cpufreq: gps && rxerr: port=1 ch=0x24, rxs=0x00000001

Nils Faerber nils.faerber at kernelconcepts.de
Tue Jun 16 16:11:28 CEST 2009

Timo Juhani Lindfors schrieb:
> Sebastian Krzyszkowiak <seba.dos1 at gmail.com> writes:
>> I think this issue isn't related to cpufreq - I have that on standard
>> andy-tracking kernel. And for me it happens always - I'm unable to use
>> GPS at all...
> But for me this has never happened before. So even if cpufreq does not
> cause it it will at least make it more probable to occur.

I have no idea about the internals of the Samsung CPU but I know from
other microcontrollers that the baudrates are generated by specifying a
divisor that will be applied to the clock frequency. You will need a
calculation formula that calculates the divisor if your clock changes.

If this is the same process here then we might have a problem in that
divisor calculation. The clock can always differ a little and an UART is
a little tolerant. If the formula's error is right at the limit of the
UART's tolerance then small changes can trigger the problem, like a
clock change or a slightly wrong core clock.
On Atmel AVR e.g. there are certain clock/baudrate combinations that are
simply not possible since there is no matching divisor to devide the
clock to that baudrate.

  nils faerber

kernel concepts GbR        Tel: +49-271-771091-12
Sieghuetter Hauptweg 48    Fax: +49-271-771091-19
D-57072 Siegen             Mob: +49-176-21024535

More information about the openmoko-kernel mailing list