[U-Boot] [PATCH] TQM8260: environment in flash instead EEPROM, baudrate 115k

[Wolfgang Denk]

Dear Heiko Schocher,

In message <48FC1FFC.2000104 at denx.de> you wrote:
> 
> Is this somewhere reproducible?

Kind of. It happens pretty r4eliably on weak power supplies (too high
impadance, too slow voltage rise times) where the TQM8260 whould go
through a series of quick startup - reset sequences when powered on.
On such systems, a power-on often consisted of 3...5 start attempts
that were cut off by the voltage monitor on the module. This often
lead to EEPROM content corruption. 

Actually this configuration was the reason  the  I2C  edge  condition
work  around  was  implemented,  a long, long time ago. But then,, it
seems it doesn;t work absolutely reliable in all situations.

> Maybe we can try my patch from:
> 
> http://lists.denx.de/pipermail/u-boot/2008-October/041817.html

I thought about this, too, but unfortunately no such test  system  is
available  to  me (nor to the customer any more; he fixed the obvious
problem with the PSU; yet he still sees occasional corruptions).

Aslo boot time was a strong reason to switch to a flash based
environment (and to a higher console baudrate).

> also for this board. The I2C deblocking is in i2c_make_abort () ...
> If this works better, maybe we use this I2C deblocking mechanism as
> new standard?

I'm afraid I have no way to test this...

> > --- a/board/tqc/tqm8260/flash.c
> > +++ b/board/tqc/tqm8260/flash.c
> 
> Couldnt use this board also the CFI driver?

Yes, it could, if the CFI driver was working on this board. It is  my
intention to change this as soon as the CFI driver is working (see my
other thread: "CFI Driver Problem: flash not ready").

Best regards,

Wolfgang Denk

-- 
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de
                  Nail here --X-- for new monitor.