency cannot be configured in the devicetree. Luckily, the
default is 1 Hz, which is what we want.

Signed-off-by: Ezra Buehler <ezra.buehler@husqvarnagroup.com>
---
 arch/mips/boot/dts/ralink/gardena_smart_gateway_mt7688.dts | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/arch/mips/boot/dts/ralink/gardena_smart_gateway_mt7688.dts b/arch/mips/boot/dts/ralink/gardena_smart_gateway_mt7688.dts
index 7743d014631a..0bfb1dde9764 100644
--- a/arch/mips/boot/dts/ralink/gardena_smart_gateway_mt7688.dts
+++ b/arch/mips/boot/dts/ralink/gardena_smart_gateway_mt7688.dts
@@ -56,7 +56,7 @@ led-power-blue {
 		led-power-green {
 			label = "smartgw:power:green";
 			gpios = <&gpio 19 GPIO_ACTIVE_HIGH>;
-			default-state = "off";
+			linux,default-trigger = "timer";
 		};
 
 		led-power-red {
-- 
2.43.0

Re: [PATCH v1 1/3] MIPS: dts: ralink: mt7628a: Fix sysc's compatible property for MT7688

[Stefan Roese]

On 11.06.25 21:47, Ezra Buehler wrote:
> From: Ezra Buehler <ezra.buehler@husqvarnagroup.com>
> 
> Otherwise, the MT7688-based GARDENA smart Gateway will fail to boot
> printing "Kernel panic - not syncing: unable to get CPU clock, err=-2".
> 
> Signed-off-by: Ezra Buehler <ezra.buehler@husqvarnagroup.com>

Reviewed-by: Stefan Roese <sr@denx.de>

Thanks,
Stefan

> ---
>   arch/mips/boot/dts/ralink/mt7628a.dtsi | 2 +-
>   1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/arch/mips/boot/dts/ralink/mt7628a.dtsi b/arch/mips/boot/dts/ralink/mt7628a.dtsi
> index 0212700c4fb4..10221a41f02a 100644
> --- a/arch/mips/boot/dts/ralink/mt7628a.dtsi
> +++ b/arch/mips/boot/dts/ralink/mt7628a.dtsi
> @@ -33,7 +33,7 @@ palmbus@10000000 {
>   		#size-cells = <1>;
>   
>   		sysc: syscon@0 {
> -			compatible = "ralink,mt7628-sysc", "syscon";
> +			compatible = "ralink,mt7628-sysc", "ralink,mt7688-sysc", "syscon";
>   			reg = <0x0 0x60>;
>   			#clock-cells = <1>;
>   			#reset-cells = <1>;

Viele Grüße,
Stefan Roese

-- 
DENX Software Engineering GmbH,      Managing Director: Erika Unter
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-51 Fax: (+49)-8142-66989-80 Email: sr@denx.de

Re: [PATCH v1 2/3] MIPS: dts: ralink: mt7628a: Update watchdog node according to bindings

[Stefan Roese]

On 11.06.25 21:47, Ezra Buehler wrote:
> From: Ezra Buehler <ezra.buehler@husqvarnagroup.com>
> 
> Most notably, add the mediatek,sysctl phandle and remove the redundant
> reset/interrupt-related properties from the watchdog node. This is in
> line with the corresponding devicetree (mt7628an.dtsi) used by the
> OpenWrt project.
> 
> This has been tested on the MT7688-based GARDENA smart Gateway.
> 
> Signed-off-by: Ezra Buehler <ezra.buehler@husqvarnagroup.com>

Reviewed-by: Stefan Roese <sr@denx.de>

Thanks,
Stefan

> ---
>   arch/mips/boot/dts/ralink/mt7628a.dtsi | 9 ++-------
>   1 file changed, 2 insertions(+), 7 deletions(-)
> 
> diff --git a/arch/mips/boot/dts/ralink/mt7628a.dtsi b/arch/mips/boot/dts/ralink/mt7628a.dtsi
> index 10221a41f02a..5d7a6cfa9e2b 100644
> --- a/arch/mips/boot/dts/ralink/mt7628a.dtsi
> +++ b/arch/mips/boot/dts/ralink/mt7628a.dtsi
> @@ -134,13 +134,8 @@ pinmux_p4led_an_gpio: p4led-an-gpio-pins {
>   
>   		watchdog: watchdog@100 {
>   			compatible = "mediatek,mt7621-wdt";
> -			reg = <0x100 0x30>;
> -
> -			resets = <&sysc 8>;
> -			reset-names = "wdt";
> -
> -			interrupt-parent = <&intc>;
> -			interrupts = <24>;
> +			reg = <0x100 0x100>;
> +			mediatek,sysctl = <&sysc>;
>   
>   			status = "disabled";
>   		};

Viele Grüße,
Stefan Roese

-- 
DENX Software Engineering GmbH,      Managing Director: Erika Unter
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-51 Fax: (+49)-8142-66989-80 Email: sr@denx.de

Re: [PATCH v1 3/3] MIPS: dts: ralink: gardena_smart_gateway_mt7688: Fix power LED

[Stefan Roese]

On 11.06.25 21:47, Ezra Buehler wrote:
> From: Ezra Buehler <ezra.buehler@husqvarnagroup.com>
> 
> When starting up, the GARDENA smart Gateway's power LED should be
> flashing green. It is unclear why it was initially set to "off".
> 
> The LED frequency cannot be configured in the devicetree. Luckily, the
> default is 1 Hz, which is what we want.
> 
> Signed-off-by: Ezra Buehler <ezra.buehler@husqvarnagroup.com>

Reviewed-by: Stefan Roese <sr@denx.de>

Thanks,
Stefan

> ---
>   arch/mips/boot/dts/ralink/gardena_smart_gateway_mt7688.dts | 2 +-
>   1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/arch/mips/boot/dts/ralink/gardena_smart_gateway_mt7688.dts b/arch/mips/boot/dts/ralink/gardena_smart_gateway_mt7688.dts
> index 7743d014631a..0bfb1dde9764 100644
> --- a/arch/mips/boot/dts/ralink/gardena_smart_gateway_mt7688.dts
> +++ b/arch/mips/boot/dts/ralink/gardena_smart_gateway_mt7688.dts
> @@ -56,7 +56,7 @@ led-power-blue {
>   		led-power-green {
>   			label = "smartgw:power:green";
>   			gpios = <&gpio 19 GPIO_ACTIVE_HIGH>;
> -			default-state = "off";
> +			linux,default-trigger = "timer";
>   		};
>   
>   		led-power-red {

Viele Grüße,
Stefan Roese

-- 
DENX Software Engineering GmbH,      Managing Director: Erika Unter
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-51 Fax: (+49)-8142-66989-80 Email: sr@denx.de

Re: [PATCH v1 1/3] MIPS: dts: ralink: mt7628a: Fix sysc's compatible property for MT7688

[Krzysztof Kozlowski]

On 11/06/2025 21:47, Ezra Buehler wrote:
> From: Ezra Buehler <ezra.buehler@husqvarnagroup.com>
> 
> Otherwise, the MT7688-based GARDENA smart Gateway will fail to boot
> printing "Kernel panic - not syncing: unable to get CPU clock, err=-2".
> 
> Signed-off-by: Ezra Buehler <ezra.buehler@husqvarnagroup.com>
> ---
>  arch/mips/boot/dts/ralink/mt7628a.dtsi | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/arch/mips/boot/dts/ralink/mt7628a.dtsi b/arch/mips/boot/dts/ralink/mt7628a.dtsi
> index 0212700c4fb4..10221a41f02a 100644
> --- a/arch/mips/boot/dts/ralink/mt7628a.dtsi
> +++ b/arch/mips/boot/dts/ralink/mt7628a.dtsi
> @@ -33,7 +33,7 @@ palmbus@10000000 {
>  		#size-cells = <1>;
>  
>  		sysc: syscon@0 {
> -			compatible = "ralink,mt7628-sysc", "syscon";
> +			compatible = "ralink,mt7628-sysc", "ralink,mt7688-sysc", "syscon";
This is in contradiction to bindings, so you need to fix bindings first
- with proper justification. If this happened in separate patchset, then
the DTS thread MUST provide lore link to that.

Best regards,
Krzysztof

Re: [PATCH v1 1/3] MIPS: dts: ralink: mt7628a: Fix sysc's compatible property for MT7688

[Ezra Buehler]

On Thu, Jun 12, 2025 at 1:22 PM Krzysztof Kozlowski <krzk@kernel.org> wrote:
>
> On 11/06/2025 21:47, Ezra Buehler wrote:
> > From: Ezra Buehler <ezra.buehler@husqvarnagroup.com>
> >
> > Otherwise, the MT7688-based GARDENA smart Gateway will fail to boot
> > printing "Kernel panic - not syncing: unable to get CPU clock, err=-2".
> >
> > Signed-off-by: Ezra Buehler <ezra.buehler@husqvarnagroup.com>
> > ---
> >  arch/mips/boot/dts/ralink/mt7628a.dtsi | 2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/arch/mips/boot/dts/ralink/mt7628a.dtsi b/arch/mips/boot/dts/ralink/mt7628a.dtsi
> > index 0212700c4fb4..10221a41f02a 100644
> > --- a/arch/mips/boot/dts/ralink/mt7628a.dtsi
> > +++ b/arch/mips/boot/dts/ralink/mt7628a.dtsi
> > @@ -33,7 +33,7 @@ palmbus@10000000 {
> >               #size-cells = <1>;
> >
> >               sysc: syscon@0 {
> > -                     compatible = "ralink,mt7628-sysc", "syscon";
> > +                     compatible = "ralink,mt7628-sysc", "ralink,mt7688-sysc", "syscon";
> This is in contradiction to bindings, so you need to fix bindings first
> - with proper justification. If this happened in separate patchset, then
> the DTS thread MUST provide lore link to that.

As the MT7628 and MT7688 are identical in most respects, mt7628a.dtsi is
used for both SoCs. Therefore, I'd rather suggest adapting the driver to
return "ralink,mt7628-sysc" in both cases and remove "ralink,mt7688-sysc"
from the DT bindings.

I'd love to hear Sergio's (or any other) opinion on this matter.

Cheers,
Ezra

Re: [PATCH v1 1/3] MIPS: dts: ralink: mt7628a: Fix sysc's compatible property for MT7688

[Krzysztof Kozlowski]

On 13/06/2025 13:50, Ezra Buehler wrote:
>>> diff --git a/arch/mips/boot/dts/ralink/mt7628a.dtsi b/arch/mips/boot/dts/ralink/mt7628a.dtsi
>>> index 0212700c4fb4..10221a41f02a 100644
>>> --- a/arch/mips/boot/dts/ralink/mt7628a.dtsi
>>> +++ b/arch/mips/boot/dts/ralink/mt7628a.dtsi
>>> @@ -33,7 +33,7 @@ palmbus@10000000 {
>>>               #size-cells = <1>;
>>>
>>>               sysc: syscon@0 {
>>> -                     compatible = "ralink,mt7628-sysc", "syscon";
>>> +                     compatible = "ralink,mt7628-sysc", "ralink,mt7688-sysc", "syscon";
>> This is in contradiction to bindings, so you need to fix bindings first
>> - with proper justification. If this happened in separate patchset, then
>> the DTS thread MUST provide lore link to that.
> 
> As the MT7628 and MT7688 are identical in most respects, mt7628a.dtsi is
> used for both SoCs. Therefore, I'd rather suggest adapting the driver to
> return "ralink,mt7628-sysc" in both cases and remove "ralink,mt7688-sysc"
> from the DT bindings.


Not sure what you propose here, but just in case: devices need specific
compatibles.

> 
> I'd love to hear Sergio's (or any other) opinion on this matter.
> 
> Cheers,
> Ezra


Best regards,
Krzysztof

Re: [PATCH v1 1/3] MIPS: dts: ralink: mt7628a: Fix sysc's compatible property for MT7688

[Ezra Buehler]

On Fri, Jun 13, 2025 at 2:11 PM Krzysztof Kozlowski <krzk@kernel.org> wrote:
> Not sure what you propose here, but just in case: devices need specific
> compatibles.

Something like this

diff --git a/arch/mips/ralink/clk.c b/arch/mips/ralink/clk.c
index 9db73fcac522..7ce9acec3a01 100644
--- a/arch/mips/ralink/clk.c
+++ b/arch/mips/ralink/clk.c
@@ -48,11 +48,9 @@ static const char *clk_cpu(int *idx)
                *idx = 2;
                return "ralink,mt7620-sysc";
        case MT762X_SOC_MT7628AN:
-               *idx = 1;
-               return "ralink,mt7628-sysc";
        case MT762X_SOC_MT7688:
                *idx = 1;
-               return "ralink,mt7688-sysc";
+               return "ralink,mt7628-sysc";
        default:
                *idx = -1;
                return "invalid";

while leaving mt7628a.dtsi as it is, resolves the issue.

Is this not an option?

Cheers,
Ezra

Re: [PATCH v1 1/3] MIPS: dts: ralink: mt7628a: Fix sysc's compatible property for MT7688

[Krzysztof Kozlowski]

On 13/06/2025 14:20, Ezra Buehler wrote:
> On Fri, Jun 13, 2025 at 2:11 PM Krzysztof Kozlowski <krzk@kernel.org> wrote:
>> Not sure what you propose here, but just in case: devices need specific
>> compatibles.
> 
> Something like this
> 
> diff --git a/arch/mips/ralink/clk.c b/arch/mips/ralink/clk.c
> index 9db73fcac522..7ce9acec3a01 100644
> --- a/arch/mips/ralink/clk.c
> +++ b/arch/mips/ralink/clk.c
> @@ -48,11 +48,9 @@ static const char *clk_cpu(int *idx)
>                 *idx = 2;
>                 return "ralink,mt7620-sysc";
>         case MT762X_SOC_MT7628AN:
> -               *idx = 1;
> -               return "ralink,mt7628-sysc";
>         case MT762X_SOC_MT7688:
>                 *idx = 1;
> -               return "ralink,mt7688-sysc";
> +               return "ralink,mt7628-sysc";
>         default:
>                 *idx = -1;
>                 return "invalid";
> 
> while leaving mt7628a.dtsi as it is, resolves the issue.
> 
> Is this not an option?
That's not a binding, but driver, so obviously it is fine since you are
not removing it from bindings.

Best regards,
Krzysztof

Re: [PATCH v1 1/3] MIPS: dts: ralink: mt7628a: Fix sysc's compatible property for MT7688

[Ezra Buehler]

On Fri, Jun 13, 2025 at 2:41 PM Krzysztof Kozlowski <krzk@kernel.org> wrote:
> That's not a binding, but driver, so obviously it is fine since you are
> not removing it from bindings.

And, if we also remove all occurrences of "ralink,mt7688-sysc" from the
code, as it is not needed from a technical standpoint, can we remove it
from mediatek,mtmips-sysc.yaml or is there no going back?

Cheers,
Ezra.

Re: [PATCH v1 1/3] MIPS: dts: ralink: mt7628a: Fix sysc's compatible property for MT7688

[Krzysztof Kozlowski]

On 13/06/2025 14:56, Ezra Buehler wrote:
> On Fri, Jun 13, 2025 at 2:41 PM Krzysztof Kozlowski <krzk@kernel.org> wrote:
>> That's not a binding, but driver, so obviously it is fine since you are
>> not removing it from bindings.
> 
> And, if we also remove all occurrences of "ralink,mt7688-sysc" from the
> code, as it is not needed from a technical standpoint, can we remove it
> from mediatek,mtmips-sysc.yaml or is there no going back?
But it is needed. MT7688 needs it to fulfill the DT bindings requirement
(see writing bindings): specific compatible.

Best regards,
Krzysztof

Re: [PATCH v1 1/3] MIPS: dts: ralink: mt7628a: Fix sysc's compatible property for MT7688

[Ezra Buehler]

On Fri, Jun 13, 2025 at 3:13 PM Krzysztof Kozlowski <krzk@kernel.org> wrote:
> On 13/06/2025 14:56, Ezra Buehler wrote:
> > On Fri, Jun 13, 2025 at 2:41 PM Krzysztof Kozlowski <krzk@kernel.org> wrote:
> >> That's not a binding, but driver, so obviously it is fine since you are
> >> not removing it from bindings.
> >
> > And, if we also remove all occurrences of "ralink,mt7688-sysc" from the
> > code, as it is not needed from a technical standpoint, can we remove it
> > from mediatek,mtmips-sysc.yaml or is there no going back?
> But it is needed. MT7688 needs it to fulfill the DT bindings requirement
> (see writing bindings): specific compatible.

I see. In that case, it seems to me that the only reasonable solution is
to adapt the bindings (mediatek,mtmips-sysc.yaml) and go forward with
the initial patch.

Cheers,
Ezra.

Re: [PATCH v1 2/3] MIPS: dts: ralink: mt7628a: Update watchdog node according to bindings

[Sergio Paracuellos]

On Wed, Jun 11, 2025 at 9:47 PM Ezra Buehler <ezra@easyb.ch> wrote:
>
> From: Ezra Buehler <ezra.buehler@husqvarnagroup.com>
>
> Most notably, add the mediatek,sysctl phandle and remove the redundant
> reset/interrupt-related properties from the watchdog node. This is in
> line with the corresponding devicetree (mt7628an.dtsi) used by the
> OpenWrt project.
>
> This has been tested on the MT7688-based GARDENA smart Gateway.
>
> Signed-off-by: Ezra Buehler <ezra.buehler@husqvarnagroup.com>
> ---
>  arch/mips/boot/dts/ralink/mt7628a.dtsi | 9 ++-------
>  1 file changed, 2 insertions(+), 7 deletions(-)

Reviewed-by: Sergio Paracuellos <sergio.paracuellos@gmail.com>

Re: [PATCH v1 3/3] MIPS: dts: ralink: gardena_smart_gateway_mt7688: Fix power LED

[Sergio Paracuellos]

On Wed, Jun 11, 2025 at 9:47 PM Ezra Buehler <ezra@easyb.ch> wrote:
>
> From: Ezra Buehler <ezra.buehler@husqvarnagroup.com>
>
> When starting up, the GARDENA smart Gateway's power LED should be
> flashing green. It is unclear why it was initially set to "off".
>
> The LED frequency cannot be configured in the devicetree. Luckily, the
> default is 1 Hz, which is what we want.
>
> Signed-off-by: Ezra Buehler <ezra.buehler@husqvarnagroup.com>
> ---
>  arch/mips/boot/dts/ralink/gardena_smart_gateway_mt7688.dts | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)

Reviewed-by: Sergio Paracuellos <sergio.paracuellos@gmail.com>

Re: [PATCH v1 1/3] MIPS: dts: ralink: mt7628a: Fix sysc's compatible property for MT7688

[Sergio Paracuellos]

Hi Ezra,

On Fri, Jun 13, 2025 at 7:40 PM Ezra Buehler <ezra@easyb.ch> wrote:
>
> On Fri, Jun 13, 2025 at 3:13 PM Krzysztof Kozlowski <krzk@kernel.org> wrote:
> > On 13/06/2025 14:56, Ezra Buehler wrote:
> > > On Fri, Jun 13, 2025 at 2:41 PM Krzysztof Kozlowski <krzk@kernel.org> wrote:
> > >> That's not a binding, but driver, so obviously it is fine since you are
> > >> not removing it from bindings.
> > >
> > > And, if we also remove all occurrences of "ralink,mt7688-sysc" from the
> > > code, as it is not needed from a technical standpoint, can we remove it
> > > from mediatek,mtmips-sysc.yaml or is there no going back?
> > But it is needed. MT7688 needs it to fulfill the DT bindings requirement
> > (see writing bindings): specific compatible.
>
> I see. In that case, it seems to me that the only reasonable solution is
> to adapt the bindings (mediatek,mtmips-sysc.yaml) and go forward with
> the initial patch.

Compatibles in ralink are a bit messy historically since there are
platforms that are pretty similar and were developed before having
proper bindings mainlined, For example, as you said, mt7628 and mt7688
are pretty similar but having wildcard mt76x8 was not an option since
specific compatibles are mandatory, so I ended introducing also the
mt7688-sysc in bindings to match MT762X_SOC_MT7688 type [0].

I have checked also openWRT tree and I think just modifying clk.c [1]
returning the expected sysc would be a valid and functional approach:

 diff --git a/arch/mips/ralink/clk.c b/arch/mips/ralink/clk.c
 index 9db73fcac522..7ce9acec3a01 100644
 --- a/arch/mips/ralink/clk.c
 +++ b/arch/mips/ralink/clk.c
 @@ -48,11 +48,9 @@ static const char *clk_cpu(int *idx)
                 *idx = 2;
                  return "ralink,mt7620-sysc";
          case MT762X_SOC_MT7628AN:
                  *idx = 1;
                  return "ralink,mt7628-sysc";
         case MT762X_SOC_MT7688:
                 *idx = 1;
 -               return "ralink,mt7688-sysc";
 +               return "ralink,mt7628-sysc";
         default:
                 *idx = -1;
                 return "invalid";

Since you are not touching bindings at all but driver code, it should be ok.

Obviously if you want to adapt the bindings I will also check them
when you do so.

Thanks,
    Sergio Paracuellos

[0]: https://elixir.bootlin.com/linux/v6.15.1/source/arch/mips/include/asm/mach-ralink/ralink_regs.h#L28
[1]: https://elixir.bootlin.com/linux/v6.15.1/source/arch/mips/ralink/clk.c#L55
>
> Cheers,
> Ezra.

Re: [PATCH v1 1/3] MIPS: dts: ralink: mt7628a: Fix sysc's compatible property for MT7688

[Ezra Buehler]

On Sat, Jun 14, 2025 at 7:01 AM Sergio Paracuellos
<sergio.paracuellos@gmail.com> wrote:
>
> Hi Ezra,
>
> On Fri, Jun 13, 2025 at 7:40 PM Ezra Buehler <ezra@easyb.ch> wrote:
> >
> > On Fri, Jun 13, 2025 at 3:13 PM Krzysztof Kozlowski <krzk@kernel.org> wrote:
> > > On 13/06/2025 14:56, Ezra Buehler wrote:
> > > > On Fri, Jun 13, 2025 at 2:41 PM Krzysztof Kozlowski <krzk@kernel.org> wrote:
> > > >> That's not a binding, but driver, so obviously it is fine since you are
> > > >> not removing it from bindings.
> > > >
> > > > And, if we also remove all occurrences of "ralink,mt7688-sysc" from the
> > > > code, as it is not needed from a technical standpoint, can we remove it
> > > > from mediatek,mtmips-sysc.yaml or is there no going back?
> > > But it is needed. MT7688 needs it to fulfill the DT bindings requirement
> > > (see writing bindings): specific compatible.
> >
> > I see. In that case, it seems to me that the only reasonable solution is
> > to adapt the bindings (mediatek,mtmips-sysc.yaml) and go forward with
> > the initial patch.
>
> Compatibles in ralink are a bit messy historically since there are
> platforms that are pretty similar and were developed before having
> proper bindings mainlined, For example, as you said, mt7628 and mt7688
> are pretty similar but having wildcard mt76x8 was not an option since
> specific compatibles are mandatory, so I ended introducing also the
> mt7688-sysc in bindings to match MT762X_SOC_MT7688 type [0].
>
> I have checked also openWRT tree and I think just modifying clk.c [1]
> returning the expected sysc would be a valid and functional approach:
>
>  diff --git a/arch/mips/ralink/clk.c b/arch/mips/ralink/clk.c
>  index 9db73fcac522..7ce9acec3a01 100644
>  --- a/arch/mips/ralink/clk.c
>  +++ b/arch/mips/ralink/clk.c
>  @@ -48,11 +48,9 @@ static const char *clk_cpu(int *idx)
>                  *idx = 2;
>                   return "ralink,mt7620-sysc";
>           case MT762X_SOC_MT7628AN:
>                   *idx = 1;
>                   return "ralink,mt7628-sysc";
>          case MT762X_SOC_MT7688:
>                  *idx = 1;
>  -               return "ralink,mt7688-sysc";
>  +               return "ralink,mt7628-sysc";
>          default:
>                  *idx = -1;
>                  return "invalid";
>
> Since you are not touching bindings at all but driver code, it should be ok.
>

The issue I see here is that if someone decides to put

compatible = "ralink,mt7688-sysc", "syscon";

into their DT, which seems very reasonable, the system will not boot.
Therefore, I would not go with that change unless one removes support
for ralink,mt7688-sysc all together.

Cheers,
Ezra.

Re: [PATCH v1 1/3] MIPS: dts: ralink: mt7628a: Fix sysc's compatible property for MT7688

[Sergio Paracuellos]

On Sat, Jun 14, 2025 at 7:47 AM Ezra Buehler <ezra@easyb.ch> wrote:
>
> On Sat, Jun 14, 2025 at 7:01 AM Sergio Paracuellos
> <sergio.paracuellos@gmail.com> wrote:
> >
> > Hi Ezra,
> >
> > On Fri, Jun 13, 2025 at 7:40 PM Ezra Buehler <ezra@easyb.ch> wrote:
> > >
> > > On Fri, Jun 13, 2025 at 3:13 PM Krzysztof Kozlowski <krzk@kernel.org> wrote:
> > > > On 13/06/2025 14:56, Ezra Buehler wrote:
> > > > > On Fri, Jun 13, 2025 at 2:41 PM Krzysztof Kozlowski <krzk@kernel.org> wrote:
> > > > >> That's not a binding, but driver, so obviously it is fine since you are
> > > > >> not removing it from bindings.
> > > > >
> > > > > And, if we also remove all occurrences of "ralink,mt7688-sysc" from the
> > > > > code, as it is not needed from a technical standpoint, can we remove it
> > > > > from mediatek,mtmips-sysc.yaml or is there no going back?
> > > > But it is needed. MT7688 needs it to fulfill the DT bindings requirement
> > > > (see writing bindings): specific compatible.
> > >
> > > I see. In that case, it seems to me that the only reasonable solution is
> > > to adapt the bindings (mediatek,mtmips-sysc.yaml) and go forward with
> > > the initial patch.
> >
> > Compatibles in ralink are a bit messy historically since there are
> > platforms that are pretty similar and were developed before having
> > proper bindings mainlined, For example, as you said, mt7628 and mt7688
> > are pretty similar but having wildcard mt76x8 was not an option since
> > specific compatibles are mandatory, so I ended introducing also the
> > mt7688-sysc in bindings to match MT762X_SOC_MT7688 type [0].
> >
> > I have checked also openWRT tree and I think just modifying clk.c [1]
> > returning the expected sysc would be a valid and functional approach:
> >
> >  diff --git a/arch/mips/ralink/clk.c b/arch/mips/ralink/clk.c
> >  index 9db73fcac522..7ce9acec3a01 100644
> >  --- a/arch/mips/ralink/clk.c
> >  +++ b/arch/mips/ralink/clk.c
> >  @@ -48,11 +48,9 @@ static const char *clk_cpu(int *idx)
> >                  *idx = 2;
> >                   return "ralink,mt7620-sysc";
> >           case MT762X_SOC_MT7628AN:
> >                   *idx = 1;
> >                   return "ralink,mt7628-sysc";
> >          case MT762X_SOC_MT7688:
> >                  *idx = 1;
> >  -               return "ralink,mt7688-sysc";
> >  +               return "ralink,mt7628-sysc";
> >          default:
> >                  *idx = -1;
> >                  return "invalid";
> >
> > Since you are not touching bindings at all but driver code, it should be ok.
> >
>
> The issue I see here is that if someone decides to put
>
> compatible = "ralink,mt7688-sysc", "syscon";
>
> into their DT, which seems very reasonable, the system will not boot.
> Therefore, I would not go with that change unless one removes support
> for ralink,mt7688-sysc all together.

Yes, you are right. Properly fixing it would be the correct thing to
do. I was talking only about the already available platforms in the
openWRT tree which seems to not be using that compatible at all.

I'll wait for your patches, then :-).

Thanks,
    Sergio Paracuellos
>
> Cheers,
> Ezra.