[PATCH v2 2/2] arm64: dts: qcom: sdm845-db845c: Use pad fn instead of defining own

David Heidelberg via B4 Relay posted 2 patches 1 month, 3 weeks ago
[PATCH v2 2/2] arm64: dts: qcom: sdm845-db845c: Use pad fn instead of defining own
Posted by David Heidelberg via B4 Relay 1 month, 3 weeks ago
From: David Heidelberg <david@ixit.cz>

Instead of defining own pad function for master clock, pick one offered
by sdm845 device-tree include.

Reviewed-by: Vladimir Zapolskiy <vladimir.zapolskiy@linaro.org>
Signed-off-by: David Heidelberg <david@ixit.cz>
---
 .../dts/qcom/sdm845-db845c-navigation-mezzanine.dtso     |  2 +-
 arch/arm64/boot/dts/qcom/sdm845-db845c.dts               | 16 ----------------
 2 files changed, 1 insertion(+), 17 deletions(-)

diff --git a/arch/arm64/boot/dts/qcom/sdm845-db845c-navigation-mezzanine.dtso b/arch/arm64/boot/dts/qcom/sdm845-db845c-navigation-mezzanine.dtso
index dbe1911d8e470..f6c2c98970d76 100644
--- a/arch/arm64/boot/dts/qcom/sdm845-db845c-navigation-mezzanine.dtso
+++ b/arch/arm64/boot/dts/qcom/sdm845-db845c-navigation-mezzanine.dtso
@@ -40,7 +40,7 @@ camera@10 {
 		/* CAM0_RST_N */
 		reset-gpios = <&tlmm 9 GPIO_ACTIVE_LOW>;
 		pinctrl-names = "default";
-		pinctrl-0 = <&cam0_default>;
+		pinctrl-0 = <&cam0_default &cam_mclk0_default>;
 
 		clocks = <&clock_camcc CAM_CC_MCLK0_CLK>;
 		clock-names = "xvclk";
diff --git a/arch/arm64/boot/dts/qcom/sdm845-db845c.dts b/arch/arm64/boot/dts/qcom/sdm845-db845c.dts
index ce23f87e0316b..518d2ec657b13 100644
--- a/arch/arm64/boot/dts/qcom/sdm845-db845c.dts
+++ b/arch/arm64/boot/dts/qcom/sdm845-db845c.dts
@@ -877,14 +877,6 @@ rst-pins {
 			drive-strength = <16>;
 			bias-disable;
 		};
-
-		mclk0-pins {
-			pins = "gpio13";
-			function = "cam_mclk";
-
-			drive-strength = <16>;
-			bias-disable;
-		};
 	};
 
 	cam3_default: cam3-default-state {
@@ -895,14 +887,6 @@ rst-pins {
 			drive-strength = <16>;
 			bias-disable;
 		};
-
-		mclk3-pins {
-			function = "cam_mclk";
-			pins = "gpio16";
-
-			drive-strength = <16>;
-			bias-disable;
-		};
 	};
 
 	dsi_sw_sel: dsi-sw-sel-state {

-- 
2.51.0
Re: [PATCH v2 2/2] arm64: dts: qcom: sdm845-db845c: Use pad fn instead of defining own
Posted by Konrad Dybcio 1 month, 3 weeks ago
On 12/17/25 12:39 PM, David Heidelberg via B4 Relay wrote:
> From: David Heidelberg <david@ixit.cz>
> 
> Instead of defining own pad function for master clock, pick one offered
> by sdm845 device-tree include.

I'm a little bitter about the wording - the pad function here is "cam_mclk",
whereas what you're doing is inheriting a common pinmux/pincfg node that
refers to that function

[...]

> -		mclk0-pins {
> -			pins = "gpio13";
> -			function = "cam_mclk";
> -
> -			drive-strength = <16>;

This patch changes the drive-strength (16 -> 2 mA)

FWIW it's 2 on reference designs and Sony boards, check your
downstream kernel

Konrad
Re: [PATCH v2 2/2] arm64: dts: qcom: sdm845-db845c: Use pad fn instead of defining own
Posted by David Heidelberg 1 month, 2 weeks ago
On 18/12/2025 13:25, Konrad Dybcio wrote:
> On 12/17/25 12:39 PM, David Heidelberg via B4 Relay wrote:
>> From: David Heidelberg <david@ixit.cz>
>>
>> Instead of defining own pad function for master clock, pick one offered
>> by sdm845 device-tree include.
> 
> I'm a little bitter about the wording - the pad function here is "cam_mclk",
> whereas what you're doing is inheriting a common pinmux/pincfg node that
> refers to that function
> 
> [...]
> 
>> -		mclk0-pins {
>> -			pins = "gpio13";
>> -			function = "cam_mclk";
>> -
>> -			drive-strength = <16>;
> 
> This patch changes the drive-strength (16 -> 2 mA)
> 
> FWIW it's 2 on reference designs and Sony boards, check your
> downstream kernel

I don't have any. I'm fine with this patch not getting applied, but I'll 
CC Robert who added the support, maybe he can verify.

David

> 
> Konrad

-- 
David Heidelberg
Re: [PATCH v2 2/2] arm64: dts: qcom: sdm845-db845c: Use pad fn instead of defining own
Posted by Konrad Dybcio 1 month, 1 week ago
On 12/22/25 3:01 PM, David Heidelberg wrote:
> On 18/12/2025 13:25, Konrad Dybcio wrote:
>> On 12/17/25 12:39 PM, David Heidelberg via B4 Relay wrote:
>>> From: David Heidelberg <david@ixit.cz>
>>>
>>> Instead of defining own pad function for master clock, pick one offered
>>> by sdm845 device-tree include.
>>
>> I'm a little bitter about the wording - the pad function here is "cam_mclk",
>> whereas what you're doing is inheriting a common pinmux/pincfg node that
>> refers to that function
>>
>> [...]
>>
>>> -        mclk0-pins {
>>> -            pins = "gpio13";
>>> -            function = "cam_mclk";
>>> -
>>> -            drive-strength = <16>;
>>
>> This patch changes the drive-strength (16 -> 2 mA)
>>
>> FWIW it's 2 on reference designs and Sony boards, check your
>> downstream kernel
> 
> I don't have any. I'm fine with this patch not getting applied, but I'll CC Robert who added the support, maybe he can verify.

FWIW this can be verified at runtime as well (running downstream) with

cat /sys/kernel/debug/gpio

Konrad