Re: [PATCH v2] usb: musb: da8xx: depopulate OF children on teardown

[Greg Kroah-Hartman]

On Wed, Jul 15, 2026 at 09:57:50PM +0800, Pengpeng Hou wrote:
> da8xx_probe() populates OF child devices before registering the MUSB
> platform device. If the later MUSB device registration fails, those
> children remain registered. The remove path also leaves the populated
> children behind.
> 
> Depopulate the children when MUSB device registration fails and after
> unregistering the MUSB device during remove.
> 
> Signed-off-by: Pengpeng Hou <pengpeng@xxxxxxxxxxx>
> ---
> Changes since v1: https://lore.kernel.org/all/20260616005819.11232-1-pengpeng@xxxxxxxxxxx/
> - keep the existing of_platform_populate() error unwind target, so both
>   labels remain used and the patch does not introduce an unused-label
>   warning
> - limit the commit message to paths where population has succeeded
> ---
>  drivers/usb/musb/da8xx.c | 5 ++++-
>  1 file changed, 4 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/usb/musb/da8xx.c b/drivers/usb/musb/da8xx.c
> index 4209f43..0c098a4 100644
> --- a/drivers/usb/musb/da8xx.c
> +++ b/drivers/usb/musb/da8xx.c
> @@ -577,11 +577,13 @@ static int da8xx_probe(struct platform_device *pdev)
>  	ret = PTR_ERR_OR_ZERO(glue->musb);
>  	if (ret) {
>  		dev_err(&pdev->dev, "failed to register musb device: %d\n", ret);
> -		goto err_unregister_phy;
> +		goto err_depopulate;
>  	}
>  
>  	return 0;
>  
> +err_depopulate:
> +	of_platform_depopulate(&pdev->dev);
>  err_unregister_phy:
>  	usb_phy_generic_unregister(glue->usb_phy);
>  	return ret;
> @@ -592,6 +594,7 @@ static void da8xx_remove(struct platform_device *pdev)
>  	struct da8xx_glue		*glue = platform_get_drvdata(pdev);
>  
>  	platform_device_unregister(glue->musb);
> +	of_platform_depopulate(&pdev->dev);

This implies that this has always been a bug for the normal remove path?
How has this never shown up anywhere?

So shouldn't this be 2 different patches, there are 2 different bugs
here, right?

thanks,

greg k-h