**** BEGIN LOGGING AT Wed Aug 01 02:59:59 2012 Aug 01 05:20:03 SHR: 03shr-devel 07buildhistory * r9f670cb14284 10/packages/ (283 files in 283 dirs): packages: Build 201208010147 of shr 20120731 for machine om-gta02 on opmbuild Aug 01 05:44:20 SHR: 03shr-devel 07buildhistory * r5c0f059a4828 10/packages/armv7a-vfp-neon-oe-linux-gnueabi/emacs/ (8 files in 8 dirs): packages: Build 201208010720 of shr 20120801 for machine nokia900 on opmbuild Aug 01 07:08:53 plotr: hey :) can you please try to fix your irc connection? Aug 01 07:09:19 I'll try Aug 01 07:09:21 sorry Aug 01 07:16:42 SHR: 03Martin.Jansa 07meta-smartphone * rc4352ce636f2 10/meta-nokia/recipes-bsp/u-boot/u-boot_git.bb: u-boot: remove 0001-config-Always-use-GNU-ld.patch from SRC_URI temporary Aug 01 07:16:52 SHR: 03Martin.Jansa 07meta-smartphone * rb9060cf17065 10/meta-shr/recipes-shr/shr/e-wm-config-illume2-shr_git.bb: Revert "e-wm-config-illume2-shr: replace illume-home with elfe" Aug 01 07:37:45 SHR: 03shr-devel 07buildhistory * r63566b8dba81 10/packages/ (16 files in 16 dirs): packages: Build 201208010919 of shr 20120801 for machine om-gta02 on opmbuild Aug 01 07:37:53 SHR: 03shr-devel 07buildhistory * r265622d4447e 10/images/om_gta02/eglibc/shr-image/ (11 files): images: Build 201208010919 of shr 20120801 for machine om-gta02 on opmbuild Aug 01 07:56:11 SHR: 03shr-devel 07buildhistory * r0089b4333c09 10/images/nokia900/eglibc/shr-image/ (11 files): images: Build 201208010939 of shr 20120801 for machine nokia900 on opmbuild Aug 01 07:56:12 SHR: 03shr-devel 07buildhistory * rd37986616b73 10/packages/armv7a-vfp-neon-oe-linux-gnueabi/e-wm-config-illume2-shr/ (9 files in 9 dirs): packages: Build 201208010939 of shr 20120801 for machine nokia900 on opmbuild Aug 01 08:11:04 SHR: 03shr-devel 07buildhistory * rccb4b70b36ac 10/images/om_gta04/eglibc/shr-image/ (11 files): images: Build 201208010957 of shr 20120801 for machine om-gta04 on opmbuild Aug 01 08:11:14 SHR: 03shr-devel 07buildhistory * rb17cc138af26 10/packages/om_gta04-oe-linux/udev-extraconf/ (7 files in 7 dirs): packages: Build 201208010957 of shr 20120801 for machine om-gta04 on opmbuild Aug 01 08:13:45 SHR: 03Martin.Jansa 07shr-themes * r1e5ab90f971e 10/e-wm/e-wm-config-illume2-shr/ (6 files in 2 dirs): e-wm-config-illume2-shr: downgrade E17 config for version which is in SHR feed Aug 01 08:27:30 SHR: 03shr-devel 07buildhistory * rabe3da033148 10/images/crespo/eglibc/shr-image/ (11 files): images: Build 201208011012 of shr 20120801 for machine crespo on opmbuild Aug 01 08:31:03 SHR: 03Martin.Jansa 07meta-smartphone * r61e105bc1227 10/meta-shr/recipes-shr/shr/e-wm-config-illume2-shr_git.bb: e-wm-config-illum2-shr: downgrade config version and add elfe again Aug 01 08:40:16 SHR: 03Martin.Jansa 07shr-chroot * r67bad2b467d4 10/ (5 files in 4 dirs): small fixes for bitbake Aug 01 08:40:26 SHR: 03Martin.Jansa 07shr-chroot * r9ef77cddc55f 10/ (OE/bin/shr_core_staging_new.sh var/run/utmp): staging scripts: prefix STAGING_NUMBERS with 1 to force test to compare them as decimal numbers (even with leading zeros) Aug 01 08:40:26 SHR: 03Martin.Jansa 07shr-chroot * r174d9a09a4df 10/ (135 files in 12 dirs): system upgrade Aug 01 08:47:30 someone with current shr-core image willing to test newer illume2 config with elfe? Aug 01 08:53:47 jake42: any idea about that PIN? maybe phoneuid not running again? Aug 01 09:14:58 SHR: 03Martin.Jansa 07meta-smartphone * r7c6013a9f8c8 10/meta-shr/recipes-shr/shr/e-wm-config-illume2-shr_git.bb: e-wm-config-illume2-shr: fix SRCREV Aug 01 09:27:01 SHR: 03Martin.Jansa 07shr-themes * r97c94207298b 10/e-wm/e-wm-config-illume2-shr/Makefile.am: e-wm-config-illume2-shr: fix WANT_MACHINE_PALMPRE Aug 01 09:39:44 SHR: 03Martin.Jansa 07meta-smartphone * r013316b838d8 10/meta-shr/recipes-shr/shr/e-wm-config-illume2-shr_git.bb: e-wm-config-illume2-shr: bump SRCREV Aug 01 09:42:04 JaMa, what watches the repo and kicks of the builds? Aug 01 09:45:36 ACTION  Aug 01 09:46:43 poor JaMa Aug 01 10:03:08 morphis: ping Aug 01 10:04:33 JaMa: just install latest staging and test? Aug 01 10:59:11 SHR: 03shr-devel 07buildhistory * r6b2fa7455906 10/packages/ (11 files in 11 dirs): packages: Build 201208011251 of shr 20120801 for machine om-gta02 on opmbuild Aug 01 11:05:45 hi JaMa Aug 01 11:06:14 I failed at changing the format of the GPS of the GTA02 to ubx with gpsd..... Aug 01 11:10:47 SHR: 03shr-devel 07buildhistory * rc29bb0063b61 10/packages/ (11 files in 11 dirs): packages: Build 201208011306 of shr 20120801 for machine nokia900 on opmbuild Aug 01 11:12:12 maybe there is a different non-ubx-standard init sequence? Aug 01 11:12:21 JaMa: It worked with sending pin via dbus Aug 01 11:12:48 I will test more later today Aug 01 11:13:03 I'm willing to test elfe image too Aug 01 11:13:17 hi everyone! fso-specs doesn't compile. http://pastie.org/4370795 anyone? Aug 01 11:13:52 GNUtoo: I got upto fso-specs, you see :) Aug 01 11:15:31 ok Aug 01 11:18:18 ildar, ask morphis or JaMa Aug 01 11:18:26 more morphis than JaMa Aug 01 11:18:55 ok. Hope they'll notice my message Aug 01 11:19:01 ildar: it's using host python and it works in shr-chroot or Ubuntu.. you can try to add "inherit pythonnative" to that recipe and rebuild to force usage of staged native python Aug 01 11:19:43 jake42: it's already in 084 staging (ipk file) so you can opkg upgrade from there Aug 01 11:20:24 JaMa: why it's using host python when it already built it's own? Aug 01 11:20:35 can't find a logic here/ Aug 01 11:21:46 because it doesn't have that inherit Aug 01 11:22:43 SHR: 03shr-devel 07buildhistory * re7a53c2b5cc6 10/packages/om_gta04-oe-linux-gnueabi/shr-version/ (latest shr-version/latest): packages: Build 201208011318 of shr 20120801 for machine om-gta04 on opmbuild Aug 01 11:23:06 before it was able to pick what every native python was available (sometimes from host sometims from sysroot) now it's strictly separated http://lists.linuxtogo.org/pipermail/openembedded-core/2012-July/026888.html Aug 01 11:23:44 ildar: and only packages which really need concrete python (the one we provide in sysroot) are being updated to inherit pythonnative Aug 01 11:24:00 IC, thanx Aug 01 11:24:04 I'll try that Aug 01 11:25:08 * JaMa back to adding PyQT Aug 01 11:27:49 JaMa: It worked, thanx a lot Aug 01 11:29:15 send patch :) Aug 01 11:30:49 ) Aug 01 11:34:33 hmmmm Aug 01 11:34:46 how does the GTA02 UBX GPS work? Aug 01 11:34:48 I tried that: Aug 01 11:35:24 SHR: 03shr-devel 07buildhistory * r7250b733a33f 10/packages/crespo-oe-linux-gnueabi/shr-version/ (latest shr-version/latest): packages: Build 201208011330 of shr 20120801 for machine crespo on opmbuild Aug 01 11:35:42 wiki.openmoko.org/wiki/Neo_FreeRunner_GPS#Configuration_for_a_higher_sampling_rate Aug 01 11:35:46 and it indeed worked Aug 01 11:35:53 but the output is still NMEA Aug 01 11:36:12 worked means that there is GPGGA GPRMC and that's all Aug 01 11:36:50 mrmoku: latest + 1 (084 is not yet closed) Aug 01 11:37:34 mrmoku: you can reflash 083 then opkg upgrade, systemctl stop xserver-nodm.service; rm -rf /root/.e; systemctl start xserver-nodm.service Aug 01 11:40:21 ka6sox: our buildhost is quite slow and disk always almost full, so automated builds like on jenkins would kill it soon :/ Aug 01 11:42:50 [BUG] libfsoframework recipe does not depend on libnl but requires it Aug 01 11:44:26 wasn't that fixed few days ago? Aug 01 11:44:59 ildar: http://git.shr-project.org/git/?p=meta-smartphone.git;a=commit;h=b1d42a9a66708883abd6b88be78f1d8227cff668 Aug 01 11:45:09 ildar: aren't you using master branch instead of shr branches? Aug 01 11:45:39 I use master but didn't update today Aug 01 11:45:45 just yesterday Aug 01 11:46:09 master is not updated as often as shr branch is Aug 01 11:46:26 hmmm Aug 01 11:46:33 I've merged shr there today, but only because all needed changes landed in oe-core/meta-oe already Aug 01 11:46:43 should I switch to shr branch then? Aug 01 11:46:49 ildar: yes.. Aug 01 11:47:00 ildar: for all layers Aug 01 11:47:01 hmm. Not obvious Aug 01 11:47:11 that's why we have shr-makefile :) Aug 01 11:47:40 eeem. Aug 01 11:47:52 is there a note somewhere? Aug 01 11:48:14 note about shr makefile or branches? Aug 01 11:48:27 any/both ? Aug 01 11:48:32 http://shr-project.org/trac/wiki/Building%20SHR Aug 01 11:48:42 ah!! Aug 01 11:48:48 both Aug 01 11:48:53 need to read it carefully, thanks Aug 01 11:49:00 ))) Aug 01 11:49:07 sorry for the noice Aug 01 12:06:49 JaMa: ok Aug 01 12:07:05 JaMa: office waiting for me first though Aug 01 12:07:36 hi lindi- Aug 01 12:07:39 GNUtoo, JaMa: what is that breakage in phonefsod? Aug 01 12:08:02 mrmoku, no idea, I'm working on specific stuff and not using the phones all day Aug 01 12:08:14 GNUtoo: ok Aug 01 12:08:19 hoping you'll fix the breakages..... Aug 01 12:08:21 so morphis is the one to talk to Aug 01 12:08:25 yeah, I will Aug 01 12:08:30 thanks a lot!!!! Aug 01 12:08:34 phonefsod is what I know :P Aug 01 12:08:39 if it's phonefsod breakage by morphis I know about it Aug 01 12:08:54 that's what I read in the backlog Aug 01 12:08:58 ok Aug 01 12:09:02 the backlog has links Aug 01 12:09:05 to fso wiki page Aug 01 12:09:06 some fsogsmd api change Aug 01 12:09:08 ahh ,ok Aug 01 12:09:12 that describe the fix Aug 01 12:10:29 mrmoku: phonefsod? Aug 01 12:10:30 DeviceStatusRework Aug 01 12:10:34 found it Aug 01 12:10:37 ah Aug 01 12:10:42 JaMa: ŝome API changes morphis is planning Aug 01 12:10:46 http://wiki.freesmartphone.org/index.php/Implementations/fsogsmd/DeviceStatusRework Aug 01 12:11:02 yup I was thinking about breakage in feeds :) Aug 01 12:11:04 what about the 2 resources bug? Aug 01 12:11:51 1) fsoraw -r $resource ls -> doesn't exit ever Aug 01 12:12:07 2) enlightenement reloading when requesting a resource Aug 01 12:12:29 for 1) I looked at fsoraw code and from there it is not understandable how that ever could happen Aug 01 12:12:33 (at least for me) Aug 01 12:12:41 it's quite simply just doing a wait Aug 01 12:13:10 and I built fsoraw on my laptop and there it did not hang Aug 01 12:13:22 so it might be some toolchain or vala thing Aug 01 12:13:48 ok Aug 01 12:13:57 did you tell that in the bugreport? Aug 01 12:14:11 did you compare the 2 C generated code? Aug 01 12:18:07 GNUtoo: hi Aug 01 12:21:23 lindi-, hi I'm trying to make gpsd 3.5 works with ubx on the gta02: Aug 01 12:21:27 I tried that: Aug 01 12:22:11 http://wiki.openmoko.org/wiki/Neo_FreeRunner_GPS#Configuration_for_a_higher_sampling_rate Aug 01 12:22:30 but that only changed the NMEA stream Aug 01 12:22:40 that is to say there is only GPRMC and GPGGA now Aug 01 12:24:32 GNUtoo: well which one did you try? you can't set both 4 Hz and 2 Hz? Aug 01 12:24:49 both Aug 01 12:24:50 it works Aug 01 12:24:57 but it doesn't switch to ubx binary mode Aug 01 12:25:28 or did I understand badly? Aug 01 12:25:56 heh http://klang.eudyptula.org/ and WTF! Aug 01 12:26:51 GNUtoo: ? Aug 01 12:27:15 GNUtoo: why would you expect those commands to enable reports in ubx protocol? Aug 01 12:27:39 lindi-, I want to replace ogpsd by gpsd Aug 01 12:27:47 but it need to have the same functionality Aug 01 12:27:58 so I'm looking for how to do it Aug 01 12:28:03 GNUtoo: good luck, gpsd upstream wasn't really interested :/ Aug 01 12:28:10 ah ouch Aug 01 12:28:11 GNUtoo: at least you should read the ubx spec first Aug 01 12:28:19 yes Aug 01 12:28:33 which one should I read? Aug 01 12:28:46 GNUtoo: set bit 0 in Out_proto_mask to enable reports in UBX Aug 01 12:28:54 GNUtoo: set bit 1 in Out_proto_mask to enable reports in NMEA Aug 01 12:28:58 ok thanks Aug 01 12:29:03 GNUtoo: there's only one spec afaik Aug 01 12:29:45 GNUtoo: http://lindi.iki.fi/lindi/gps/ubx.html is my own documentation but it's incomplete Aug 01 12:29:56 http://wiki.openmoko.org/wiki/Neo_FreeRunner_GPS#UBX_protocol has multiples specs Aug 01 12:30:09 but do you think it's easy/fast and worth to switch to gpsd? Aug 01 12:30:59 GNUtoo: ANTARIS_Protocol_Specification(GPS.G3-X-03002).chm Aug 01 12:31:09 ok Aug 01 12:31:24 GNUtoo: not really, gpsd is not designed for mobile use Aug 01 12:31:56 but it has hooks now, at least that's what PaulFertser told me if I remember well Aug 01 12:31:58 there's no real support for any kind of power management or agps Aug 01 12:32:07 yeah something limited Aug 01 12:32:24 for AGPS we could use hooks..... Aug 01 12:32:27 just don't expect it to be easy to add it to upstream Aug 01 12:32:31 I'm not sure where the hooks are Aug 01 12:32:40 but I saw some stuff in gpsctl Aug 01 12:32:55 gpsctl is also quite a hack :/ Aug 01 12:32:59 ok Aug 01 12:33:22 hmmm Aug 01 12:34:37 GNUtoo: surely it is possible to modify it but don't expect it to be completely straightfoward Aug 01 12:34:40 forward Aug 01 12:34:53 ok Aug 01 12:35:03 maybe I should keep ogpsd for now then Aug 01 12:35:28 fixing suspend for 2.6.39 would be bit more important imho :) Aug 01 12:38:26 yes indeed Aug 01 12:38:39 but I think GPS was broken not so long ago in gta02 Aug 01 12:38:48 not sure about the current state....it's still compiling Aug 01 12:40:59 GNUtoo: https://developer.berlios.de/bugs/?func=detailbug&bug_id=17431&group_id=2116 Aug 01 12:44:38 ah nice Aug 01 12:45:53 I wonder where are theses script hooks tough Aug 01 12:46:31 GNUtoo: /etc/gpsd/device-hook but it's really primitive Aug 01 12:46:46 the policy is hardcoded Aug 01 12:46:48 ok Aug 01 12:46:55 when first client connects it wants to activate Aug 01 12:47:06 ok Aug 01 12:47:12 so if you want to turn the gps off you need to close all your clients Aug 01 12:47:22 ok Aug 01 13:08:36 lindi-: hi, "fixing suspend for 2.6.39", is it for gta02? Any details? Which branch? Aug 01 13:10:10 lindi-, the UBX protocol is not online anymore and yours seem incomplete... Aug 01 13:12:01 lindi-, boot my GTA02 recently, while n900 discharged (one more time), and decided to do some hacking =) Aug 01 13:12:55 I'll try to find it harder Aug 01 13:14:37 lindi-, or there is something better to hack? Aug 01 14:09:24 SHR: 03shr-devel 07buildhistory * r1812e897bf17 10/packages/armv4t-oe-linux-gnueabi/python-sip/ (5 files in 5 dirs): packages: Build 201208011607 of shr 20120801 for machine om-gta02 on opmbuild Aug 01 14:10:43 gena2x: 2.6.39 yes http://lists.openmoko.org/pipermail/openmoko-kernel/2012-January/011386.html Aug 01 14:32:50 lindi-, i see. which git tree should i use and which kernel config? Aug 01 14:48:56 SHR: 03shr-devel 07buildhistory * rcd7f0a79deb8 10/packages/armv4t-oe-linux-gnueabi/ (10 files in 10 dirs): packages: Build 201208011636 of shr 20120801 for machine om-gta02 on opmbuild Aug 01 14:49:27 GNUtoo, hello Aug 01 15:14:40 JaMa, Automated builds are brutal on hardware. sometime we should talk. Aug 01 15:24:44 ayaka, hi Aug 01 15:25:13 GNUtoo, could you watch log today? Aug 01 15:25:36 Ive time todqy Aug 01 15:26:06 https://sumomo.r.gd/l0 Aug 01 15:27:04 ok Aug 01 15:27:15 what happened in that log exactly Aug 01 15:27:19 because it's quite long Aug 01 15:28:21 no thing happen, init doesn't run(at lease rsyslog doesn't), it seems initramfs doesn't be loaded either Aug 01 15:28:40 ha? Aug 01 15:28:49 but I am not sure it, that is why I need your help Aug 01 15:28:57 ah ok Aug 01 15:30:10 maybe I need to re do the base file Aug 01 15:30:26 you mean the board file? Aug 01 15:31:54 I can't understand, I mean the linux system file Aug 01 15:33:55 basically it mounts the rootfs Aug 01 15:34:00 it switches to it Aug 01 15:34:42 [ 5090.611022] ### Show Blocked State ### Aug 01 15:34:52 after that it reboots Aug 01 15:35:23 so look in kernel hacking section Aug 01 15:35:29 if it's forced to reboot on hangup Aug 01 15:35:32 and similar stuff Aug 01 15:35:37 else look in the code Aug 01 15:38:10 GNUtoo, it is me reboot it Aug 01 15:39:16 I press power, vol+ and vol- then it will reboot, but if init has been runed ,I need press more longer Aug 01 15:44:59 SHR: 03shr-devel 07buildhistory * rc1276fa32808 10/packages/armv4t-oe-linux-gnueabi/python-pyqt/ (5 files in 5 dirs): packages: Build 201208011728 of shr 20120801 for machine om-gta02 on opmbuild Aug 01 15:48:53 I have to sleep, good byte, but bouncer is still here Aug 01 15:51:57 ayaka, ok I'm not sure for init Aug 01 15:51:59 ok Aug 01 16:40:17 morphis, hi Aug 01 16:40:25 GNUtoo: heyho Aug 01 16:41:39 compilation take a too huge time for me here since I cannot compile the night anymore Aug 01 16:42:39 I wonder what I can do in between Aug 01 16:42:55 I cannot bisect the kernel since that requires to compile fast Aug 01 16:43:12 so gta02 kernel work cannot be done Aug 01 16:44:11 maybe it would take less time if I change the settings Aug 01 16:44:17 like parallel_make -- Aug 01 16:44:23 and threads++ Aug 01 16:44:53 why you can build at night anymore? Aug 01 16:44:57 s/can/cannot/ Aug 01 16:44:58 morphis meant: why you cannot build at night anymore? Aug 01 16:47:14 morphis, see PM Aug 01 16:48:49 GNUtoo: hm, so you're searching something to do? Aug 01 16:51:01 yes but something that doesn't require any heavy compilation Aug 01 16:51:05 hm Aug 01 16:51:16 powervr reverse engineering :D Aug 01 16:51:16 indeed Aug 01 16:51:19 lol Aug 01 16:52:37 hm Aug 01 16:53:12 maybe writing some python app Aug 01 16:53:36 if you have fun with that :) Aug 01 16:53:46 btw. webos UI components are now open source Aug 01 16:54:52 ok Aug 01 16:55:10 will there be a new kernel for older palm devices? Aug 01 16:55:32 no Aug 01 16:55:39 ok Aug 01 16:55:43 there will be no support for any devices Aug 01 16:55:44 and no indications of serial Aug 01 16:55:53 *on Aug 01 16:56:01 GNUtoo: http://blog.openwebosproject.org/post/28423182686/open-webos-july-edition Aug 01 16:56:29 ok let me look Aug 01 16:59:58 so basically they are looking for a fully supported device with 3.3+ kernel and free software support for drivers and userspace Aug 01 17:00:09 and they want to target that for webos Aug 01 17:00:20 it's doable on newer devices Aug 01 17:00:26 but there is still the 3d issue Aug 01 17:00:36 morning Aug 01 17:00:44 because newer devices have SIRF GPS, like galaxy nexus, SGS2 etc... Aug 01 17:00:51 hi Aug 01 17:01:18 3D will always be an issue Aug 01 17:01:23 indeed Aug 01 17:01:28 for now tough Aug 01 17:01:44 yup...so we ignore it since we can't do anything about it. Aug 01 17:02:07 yes Aug 01 17:02:42 many more bits to get working before that one is the last one to sqauash. Aug 01 17:03:05 webOS is upstart based Aug 01 17:03:14 yes but we looked at newer devices with paulk and other people.... Aug 01 17:03:18 and they seem doable Aug 01 17:03:32 good Aug 01 17:03:35 camera,framebuffer,audio,gps,etc.. seem doable Aug 01 17:03:43 and paulk has a sgs2 Aug 01 17:03:48 how is MT support? Aug 01 17:04:05 hmmm in crespo multitouch works as simple touch Aug 01 17:04:10 *single Aug 01 17:04:20 in GNU/Linux Aug 01 17:04:31 I guess there is a way to integrate it Aug 01 17:05:33 and modem is samsung in theses devices and there is fsogsmd+ libsamsung-ipc Aug 01 17:05:38 need to walk before we run. Aug 01 17:05:42 nice Aug 01 17:05:43 +samsung-ril in the case of android Aug 01 17:05:44 ok Aug 01 17:06:09 else there is the GTA04 Aug 01 17:06:21 there is a standard 3.5 kernel Aug 01 17:06:27 even @ 1.0 Open-WebOS will be "rough" and missing pieces. Aug 01 17:06:32 ok Aug 01 17:06:46 GNUtoo: on sgs2, you can either use the upstream kernel touchscreen driver (atmel mxt ts) which is the best multitouch driver in linux up to date, implementing newest frameworks. or you can use my cm9-based kernel which has hacks to make single-touch mode work in SHR Aug 01 17:06:52 and the big problem: it's based on framebuffer Aug 01 17:07:07 Alex[sp3dev], ah nice Aug 01 17:07:15 it's atmel mxt ts Aug 01 17:07:17 old was framebuffer Aug 01 17:07:21 same than nexus S? Aug 01 17:07:23 new should use GLES Aug 01 17:07:35 ah? Aug 01 17:07:41 so unusable without 3d? Aug 01 17:08:03 btw. maybe we should try to fix llvm@arm and make llvmpipe work on arm? with neon it should give decent performance for 2d ui Aug 01 17:08:31 let me ask a few questions about exactly what they are doing. Aug 01 17:08:39 (I'm more of a low level guy) Aug 01 17:08:44 ok Aug 01 17:09:16 GNUtoo: btw, if you were to paravirtualize linux (like l4android), would you a) port linux drivers to a hypervisor kernel or b) run multiple linuces each acting like a driver and replacing memory access with IPC calls? I just want to know what different developers think the easiest way is Aug 01 17:10:41 Alex[sp3dev], I've no idea sorry...hmm but you need an userspace solution rather? Aug 01 17:10:44 I'm told there is work on making the Android drivers work with linux Aug 01 17:10:44 like selinux? Aug 01 17:11:34 Alex[sp3dev], I guess you have free kernel drivers for sgs2 Aug 01 17:11:46 you rather need to isolate the userspace 3d blob right? Aug 01 17:12:02 GNUtoo: I'm not yet working on it but I got very interested by the idea and even plan to work on it in my free time for my projects besides work. The idea is that the minimal amount of trusted code (around 50 SLOC) runs in kernel mode and everything, including device drivers, runs in usermode Aug 01 17:12:20 ok Aug 01 17:12:23 GNUtoo: if we find an easy way to run unmodified drivers under a hypervisor, we can reuse any blobs Aug 01 17:12:44 yes but do you have any kernel blobs? Aug 01 17:12:48 I don't think so.... Aug 01 17:12:58 GNUtoo: nope. the only closed thing on sgs2 is the gpu Aug 01 17:13:30 galaxy nexus is another story. there's the closed-source dsp firmware controlling camera. Aug 01 17:14:16 ah their gpu driver is non-free at kerenl level? Aug 01 17:14:31 I tough they usually used a small pass-trough driver Aug 01 17:14:39 GNUtoo: no, kernel side is opensource for both mali and powervr, but the userspace is a blob Aug 01 17:14:45 ok Aug 01 17:15:01 ah and you want to prevent it to do bad thing Aug 01 17:15:06 like do kenrel level stuff Aug 01 17:15:10 trough the pass-trough Aug 01 17:15:11 ? Aug 01 17:17:38 GNUtoo: btw, imho it would be easy for somebody who knows python to add AGPS support using the lindi-'s ubx tools. I think it would work good enough in most situations. gpsd doesn't allow fine-grained GPS power control but is it really needed? Aug 01 17:17:43 GNUtoo: well, the closed gpu and drivers are not much of a problem in the short run. the idea of microkernels is to run everything in userspace and use fixed-size IPC to prevent buffer overwriting attacks which could allow a usermode app raise priviliges to kernel mode. But I still have some projects to finish before I can start hacking l4 Aug 01 17:18:05 PaulFertser, ok Aug 01 17:18:16 PaulFertser, I know python.... Aug 01 17:18:26 and I want to do something in python Aug 01 17:18:41 because it's compiling really hard right now Aug 01 17:18:49 GNUtoo: btw, we had some lectures on l4 and home assignments so I had some experience coding for l4. It's kinda fun that l4linux runs as a userspace app under L4 and links with L4Re libraries. I linked a C++ library against linux kernel and called C++ code from my modules ^_^ Aug 01 17:19:04 GNUtoo: i might be short-sighted but to me it seems that just adding the hooks to gpsd would be nice enough for smartphone usage. Aug 01 17:19:10 Alex[sp3dev], lol ok Aug 01 17:19:26 PaulFertser, the hooks can be done in python? Aug 01 17:19:55 GNUtoo: sure Aug 01 17:20:02 nice Aug 01 17:20:08 is there some docs for that? Aug 01 17:20:16 or some general hints? Aug 01 17:20:20 GNUtoo: yes, the gpsd manpage :) Aug 01 17:20:25 ok Aug 01 17:21:48 how far along is a Google Nexus BSP? Aug 01 17:22:51 ka6sox: for what? Aug 01 17:23:29 ka6sox, not so far, it just has to be added in meta-samsung and supported in fso Aug 01 17:23:38 JaMa|Off: elfe works on first glance, though i get the "settings are newer than e" message on startup Aug 01 17:23:45 I don't have a galaxy nexus so I cannot do it Aug 01 17:24:04 playya has one and was starting to get intial support for it in meta-samsung Aug 01 17:24:15 ka6sox: we need machine config + kernel recipe at least for it Aug 01 17:24:33 ok nice Aug 01 17:25:03 since its all 3.x already that shouldn't be too hard to extract from AOSP. Aug 01 17:25:46 I'll poke around meta-samsung todya Aug 01 17:26:05 ka6sox: there is a omap-tuna kernel tree in AOSP which we should use Aug 01 17:26:09 PaulFertser, with the hooks I should set the sysfs to enable modem + store/restore agps? Aug 01 17:27:06 GNUtoo: yes, it can work that way, i think it's better to use "om" though. Aug 01 17:27:12 ah ok Aug 01 17:27:50 morphis, that is what I was looking for. Aug 01 17:28:20 ok Aug 01 17:28:28 ka6sox: look here https://source.android.com/source/building-kernels.html Aug 01 17:29:52 I'll get this working on my ES and then deal with other platforms Aug 01 17:31:21 ok Aug 01 17:33:35 morphis: you'd better use cm10 kernel. and actually never use aosp. because stuff from there tends to only compile properly on a 64-bit ubuntu 10.10 with sun java5 :( Aug 01 17:33:57 Alex[sp3dev]: but not only the kernel :) Aug 01 17:34:44 at least aosp has full commit history unlike source code drops from OEMs Aug 01 17:43:16 morphis: regarding the fsogsmd API... is that already merged? Am I supposed to be fast because it already is broken? :-P Aug 01 17:44:16 mrmoku: no, it's not merged yet Aug 01 17:44:20 and it's not finished Aug 01 17:46:06 ok Aug 01 17:52:07 morphis: any estimate? Aug 01 17:53:04 mrmoku: not really Aug 01 17:53:13 but you can test with the morphis/device-status branch Aug 01 17:54:11 ok, that is supposed to work somehow then? Aug 01 17:56:17 in any way that should not be to bad to adopt to... Aug 01 17:56:24 s/to bad/too bad/ Aug 01 17:56:24 mrmoku meant: in any way that should not be too bad to adopt to... Aug 01 18:10:43 Too bad SD drive is set automatically in 2.6.39. I still suspect that's the root cause of my inability to use GPS now :( Aug 01 18:16:17 Just several minutes ago i was able to "see" three sats and now a gpsd compilation is ongoing and i restarted the gps chip and i see 0. Aug 01 18:16:50 hmmm Aug 01 18:16:57 mrmoku: so you like the new state-handling? Aug 01 18:17:03 gps sucks without agps Aug 01 18:18:17 Alex[sp3dev]: not the gta02's Aug 01 18:18:32 Alex[sp3dev]: it can get a fix in 45s from _cold start_ Aug 01 18:19:19 PaulFertser: which gps chip is that? quite impressive Aug 01 18:19:23 Alex[sp3dev]: ublox5 Aug 01 18:19:26 iirc Aug 01 18:20:32 Alex[sp3dev]: it can receive from many sats simultaneously, so with a clear view it's enough to download ephemeris from them. (every sat transmits its ephemeri every 30s) Aug 01 18:21:07 I couldn't believe it but then i tested and was convinced. Aug 01 18:21:30 morphis: it has some logic :-) Aug 01 18:21:32 PaulFertser: why do mainstream chips like CSR or Sirf are so slow? Aug 01 18:21:42 oops, my grammar Aug 01 18:21:48 I think i got something under 1 minute without any aid with fr sitting at the window even. Aug 01 18:21:52 anyway... got to go now... bbl Aug 01 18:22:10 Alex[sp3dev]: depends on the chip generation and also on the antenna used i think. Aug 01 18:22:19 lindi- surely can tell you more about it. Aug 01 18:22:30 mrmoku: :) Aug 01 18:22:37 mrmoku: it's already working quite nice Aug 01 18:22:44 I can't even get time right atm while the compilation is still ongoing :( Aug 01 18:23:00 That's enraging, especially given that it used to work so fine in every regard with sd_drive set lower. Aug 01 18:23:09 mrmoku: but I need to rework status handling for SIM and Network subsystems Aug 01 18:23:46 I asked larsc about the decision but he assured me it's allright. Now that it looks that it's not. Well, either that, or my internal antenna went bad. Aug 01 19:08:14 PaulFertser: stop compiler, recheck Aug 01 19:09:39 DocScrutinizer51: stopping various things, rechecking :) Aug 01 19:09:59 hi btw Aug 01 19:12:09 DocScrutinizer51: hi :) are you doing all right? Aug 01 19:13:19 yeah, trying to get drunk to celebrate 1st real vacancy eve (I forgot I had vacancy since monday already) Aug 01 19:13:33 and you? Aug 01 19:14:52 you can go under 40 sec for gps fix only with agps , ie if the gps is feed the hemisphere Aug 01 19:15:09 DocScrutinizer51: fine too :) Cycling, working, running OpenWrt etc Aug 01 20:03:21 hi there, I updated my FR this week. I tried the shr-core based image (not staging) ubifs did not boot at all. jffs booted everything went fine until I rebooted. After that I had strange window behaviour. The pin input always disappeared after a few seconds. I was fast enough, but the pin dialog kept coming up from time to time. Aug 01 20:05:54 then I tried before 110822, no such problem, but no sound in calls. Is there any image without show stopper? Or can anybody help me to get one of these images going,that I can use SHR as a phone again? Aug 01 21:13:22 jake42: hmm I had this warning in morning.. that's why I recreated config in offical image :/ Aug 01 21:15:40 ESI: if somebody reads the irc logs it might be useful to include md5sums and direct links to the images you used Aug 01 21:40:10 ah, sorry, I used 0e26a648fe868edbc887dbf17ada7cd3 shr-full-eglibc-ipk--20110822-om-gta02.rootfs.jffs2 as before 110822 Aug 01 21:42:33 ESI: I don't usually use ubifs and official images.. but some people reported weird behavior with ubifs before Aug 01 21:43:32 and 096967683f3a48fd5fca8ddb114ef864 shr-image-om-gta02-20120727155424.rootfs.jffs2 a shr-core Aug 01 21:45:08 Hi JaMa, I tried ubifs, but I did not get it to boot... I flashed qi for ubifs and the ubifs rootfs, (and kernel) but it did not boot. Aug 01 21:47:45 I flashed it twice, but nothing, so I switched to jffs2, and had the effect described above. Aug 01 21:49:55 ESI: have you tried to enable debug (holding POWER button?) Aug 01 21:51:28 because "did not boot" could be caused by many things Aug 01 21:52:05 ESI: and do you know the difference between .ubifs and .ubi right? Aug 01 21:54:27 ah no... I used ubi Aug 01 21:55:46 you need ubinized image which is .ubi, writing .ubifs with dfu-util wont work Aug 01 21:57:52 I did, The effect was: After the kernel unpacked the the screen backlight was on, but the screen was empty, pressing the "shoulder button" produced an escape charater, but nothing happened, I waited for several minutes... Aug 01 21:58:46 The same image, with jffs2 worked, but had the strange effects after second boot.... Aug 01 21:58:50 23:49:59 < JaMa|Off> ESI: have you tried to enable debug (holding POWER button?) Aug 01 22:00:34 JaMa, no I did not, I wasn't aware of this. Right now, I have the image (before shr-core) flashed, but without sound during call. Aug 01 22:01:20 JaMa, Should I flash shr-core again and acivate debug? Aug 01 22:17:50 if you want to know why it didn't boot for you.. Aug 01 22:30:14 JaMa, reflashing now, then remove SD-Card and try to boot holding the power button, right? Aug 01 22:33:04 ESI: yes or you can flash it from distro running on uSD, as described here http://wiki.openmoko.org/wiki/Ubifs to be sure it's written ok Aug 01 22:38:34 SHR: 03shr-devel 07buildhistory * rc3e90e93eadf 10/images/om_gta02/eglibc/shr-image/ (11 files): images: Build 201208020022 of shr 20120801 for machine om-gta02 on opmbuild Aug 01 22:53:48 SHR: 03shr-devel 07buildhistory * rb3d28969b99c 10/images/nokia900/eglibc/shr-image/ (11 files): images: Build 201208020039 of shr 20120801 for machine nokia900 on opmbuild Aug 01 22:53:56 SHR: 03shr-devel 07buildhistory * rcfaf66bbc4e7 10/packages/armv7a-vfp-neon-oe-linux-gnueabi/phonet-utils/ (7 files in 7 dirs): packages: Build 201208020039 of shr 20120801 for machine nokia900 on opmbuild Aug 01 23:00:47 JaMa: It booted, I cannot tell, what I did different this time... Aug 01 23:07:49 SHR: 03shr-devel 07buildhistory * rc65eab217b5c 10/images/om_gta04/eglibc/shr-image/ (11 files): images: Build 201208020054 of shr 20120801 for machine om-gta04 on opmbuild Aug 01 23:09:12 it worked also on second boot, maybe something on my sdcard was messing with SHR. Thank you for your help and all your effort you are doing here... Aug 01 23:20:46 SHR: 03shr-devel 07buildhistory * r04a155cae7e5 10/images/crespo/eglibc/shr-image/ (11 files): images: Build 201208020108 of shr 20120801 for machine crespo on opmbuild Aug 02 01:21:24 anybody near north bavaria? ping me for BBQ tomorrow eve **** ENDING LOGGING AT Thu Aug 02 02:59:58 2012