ONKEY interrupt is trapped in PMU occasionally on resuming

[Andy Green]

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

Somebody in the thread at some point said:

> Andy Green wrote:
>> 	set_irq_type(irq, IRQT_LOW);
> 
> Level-triggered sounds good to me.

Yep but if it makes any difference to the behaviour, we should maybe see
if we can discard the workqueue concept with its disable-based
serialization and go back to edge triggered that is never disabled.

>> in drivers/i2c/chips/pcf50633.c
>>
>> Still it doesn't explain the 0xff INT2 register.
> 
> A bit weird that one, particularly if it's not "just" a communication

I believe the PMU powers up with funny bits in there, I saw it earlier
myself

[    3.230000] pcf50633_work: INT1=0x80 INT2=0xff INT3=0x00 INT4=0xc6
INT5=0x00

We should probably clean out the interrupt flags ourselves on driver init.

I also saw one of these exciting messages early in boot just before death.

[    3.290000] LOWBAT SIGPWR(init) REGULATOR_FAIL

If those bits fire in the PMU it means the voltage went below 90% of
what was asked for... whether its because the rail was being powered
down already or its a clue I don't know.

> issue. What other problems did you hit ?

Right now without a battery, the A5 is pretty spastic, shutting down and
rebooting like a good 'un, in U-Boot, partway through the boot, all
sorts.  It's a good platform for testing what that problem is for sure!

With the battery in, it does not seem to die in U-Boot even when left
for 30 minutes, but it shuts itself down after some longer period, maybe
during Linux boot or a minute after hitting the bash prompt.  Other
times it stays up much longer, I don't know what the variable is.

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

iD8DBQFHp12cOjLpvpq7dMoRAhWvAJ9Gs06si8indtoDWQi4qntg4I6bvwCeN3bD
iaITRFKBy2rE+2xmBPmmJls=
=Wm8b
-----END PGP SIGNATURE-----