**** BEGIN LOGGING AT Sun Apr 25 02:59:56 2021 Apr 25 08:02:45 Morning! Apr 25 09:12:02 Morning! Apr 25 09:12:49 Herrie: I lost my znc history again, looks like this isn't my best week-end for a continuous irc log :p Apr 25 09:13:04 Tofe: Ah :s Apr 25 09:13:18 Well nothing much happened Apr 25 09:13:26 well, I'll just fetch and read the last logs on nslu2 Apr 25 09:13:42 I see that, good Apr 25 09:13:54 Was trying to help someone in #halium with Razer Phone 1 port but a bit stuck with kernel at our side Apr 25 09:14:10 Will paste some log in a bit, maybe you have some thoughts Apr 25 09:26:46 It's still this "python" issue ? Apr 25 09:56:53 Tofe: No this is a different target ;) https://paste.ubuntu.com/p/GR3HsKKnb5/ Apr 25 09:56:59 But yeah still python for Mido Apr 25 10:01:23 ah, just don't activate the RNDIS gadget, if adb is available it should be enough Apr 25 10:02:58 Tofe: Well I try, but it seems to hard force it to Y even when I use the script to force it to N ;) Apr 25 10:03:36 or try that : https://patchwork.kernel.org/project/linux-kbuild/patch/1510818396-17443-1-git-send-email-yamada.masahiro@socionext.com/ Apr 25 10:03:58 I tried that and got another error ;) Apr 25 10:04:03 damn :p Apr 25 11:16:24 Tofe: I managed: https://github.com/Herrie82/android_kernel_razer_msm8998/commit/3e3285b2cb0b5abfdf3f217b3f2a4ce53e53c4a7 Apr 25 11:16:35 Found something from mal in logs from SFOS :P Apr 25 11:23:58 Herrie: good, well done Apr 25 13:46:46 Tofe: OK got an image for this Razer phone, doesn't look too bad as a first stab... You have any thoughts on what might be the issue? https://paste.ubuntu.com/p/gc2z7DrS9W/ Apr 25 13:47:08 I see [ 4.074466] init: [libfs_mgr]check_fs(): mount(/dev/block/by-name/system_a,/mnt/scratch,ext4)=-1: Permission denied which I guess is fatal somehow Apr 25 13:47:29 Though later it seems to mount successfully? [ 4.096044] init: [libfs_mgr]__mount(source=/dev/block/by-name/system_a,target=/mnt/scratch,type=ext4)=0: Success Apr 25 16:06:53 Tofe: Different mido 4.9 kernel looks better, but doesn't display UI Apr 25 16:06:55 https://paste.ubuntu.com/p/zTFTxG73Gv/ Apr 25 16:07:01 Seems android container is running Apr 25 16:07:30 No logcat though :( https://paste.ubuntu.com/p/hRgqPF2m69/ Apr 25 16:07:46 Thought? Seems LiveDisplay might be an issue? Apr 25 16:35:54 Tofe: Ah seems I forgot to add: https://github.com/Herrie82/android_kernel_xiaomi_msm8953-1/commit/3b511b24c3ab209b1ba42bc3443be651e70d22fd Apr 25 16:35:58 Let me retry :p Apr 25 16:45:57 Nope, seems LXC container starts, but lxc-checkconfig has some missing values, so let me fix those first Apr 25 17:17:05 Herrie: back home Apr 25 17:18:33 Herrie: do you have lxc's log ? Apr 25 17:25:42 Tofe: How to get lxc log again? Apr 25 17:25:55 Ah found it Apr 25 17:25:58 /var/log/lxc/android.log I think Apr 25 17:27:22 Herrie: and if you see LuneOS/Halium in dmesg, then you don't need the initramfs hack Apr 25 17:28:58 you might need to add androidboot.keymaster=1 in the cmdline later on Apr 25 17:30:48 Tofe: https://paste.ubuntu.com/p/B38c434PyX/ Apr 25 17:31:14 What does androidboot.keymaster=1 do? Apr 25 17:32:06 iirc it helps with choosing the security profile Apr 25 17:32:19 I thnk I had an issue with that on rosy Apr 25 17:34:26 Herrie: "init: Couldn't open /sys/fs/selinux/null: No such file or directory" ... I've seen that before too Apr 25 17:35:50 yes, was on tissot back in Feb 11th Apr 25 17:37:07 OK I'll add the selinux permissive nad keymaster Apr 25 17:37:12 THose are in Tissot but not in Mido Apr 25 17:37:26 For cmdline Apr 25 17:37:57 Also, make sure it's using the correct bootloader version; on tissot I had several issues related to the fact that my bootloader hadn't been upgraded Apr 25 17:38:26 i.e. flash the latest stock rom <= 9.0 before proceeding Apr 25 17:38:38 Well I had 3.18 kernel booting before Apr 25 17:39:05 ah, wait, I'm lost here: we're talking about mido? Apr 25 17:39:15 Yes Apr 25 17:39:17 ok Apr 25 17:39:20 I had 3.18 booting on Mido, trying 4.9 now Apr 25 17:39:25 ok Apr 25 17:39:34 Found another 4.9 compared to the one that had issues yesterday Apr 25 17:39:38 Gave up on 4.19 for now Apr 25 17:39:45 This one seems to behave better at least Apr 25 17:39:54 let's first solve 4.9 yes Apr 25 17:41:14 Then can also build it on Jenkins and push proper manifest to halium-devices Apr 25 17:45:44 Still the same it seems: https://paste.ubuntu.com/p/nS98GKDpTy/ Apr 25 17:46:12 https://paste.ubuntu.com/p/cMkcGXVnfs/ Apr 25 18:02:38 Herrie: hard to say... Apr 25 18:03:24 Any thoughts as to what to debug? Apr 25 18:03:31 Kernel in general seems fine Apr 25 18:03:41 Seems more something on Android side? Apr 25 18:05:13 yes, it's closer to android Apr 25 18:05:52 I'd say the selinux thing is a hint, but not sure how to interpret it Apr 25 18:07:08 Will compare tissot and mido defconfigs Apr 25 18:07:14 Maybe some clue there Apr 25 18:08:13 but on tissot, is it a 4.9 kernel? Apr 25 18:08:16 With regards to audit, selinux etc Apr 25 18:08:30 Yeah we have 4.9 for Tissot now as well I think? Apr 25 18:09:32 I don't remember tbh :) Apr 25 18:09:50 you gave me also this hint, in February: https://irclogs.sailfishos.org/logs/%23sailfishos-porters/%23sailfishos-porters.2020-06-07.log.html Apr 25 18:10:44 https://github.com/Tofee/tissot/commit/95552bdc81cecda0accf85fb4dfda097d6570859 ! Apr 25 18:11:25 so you need AUDIT=y Apr 25 18:11:27 OK that looks like what I'm having Apr 25 18:12:15 https://github.com/Herrie82/android_kernel_xiaomi_msm8953-1/blob/27c0576ad170b562f99c60fccdd905cd94d54edc/arch/arm64/configs/mido_defconfig Apr 25 18:12:19 Unset in line 4 Apr 25 18:12:48 yep, that's the mistake Apr 25 18:15:18 Building updated kernel Apr 25 18:15:29 Curious to see how it'll behave now Apr 25 18:16:03 Yeah me too Apr 25 18:16:17 Logs looked pretty decent for the rest Apr 25 18:17:14 and we now mido can behave quite well Apr 25 18:22:51 device trees are pretty decent in general Apr 25 18:24:52 SELinux error is gone, but still lxc restarting all the time: https://paste.ubuntu.com/p/scks2zkR7X/ Apr 25 18:25:17 "init: Could not open uevent socket" hmm... Apr 25 18:26:39 Could be this fwhelper stuff? Apr 25 18:26:51 But I think I have it in defconfig Apr 25 18:26:57 Anyway first some food Apr 25 18:27:48 no, I think it's something else, maybe a conflict between udev and ueventd? I remeber pointing that risk out in the past Apr 25 18:31:52 I'm pretty sure I proposed a patch for this somewhere, to patch the socket address used by ueventd Apr 25 18:35:35 it was about patching this kind of line: https://android.googlesource.com/platform/system/core/+/android-4.4_r1/libcutils/uevent.c#104 Apr 25 18:42:53 halium.20210106.txt:Jan 05 12:23:27 there is an error of "init: Could not open uevent socket" if I don't use the getpid() patch, https://paste.opensuse.org/24828232 Apr 25 18:43:00 halium.20210103.txt:Jan 02 14:38:52 NotKit: I didn't push the patch yet, but roughly it's just changing https://github.com/Halium/android_system_core/blob/halium-8.1/libcutils/uevent.c#L112 to something else, like "0x8000000 + getpid()" Apr 25 18:44:52 Seems you never PR-ed it to hybris-patches? https://github.com/Halium/android_system_core/pull/10/files Apr 25 18:45:08 Our Tissot would fail now as well probably with Halium build from Jenkins ;) Apr 25 18:45:37 maybe :/ Apr 25 18:46:11 but in the end, I think I got it to work even without the patch, but I don't recall why it worked Apr 25 18:46:42 Herrie: you can always try to patch manually Apr 25 18:46:53 and we can push it to halium Apr 25 19:54:09 Tofe; Let me try that at least Apr 25 21:39:36 Tofe: Hmmz LXC behaves better now, but it seems it's starting ACTUAL LOS now instead of our Android container? Apr 25 21:39:43 I get LOS bootlogo Apr 25 21:39:49 Where the circle is moving left to right Apr 25 21:40:00 Ehm right to left Apr 25 21:40:34 This is my lxc log: https://paste.ubuntu.com/p/Yv4WMB2fNQ/ Apr 25 21:42:48 console-ramoops: https://paste.ubuntu.com/p/yX57HfvVPP/ **** ENDING LOGGING AT Mon Apr 26 03:00:18 2021