**** BEGIN LOGGING AT Wed Aug 21 02:59:58 2019 Aug 21 06:41:05 Morning! Aug 21 06:41:35 Tofe: Yeah but seems they're on oFono 1.19 and not 1.21 like us? Aug 21 06:41:50 Or we didn't upgrade to Mer's 1.21 yet? In the train now and laptop battery died ;) Aug 21 06:41:56 So cannot check easily Aug 21 07:29:48 Herrie: we tried Mer's 1.21, oFono 1.26, 1.29 and 1.30 iirc Aug 21 07:30:44 I'll look for any patch from ubports on oFono Aug 21 07:34:27 Tofe: Not really sure what they're using nowadays: https://github.com/ubports/ofono/branches Aug 21 07:34:35 Seems this is the old rilmodem, not Mer's oFono :S Aug 21 07:35:02 Though there's a branch with merofono too Aug 21 07:35:04 Well we never used Mer's "ril" modem Aug 21 07:35:47 I mean their branch is a fork of https://github.com/rilmodem/ofono Aug 21 07:35:52 oh... Aug 21 07:36:15 But let's figure out what they're actually building Aug 21 07:36:18 that's the one that got merged some years ago Aug 21 07:36:58 Yup, so not really sure what they're building. Having a look Aug 21 07:37:28 On my side I'll check if their fixes are already present upstream Aug 21 07:39:25 Based on the filename of the .deb I would say they're just building: https://github.com/ubports/ofono (ofono-1.18.0+ubports1+0~20181106172057.3~1.gbp518afa) 518afa is HEAD on xenial branch Aug 21 07:40:22 Though seems that's even 1.17, even though their filename suggests 1.18. Changelog is for 1.17 Aug 21 07:40:40 Ah it's bumped, but changelog not updated: https://github.com/ubports/ofono/commit/1a2fdffbd18d58e9dc04b1d729d4b85053bcbfa4 Aug 21 07:43:09 I don't understand on which source code they're based exactly; it doesn't match https://git.kernel.org/pub/scm/network/ofono/ofono.git/tree/drivers/rilmodem?h=1.18 Aug 21 07:44:59 ah ok "This git repository is a temporary fork of ofono 1.12 meant as a staging area for work on rilmodem, an ofono driver and associated plugins used to provide a telephony stack based upon an Android RILD instance ( vs. a native modem driver ). This repository is *also* the upstream for Ubuntu's version of ofono." Aug 21 07:47:26 so they only have 4-5 commits ahead of 1.12 rilmodem version, in the end Aug 21 07:49:16 ooh that's OLD Aug 21 07:49:38 I would expect that Mer's version should be a lot better to go with. Also seeing that UBPorts uses Telepathy which we don't Aug 21 07:49:47 Stack wise we're a lot closer to Mer I'd say Aug 21 07:52:13 And they have working Android 7/8/9 based targets Aug 21 07:55:06 but with their own driver Aug 21 07:59:33 True Aug 21 07:59:40 Well we could try to use their driver? Aug 21 08:00:11 Yes, but it's either we clean up unwanted dependencies (mce, sailfish-manager) or we also build them Aug 21 08:00:39 for the sailfish-manager I think it's in the same source code Aug 21 08:01:52 Didn't we disable most MCE & sailfish-manager bits or was that elsewhere in the code only? Aug 21 08:02:03 We can make them conditional, shouldn't be too hard hopefully Aug 21 08:02:24 it could be tricky; ok I'll try a cleanup Aug 21 08:02:26 Ideally there should be a working RILMODEM driver upstream but that can be a second step Aug 21 08:02:54 Once it's cleaned up, upstreaming should be fairly straightforward Aug 21 08:35:27 novaldex|away: ping Aug 21 09:04:00 Tofe: THere was a little bump upstream that might be useful for us too (at least for the RILMODEM targets): https://git.merproject.org/mer-core/ofono/commit/e38a63d179bb25e1fbde94288ae407fecabce5fa Aug 21 09:20:32 Herrie: same status with the "ril" modem, unfortunately Aug 21 09:21:04 (I chose not to cleanup and include libmce-glib) Aug 21 09:22:22 Herrie: yes, we alraedy have that fix included in one of our patches https://github.com/webOS-ports/meta-webos-ports/blob/warrior/meta-luneos/recipes-connectivity/ofono/ofono/0001-Enable-the-various-modem-plugins-we-support-again.patch#L27 Aug 21 09:27:55 the logs look maybe a bit nicer with Mer's driver Aug 21 09:34:06 Tofe: Ah that's probably the issue I had with upstream oFono. I didn't remove the connman bits i guess Aug 21 09:47:52 Tofe: What's libmce anyway? Aug 21 09:48:44 libmce-glib helps Mer's oFono driver communicate with mce over dbus Aug 21 09:48:44 mce = Mer specific no? Their equivalent for much we have in LunaSysMgr? Aug 21 09:49:19 I didn't install mce, of course; but it's just some power management daemon iirc Aug 21 09:49:39 Yeah their documentation in repos is not great as usual :S https://git.merproject.org/mer-core/mce Aug 21 09:50:01 "mce - daemon that provides mode control functionality" Aug 21 09:50:12 Mode Control Entity yes Aug 21 09:50:24 low-power, locked, etc I guess Aug 21 09:50:47 Yeha so basically our powerd and lunasysmgr I guess Aug 21 09:51:23 yes, something like that Aug 21 09:52:30 but to me it does look like we're chasing the wrong rabbit Aug 21 09:53:45 each time, we're doing the right thing with oFono, but the Android side doesn't follow Aug 21 10:02:09 Maybe we're missing some of the glue bits? Aug 21 11:34:40 Tofe: Having a look into Mer's Android build to see if we're missing something in Halim 7.1 Aug 21 11:34:44 +u Aug 21 11:45:51 This might be of help, was recently merged in Halium: https://github.com/Halium/android_build/commit/85409672a6bbbc17e45b2e8613e3f9882bfce800 Aug 21 12:19:40 Herrie: well, they're switching from libcompat to droidmedia, it seems Aug 21 12:25:18 Tofe: Yeah, but some bits seem needed for audio calls as well Aug 21 12:29:57 At least miniaf according to https://github.com/Halium/android_build/pull/17 Aug 21 12:40:31 Tofe: I merged some of the pending Halium PR's for the various bits. Just not very familiar with this part. Looks OK to you? Or this is UB Ports specific? https://github.com/Halium/android_frameworks_native/pull/5/commits/2541bacd0ef1c5ec5d841831a6ebb8174a6276dc Aug 21 12:43:02 Herrie: lgtm yes, we don't use SurfaceFlinger either, same for all Halium users theoretically Aug 21 12:48:02 OK everything related to camera seems to have merged now Aug 21 12:48:12 Let me kick off a Halium-7.1 build on Jenkins Aug 21 13:05:38 Ah failing.... :S Aug 21 13:05:43 http://jenkins.nas-admin.org/job/LuneOS/job/halium-luneos-7.1-build/33/console Aug 21 13:05:49 I guess need JaMa or Novaldex for that Aug 21 13:06:34 Let me try locally at least ;) Aug 21 13:08:17 Tofe: You're working on Tissot right? Aug 21 13:08:27 rigt Aug 21 13:08:29 +h Aug 21 13:11:49 https://github.com/webOS-ports/jenkins-jobs/blob/master/jenkins-job.sh#L125 here we lack the "LuneOS/" prefix Aug 21 13:12:47 I can try a modification Aug 21 13:14:59 https://github.com/webOS-ports/jenkins-jobs/pull/11 Aug 21 13:17:05 Tofe: Ah true, can you bump the script version ? BUILD_SCRIPT_VERSION="2.5.32" -> BUILD_SCRIPT_VERSION="2.5.33"? Aug 21 13:18:58 Done, though in a separate commit, I used the web interface Aug 21 13:19:53 No worries, squashed while merging by web interface ;) Aug 21 13:19:57 New build looks better Aug 21 13:28:38 good! Aug 21 13:29:55 Locally I have some issues though, so I guess ours will also fail Aug 21 13:30:11 Seems there were some recent changes in LineageOS for CM 14.1 :S Aug 21 13:30:38 Seems this one causing havoc: https://github.com/LineageOS/android_frameworks_av/commit/f98ed1833cc7625c9f96e8de0843d3ec9da82d2f Aug 21 13:32:06 I'm getting https://bpaste.net/show/JuX- Aug 21 13:32:14 Should be easy enough to fix, looking into it now Aug 21 13:33:06 are you rebuilding from scratch? Aug 21 13:33:35 Tofe: Yup Aug 21 13:34:32 Connection speed & CPU speed no issues here :P Aug 21 13:34:35 then it looks like hardware/qcom/audio-caf needs to be adapted, and wasn't Aug 21 13:36:39 Other way around. https://github.com/LineageOS/android_hardware_qcom_audio/commit/f61cda38040e5f5466aea79f00d8ff0ddd13f74b was adopted however Halium keeps a fork of frameworks/av which wasn't: https://github.com/webOS-ports/android/blob/luneos-halium-7.1/halium/default.xml#L143 Aug 21 13:37:01 For the CAF audio we use the upstream LOS directly without patches Aug 21 13:37:15 So the patch needs to be cherry picked to Halium, will sort that Aug 21 13:37:32 ok Aug 21 13:49:44 OK simply cherry picked that commit for now Aug 21 13:50:05 There are a couple of dozen more commits it seems, don't feel like rebasing and rechecking everything today :P Aug 21 13:54:53 OK succeeded locally Aug 21 13:54:58 You want me to put it on my FTP? Aug 21 13:56:43 Tofe: OK halium-luneos-7.1-20190821-1-tissot.tar.bz2 is there Aug 21 13:56:59 I'll fetch it :) Aug 21 13:57:41 it's quite small Aug 21 13:58:56 Yup ;) Aug 21 13:58:59 183MB Aug 21 14:05:32 also I'll push my little modifications that lead to using Mer's ofono RIL driver Aug 21 14:05:39 as a separate branch Aug 21 14:09:01 Having the updated Halium bits might help a bit for getting the various gstreamer bits added later :) Aug 21 14:09:19 done; https://github.com/Tofee/meta-webos-ports/commit/089d32e138dfea8fee0add65100991c80a00ca50 Aug 21 14:09:24 Since libcompat anyway wouldn't work with 7.1 if I understood correctly? Aug 21 14:09:46 yes, that would be a necessary step Aug 21 14:10:40 I've now installed the new halium image, let's see... Aug 21 14:24:49 nope, same status :/ Aug 21 14:27:31 Any logs? logcat, journalctl ? Aug 21 15:04:56 well, yes, always the same ones Aug 21 15:06:17 like this https://paste2.org/5N4cPYYK Aug 21 15:11:21 Tofe: For the no space errors I came across https://stackoverflow.com/questions/28348627/echo-tasks-gives-no-space-left-on-device-when-trying-to-use-cpuset and "mkdir /dev/cpuset && mount -t cpuset cpuset /dev/cpuset" for possible solutions Aug 21 15:13:15 seems to work interactively, let me insert it during boot Aug 21 15:18:54 Tofe: That should kill some logspam at least which is always good ;) Aug 21 15:22:43 mmmh it's still there, though the mount exists Aug 21 16:40:36 Tofe: Well was worth a try Aug 21 16:44:12 yes, it looked promising... I don't know why it didn't work Aug 21 17:23:13 Herrie: we only have xiaomi devices as halium-7 devices, currently, isn't it ? all quite alike ? Aug 21 17:29:24 Seems so yes Aug 21 17:29:50 Elvispre has an Onyx Aug 21 17:29:58 Which is OnePlus Aug 21 17:30:17 yep Aug 21 17:30:29 but does it work ? Aug 21 17:41:34 Well it boots, not sure about telephony status Aug 21 17:41:48 elvispre: ^ Aug 21 17:53:08 I'd be interested in some logs, eventually :p Aug 21 17:54:57 elvispre: if you have some not-too-important SIM (like a free lycamobile one) for onyx, and if you can try to call the device from another phone, I'd be interested in: 1) output of "/system/bin/logcat -b radio" 2) /var/log/legacy-log 3) journalctl output Aug 21 17:59:56 Tofe: Well I guess I could mock Hammerhead 7.1 together as well maybe? Aug 21 18:01:22 maybe, I don't know how time-costly it would be Aug 21 18:01:41 on my side I'm trying to revive rosy a bit Aug 21 18:01:49 though it's basically also similar Aug 21 18:09:10 Tofe: Hi. What am I trying? :) Aug 21 18:09:44 My onyx does have a SIM in it, but LuneOS has not booted it (not so you can tell) for months. Aug 21 18:10:04 elvispre: that's plenty enough Aug 21 18:10:20 I'm fine with an old build of luneos too Aug 21 18:11:26 elvispre: what is you best status about telephony on onyx? Aug 21 18:11:30 your* Aug 21 18:12:02 on tissot we can't do calls at all, and our logs aren't very satisfying Aug 21 18:14:04 Well onyx stopped booting when (maybe even before) we started upgrading Yocto. Aug 21 18:14:15 So I'll have to find quite an old image. Aug 21 18:14:34 But it used to be my best behaved device (so long as you didn't want BT). Aug 21 18:15:24 elvispre: so... calling worked? Aug 21 18:16:19 Yes, so far as I remember. I'm pretty sure you could place a call in both directions. Getting an audio stream s a different question. Aug 21 18:16:28 s/s/i Aug 21 18:16:30 s/s/is Aug 21 18:18:27 ok, that's interesting, it means the issues we have on tissot are probably device-specific Aug 21 18:19:06 Why onyx? mako still works for calls so far as I know. Aug 21 18:19:16 onyx is using halium-7.1 Aug 21 18:19:27 afaik Aug 21 18:19:45 * Tofe double-checks... Aug 21 18:20:14 yes, it doeqs Aug 21 18:21:15 and I'm still struggling to get my rosy flashed Aug 21 18:22:14 Tofe: Best I can do right now is flash the onyx build I did yesterday and see if it works at all. Aug 21 18:22:34 I'll be surprised if it does. Aug 21 18:23:30 But it's worth a try. Aug 21 18:30:52 OK. I have tried that. The same result as always: The Android boot screen and no adb. To me, this is not really booting at all. Aug 21 18:33:31 I'll try the official images off /luneos-testing-warrior/images/onyx/ in a bit. Back later... Aug 21 18:34:15 It could be something in kernel Aug 21 18:34:30 We might be able to pull some logs from recovery Aug 21 19:00:59 or in recovery... somehow Aug 21 19:01:07 s/recovery/initrd/ Aug 21 19:23:05 Yeah, maybe. Since it stopped working, I have seen / found no evidence of activity at all. Aug 21 19:23:48 One day it was working, then a couple of weeks later I came back and it was just a desk ornament. Aug 21 19:43:50 Tofe: Some discussions in Halium channel now. Suggestion from their side: [21:35] (ubptgbot) Herrie have you tried running ofono manually with debugging enabled? (`sudo env OFONO_DEBUG=-d OFONO_RIL_DEVICE=ril OFONO_RIL_NUM_SIM_SLOTS=1 ofonod -d drivers/* -n -P provision,udev*,dun*,smart*,hfp_bluez5,stktest,sap`) **** ENDING LOGGING AT Thu Aug 22 02:59:56 2019