**** BEGIN LOGGING AT Tue Feb 21 02:59:58 2012 Feb 21 04:54:10 SHR: 03shr-devel 07buildhistory * rc37b6df6cb37 10/packages/ (112 files in 112 dirs): Build 201202210108 of shr 20120221 for machine om-gta04 on shr-chroot Feb 21 08:51:31 SHR: 03shr-devel 07buildhistory * r18daf53d50ce 10/packages/ (141 files in 141 dirs): Build 201202210558 of shr 20120221 for machine palmpre on shr-chroot Feb 21 09:41:21 blist Feb 21 11:11:27 SHR: 03shr-devel 07buildhistory * r2933366b282a 10/packages/palmpre2-oe-linux-gnueabi/ (113 files in 113 dirs): Build 201202210956 of shr 20120221 for machine palmpre2 on shr-chroot Feb 21 12:19:21 GNUtoo|laptop, hi Feb 21 12:19:28 I tried the latest crespo shr builds Feb 21 12:19:33 sound doesn't work Feb 21 12:23:58 what you mean by latest? Feb 21 12:40:41 paulk-desktop, it should work if you install fsoaudiod's /etc/asound.conf Feb 21 12:40:55 paulk-desktop, do you have /etc/asound.conf Feb 21 12:40:56 ? Feb 21 12:45:39 I'll check Feb 21 12:45:44 JaMa, latest shr-core-staging Feb 21 12:46:27 paulk-desktop: thanks, 028 is not yet finished for crespo Feb 21 12:46:46 it is: shr-image-crespo-20120210002241.rootfs Feb 21 12:47:10 GNUtoo|laptop, yes, it's there Feb 21 12:47:57 also, when the screen goes off and resumes, it gets to dmesg output Feb 21 12:48:01 not to X Feb 21 12:48:03 so I have to restart X Feb 21 12:48:16 hmm that's not where "latest" points.. shr-image-crespo-20120210002241.rootfs is 028 Feb 21 12:48:52 anyway, I have configured opkg to use the literally "latest" feeds Feb 21 12:48:59 shr-image-crespo-20120206013244.rootfs.tar.gz is latest closed Feb 21 12:49:12 anyway shr-image-crespo-20120210002241.rootfs should be broken at least by OOM issue Feb 21 12:49:41 ouch Feb 21 12:49:42 and opkg pointing to latest feeds won't downgrade it to right versions.. Feb 21 12:49:50 when was it right? Feb 21 12:50:04 027 is right and "latest" Feb 21 12:50:04 perhaps I should just take the shr-core image Feb 21 12:50:28 how is aurora btw? Feb 21 12:50:29 028 is WIP and probably will finish build tonight or tomorrow morning Feb 21 12:50:49 I see Feb 21 12:50:51 thanks JaMa Feb 21 12:51:00 I'll use 027 Feb 21 12:51:50 GNUtoo|laptop, btw, I understood why we have b&w camera preview on replicant Feb 21 12:52:31 it's because opengl (via surfaceflinger) only shows the Y plane of YUV feeds Feb 21 12:52:41 so I'm trying to get RGB out of the camera Feb 21 12:52:49 and now I have a way better understanding of it Feb 21 12:53:04 finally, it's standard V4L2, the non-standard stuff is done in a very clean way Feb 21 12:53:22 like with an appropriate v4l2 ioctl Feb 21 12:53:32 I tried with mplayer Feb 21 12:53:55 and it fails because it needs one of these non-standard requests to select which camera to use Feb 21 12:54:17 we could easily workaround the driver to make it assume one by default Feb 21 12:54:37 paulk-desktop: mplayer will fail anyway because these drivers use mutiplanar api, no? Feb 21 12:54:49 hi btw Feb 21 12:55:01 no, it's really standard Feb 21 12:55:29 the error message on fimc is that it didn't set the camera to use prior of asking the feed Feb 21 12:56:04 paulk-desktop: i have not looked at 3.0 android tree but what we now have in vanilla kernel is a new api, which is not compatible with old tools. Feb 21 12:56:04 but as I looked, the android libcamera really does the same as any v4l2 sample app Feb 21 12:57:02 ok, i won't misdirect you. just keep in mind there's some new mplane api extensions and media controller api in vanilla and i've not figured it out yet Feb 21 12:57:51 ok Feb 21 12:58:25 ok Feb 21 12:58:47 SHR: 03shr-devel 07buildhistory * r522f817c6328 10/packages/crespo-oe-linux-gnueabi/ (117 files in 117 dirs): Build 201202211212 of shr 20120221 for machine crespo on shr-chroot Feb 21 12:58:54 any news on sgs2 modem, btw? i hope i will be able to port svnet to vanilla if i have time tonight Feb 21 12:59:11 heyho Feb 21 12:59:26 heyho Feb 21 12:59:38 Alex[sp3dev], someone named grindars have started to write software to boot it Feb 21 12:59:40 hey morphis Feb 21 12:59:53 JaMa: I build eglibc only in shr-root and then afterwards a full shr-lite-image Feb 21 13:00:00 that worked for me Feb 21 13:00:31 ok, i don't care about modem actually. sound and mali are more important. anyway, back to work Feb 21 13:00:32 morphis: weird Feb 21 13:00:42 morphis, I'm doing the manual/automatic network selection stuff for nexus s on replicant RIL. You were asking something related the other day, weren't you? Feb 21 13:01:01 morphis: but if something -native was breaking your build then it's possible the cause I'm not seeing it (as I always build smaller parts like gcc first) Feb 21 13:01:18 because of limited space on tmp partition.. Feb 21 13:01:26 JaMa: ok Feb 21 13:01:45 JaMa: I will try a little bit harder when I have more time Feb 21 13:01:58 paulk-desktop: ok, I am currently doing RFS integration in fsogsmd Feb 21 13:02:08 nice Feb 21 13:02:21 paulk-desktop: I was asking about network/sim availbility in LPM mode Feb 21 13:02:25 we discovered that it can't go in low power mode without RFS correctly answered Feb 21 13:02:30 yes Feb 21 13:02:34 ah ok Feb 21 13:02:36 that I found out too :) Feb 21 13:02:47 thats why I am doing the RFS stuff now Feb 21 13:03:00 the ICS radio log is a little bit confusing there Feb 21 13:03:16 IIRc it's less verbose than the GB one, isn't it? Feb 21 13:03:49 it's different Feb 21 13:03:58 sometimes it offers more details Feb 21 13:04:07 ok Feb 21 13:04:10 but for example hexdumps are limitied to a number of bytes Feb 21 13:04:17 so you never get the complete dump Feb 21 13:04:21 ouch Feb 21 13:04:23 yes Feb 21 13:05:56 I am at launch now Feb 21 13:06:01 let's talk later Feb 21 13:06:18 ok Feb 21 14:24:44 JaMa, did you see my conversation with lindi- last night? Feb 21 14:24:53 about thumb interworking Feb 21 14:25:01 not whole Feb 21 14:25:33 eglibc and bash were compiled differently Feb 21 14:25:36 one with -marm Feb 21 14:25:40 and one with -mthumb Feb 21 14:25:59 that's normal Feb 21 14:26:50 thumb is default for armv4t and whatever can override it with ARM_INSTRUCTION_SET Feb 21 14:26:55 in recipe Feb 21 14:32:53 yes that's when thumb interworking works Feb 21 14:34:02 but according to lindi's arm manual ldr pc, [foo] is not a valid change from arm to thumb Feb 21 14:34:12 and that's exactly what it does Feb 21 14:34:27

GNUtoo-desktop: on armv4t you can't use ldr to switch to thumb mode Feb 21 14:49:45 have you tried to rebuild it with bfd? Feb 21 14:50:37 there is also patchset from khem changing thumb overrides maybe he had seen some issues too Feb 21 14:53:01 and we had some patches for thumb support in bfd which were later applied upstream, but maybe only to bfd not gold... like this one http://git.openembedded.org/openembedded/tree/recipes/binutils/binutils-2.19.51/thumb-func.patch Feb 21 14:54:51 I was trying to find where it was applied but upstream CVS repo log is a bit messy Feb 21 14:55:31 or at least hard to read as most commit messages are just bug numbers Feb 21 15:08:04 JaMa|Off, not yet with bfd, how should I do that? Feb 21 15:08:24 ok Feb 21 15:08:37 JaMa|Off, maybe use git to clone the svn (very slow) Feb 21 15:10:39 ? Feb 21 15:10:44 binutils are in CVS not svn Feb 21 15:11:07 and see the commit where I've enabled gold to see how to disable it Feb 21 15:11:35 http://git.shr-project.org/git/?p=meta-smartphone.git;a=commit;h=8b1a5c8c2443660837e1ef7989e8ae5bd68c84ad Feb 21 15:12:00 JaMa|Off, ah ok but you can clone cvs with git too if I remember well Feb 21 15:12:47 or something like http://git.shr-project.org/git/?p=meta-smartphone.git;a=commit;h=336d0ca3f2d54cd56122f39e0771d1cd06fa4abd for single recipe Feb 21 15:12:59 well and how it will improve commit messages? Feb 21 15:13:18 it won't Feb 21 15:13:32 but at least it will give you full git power Feb 21 15:14:05 with git blame Feb 21 15:14:07 etc... Feb 21 15:14:11 is not going to rebuild git with cvs support just because of binutils :) Feb 21 15:14:20 but maybe with next git upgrade I'll enable it Feb 21 15:14:26 ok Feb 21 15:15:08 finally at least 50G free on buildhost, now time to build images Feb 21 15:15:19 ok Feb 21 15:18:19 SHR: 03shr-devel 07buildhistory * r16580d615a74 10/packages/nokia900-oe-linux-gnueabi/shr-version/ (latest shr-version/latest): Build 201202211406 of shr 20120221 for machine nokia900 on shr-chroot Feb 21 15:42:18 Heya Feb 21 15:47:28 GNUtoo|laptop, uhm, this is very strange, I'm trying to determine a sequence to initilize the modem on geeksphone one, but no matter what I do, it ignores it, until 15 minutes or so, it starts working automagically Feb 21 15:47:47 without any sequence Feb 21 15:48:34 what can be happening? the sim is autoconfiguring the modem? the network can configure the modem? Feb 21 15:49:41 after the modem get's in this status, fsogsmd works, before the modem gets in it, it always error when doing AT+CPIN? Feb 21 15:50:24 ouch Feb 21 15:50:45 define initialize Feb 21 15:51:05 by initialize you mean until what state? Feb 21 15:51:05 GNUtoo|laptop, ok, first I boot the phone Feb 21 15:51:12 until it respond to AT commands? Feb 21 15:51:17 GNUtoo|laptop, yes Feb 21 15:51:19 or later states Feb 21 15:51:20 ok Feb 21 15:51:27 GNUtoo|laptop, but, for example AT works always Feb 21 15:51:30 ATI Feb 21 15:51:35 etc... Feb 21 15:51:48 others doesn't work for example AT+CSCS=? Feb 21 15:51:49 what doesn't work then Feb 21 15:51:54 until the modem reachs this state Feb 21 15:51:59 AT+CPIN? Feb 21 15:51:59 ok Feb 21 15:52:04 AT+CRC Feb 21 15:52:15 AT+CFUN=1? Feb 21 15:52:24 AT+CFUN always works Feb 21 15:52:36 maybe you need the right combinaison of AT commands Feb 21 15:52:52 GNUtoo|laptop, I tried a lot of combinations Feb 21 15:54:52 maybe you should ask a modem expert Feb 21 15:54:54 hi radekp Feb 21 15:54:54 so no idea of what can be happening I tried reseting the modem with ATZ wait 2 seconds then ATE0V1 again ATE0V1 then AT+CFUN=1 wait 2 seconds Feb 21 15:55:11 but it starts working when it wants Feb 21 15:55:17 hmmm Feb 21 15:55:26 GNUtoo|laptop: hi Feb 21 15:55:53 radekp, about echo, I guess you didn't look in oslec yet Feb 21 15:56:21 GNUtoo|laptop: i saw, but i didnt like it works with kernel Feb 21 15:56:36 ok Feb 21 15:56:50 maybe we should userspace it then? Feb 21 15:56:57 GNUtoo|laptop, ok, who knows about modems here? Feb 21 15:57:00 GNUtoo|laptop: i dont get why this should need any kernel driver... it's just some computing on array of bytes... Feb 21 15:57:18 VQuickSilver, mickey_office does Feb 21 15:57:25 but he's hard to catch Feb 21 15:57:26 GNUtoo|laptop: the walkie talkie algorithm works quite nice - i had 2 calls today and they were ok Feb 21 15:57:36 ok Feb 21 15:57:55 GNUtoo|laptop, ok, thanks Feb 21 15:58:22 GNUtoo|laptop: i need to try something with speex - which looks exactly what we need, but today i am busy at paid work and tomorrow i am leaving for 5 days Feb 21 15:59:28 ok Feb 21 16:00:13 radekp, I'll re-read your mail Feb 21 16:01:07 GNUtoo|laptop: btw i just pushed latest version which works for me ok - with qtmoko's earpiece state Feb 21 16:03:27 ok Feb 21 16:03:37 I just tried last version and same error Feb 21 16:03:46 does qtmoko have an /etc/asound Feb 21 16:03:51 */etc/asound.conf Feb 21 16:03:53 no Feb 21 16:03:54 or .asoundrc Feb 21 16:03:56 ok Feb 21 16:04:19 VQuickSilver: why don't you trace proprietary ril for proper init? anyway, most msm modems need some at commands on cold boot Feb 21 16:04:40 Alex[sp3dev], because the ril don't use AT Feb 21 16:05:05 Alex[sp3dev], yeah the ril doesn't even use /dev/smd0 Feb 21 16:05:15 radekp, ok works without /etc/asound.conf Feb 21 16:05:17 let me try Feb 21 16:05:25 it's using a rpc protocol I think Feb 21 16:06:16 you can add a printk to rpc driver. and you can hook write syscall via LD_PRELOAD to /dev/rpc stuff probably Feb 21 16:06:22 anyway Feb 21 16:06:23 https://gist.github.com/CyanogenMod/android_device_htc_leo/blob/gingerbread/libhtc_ril_wrapper/libhtc_ril_wrapper.c#L713 Feb 21 16:06:50 not sure if your device needs it, but some htc modems misbehave without this 3 lines Feb 21 16:08:21 Alex[sp3dev], uhm, no idea, going to test it Feb 21 16:10:49 radekp, however it often returns Feb 21 16:11:16 http://www.pastie.org/private/dmcbdcvl1defmhvuuqzdg Feb 21 16:11:37 GNUtoo|laptop: that's ok Feb 21 16:11:44 it means hangup Feb 21 16:11:50 ok Feb 21 16:12:33 i think we can later disable the error messages, but they are quite useful now... Feb 21 16:12:54 i had another call now and it was working nice :) Feb 21 16:13:00 Alex[sp3dev], nope it doesn't work Feb 21 16:13:12 Alex[sp3dev], only AT+CFUN=0 returns OK Feb 21 16:17:55 morphis: still here? Feb 21 16:18:27 radekp, do you have DTMF working btw Feb 21 16:19:26 GNUtoo|laptop: that's the thing where you press numbers e.g. for operator automats? Feb 21 16:19:46 morphis: can we move vala-0.14.2 to meta-oe where is 0.12.1 only or do you need to keep it in meta-fso? I have small patch with prevents incorect -dirty append (for which I have stupid workaround in midori) Feb 21 16:19:58 yes Feb 21 16:20:05 GNUtoo|laptop: yes, i think it works Feb 21 16:21:10 morphis, hi Feb 21 16:25:12 morphis, 1) about the fsoaudiod routing plugin, should we integrate radekp's work (no config files) or should we do something else(like add config to radek's work or to improve the current one ) Feb 21 16:25:26 2) opkg files on fsoaudiod-config gives: Feb 21 16:25:33 /etc/freesmartphone/conf/palm_pre/fsoaudiod.conf Feb 21 16:25:34 /etc/asound.conf Feb 21 16:25:37 on a gta04.... Feb 21 16:55:35 JaMa|Off: you can move it to meta-oe Feb 21 16:55:47 SHR: 03Martin.Jansa 07meta-smartphone * re06f1212b1b1 10/meta-fso/recipes-devtools/vala/ (2 files in 2 dirs): vala: add patch which prevent incorrect -dirty append to reported version Feb 21 16:56:01 GNUtoo|laptop: you mean the alsa routing thing? Feb 21 16:57:43 morphis: ok will do with next pull request Feb 21 16:58:04 JaMa|Off: thanks a lot Feb 21 17:00:22 JaMa|Off: what do you think is left to get the 2012.01 release of SHR out? Feb 21 17:03:17 morphis: OOM issue is hopefully fixed ( but I see another issues now - but I have newer EFL, haven't tested official image yet as it's still building) Feb 21 17:03:37 morphis: and more pressing issue is SIGILL all over armv4t images.. GNUtoo|laptop is trying to fix it Feb 21 17:05:52 JaMa|Off: ok Feb 21 17:07:02 morphis: do you have some patches for vala upstream? (that you could include my commit which I just added to meta-oe?) Feb 21 17:08:00 JaMa|Off: no, all patches are upstream with 0.14.2 Feb 21 17:08:23 ok then I'll send it upstream myself Feb 21 17:08:28 JaMa|Off: but don fry to report it upstream Feb 21 17:08:58 ? Feb 21 17:09:52 morphis, yes Feb 21 17:10:24 GNUtoo|laptop: is you approach not working? Feb 21 17:10:51 no echo cancelation yet Feb 21 17:10:58 and sometimes huge cpu usage Feb 21 17:10:59 s/don fry/don't try/ ? Feb 21 17:11:25 because of buffer underrun/overrun issues Feb 21 17:11:51 but I can try to import radek improvements in alsaloop Feb 21 17:12:29 JaMa|Off: yes Feb 21 17:12:36 why? Feb 21 17:12:45 do they eat strangers? Feb 21 17:12:45 basically he has echo cancelation Feb 21 17:13:00 and he has a way to start only when the pcm is ready Feb 21 17:13:04 not only the call Feb 21 17:13:06 but the pcm Feb 21 17:13:59 JaMa|Off: no :) Feb 21 17:14:33 GNUtoo|laptop: so his approach is the better one? Feb 21 17:14:43 he has more features Feb 21 17:15:05 * echo cancelation based on a walkie-talkie algorithm Feb 21 17:15:25 * starts when the modem pcm is ready Feb 21 17:15:31 but we have a config file Feb 21 17:15:39 he doesn't have a config file Feb 21 17:15:40 GNUtoo|laptop: feature are ok, I just don't want to swap the implementation of the routing plugin from time to time Feb 21 17:15:46 and we are more tweakable Feb 21 17:15:53 ok Feb 21 17:15:55 then use it Feb 21 17:16:09 maybe I'll try importing his features then Feb 21 17:16:21 and see how it works Feb 21 17:17:20 ok Feb 21 17:43:22 mrmoku, hi Feb 21 17:44:04 GNUtoo|laptop: btw. I saw tslib has support for the multitouch protocol too Feb 21 17:44:12 ok Feb 21 17:44:13 so we can move on from xf86-input-mtdev to tslib Feb 21 17:44:19 maybe, when it's beter Feb 21 17:44:32 what's the advantage beside having it for framebuffer? Feb 21 17:46:48 SHR: 03shr-devel 07buildhistory * r2acc876c0c3b 10/packages/armv7a-vfp-neon-oe-linux-gnueabi/midori/ (44 files in 44 dirs): Build 201202211844 of shr 20120221 for machine nokia900 on shr-chroot Feb 21 18:04:11 SHR: 03shr-devel 07buildhistory * r5225d33356aa 10/packages/armv7a-vfp-neon-oe-linux-gnueabi/vala/ (7 files in 7 dirs): Build 201202211902 of shr 20120221 for machine nokia900 on shr-chroot Feb 21 18:04:51 SHR: 03Martin.Jansa 07meta-smartphone * r02f360046234 10/meta-fso/recipes-devtools/vala/ (2 files in 2 dirs): meta-fso: drop vala, moved to meta-oe Feb 21 18:30:27 * GNUtoo|laptop tought mrmoku would be back Feb 21 18:30:49 sigh Feb 21 18:51:19 freesmartphone.org: 03GNUtoo 07cornucopia * ref4a272e1cdf 10/fsoaudiod/src/plugins/gsmvoice_alsa_forwarder/alsaloop.c: Feb 21 18:51:19 freesmartphone.org: fsoaudiod: gsmvoice_alsa_fowarder: fix killing of fsogsmd Feb 21 18:51:19 freesmartphone.org: Before, running the alsaloop plugin prevented fsogsmd from Feb 21 18:51:19 freesmartphone.org: beeing killed, this fixes it. Feb 21 18:51:19 freesmartphone.org: Signed-off-by: Denis 'GNUtoo' Carikli Feb 21 18:52:13 mrmoku, are you back and should we work on gta04 or should I abandon and continue waiting for you? Feb 21 18:52:35 ouch Feb 21 18:52:42 I did a bad commit message Feb 21 18:52:49 it's fix killing of fsoaudiod Feb 21 18:52:51 not fsogsmd Feb 21 18:52:56 sigh Feb 21 18:54:04 is there someone here? Feb 21 19:22:50 lindi-: "1. Debian will not knowingly distribute software encumbered by patents" Ahem, how can it be so, isn't Linux itself affected? Feb 21 21:19:54 PaulFertser: no comment Feb 21 22:03:28 SHR: 03shr-devel 07buildhistory * r2fa88289fb07 10/packages/ (85 files in 85 dirs): Build 201202212256 of shr 20120221 for machine nokia900 on opmbuild Feb 21 22:13:51 SHR: 03shr-devel 07buildhistory * r37b66966c878 10/packages/armv7a-vfp-neon-oe-linux-gnueabi/midori/ (44 files in 44 dirs): Build 201202212311 of shr 20120221 for machine nokia900 on opmbuild Feb 21 22:24:19 PaulFertser: I'm not a lawyer and don't think that my speculation would do any good here :) Feb 21 23:41:51 SHR: 03shr-devel 07buildhistory * re8df6491552e 10/ (35 files in 25 dirs): Build 201202212314 of shr 20120221 for machine nokia900 on opmbuild Feb 22 01:21:51 Project shr-core-deploy-cleanup build #10: SUCCESS in 12 min: http://norman-schleicher.de/jenkins/job/shr-core-deploy-cleanup/10/ Feb 22 01:24:14 SHR: 03shr-devel 07buildhistory * r894eacdf9718 10/ (157 files in 147 dirs): Build 201202220056 of shr 20120221 for machine nokia900 on opmbuild **** ENDING LOGGING AT Wed Feb 22 02:59:58 2012