[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <4E39C7A7.3000302@cavium.com>
Date: Wed, 03 Aug 2011 15:11:51 -0700
From: David Daney <david.daney@...ium.com>
To: devicetree-discuss@...ts.ozlabs.org,
Grant Likely <grant.likely@...retlab.ca>
CC: Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
linux-mips <linux-mips@...ux-mips.org>
Subject: Re: How to describe I2C and MDIO bus multiplexers with the device
tree?
On 08/03/2011 03:04 PM, David Daney wrote:
>
> I have a board (ASCII art representation below) that has a bunch of
> PHY devices that are connected behind a MDIO bus switch/multiplexer.
> This multiplexer is in turn controlled by GPIO pins on a I2C GPIO port
> expander. which is in turn on a I2C bus switch/multiplexer.
>
> There is another switch/multiplexer that is controlled directly via I2C.
>
> Q: How can this be represented?
>
> My attempt is below...
>
> I2C-0 MDIO-1 MDIO-0
> | | |
> | |--------------| | |
Crap, my mailer ate the ASCII art.
Attached is the text file that should make it through.
>
> Some notes/questions:
>
> o mdio-mux@0 is a child of soc@0, but conceptually it might just as
> well be floating in space.
>
> o The parent-bus property of both mdio-mux@0 and mdio-mux@77 denotes
> the connection of the multiplexer to its 'parent' bus.
>
> o The cell-index property of the multiplexed bus instances denotes the
> sub-bus of the switch/multiplexer the children are connected to.
> Would a different property name be better here?
>
> o For mdio-mux@0 the driver will use the cell-index to determine how
> to program the gpios that control the switch/multiplexer.
>
> Am I totally insane?
> David Daney
View attachment "multiplexer-1.txt" of type "text/plain" (9098 bytes)
Powered by blists - more mailing lists