Re: [PATCH v2] usb: dwc3: imx8mp: make dwc3_imx_glue_ops static and rename to imx8mp

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:

Pre-existing issues:
- [High] The `dwc3->glue_ops` pointer is assigned after `of_platform_populate()` returns, creating a race condition where role switches can occur before the `pre_set_role` callback is registered.
--

commit b965a813ee62bb063a7c9f8c07f1ddfe408ebc85
Author: Ben Dooks <ben.dooks@codethink.co.uk>

usb: dwc3: imx8mp: make dwc3_imx_glue_ops static and rename to imx8mp

This commit renames dwc3_imx_glue_ops to dwc3_imx8mp_glue_ops and makes
it static to fix a sparse warning about an undeclared symbol.

> diff --git a/drivers/usb/dwc3/dwc3-imx8mp.c b/drivers/usb/dwc3/dwc3-imx8mp.c
> index 1cf96540b66e4..de8c17bc940d2 100644
> --- a/drivers/usb/dwc3/dwc3-imx8mp.c
> +++ b/drivers/usb/dwc3/dwc3-imx8mp.c
[ ... ]
> @@ -266,7 +266,7 @@ static int dwc3_imx8mp_probe(struct platform_device *pdev)
>  		goto put_dwc3;
>  	}
>  
> -	dwc3->glue_ops = &dwc3_imx_glue_ops;
> +	dwc3->glue_ops = &dwc3_imx8mp_glue_ops;

[Severity: High]
This isn't a bug introduced by this patch, but is it possible for a role
switch to occur before glue_ops is registered?

In dwc3_imx8mp_probe(), of_platform_populate() runs earlier and probes the
dwc3 core, which registers its USB role switch. If a concurrent Type-C or
extcon event triggers a role switch before dwc3->glue_ops is assigned here,
dwc3_imx8mp_pre_set_role() won't be called.

Does this leave pm_runtime autosuspend enabled for the host role, potentially
causing the XHCI controller to clear RUN_STOP and disable wakeups so it
becomes unresponsive to device insertions?

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260623101043.748359-1-ben.dooks@codethink.co.uk?part=1