**** BEGIN LOGGING AT Mon Jun 01 02:59:57 2020 Jun 01 03:37:00 This part of error messages in journal/system-log are being looped consistently.... https://pastebin.com/5bp00cSd. Seems screen could be related to this. Jun 01 03:38:17 init: couldn't write 1962 to /dev/cpuset/system-background/tasks: No space left on device Jun 01 03:39:53 imscmservice, qseeproxydaemon and time_daemon. These are trying to start and being killed. The beep i hear each one or two second might be.... Jun 01 03:40:51 when manually viewed, time on top bar is also not consistent. suddenly it jumps two minutes :) Jun 01 05:08:08 https://www.dropbox.com/s/6r745i7spuhewvv/err.txt?dl=0 Jun 01 05:09:04 above file is for this command (filtered by error, last boot log): journalctl -b 0 -p err --no-pager --file=systemlune.journal > err.txt Jun 01 05:10:48 When filtered by critical using -crit parameter with journalctl, this is output: https://pastebin.com/wAZpZxcg Jun 01 05:15:43 Morning! Jun 01 05:16:23 Don't worry about init: couldn't write 1962 to /dev/cpuset/system-background/tasks: No space left on device it seems we have this on many devices in different OS-es Jun 01 05:16:33 It's not a real thing to be worried about Jun 01 05:17:00 Does ADB work now or still not? Jun 01 05:18:52 It might be good to check logcat once ADB is working to see how things are on Android side. Jun 01 05:29:01 Let me see if I can put together a kernel with less spam in the logs Jun 01 05:33:46 BT40: ^ Jun 01 05:37:15 Herrie, wait a moment Jun 01 05:38:44 It seems we might want to tweak Android side in Halium build as well at some point, nothing too critical but it's merely to prevent Android's init from starting things and throwing errors (they don't really have much effect on LuneOS itself though). For Mido I have the following changes: https://github.com/Herrie82/android_device_xiaomi_mido/commits/pgz-14.1 Jun 01 05:39:06 Basically the last 6 commits there (from piggz and herrie82) Jun 01 05:41:08 https://databasefaq.com/index.php/answer/102767/i2c-probing-how-to-fix-error-on-clk-getcore-clk-during-probing-driver-i2c-msm-v2 Jun 01 05:41:35 can we fix above error also of i2c as suggested in above thread? Jun 01 05:43:41 There is also --restart-- written after 3-4 lines in main log. I am sure this isn't complete reboot. Is this related to screen blink? Jun 01 05:45:48 these are reboot messages i am talking about : https://pastebin.com/vmqZaB8S Jun 01 05:46:24 Not sure Jun 01 05:46:32 Once ADB is working it will be a lot easier to debug Jun 01 05:48:02 even to power on, i have to press power key for 10-12 seconds :) Jun 01 05:48:15 will give one more try to adb Jun 01 05:48:20 if it detects Jun 01 05:49:04 no detection: Jun 01 05:49:20 adb list of devices show nothing as output Jun 01 05:50:22 try "adb kill-server" and retry "adb devices" Jun 01 05:50:25 its not even charging Jun 01 05:50:27 Sometimes adb is funny Jun 01 05:50:47 Could also be you have a weird USB cable that's acting up a bit Jun 01 05:51:11 i have another one. will give it a try Jun 01 05:51:15 give me a minute Jun 01 05:52:38 same result, no devices detected. noteven battery is charging when connected cable Jun 01 05:54:30 checked one more thing, when connected to wall charger even with new original xiaomi cable, there is no power indicationg charging. Jun 01 05:54:54 but everything works fine in twrp recovery mode. it will charge, adb will work Jun 01 05:59:30 BT40: Well that could be minor config on LuneOS side for battery Jun 01 05:59:56 I'm making a new kernel now with the i2c/clk change, some less logging & added defconfig for USB Jun 01 05:59:59 Maybe that iwll solve things Jun 01 06:00:59 is it possible that usb is not configured properly and cable not detected? and this is why neither charging work neither adb? Jun 01 06:02:16 Could be but owuld be weird, since we didn't change anything really in that regard v.s. LineageOS Jun 01 06:13:17 BT40: New kernel to try: Image.gz-dtb--3.18.31+gitr2+b8b1ecb00d-r0.2-land-0-0.fastboot Jun 01 06:13:20 On your Dropbox Jun 01 06:13:32 ok Jun 01 06:18:22 now charging...indicator in top bar appears Jun 01 06:18:55 adb devices not reflecting Jun 01 06:19:46 similar screen flickering Jun 01 06:28:08 I'm a bit at a loss for why adb doesn't work, Tofe might have some more ideas Jun 01 06:28:40 void DisplayManager::displayOff() Jun 01 06:28:45 Can you check logs again to see if the i2c is gone now and the batter power spam? Jun 01 06:28:56 virtual void DisplayOff::enter(DisplayState, DisplayEvent, sptr): entering state Jun 01 06:29:08 these appear each second multiple time Jun 01 06:29:13 BT40: This is something on LuneOS side for sure Jun 01 06:29:19 each minute sorry Jun 01 06:29:29 this is turning screen on off? Jun 01 06:29:39 That's "normal" I guess, I would say so Jun 01 06:30:00 If sensors don't work, input doesn't work it will turn the screen off Jun 01 06:36:31 i2c seems resolved. i am not sure. although it doesnot appear as per time stamps of new boot, but it is present in file even though i passed on parameters to display current boot only Jun 01 06:36:49 so i am confused Jun 01 06:37:01 will share new log file Jun 01 06:37:34 i was using this command journalctl -b 0 --no-pager --file=systemlune2.journal > my_log4.txt Jun 01 06:38:08 will share output of full command without b parameter, same as tofe advised. Jun 01 06:45:38 https://www.dropbox.com/s/96q9g6vbk8bf36y/my_log6.txt?dl=0 Jun 01 06:46:05 above is output for journalctl --no-pager --file=systemlune2.journal > my_log6.txt Jun 01 06:47:48 https://www.dropbox.com/s/1vfvazteuuzwpvk/my_log3-err.txt?dl=0 This is output with error filter : Command>> journalctl -b 0 -err --no-pager --file=systemlune2.journal > my_log3-err.txt Jun 01 07:06:04 BT40: I don't see a whole lot unusual in there Jun 01 07:07:30 There is some bits on Android side that would need a little tweaking though Jun 01 07:08:38 ok Jun 01 07:08:46 meems positive news Jun 01 07:08:53 seems* Jun 01 07:08:55 libtinyxml missing rings a bell for me Jun 01 07:09:18 Herrie, you reffered error filtered or normal? Jun 01 07:09:26 The normal one Jun 01 07:09:30 two links i had provided Jun 01 07:09:33 okay Jun 01 07:09:44 Only thing that I see there is unusual is the "tainted" messages for some of LuneOS bits Jun 01 07:09:58 But I suspect this has to do with AppArmor/Audit somehow Jun 01 07:10:08 have u seen once the error filtered? is it okay? Jun 01 07:10:14 Tofe probably knows more Jun 01 07:10:28 I'm not that great with low level kernel stuff ;) Jun 01 07:10:39 will wait for him to arrive Jun 01 07:12:01 also the network towers indicator is there, but i cant make or receive calls. when some other person dials my number, it says number is switched off. so i think network doesn't work Jun 01 07:12:34 BT40: Well with the current state of your build I'm not surprised, some things need tweaking still Jun 01 07:12:50 But getting screen to work so early on is good already Jun 01 07:12:56 It took me weeks to get to that point ;) Jun 01 07:13:21 Also because when I did Mido it was the first device with Android 7.1, before we only had Android 5.1 targets Jun 01 07:13:25 So it was tricky Jun 01 07:19:24 can i try the same image in virtual device emulator? Jun 01 07:19:56 will virtual device emulator also give errors if configured as my device? Jun 01 07:21:09 most chances are that not, physical device is only option. Still asking if any chances, i'd be able to use my phone until then with android and receive phones and messages. Otherwise i have to arrange another phone Jun 01 07:50:43 BT40: Our VirtualBox image is x86-64 CPU architecture and won't show anything specific relating to your device Jun 01 07:50:58 So it won't help much Jun 01 07:51:12 Any development is always better on a not daily driver phone Jun 01 07:57:19 okay, not an issue Jun 01 08:08:34 I'll try to get my Halium setup later today/tomorrow so I can rebuild the Android bits with some fixes Jun 01 08:12:02 Seems my replacement memory modules might arrive tomorrow Jun 01 09:05:34 i think we may try also CONFIG_USB_F_FS=y Jun 01 09:13:55 nizovn: Can add that one yes Jun 01 09:14:37 It could also be something on Android init side that's the issue I think? Jun 01 09:15:49 For mido I saw this: https://github.com/Herrie82/android_device_xiaomi_mido/commit/405ac1ca494cd21e30b370bdaae9731393bdb15e Jun 01 09:16:01 I recall adb works fine for mido though Jun 01 09:16:14 We don't seem to have this disabled for Tissot though Jun 01 09:18:10 we should have adb without working android container, but true, there is android_usb error in android init as well Jun 01 09:18:13 init: write_file: Unable to open '/sys/class/android_usb/android0/f_ffs/aliases': No such file or directory Jun 01 09:53:29 BT40: New kernel with suggested change from nizovn: Image.gz-dtb--3.18.31+gitr3+5390567904-r0.2-land-0-0.fastboot Jun 01 09:53:35 On your dropbox Jun 01 09:55:39 great Jun 01 09:55:50 willimplement in 10 minutes Jun 01 10:07:29 flashed... rebooting Jun 01 10:08:42 adb not working, although charging icon is there Jun 01 10:10:50 yes we progressed :) Jun 01 10:11:18 Now screen does not blink, i mean it doesnot automatically on/off. it is in my control now :) Jun 01 10:11:43 apps also crash less Jun 01 10:11:56 autorotate works Jun 01 10:13:09 BT40: weird such a small defconfig flag does all that Jun 01 10:13:17 hahaha Jun 01 10:13:18 Do you have sound also? Jun 01 10:13:28 what did you changed, can u share Jun 01 10:13:32 yes Jun 01 10:13:50 volume buttons are working, sound is coming Jun 01 10:22:25 It's on my github Jun 01 10:22:55 okay Jun 01 10:23:19 https://github.com/Herrie82/android_kernel_xiaomi_msm8937/commits/halium-7.1 Jun 01 10:26:11 CONFIG_SDCARD_FS=n.................this seems strange Jun 01 10:26:30 wouldn't this disable memory card? Jun 01 10:34:35 screen turn off started again, automatically turn on, with unnecessary beeps. Rebooted phone to verify this. yes again it has started. Jun 01 10:35:46 power button also not working on single click (screen doesn't get off). neither boot menu comes. 10 secont long press restarts the phone Jun 01 10:41:24 power cut, will join later on. Jun 01 11:55:02 * BT40 is ready to go ahead Jun 01 12:20:03 BT40: The SDCard FS I disabled for now because kernel would fail to compile with it enabled and I didn't feel like debugging that Jun 01 12:20:13 It's not that important for now Jun 01 12:20:22 okay Jun 01 12:21:08 Power button not working is another issue which should be very easy to solve once ADB is working Jun 01 12:23:28 Morning! Jun 01 12:25:42 Tofe: Morning! Jun 01 12:25:54 We got a bit further with different defconfig options, but still no ADB it seems Jun 01 12:26:01 Sound, rotation work Jun 01 12:26:05 you enabled FUNCTIONFS ? Jun 01 12:26:08 Also charging now works Jun 01 12:26:42 https://github.com/Herrie82/android_kernel_xiaomi_msm8937/commits/halium-7.1 Jun 01 12:26:53 Last 5 commits from today Jun 01 12:27:18 The power one I took from another Xiaomi target, instead of your suggested change. I guess it would do something similar Jun 01 12:28:01 looks good Jun 01 12:28:52 I noticed that Android's init for Mido & Tissot are different.. I.e. on Mido I seem to have more disabled vs Tissot Jun 01 12:29:20 This is what I have on Mido: https://github.com/Herrie82/android_device_xiaomi_mido/commits/pgz-14.1 Jun 01 12:29:44 The changes from piggz seem to disable also qcom usb init Jun 01 12:31:40 we could replicate that on land, and maybe even on tissot Jun 01 12:32:03 See https://github.com/Herrie82/android_device_xiaomi_mido/commit/405ac1ca494cd21e30b370bdaae9731393bdb15e Jun 01 12:32:19 Not sure it's needed or not Jun 01 12:32:29 I remember logcat being pretty clean on my Mido Jun 01 12:32:33 charging it now to verify Jun 01 12:32:33 I'm not sure either Jun 01 12:32:56 but if mido works well, that would mean it's not needed, or even could conflict with luneos Jun 01 12:33:14 I haven't checked logcat on Tissot for a while Jun 01 12:33:33 I think I only had 1 constant nag in the logs which was for netd or netlistener or something Jun 01 12:35:01 Maybe you can check logcat on Tissot? Jun 01 12:35:19 See what's there to see if it could be improved as well... Because I remember I cleaned it for Mido pretty well Jun 01 12:35:48 Herrie: you've maybe also noticed that "android-gadget-setup[972]: /usr/bin/android-gadget-setup: line 27: /sys/class/android_usb/android0/iSerial: No such file or directory" as the *first* error from that script means that the "grep -q functionfs /proc/filesystems" just before has probably failed Jun 01 12:35:50 Maybe this way we'll have some common things to disable in init for Xiaomi targets Jun 01 12:36:16 Herrie: it's a busy day today, I'll see what I can do Jun 01 12:37:10 CONFIG_PROC_FS=y seems OK Jun 01 12:40:30 Herrie: yes, but what about functionfs Jun 01 12:41:45 looking into kernel sources, it seems just CONFIG_USB_G_ANDROID=y should enable all requirements, so if something changes with new configs it's strange Jun 01 12:41:48 Herrie: can you check, in the build directory of the land kernel, what is the content of .config ? sometimes an enabled option is ignored because it depends on another disabled option, and in that case it's quite silent about it Jun 01 12:46:37 Tofe: Where to see this exactly? I don't seem to have .config Jun 01 12:48:56 Ah wait Jun 01 12:49:06 FTP client doesn't show the hidden files... It's there Jun 01 12:49:25 https://paste.ubuntu.com/p/47rVcQmxCk/ Jun 01 12:50:12 This is logcat on Mido, not so clean actually after a while: https://paste.ubuntu.com/p/Vj2dH3VBbR/ Jun 01 13:31:30 For you guys this might be a routine work. But the fact is, you guys are putting so much efforts just to make luneos run on my device.... This is hurting me. Yes we are connected and share enthusiasm for open source. but you people can use this time somewhere else which is more productive either to your family or to luneOS/qt5 community. Instead i feel i am wasting your time...ethically. Jun 01 13:31:45 For me, this experience is learning. I will still gain. But for you, since u are already experienced, nothing much to gain..only donating efforts. I dont have words to thank. Jun 01 13:32:14 Can i do something to contribute? Maybe not on kernel side, but other stuff which may somewhat boost luneOS community. That would be relaxing for me. Jun 01 13:33:02 BT40: We're still learning every day too ;) Jun 01 13:33:04 Do we have a tool that can scan all the parameters in kernel config file, then compare the conflicts, and give the recommendations? It will save some time of diagnose. For example, a script which is optimised for luneos, scans all entries in kernel of a new port, gives recommendation so that chances of messing are reduced. Jun 01 13:33:30 We used the Mer one mainly Jun 01 13:33:45 even 20% reduction in messing will save a lot of time Jun 01 13:33:53 Those work for us: https://github.com/mer-hybris/mer-kernel-check Jun 01 13:34:07 ok, so already work done :) Jun 01 13:34:10 Help with documentation is always welcome, app development etc ;) Jun 01 13:34:21 and tool to scan the log files? Jun 01 13:34:40 Log files are different on each target Jun 01 13:34:44 No real generic approach there Jun 01 13:35:02 What we basically check is journalctl and logcat on Android side Jun 01 13:35:09 Logcat should normally be fairly clean Jun 01 13:36:26 BT40: it's some work for a couple of days, but once it's solved we rarely have regressions Jun 01 13:36:42 app development side, any specific priorities/targets for luneos? I mean instead off developing games etc, the work on essentials would be more practical. so i will focus on that area as priority Jun 01 13:36:54 But Android side is mainly for Halium. But since Halium makes our life easier, we happily contribute things back there Jun 01 13:37:11 BT40: I would say Messaging app Jun 01 13:37:20 It's really basic at the moment Jun 01 13:37:22 okay Jun 01 13:37:40 There is a working one in HP's SDK, but they never open sourced it..... Jun 01 13:37:46 any documentation or running project on git? Jun 01 13:37:54 So it's status is a bit grey area ;) Jun 01 13:38:11 Since you can download SDK, code is in there but not formally open sourced Jun 01 13:38:38 So we cannot distribute it, but you can install it on LuneOS with minor changes and it will work and give full IM/SMS/MMS capabilities ;) Jun 01 13:38:41 can we integrate some other open source into our luneos instead of writing from scratch? Jun 01 13:38:58 making it standard one? Jun 01 13:39:01 BT40: We already do that a lot ;) Jun 01 13:39:14 Share between Halium, UBPorts, Mer etc ;) Jun 01 13:39:36 We're not reinventing the wheel unless necessary ;0 Jun 01 13:39:38 ;) Jun 01 13:39:52 hmm... Jun 01 14:01:37 Connect your device with a LuneOS developer mode build on it to your development machine with a USB..... This line is written in documentation page. Jun 01 14:02:06 Is our build in developer mode? is debugging enabled by default? Jun 01 14:02:36 even in android we dont get adb devices until we allow permissions by unlocking usb debugging. Jun 01 14:56:29 Yes enabled by default Jun 01 15:54:49 almost all builds with ACG and without those unneccessary bbclasses done of fileserver, unless you have something else for meta-webos-ports I'll trigger unstable/dunfell build Jun 01 17:24:05 JaMa: please go ahead Jun 01 17:30:33 need to fix ls2 first again Jun 01 17:31:28 Herrie: I guess this wasn't intentional part of ACG, right? https://github.com/webOS-ports/meta-webos-ports/commit/47b6d74a1bda1dd39b8db2eb22dc7a08ae01a007#diff-b5bb09bcdff3667001a723055d0bb360 Jun 01 17:31:48 it just reverts my earlier SRCREV bump from https://github.com/webOS-ports/meta-webos-ports/commit/272910dc87c093fe85b11772840ac5bcc5114802 Jun 01 17:32:40 Somehow must have missed that one... Jun 01 17:32:40 Not intentional no Jun 01 17:36:35 the same with mediaindexer it seems https://github.com/webOS-ports/meta-webos-ports/commit/47b6d74a1bda1dd39b8db2eb22dc7a08ae01a007#diff-9aa4ba7690d876817648eb27c7d7b936 just to different SRCREV than mine https://github.com/webOS-ports/meta-webos-ports/commit/7517d319d6096aefa6f21167ff9b97a721eb138a Jun 01 17:38:22 will re-test with gatesgarth build then another big rebuild Jun 01 17:40:56 libqofono failed as well, but that looks like random issue which happens from time to time already (independently on ACG/dunfell/gatesgarth) Jun 01 17:40:59 qdbusxml2cpp: Got unknown type `a{sv}' processing '../../git/src/dbus/ofono_manager.xml' Jun 01 17:41:02 You should add to the XML description Jun 01 17:41:18 qdbusxml2cpp: Got unknown type `a{sv}' processing '../../git/src/dbus/ofono_connection_manager.xml' Jun 01 17:41:21 You should add to the XML description Jun 01 17:41:24 qdbusxml2cpp: Got unknown type `a(oa{sv})' processing '../../git/src/dbus/ofono_network_registration.xml' Jun 01 17:41:27 You should add to the XML description Jun 01 17:47:45 Load average: 217.79 167.03 99.49 with 3 qtwebengine builds at the same time (and few smaller ones as well) and still perfectly usable system with enough ram :) Jun 01 17:51:02 just tail -f of all 3 builds is too fast to read now (compared with R5 1600AF) Jun 01 17:51:16 ==> /OE/build/luneos-master/webos-ports/tmp-glibc/work/core2 Jun 01 17:51:16 [11023/21175] CXX obj/extensions/common/mojom_shared_cpp_sou Jun 01 17:51:16 ==> /OE/build/luneos-dunfell/webos-ports/tmp-glibc/work/core Jun 01 17:51:16 [5834/21175] CXX obj/third_party/icu/bundled_icuuc/ustring.o Jun 01 17:51:16 ==> /OE/build/luneos-zeus/webos-ports/tmp-glibc/work/core2-6 Jun 01 17:51:19 [8307/21175] CXX obj/third_party/webrtc/modules/video_coding Jun 01 17:59:45 it was fine until qtwebengine started linking (all 3 at the same time Jun 01 18:00:07 128G ram wasn't enough for that Jun 01 18:00:26 OOMK to the rescue Jun 01 18:03:17 8x chrome and 3x cc1plus killed and still it managed to somehow break barriers so that it stopped working and even responding to kill Jun 01 18:03:45 that will teach me how far to push the new builder :) Jun 01 18:39:16 JaMa: Seems replacement modules arrived in Amsterdam today but no deliveries due to public holidays so I guess tomorrow I'll get them and can tesk Jun 01 18:39:35 s/tesk/test Jun 01 18:58:39 Herrie: did you have to send the faulty ones first (so now building on some spare ram you had around) or did they ship replacement kit in advance? Jun 01 19:01:29 Replacement in advance with blocked amount on credit card Jun 01 19:02:34 ChanServ! Jun 01 20:43:31 Herrie: wendell made new video with his 256G kit used in test-oe-build-time https://www.youtube.com/watch?v=GSIFsN1Ofqw Jun 01 20:49:09 JaMa: oh, amd-pr0n :) Jun 01 20:56:35 JaMa: This is good stuff Jun 01 20:56:51 I'll stick with my 2nd gen for a while though :D Jun 02 01:53:50 Tofe: if you want some HW pr0n check this channel https://www.youtube.com/channel/UCGxtYMx16kD7jK0RKSxsVrw not just because of his handlebar moustache, but the mods he is doing to either amd and intel builds, not very practical, but I would call this art :) **** ENDING LOGGING AT Tue Jun 02 02:59:57 2020