[ECOS] AT91 I2C driver

[ECOS] AT91 I2C driver

[Steven Clugston]

Is anybody out there interested in having/testing an AT91 I2C driver?

I'm currently having a stab at writing one using the lpc2xxx one as a reference, although I'm not sure if I'll actually have time to see it through to completion yet.

Just though I would post my intentions.

Regards,

Steven

--
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss

Re: [ECOS] AT91 I2C driver

[Andrew Lunn]

On Wed, Oct 28, 2009 at 11:27:36AM +0000, Steven Clugston wrote:
> Is anybody out there interested in having/testing an AT91 I2C driver?
> 
> I'm currently having a stab at writing one using the lpc2xxx one as
> a reference, although I'm not sure if I'll actually have time to see
> it through to completion yet.

A word of caution. The Linux kernel guys consider the AT91 I2C device
to be FUBAR. 

http://cateee.net/lkddb/web-lkddb/I2C_AT91.html
http://www.mail-archive.com/git-commits-head@vger.kernel.org/msg24797.html
http://www.mail-archive.com/i2c@lm-sensors.org/msg02209.html

I think it is O.K. for single byte transfers, but anything more than
that could lead to problems.

Maybe you should consider bit banging the pins?

      Andrew

-- 
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss

RE: [ECOS] AT91 I2C driver

[Steven Clugston]

>> Is anybody out there interested in having/testing an AT91 I2C driver?
>>
>> I'm currently having a stab at writing one using the lpc2xxx one as
>> a reference, although I'm not sure if I'll actually have time to see
>> it through to completion yet.

> A word of caution. The Linux kernel guys consider the AT91 I2C device
> to be FOOBAR.

> http://cateee.net/lkddb/web-lkddb/I2C_AT91.html
> http://www.mail-archive.com/git-commits-head@vger.kernel.org/msg24797.html
> http://www.mail-archive.com/i2c@lm-sensors.org/msg02209.html

> I think it is O.K. for single byte transfers, but anything more than
> that could lead to problems.

> Maybe you should consider bit banging the pins?

>      Andrew

Andrew,

Thanks very much for pointing this out. It must explain why nobody has written the driver before.
I was in two minds whether to just try the generic bit bang driver instead, but I thought it was worth an attempt at using the chip's hardware first particularly if anybody else was interested.

I've shot myself in the foot now because it turns out I do need multi-byte transfers and I've used a part that has a timeout of 35ms on the clock which might make bit-bang problematic if its too slow for some reason.

Steven

--
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss