[PATCH 0/2] Modularize SX150X GPIO expander

[PATCH 0/2] Modularize SX150X GPIO expander

[Fange Zhang]

The ANX7625 relies on SX150X-controlled GPIOs for reset and HPD for
the QCS615 Ride board

Signed-off-by: Fange Zhang <fange.zhang@oss.qualcomm.com>
---
Fange Zhang (2):
      pinctrl: sx150x: Make the driver tristate
      arm64: defconfig: Modularize SX150X GPIO expander

 arch/arm64/configs/defconfig     | 1 +
 drivers/pinctrl/Kconfig          | 2 +-
 drivers/pinctrl/pinctrl-sx150x.c | 4 ++++
 3 files changed, 6 insertions(+), 1 deletion(-)
---
base-commit: 1357b2649c026b51353c84ddd32bc963e8999603
change-id: 20250818-modularize-sx150x-gpio-expander-cef7fd0a7511

Best regards,
-- 
Fange Zhang <fange.zhang@oss.qualcomm.com>

[PATCH 1/2] pinctrl: sx150x: Make the driver tristate

[Fange Zhang]

Set PINCTRL_SX150X config option as a tristate and add
MODULE_DEVICE_TABLE()/MODULE_LICENSE() to export appropriate information.

Signed-off-by: Fange Zhang <fange.zhang@oss.qualcomm.com>
---
 drivers/pinctrl/Kconfig          | 2 +-
 drivers/pinctrl/pinctrl-sx150x.c | 4 ++++
 2 files changed, 5 insertions(+), 1 deletion(-)

diff --git a/drivers/pinctrl/Kconfig b/drivers/pinctrl/Kconfig
index ddd11668457ceded9861438768f89945f2307e31..3d06f02a3da549317a9364540d1b598051b926f4 100644
--- a/drivers/pinctrl/Kconfig
+++ b/drivers/pinctrl/Kconfig
@@ -550,7 +550,7 @@ config PINCTRL_STMFX
 	  interrupt-controller.
 
 config PINCTRL_SX150X
-	bool "Semtech SX150x I2C GPIO expander pinctrl driver"
+	tristate "Semtech SX150x I2C GPIO expander pinctrl driver"
 	depends on I2C=y
 	select PINMUX
 	select PINCONF
diff --git a/drivers/pinctrl/pinctrl-sx150x.c b/drivers/pinctrl/pinctrl-sx150x.c
index 53cf8168b274c5310706512dcd3ba8d4d25f5049..b613568b42b73d1c1b220b641e292055e23f4de2 100644
--- a/drivers/pinctrl/pinctrl-sx150x.c
+++ b/drivers/pinctrl/pinctrl-sx150x.c
@@ -863,6 +863,7 @@ static const struct of_device_id sx150x_of_match[] = {
 	{ .compatible = "semtech,sx1509q", .data = &sx1509q_device_data },
 	{},
 };
+MODULE_DEVICE_TABLE(of, sx150x_of_match);
 
 static int sx150x_reset(struct sx150x_pinctrl *pctl)
 {
@@ -1266,3 +1267,6 @@ static int __init sx150x_init(void)
 	return i2c_add_driver(&sx150x_driver);
 }
 subsys_initcall(sx150x_init);
+
+MODULE_DESCRIPTION("Semtech SX150x I2C GPIO expander pinctrl driver");
+MODULE_LICENSE("GPL");

-- 
2.34.1

Re: [PATCH 1/2] pinctrl: sx150x: Make the driver tristate

[Linus Walleij]

On Mon, Aug 18, 2025 at 6:41 AM Fange Zhang
<fange.zhang@oss.qualcomm.com> wrote:

> Set PINCTRL_SX150X config option as a tristate and add
> MODULE_DEVICE_TABLE()/MODULE_LICENSE() to export appropriate information.
>
> Signed-off-by: Fange Zhang <fange.zhang@oss.qualcomm.com>

Fair enough!

This patch 1/2 applied to the pinctrl tree.

Please funnel patch 2/2 through the SoC tree.

> subsys_initcall(sx150x_init

I seriously wonder about this. It feels like this driver should
be a pure module_init() device. But other users may disagree.

Yours,
Linus Walleij

Re: [PATCH 1/2] pinctrl: sx150x: Make the driver tristate

[Fange Zhang]

On 8/19/2025 8:55 PM, Linus Walleij wrote:
> On Mon, Aug 18, 2025 at 6:41 AM Fange Zhang
> <fange.zhang@oss.qualcomm.com> wrote:
> 
>> Set PINCTRL_SX150X config option as a tristate and add
>> MODULE_DEVICE_TABLE()/MODULE_LICENSE() to export appropriate information.
>>
>> Signed-off-by: Fange Zhang <fange.zhang@oss.qualcomm.com>
> 
> Fair enough!
> 
> This patch 1/2 applied to the pinctrl tree.

Thanks

> 
> Please funnel patch 2/2 through the SoC tree.

Hi Bjorn, Konrad,

Would it be acceptable to include this defconfig change, or is there a 
preferred process I should follow to get it reviewed and accepted?

Thanks,
Fange Zhang

> 
>> subsys_initcall(sx150x_init
> 
> I seriously wonder about this. It feels like this driver should
> be a pure module_init() device. But other users may disagree.
> 
> Yours,
> Linus Walleij

Re: [PATCH 1/2] pinctrl: sx150x: Make the driver tristate

[Krzysztof Kozlowski]

On 20/08/2025 04:55, Fange Zhang wrote:
> 
> 
> On 8/19/2025 8:55 PM, Linus Walleij wrote:
>> On Mon, Aug 18, 2025 at 6:41 AM Fange Zhang
>> <fange.zhang@oss.qualcomm.com> wrote:
>>
>>> Set PINCTRL_SX150X config option as a tristate and add
>>> MODULE_DEVICE_TABLE()/MODULE_LICENSE() to export appropriate information.
>>>
>>> Signed-off-by: Fange Zhang <fange.zhang@oss.qualcomm.com>
>>
>> Fair enough!
>>
>> This patch 1/2 applied to the pinctrl tree.
> 
> Thanks
> 
>>
>> Please funnel patch 2/2 through the SoC tree.
> 
> Hi Bjorn, Konrad,
> 
> Would it be acceptable to include this defconfig change, or is there a 
> preferred process I should follow to get it reviewed and accepted?

You need to send it to them... Cc-ing people on random email does not
solve the problem. You need to send the patch to proper maintainers.

Best regards,
Krzysztof

Re: [PATCH 1/2] pinctrl: sx150x: Make the driver tristate

[Fange Zhang]

On 8/20/2025 2:18 PM, Krzysztof Kozlowski wrote:
> On 20/08/2025 04:55, Fange Zhang wrote:
>>
>>
>> On 8/19/2025 8:55 PM, Linus Walleij wrote:
>>> On Mon, Aug 18, 2025 at 6:41 AM Fange Zhang
>>> <fange.zhang@oss.qualcomm.com> wrote:
>>>
>>>> Set PINCTRL_SX150X config option as a tristate and add
>>>> MODULE_DEVICE_TABLE()/MODULE_LICENSE() to export appropriate information.
>>>>
>>>> Signed-off-by: Fange Zhang <fange.zhang@oss.qualcomm.com>
>>>
>>> Fair enough!
>>>
>>> This patch 1/2 applied to the pinctrl tree.
>>
>> Thanks
>>
>>>
>>> Please funnel patch 2/2 through the SoC tree.
>>
>> Hi Bjorn, Konrad,
>>
>> Would it be acceptable to include this defconfig change, or is there a
>> preferred process I should follow to get it reviewed and accepted?
> 
> You need to send it to them... Cc-ing people on random email does not
> solve the problem. You need to send the patch to proper maintainers.

Got it, will update in next patch

> 
> Best regards,
> Krzysztof

[PATCH 2/2] arm64: defconfig: Modularize SX150X GPIO expander

[Fange Zhang]

Modularize the SX150X GPIO expander which is equipped on the QCS615 Ride
board to allow the ANX7625 external bridge to function correctly.

Signed-off-by: Fange Zhang <fange.zhang@oss.qualcomm.com>
---
 arch/arm64/configs/defconfig | 1 +
 1 file changed, 1 insertion(+)

diff --git a/arch/arm64/configs/defconfig b/arch/arm64/configs/defconfig
index 667c4859ecc3cda9b0a3dae7abdd8587fc4f8dbe..dce4450fa9fa58fab0556a6e84b899a3306b4db3 100644
--- a/arch/arm64/configs/defconfig
+++ b/arch/arm64/configs/defconfig
@@ -598,6 +598,7 @@ CONFIG_PINCTRL_DA9062=m
 CONFIG_PINCTRL_MAX77620=y
 CONFIG_PINCTRL_RK805=m
 CONFIG_PINCTRL_SINGLE=y
+CONFIG_PINCTRL_SX150X=m
 CONFIG_PINCTRL_OWL=y
 CONFIG_PINCTRL_S700=y
 CONFIG_PINCTRL_S900=y

-- 
2.34.1

Re: [PATCH 2/2] arm64: defconfig: Modularize SX150X GPIO expander

[Krzysztof Kozlowski]

On 18/08/2025 06:41, Fange Zhang wrote:
> Modularize the SX150X GPIO expander which is equipped on the QCS615 Ride

Qualcomm QCS615 Ride

You are changing defconfig for all platforms, it's not your personal or
company defconfig.

> 


Best regards,
Krzysztof

Re: [PATCH 2/2] arm64: defconfig: Modularize SX150X GPIO expander

[Fange Zhang]

On 8/18/2025 2:33 PM, Krzysztof Kozlowski wrote:
> On 18/08/2025 06:41, Fange Zhang wrote:
>> Modularize the SX150X GPIO expander which is equipped on the QCS615 Ride
> 
> Qualcomm QCS615 Ride
> 
> You are changing defconfig for all platforms, it's not your personal or
> company defconfig.

Thank you for the feedback. Yes, The change is intended to support the 
Qualcomm QCS615 Ride platform, which is now upstream-ready and actively 
maintained.
I believe enabling it as a module ensures it's available without 
impacting other platforms. While it's not currently used directly in the 
kernel, having it built as a module allows flexibility for platforms 
like QCS615 that may require it in future use cases.

> 
>>
> 
> 
> Best regards,
> Krzysztof

Re: [PATCH 2/2] arm64: defconfig: Modularize SX150X GPIO expander

[Krzysztof Kozlowski]

On 18/08/2025 09:10, Fange Zhang wrote:
> 
> 
> On 8/18/2025 2:33 PM, Krzysztof Kozlowski wrote:
>> On 18/08/2025 06:41, Fange Zhang wrote:
>>> Modularize the SX150X GPIO expander which is equipped on the QCS615 Ride
>>
>> Qualcomm QCS615 Ride
>>
>> You are changing defconfig for all platforms, it's not your personal or
>> company defconfig.
> 
> Thank you for the feedback. Yes, The change is intended to support the 
> Qualcomm QCS615 Ride platform, which is now upstream-ready and actively 
> maintained.
> I believe enabling it as a module ensures it's available without 
> impacting other platforms. While it's not currently used directly in the 
> kernel, having it built as a module allows flexibility for platforms 

Hm? Not used? So this rationale is incorrect.

> like QCS615 that may require it in future use cases.

No, we do not make defconfig changes for future.


Best regards,
Krzysztof

Re: [PATCH 2/2] arm64: defconfig: Modularize SX150X GPIO expander

[Fange Zhang]

On 8/18/2025 4:02 PM, Krzysztof Kozlowski wrote:
> On 18/08/2025 09:10, Fange Zhang wrote:
>>
>>
>> On 8/18/2025 2:33 PM, Krzysztof Kozlowski wrote:
>>> On 18/08/2025 06:41, Fange Zhang wrote:
>>>> Modularize the SX150X GPIO expander which is equipped on the QCS615 Ride
>>>
>>> Qualcomm QCS615 Ride
>>>
>>> You are changing defconfig for all platforms, it's not your personal or
>>> company defconfig.
>>
>> Thank you for the feedback. Yes, The change is intended to support the
>> Qualcomm QCS615 Ride platform, which is now upstream-ready and actively
>> maintained.
>> I believe enabling it as a module ensures it's available without
>> impacting other platforms. While it's not currently used directly in the
>> kernel, having it built as a module allows flexibility for platforms
> 
> Hm? Not used? So this rationale is incorrect.

Yes, still not used yet, but we're actively working to upstream full 
support for the QCS615 platform, and this change is part of that effort.
(see 
https://lore.kernel.org/all/20250818-add-display-support-for-qcs615-platform-v6-2-62aad5138a78@oss.qualcomm.com/)

> 
>> like QCS615 that may require it in future use cases.
> 
> No, we do not make defconfig changes for future.

Thank you for the clarification — I understand and respect the policy.
That said, would it be acceptable to consider the defconfig change 
together with the platform support patch, once it's accepted?

> 
> 
> Best regards,
> Krzysztof