[PATCH] GTA02 power manager for bluetooth

[Andy Green]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Somebody in the thread at some point said:
> Andy Green wrote:
>> Update on this: overnight the second device transferred 34GB without
>> error and in fact has never shown the error that was seen on the first
>> device.
> 
> Excellent ! Seems that we have a winner !
> 
>> BTW I currently look at suspend/resume since Matt seems to not be
>> available for some time one way or another.
> 
> He told me that he's trying to keep on working on this a bit, but
> that he's only available part-time at the moment, due to other tasks.
> (And then on vacation for the second half of next week.)

Matt, I found that the failure of resume on normal boots is down to gsmd
starting.  I am looking into that now.

> One issue he found is that the device resets randomly if running from
> the USB wall charger (but with battery present) if the debug board is
> connected, where it can take up to 5 minutes until such a reset
> happens.
> 
> Sounds like a fun problem to debug ...

This sounds like part of a larger collection of poorly understood power
management bugs, including some boards (?) that auto-reset when cold (!)

Why do I sometimes get a one-a-second PMU driver spew on the console
until the power button is pressed?  Why not never or always?  Why does
the LCM backlight sometimes stay off forever until power is cycled
rather than hard reset?

I guess we're hoping they magically went away on A5, which is possible
due to developments on the PMU chip from the vendor: we will have to
mount a sustained attack on them if they didn't.

- -Andy
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.7 (GNU/Linux)
Comment: Using GnuPG with Fedora - http://enigmail.mozdev.org

iD8DBQFHkIL6OjLpvpq7dMoRAtRcAJwMW8gHoRa0wVSmCQBfC+p6+lQ7+QCfdK20
0+R1FQdO1/YPvq2b78RDNGs=
=84Y0
-----END PGP SIGNATURE-----