On Thu, Mar 28 2019, Sergio Paracuellos wrote: > Device tree has been simplified to don't use child nodes and use > the #phy-cells property instead. Change the driver accordly implementing > custom 'xlate' function to return the correct phy for each port. > > Signed-off-by: Sergio Paracuellos > --- > .../staging/mt7621-pci-phy/pci-mt7621-phy.c | 26 ++++++++++++++++--- > 1 file changed, 22 insertions(+), 4 deletions(-) > > diff --git a/drivers/staging/mt7621-pci-phy/pci-mt7621-phy.c b/drivers/staging/mt7621-pci-phy/pci-mt7621-phy.c > index 98c06308143c..557e6a53fc1e 100644 > --- a/drivers/staging/mt7621-pci-phy/pci-mt7621-phy.c > +++ b/drivers/staging/mt7621-pci-phy/pci-mt7621-phy.c > @@ -78,6 +78,8 @@ > > #define RG_PE1_FRC_MSTCKDIV BIT(5) > > +#define MAX_PHYS 2 > + > /** > * struct mt7621_pci_phy_instance - Mt7621 Pcie PHY device > * @phy: pointer to the kernel PHY device > @@ -289,6 +291,20 @@ static const struct phy_ops mt7621_pci_phy_ops = { > .owner = THIS_MODULE, > }; > > +static struct phy *mt7621_pcie_phy_of_xlate(struct device *dev, > + struct of_phandle_args *args) > +{ > + struct mt7621_pci_phy *mt7621_phy = dev_get_drvdata(dev); > + > + if (args->args_count == 0) > + return mt7621_phy->phys[0]->phy; > + > + if (WARN_ON(args->args[0] >= MAX_PHYS)) > + return ERR_PTR(-ENODEV); > + > + return mt7621_phy->phys[args->args[0]]->phy; > +} > + > static int mt7621_pci_phy_probe(struct platform_device *pdev) > { > struct device *dev = &pdev->dev; > @@ -299,6 +315,7 @@ static int mt7621_pci_phy_probe(struct platform_device *pdev) > struct resource res; > int port, ret; > void __iomem *port_base; > + u32 phy_num; > > phy = devm_kzalloc(dev, sizeof(*phy), GFP_KERNEL); > if (!phy) > @@ -325,8 +342,9 @@ static int mt7621_pci_phy_probe(struct platform_device *pdev) > return PTR_ERR(port_base); > } > > - port = 0; > - for_each_child_of_node(np, child_np) { This isn't the old place that you depend on the children nodes. A little earlier, you have phy->nphys = of_get_child_count(np); which now sets nphys to zero. I changed this to phy->nphys = MAX_PHYS; > + of_property_read_u32(dev->of_node, "#phy-cells", &phy_num); > + > + for (port = 0; port < phy_num + 1; port++) { I don't think it is correct to use #phy-cells as the number of phys. #phy-cells is the number of arguments - the args_count seen in mt7621_pci_phy_probe. I think you should either assume phy_num is MAX_PHYS, or have built-in knowledge of when it is 1 and when it is too. There is minimal cost to allocating an extra phy, so I would go with MAX_PHYS. > struct mt7621_pci_phy_instance *instance; > struct phy *pphy; > > @@ -338,7 +356,7 @@ static int mt7621_pci_phy_probe(struct platform_device *pdev) > > phy->phys[port] = instance; > > - pphy = devm_phy_create(dev, child_np, &mt7621_pci_phy_ops); > + pphy = devm_phy_create(dev, dev->of_node, &mt7621_pci_phy_ops); > if (IS_ERR(phy)) { > dev_err(dev, "failed to create phy\n"); > ret = PTR_ERR(phy); > @@ -352,7 +370,7 @@ static int mt7621_pci_phy_probe(struct platform_device *pdev) > port++; This "port++" duplicates the "port++" that you introduced in the "for" loop above. Thanks, NeilBrown > } > > - provider = devm_of_phy_provider_register(dev, of_phy_simple_xlate); > + provider = devm_of_phy_provider_register(dev, mt7621_pcie_phy_of_xlate); > > return PTR_ERR_OR_ZERO(provider); > > -- > 2.19.1