Geode SC1100 i2c bus

Milan Obuch freebsd-hackers at dino.sk
Sun Feb 4 08:20:01 UTC 2007


On Sunday 04 February 2007 01:21, perryh at pluto.rain.com wrote:
> > > The ACB is at the same level than Uart, so I had a look at sio
> > > source and got a big headache. I'm not really used to drivers
> > > nor kernel programming stuff, that's why I need your help.
> >
> > Why do you compare ACB and UART? In FreeBSD, sio is serial i/o
> > controller, not superIO, maybe you are confused... ACB and UART
> > does not have much in common...
>
> As I read the OP, both the ACB and the serial are in the same chip
> (the superIO), so it would seem reasonable to expect that the method
> of accessing the device registers in the (currently unsupported) ACB
> might be similar to the method of accessing the device registers in
> the (supported) serial port.
>

You are right from the point of view 'how to access registers'. But what OP 
wrote could be understand another way, too - he tried to modify sio to create 
acb driver, which would not achieve the correct results. There is i2c 
infrastructure in FreeBSD, acb driver should fit into its place there and it 
is not that easy - I tried to understand it, but nobody could/was willing 
to/did not care to help me understand it.

> Note I said "reasonable" -- it may not be "accurate" -- and
> this sort of analysis applies only to getting at the hardware.
> Certainly the means of communicating with the rest of the OS
> would be different, unless one wants the ACB to show up as sio7
> or some such :)
>

At hardware level, I consider it to be 'accurate', at OS level, they are 
completely different. Globally, at looks to me we basically do not disagree 
on subject, but that's just langauage ambiguity out there...

Regards,
Milan

-- 
No need to mail me directly. Just reply to mailing list, please.


More information about the freebsd-hackers mailing list