[PATCH 1/4] powerpc: document max-speed and interface-type properties

[PATCH 1/4] powerpc: document max-speed and interface-type properties

[Kim Phillips]

Since ucc_geth is being migrated to use the phylib, the existing
(undocumented) 'interface' property is being deprecated in favour
of unconjoined variations 'max-speed' and 'interface-type'.

max-speed is necessary for QE UEC devices that do not make that
information available through their programming model.

interface-type is now maintained one-to-one with definitions
in include/linux/phy.h.

Signed-off-by: Kim Phillips <kim.phillips@freescale.com>
---
please consider for 2.6.22

 Documentation/powerpc/booting-without-of.txt |    5 +++++
 1 files changed, 5 insertions(+), 0 deletions(-)

diff --git a/Documentation/powerpc/booting-without-of.txt b/Documentation/powerpc/booting-without-of.txt
index 88cdb59..92fa74c 100644
--- a/Documentation/powerpc/booting-without-of.txt
+++ b/Documentation/powerpc/booting-without-of.txt
@@ -1556,6 +1556,9 @@ platforms are moved over to use the flattened-device-tree model.
    Required properties for network device_type:
    - mac-address : list of bytes representing the ethernet address.
    - phy-handle : The phandle for the PHY connected to this controller.
+   - max-speed : The maximum speed supported by the controller (in Mbit/s)
+   - interface-type : a string naming the controller/PHY interface type,
+     i.e., one of "mii", "gmii", "sgmii", "tbi", "rmii", "rgmii", or "rtbi".
 
    Recommended properties:
    - linux,network-index : This is the intended "index" of this
@@ -1573,6 +1576,8 @@ platforms are moved over to use the flattened-device-tree model.
 		interrupts = <a0 0>;
 		interrupt-parent = <700>;
 		mac-address = [ 00 04 9f 00 23 23 ];
+		interface-type = "rgmii";
+		max-speed = <d#1000>;
 		rx-clock = "none";
 		tx-clock = "clk9";
 		phy-handle = <212000>;
-- 
1.5.0.1

Re: [PATCH 1/4] powerpc: document max-speed and interface-type properties

[Olof Johansson]

On Tue, Apr 10, 2007 at 04:56:39PM -0500, Kim Phillips wrote:
> Since ucc_geth is being migrated to use the phylib, the existing
> (undocumented) 'interface' property is being deprecated in favour
> of unconjoined variations 'max-speed' and 'interface-type'.
> 
> max-speed is necessary for QE UEC devices that do not make that
> information available through their programming model.

Do you really need to set max speed for devices that are at the default
max speed for the interface type? Seems redundant.

If for example there's an rgmii interface that for some reason only can
handle 100Mbit, that'd be different.


-Olof

Re: [PATCH 1/4] powerpc: document max-speed and interface-type properties

[Kim Phillips]

On Tue, 10 Apr 2007 17:48:09 -0500
olof@lixom.net (Olof Johansson) wrote:

> 
> Do you really need to set max speed for devices that are at the default
> max speed for the interface type? Seems redundant.
> 
> If for example there's an rgmii interface that for some reason only can
> handle 100Mbit, that'd be different.

rgmii@100/10Mbit are supported interface/speed combinations for our MDS boards; see pdf p. 72 in:

http://www.freescale.com/files/32bit/doc/user_guide/MPC8360EAMDSUM.pdf

Kim

Re: [PATCH 1/4] powerpc: document max-speed and interface-type properties

[Olof Johansson]

On Tue, Apr 10, 2007 at 06:19:49PM -0500, Kim Phillips wrote:
> On Tue, 10 Apr 2007 17:48:09 -0500
> olof@lixom.net (Olof Johansson) wrote:
> 
> > 
> > Do you really need to set max speed for devices that are at the default
> > max speed for the interface type? Seems redundant.
> > 
> > If for example there's an rgmii interface that for some reason only can
> > handle 100Mbit, that'd be different.
> 
> rgmii@100/10Mbit are supported interface/speed combinations for our MDS boards; see pdf p. 72 in:
> 
> http://www.freescale.com/files/32bit/doc/user_guide/MPC8360EAMDSUM.pdf

Right, my bad. Sorry for the confusion -- That comment was meant for the
next patch in the series where you specify the speed even for interfaces
that support the full max.


-Olof

Re: [PATCH 1/4] powerpc: document max-speed and interface-type properties

[Kim Phillips]

On Tue, 10 Apr 2007 20:49:58 -0500
olof@lixom.net (Olof Johansson) wrote:

> On Tue, Apr 10, 2007 at 06:19:49PM -0500, Kim Phillips wrote:
> > On Tue, 10 Apr 2007 17:48:09 -0500
> > olof@lixom.net (Olof Johansson) wrote:
> > 
> > > 
> > > Do you really need to set max speed for devices that are at the default
> > > max speed for the interface type? Seems redundant.
> > > 
> > > If for example there's an rgmii interface that for some reason only can
> > > handle 100Mbit, that'd be different.
> > 
> > rgmii@100/10Mbit are supported interface/speed combinations for our MDS boards; see pdf p. 72 in:
> > 
> > http://www.freescale.com/files/32bit/doc/user_guide/MPC8360EAMDSUM.pdf
> 
> Right, my bad. Sorry for the confusion -- That comment was meant for the
> next patch in the series where you specify the speed even for interfaces
> that support the full max.
> 
just to be clear, you want the logic in ucc_geth (phylib patch) to be enhanced such that it sets the maximum max-speed based on the interface-type, e.g. if rgmii specified and max-speed not specified, assign max-speed = 1000?

Kim

Re: [PATCH 1/4] powerpc: document max-speed and interface-type properties

[Olof Johansson]

On Wed, Apr 11, 2007 at 01:09:13PM -0500, Kim Phillips wrote:
> On Tue, 10 Apr 2007 20:49:58 -0500
> olof@lixom.net (Olof Johansson) wrote:
> 
> > On Tue, Apr 10, 2007 at 06:19:49PM -0500, Kim Phillips wrote:
> > > On Tue, 10 Apr 2007 17:48:09 -0500
> > > olof@lixom.net (Olof Johansson) wrote:
> > > 
> > > > 
> > > > Do you really need to set max speed for devices that are at the default
> > > > max speed for the interface type? Seems redundant.
> > > > 
> > > > If for example there's an rgmii interface that for some reason only can
> > > > handle 100Mbit, that'd be different.
> > > 
> > > rgmii@100/10Mbit are supported interface/speed combinations for our MDS boards; see pdf p. 72 in:
> > > 
> > > http://www.freescale.com/files/32bit/doc/user_guide/MPC8360EAMDSUM.pdf
> > 
> > Right, my bad. Sorry for the confusion -- That comment was meant for the
> > next patch in the series where you specify the speed even for interfaces
> > that support the full max.
> > 
> just to be clear, you want the logic in ucc_geth (phylib
> patch) to be enhanced such that it sets the maximum max-speed
> based on the interface-type, e.g. if rgmii specified and
> max-speed not specified, assign max-speed = 1000?

Yes, that would seem logical, wouldn't it?


-Olof

RE: [PATCH 1/4] powerpc: document max-speed and interface-type properties

[Joakim Tjernlund]

> -----Original Message-----
> From: 
> linuxppc-dev-bounces+joakim.tjernlund=transmode.se@ozlabs.org 
> [mailto:linuxppc-dev-bounces+joakim.tjernlund=transmode.se@ozl
> abs.org] On Behalf Of Kim Phillips
> Sent: den 10 april 2007 23:57
> To: linuxppc-dev@ozlabs.org
> Subject: [PATCH 1/4] powerpc: document max-speed and 
> interface-type properties
> 
> Since ucc_geth is being migrated to use the phylib, the existing
> (undocumented) 'interface' property is being deprecated in favour
> of unconjoined variations 'max-speed' and 'interface-type'.
> 
> max-speed is necessary for QE UEC devices that do not make that
> information available through their programming model.
> 
> interface-type is now maintained one-to-one with definitions
> in include/linux/phy.h.
> 
> Signed-off-by: Kim Phillips <kim.phillips@freescale.com>
> ---
> please consider for 2.6.22
> 
>  Documentation/powerpc/booting-without-of.txt |    5 +++++
>  1 files changed, 5 insertions(+), 0 deletions(-)
> 
> diff --git a/Documentation/powerpc/booting-without-of.txt 
> b/Documentation/powerpc/booting-without-of.txt
> index 88cdb59..92fa74c 100644
> --- a/Documentation/powerpc/booting-without-of.txt
> +++ b/Documentation/powerpc/booting-without-of.txt
> @@ -1556,6 +1556,9 @@ platforms are moved over to use the 
> flattened-device-tree model.
>     Required properties for network device_type:
>     - mac-address : list of bytes representing the ethernet address.
>     - phy-handle : The phandle for the PHY connected to this 
> controller.
> +   - max-speed : The maximum speed supported by the 
> controller (in Mbit/s)
> +   - interface-type : a string naming the controller/PHY 
> interface type,
> +     i.e., one of "mii", "gmii", "sgmii", "tbi", "rmii", 
> "rgmii", or "rtbi".

Hmm, what if I have a direct MII to MII connection without a PHY in between?
Will this support such a "fixed" connection?
How do I express that in OF speak?

 Jocke

Re: [PATCH 1/4] powerpc: document max-speed and interface-type properties

[Kim Phillips]

On Wed, 11 Apr 2007 23:45:36 +0200
"Joakim Tjernlund" <joakim.tjernlund@transmode.se> wrote:

> 
> > +   - interface-type : a string naming the controller/PHY 
> > interface type,
> > +     i.e., one of "mii", "gmii", "sgmii", "tbi", "rmii", 
> > "rgmii", or "rtbi".
> 
> Hmm, what if I have a direct MII to MII connection without a PHY in between?

that would depend on the connection itself (and how you wanted your UCC configured).  It would still be one of the above, since those are what the UCC supports.

> Will this support such a "fixed" connection?

I don't see why not.

> How do I express that in OF speak?

omit the PHY node?

Kim

Re: [PATCH 1/4] powerpc: document max-speed and interface-type properties

[Joakim Tjernlund]

On Wed, 2007-04-11 at 17:56 -0500, Kim Phillips wrote:
> On Wed, 11 Apr 2007 23:45:36 +0200
> "Joakim Tjernlund" <joakim.tjernlund@transmode.se> wrote:
> 
> > 
> > > +   - interface-type : a string naming the controller/PHY 
> > > interface type,
> > > +     i.e., one of "mii", "gmii", "sgmii", "tbi", "rmii", 
> > > "rgmii", or "rtbi".
> > 
> > Hmm, what if I have a direct MII to MII connection without a PHY in between?
> 
> that would depend on the connection itself (and how you wanted your UCC configured).  It would still be one of the above, since those are what the UCC supports.
> 
Right, should be mii then.

> > Will this support such a "fixed" connection?
> 
> I don't see why not.
Great.
> 
> > How do I express that in OF speak?
> 
> omit the PHY node?
> 
hmm, don't think so. I need to define a pseudo PHY of sorts but I
don't know how yet.

 Jocke

> Kim
> 
> 

Re: [PATCH 1/4] powerpc: document max-speed and interface-type properties

[Segher Boessenkool]

> +   - max-speed : The maximum speed supported by the controller (in 
> Mbit/s)
> +   - interface-type : a string naming the controller/PHY interface 
> type,
> +     i.e., one of "mii", "gmii", "sgmii", "tbi", "rmii", "rgmii", or 
> "rtbi".

This information should be in the PHY node, instead.
Both property values would normally be implied by
"compatible" in that node (and this node, for the
max speed).


Segher

Re: [PATCH 1/4] powerpc: document max-speed and interface-type properties

[Andy Fleming]

On Apr 12, 2007, at 02:06, Segher Boessenkool wrote:

>> +   - max-speed : The maximum speed supported by the controller (in
>> Mbit/s)
>> +   - interface-type : a string naming the controller/PHY interface
>> type,
>> +     i.e., one of "mii", "gmii", "sgmii", "tbi", "rmii", "rgmii", or
>> "rtbi".
>
> This information should be in the PHY node, instead.
> Both property values would normally be implied by
> "compatible" in that node (and this node, for the
> max speed).


Why should it be in the PHY node?  It's not a property of the PHY.   
It's a property of the connection between the PHY and the ethernet  
controller.  The PHY Lib API has the ethernet provide the connection  
type as an argument in the phy_connect() call, so it makes sense for  
this property to reside in the ethernet node.  I think ethernet  
drivers typically have the most power in this situation, and so it's  
best (IMHO) to let the ethernet node provide this information.

Andy

Re: [PATCH 1/4] powerpc: document max-speed and interface-type properties

[Kumar Gala]

On Apr 16, 2007, at 12:28 PM, Andy Fleming wrote:

>
> On Apr 12, 2007, at 02:06, Segher Boessenkool wrote:
>
>>> +   - max-speed : The maximum speed supported by the controller (in
>>> Mbit/s)
>>> +   - interface-type : a string naming the controller/PHY interface
>>> type,
>>> +     i.e., one of "mii", "gmii", "sgmii", "tbi", "rmii",  
>>> "rgmii", or
>>> "rtbi".
>>
>> This information should be in the PHY node, instead.
>> Both property values would normally be implied by
>> "compatible" in that node (and this node, for the
>> max speed).
>
>
> Why should it be in the PHY node?  It's not a property of the PHY.
> It's a property of the connection between the PHY and the ethernet
> controller.  The PHY Lib API has the ethernet provide the connection
> type as an argument in the phy_connect() call, so it makes sense for
> this property to reside in the ethernet node.  I think ethernet
> drivers typically have the most power in this situation, and so it's
> best (IMHO) to let the ethernet node provide this information.

The interface type should be a property of the phy because its a  
property of the physical 'phy' device on the board.

- k