[PATCH] ARM: dts: imx6qdl-dhcom: Add USB overcurrent pin on SoM layer

Christoph Niedermaier cniedermaier at dh-electronics.com
Thu Dec 9 23:58:15 PST 2021


From: Marek Vasut [mailto:marex at denx.de]
Sent: Thursday, December 9, 2021 11:26 PM
> On 12/9/21 10:54, Christoph Niedermaier wrote:
>> From: Marek Vasut
>> Sent: Thursday, December 9, 2021 1:23 AM
>>>
>>> On 12/8/21 16:15, Christoph Niedermaier wrote:
>>>> Add USB overcurrent pin muxing on SoM layer, but disable it
>>>> by default. If a board has connected this pin like the
>>>> picoITX, this property should be removed in the board file.
>>>>
>>>> Signed-off-by: Christoph Niedermaier <cniedermaier at dh-electronics.com>
>>>> Cc: Shawn Guo <shawnguo at kernel.org>
>>>> Cc: Fabio Estevam <festevam at gmail.com>
>>>> Cc: Marek Vasut <marex at denx.de>
>>>> Cc: NXP Linux Team <linux-imx at nxp.com>
>>>> Cc: kernel at dh-electronics.com
>>>> To: linux-arm-kernel at lists.infradead.org
>>>> ---
>>>>    arch/arm/boot/dts/imx6qdl-dhcom-picoitx.dtsi | 4 ++++
>>>>    arch/arm/boot/dts/imx6qdl-dhcom-som.dtsi     | 2 ++
>>>>    2 files changed, 6 insertions(+)
>>>>
>>>> diff --git a/arch/arm/boot/dts/imx6qdl-dhcom-picoitx.dtsi b/arch/arm/boot/dts/imx6qdl-dhcom-picoitx.dtsi
>>>> index 4cd4cb9543c8..a67682bfe7bd 100644
>>>> --- a/arch/arm/boot/dts/imx6qdl-dhcom-picoitx.dtsi
>>>> +++ b/arch/arm/boot/dts/imx6qdl-dhcom-picoitx.dtsi
>>>> @@ -48,6 +48,10 @@
>>>>                "", "", "", "", "", "", "", "";
>>>>    };
>>>>
>>>> +&usbh1 { /* USB overcurrent pin is connected */
>>>> +     /delete-property/ disable-over-current;
>>>> +};
>>>> +
>>>>    &iomuxc {
>>>>        pinctrl-0 = <
>>>>                        /*
>>>> diff --git a/arch/arm/boot/dts/imx6qdl-dhcom-som.dtsi b/arch/arm/boot/dts/imx6qdl-dhcom-som.dtsi
>>>> index 5d10c40313cb..e4fdce016c34 100644
>>>> --- a/arch/arm/boot/dts/imx6qdl-dhcom-som.dtsi
>>>> +++ b/arch/arm/boot/dts/imx6qdl-dhcom-som.dtsi
>>>> @@ -385,6 +385,7 @@
>>>>    };
>>>>
>>>>    &usbh1 {
>>>> +     disable-over-current;
>>>>        dr_mode = "host";
>>>>        pinctrl-0 = <&pinctrl_usbh1>;
>>>>        pinctrl-names = "default";
>>>> @@ -728,6 +729,7 @@
>>>>        pinctrl_usbh1: usbh1-grp {
>>>>                fsl,pins = <
>>>>                        MX6QDL_PAD_EIM_D31__GPIO3_IO31          0x120b0
>>>> +                     MX6QDL_PAD_EIM_D30__USB_H1_OC           0x1b0b1
>>>>                >;
>>>>        };
>>>
>>> Shouldn't this be the other way around -- boards with unused USB
>>> overcurrent detection should disable it ? In this case, PDK2 and DRC02
>>> should add the 'disable-over-current' DT property and pinmux, while
>>> picoitx should add the USB OC pinmux .
>>
>> This pin is defined by the DHCOM standard, therefore no other function
>> can be used on this pin. That is why it should be configured in the SoM
>> file.
> 
> Then the pinmux in the SoM dtsi is OK.
> 
>> The first internal version was the other way around, but then we had a
>> discussion about accidentally enabling the overcurrent detection, because
>> it is enabled by default. Since most customers do not use overcurrent
>> detection, we have decided to disable it by default. So if a customer
>> uses that pin, he has to actively remove the DT property as for example
>> shown in the PicoITX board file. In this way, the source of issues should
>> be reduced.
> 
> It seems to me that if the SoM has a dedicated pin for USB OC, then the
> SoM dtsi should keep the default configuration of USB OC (i.e. enabled).
> If a board does not use the USB OC (e.g. because there is a USB hub on
> it), then the board should add the 'disable-over-current' property,
> because this is clearly a board property, not a SoM property.
> 
> Besides, on systems without a USB hub, you likely want to make sure the
> OC detection is not accidentally forgotten disabled, as that might lead
> to damage to the port.
> 
> So I would say, keep the pinmux settings in the SoM dtsi, and add
> disable-over-current property on board level dts.

I am with you, it is a board property. But I don't want to enable it by
default, because here I rate the accidental damage of the port higher.
So if you need it you can enable it on board layer. Because of the negative
logic have to do it this way. What is the argument against it?

Regards
Christoph


More information about the linux-arm-kernel mailing list