**** BEGIN LOGGING AT Tue Dec 03 02:59:58 2013
Dec 03 04:10:40 <fling> What is EAGLE?
Dec 03 04:37:36 <DocScrutinizer05> EDA
Dec 03 04:37:44 <DocScrutinizer05> cadsoft eagle
Dec 03 04:39:07 <DocScrutinizer05> though I guess wikipedia disambiguation has a few other meanings as well
Dec 03 04:45:55 <fling> ok
Dec 03 08:55:56 <PaulFertser> Heh, enlightment team reinvents GObject?
Dec 03 09:39:04 <pabs3> PaulFertser: ?
Dec 03 09:41:15 <PaulFertser> pabs3: "Inside of EFL a new Object model is being created, called Eo. This starts to unify the ad-hoc object systems EFL has had and begins a journey on the way to EFL 2.0, providing interfaces, multiple-inheritance, unified callback prototypes, introspection and weak and strong references among other things..." https://phab.enlightenment.org/phame/live/3/post/efl_1_8/
Dec 03 09:57:24 <pabs3> interesting
Dec 03 10:31:10 <jake42> DocScrutinizer05: sorry, shouldn't project from me to others "sollte nicht von mir auf andere schließen" :)
Dec 03 12:22:23 <rah> http://lists.nongnu.org/archive/html/fsuk-manchester/2013-12/msg00005.html
Dec 03 12:22:29 * rah gets himself added to terrorist watch lists
Dec 03 13:39:22 <radekp> PaulFertser: hi, it seems that 2.6.39 now works very well - maybe it's time for me to start using Freerunner again, so thanks a lot :)
Dec 03 13:39:48 <radekp> PaulFertser: nice surprise is also the power in suspend - i get 7mA which is really nice
Dec 03 13:40:10 <PaulFertser> radekp: I didn't think 2.6.39 was that good :) It's now quite stable for me too.
Dec 03 13:40:40 <radekp> PaulFertser: i charged it on Sunday evening, now the battery still shows charge_now 750mA
Dec 03 13:41:13 <radekp> that looks very usable...
Dec 03 13:42:30 <PaulFertser> radekp: nice to hear, I thought I was the only one left who wanted 2.6.39 to work (I was actually using it for half a year suffering from resume issues).
Dec 03 13:43:28 <radekp> PaulFertser: hmm my plan was to switch to gta04, but it's just useless toy until the power management is fixed
Dec 03 13:44:15 <PaulFertser> radekp: I thought the kernel devs (Neil?) are actively working on it.
Dec 03 13:45:19 <radekp> PaulFertser: well it's been 2 years - Neil did all the work and shifted it from completely unusable (battery life ~4hours) to barely usable (~1day)
Dec 03 13:47:13 <radekp> PaulFertser: but i doubt that it can be solved just by one kernel dev no matter how smart he is... my guess is that some HW measuring must be done - proably by Nikolaus
Dec 03 13:47:41 <PaulFertser> radekp: right but I thought the schematics incorporated enough R0's and Neil has all the stuff needed to measure individual components?
Dec 03 13:47:59 <radekp> PaulFertser: but he does not seem to understand the problem - instead goldelico is producing useless android rootfs and other strange projects...
Dec 03 13:49:34 <radekp> PaulFertser: i doubt that normal men can measure the inidividual components - there are no pads for it, we only measured current using bq battery
Dec 03 13:50:28 <radekp> PaulFertser: another funny thing with GTA04 current is that it's completely random - sometimes it's 21mA which is quite ok (~days standby)
Dec 03 13:50:58 <radekp> PaulFertser: but somtimes it's 80 mA and you realize after 10 hours that the device is discharged
Dec 03 13:54:57 <PaulFertser> radekp: do you mean it's not possible to desolder 0R resistors and measure with a DMM?
Dec 03 13:55:50 <radekp> PaulFertser: i am no aware that anyone did it...
Dec 03 13:56:22 <radekp> PaulFertser: i havent seen the schematics and i'd not understand them anyways :)
Dec 03 13:57:44 <PaulFertser> radekp: I was going to do that for gta01 but when I realised the main shield is soldered, I gave up.
Dec 03 13:59:42 <radekp> hmm qtmoko is now running quite nice on 2.6.39, time for new release :)
Dec 03 14:09:34 <JaMa> PaulFertser: both patches are also in latest shr-staging, thanks again
Dec 03 15:15:55 <PaulFertser> JaMa: thank you :)
Dec 03 15:22:30 <PaulFertser> radekp: thank you too for your hard work on qtmoko, it's amazing you're doing most of that all alone.
Dec 03 17:39:10 <DocScrutinizer05> radekp: go for maemo, or at very least for non-suspend but zero clock paradigm
Dec 03 17:39:38 <DocScrutinizer05> radekp: I'm pretty sure suspend causing more issues than it solves, on OMAP3
Dec 03 17:40:51 <DocScrutinizer05> in the end when everything correctly managed, a completely clocked-down OMAP3 board should have insignificantly higher power usage than when suspended to RAM
Dec 03 17:41:22 <DocScrutinizer05> but you have none of the resume-headaches
Dec 03 17:41:44 <DocScrutinizer05> aaaand: yiou can *debug* da shite
Dec 03 17:42:20 <DocScrutinizer05> which is obviously pretty difficult when CPU goes completely zombie
Dec 03 17:43:03 <DocScrutinizer05> plus: maemo5 did it (actually maemo4 did it, already, though for omap2)
Dec 03 17:43:55 <DocScrutinizer05> the approach to treat embedded like a tiny desktop is a delusion
Dec 03 17:44:34 <DocScrutinizer05> big mistake
Dec 03 17:45:23 <DocScrutinizer05> you may want to make it LOOK and FEEL like a tiny desktop, but the core has to be designed according to quite different paradigms
Dec 03 17:46:19 <DocScrutinizer05> you wouldn't hope to send your desktop to suspend-to-ram twice per minute, would you?
Dec 03 17:46:46 <DocScrutinizer05> particularly when it doesn't yield *any* benefit
Dec 03 17:52:16 <fling> >> insignificantly higher << how higher?
Dec 03 17:53:18 <PaulFertser> radekp: but really, what's the reason for gta04 to consume any significant amount of power when it's not suspended? Afaict the kernel should put CPU in the low power state automatically, fully transparent for all the apps.
Dec 03 17:53:28 <PaulFertser> DocScrutinizer05: I think android device are all doing the same now.
Dec 03 17:53:59 <DocScrutinizer05> great argument why it MUST NOT get done this way ;-)
Dec 03 17:54:17 <fling> PaulFertser: network will continue working so I like the idea.
Dec 03 17:54:23 <DocScrutinizer05> fling: like 5 instead of 3 mA
Dec 03 17:54:49 <PaulFertser> DocScrutinizer05: the nice thing about mobile SoCs is that they have very low power state that's the same as the other freq states, requiring very little extra work from the kernel.
Dec 03 17:54:57 <PaulFertser> DocScrutinizer05: i meant android does the same as maemo.
Dec 03 17:55:10 <DocScrutinizer05> oooh
Dec 03 17:55:12 <PaulFertser> fling: we all like, I wonder why gta04 is not able to work that way.
Dec 03 17:55:39 <dos1> did anyone try it at all?
Dec 03 17:55:57 <PaulFertser> DocScrutinizer05: take a look at any recent android device. It doesn't look like it really suspends/resume ever, it's just transparent, the feel is the same as with maemo, and I think the underlying mechanism too.
Dec 03 17:56:00 <DocScrutinizer05> fling: and that's probably even incl modem
Dec 03 17:56:05 <fling> qtmoko needs more developers :>
Dec 03 17:56:07 <PaulFertser> dos1: tried what? not suspending gta04?
Dec 03 17:56:31 <dos1> PaulFertser: tried looking at zero-clocking with gta04
Dec 03 17:57:30 <dos1> maybe it just needs to be enabled somehow? I haven't seen anyone else than DocScrutinizer05 to talk about it in context of gta04
Dec 03 17:57:38 <DocScrutinizer05> aiui there are fugures of power consumption in idle which are waaay too high but nobody cared since... "we will suspend anyway"
Dec 03 17:58:03 <PaulFertser> dos1: it should just work, nothing needs to be enabled, the kernel manages that state as any other cpu frequency/power state afaict.
Dec 03 17:58:16 <DocScrutinizer05> yes
Dec 03 17:58:22 <DocScrutinizer05> PM-kernel
Dec 03 17:58:30 <DocScrutinizer05> needs a few patches for OMAP
Dec 03 17:58:42 <DocScrutinizer05> but they should all be upstream by now
Dec 03 17:58:49 <PaulFertser> So the kernel as used by qtmoko for gta04 just lacks some proper kernel patches?
Dec 03 17:58:59 <DocScrutinizer05> *shrug*
Dec 03 17:59:18 <DocScrutinizer05> you tell me ;-)
Dec 03 17:59:34 <PaulFertser> I agree suspending makes little sense if the SoC consumes about the same in suspend as in the P0/C0 (or how they call it).
**** ENDING LOGGING AT Wed Dec 04 02:59:58 2013