Re: [PATCH 1/2] phy: qcom: phy-qcom-snps-eusb2: Make repeater optional

[Dmitry Baryshkov]

On Wed, Mar 19, 2025 at 11:35:00AM +0100, Stephan Gerhold wrote:
> On Wed, Mar 19, 2025 at 12:20:07PM +0200, Dmitry Baryshkov wrote:
> > On Tue, Mar 18, 2025 at 10:22:56PM -0500, Bjorn Andersson via B4 Relay wrote:
> > > From: Bjorn Andersson <bjorn.andersson@xxxxxxxxxxxxxxxx>
> > > 
> > > In a multiport configuration based on the SNPS eUSB2 PHY it's not
> > > necessary that all ports are connected to something.
> > > 
> > > While this is allowed by the Devicetree binding, the implementation
> > > current fails probing for such PHYs, which also prevents the multiport
> > > controller from probing.
> > > 
> > > The lack of repeater does not alter the fact that the PHY is there and
> > > attempts at describing only the used PHYs in Devicetree results in
> > > failures to initialize the USB controller.
> > > 
> > > Make the repeater optional, to allow the these PHYs to be described in
> > > the DeviceTree and for the associated multiport controller to operate
> > > the other ports.
> > > 
> > > Signed-off-by: Bjorn Andersson <bjorn.andersson@xxxxxxxxxxxxxxxx>
> > > ---
> > >  drivers/phy/qualcomm/phy-qcom-snps-eusb2.c | 10 +++++++---
> > >  1 file changed, 7 insertions(+), 3 deletions(-)
> > > 
> > > diff --git a/drivers/phy/qualcomm/phy-qcom-snps-eusb2.c b/drivers/phy/qualcomm/phy-qcom-snps-eusb2.c
> > > index 1484691a41d59a7eaf257ef44300827c668bf7e0..8897d2072ccfcaa5b4a510c17761dcdeed5bad0f 100644
> > > --- a/drivers/phy/qualcomm/phy-qcom-snps-eusb2.c
> > > +++ b/drivers/phy/qualcomm/phy-qcom-snps-eusb2.c
> > > @@ -401,9 +401,13 @@ static int qcom_snps_eusb2_hsphy_probe(struct platform_device *pdev)
> > >  				     "failed to get regulator supplies\n");
> > >  
> > >  	phy->repeater = devm_of_phy_get_by_index(dev, np, 0);
> > > -	if (IS_ERR(phy->repeater))
> > > -		return dev_err_probe(dev, PTR_ERR(phy->repeater),
> > > -				     "failed to get repeater\n");
> > > +	if (IS_ERR(phy->repeater)) {
> > > +		if (PTR_ERR(phy->repeater) == -ENODEV)
> > > +			phy->repeater = NULL;
> > > +		else
> > > +			return dev_err_probe(dev, PTR_ERR(phy->repeater),
> > > +					     "failed to get repeater\n");
> > 
> > Can you use devm_of_phy_optional_get() or devm_phy_optional_get()
> > instead?
> > 
> 
> There is such a patch from Ivaylo already [1].

I will respond there.

> 
> @Ivaylo: Are you planning to re-spin that patch set? Might be even worth
> putting that patch first / sending it separately, since Neil pointed out
> there that the bindings already have the repeater as non-required.
> 
> Thanks,
> Stephan
> 
> [1]: https://lore.kernel.org/linux-arm-msm/20250223122227.725233-6-ivo.ivanov.ivanov1@xxxxxxxxx/

-- 
With best wishes
Dmitry