LinuxLists.cc - [PATCH net] net: stmmac: Handle disabled MDIO busses from devicetree

2023-12-11 21:31:59

Subject: [PATCH net] net: stmmac: Handle disabled MDIO busses from devicetree

Many hardware configurations have the MDIO bus disabled, and are instead
using some other MDIO bus to talk to the MAC's phy.

of_mdiobus_register() returns -ENODEV in this case. Let's handle it
gracefully instead of failing to probe the MAC.

Fixes: 47dd7a540b8a (net: add support for STMicroelectronics Ethernet controllers.")
Signed-off-by: Andrew Halaney <[email protected]>
---
drivers/net/ethernet/stmicro/stmmac/stmmac_mdio.c | 8 +++++++-
1 file changed, 7 insertions(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/stmicro/stmmac/stmmac_mdio.c b/drivers/net/ethernet/stmicro/stmmac/stmmac_mdio.c
index fa9e7e7040b9..a39be15d41a8 100644
--- a/drivers/net/ethernet/stmicro/stmmac/stmmac_mdio.c
+++ b/drivers/net/ethernet/stmicro/stmmac/stmmac_mdio.c
@@ -591,7 +591,13 @@ int stmmac_mdio_register(struct net_device *ndev)
new_bus->parent = priv->device;

err = of_mdiobus_register(new_bus, mdio_node);
- if (err != 0) {
+ if (err) {
+ if (err == -ENODEV) {
+ /* The bus is disabled in the devicetree, that's ok */
+ mdiobus_free(new_bus);
+ return 0;
+ }
+
dev_err_probe(dev, err, "Cannot register the MDIO bus\n");
goto bus_register_fail;
}

---
base-commit: bbd220ce4e29ed55ab079007cff0b550895258eb
change-id: 20231211-b4-stmmac-handle-mdio-enodev-82168de68c6a

Best regards,
--
Andrew Halaney <[email protected]>

2023-12-12 10:59:42

by Serge Semin

[permalink] [raw]

Subject: Re: [PATCH net] net: stmmac: Handle disabled MDIO busses from devicetree

On Mon, Dec 11, 2023 at 03:31:17PM -0600, Andrew Halaney wrote:
> Many hardware configurations have the MDIO bus disabled, and are instead
> using some other MDIO bus to talk to the MAC's phy.
>
> of_mdiobus_register() returns -ENODEV in this case. Let's handle it
> gracefully instead of failing to probe the MAC.
>
> Fixes: 47dd7a540b8a (net: add support for STMicroelectronics Ethernet controllers.")
> Signed-off-by: Andrew Halaney <[email protected]>
> ---
> drivers/net/ethernet/stmicro/stmmac/stmmac_mdio.c | 8 +++++++-
> 1 file changed, 7 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/net/ethernet/stmicro/stmmac/stmmac_mdio.c b/drivers/net/ethernet/stmicro/stmmac/stmmac_mdio.c
> index fa9e7e7040b9..a39be15d41a8 100644
> --- a/drivers/net/ethernet/stmicro/stmmac/stmmac_mdio.c
> +++ b/drivers/net/ethernet/stmicro/stmmac/stmmac_mdio.c
> @@ -591,7 +591,13 @@ int stmmac_mdio_register(struct net_device *ndev)
> new_bus->parent = priv->device;
>
> err = of_mdiobus_register(new_bus, mdio_node);
> - if (err != 0) {
> + if (err) {
> + if (err == -ENODEV) {
> + /* The bus is disabled in the devicetree, that's ok */
> + mdiobus_free(new_bus);
> + return 0;
> + }
> +
> dev_err_probe(dev, err, "Cannot register the MDIO bus\n");
> goto bus_register_fail;
> }

This can be implemented a bit simpler, more maintainable and saving
one indentations level:

err = of_mdiobus_register(new_bus, mdio_node);
if (err == -ENODEV) {
err = 0;
dev_warn(dev, "MDIO bus is disabled\n");
goto bus_register_fail;
} else if (err) {
dev_err_probe(dev, err, "Cannot register the MDIO bus\n");
goto bus_register_fail;
}

-Serge(y)

>
> ---
> base-commit: bbd220ce4e29ed55ab079007cff0b550895258eb
> change-id: 20231211-b4-stmmac-handle-mdio-enodev-82168de68c6a
>
> Best regards,
> --
> Andrew Halaney <[email protected]>
>

2023-12-12 14:28:58

by Andrew Halaney

[permalink] [raw]

Subject: Re: [PATCH net] net: stmmac: Handle disabled MDIO busses from devicetree

On Tue, Dec 12, 2023 at 01:59:25PM +0300, Serge Semin wrote:
> On Mon, Dec 11, 2023 at 03:31:17PM -0600, Andrew Halaney wrote:
> > Many hardware configurations have the MDIO bus disabled, and are instead
> > using some other MDIO bus to talk to the MAC's phy.
> >
> > of_mdiobus_register() returns -ENODEV in this case. Let's handle it
> > gracefully instead of failing to probe the MAC.
> >
> > Fixes: 47dd7a540b8a (net: add support for STMicroelectronics Ethernet controllers.")
> > Signed-off-by: Andrew Halaney <[email protected]>
> > ---
> > drivers/net/ethernet/stmicro/stmmac/stmmac_mdio.c | 8 +++++++-
> > 1 file changed, 7 insertions(+), 1 deletion(-)
> >
> > diff --git a/drivers/net/ethernet/stmicro/stmmac/stmmac_mdio.c b/drivers/net/ethernet/stmicro/stmmac/stmmac_mdio.c
> > index fa9e7e7040b9..a39be15d41a8 100644
> > --- a/drivers/net/ethernet/stmicro/stmmac/stmmac_mdio.c
> > +++ b/drivers/net/ethernet/stmicro/stmmac/stmmac_mdio.c
> > @@ -591,7 +591,13 @@ int stmmac_mdio_register(struct net_device *ndev)
> > new_bus->parent = priv->device;
> >
> > err = of_mdiobus_register(new_bus, mdio_node);
> > - if (err != 0) {
> > + if (err) {
> > + if (err == -ENODEV) {
> > + /* The bus is disabled in the devicetree, that's ok */
> > + mdiobus_free(new_bus);
> > + return 0;
> > + }
> > +
> > dev_err_probe(dev, err, "Cannot register the MDIO bus\n");
> > goto bus_register_fail;
> > }
>
> This can be implemented a bit simpler, more maintainable and saving
> one indentations level:
>
> err = of_mdiobus_register(new_bus, mdio_node);
> if (err == -ENODEV) {
> err = 0;
> dev_warn(dev, "MDIO bus is disabled\n");

Thanks for all your reviews, I agree this is cleaner!

I'm going to opt to use dev_info() here as this isn't something that's
wrong, just worth noting.

> goto bus_register_fail;
> } else if (err) {
> dev_err_probe(dev, err, "Cannot register the MDIO bus\n");
> goto bus_register_fail;
> }
>
> -Serge(y)
>
> >
> > ---
> > base-commit: bbd220ce4e29ed55ab079007cff0b550895258eb
> > change-id: 20231211-b4-stmmac-handle-mdio-enodev-82168de68c6a
> >
> > Best regards,
> > --
> > Andrew Halaney <[email protected]>
> >
>

2023-12-12 18:48:27

by Simon Horman

[permalink] [raw]