GPS problems, summary

[Flemming Richter Mikkelsen]

On 2008-07-16, Joerg Reisenweber <joerg at openmoko.org> wrote:
> Am Mi  16. Juli 2008 schrieb Yorick Moko:
> > no fault or wrong conclusion but I just want to add that Andy also
> > made a patch to crank up the voltage a bit for the GPS unit
> >
> Yep, but this one didn't help at all it seems.
>
> One flaw in OP's summary: the recent andy-sd_clock-patch doesn't block sd-card
> for benefit of GPS FF. It simply switches off clock of sdcard as long as we
> don't read from it (or write). So you still can spoil TTFF by *constantly*
> reading from SDcard, but GPS seems to be fine as long as you don't do this.
> Stopping sd-access during TTFF is a userland issue. Kernel mustn't block
> sd-reads for extended periods, we might see all kinds of deadlocks if we did
> it this way. For sure answering an inbound call is higher priority than
> shortest possible time for this particular FF, no? So maybe you have your
> ringtones on sd. U see?
>
> rootfs on sdcard might be an issue, we (community?) should give this one a
> test
I think the Freerunner has enough RAM to actually make that possible
(ringtones and etc. may be stored in tmpfs) :)