[ECOS] AT91SAM and I2C

[ECOS] AT91SAM and I2C

[Jan Jacobs]

Hi All,

I need an I2C (TWI) driver for the AT91SAM7X. I have started doing 
something, but want to find out if somebody has done one or has any advice?

Secondly, I think there might be a flaw with the I2C IO package.
If the device you communication to does not use internal addressing (RTC, 
Sensor, etc) then the I2C IO API works.
If the device requires internal addressing (EEPROM) I do not see how that 
gets conveyed via the I2C IO API?

I must state that I am new to ecos and the AT91SAM7x, so if I ask stupid 
questions humour me :)

Cheers
Jan 


-- 
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] AT91SAM and I2C

[Andrew Lunn]

On Sun, Jan 06, 2008 at 11:01:51PM +0200, Jan Jacobs wrote:
> Hi All,
>
> I need an I2C (TWI) driver for the AT91SAM7X. I have started doing 
> something, but want to find out if somebody has done one or has any advice?

I think somebody is working on such a driver. It was mentioned last
month. Take a look in the mail archive.

> Secondly, I think there might be a flaw with the I2C IO package.
> If the device you communication to does not use internal addressing (RTC, 
> Sensor, etc) then the I2C IO API works.
> If the device requires internal addressing (EEPROM) I do not see how that 
> gets conveyed via the I2C IO API?

I don't really follow. Do you want to make it look like the EEPROM is
memory mapped?

       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] AT91SAM and I2C

[Laurie Gellatly]

-----Original Message-----
From: ecos-discuss-owner@ecos.sourceware.org
[mailto:ecos-discuss-owner@ecos.sourceware.org]On Behalf Of Andrew Lunn
Sent: Monday, 7 January 2008 8:12 AM
To: Jan Jacobs
Cc: ecos-discuss@ecos.sourceware.org
Subject: Re: [ECOS] AT91SAM and I2C


On Sun, Jan 06, 2008 at 11:01:51PM +0200, Jan Jacobs wrote:
> Hi All,
>
> I need an I2C (TWI) driver for the AT91SAM7X. I have started doing
> something, but want to find out if somebody has done one or has any
advice?

>I think somebody is working on such a driver. It was mentioned last
>month. Take a look in the mail archive.

> Secondly, I think there might be a flaw with the I2C IO package.
> If the device you communication to does not use internal addressing (RTC,
> Sensor, etc) then the I2C IO API works.
> If the device requires internal addressing (EEPROM) I do not see how that
> gets conveyed via the I2C IO API?

>I don't really follow. Do you want to make it look like the EEPROM is
>memory mapped?

Jan,
When you talk to a EEPROM then part of the I2C message is a command followed
by the memory address then the data is exchanged all across teh I2C bus.
You should see this in the Datasheet.

(apologies for duplicate message Andrew).

			...Laurie:{)


-- 
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] AT91SAM and I2C

[Tom Deconinck]

On Jan 6, 2008 10:28 PM, Laurie Gellatly <laurie.gellatly@netic.com> wrote:
> -----Original Message-----
> From: ecos-discuss-owner@ecos.sourceware.org
> [mailto:ecos-discuss-owner@ecos.sourceware.org]On Behalf Of Andrew Lunn
> Sent: Monday, 7 January 2008 8:12 AM
> To: Jan Jacobs
> Cc: ecos-discuss@ecos.sourceware.org
> Subject: Re: [ECOS] AT91SAM and I2C
>
>
> On Sun, Jan 06, 2008 at 11:01:51PM +0200, Jan Jacobs wrote:
> > Hi All,
> >
> > I need an I2C (TWI) driver for the AT91SAM7X. I have started doing
> > something, but want to find out if somebody has done one or has any
> advice?
>
> >I think somebody is working on such a driver. It was mentioned last
> >month. Take a look in the mail archive.
>
> > Secondly, I think there might be a flaw with the I2C IO package.
> > If the device you communication to does not use internal addressing (RTC,
> > Sensor, etc) then the I2C IO API works.
> > If the device requires internal addressing (EEPROM) I do not see how that
> > gets conveyed via the I2C IO API?
>
> >I don't really follow. Do you want to make it look like the EEPROM is
> >memory mapped?
>
> Jan,
> When you talk to a EEPROM then part of the I2C message is a command followed
> by the memory address then the data is exchanged all across teh I2C bus.
> You should see this in the Datasheet.
>
> (apologies for duplicate message Andrew).
>
>                         ...Laurie:{)
>
>
>
> --
> Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
> and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss
>
>
Hi,

I'm currently working on and off an I2C driver for the AT91SAM7S family.
I have started with the framework provided by the Opencores I2C driver
(basically, because that's a driver someone developed at the company I
work and he's only sitting 5m away ;) ) and I've adapted that code for
use with the AT91 I2C controller.
The write side is working: interrupt driven, no major data transfer
tested yet, this could be cumbersome as Andrew warned me, but I
haven't got the time to test this yet.

I have problems with reading, because right now, I seem to miss the
data received interrupts of the TWI controller. I've got some support
code from Atmel, but again, due to lack of time I didn't got the
chance to look into the issue any further. Right now, I can start a
read transfer, but it just keeps on reading (that's due to the rather
strange implementations decisions Atmel has made.... )
No CDL or anything alike yet.

For information purposes: I hooked up an Atmel EEPROM (AT24C64A) to
the CPU to test my driver code.

If someone's interested in the code I've got this far, I'm willing to
share ;), but keep in mind that all is still very raw.

Regards,
Tom

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