[RFC] staging: greybus: arche still depends on missing USB3613 provider

[RFC] staging: greybus: arche still depends on missing USB3613 provider

[Pengpeng Hou]

Hi,

while auditing conditional provider/header contracts, I noticed that Greybus
Arche still appears to describe a USB3613 provider world that is absent from
current mainline.

drivers/staging/greybus/Kconfig still has:

depends on USB_HSIC_USB3613 || COMPILE_TEST

and drivers/staging/greybus/arche-platform.c still conditionally includes
the USB3613 header and calls usb3613_hub_mode_ctrl() when
CONFIG_USB_HSIC_USB3613 is enabled.  However, the current tree does not appear
to provide include/linux/usb/usb3613.h or a Kconfig provider for
USB_HSIC_USB3613.

I am not sending a patch yet because this is staging/hardware policy sensitive.
The possible directions seem to be:

1. restore or move the USB3613 provider/header if the hardware path is still
   intended;
2. remove the stale USB3613 integration path and rely on the local stub;
3. change the Kconfig dependency to describe only current supported worlds; or
4. keep the contract if an out-of-tree provider is intentionally expected.

Could you advise which direction is expected for Arche?

This is static source/Kconfig/header analysis only.  I have not tested Arche
hardware.

Thanks,
Pengpeng

Re: [RFC] staging: greybus: arche still depends on missing USB3613 provider

[Greg Kroah-Hartman]

On Thu, Jun 25, 2026 at 02:12:08PM +0800, Pengpeng Hou wrote:
> Hi,
> 
> while auditing conditional provider/header contracts, I noticed that Greybus
> Arche still appears to describe a USB3613 provider world that is absent from
> current mainline.
> 
> drivers/staging/greybus/Kconfig still has:
> 
> depends on USB_HSIC_USB3613 || COMPILE_TEST
> 
> and drivers/staging/greybus/arche-platform.c still conditionally includes
> the USB3613 header and calls usb3613_hub_mode_ctrl() when
> CONFIG_USB_HSIC_USB3613 is enabled.  However, the current tree does not appear
> to provide include/linux/usb/usb3613.h or a Kconfig provider for
> USB_HSIC_USB3613.
> 
> I am not sending a patch yet because this is staging/hardware policy sensitive.
> The possible directions seem to be:
> 
> 1. restore or move the USB3613 provider/header if the hardware path is still
>    intended;
> 2. remove the stale USB3613 integration path and rely on the local stub;
> 3. change the Kconfig dependency to describe only current supported worlds; or
> 4. keep the contract if an out-of-tree provider is intentionally expected.
> 
> Could you advise which direction is expected for Arche?

I would just leave it as-is for now.  Hopefully once the greybus code
for the beaglebone devices gets more integrated we can get all of this
code out of staging, which will delete the unused stuff properly.

thanks,

greg k-h