Re: New I2C and machine probing method
On Sat, 22 Jul 2006 13:22:11 +0200
Johannes Berg <johannes@sipsolutions.net> wrote:
> On Fri, 2006-07-21 at 17:33 +0200, Johannes Berg wrote:
> > I2C: 'i2c-7', 'uni-n 0' probing /dev/i2c-7 ... gotcha, this is the
> > LMU device I2C: 'i2c-6', 'mac-io 0'
>
> Btw, this makes it seem that you don't understand how the i2c devices
> work; /dev/i2c-7 isn't the LMU device, /dev/i2c-7 is the device node for
> the bus that the LMU is on.
What are you trying to tell me? Do you want to argue about the wording
in my test program? In this case you will win. ;-)
I surely know that the /dev/i2c-7 is not the LMU device itself. It's
the interface to the I2C controller built into the uni-n memory
controller and bus bridge IC the LMU is connected to. But this makes no
difference for the normal user who wants that this thing simply works.
I2C is a two line serial bus invented and patented by Philips and
up to 400 KBit fast, where you can attach multiply devices to and each
device could be addressed by a unique (on that specific bus) device
address. I read this address from the open firmware device-tree and use
it to connect to the LMU via /dev/i2c-7. It would be much better if I
could read the address somewhere from /sys, but it isn't provided
by /sys yet.
Any further questions? ;-)
Best regards
Matthias
Reply to: