RE: [RFC PATCH 0/2] net: ethernet: Add support for gmii2rgmii converter

[Appana Durga Kedareswara Rao]

Hi Florian,

	Thanks for the review...

> Le 30/06/2016 23:20, Kedareswara rao Appana a Ãcrit :
> > This patch series does the following
> > ---> Add support for gmii2rgmii converter.
> > ---> Add support for gmii2rgmii converter in the macb driver.
> >
> > Earlier sent one RFC patch https://patchwork.kernel.org/patch/9186615/
> > which includes converter related stuff also in macb driver.
> > This patch series fixes this issue.
> 
> This still seems very adhoc and not completely explained. Can you clarify how
> the gmmi2rgmii converter gets used?

Sorry I should have been explained it clearly in the patch.
Will fix it in the v2.

This converter works like below

MACB <==> GMII2RGMII <==> RGMII_PHY

	MDIO	    <========> GMII2RGMII
MCAB<=======> 
		  <========> RGMII

Using MACB MDIO bus we can access both the converter and the external PHY.
We need to program the line speed of the converter during run time based on the
External phy negotiated speed. 

MDIO interface is used to set operating speed (line speed) 

The converter has only one register (0x10) that we need to program to set the operating speed.
The composition of this register is similar to the IEEE standard 802.3 MDIO control register 0x0.

Please let me know if you still need not clear with how this converter works.

IP data sheet is available here @ http://www.xilinx.com/support/documentation/ip_documentation/gmii_to_rgmii/v3_0/pg160-gmii-to-rgmii.pdf 

		   
> 
> Is the expectation that a MACB Ethernet adapter will be connected to a RGMII
> PHY like this:
> 
> MACB <=> GMII2RGMII <=> RGMII PHY
> MACB MDIO <===========> RGMII_PHY
> 
> and still have the ability to control both the PHY device's MDIO registers and the
> GMII2RGMII converter and we need to make sure both have matching settings,
> or is it something like this:
> 
> MACB <=> GMII2RGMII <=> RGMII PHY
> MACB MDIO unconnected
> 
> and we lose the ability to query the PHY via MDIO?
> 
> As Nicolas pointed out, providing a binding documentation update may help
> reviewers understand what is being accomplished here.

Sure will fix in v2.

Regards,
Kedar.