**** BEGIN LOGGING AT Sun Sep 09 02:59:58 2012 Sep 09 03:35:05 how does one rebuild an ipk from a "LOCAL" ? (ie there is no .bb , only local-builds tweak) Sep 09 03:35:42 I rm the stamps, the work subdir , -c clean -f , tried 20th variations no go Sep 09 03:36:02 it seems it build a cache somewhere Sep 09 05:05:14 ok devshell and "make DESTDIR= install" is more efficient than fighting bitbake :) Sep 09 05:07:08 ugh shr uses renice everywhere . what the point ? it reminds me there was a thread about that for desktop environments like 7 years ago Sep 09 05:08:16 I just remember the conclusion was that it hurts more than it helps (except when applied on very specific programs ... here it is all that is running under Xorg + all shr daemons and notifiers Sep 09 05:09:29 ie it is in short fighting the scheduler by hand .... I commented out most of the renice I found locally ... the ui is more responsive on gta02 Sep 09 05:10:13 ie I could hardly scroll the slider to seek in a song in enjoy beforehand Sep 09 12:00:14 about my 3G card not working anymore . I revert commit 1d41feb of cornucopia (was still not working) then disabled sim protection on my other phone and put the sim back , this time pushing it the furthest in the holder I could ... one of those three items made it works again (or them all) Sep 09 12:01:28 ie the last two items I did at the same time (I was getting an error 10 , ie no sim card ) Sep 09 12:06:18 also I still do not know how to force the rebuild of an ipk from a change in a "LOCAL" source (ie via local-builds.inc) Sep 09 12:07:48 clean+build is getting the old build data ( -c compile properly rebuild the binaries in the source tree but still get the old binaries when I later force the rebuild of the ipk ... via rm the stamps, work subdirs , etc Sep 09 12:29:32 I resorted to -c devshell and install to destdir Sep 09 13:04:16 prahal2, for 3g complain to morphis Sep 09 13:04:20 he tried to fix but failed Sep 09 13:11:02 hi JaMa Sep 09 13:11:15 I need to push prahal2's patch Sep 09 13:11:20 so in order to do that: Sep 09 13:11:28 *) I've to add a patch header I guess Sep 09 13:11:37 2) should I bump PR? Sep 09 13:11:42 we're using BasicHash Sep 09 13:11:49 but maybe upstream oe is not anymore Sep 09 13:12:32 yes PR bump is always needed Sep 09 13:12:40 ok Sep 09 13:12:43 even with OEBasicHash to upgrade it on target Sep 09 13:12:51 ok Sep 09 13:13:51 I planned to apply it today, but if you want it now, go ahead (I'm cooking lunch now..) Sep 09 13:14:23 ok thanks a lot Sep 09 13:14:29 if you want to do it you can Sep 09 13:14:44 however if you prefer that I do it I'll do it Sep 09 13:14:49 I can in an hour or so Sep 09 13:23:42 ok Sep 09 13:26:35 JaMa: I have 0.12.0-rc2 tarballs ready here Sep 09 13:26:42 will push them later Sep 09 13:35:52 morphis: ok I'll upgrade recipes Sep 09 13:37:50 JaMa: will ping you when tarballs are uploaded Sep 09 13:38:21 ok Sep 09 13:53:16 GNUtoo: so you you want me to apply that patch? Sep 09 13:53:19 I have few mins Sep 09 13:53:33 JaMa, ok do it then, I'm already testing it Sep 09 13:53:51 if you're testing it, then push your version Sep 09 13:53:57 I'm running test-lms right now Sep 09 13:54:15 I'll prepare rc2 recipes instead Sep 09 13:55:25 JaMa: http://downloads.freesmartphone.org/sources/cornucopia/0.12-rc2/ Sep 09 13:55:51 wow!!!!!!! it seem to work Sep 09 13:55:58 did someone test a lot the images btw Sep 09 13:56:02 because I had another bug Sep 09 13:56:16 difficulties to go back to main screen Sep 09 13:56:26 you press home nothing happen Sep 09 13:56:59 you click on shr_elm_softkey the next button, it only switches between existings apps, not the home screen Sep 09 13:57:18 you select illume_home or elfe in the list, doesn't work Sep 09 13:57:21 but you can use home icon, cannot you? Sep 09 13:57:25 no Sep 09 13:57:30 it has no effect Sep 09 13:57:42 I've tested both elfe and illume-home Sep 09 13:57:47 none of both work Sep 09 13:58:26 altough the split button works Sep 09 13:58:35 so maybe there is a way to use it trough the split button Sep 09 13:59:37 hi Sep 09 13:59:40 morphis, ping Sep 09 14:05:57 SHR: 03Martin.Jansa 07meta-smartphone * rcaf4d13a65fb 10/meta-fso/recipes-freesmartphone/cornucopia/ (22 files): meta-fso: upgrade cornucopia to 0.12.0-rc2 Sep 09 14:06:57 paulk-desktop: pong Sep 09 14:07:18 morphis, hi Sep 09 14:07:31 are you going to make more changes to libsamsung-ipc in the next few hours? Sep 09 14:07:45 because I'm in the process of building new official images for replicant Sep 09 14:08:53 paulk-desktop: no, not today Sep 09 14:09:11 paulk-desktop: why are you using still HEAD of libsamsung-ipc and not a fixed SRCREV? Sep 09 14:09:11 morphis, ok thanks Sep 09 14:09:41 morphis, because I would be way too lazy to update the rev in the manifest everytime Sep 09 14:10:20 hm Sep 09 14:10:25 your choice :) Sep 09 14:38:34 morphis: fsogsmd: Sep 09 14:38:35 | pdphandler.c:123:30: error: field 'contextActMessage' has incomplete type Sep 09 14:38:38 | pdphandler.c:129:30: error: field '_tmp8_' has incomplete type Sep 09 14:38:41 | pdphandler.c:151:30: error: field 'contextDeactMessage' has incomplete type Sep 09 14:38:44 | pdphandler.c:153:30: error: field '_tmp4_' has incomplete type Sep 09 14:39:00 | pdphandler.c:590:2: error: too few arguments to function 'ipc_gprs_pdp_context_setup' Sep 09 14:39:08 needs newer samsung ipc? Sep 09 14:39:47 it should not Sep 09 14:39:54 tested with libsamsung-ipc 0.1.0 Sep 09 14:40:10 do not use git HEAD! Sep 09 14:49:27 JaMa, for the illume-home I've found a workarround: Sep 09 14:49:44 e-settings->illume->windows Sep 09 14:49:52 there is home Sep 09 14:49:57 I uncheck all buttons Sep 09 14:50:02 and I can go right and left Sep 09 16:26:03 I need to go, bye Sep 09 17:35:05 morphis: that fsogsmd seems to fail only on gta02.. I'll test it on shr buildhost too first Sep 09 18:47:36 I enabled preempt and statistics on gta02 ... now I get why those were not ... but I would welcome any clue about the root : PC is vprintk , lr is hdq_read (all those from bq27x00) and this oops tell FIQ_32 ie fiq context (thus printk is bad) . Sadly I did not yet enabled stacktrace Sep 09 18:48:10 though it looks close to the msleep while waiting for fiq handler to complete Sep 09 18:48:12 prahal2: what are you testing? Sep 09 18:48:32 preempt + collect stats on gta02 Sep 09 18:49:07 prahal2: but what kernel version? Sep 09 18:49:11 I saw your patch about function_tracer . Sadly this is disabled here (well I have the HAVE_FUNCTION_TRACER but not FUNCTION_TRACER per se Sep 09 18:49:15 2.6.39 Sep 09 18:49:19 -r13 Sep 09 18:49:45 prahal2: what sort of statistics are you trying to get? Sep 09 18:49:50 powertop Sep 09 18:50:09 prahal2: if you want to save energy then cpufreq support is probably the best goal? Sep 09 18:50:10 I mean for powertop to tell me the wake Sep 09 18:50:42 preempt is only because it looks more sane Sep 09 18:50:49 prahal2: yeah but why are you looking at powertop? Sep 09 18:51:51 to find out what is causing lags in enjoy playback on gta02 (was kind of helpfull on gt04 -> ie bma sensor driver was polling like mad and snd jack was not so kind too Sep 09 18:52:17 prahal2: enjoy? Sep 09 18:52:37 triggering omap xfer at every poll thus pm_wkup (enjoy the illume/e17 music player Sep 09 18:52:52 prahal2: ah, playing mp3 or vorbis? Sep 09 18:52:57 mp3 Sep 09 18:53:21 though on gta04 that was (I did not managed to get powertop run on gta02 Sep 09 18:54:05 around 5000 pm_wkup interrupt per second it started to play less well Sep 09 18:54:45 intone+mplayer was leading to complete kernel lag to the point that all I could do was force poweroff Sep 09 18:55:41 though there was a slider+seek issue with enjoy and intone (still is there for intone) that meant they were playing weird too) Sep 09 18:56:09 prahal2: I don't recall noticing any audio issues on 2.6.39 Sep 09 18:56:35 gta04 -> 3.4 Sep 09 18:57:02 the audio issue I have with enjoy on 2.6.39 is while touching the screen Sep 09 18:57:39 different though I wanted to try the same therapy -> powertop and perf if needs be Sep 09 18:57:48 SHR: 03shr-devel 07buildhistory * rd77cfae1785c 10/ (490 files in 490 dirs): packages: Build 201209091950 of shr 20120909 for machine nokia900 on opmbuild Sep 09 18:58:32 I found out I could barely seek in enjoy player with all the renice applied to various items Sep 09 19:00:26 so I removed them all (the enlightenment_start, opimd-notifier, and probably one other I forgot) ... now I can seek fine . but the whole phone is more sensitive about ssh traffic Sep 09 19:00:48 so I gave a shot at preempt instead of nopreempt Sep 09 19:03:49 by the way fiq seems so pedantic it makes it easy to see if any code is doing something wrong (not like those ghosts bug that appear and disappear randomly Sep 09 19:21:04 SHR: 03shr-devel 07buildhistory * r53d2b16f58ce 10/packages/om_gta04-oe-linux-gnueabi/ (6 files in 6 dirs): packages: Build 201209092107 of shr 20120909 for machine om-gta04 on opmbuild Sep 09 19:27:47 I now wonder if it comes from the use of msleep (well there might be another issue before this one) Sep 09 19:28:35 ie it always fails on mov r5, #5 which looks wrong ... but the line above is the one that set pc (thus that ought to be the one that set pc to vprintk Sep 09 19:30:46 sorry ... I am learning arm assembly as I go Sep 09 19:42:54 SHR: 03shr-devel 07buildhistory * r9b3f365cc2d0 10/packages/crespo-oe-linux-gnueabi/ (6 files in 6 dirs): packages: Build 201209092130 of shr 20120909 for machine crespo on opmbuild Sep 09 20:04:54 SHR: 03shr-devel 07buildhistory * rd7bb37595fdf 10/packages/tuna-oe-linux-gnueabi/ (6 files in 6 dirs): packages: Build 201209092152 of shr 20120909 for machine tuna on opmbuild Sep 09 21:21:02 SHR: 03shr-devel 07buildhistory * r84eb67d22667 10/packages/ (489 files in 489 dirs): packages: Build 201209092214 of shr 20120909 for machine om-gta02 on opmbuild Sep 09 21:55:29 the printk comes from power_supply_show_property (it is called in the middle of the bq27x00_battery probe via power_supply_register->device_add->*uevent Sep 09 22:02:18 I bet because it cannot read the property of battery driver before it is register thus print an error **** ENDING LOGGING AT Mon Sep 10 02:59:58 2012