**** BEGIN LOGGING AT Thu Nov 06 03:00:00 2014
Nov 06 10:29:58 <lumag> ant_work, hello
Nov 06 10:30:02 <lumag> bluelightning, hello
Nov 06 10:30:54 <ant_work> hi lumag
Nov 06 10:32:50 <lumag> ant_work: Regarding your question on spi-pxa2xx.
Nov 06 10:33:17 <lumag> It might be worth adding debug before all clk_prepare_enable()/clk_disable_unprepare() calls.
Nov 06 10:33:32 <lumag> However I see no issues with that code.
Nov 06 10:33:58 <lumag> Except maybe an spi-pxa2xx being runtime-suspended and only then suspended.
Nov 06 10:37:26 <lumag> ant_work, I have sent you the link to asana project that I'm trying to use to track the patches/issues.
Nov 06 10:38:20 <ant_work> yes, I'll check it out, thx
Nov 06 10:38:55 <ant_work> hmm does poodle need CONFIG_PM_RUNTIME?
Nov 06 10:39:17 <bluelightning> hi lumag, ant_work
Nov 06 10:39:59 <lumag> ant_work, It should be benefitable to all battery-powered devices
Nov 06 10:40:05 <lumag> If it works.
Nov 06 10:42:23 <ant_work> fwiw I've tried a first debug adding some checks like if (!clk_prepare_enable(ssp-clk)) and it triggers once
Nov 06 10:43:28 <ant_work> I even tried something naive: printk of the return value of linux/clk.h clk_prepare_enable
Nov 06 10:43:32 <ant_work> ..always 0
Nov 06 10:44:32 <lumag> The return value should be 0 (succeeds).
Nov 06 10:44:40 <lumag> Somebody is disabling the clock later.
Nov 06 10:48:59 <lumag> Yes, it is runtime pm
Nov 06 10:49:03 <lumag> It looks so.
Nov 06 10:49:09 <lumag> Trying to reproduce the issue
Nov 06 10:52:07 <ant_work> afair runtime pm is disabled in the defconfigs we use atm
Nov 06 10:52:14 <lumag> Hmm.
Nov 06 10:55:53 <lumag> Got a nice backtrace in cpu_xscale_do_suspend in QEmu
Nov 06 10:56:41 <lumag> ant_work, I can reproduce backtrace with PM_RUNTIME and can not w/o it.
Nov 06 10:57:12 <ant_work> heh, I was almost sure it was still disabled
Nov 06 10:58:20 <ant_work> about power, I think we can trick poodle firmware and get a working power_off
Nov 06 10:58:36 <ant_work> #define POODLE_GPIO_MAIN_BAT_LOW	(13)
Nov 06 10:58:36 <ant_work> #define POODLE_GPIO_BAT_COVER	(13)
Nov 06 10:59:16 <ant_work> ^ this gpio must be toggled (manually...) on power_off, a bit like corgi does with green led
Nov 06 11:00:44 <ant_work> I guess what happens is it sees a low battery and stays 'suspended' (unsure pxa does like sa1100 but seems similar)
Nov 06 11:03:03 <lumag> ant_work, hopefully I have a fix for spi issue.
Nov 06 11:05:42 <ant_work> that was fast...
Nov 06 11:08:47 <lumag> In your inbox
Nov 06 11:20:23 <ant_work> not yet
**** ENDING LOGGING AT Fri Nov 07 03:00:00 2014