**** BEGIN LOGGING AT Thu May 23 03:00:36 2019 May 23 04:08:13 Morning! May 23 04:31:46 JaMa: I'm getting this when trying to build Hammerhead kernel after Mako kernel (both Android kernels, not the mainline one): https://bpaste.net/show/1a0b107ca4b6 May 23 04:47:33 I've seen this before May 23 04:48:28 Maybe we need to make https://github.com/webOS-ports/meta-webos-ports/blob/warrior/meta-luneos/recipes-android/android-tools/android-tools-conf_1.0.bbappend MACHINE_ARCH? May 23 04:49:30 Or change do_install_prepend into do_install_prepend_tenderloin ? **** BEGIN LOGGING AT Thu May 23 06:10:05 2019 May 23 06:14:49 JaMa: Seems that the same 2 patches work for Hammerhead ;) May 23 06:14:55 Next to test Tenderloin May 23 06:15:14 I have Hammerhead & Tenderloin with me today, so I can actually test it on device, not only build test :P May 23 06:58:35 Morning! May 23 07:03:19 Tofe: Morning! May 23 07:55:58 my fix for geoclue which worked in master seems to break it in warrior :/ http://jenkins.nas-admin.org/job/LuneOS/view/testing/job/luneos-testing_qemux86-64/118/consoleFull May 23 08:01:26 JaMa: I tested in Warrior and my code worked there May 23 08:01:48 I think we might need a backport from Zeus to Warrior of the meson bbclass to get it to work with your fix? May 23 08:14:20 or the gtkdoc might be fixed in the newer version May 23 08:30:05 Herrie: the upgrade wasn't enough, fixed in warrior now May 23 08:32:12 JaMa: OK. GitHub is laggy here.. At your end too? May 23 08:34:13 JaMa: Like I thought more or less. Anyway good it's sorted :D May 23 08:37:48 for the android-tools-conf issue, it's IMHO because do_install was re-executed with the patch already applied in the same $D May 23 08:38:11 JaMa: We only should apply the patch for tenderloin May 23 08:38:12 it's already MACHINE_ARCH and it's applied for all except tenderloin May 23 08:38:26 Not really sure how to fix this properly May 23 08:41:08 it's really short script and patch, it might be easier to just provide the patched script in meta-webos-ports May 23 08:41:17 and let tenderloin to use different one again May 23 08:42:59 I'll fix that May 23 08:47:51 JaMa: Thnx May 23 08:48:16 Got three 3.4 kernels to build with those 2 patches so far. Soon testing Hammerhead on device ;) May 23 08:49:33 Herrie: Tofe: https://github.com/webOS-ports/meta-webos-ports/commit/e768cd407f18868af5b22383171a345886893aea May 23 08:49:46 Herrie: GH doesn't seem to be slugish here May 23 08:50:06 with just 1 build running on my desktop it actually seems pretty good :) May 23 08:52:22 GH is back to normal it seems. Was a few minutes where I would get the unicorn and it was slow ;) May 23 08:55:45 OK Hammerhead boots with GCC 8 kernel ;) May 23 08:56:46 May 23 05:57:29 hammerhead kernel[885]: [ 0.000000] Linux version 3.4.0 (oe-user@oe-host) (gcc version 8.3.0 (GCC) ) #1 SMP PREEMPT Thu May 23 06:11:57 UTC 2019 May 23 08:57:15 FirstUse starts, so that's good enough I would say for confirming the kernel works ;) May 23 08:58:50 good ! May 23 08:58:58 JaMa: I'm fine with that May 23 09:04:32 ok, will also upstream it to meta-oe, so we can drop it later May 23 09:32:23 JaMa: I've sent them upstream to Halium & UB Ports for Hammerhead & Mako as well May 23 09:33:02 bshah: ^ https://github.com/Halium/android_kernel_lge_hammerhead/pull/4 May 23 09:33:17 JaMa: let me build test them for tissot, rosy & mido as well. May 23 09:33:23 And Onyx May 23 09:36:41 JaMa: Finally got to have a look into this one as well https://github.com/webOS-ports/meta-webos-ports/pull/307 Hope to have an update on testing for that shortly as well May 23 09:42:13 JaMa: Updated with my findings May 23 09:59:18 me too, but I still don't belive this is the right fix May 23 09:59:49 We can set the default value in to "hardware" in meta-luneos/classes/webos_machine_impl_dep.bbclass but all recipes should parse correctly without luneos.inc included May 23 10:00:11 so they should inherit webos_machine_impl_dep whenever they want to use WEBOS_MACHINE_IMPL_DEP variable May 23 10:00:35 WEBOS_TARGET_MACHINE_IMPL_qemuall should be moved to that bbclass as well May 23 10:01:02 s/WEBOS_MACHINE_IMPL_DEP/WEBOS_TARGET_MACHINE_IMPL/g May 23 10:18:53 I'll fix that in OSE as well May 23 10:20:06 JaMa: OK. Just not sure what would be the best approach. Setting it in each machine.conf seems to be a bit overdone as well? May 23 10:20:23 that's now that I was suggesting :) May 23 10:39:49 Herrie: https://github.com/webOS-ports/meta-webos-ports/commit/e1fec3a81c9e87c18f5324d3ab5571dbfb92eddd May 23 10:40:59 JaMa: OK May 23 10:41:09 We still need to update the .confs right? May 23 10:49:52 why? May 23 10:57:22 Well I need to see if there's something reading it, probably not ;) May 23 10:57:29 Would be luna-sysmgr normally May 23 11:20:18 Tofe: I think we should have a look at some point for a single kernel for mido, rosy & tissot. Seems that they use the same kernel in LineageOS, so we should be able to do as well, just providing different defconfigs for each of them ? May 23 13:21:58 that would be nice if it's possible yes May 23 13:47:05 JaMa: For the uriparser: See https://sourceforge.net/projects/uriparser/ for reason to switch from tarball to Git :) "PLEASE NOTE that we are in the process of moving to GitHub: https://github.com/uriparser/uriparser" May 23 13:47:21 Tofe: Seems the LineageOS is Tissot/Mido only. No mention of Rosy anywhere May 23 13:47:28 Anyway if we can bring 2 down to 1 is not bad either May 23 13:55:56 FWIW the 3.18 Xiaomi kernels build without any issues or patches on GCC 8.x it seems :) May 23 14:56:45 Herrie: there are also developer uploaded tarballs on github for uriparser May 23 14:57:03 Herrie: yes, the issue was only with the older kernels May 23 14:59:09 https://github.com/uriparser/uriparser/releases May 23 14:59:43 JaMa: But there we would have the MD5 issues at times no? May 23 14:59:49 Due to the regeneration ;) May 23 14:59:54 no, releases != archives May 23 15:01:30 JaMa: Ah Ok :) May 23 15:02:57 details in e.g. https://github.com/bmwcarit/meta-ros/issues/609 May 23 15:03:35 JaMa: Onyx is building now too. PR-ed all GCC8 patches upstream. As soon as they merge them we can switch GCC8 on I guess May 23 15:03:45 Tofe would need to test Pinephone still with GCC8 but should be fine May 23 15:48:22 Herrie: don't block anything just because of Pinephone :) it's mainline, so the risk is very small here May 23 15:51:35 Tofe: Well we're waiting for UBPorts, bshah and elvispre for their merges to kernel, so that will be a few days I guess May 23 15:51:45 Before we can bump the srcrevs May 23 16:07:38 ok May 23 16:21:00 Can confirm that Tenderloin also boots with GCC 8 built kernel :) May 23 16:49:05 good! :) May 23 16:56:29 It's very slow though May 23 17:06:15 But could be first boot May 23 17:06:25 Didn't check in much detail May 23 19:12:57 Herrie: new warnigs from ls2 security (after fixing the WEBOS_TARGET_MACHINE_IMPL): http://paste.ubuntu.com/p/Y9VcRZVNr9/ May 23 19:13:54 from meta-luneos/classes/webos_image.bbclass:ROOTFS_POSTPROCESS_COMMAND += '${@oe.utils.conditional("WEBOS_TARGET_MACHINE_IMPL", "hardware", "luna_service2_check_permissions ; ", "", d)}' May 23 19:20:00 JaMa: Well that's what it should do in general. It might actually help us in solving all those LS2 permission errors ;) May 23 19:22:23 that's why I'm sharing it May 23 19:22:31 And give us some more clues as to what needs doing in terms of those. I remember seeing a flag to say it are not production devices which would disable this LS2 security again ;) May 23 19:22:54 At least if my memory serves me right May 23 19:23:42 WEBOS_DISABLE_LS2_SECURITY May 23 19:23:49 but it's not set automagically May 23 19:25:37 Yes May 23 19:25:53 But it's good to tackle the LS2 in general May 23 19:26:20 It's one of the key things to do to get various missing/non-working bits working it seems May 23 19:27:21 Saw your patches to mailing list: Thanks for those :) May 23 19:51:20 wait to see my OSE patches :) May 23 19:53:02 few more warnings in tissot dev-image: http://paste.ubuntu.com/p/PtTVJYyJgW/ May 23 19:57:00 OSE patches? Hmmz you make me curious May 23 19:57:22 Will also be upstreamed? May 23 19:59:56 I mean the Yocto upgrade when it gets merged May 23 20:00:48 Ah ok May 23 20:01:14 Planned for this month's OSE release already? That would be quick :P **** ENDING LOGGING AT Fri May 24 03:00:35 2019