On 08/10/2012 08:31 AM, John Rippetoe wrote:
Marc,
No problem at all, better for it to be accessible to others in case they run across similar issues.
No kidding? As far as I know the at91rm9200 is essentially the grandfather of the at91sam9g20 I am using now; similar design but two revisions older.
As of yesterday afternoon I myself have begun to suspect there could be a problem with the spi controller hardware within the mcu. I have also been working with the board manufacturer in solving this problem as well. In an attempt to troubleshoot your driver I was comparing it with a different kernel driver recommended by my board manufacturer (Micromint) that was written by another customer. This particular driver has apparently been well received, however, I have yet to have any luck getting it to work. Even when a representative from the company sent it to me precompiled and tested, I still had no lluck getting it to work in an identical environment.
Now I know the ADC works because I have userland code which directly accesses the ADC and I get correct output. Note that this code completely bypasses the atmel spi controller and operates directly on the ADC128S052s registers. Suspicious right?
Strike that, the userland example doesn't access the ADC directly. It
actually accesses the spi controller directly. I didn't have a copy of
the code on hand when I wrote that and was mistaken. Sorry about that.
I assume you are using the atmel-spi driver as the master controller driver correct? If so, did you make any modifications to it at all?
Thanks for your help guys.
John
Marc Pignat <marc.pignat@xxxxxxx> wrote:
Hi John,
(You asked me in private, but I prefer to respond on the mailing list).
This driver was working, I wrote (and tested it) in 2008 ;)
The hardware it was running on was a custom at91rm9200 based board with
a adc084s051.
The last time a checked it, it was on 2.6.28, and it was still working fine.
I just had a look to the adc128s052 datasheet, and it looks compatible
with this driver.
Guenter, the spi_write_then_read() function is really the good one.
In brief, the write selects the channel, then the read reads the data.
I think the problem you're facing is either in the spi driver or in the
hardware.
Best regards
Marc
_______________________________________________
lm-sensors mailing list
lm-sensors@xxxxxxxxxxxxxx
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors
--
Respectfully,
John Rippetoe
Communication Concepts
813-888-6611
_______________________________________________
lm-sensors mailing list
lm-sensors@xxxxxxxxxxxxxx
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors
