Reason for GPS problems found!

[Timo Jyrinki]

2008/7/20 Andy Green <andy at openmoko.com>:
> Thanks for this report, I went and looked and found I didn't take care
> of the case where last packet transferred was not a bulk read packet:
> that what happens on resume.  I added a patch to stable branch that
> should be out tomorrow hopefully and give the same SD_CLK performance
> with or without resumes.

This is now starting to look great! I seem to be unable to get the
device into mode where it wouldn't be getting any fixes - suspend,
idle, power off, agps ui, switching GPS on and off, after everything
it's still ready for GPS action.

So, huge thanks for finding the glitch so that the GPS is now finally
really usable even without hardware fix!

It would be nice to know whether the S/N ratio is about as good with
the software fix as it's without SD card - it might not be, it might
be. If I get into situation that I can't get signal or it takes very
long to get a fix, it's easy to wonder whether the situation could be
even better...

Anyway, next step for me personally would be more to find out what's
wrong with GPRS functionality via which I could get the agps data to
speed up GPS even more.

-Timo