**** BEGIN LOGGING AT Wed Nov 17 02:59:57 2010 Nov 17 07:23:38 morning Nov 17 08:22:33 daniele_athome: morning Nov 17 08:23:11 hi mrmoku Nov 17 08:34:25 daniele_athome: quest'anno una festa in piu a dicembre? :P Nov 17 08:34:47 o si salvera la testa ancora un'altra volta? Nov 17 08:34:51 mrmoku, che intendi? Nov 17 08:35:10 testa d'asfalto Nov 17 08:35:18 mmm Nov 17 08:35:28 scusa mrmoku non ho capito :( Nov 17 08:35:34 politica.... Nov 17 08:35:40 aaaaaah Nov 17 08:35:47 i got it now :D Nov 17 08:35:49 :) Nov 17 08:35:59 it's time for berlusca to go, at last i'd say Nov 17 08:36:01 :) Nov 17 08:36:16 they say the current government will last until march 2011... Nov 17 08:36:28 i want to vote again :) Nov 17 08:36:46 hehe Nov 17 08:37:05 * mrmoku will send his wife voting ;) Nov 17 08:38:18 lol Nov 17 08:38:22 :) Nov 17 08:38:39 i'm exploring the TeX format world Nov 17 08:38:45 it's really amazing Nov 17 08:38:51 i'm writing a resume with it lol Nov 17 08:39:02 in fact it's my resume... Nov 17 08:40:15 * mrmoku goes down into the basement, searches his old TeX books and sends them to daniele_athome to get a very impressive resume :) Nov 17 08:40:32 :) Nov 17 08:40:49 ehm i'm running out of memory Nov 17 08:40:50 sec. Nov 17 12:24:06 JaMa: the OE december release would be a nice occasion for shr-stable :) Nov 17 12:30:58 but it will be just tag, so I think it would be better to branch shr testing branch from it instead Nov 17 12:31:07 or do you feel current shr-u so great for shr-s? :) Nov 17 12:40:44 heh, still many things todo :P Nov 17 12:44:57 probably to many for shr-stable :/ Nov 17 12:46:14 yes.. but IMHO good enough for shr-t :) Nov 17 12:49:58 yup Nov 17 13:10:35 I log out.. pre2 as hospot deos not stop to deconnect reconnect.. cu Nov 17 13:13:53 mrmoku: i have a local shr/testing2011.1 branche Nov 17 13:14:31 but, there is no use without fixed bugs in fso Nov 17 13:17:19 Heinervdm: which are the showstopper bugs in fso? Nov 17 13:17:48 i think not working shutdown Nov 17 13:17:59 and i think this causes also some other bugs Nov 17 13:19:42 the not working shutdown we could work around... but you're probably right that it's just one symptom :/ Nov 17 13:19:46 and i couldn't test gps after suspend Nov 17 13:20:01 and also gprs Nov 17 13:21:33 And i think for shr-testing we should use a no-drm kernel, as it seems that drm causes the sd-card bug Nov 17 13:23:17 gps is fixed by newer u-boot afair Nov 17 13:25:12 And this is a bit annoying too: http://trac.shr-project.org/trac/ticket/1241 Nov 17 13:25:13 Heinervdm: I just tried gprs after resume... works for me with current shr-u Nov 17 13:25:24 mrmoku: ok Nov 17 13:25:58 mrmoku: my opinion is, that we should start a branch now, but release the image next month Nov 17 13:26:09 yeah, #1241 should be fixable though Nov 17 13:26:18 and for now we make beta or rc images Nov 17 13:26:50 those can be tested by the users Nov 17 13:26:50 if we branch... we should branch based on the oe release tag Nov 17 13:27:41 ahh... wsod again :) Nov 17 13:27:53 is another thing Nov 17 13:28:02 btw is Tim and other people who wrote about shr-t maintainership in contact with you Heinervdm ? Nov 17 13:28:14 JaMa: currently not Nov 17 13:28:24 I should write a mail to them Nov 17 13:29:34 just asking because someone wrote that they will just ignore "my" build form end of may, because they are more used to build from start of may (still with ogsmd) Nov 17 13:30:48 so something so bleeding-edge could be problem for them, but I belive more in your judgement :) Nov 17 13:31:43 JaMa: because of that i think we should make rc or beta images Nov 17 13:32:09 JaMa: so they can test the images and can't complain, because this aren't the final images ;) Nov 17 13:32:30 well those end of may images were for testing for quite long :) Nov 17 13:32:45 and iirc only you and Rui tested them :) Nov 17 13:32:54 but they were hidden Nov 17 14:35:04 JaMa: somewhat working atm on gta01 patches, i'll probably have something to show in the evening. Am i the only one here able to test? Nov 17 14:45:08 PaulFertser: I've a gta01, I can spend ~ 1/2 hour day on the test Nov 17 14:46:46 PaulFertser: I don't have gta01 :/, so I can only build newer gta01 images with it Nov 17 14:54:42 does e17 segfault for someone quite often (shows dialog Recover/Exit) then restarts and segfaults again after a while? Nov 17 15:00:23 PaulFertser: i am able to test too Nov 17 15:00:54 PaulFertser: but i am at work now and not sure if i will have time later... Nov 17 15:08:14 jluis: radekp: ok, TIA :) Nov 17 15:08:29 In fact i have gta01 myself but i'm unsure about the battery condition etc. Nov 17 15:08:48 JaMa: btw, it looks like 2.6.31 version for gta01 lacked at least battery driver. Nov 17 15:09:23 PaulFertser: not sure if someone even tried to boot .31 on gta01 :/ Nov 17 15:09:31 JaMa: I've seen e17 segfaults with randr on weekend, but updated yesterday and was no longer able to reproduce it Nov 17 15:10:56 I have it only in my images, on 2nd partition with image from shr buildhost I haven't seen it.. so wondering where is the difference Nov 17 15:11:18 btw. I've seen WS(maybe not OD) twice since yesterday Nov 17 15:12:05 there was a kernel update recently, right? Nov 17 15:13:12 with a ~ week old image I was running off SD card w/o issues, but after upgrading yesterday I've again seem the issue I was seeing with 2.6.32(I believe) - system running off SD would eventually sloooow down and grind to a halt Nov 17 15:13:30 I never managed to diagnose it, though. But had a feeling it has to do with SD access Nov 17 15:14:02 and it seems it's back (at least on my phone with my SD card) Nov 17 15:14:03 zub: but nothing dangerous in kernel diff Nov 17 15:14:27 ah, ok Nov 17 15:14:45 zub: xf86-video-glamo change is pushing drm cmdq more often so it could be related Nov 17 15:14:50 I'll see if I see it again, if so I'll make some more noise here :) ... and will try to find out more Nov 17 17:35:20 mrmoku: it's definitely too early for an shr-stable, considering that latest testing has a good number of regressions compared to unstable from a year back... Nov 17 17:36:09 guess I should check whether some of them at least are fixed in current unstable... though from I hear there seem to be new ones too Nov 17 18:28:58 PaulFertser: I've some time now to play with the gta01 Nov 17 18:29:30 jluis|home: the branches are not ready yet, sorry :( Nov 17 18:33:09 no problem. I ve just found that the usb isn't recongized on theo host so I will try to load it with a known working image Nov 17 18:51:56 Do you guys also get warnings about functions from plat-samsung and plat-24xx being __init and EXPORTed at the same time? JaMa|Off ? Nov 17 19:30:15 PaulFertser: my log.do_compile doesn't show any Nov 17 19:30:30 hi mickeyl Nov 17 19:30:41 I'm working on n900 modem activator Nov 17 19:30:46 PaulFertser: greping for EXPORT, __init, plat Nov 17 19:30:59 mickeyl, you're working on a g_dbus branch right? Nov 17 19:31:32 JaMa: http://paste.debian.net/100068/ Nov 17 19:36:38 PaulFertser: only warning in my log is WARNING: modpost: Found 2 section mismatch(es). To see full details build your kernel with: 'make CONFIG_DEBUG_SECTION_MISMATCH=y' Nov 17 19:41:29 JaMa: oh yeah, those are most probably i'm talking about. Nov 17 19:43:03 already rebuilding with it, mmt Nov 17 19:48:56 PaulFertser: yes __ksymtab_s3c24xx_ts_set_platdata __ksymtab_s3c_nand_set_platdata Nov 17 20:09:42 JaMa: any idea how defconfigs are prepared nowadays? Nov 17 20:11:03 PaulFertser: in OE? I'm using menuconfig and then copy saved .config as defconfig to OE Nov 17 20:12:11 PaulFertser: because ie linux.inc is changing few items in config (like logo, cmdline, log verbosity) and I wan't it as similar as possible in defconfig file and resulting .config (for easier compare in new version) Nov 17 20:12:56 hi mickeyl I made an fsogsmd lowlevel plugin as sre requested, should I commit it? oe_runmake fails with a libgee issue Nov 17 20:13:00 JaMa: the config larsc committed doesn't look like a config generated with "oldconfig". Nov 17 20:15:42 PaulFertser: if you're talking about those in om branches they were just copies from .32 first Nov 17 20:15:48 iirc Nov 17 20:18:04 JaMa: that doesn't explain how they were created :) Nov 17 20:19:36 PaulFertser: are you talking about ie this one? http://git.openmoko.org/?p=kernel.git;a=commit;h=8b604bd0f2f3e097bb2fc8b558a0f4e6ca7bdb13 Nov 17 20:20:26 mickeyl, also ofono is GPL, fsogsmd is LGPL Nov 17 20:20:29 is it an issue? Nov 17 20:20:40 as the LGPL uses GPL source code Nov 17 20:21:02 what should I do with that? make an --enable-n900? Nov 17 20:21:06 that's imho just cherry-pick from 2.6.32 branch and those in d63ade17a0274b3b3f662fa381c4bb496bbf2ca9 I've created with OE (where oldconfig is used) Nov 17 20:21:19 JaMa: http://git.openmoko.org/?p=kernel.git;a=commit;h=5a4f431217881c3066e563098323b67fbd290d54 Nov 17 20:22:27 ah this old one.. no idea Nov 17 20:24:17 but even gta02 defconfig in 2.6.31 looks "normal" http://git.openmoko.org/?p=kernel.git;a=commit;h=1431d829b041070707a1a18516df88b3f9f9d4ce Nov 17 20:24:53 JaMa: ok, nvm, probably we should start committing "optimized" defconfigs now with Uwe's script. Nov 17 20:25:07 why not start with gta02 deconfig in 2.6.34 and after oldconfig just update those items where gta01 is different? Nov 17 20:25:48 PaulFertser: I expected that in 2.6.36 branches if they will be in git.om.org someday Nov 17 20:27:56 JaMa: hm, probably but that wouldn't be a minimal config. Nov 17 20:29:08 ? Nov 17 20:29:46 I meant that with those minimal defconfigs now in linus tree we could reuse some parts and have our configs even smaller Nov 17 20:30:07 but I haven't looked how it's done in linus tree now exactly Nov 17 20:30:28 only remember some discusion about hierarchical defconfigs on lkml Nov 17 20:32:42 JaMa: ah, ok, i am most probably not aware of the details. Nov 17 20:38:19 PaulFertser: looking on git checkout it seems that hierarchical option wasn't implemented in the end Nov 17 20:50:46 JaMa: should i push something that compiles for me even if i haven't tested it? Nov 17 20:52:09 depends on your confidence in it :), but from SHR point of view anything which boots will be better than current state Nov 17 20:53:01 but if you push it, then it will be easier for someone else with gta01 to test it :) Nov 17 20:53:34 * JaMa updating shr-kms build dir to test it there Nov 17 21:11:30 JaMa: ok, pushed, haven't tried to boot yet Nov 17 21:11:41 JaMa: i do not even have serial console atm, so... Nov 17 21:13:02 PaulFertser: wow great work (looking on patches..) Nov 17 21:13:26 JaMa: rather routine work and not even tested :| Nov 17 21:15:26 Feeling kinda tired so i can't say i will test it until next evening, and probably i'll do not have the serial console here with me anyway (i suppose it's at my student's, so i get it only on friday the earliest). Nov 17 21:17:23 PaulFertser: one more thing, can you merge om-gta01-2.6.34 and om-gta02-2.6.34 to new om-2.6.34 as it was in 2.6.31? Nov 17 21:17:51 JaMa: i can try but i'm not sure if the resulting image would be bootable even on gta02 :) Nov 17 21:17:53 PaulFertser: that way I would get single sources for both builds Nov 17 21:18:23 I can help with testing on gta02 :) Nov 17 21:20:41 PaulFertser: and wrt great work I was looking on b43e8c5823633b8fe0c588b4ceb79fca730b6c4d and even with colored diff it was hard to notice where the real change is, no idea how you could see it in that source :) Nov 17 21:21:18 JaMa: in fact that was trivial as all of those produced compile errors :D Nov 17 21:22:42 hehe, lucky :) Nov 17 21:47:30 JaMa: pushed, compile-tested for gta01 Nov 17 21:47:40 thanks Nov 17 22:07:21 PaulFertser: any idea about ubi on gta01? Nov 17 22:07:45 PaulFertser: we're building ubi images for it so I should probably enable it in OE defconfig Nov 17 22:08:25 JaMa: hm, i do not remember if you're using appropriate parameters for the ubi images on gta01. Nov 17 22:08:48 JaMa: but those must be different from gta02, that's for sure. Nov 17 22:09:10 PaulFertser: I hope so.. last time I checked with some gta01 user with 2.6.29.. Nov 17 22:09:48 JaMa: also i do not exactly understand the result of the discussions about hardware ECC, and so i've no idea if/how it should be enabled/disabled. I seem to remember it's not recommended to enable it on 2410 but i'm not sure. Nov 17 22:10:18 first is gta01 Nov 17 22:10:19 UBINIZE_ARGS = "-m 512 -p 16KiB -s 512 -O 512" Nov 17 22:10:19 UBINIZE_ARGS = "-m 2048 -p 128KiB -s 2048 -O 2048" Nov 17 22:10:26 MKUBIFS_ARGS = "-m 512 -e 15360 -c 2376" Nov 17 22:10:26 MKUBIFS_ARGS = "-m 2048 -e 126976 -c 2047" Nov 17 22:11:05 Flashing will show Nov 17 22:11:24 :) Nov 17 22:11:30 one more reason to enable it in kernel :) Nov 17 22:14:42 ehlo Nov 17 22:14:58 anyone know of a fake gps device that you could just set a position and have it run? Nov 17 22:18:26 right, got it Nov 17 22:19:11 should I enable CONFIG_FPE_NWFPE=y in gta02? Nov 17 22:20:41 JaMa: i'd say no, all fpe emulation should be done in userspace. Nov 17 22:23:08 See you tomorrow, good night :) Nov 17 22:28:21 gnite Nov 17 22:31:09 also don't know if OPENMOKO_RESUME_REASON should be gta02-only, but needs small update in mach-gta01.c Nov 17 22:31:17 PaulFertser: ^ Nov 17 22:37:13 PaulFertser: also TOUCHSCREEN_S3C2410 is not used on gta01? Nov 18 00:32:54 PaulFertser: fyi here is my SHR defconfig for gta01 http://gitorious.org/~jama/htc-msm-2-6-32/openmoko-kernel/commit/b8595e58a649181af948085aeee50d0f1422f3ed **** ENDING LOGGING AT Thu Nov 18 02:59:58 2010