**** BEGIN LOGGING AT Wed Jan 08 02:59:58 2020 Jan 08 08:06:28 Morning! Jan 08 08:57:48 Morning! Jan 08 10:31:35 JaMa: Hi! Did the new mainboard arrive? Jan 08 10:53:23 Tofe: OK got the Lineage 15.1 kernel building now... Basically needed to update defconfig & cherry pick most of our patches from 5.1 ;) Jan 08 10:54:48 Here it is: Seems their pidns_install was broken, so picked your version of the function and it worked LOL ;) https://github.com/Herrie82/android_kernel_lge_hammerhead-1/commits/halium-8.1 Jan 08 10:54:54 Now let's build & test Jan 08 11:54:03 Herrie: yes, I have temporary board and cpu :) Jan 08 11:54:41 JaMa: Ah ok Jan 08 11:57:50 JaMa: For when you have a few minutes: https://github.com/shr-distribution/meta-smartphone/pull/110 Jan 08 12:00:12 merged before I forget (again) Jan 08 12:07:44 Thnx Jan 08 12:17:24 Herrie: ah, good! I hope that this kernel will work out Jan 08 12:35:45 Tofe: Well it's acting up a bit, not really sure why.... Jan 08 12:36:17 I forgot CONFIG_MODULES=y in defconfig which caused memnotify not to work, added that now but then the backlight driver is acting up Jan 08 12:38:20 I.e. https://github.com/Herrie82/android_kernel_lge_hammerhead-1/blob/92e22d6dc1dad374fc815e572790e57a9e96912c/drivers/video/backlight/lm3630_bl.c#L616 saying lge_get_boot_mode is not defined. Jan 08 12:40:19 Which it should get from https://github.com/Herrie82/android_kernel_lge_hammerhead-1/blob/92e22d6dc1dad374fc815e572790e57a9e96912c/arch/arm/mach-msm/include/mach/board_lge.h#L48 and should be enabled when I set the CONFIG_MACH_LGE in defconfig.... Jan 08 12:58:54 So trying to work through all those now Jan 08 13:02:32 These 2 commits give headaches, so working around these now: https://github.com/Herrie82/android_kernel_lge_hammerhead-1/commit/0b4a555ee3f6d5234f0eb01b48182a06352517db#diff-16a45f9920811cb1a991025d171edcf5 and https://github.com/Herrie82/android_kernel_lge_hammerhead-1/commit/42fe48d30c41874e679b6552798de94797c5f402#diff-2f28728c958b4e2eae9b089a1668b46a Jan 08 13:05:15 OK have something building now it seems :S Jan 08 13:08:47 Let's flash ;) Jan 08 13:26:50 Good luck ! Jan 08 13:42:34 Boot loop... Jan 08 13:45:49 last_kmsg: https://paste.ubuntu.com/p/sxYSH2xWfc/ Jan 08 13:45:59 "Unable to handle kernel NULL pointer dereference at virtual address 00000000" Jan 08 13:46:06 That's not very helpful Jan 08 13:52:30 try again, maybe the kmsg will be readable next time Jan 08 13:52:43 there should be a stack after the panic Jan 08 13:55:45 Ah seems proc_ns_follow_link Jan 08 13:55:48 We've seen that before Jan 08 13:55:57 Maybe I forgot a commit somewhere Jan 08 13:58:43 yes, it's another namespace-related commit Jan 08 14:08:13 Seems there are 3 commits added to /fs/proc/namespaces.c. I will try to just go back to the namespaces.c we were using before and that worked Jan 08 14:11:11 ok Jan 08 14:30:32 Tofe: I should be able to just use the zImage right? Jan 08 14:36:08 hammerhead uses zImage-dtb, iirc Jan 08 14:37:17 also, you'll have to wrap it in a fastboot image Jan 08 14:37:36 but virtual/kernel recipe will do that for you Jan 08 14:40:58 Visually not a whole lot difference but no more null pointer: https://paste.ubuntu.com/p/RmyPyQC64Z/ Jan 08 14:52:56 no adb ? Jan 08 14:53:22 Morning Jan 08 14:54:06 ka6sox: Morning! Jan 08 14:55:26 When is the next time the AutoBuilder is needed? I'm working on some changes to share Sources/Log Locations/place to store SSTATES that won't constantly require us to increase space. Jan 08 14:56:48 There no big need the few coming days, I'd say Jan 08 14:57:05 +is Jan 08 14:57:35 I can completely live without jenkins for a week or so :) Jan 08 15:02:30 ka6sox: We're good without Jenkins for a week or so like Tofe said. It's more JaMa and Khem that you would need to worry about Jan 08 15:02:35 Tofe: Yeah no ADB Jan 08 15:03:55 I've been working with khem to move the OE bits already Jan 08 15:04:06 trying to connect with JaMa... Jan 08 15:04:17 I have a 10TB partition for the builders to use setup. Jan 08 15:07:31 Herrie: then you can try this: go to recovery, remove the "/data/luneos-data/var/log" symlink and create a real "var/log/journal" directory. Then boot once, and go back to recovery. Pull the whole journal, and read it on your machine Jan 08 15:12:13 Tofe: That doesn't seem to work Jan 08 15:17:55 Seems I might be missing a few more defconfig values. Checking now Jan 08 15:32:33 Quite a few it seems... Seems I overlooked initial commit from UBPorts for the defconfig: https://github.com/Herrie82/android_kernel_lge_hammerhead-1/commit/41c4d4803f96920eff9567b8bea4e0868a994498 Jan 08 15:33:48 ka6sox: I'm building webOS OSE 2.2 release right now (qemux86 build already done, rpi3 running now, rpi4 in queue), but I don't mind if they are killed and finished couple days later Jan 08 15:34:28 I'll let them finish. Jan 08 15:34:45 I'm working on a change to OE's builds right now... Jan 08 15:34:52 we are moving the build logs for them. Jan 08 15:35:09 where are you storing the buildlogs for LuneOS and WOSE? Jan 08 15:35:23 (Publically they will still be the same URL) Jan 08 15:39:35 we don't store buildlogs for them Jan 08 15:39:54 other than the jenkins jobs logs on jenkins itself Jan 08 15:43:33 ah, that's why I can't find them...okay Jan 08 16:05:19 Tofe: OK I now see Android init etc in the logs, still no ADB and lots of audit failures, but that could be simply another defconfig flag I need to check Jan 08 16:33:13 This is the log: https://paste.ubuntu.com/p/7TNz2r9Vzm/ Jan 08 17:19:30 Tofe: Well seems my Titan is ready so can use Mido for 8.1 testing with newer kernel Jan 08 17:19:38 To exclude 3.4 being the issue Jan 08 17:46:31 Herrie: ok :) Jan 08 17:46:38 I'll have a look at the logs Jan 08 17:47:37 Herrie: as soon as you see "/usr/libexec/droid-hybris", you can deduce there's a Mer modification that shouldn't have occured Jan 08 17:48:04 We don't use that path at all for Halium or LuneOS Jan 08 17:59:14 Tofe: Yeah I know Jan 08 17:59:24 I didn't manage to get these 3 services in image yet Jan 08 17:59:34 I guess we'd need them? Or can I disable them completely? Jan 08 17:59:39 Seems needed for media stuff? Jan 08 17:59:48 They're simply in init.rc Jan 08 18:00:20 Ehm scrap that Jan 08 18:00:23 In another easy location Jan 08 18:00:25 I've seen them Jan 08 18:00:44 But those shouldn't be critical for now Jan 08 18:00:58 The tzdata binary is more interesting for now I'd say Jan 08 18:07:14 well I'm not sure how critical they are for the init of android Jan 08 18:09:05 looks like everything get killed by audit ? **** ENDING LOGGING AT Thu Jan 09 02:59:58 2020