Reason for GPS problems found!

[Flemming Richter Mikkelsen]

On 2008-07-18, Timo Jyrinki <timo.jyrinki at gmail.com> wrote:
> 2008/7/18 Timo Jyrinki <timo.jyrinki at gmail.com>:
> > It seems it's not as good as without SD card, but definitely a visible
> > improvement.
>
> Sorry for spamming, I just try to describe what I'm experiencing :)
>
> Does anyone else have a reliability problem now with the new kernel
> and GPS when SD card is used? I went properly outside but didn't seem
> to get a fix, maybe 1-2 satellites max and no fixes. I drove around a
> bit and nothing. I came back, then rebooted Neo and tried again -
> almost instant fix in ca. 30-45 seconds with 6-8 satellites in the
> same spot I got nothing just before reboot.
>
> So it might be worth checking if the SD card clock is properly
> adjusted in all needed cases, or it if might be left unchanged in some
> situations. When I didn't have fixes, it behaved basically as bad as
> before the new kernel. But when the new kernel worked as it should, it
> was great.
If you want to test if the SD card is used, you can patch your kernel
to turn on a LED when the SD clock is active and switch off the LED
when it has been inactive for more than 0.3 sec. It is a good way
to test.

If you do not like the LED idea, you can use
printk(KERN_WARNING, "SD_CLK ON");
whenever the SD clock is activated, and
printk(KERN_WARNING, "SD_CLK OFF");
when it is disabled. Maybe also add a timestamp.

This way, you will know if the SD card was in use when you tried
to get your fix.