Skip to contentSkip to conversation

Add helper function for linking a DM Eth device to a PHY

PATCH Superseded Alex Marginean 4 participants 17 messages Nov 14, 2019 – Nov 24, 2019

Subsystems (2)

v4 of 5 Superseded Latest is v5 →

lore
Share & tools
lore
#0 lore raw

v4 cover Add helper function for linking a DM Eth device to a PHY

The patch set introduces dm_eth_phy_connect which takes in an ethernet device and uses DT information to find the associated PHY and connect the Ethernet interface to it. This should simplify similar code in ethernet drivers. A new binding is introduced for scanning an MDIO bus, useful if the bus is private to the Ethernet interface. We're going to try to get this binding accepted in Linux too.
Finally the patch set updates fsl_enetc driver to use the new helper function.

This patch set supersedes v2 series:
https://patchwork.ozlabs.org/project/uboot/list/?series=124537
and v3 series:
https://patchwork.ozlabs.org/project/uboot/list/?series=140114

Changes in v4:
- rebased on current head

Changes in v3:
- added cover letter
- check for null PHY pointer before using it in dm_eth_connect_phy_handle
- moved the code dealing with MDIO scanning into a separate patch
- renames several arguments and variables for a bit more clarity and consistency

Changes in v2:
- Moved MDIO scan code into dm_mdio_phy_scan which is also exported
- Use interface instead of if_type for consistency
- don't use phy pointer if NULL in fsl_enetc code

Alex Marginean (5):
net: mdio-uclass: rename arguments of dm_mdio_phy_connect for
consistency
net: mdio-uclass: add dm_eth_phy_connect helper function
net: mdio-uclass: Add helper functions for scanning the MDIO bus
doc: bindings: add mdio-handle property to ethernet nodes
drivers: net: fsl_enetc: use the new MDIO DM helper functions

 doc/device-tree-bindings/net/ethernet.txt |   2 +
 drivers/net/fsl_enetc.c                   |  52 ++------
 drivers/net/fsl_enetc.h                   |   1 +
 include/miiphy.h                          |  32 ++++-
 net/mdio-uclass.c                         | 139 +++++++++++++++++++++-
 5 files changed, 175 insertions(+), 51 deletions(-)
Signature · 1 line
2.17.1
#1 lore raw

PATCH v4 1/5 net: mdio-uclass: rename arguments of dm_mdio_phy_connect for clarity

Renamed dm_mdio_phy_connect arguments dev to mdiodev and addr to phyaddr for a bit more clarity and consistency with the following patches. Also use NULL instead of 0 on error return path.

---
 include/miiphy.h  |  6 +++---
 net/mdio-uclass.c | 10 +++++-----
 2 files changed, 8 insertions(+), 8 deletions(-)
include/miiphy.h +3 -3
index 9b97d09f18..94bf0da24a 100644
@@ -154,14 +154,14 @@ void dm_mdio_probe_devices(void);
 /**
  * dm_mdio_phy_connect - Wrapper over phy_connect for DM MDIO
  *
- * @dev: mdio dev
- * @addr: PHY address on MDIO bus
+ * @mdiodev: mdio device the PHY is accesible on
+ * @phyaddr: PHY address on MDIO bus
  * @ethdev: ethernet device to connect to the PHY
  * @interface: MAC-PHY protocol
  *
  * @return pointer to phy_device, or 0 on error
  */
-struct phy_device *dm_mdio_phy_connect(struct udevice *dev, int addr,
+struct phy_device *dm_mdio_phy_connect(struct udevice *mdiodev, int phyaddr,
 				       struct udevice *ethdev,
 				       phy_interface_t interface);
 
net/mdio-uclass.c +5 -5
index 6f922e80b6..7a5f1d6dcc 100644
@@ -104,16 +104,16 @@ static int dm_mdio_pre_remove(struct udevice *dev)
 	return 0;
 }
 
-struct phy_device *dm_mdio_phy_connect(struct udevice *dev, int addr,
+struct phy_device *dm_mdio_phy_connect(struct udevice *mdiodev, int phyaddr,
 				       struct udevice *ethdev,
 				       phy_interface_t interface)
 {
-	struct mdio_perdev_priv *pdata = dev_get_uclass_priv(dev);
+	struct mdio_perdev_priv *pdata = dev_get_uclass_priv(mdiodev);
 
-	if (device_probe(dev))
-		return 0;
+	if (device_probe(mdiodev))
+		return NULL;
 
-	return phy_connect(pdata->mii_bus, addr, ethdev, interface);
+	return phy_connect(pdata->mii_bus, phyaddr, ethdev, interface);
 }
 
 UCLASS_DRIVER(mdio) = {
Signature · 1 line
2.17.1
#2 lore raw

PATCH v4 2/5 net: mdio-uclass: add dm_eth_phy_connect helper function

In reply to Alex Marginean

The function connects an ethernet device to a PHY using DT information. This API is only available for eth devices with an associated device tree node.

---
 include/miiphy.h  | 12 ++++++++
 net/mdio-uclass.c | 72 +++++++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 84 insertions(+)
include/miiphy.h +12 -0
index 94bf0da24a..e591d93416 100644
@@ -165,6 +165,18 @@ struct phy_device *dm_mdio_phy_connect(struct udevice *mdiodev, int phyaddr,
 				       struct udevice *ethdev,
 				       phy_interface_t interface);
 
+/**
+ * dm_eth_phy_connect - Connect an Eth device to a PHY based on device tree
+ *
+ * Picks up the DT phy-handle/mdio-handle and phy-mode from ethernet device node
+ * and connects the ethernet device to the linked PHY.
+ *
+ * @ethdev: ethernet device
+ *
+ * @return pointer to phy_device, or 0 on error
+ */
+struct phy_device *dm_eth_phy_connect(struct udevice *ethdev);
+
 #endif
 
 #ifdef CONFIG_DM_MDIO_MUX
net/mdio-uclass.c +72 -0

78 lines not on this page · raw message

Signature · 1 line
2.17.1
#3 lore raw

PATCH v4 3/5 net: mdio-uclass: Add helper functions for scanning the MDIO bus

In reply to Alex Marginean

The patch introduces dm_mdio_phy_scan which is a DM wrapper over phy_find_by_mask. It also updates dm_eth_phy_connect to scan a MDIO bus for PHYs if the ethernet node contains a mdio-handle property.

---
 include/miiphy.h  | 14 ++++++++++++
 net/mdio-uclass.c | 57 +++++++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 71 insertions(+)
include/miiphy.h +14 -0
index e591d93416..a5bdc29a47 100644
@@ -165,6 +165,20 @@ struct phy_device *dm_mdio_phy_connect(struct udevice *mdiodev, int phyaddr,
 				       struct udevice *ethdev,
 				       phy_interface_t interface);
 
+/**
+ * dm_mdio_phy_scan - Scan for a PHY on the given MDIO bus
+ *
+ * @mdiodev: mdio device to scan
+ * @ethdev: ethernet device to connect to the PHY
+ * @interface: MAC-PHY protocol
+ * @addrmask: bitmask of PHY addresses to scan, use all Fs to scan all addresses
+ *
+ * @return pointer to phy_device, or 0 on error
+ */
+struct phy_device *dm_mdio_phy_scan(struct udevice *mdiodev,
+				    struct udevice *ethdev,
+				    phy_interface_t interface, uint addrmask);
+
 /**
  * dm_eth_phy_connect - Connect an Eth device to a PHY based on device tree
  *
net/mdio-uclass.c +57 -0

77 lines not on this page · raw message

Signature · 1 line
2.17.1
#4 lore raw

PATCH v4 4/5 doc: bindings: add mdio-handle property to ethernet nodes

In reply to Alex Marginean

Adds an optional mdio-handle property which identifies a MDIO bus which can be scanned to find the relevant PHY. The property is ignored if phy-handle is also present.

---
 doc/device-tree-bindings/net/ethernet.txt | 2 ++
 1 file changed, 2 insertions(+)
doc/device-tree-bindings/net/ethernet.txt +2 -0
index 3fc360523b..9f9629f8d6 100644
@@ -9,6 +9,8 @@ The following properties are common to the Ethernet controllers:
 - max-speed: number, specifies maximum speed in Mbit/s supported by the device;
 - max-frame-size: number, maximum transfer unit (IEEE defined MTU), rather than
   the maximum frame size (there's contradiction in ePAPR).
+- mdio-handle: phandle, specifies a reference to a MDIO bus to be scanned to
+  find the PHY device.  Ignored if phy-handle is also present.
 - phy-mode: string, operation mode of the PHY interface; supported values are
   "mii", "gmii", "sgmii", "qsgmii", "tbi", "rev-mii", "rmii", "rgmii", "rgmii-id",
   "rgmii-rxid", "rgmii-txid", "rtbi", "smii", "xgmii"; this is now a de-facto
Signature · 1 line
2.17.1
#5 lore raw

PATCH v4 4/5 doc: bindings: add mdio-handle property to ethernet nodes

Quoted 14 lines · Alex Marginean

Adds an optional mdio-handle property which identifies a MDIO bus which can be scanned to find the relevant PHY. The property is ignored if phy-handle is also present.

---
  doc/device-tree-bindings/net/ethernet.txt | 2 ++
  1 file changed, 2 insertions(+)
doc/device-tree-bindings/net/ethernet.txt +0 -0
index 3fc360523b..9f9629f8d6 100644

doc/device-tree-bindings/net/ethernet.txt The following properties are common to the Ethernet controllers: · line 13

Quoted 5 lines
doc/device-tree-bindings/net/ethernet.txt +2 -0
@@ -9,6 +9,8 @@ The following properties are common to the Ethernet controllers:
  - max-speed: number, specifies maximum speed in Mbit/s supported by the device;
  - max-frame-size: number, maximum transfer unit (IEEE defined MTU), rather than
    the maximum frame size (there's contradiction in ePAPR).
+- mdio-handle: phandle, specifies a reference to a MDIO bus to be scanned to
+  find the PHY device.  Ignored if phy-handle is also present.

Sry, but it looks redundant. The Ethernet-controller bindings expects to use phy-handle which, in turn, allows to get MDIO node.

So, if your platform is DT based and can use DT then it's reasonable to follow standard binding, which, in addition, allows to specify Ethernet PHY properties.
More over, your series does not provide user for this new property.

Personally I do not see even reasons to have doc/device-tree-bindings/net/ethernet.txt in u-boot and think we should follow [1]

doc/device-tree-bindings/net/ethernet.txt The following properties are common to the Ethernet controllers: · line 16

Quoted 4 lines
doc/device-tree-bindings/net/ethernet.txt +0 -0
  - phy-mode: string, operation mode of the PHY interface; supported values are
    "mii", "gmii", "sgmii", "qsgmii", "tbi", "rev-mii", "rmii", "rgmii", "rgmii-id",
    "rgmii-rxid", "rgmii-txid", "rtbi", "smii", "xgmii"; this is now a de-facto

[1] https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/Documentation/devicetree/bindings/net/ethernet-controller.yaml

Signature · 2 lines
Best regards,
grygorii
#6 lore raw

PATCH v4 4/5 doc: bindings: add mdio-handle property to ethernet nodes

Quoted 24 lines · Grygorii Strashko
Quoted 21 lines · Alex Marginean

Adds an optional mdio-handle property which identifies a MDIO bus which can
be scanned to find the relevant PHY.  The property is ignored if phy-handle
is also present.

---
  doc/device-tree-bindings/net/ethernet.txt | 2 ++

  1 file changed, 2 insertions(+)
doc/device-tree-bindings/net/ethernet.txt +0 -0
index 3fc360523b..9f9629f8d6 100644
@@ -9,6 +9,8 @@ The following properties are common to the Ethernet 

controllers:
  - max-speed: number, specifies maximum speed in Mbit/s supported by the device;
  - max-frame-size: number, maximum transfer unit (IEEE defined MTU),

Quoted 8 lines
Quoted 5 lines

rather than
    the maximum frame size (there's contradiction in ePAPR). +- mdio-handle: phandle, specifies a reference to a MDIO bus to be scanned to
+  find the PHY device.  Ignored if phy-handle is also present.

Sry, but it looks redundant. The Ethernet-controller bindings expects to use phy-handle which, in turn, allows to get MDIO node.

The problem I'm trying to solve is not that I don't want to get the parent MDIO bus of a PHY referenced though phy-handle, but rather that I want to avoid having a specific U-Boot DT for each PHY plug-in card that could be used in our systems. I'll explain that in more detail.

We have these QDS systems which support plug-in cards with PHYs on them. The way it works is that a given ethernet interface is associated with one of the QDS slots that should contain a plug-in card. Each of the slots has a MDIO bus associated with it, this bus is accessed through a MDIO MUX on the QDS board.
The slot may be populated with one of several types of cards, which are different designs, have different types of PHYs and more importantly use different PHY addresses on the MDIO bus.

So now the summary is that an Ethernet interface is associated with a MDIO bus that could be scanned to find the relevant PHY, while using the existing phy-handle binding requires that U-Boot DT is edited whenever a different card is plugged in, which is not practical and avoidable.

Quoted 4 lines

So, if your platform is DT based and can use DT then it's reasonable to follow standard binding,
which, in addition, allows to specify Ethernet PHY properties.

Sure, and we do that on boards and for PHYs which are at fixed addresses on MDIO bus, but that is impractical with the swappable cards.

Quoted 1 line

More over, your series does not provide user for this new property.

That's true, I was planning to send the QDS DT after this was merged, but if it helps I can add it to this patch set.

Alex

Quoted 16 lines

Personally I do not see even reasons to have doc/device-tree-bindings/net/ethernet.txt in u-boot and think we should follow [1]

Quoted 7 lines

  - phy-mode: string, operation mode of the PHY interface; supported values are
    "mii", "gmii", "sgmii", "qsgmii", "tbi", "rev-mii", "rmii", "rgmii", "rgmii-id",
    "rgmii-rxid", "rgmii-txid", "rtbi", "smii", "xgmii"; this is now a de-facto

[1]
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/Documentation/devicetree/bindings/net/ethernet-controller.yaml

#7 lore raw

PATCH v4 4/5 doc: bindings: add mdio-handle property to ethernet nodes

Quoted 42 lines · Alexandru Marginean
Quoted 12 lines · Grygorii Strashko
Quoted 9 lines · Alex Marginean

Adds an optional mdio-handle property which identifies a MDIO bus which can be scanned to find the relevant PHY.  The property is ignored if phy-handle is also present.

---
  doc/device-tree-bindings/net/ethernet.txt | 2 ++

  1 file changed, 2 insertions(+)
Quoted 13 lines
Quoted 10 lines
doc/device-tree-bindings/net/ethernet.txt +0 -0
index 3fc360523b..9f9629f8d6 100644
@@ -9,6 +9,8 @@ The following properties are common to the Ethernet controllers:
   - max-speed: number, specifies maximum speed in Mbit/s supported by the device;

  - max-frame-size: number, maximum transfer unit (IEEE defined MTU), rather than the maximum frame size (there's contradiction in ePAPR).
+- mdio-handle: phandle, specifies a reference to a MDIO bus to be scanned to +  find the PHY device.  Ignored if phy-handle is also present.

Sry, but it looks redundant. The Ethernet-controller bindings expects to use phy-handle which, in turn, allows to get MDIO node.

The problem I'm trying to solve is not that I don't want to get the parent MDIO bus of a PHY referenced though phy-handle, but rather that I want to avoid having a specific U-Boot DT for each PHY plug-in card that could be used in our systems.  I'll explain that in more detail.

We have these QDS systems which support plug-in cards with PHYs on them. The way it works is that a given ethernet interface is associated with one of the QDS slots that should contain a plug-in card.  Each of the slots has a MDIO bus associated with it, this bus is accessed through a MDIO MUX on the QDS board.
The slot may be populated with one of several types of cards, which are different designs, have different types of PHYs and more importantly use different PHY addresses on the MDIO bus.

Quoted 4 lines

So now the summary is that an Ethernet interface is associated with a MDIO bus that could be scanned to find the relevant PHY, while using the existing phy-handle binding requires that U-Boot DT is edited whenever a different card is plugged in, which is not practical and avoidable.

Thank you for explanation. It's clear now.
I think above description is good to have as part of commit message.

Quoted 25 lines
Quoted 4 lines

So, if your platform is DT based and can use DT then it's reasonable to follow standard binding, which, in addition, allows to specify Ethernet PHY properties.

Sure, and we do that on boards and for PHYs which are at fixed addresses on MDIO bus, but that is impractical with the swappable cards.

Quoted 1 line

More over, your series does not provide user for this new property.

That's true, I was planning to send the QDS DT after this was merged, but if it helps I can add it to this patch set.

Alex

Quoted 10 lines

Personally I do not see even reasons to have doc/device-tree-bindings/net/ethernet.txt in u-boot and think we should follow [1]

Quoted 4 lines

  - phy-mode: string, operation mode of the PHY interface; supported values are "mii", "gmii", "sgmii", "qsgmii", "tbi", "rev-mii", "rmii", "rgmii", "rgmii-id", "rgmii-rxid", "rgmii-txid", "rtbi", "smii", "xgmii"; this is now a de-facto

[1] https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/Documentation/devicetree/bindings/net/ethernet-controller.yaml

Signature · 2 lines
Best regards,
grygorii
#8 lore raw

PATCH v4 4/5 doc: bindings: add mdio-handle property to ethernet nodes

Quoted 45 lines · Grygorii Strashko
Quoted 42 lines · Alexandru Marginean
Quoted 12 lines · Grygorii Strashko
Quoted 9 lines · Alex Marginean

Adds an optional mdio-handle property which identifies a MDIO bus which can be scanned to find the relevant PHY.  The property is ignored if phy-handle is also present.

---
  doc/device-tree-bindings/net/ethernet.txt | 2 ++

  1 file changed, 2 insertions(+)
Quoted 13 lines
Quoted 10 lines
doc/device-tree-bindings/net/ethernet.txt +0 -0
index 3fc360523b..9f9629f8d6 100644
@@ -9,6 +9,8 @@ The following properties are common to the Ethernet controllers:
   - max-speed: number, specifies maximum speed in Mbit/s supported by the device;

  - max-frame-size: number, maximum transfer unit (IEEE defined MTU), rather than the maximum frame size (there's contradiction in ePAPR).
+- mdio-handle: phandle, specifies a reference to a MDIO bus to be scanned to +  find the PHY device.  Ignored if phy-handle is also present.

Sry, but it looks redundant. The Ethernet-controller bindings expects to use phy-handle which, in turn, allows to get MDIO node.

The problem I'm trying to solve is not that I don't want to get the parent MDIO bus of a PHY referenced though phy-handle, but rather that I want to avoid having a specific U-Boot DT for each PHY plug-in card that could be used in our systems.  I'll explain that in more detail.

We have these QDS systems which support plug-in cards with PHYs on them. The way it works is that a given ethernet interface is associated with one of the QDS slots that should contain a plug-in card.  Each of the slots has a MDIO bus associated with it, this bus is accessed through a MDIO MUX on the QDS board.
The slot may be populated with one of several types of cards, which are different designs, have different types of PHYs and more importantly use different PHY addresses on the MDIO bus.

Quoted 16 lines
Quoted 4 lines

So now the summary is that an Ethernet interface is associated with a MDIO bus that could be scanned to find the relevant PHY, while using the existing phy-handle binding requires that U-Boot DT is edited whenever a different card is plugged in, which is not practical and avoidable.

Thank you for explanation. It's clear now.
I think above description is good to have as part of commit message.

Quoted 8 lines
Quoted 4 lines

So, if your platform is DT based and can use DT then it's reasonable to follow standard binding, which, in addition, allows to specify Ethernet PHY properties.

Sure, and we do that on boards and for PHYs which are at fixed addresses on MDIO bus, but that is impractical with the swappable cards.

Some thought, which i think might help.
- u-boot allows phy node not to have "reg" property.
- phy_connect() will return first PHY discovered on the MDIO bus if addr<=0

So, if MDIO assignment per ethernet interface/slot is fixed DT can look like

mdioX {
	mdiox_phy0: ethernet-phy@0 {
	};
}
ethernetX {
	phy-handle = <&mdiox_phy0>;
}

and so on. driver's parsing code can ignore PHY "reg" prop in phy node and pass addr<=0 to phy_connect().

It seems, current Linux approach is to have PHY "reg" property as required, but your case might be the reason to change this.

Signature · 2 lines
Best regards,
grygorii
#9 lore raw

PATCH v4 4/5 doc: bindings: add mdio-handle property to ethernet nodes

Quoted 85 lines · Grygorii Strashko
Quoted 52 lines · Grygorii Strashko
Quoted 49 lines · Alexandru Marginean
Quoted 24 lines · Grygorii Strashko
Quoted 21 lines · Alex Marginean

Adds an optional mdio-handle property which identifies a MDIO bus which can
be scanned to find the relevant PHY.  The property is ignored if phy-handle
is also present.

---
  doc/device-tree-bindings/net/ethernet.txt | 2 ++

  1 file changed, 2 insertions(+)
doc/device-tree-bindings/net/ethernet.txt +0 -0
index 3fc360523b..9f9629f8d6 100644
@@ -9,6 +9,8 @@ The following properties are common to the Ethernet 

controllers:
  - max-speed: number, specifies maximum speed in Mbit/s supported by the device;
  - max-frame-size: number, maximum transfer unit (IEEE defined

Quoted 8 lines
Quoted 5 lines

MTU), rather than
    the maximum frame size (there's contradiction in ePAPR). +- mdio-handle: phandle, specifies a reference to a MDIO bus to be scanned to
+  find the PHY device.  Ignored if phy-handle is also present.

Sry, but it looks redundant. The Ethernet-controller bindings expects to use phy-handle which, in turn, allows to get MDIO node.

The problem I'm trying to solve is not that I don't want to get the parent MDIO bus of a PHY referenced though phy-handle, but rather that I want to avoid having a specific U-Boot DT for each PHY plug-in card that could be used in our systems.  I'll explain that in more detail.

We have these QDS systems which support plug-in cards with PHYs on them. The way it works is that a given ethernet interface is associated with one of the QDS slots that should contain a plug-in card.  Each of the slots has a MDIO bus associated with it, this bus is accessed through a MDIO MUX on the QDS board.
The slot may be populated with one of several types of cards, which are different designs, have different types of PHYs and more importantly use different PHY addresses on the MDIO bus.

Quoted 17 lines
Quoted 4 lines

So now the summary is that an Ethernet interface is associated with a MDIO bus that could be scanned to find the relevant PHY, while using the existing phy-handle binding requires that U-Boot DT is edited whenever a different card is plugged in, which is not practical and avoidable.

Thank you for explanation. It's clear now.
I think above description is good to have as part of commit message.

Quoted 9 lines
Quoted 5 lines

So, if your platform is DT based and can use DT then it's reasonable to follow standard binding,
which, in addition, allows to specify Ethernet PHY properties.

Sure, and we do that on boards and for PHYs which are at fixed addresses on MDIO bus, but that is impractical with the swappable cards.

Some thought, which i think might help.
- u-boot allows phy node not to have "reg" property.
- phy_connect() will return first PHY discovered on the MDIO bus if addr<=0

So, if MDIO assignment per ethernet interface/slot is fixed DT can look like

mdioX {
    mdiox_phy0: ethernet-phy at 0 {
    };
}

Quoted 6 lines

ethernetX {
    phy-handle = <&mdiox_phy0>;
}

and so on. driver's parsing code can ignore PHY "reg" prop in phy node and pass addr<=0 to phy_connect().

Functionally that's what I'm looking for, yes. There is the problem of not strictly following the kernel binding and@the end of the day I will have the same problem with the kernel too, so I should take this topic to the netdev list anyway.

Having 'reg' made optional is actually an interesting idea. I think 'mdio-handle' is a bit more clear what it is, but having an actual PHY node allows passing DT properties on to the PHY driver which is certainly useful for some PHYs. It's like saying I don't know what PHY I'm going to find, but if it's a PHY that has these configurable properties here is your configuration.

Quoted 3 lines

It seems, current Linux approach is to have PHY "reg" property as required, but your case might be the reason to change this. >

I'll have to do a better write-up for the kernel list and we'll see how this will go.
In the meantime any feedback on the other patches, except the one introducing mdio-handle? For instance I should deal with fixed links in the dm_eth_phy_connect helper too (calling phy_connect_fixed), my intent is to take some code that is otherwise pretty generic out of the drivers.

Thanks!
Alex

#10 lore raw

PATCH v4 4/5 doc: bindings: add mdio-handle property to ethernet nodes

Quoted 88 lines · Alexandru Marginean
Quoted 76 lines · Grygorii Strashko
Quoted 45 lines · Grygorii Strashko
Quoted 42 lines · Alexandru Marginean
Quoted 12 lines · Grygorii Strashko
Quoted 9 lines · Alex Marginean

Adds an optional mdio-handle property which identifies a MDIO bus which can be scanned to find the relevant PHY.  The property is ignored if phy-handle is also present.

---
  doc/device-tree-bindings/net/ethernet.txt | 2 ++

  1 file changed, 2 insertions(+)
Quoted 13 lines
Quoted 10 lines
doc/device-tree-bindings/net/ethernet.txt +0 -0
index 3fc360523b..9f9629f8d6 100644
@@ -9,6 +9,8 @@ The following properties are common to the Ethernet controllers:
   - max-speed: number, specifies maximum speed in Mbit/s supported by the device;

  - max-frame-size: number, maximum transfer unit (IEEE defined MTU), rather than the maximum frame size (there's contradiction in ePAPR).
+- mdio-handle: phandle, specifies a reference to a MDIO bus to be scanned to +  find the PHY device.  Ignored if phy-handle is also present.

Sry, but it looks redundant. The Ethernet-controller bindings expects to use phy-handle which, in turn, allows to get MDIO node.

The problem I'm trying to solve is not that I don't want to get the parent MDIO bus of a PHY referenced though phy-handle, but rather that I want to avoid having a specific U-Boot DT for each PHY plug-in card that could be used in our systems.  I'll explain that in more detail.

We have these QDS systems which support plug-in cards with PHYs on them. The way it works is that a given ethernet interface is associated with one of the QDS slots that should contain a plug-in card.  Each of the slots has a MDIO bus associated with it, this bus is accessed through a MDIO MUX on the QDS board.
The slot may be populated with one of several types of cards, which are different designs, have different types of PHYs and more importantly use different PHY addresses on the MDIO bus.

Quoted 16 lines
Quoted 4 lines

So now the summary is that an Ethernet interface is associated with a MDIO bus that could be scanned to find the relevant PHY, while using the existing phy-handle binding requires that U-Boot DT is edited whenever a different card is plugged in, which is not practical and avoidable.

Thank you for explanation. It's clear now.
I think above description is good to have as part of commit message.

Quoted 8 lines
Quoted 4 lines

So, if your platform is DT based and can use DT then it's reasonable to follow standard binding, which, in addition, allows to specify Ethernet PHY properties.

Sure, and we do that on boards and for PHYs which are at fixed addresses on MDIO bus, but that is impractical with the swappable cards.

Some thought, which i think might help.
- u-boot allows phy node not to have "reg" property.
- phy_connect() will return first PHY discovered on the MDIO bus if addr<=0

So, if MDIO assignment per ethernet interface/slot is fixed DT can look like

mdioX {
     mdiox_phy0: ethernet-phy at 0 {
     };
}

Quoted 6 lines

ethernetX {
     phy-handle = <&mdiox_phy0>;
}

and so on. driver's parsing code can ignore PHY "reg" prop in phy node and pass addr<=0 to phy_connect().

Functionally that's what I'm looking for, yes.  There is the problem of not strictly following the kernel binding and at the end of the day I will have the same problem with the kernel too, so I should take this topic to the netdev list anyway.

Quoted 6 lines

Having 'reg' made optional is actually an interesting idea.  I think 'mdio-handle' is a bit more clear what it is, but having an actual PHY node allows passing DT properties on to the PHY driver which is certainly useful for some PHYs.  It's like saying I don't know what PHY I'm going to find, but if it's a PHY that has these configurable properties here is your configuration.

Another good question is "phy-connection-type"/"phy-mode". Are all your cards works in the same mode? And how is this solved?

Is there anything common with SFP? Linux: bindings/net/sff,sfp.txt

Quoted 10 lines
Quoted 3 lines

It seems, current Linux approach is to have PHY "reg" property as required, but your case might be the reason to change this. >

I'll have to do a better write-up for the kernel list and we'll see how this will go.
In the meantime any feedback on the other patches, except the one introducing mdio-handle?  For instance I should deal with fixed links in the dm_eth_phy_connect helper too (calling phy_connect_fixed), my intent is to take some code that is otherwise pretty generic out of the drivers.

From one side, it look nice. From another, in my case MDIO is not a device,
so can't try it.

But, any way, I'd like to share some notes I have. It may help you or may save your time by reducing number of possible regressions.

To be honest, there is a little mess in PHY creation area :( Common way is to do in drivers:
phy_connect(mii_bus, phy_addr, dev, interface)
|-phy_find_by_mask()

      |-get_phy_device_by_mask()
	|-create_phy_by_mask()
	  |-create_phy_by_mask()
             |-phy_device_create()
		|-phy_probe()
  |-phy_connect_dev(phydev, dev);
..
#ifdef CONFIG_DM_ETH
	if (ofnode_valid(slave->data->phy_of_handle))
		phydev->node = slave->data->phy_of_handle;

#endif

As you can see from above - the PHY probe will be called when both phy_device->dev and phydev->node have not been initialized yet, so no way to perform DT parsing in PHY .probe().

Another, not obvious, thing is that DT node for fixed phy passed through the int addr parameter in create_phy_by_mask().

And finally, as phy as fixed-phy may be connected to DT node which is *not" ethernet device node and describes Port. The ethernet device is parent DT node in this case.

In my opinion, it could really simplify things if phy_device->dev and phydev->node will be configured before calling PHY .probe(), but the challenge is to make it work for !DT case.

Signature · 2 lines
Best regards,
grygorii
#11 lore raw

PATCH v4 4/5 doc: bindings: add mdio-handle property to ethernet nodes

Quoted 33 lines · Grygorii Strashko
Quoted 26 lines
Quoted 15 lines

Some thought, which i think might help.
- u-boot allows phy node not to have "reg" property.
- phy_connect() will return first PHY discovered on the MDIO bus if addr<=0

So, if MDIO assignment per ethernet interface/slot is fixed DT can look like

mdioX {
     mdiox_phy0: ethernet-phy at 0 {
     };
}

Quoted 6 lines

ethernetX {
     phy-handle = <&mdiox_phy0>;
}

and so on. driver's parsing code can ignore PHY "reg" prop in phy node and pass addr<=0 to phy_connect().

Functionally that's what I'm looking for, yes.  There is the problem of not strictly following the kernel binding and at the end of the day I will have the same problem with the kernel too, so I should take this topic to the netdev list anyway.

Quoted 6 lines

Having 'reg' made optional is actually an interesting idea.  I think 'mdio-handle' is a bit more clear what it is, but having an actual PHY node allows passing DT properties on to the PHY driver which is certainly useful for some PHYs.  It's like saying I don't know what PHY I'm going to find, but if it's a PHY that has these configurable properties here is your configuration.

Quoted 2 lines

Another good question is "phy-connection-type"/"phy-mode". Are all your cards works in the same mode? And how is this solved?

These SoCs generally support multiple protocols on the serdes lanes. The protocol is selected at power on so just using a single static DT isn't sufficient.

For Linux DT I tried the DT overlay which seems to work fine. I'm using it to fix up phy-connection-type based on serdes configuration that is known to u-boot. Of course it's possible to apply an overlay for each specific plug-in card in a given slot too, but that's not practical especially for mixes of board with many slots combined with many types of cards. I'm not sure this can be done automatically in U-Boot such that the user doesn't have to do any manual configuration when swapping cards either.

U-Boot doesn't have the luxury of having overlays applied before it starts, we don't use SPL. I'm considering ways to have platform init code apply some sort of fix-up or overlay for U-Boot DT based on serdes configuration, but I don't have something running yet. Live tree is something I want to look into to see if it would work here.

Quoted 1 line

Is there anything common with SFP? Linux: bindings/net/sff,sfp.txt

They are not SFPs, they are normal MDIO driven PHYs just that they are mounted on plug-in cards.

Quoted 32 lines
Quoted 9 lines
Quoted 2 lines

It seems, current Linux approach is to have PHY "reg" property as required, but your case might be the reason to change this. >

I'll have to do a better write-up for the kernel list and we'll see how this will go.
In the meantime any feedback on the other patches, except the one introducing mdio-handle?  For instance I should deal with fixed links in the dm_eth_phy_connect helper too (calling phy_connect_fixed), my intent is to take some code that is otherwise pretty generic out of the drivers.

From one side, it look nice. From another, in my case MDIO is not a device, so can't try it.

But, any way, I'd like to share some notes I have. It may help you or may save your time by reducing number of possible regressions.

To be honest, there is a little mess in PHY creation area :( Common way is to do in drivers:
phy_connect(mii_bus, phy_addr, dev, interface)
 |-phy_find_by_mask()
     |-get_phy_device_by_mask() |-create_phy_by_mask()
      |-create_phy_by_mask() |-phy_device_create()
        |-phy_probe()
 |-phy_connect_dev(phydev, dev);
..
#ifdef CONFIG_DM_ETH
    if (ofnode_valid(slave->data->phy_of_handle)) phydev->node = slave->data->phy_of_handle;
#endif

Quoted 3 lines

As you can see from above - the PHY probe will be called when both phy_device->dev and phydev->node have not been initialized yet, so no way to perform DT parsing in PHY .probe().

That is certainly inconsistent with the way other kinds of devices work. What I did for DT based configuration on aquantia PHYs was to read the DT properties in _config. Just from a practical stand-point that's fine, phydev->node is presumably set up by now and ethernet drivers have to call _config anyway and that gives the PHY driver a chance to check the DT properties. But conceptually it is messy, it makes phy drivers special. I suppose the solution is to make phy drivers behave like the other udevices, either actually making them udevices or at least introducing phy_bind ahead of probe.

On the current code, do you see any issues with dealing with the PHY DT properties in _config, other than this inconsistency with the other devices?

Quoted 2 lines

Another, not obvious, thing is that DT node for fixed phy passed through the int addr parameter in create_phy_by_mask().

This is a work-around for the previous issue I assume.

Quoted 3 lines

And finally, as phy as fixed-phy may be connected to DT node which is *not" ethernet device node and describes Port. The ethernet device is parent DT node in this case.

Are these switch ports or on a multi-port Ethernet card? If that's the case I would say the better solution is to actually drive ports as ethernet devices in that case, rather than driving the switch as a whole as an Ethernet interface. I realize this approach may not work in all cases, for various reasons, but then it's a matter of drivers that are different not getting the benefit of using high level helper functions that don't work for them. At least we should make sure such drivers can set things up properly if they are using custom binding or no DT at all. On the switch topic I just sent some patches to support DSA switches and the code drives ports as ethernet devices. Perhaps the code is not applicable to your device but I'd like to hear whether it makes sense at the concept level.

Quoted 3 lines

In my opinion, it could really simplify things if phy_device->dev and
phydev->node will be configured before calling PHY .probe(), but the
challenge is to make it work for !DT case.

I think the nice way to fix that would be to introduce a bind function for phy drivers, I could try to look into that after I free up my todo list a little bit.

Thanks!
Alex

#12 lore raw

PATCH v4 5/5 drivers: net: fsl_enetc: use the new MDIO DM helper functions

In reply to Alex Marginean

Uses the new dm_eth_phy_connect helper to connect to the PHY to simplify the code.

---
 drivers/net/fsl_enetc.c | 53 +++++++----------------------------------
 drivers/net/fsl_enetc.h |  1 +
 2 files changed, 10 insertions(+), 44 deletions(-)
drivers/net/fsl_enetc.c +9 -44
index 0ca7e838a8..e1713a3337 100644
@@ -203,57 +203,20 @@ static void enetc_start_pcs(struct udevice *dev)
 }
 
 /* Configure the actual/external ethernet PHY, if one is found */
-static void enetc_start_phy(struct udevice *dev)
+static void enetc_config_phy(struct udevice *dev)
 {
 	struct enetc_priv *priv = dev_get_priv(dev);
-	struct udevice *miidev;
-	struct phy_device *phy;
-	u32 phandle, phy_id;
-	ofnode phy_node;
 	int supported;
 
-	if (!ofnode_valid(dev->node)) {
-		enetc_dbg(dev, "no enetc ofnode found, skipping PHY set-up\n");
-		return;
-	}
-
-	if (ofnode_read_u32(dev->node, "phy-handle", &phandle)) {
-		enetc_dbg(dev, "phy-handle not found, skipping PHY set-up\n");
-		return;
-	}
-
-	phy_node = ofnode_get_by_phandle(phandle);
-	if (!ofnode_valid(phy_node)) {
-		enetc_dbg(dev, "invalid phy node, skipping PHY set-up\n");
-		return;
-	}
-	enetc_dbg(dev, "phy node: %s\n", ofnode_get_name(phy_node));
+	priv->phy = dm_eth_phy_connect(dev);
 
-	if (ofnode_read_u32(phy_node, "reg", &phy_id)) {
-		enetc_dbg(dev,
-			  "missing reg in PHY node, skipping PHY set-up\n");
+	if (!priv->phy)
 		return;
-	}
-
-	if (uclass_get_device_by_ofnode(UCLASS_MDIO,
-					ofnode_get_parent(phy_node),
-					&miidev)) {
-		enetc_dbg(dev, "can't find MDIO bus for node %s\n",
-			  ofnode_get_name(ofnode_get_parent(phy_node)));
-		return;
-	}
-
-	phy = dm_mdio_phy_connect(miidev, phy_id, dev, priv->if_type);
-	if (!phy) {
-		enetc_dbg(dev, "dm_mdio_phy_connect returned null\n");
-		return;
-	}
 
 	supported = GENMASK(6, 0); /* speeds up to 1G & AN */
-	phy->advertising = phy->supported & supported;
-	phy->node = phy_node;
-	phy_config(phy);
-	phy_startup(phy);
+	priv->phy->advertising = priv->phy->supported & supported;
+
+	phy_config(priv->phy);
 }
 
 /*
@@ -468,7 +431,9 @@ static int enetc_start(struct udevice *dev)
 	enetc_setup_rx_bdr(dev);
 
 	enetc_start_pcs(dev);
-	enetc_start_phy(dev);
+	enetc_config_phy(dev);
+	if (priv->phy)
+		phy_startup(priv->phy);
 
 	return 0;
 }
drivers/net/fsl_enetc.h +1 -0
index 0bb4cdff47..9a36cdad80 100644
@@ -154,6 +154,7 @@ struct enetc_priv {
 
 	int if_type;
 	struct mii_dev imdio;
+	struct phy_device *phy;
 };
 
 /* register accessors */
Signature · 1 line
2.17.1
#13 lore raw

v4 cover Add helper function for linking a DM Eth device to a PHY

In reply to Alex Marginean

Hi,

sorry for being so late..

Quoted 1 line

We're going to try to get this binding accepted in Linux too.

shouldn't we try to get it accepted to linux first? To avoid any incompatibilities?

-michael

#14 lore raw

v4 cover Add helper function for linking a DM Eth device to a PHY

Hi Michael,

Quoted 8 lines · Michael Walle

Hi,

sorry for being so late..

Quoted 1 line

We're going to try to get this binding accepted in Linux too.

shouldn't we try to get it accepted to linux first? To avoid any incompatibilities?

I can propagate any changes kernel crows asks us to do here, that shouldn't be too difficult. Of course any feedback from the U-Boot community is welcome too, in the meantime :) Alex

Quoted 7 lines

-michael
_______________________________________________
U-Boot mailing list
U-Boot at lists.denx.de
https://lists.denx.de/listinfo/u-boot

#15 lore raw

v4 cover Add helper function for linking a DM Eth device to a PHY

Quoted 6 lines · Alexandru Marginean

Hi Michael,

Quoted 3 lines · Michael Walle

Hi,

Quoted 7 lines
Quoted 5 lines

sorry for being so late..

Quoted 1 line

We're going to try to get this binding accepted in Linux too.

shouldn't we try to get it accepted to linux first? To avoid any incompatibilities?

I can propagate any changes kernel crows asks us to do here, that

that was a weird typo. I meant 'kernel crowd' :)

Quoted 11 lines

shouldn't be too difficult.  Of course any feedback from the U-Boot community is welcome too, in the meantime :) Alex

Quoted 7 lines

-michael
_______________________________________________
U-Boot mailing list
U-Boot at lists.denx.de
https://lists.denx.de/listinfo/u-boot

#16 lore raw

v4 cover Add helper function for linking a DM Eth device to a PHY

In reply to Alex Marginean

Am 2019-11-14 16:04, schrieb Alex Marginean:

Quoted 44 lines

The patch set introduces dm_eth_phy_connect which takes in an ethernet device
and uses DT information to find the associated PHY and connect the Ethernet
interface to it. This should simplify similar code in ethernet drivers.
A new binding is introduced for scanning an MDIO bus, useful if the bus is
private to the Ethernet interface. We're going to try to get this binding
accepted in Linux too.
Finally the patch set updates fsl_enetc driver to use the new helper function.

This patch set supersedes v2 series:
https://patchwork.ozlabs.org/project/uboot/list/?series=124537
and v3 series:
https://patchwork.ozlabs.org/project/uboot/list/?series=140114

Changes in v4:
- rebased on current head

Changes in v3:
- added cover letter
- check for null PHY pointer before using it in
dm_eth_connect_phy_handle
- moved the code dealing with MDIO scanning into a separate patch
- renames several arguments and variables for a bit more clarity and consistency

Changes in v2:
- Moved MDIO scan code into dm_mdio_phy_scan which is also exported
- Use interface instead of if_type for consistency
- don't use phy pointer if NULL in fsl_enetc code

Alex Marginean (5):
net: mdio-uclass: rename arguments of dm_mdio_phy_connect for
consistency
net: mdio-uclass: add dm_eth_phy_connect helper function
net: mdio-uclass: Add helper functions for scanning the MDIO bus
doc: bindings: add mdio-handle property to ethernet nodes
drivers: net: fsl_enetc: use the new MDIO DM helper functions

Quoted 6 lines
 doc/device-tree-bindings/net/ethernet.txt |   2 +
 drivers/net/fsl_enetc.c                   |  52 ++------
 drivers/net/fsl_enetc.h                   |   1 +
 include/miiphy.h                          |  32 ++++-
 net/mdio-uclass.c                         | 139 +++++++++++++++++++++-
 5 files changed, 175 insertions(+), 51 deletions(-)

For this series: