[PATCH 1/2] net: mdio: common handling of phy reset properties

[PATCH 1/2] net: mdio: common handling of phy reset properties

[Buday Csaba]

Reset properties of an `mdio_device` are initialized in multiple
source files and multiple functions:
  - `reset_assert_delay` and `reset_deassert_delay` are in
    fwnode_mdio.c
  - `reset_gpio` and `reset_ctrl` are in mdio_bus.c, but handled by
    different functions

This patch unifies the handling of all these properties into two
functions.
mdiobus_register_gpiod() and mdiobus_register_reset() are removed,
while mdio_device_register_reset() and mdio_device_unregister_reset()
are introduced instead.
These functions handle both reset-controllers and reset-gpios, and
also read the corresponding properties from the device tree.
These changes should make tracking the reset properties easier.

The reset logic is unaltered, and should work as it did before.

Signed-off-by: Buday Csaba <buday.csaba@prolan.hu>
---
 drivers/net/mdio/fwnode_mdio.c |  5 ----
 drivers/net/phy/mdio_bus.c     | 39 ++-----------------------
 drivers/net/phy/mdio_device.c  | 52 ++++++++++++++++++++++++++++++++++
 include/linux/mdio.h           |  2 ++
 4 files changed, 56 insertions(+), 42 deletions(-)

diff --git a/drivers/net/mdio/fwnode_mdio.c b/drivers/net/mdio/fwnode_mdio.c
index 9b41d4697..ba7091518 100644
--- a/drivers/net/mdio/fwnode_mdio.c
+++ b/drivers/net/mdio/fwnode_mdio.c
@@ -92,11 +92,6 @@ int fwnode_mdiobus_phy_device_register(struct mii_bus *mdio,
 	if (fwnode_property_read_bool(child, "broken-turn-around"))
 		mdio->phy_ignore_ta_mask |= 1 << addr;
 
-	fwnode_property_read_u32(child, "reset-assert-us",
-				 &phy->mdio.reset_assert_delay);
-	fwnode_property_read_u32(child, "reset-deassert-us",
-				 &phy->mdio.reset_deassert_delay);
-
 	/* Associate the fwnode with the device structure so it
 	 * can be looked up later
 	 */
diff --git a/drivers/net/phy/mdio_bus.c b/drivers/net/phy/mdio_bus.c
index cad6ed3aa..cc3f9cfb1 100644
--- a/drivers/net/phy/mdio_bus.c
+++ b/drivers/net/phy/mdio_bus.c
@@ -33,33 +33,6 @@
 #define CREATE_TRACE_POINTS
 #include <trace/events/mdio.h>
 
-static int mdiobus_register_gpiod(struct mdio_device *mdiodev)
-{
-	/* Deassert the optional reset signal */
-	mdiodev->reset_gpio = gpiod_get_optional(&mdiodev->dev,
-						 "reset", GPIOD_OUT_LOW);
-	if (IS_ERR(mdiodev->reset_gpio))
-		return PTR_ERR(mdiodev->reset_gpio);
-
-	if (mdiodev->reset_gpio)
-		gpiod_set_consumer_name(mdiodev->reset_gpio, "PHY reset");
-
-	return 0;
-}
-
-static int mdiobus_register_reset(struct mdio_device *mdiodev)
-{
-	struct reset_control *reset;
-
-	reset = reset_control_get_optional_exclusive(&mdiodev->dev, "phy");
-	if (IS_ERR(reset))
-		return PTR_ERR(reset);
-
-	mdiodev->reset_ctrl = reset;
-
-	return 0;
-}
-
 int mdiobus_register_device(struct mdio_device *mdiodev)
 {
 	int err;
@@ -68,16 +41,9 @@ int mdiobus_register_device(struct mdio_device *mdiodev)
 		return -EBUSY;
 
 	if (mdiodev->flags & MDIO_DEVICE_FLAG_PHY) {
-		err = mdiobus_register_gpiod(mdiodev);
+		err = mdio_device_register_reset(mdiodev);
 		if (err)
 			return err;
-
-		err = mdiobus_register_reset(mdiodev);
-		if (err)
-			return err;
-
-		/* Assert the reset signal */
-		mdio_device_reset(mdiodev, 1);
 	}
 
 	mdiodev->bus->mdio_map[mdiodev->addr] = mdiodev;
@@ -91,8 +57,7 @@ int mdiobus_unregister_device(struct mdio_device *mdiodev)
 	if (mdiodev->bus->mdio_map[mdiodev->addr] != mdiodev)
 		return -EINVAL;
 
-	gpiod_put(mdiodev->reset_gpio);
-	reset_control_put(mdiodev->reset_ctrl);
+	mdio_device_unregister_reset(mdiodev);
 
 	mdiodev->bus->mdio_map[mdiodev->addr] = NULL;
 
diff --git a/drivers/net/phy/mdio_device.c b/drivers/net/phy/mdio_device.c
index f64176e0e..e72bfa5a7 100644
--- a/drivers/net/phy/mdio_device.c
+++ b/drivers/net/phy/mdio_device.c
@@ -74,6 +74,58 @@ struct mdio_device *mdio_device_create(struct mii_bus *bus, int addr)
 }
 EXPORT_SYMBOL(mdio_device_create);
 
+/**
+ * mdio_device_register_reset - Read and initialize the reset properties of
+ *				an mdio device
+ * @mdiodev: mdio_device structure
+ */
+int mdio_device_register_reset(struct mdio_device *mdiodev)
+{
+	struct reset_control *reset;
+
+	/* Read optional firmware properties */
+	fwnode_property_read_u32(dev_fwnode(&mdiodev->dev), "reset-assert-us",
+				 &mdiodev->reset_assert_delay);
+	fwnode_property_read_u32(dev_fwnode(&mdiodev->dev), "reset-deassert-us",
+				 &mdiodev->reset_deassert_delay);
+
+	/* reset-gpio, bring up deasserted */
+	mdiodev->reset_gpio = gpiod_get_optional(&mdiodev->dev, "reset",
+						 GPIOD_OUT_LOW);
+
+	if (IS_ERR(mdiodev->reset_gpio))
+		return PTR_ERR(mdiodev->reset_gpio);
+
+	if (mdiodev->reset_gpio)
+		gpiod_set_consumer_name(mdiodev->reset_gpio, "PHY reset");
+
+	reset = reset_control_get_optional_exclusive(&mdiodev->dev, "phy");
+	if (IS_ERR(reset))
+		return PTR_ERR(reset);
+
+	mdiodev->reset_ctrl = reset;
+
+	/* Assert the reset signal */
+	mdio_device_reset(mdiodev, 1);
+
+	return 0;
+}
+EXPORT_SYMBOL(mdio_device_register_reset);
+
+/**
+ * mdio_device_unregister_reset - uninitialize the reset properties of
+ *				  an mdio device
+ * @mdiodev: mdio_device structure
+ */
+int mdio_device_unregister_reset(struct mdio_device *mdiodev)
+{
+	gpiod_put(mdiodev->reset_gpio);
+	reset_control_put(mdiodev->reset_ctrl);
+
+	return 0;
+}
+EXPORT_SYMBOL(mdio_device_unregister_reset);
+
 /**
  * mdio_device_register - Register the mdio device on the MDIO bus
  * @mdiodev: mdio_device structure to be added to the MDIO bus
diff --git a/include/linux/mdio.h b/include/linux/mdio.h
index c640ba44d..e6023cf13 100644
--- a/include/linux/mdio.h
+++ b/include/linux/mdio.h
@@ -90,6 +90,8 @@ static inline void *mdiodev_get_drvdata(struct mdio_device *mdio)
 
 void mdio_device_free(struct mdio_device *mdiodev);
 struct mdio_device *mdio_device_create(struct mii_bus *bus, int addr);
+int mdio_device_register_reset(struct mdio_device *mdiodev);
+int mdio_device_unregister_reset(struct mdio_device *mdiodev);
 int mdio_device_register(struct mdio_device *mdiodev);
 void mdio_device_remove(struct mdio_device *mdiodev);
 void mdio_device_reset(struct mdio_device *mdiodev, int value);
-- 
2.39.5

[PATCH 2/2] net: mdio: reset PHY before attempting to access registers in fwnode_mdiobus_register_phy

[Buday Csaba]

When the ID of an ethernet PHY is not provided by the 'compatible'
string in the device tree, its actual ID is read via the MDIO bus.
For some PHYs this could be unsafe, since a hard reset may be
necessary to safely access the MDIO registers.
This patch makes it possible to hard-reset an ethernet PHY before
attempting to read the ID, via a new device tree property, called:
`reset-phy-before-probe`.

There were previous attempts to implement such functionality, I
tried to collect a few of these (see links).

Link: https://lore.kernel.org/lkml/1499346330-12166-2-git-send-email-richard.leitner@skidata.com/
Link: https://lore.kernel.org/all/20230405-net-next-topic-net-phy-reset-v1-0-7e5329f08002@pengutronix.de/
Link: https://lore.kernel.org/netdev/20250709133222.48802-4-buday.csaba@prolan.hu/
Signed-off-by: Buday Csaba <buday.csaba@prolan.hu>
---
 drivers/net/mdio/fwnode_mdio.c | 30 +++++++++++++++++++++++++++++-
 1 file changed, 29 insertions(+), 1 deletion(-)

diff --git a/drivers/net/mdio/fwnode_mdio.c b/drivers/net/mdio/fwnode_mdio.c
index ba7091518..6f405df98 100644
--- a/drivers/net/mdio/fwnode_mdio.c
+++ b/drivers/net/mdio/fwnode_mdio.c
@@ -114,6 +114,31 @@ int fwnode_mdiobus_phy_device_register(struct mii_bus *mdio,
 }
 EXPORT_SYMBOL(fwnode_mdiobus_phy_device_register);
 
+static int fwnode_reset_phy_before_probe(struct mii_bus *bus, u32 addr,
+					 struct fwnode_handle *phy_node)
+{
+	struct mdio_device *tmpdev;
+	int err;
+
+	tmpdev = mdio_device_create(bus, addr);
+
+	fwnode_handle_get(phy_node);
+	device_set_node(&tmpdev->dev, phy_node);
+	err = mdio_device_register_reset(tmpdev);
+	if (err)
+		return err;
+
+	mdio_device_reset(tmpdev, 1);
+	mdio_device_reset(tmpdev, 0);
+
+	mdio_device_unregister_reset(tmpdev);
+
+	mdio_device_free(tmpdev);
+	fwnode_handle_put(phy_node);
+
+	return 0;
+}
+
 int fwnode_mdiobus_register_phy(struct mii_bus *bus,
 				struct fwnode_handle *child, u32 addr)
 {
@@ -129,8 +154,11 @@ int fwnode_mdiobus_register_phy(struct mii_bus *bus,
 		return PTR_ERR(mii_ts);
 
 	is_c45 = fwnode_device_is_compatible(child, "ethernet-phy-ieee802.3-c45");
-	if (is_c45 || fwnode_get_phy_id(child, &phy_id))
+	if (is_c45 || fwnode_get_phy_id(child, &phy_id)) {
+		if (fwnode_property_present(child, "reset-phy-before-probe"))
+			fwnode_reset_phy_before_probe(bus, addr, child);
 		phy = get_phy_device(bus, addr, is_c45);
+	}
 	else
 		phy = phy_device_create(bus, addr, phy_id, 0, NULL);
 	if (IS_ERR(phy)) {
-- 
2.39.5

Re: [PATCH 1/2] net: mdio: common handling of phy reset properties

[Andrew Lunn]

> +/**
> + * mdio_device_register_reset - Read and initialize the reset properties of
> + *				an mdio device
> + * @mdiodev: mdio_device structure
> + */
> +int mdio_device_register_reset(struct mdio_device *mdiodev)
> +{
> +	struct reset_control *reset;
> +
> +	/* Read optional firmware properties */
> +	fwnode_property_read_u32(dev_fwnode(&mdiodev->dev), "reset-assert-us",
> +				 &mdiodev->reset_assert_delay);
> +	fwnode_property_read_u32(dev_fwnode(&mdiodev->dev), "reset-deassert-us",
> +				 &mdiodev->reset_deassert_delay);
> +

device_property_read_u32() would be better here. But i can see why you
kept fwnode_property_read_u32() it makes it easier to see that there
have not been any code changes. So maybe add a patch 2/3 to convert
this?

> +/**
> + * mdio_device_unregister_reset - uninitialize the reset properties of
> + *				  an mdio device
> + * @mdiodev: mdio_device structure
> + */
> +int mdio_device_unregister_reset(struct mdio_device *mdiodev)
> +{
> +	gpiod_put(mdiodev->reset_gpio);
> +	reset_control_put(mdiodev->reset_ctrl);

Both of these return void, so you should make this function a void as
well.

    Andrew

---
pw-bot: cr

Re: [PATCH 2/2] net: mdio: reset PHY before attempting to access registers in fwnode_mdiobus_register_phy

[Maxime Chevallier]

Hi,

On 13/10/2025 15:55, Buday Csaba wrote:
> When the ID of an ethernet PHY is not provided by the 'compatible'
> string in the device tree, its actual ID is read via the MDIO bus.
> For some PHYs this could be unsafe, since a hard reset may be
> necessary to safely access the MDIO registers.
> This patch makes it possible to hard-reset an ethernet PHY before
> attempting to read the ID, via a new device tree property, called:
> `reset-phy-before-probe`.
> 
> There were previous attempts to implement such functionality, I
> tried to collect a few of these (see links).
> 
> Link: https://lore.kernel.org/lkml/1499346330-12166-2-git-send-email-richard.leitner@skidata.com/
> Link: https://lore.kernel.org/all/20230405-net-next-topic-net-phy-reset-v1-0-7e5329f08002@pengutronix.de/
> Link: https://lore.kernel.org/netdev/20250709133222.48802-4-buday.csaba@prolan.hu/
> Signed-off-by: Buday Csaba <buday.csaba@prolan.hu>

This should probably be accompanied by a DT binding update,
with some justification that this is indeed HW description
and not OS confguration.

At least the use of the term "probe" in the property makes this
sound like OS configuration, maybe something like :
  "phy-id-needs-reset" ?

Maxime

Re: [PATCH 2/2] net: mdio: reset PHY before attempting to access registers in fwnode_mdiobus_register_phy

[Buday Csaba]

On Mon, Oct 13, 2025 at 04:31:10PM +0200, Maxime Chevallier wrote:
> Hi,
> 
> On 13/10/2025 15:55, Buday Csaba wrote:
> > When the ID of an ethernet PHY is not provided by the 'compatible'
> > string in the device tree, its actual ID is read via the MDIO bus.
> > For some PHYs this could be unsafe, since a hard reset may be
> > necessary to safely access the MDIO registers.
> > This patch makes it possible to hard-reset an ethernet PHY before
> > attempting to read the ID, via a new device tree property, called:
> > `reset-phy-before-probe`.
> > 
> > There were previous attempts to implement such functionality, I
> > tried to collect a few of these (see links).
> > 
> > Link: https://lore.kernel.org/lkml/1499346330-12166-2-git-send-email-richard.leitner@skidata.com/
> > Link: https://lore.kernel.org/all/20230405-net-next-topic-net-phy-reset-v1-0-7e5329f08002@pengutronix.de/
> > Link: https://lore.kernel.org/netdev/20250709133222.48802-4-buday.csaba@prolan.hu/
> > Signed-off-by: Buday Csaba <buday.csaba@prolan.hu>
> 
> This should probably be accompanied by a DT binding update,
> with some justification that this is indeed HW description
> and not OS confguration.
> 

I have the corresponding patch ready, I just wanted this to be accepted
first. My description was to be:
  "When the PHY ID is not provided with the 'compatible' string,
   reset the PHY before attempting to read its ID over MDIO."

> At least the use of the term "probe" in the property makes this
> sound like OS configuration, maybe something like :
>  "phy-id-needs-reset" ?
> 
> Maxime
> 
> 

Okay, I will change that accordingly. Maybe "phy-id-read-needs-reset"?
A bit longer, but more self-explaining.

Best regards,
Csaba

Re: [PATCH 2/2] net: mdio: reset PHY before attempting to access registers in fwnode_mdiobus_register_phy

[Russell King (Oracle)]

On Wed, Oct 15, 2025 at 11:07:03AM +0200, Buday Csaba wrote:
> On Mon, Oct 13, 2025 at 04:31:10PM +0200, Maxime Chevallier wrote:
> > Hi,
> > 
> > On 13/10/2025 15:55, Buday Csaba wrote:
> > > When the ID of an ethernet PHY is not provided by the 'compatible'
> > > string in the device tree, its actual ID is read via the MDIO bus.
> > > For some PHYs this could be unsafe, since a hard reset may be
> > > necessary to safely access the MDIO registers.
> > > This patch makes it possible to hard-reset an ethernet PHY before
> > > attempting to read the ID, via a new device tree property, called:
> > > `reset-phy-before-probe`.
> > > 
> > > There were previous attempts to implement such functionality, I
> > > tried to collect a few of these (see links).
> > > 
> > > Link: https://lore.kernel.org/lkml/1499346330-12166-2-git-send-email-richard.leitner@skidata.com/
> > > Link: https://lore.kernel.org/all/20230405-net-next-topic-net-phy-reset-v1-0-7e5329f08002@pengutronix.de/
> > > Link: https://lore.kernel.org/netdev/20250709133222.48802-4-buday.csaba@prolan.hu/
> > > Signed-off-by: Buday Csaba <buday.csaba@prolan.hu>
> > 
> > This should probably be accompanied by a DT binding update,
> > with some justification that this is indeed HW description
> > and not OS confguration.
> > 
> 
> I have the corresponding patch ready, I just wanted this to be accepted
> first. My description was to be:

Please see Documentation/devicetree/bindings/submitting-patches.rst
point 5.

Basically, don't submit code that introduces a new undocumented DT
property without also including the binding update in the patch set.

Thanks.

-- 
RMK's Patch system: https://www.armlinux.org.uk/developer/patches/
FTTP is here! 80Mbps down 10Mbps up. Decent connectivity at last!