Re: [v2] powerpc/fsl: Update fman dt binding with pcs-phy and tbi-phy

On Thu, Dec 24, 2015 at 03:42:11AM +0200, igal.liberman@freescale.com wrote:

From: Igal Liberman <redacted>

The FMan contains internal PHY devices used for SGMII connections
to external PHYs. When these PHYs are in use a reference is needed
for both the external PHY and the internal one. For the external
PHY phy-handle provides the reference. For the internal PHY a new
handle is required.
In dTSEC, the internal PHY is a TBI (Ten Bit Interface) PHY,
the handle used will be tbi-handle.
In mEMAC, the internal PHY is a PCS (Physical Coding Sublayer) PHY,
the handle used will be pcsphy-handle.

This is fairly commom for 10G eth I think. Can't you use the common PHY 
binding here in the case without internal MDIO bus? Just because you use 
it that doesn't mean you have to use the generic phy subsystem in the 
kernel. 

Perhaps phy-handle should be deprecated in favor of doing something like 
this if you need a phandle to both:

phys = <&internal-phy>, <&external-phy>;

What does the tbi0 node contain? It should be present in the example.

Rob

Hi Rob,

-----Original Message-----
From: Rob Herring [mailto:robh@kernel.org]
Sent: Wednesday, December 30, 2015 6:28 PM
To: Igal Liberman <redacted>
Cc: devicetree@vger.kernel.org; linuxppc-dev@lists.ozlabs.org; Scott Wood
[off-list ref]; Madalin-Cristian Bucur
[off-list ref]; shaohui.xie@freescale.com
Subject: Re: [v2] powerpc/fsl: Update fman dt binding with pcs-phy and tb=

i-

phy
=20
On Thu, Dec 24, 2015 at 03:42:11AM +0200, igal.liberman@freescale.com
wrote:

quoted

From: Igal Liberman <redacted>

The FMan contains internal PHY devices used for SGMII connections to
external PHYs. When these PHYs are in use a reference is needed for
both the external PHY and the internal one. For the external PHY
phy-handle provides the reference. For the internal PHY a new handle
is required.
In dTSEC, the internal PHY is a TBI (Ten Bit Interface) PHY, the
handle used will be tbi-handle.
In mEMAC, the internal PHY is a PCS (Physical Coding Sublayer) PHY,
the handle used will be pcsphy-handle.

=20
This is fairly commom for 10G eth I think. Can't you use the common PHY
binding here in the case without internal MDIO bus? Just because you use =

it

that doesn't mean you have to use the generic phy subsystem in the kernel=

.

=20

mEMAC and dTSEC always have internal MDIO bus.
I was requested to use generic PHY API for internal PHY configuration by ne=
tdev, this part was accepted.=20

Perhaps phy-handle should be deprecated in favor of doing something like
this if you need a phandle to both:
=20
phys =3D <&internal-phy>, <&external-phy>;
=20

I think that pcsphy-handle and tbi-handle represent the hardware in a good =
way.
This is the actual name of the hardware.

quoted

diff --git a/Documentation/devicetree/bindings/powerpc/fsl/fman.txt

b/Documentation/devicetree/bindings/powerpc/fsl/fman.txt
index 1fc5328..55c2c03 100644

--- a/Documentation/devicetree/bindings/powerpc/fsl/fman.txt
+++ b/Documentation/devicetree/bindings/powerpc/fsl/fman.txt

@@ -315,6 +315,16 @@ PROPERTIES
 		Value type: <phandle>
 		Definition: A phandle for 1EEE1588 timer.

+- pcsphy-handle
+		Usage required for "fsl,fman-memac" MACs
+		Value type: <phandle>
+		Definition: A phandle for pcsphy.
+
+- tbi-handle
+		Usage required for "fsl,fman-dtsec" MACs
+		Value type: <phandle>
+		Definition: A phandle for tbiphy.
+
 EXAMPLE

 fman1_tx28: port@a8000 {

@@ -340,6 +350,7 @@ ethernet@e0000 {
 	reg =3D <0xe0000 0x1000>;
 	fsl,fman-ports =3D <&fman1_rx8 &fman1_tx28>;
 	ptp-timer =3D <&ptp-timer>;
+	tbi-handle =3D <&tbi0>;

=20
What does the tbi0 node contain? It should be present in the example.
=20

There is an example in under MDIO section:

mdio@e3120 {
	compatible =3D "fsl,fman-mdio";
	reg =3D <0xe3120 0xee0>;
	fsl,fman-internal-mdio;

	tbi1: tbi-phy@8 {
		reg =3D <0x8>;
		device_type =3D "tbi-phy";
	};
};

I used different indexes to make sure that it's aligned to the other exampl=
es in the binding document.

Rob

Thank you for your feedback,=20
Igal