problems running I2C and/or SPI on 857 board

problems running I2C and/or SPI on 857 board

[Stephan Linke]

Hi!

I'm trying to get I2C and/or SPI running on a 857 board. Register settings
and BDs look good. But when the start command is issued nearly nothing
happens as if there would be no clock.
TxBDs have ready bit set. Both ports initialize the internal BD pointer. SPI
is changing the clock line to inactive level (once). But that's all.
Couldn't find an errata that fits. Only the stuff with Parameter RAM
relocation but I can live with the default location...

Is there anyone who has an idea?

Thanks, Stephan

P.S.: Are the BD offsets counted from IMMR base, DPRam base or ...?


** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

Re: problems running I2C and/or SPI on 857 board

[Wolfgang Denk]

In message <FCEAKDJJAPHPLJFINDAJCEMMCDAA.Stephan.Linke@epygi.de> you wrote:
>
> I'm trying to get I2C and/or SPI running on a 857 board. Register settings
> and BDs look good. But when the start command is issued nearly nothing
> happens as if there would be no clock.

I guess your implementation is basedon existing code? Which one?
[Hint: we have I2C running on a couple of systems...]

> Is there anyone who has an idea?

Start from some code that is known to work?

Wolfgang Denk

--
Software Engineering:  Embedded and Realtime Systems,  Embedded Linux
Phone: (+49)-8142-4596-87  Fax: (+49)-8142-4596-88  Email: wd@denx.de
The existence of god implies a violation of causality.

** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

RE: problems running I2C and/or SPI on 857 board

[Stephan Linke]

At the moment all my tries are based on the existing drivers in PPCBOOT.
Using no microcode patch, and DPRAM allocation.
DPRAM allocation pointed out a bug so I guess it's not that well tested yet
:( .
Trying fixed DPRAM offsets and microcode now...

Thanks, Stephan

> -----Original Message-----
> From: wd@denx.de [mailto:wd@denx.de]
> Sent: Montag, 25. März 2002 14:11
> To: Stephan Linke
> Cc: Linuxppc-Embedded
> Subject: Re: problems running I2C and/or SPI on 857 board
>
>
> In message <FCEAKDJJAPHPLJFINDAJCEMMCDAA.Stephan.Linke@epygi.de>
> you wrote:
> >
> > I'm trying to get I2C and/or SPI running on a 857 board.
> Register settings
> > and BDs look good. But when the start command is issued nearly nothing
> > happens as if there would be no clock.
>
> I guess your implementation is basedon existing code? Which one?
> [Hint: we have I2C running on a couple of systems...]
>
> > Is there anyone who has an idea?
>
> Start from some code that is known to work?
>
> Wolfgang Denk
>
> --
> Software Engineering:  Embedded and Realtime Systems,  Embedded Linux
> Phone: (+49)-8142-4596-87  Fax: (+49)-8142-4596-88  Email: wd@denx.de
> The existence of god implies a violation of causality.
>


** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

Re: problems running I2C and/or SPI on 857 board

[Wolfgang Denk]

In message <FCEAKDJJAPHPLJFINDAJEENDCDAA.Stephan.Linke@epygi.de> you wrote:
>
> At the moment all my tries are based on the existing drivers in PPCBOOT.

Which version of PPCBoot? The official releases (1.1.5)  should  work
just fine; they do so on several tested systems.

If you're using the top of tree on the CVS server (i. e. the  version
which includes the "Grand Unifying I2C interface patch") you may have
to  adjust a couple of things; this is not that well tested / working
yet.

> Using no microcode patch, and DPRAM allocation.

I wonder why you use the CPM code at all.  I  always  found  it  much
easier to use CONFIG_SOFT_I2C ...

> DPRAM allocation pointed out a bug so I guess it's not that well tested yet
> :( .
> Trying fixed DPRAM offsets and microcode now...

Please submit a patch!

Wolfgang Denk

--
Software Engineering:  Embedded and Realtime Systems,  Embedded Linux
Phone: (+49)-8142-4596-87  Fax: (+49)-8142-4596-88  Email: wd@denx.de
The price of curiosity is a terminal experience.
                         - Terry Pratchett, _The Dark Side of the Sun_

** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

Re: problems running I2C and/or SPI on 857 board

[Wolfgang Denk]

In message <FCEAKDJJAPHPLJFINDAJIENECDAA.Stephan.Linke@epygi.de> you wrote:
> I'm using 1.1.4 at the moment. (Can't update for now.)
> I found the problem with I2C. Looks like I don't need the microcode patch.
> But I need to do the settings as if I would use the microcode patch. I don't
> know why but maybe the 857/862 ERRATA ESAR6 says something like that...
>
> I still think that there are some buggy pathes in the code. But let me
> see...

I'll happily accept patches...

> > I wonder why you use the CPM code at all.  I  always  found  it  much
> > easier to use CONFIG_SOFT_I2C ...
> >
> Well, I don't know the soft I2C. I guess, when we started playing with the
> 8260 there was no soft I2C. Is there a timing difference? Is there a soft
> SPI too?

There has always been a soft-i2c.  The  timing  is  dictated  by  the
external  hardware  (i2c  bus),  so  no,  there  is  not  much  of  a
difference. The main difference is where the CPU burns  it's  cycles:
with  soft-i2c  it's  running  the delay loops, and with the CPM it's
setting up the CPM stuff and waiting for the transfer to be done.

Not much practical difference,  if  you  ask  me.  And  the  soft-i2c
solution is much simpler...

Wolfgang Denk

--
Software Engineering:  Embedded and Realtime Systems,  Embedded Linux
Phone: (+49)-8142-4596-87  Fax: (+49)-8142-4596-88  Email: wd@denx.de
Physician: One upon whom we set our hopes when ill and our dogs  when
well.                                                - Ambrose Bierce

** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/