**** BEGIN LOGGING AT Sun Jun 28 02:59:58 2020 Jun 28 06:02:27 Morning Jun 28 06:02:40 Tofe: BTW this is what I get on Halium 9 image with our 7.1 kernel: https://paste.ubuntu.com/p/GyvDR7GvrV/ Jun 28 06:02:45 Seems LXC fails to start there Jun 28 06:06:03 Tofe: Seems it has issues with the vendor mount: https://paste.ubuntu.com/p/bWnVvjRvpZ/ Jun 28 06:07:27 My mounts: https://paste.ubuntu.com/p/NGXswYKTgB/ Jun 28 06:22:11 Might be the same issues I had with 8.1, seems they might minimalize the vendor a bit too much here: https://github.com/Halium/hybris-patches/blob/halium-9.0/build/make/0003-hybris-Reduce-vendorimage-build-size.patch Jun 28 06:40:50 Seems I had this issue with 8.1 as well Jun 28 06:41:06 Going by the logs, let me try mka systemimage vendorimage, that seems to have worked previously Jun 28 07:51:14 hi Herrie Jun 28 07:51:28 what shall we do now. Jun 28 07:52:20 any other solution in your mind Jun 28 07:53:51 even the mobile device CONNECTION is not detected by desktop operating system by any means. Jun 28 07:56:07 atleast if it could detect mobile as unrecognized device,... Jun 28 07:59:08 BT40: so if you have bootable kernel, you can check what in blocking CONFIG_USB_G_ANDROID to enable using menuconfig. That's what I've tried before. Jun 28 08:01:05 nizovn, thanks. can we analyse this from journal log file (luneos-data/var/log)? Or there is some other log for this at different location? Jun 28 08:01:42 yes kernel is booting into luneos Jun 28 08:03:20 we can see in logs that android_usb is not loading, so we need to enable it in kernel config Jun 28 08:04:40 if we get it loaded (in logs), then we can debug further (check adb config in LuneOS) Jun 28 08:08:01 just try: make menuconfig arch/arm64/configs/lineageos_land_defconfig Jun 28 08:09:14 https://www.dropbox.com/s/92dydl0pa9rl16u/log27jun.txt?dl=0 This is the log file of luneos journalctl Jun 28 08:09:42 ok Jun 28 08:11:56 Do i need to run make menuconfig from kernel sources diretory? Jun 28 08:12:00 yes Jun 28 08:12:20 fatal error: curses.h: No such file or director Jun 28 08:12:41 hm Jun 28 08:12:42 make[1]: *** [scripts/Makefile.host:108: scripts/kconfig/mconf.o] Error 1 Jun 28 08:13:24 which kernel sources are you using? are they available online? Jun 28 08:14:08 yes Jun 28 08:14:26 just sharing Jun 28 08:15:53 https://github.com/BT40/android_kernel_xiaomi_msm8937/tree/d598d6c97cc17a78238597fc19060ffa9621b0cc This one i forked from herrie, added a few new usb options, kernel is compiled successfully. Jun 28 08:16:06 i will also share original one by herrie Jun 28 08:16:49 https://github.com/Herrie82/android_kernel_xiaomi_msm8937/tree/539056790439a43e0dfa4279a26d9065818c0224 This is original one by herrie Jun 28 08:17:11 ok, thanks Jun 28 08:18:45 will show conflicts when it's downloaded Jun 28 08:19:29 nizovn, please try herrie one, which is reliable. My version is just for testing. Jun 28 08:19:57 BT40: well, you kernel is bootable, let's start with it Jun 28 08:20:44 latest one is due for test Jun 28 08:20:51 i will test in a couple of mins Jun 28 08:20:54 flashing Jun 28 08:23:06 yes it is booting Jun 28 08:23:13 we can proceed Jun 28 08:23:34 so it says: Symbol: USB_G_ANDROID [=n] Jun 28 08:24:29 Selects: USB_F_ACM [=n] && USB_F_AUDIO_SRC [=n] && USB_F_SERIAL [=n] && USB_LIBCOMPOSITE [=y] && USB_F_FS [=n] && USB_U_SERIAL [=n] && USB_F_ECM [=n] && USB_F_NCM [=n] && USB_F_MASS_STORAGE [=n] && USB_F_UAC1 [=n] && USB_F_UAC2 [=n] && USB_F_UVC [=n] Jun 28 08:24:42 okay, seems good news for me, u are again the saviour this time. U have given the useful clue Jun 28 08:26:09 please guide, how can we enable this and troubleshoot Jun 28 08:26:19 the " USB functions configurable through configfs" is selected instead of "Android Composite Gadget" Jun 28 08:29:23 so the difference between this config and the one with enabled android_gadget is following: Jun 28 08:29:40 CONFIG_USB_CONFIGFS=y <- removed Jun 28 08:30:14 added: CONFIG_USB_F_ACM=y CONFIG_USB_U_SERIAL=y CONFIG_USB_F_SERIAL=y CONFIG_USB_F_NCM=y CONFIG_USB_F_ECM=y CONFIG_USB_F_MASS_STORAGE=y CONFIG_USB_F_FS=y CONFIG_USB_F_UAC1=y CONFIG_USB_F_UAC2=y CONFIG_USB_F_UVC=y CONFIG_USB_F_AUDIO_SRC=y CONFIG_USB_G_ANDROID=y Jun 28 08:31:54 so removing CONFIG_USB_CONFIGFS=y and adding just CONFIG_USB_G_ANDROID=y is enough (it will enable the rest automatically) Jun 28 08:37:36 ok Jun 28 08:37:39 BT40: so you can just remove this line, and it should work https://github.com/BT40/android_kernel_xiaomi_msm8937/blob/d598d6c97cc17a78238597fc19060ffa9621b0cc/arch/arm64/configs/lineageos_land_defconfig#L841 Jun 28 08:40:11 Herrie config files had this already nizovn , but it didn't enabled usb. can u please verify that file? Jun 28 08:41:19 https://github.com/Herrie82/android_kernel_xiaomi_msm8937/blob/539056790439a43e0dfa4279a26d9065818c0224/arch/arm64/configs/lineageos_land_defconfig Jun 28 08:41:34 above file is heerie's one Jun 28 08:41:45 will see in a bit Jun 28 08:41:47 which seems to be inline with your Jun 28 08:43:16 it's disabled in 539056790439a43e0dfa4279a26d9065818c0224 Jun 28 08:43:32 ok Jun 28 08:45:09 CONFIG_USB_CONFIGFS=y is selected as well Jun 28 08:45:15 there is warning: .config:857:warning: override: USB_CONFIGFS changes choice state Jun 28 08:47:21 sorry, but i cant see it in defconfig file in herrie. seems i am missing something Jun 28 08:49:42 oops, i chose wrong commit Jun 28 08:53:36 ok, in 539056790439a43e0dfa4279a26d9065818c0224 CONFIG_USB_FUNCTIONFS is selected instead of CONFIG_USB_G_ANDROID Jun 28 08:54:31 g android is also selected in that Jun 28 08:54:43 line number 448 Jun 28 08:57:47 yes, but it's blocked by following CONFIG_USB_FUNCTIONFS, if you moved CONFIG_USB_G_ANDROID to the bottom, it will enable Jun 28 08:57:54 *move Jun 28 09:00:19 do functionfs still needs to be there on top? or we can delete that? Jun 28 09:02:57 yes, we can delete it as it won't be enabled anyway, and may confuse Jun 28 09:03:44 what about CONFIG_USB_F_FS=y Jun 28 09:04:51 f_fs seems similar, so thats y asking Jun 28 09:05:06 yes, it should be enabled, it actually will be enabled by CONFIG_USB_FUNCTIONFS Jun 28 09:05:21 oops, CONFIG_USB_G_ANDROID i mean Jun 28 09:05:57 ok, so i will write g_android just anobe F_Fs. Jun 28 09:06:02 above Jun 28 09:07:07 ok Jun 28 09:14:17 * BT40 : kernel compilation started, will take 15 mins. Then will report. Jun 28 09:32:51 flashed new kernel. booted into luneos. USB not detected Jun 28 09:33:35 ok, let's see in logs if android_usb is loaded Jun 28 09:33:37 https://github.com/BT40/android_kernel_xiaomi_msm8937/commit/1a213165b30c44be0113cab27c4c1ad6944c6e8d Jun 28 09:34:03 This is commit. can u please verity as like earlier did Jun 28 09:34:12 yup Jun 28 09:34:18 meanwhile, i am fetching logs Jun 28 09:36:01 yes, CONFIG_USB_G_ANDROID is enabled (not blocked) in lineageos_land_defconfig_hm Jun 28 09:41:10 https://www.dropbox.com/s/47ilritq3vcdvvh/log28jun1.txt?dl=0 Jun 28 09:41:18 Here is the log Jun 28 09:45:57 init: write_file: Unable to open '/config/usb_gadget/g1/bcdUSB': No such file or directory Jun 28 09:46:18 Line 4172 onwards various usb errors Jun 28 09:50:51 line 15751... init: Not starting service for keychord 1 because ADB is disabled Jun 28 09:52:01 do you have fingerterm app working? Jun 28 09:52:25 logs are flooded, and cropped... Jun 28 09:53:00 will ceck fingerterm Jun 28 09:54:50 fingerterm opens, but system is not in state that i can work. Every half second screen turns off Jun 28 09:55:06 automatically Jun 28 09:57:55 so i think we can try to get rid of flood in logs, or research screen power offs Jun 28 10:00:52 sure sir Jun 28 10:03:57 thinking about the second, if it can be solved by one command, maybe you will be able to type it Jun 28 10:07:52 yeah :) Jun 28 10:08:01 ready to try Jun 28 10:10:10 i would try to disable ambient light sensor Jun 28 10:11:36 it can be done in luna-platfom.conf, did you add it for land? Jun 28 10:12:15 like this: https://github.com/webOS-ports/meta-webos-ports/blob/zeus/meta-luneos/recipes-webos/luna-sysmgr-conf/luna-sysmgr-conf/tissot/luna-platform.conf Jun 28 10:16:12 checking Jun 28 10:17:29 yes i added Jun 28 10:19:10 In this, i changed resolution of display and DPI values. Also put HasHomeButton=true. Jun 28 10:23:42 Two things i have noted, sharing. After the change of USB configs in kernel, USB_gadget errors have gone from log. No usb errors are there which is positive. Jun 28 10:24:13 downside is, phone cannot charge now. neither on wall socket, nor on PC. Jun 28 10:25:38 new kernel is effective from time 15.01 in logs Jun 28 11:00:29 have no good ideas... let's clean logs then Jun 28 11:20:59 i would try to add "#define pr_err(reason, fmt, ...)" to https://github.com/BT40/android_kernel_xiaomi_msm8937/blob/1a213165b30c44be0113cab27c4c1ad6944c6e8d/drivers/power/qpnp-smbcharger.c#L497 Jun 28 11:21:22 and check logs again Jun 28 11:23:21 ok Jun 28 11:23:31 i'll have a look Jun 28 11:35:37 added that line, started kernel compilation, will take around 20mins on my machine to build Jun 28 11:36:13 shall i delete all previous logs to reduce bloat? Jun 28 11:39:27 no, they just shifted boot messages out of recorded logs, don't bother Jun 28 11:42:54 for the while kernel is compiling, can we have a look on other issues? Or we will take care of them later on? Jun 28 11:44:32 not sure much what to do as we need adb to fix screen issue, or need to solve screen to fix adb... hope logs will give a hint Jun 28 11:45:53 nizovn, do u pro guys use luneos on regular basis? I mean on your daily routine phone, which os do you guys in this channel use? Jun 28 11:46:34 i mean is android/ios still daily phone OS? Jun 28 11:48:03 well LuneOS is not ready yet, I use hp pre3 (on webOS) as main phone. + using hp touchpad (webOS) and xiaomi tissot (android) for browsing Jun 28 11:49:58 compilation failed with some error Jun 28 11:50:33 I have HP Veer, Touchpad and an Android phone for works Jun 28 11:50:49 I plan to use LuneOS on a daily basis at some point Jun 28 11:50:52 great Jun 28 11:51:23 I saw UBPorts added some commit to kernel tree for display Jun 28 11:51:34 That might have broken things in LuneOS Jun 28 11:52:47 pr_err("Missing qcom,pmic-revid property - driver failed\n"); Jun 28 11:52:58 This is line in error log Jun 28 11:53:08 will share full log on dropbox Jun 28 11:53:52 seems i added to much arguments, should have been: "#define pr_err(fmt, ...)" Jun 28 11:54:32 BT40: how do you build kernel? i think there is a way to rebuild it faster Jun 28 11:54:44 reusing previous build Jun 28 11:55:14 MACHINE=land bb virtual/kernel Jun 28 11:55:21 using above command Jun 28 11:57:49 https://www.dropbox.com/s/zew1wh6k83iexqg/kernelfail1.log?dl=0 This is eeror log for failed kernel compilation after adding charger Jun 28 11:58:18 you can edit qpnp-smbcharger.c in tmp_glibc, and then continue previous build with -c compile Jun 28 12:00:01 so you can just edit /home/hextreme/luneos/mi-env/webos-ports/tmp-glibc/work-shared/land/kernel-source/drivers/power/qpnp-smbcharger.c file Jun 28 12:00:29 maybe add even less arguments: "#define pr_err(fmt, ...)" Jun 28 12:00:33 MACHINE=land bb virtual/kernel -c Like this? Jun 28 12:00:58 ah, i meant "#define pr_err(...)" Jun 28 12:01:16 BT40: yes, MACHINE=land bb virtual/kernel -c compile Jun 28 12:01:27 ok. will do now Jun 28 12:02:07 ah wait Jun 28 12:02:27 also add -f Jun 28 12:03:08 MACHINE=land bb virtual/kernel -c -f compile ? Jun 28 12:04:24 seems better MACHINE=land bb virtual/kernel -C compile Jun 28 12:04:40 okay Jun 28 12:04:55 started Jun 28 12:05:30 yes, it has started a forced run :) Jun 28 12:05:59 ok.. tend to forget options :) Jun 28 12:13:04 Today i tested luneos browser to check whether sim works or not cuz calling was not working. I wanted to also test internet data. Surprisingly, it worked. I m really impressed. Browser works great for me. Jun 28 12:13:22 websites were loaded fine Jun 28 12:13:28 :) Jun 28 12:14:49 once i got telephony work, i'll ditch android and make LUNEos my production phone. Rest of the things i can manage on desktop. Jun 28 12:15:13 BT40: telephony is easy to fix once we have the other things sorted Jun 28 12:16:17 Tofe made progress with telephony recently iirc Jun 28 12:17:22 compilation completed, will flash and report Jun 28 12:27:59 https://www.dropbox.com/s/j754kj74qwuglty/log28jun2.txt?dl=0 Jun 28 12:28:33 Above is new log file. I have filtered results to half hour. Used this syntax: journalctl -S -0.5h --no-pager --file=system.journal > log28jun2.txt Jun 28 12:30:15 seems we need full logs, as the clock is getting updated after relevant messages Jun 28 12:30:50 ok Jun 28 12:32:20 https://www.dropbox.com/s/8ziehpcr6cssvot/log28jun3.txt?dl=0 Here is full log Jun 28 12:35:18 compared both one, even the filtered one ase same as full one if we compare number of lines from time 5.49 to 5.50 which is around 2280 lines Jun 28 12:35:19 hm, the beginning still cropped Jun 28 12:35:57 journalctl --no-pager --file=system.journal > log28jun3.txt Jun 28 12:36:05 is the command ok i am using? Jun 28 12:38:22 charging and battery bloat seems to be gone now Jun 28 12:41:57 the command seems correct. not sure why boot logs are cropped... Jun 28 12:44:23 Herrie, any idea why logs are cropped? Jun 28 12:46:09 yesterday logs were not cropped. They started from kernel: Initializing cgroup subsys cpuset, May 10 00:01:36 land kernel: Initializing cgroup subsys cpu Jun 28 12:47:17 seems so Jun 28 12:48:08 early logs didn't have this line from which todays logs are started: Synchronized to time server for Jun 28 12:48:41 i compared to see how much part is cropped, but strange. This line was not found on previous logs Jun 28 12:53:43 hm, what have changed? sim inserted or internet data enabled? Jun 28 12:55:34 Maybe the limit for logging in defconfig? Jun 28 12:55:42 So it only logs last x lines? Jun 28 13:00:08 i will eject sim and see what happen Jun 28 13:01:16 I always just do: journalctl > /media/internal/log.txt Jun 28 13:01:47 in phone or PC? Jun 28 13:01:52 And then adb pull /media/internal/log.txt Jun 28 13:01:57 In phone Jun 28 13:01:59 BT40: are you deleting old logs from /var/logs? maybe there is no free space.. Jun 28 13:02:07 shall i do this in recovery? Jun 28 13:02:29 nizovn, nothing deleted as of now, shall i delete? Jun 28 13:03:16 well, you already have copies, it's safe to clear them Jun 28 13:03:38 Herrie, shall i run that command in recovery mode? Jun 28 13:04:01 or need to run in luneos terminal? Jun 28 13:04:22 nizovn, any recommended way to delete? Jun 28 13:06:32 nizovn, can i delete the full folder (having some long random name) in recovery file manager? will luneos automatically create newone? Jun 28 13:06:54 yes, it should Jun 28 13:07:32 not sure about recommended way, probably some journalctl command, but we can't use it Jun 28 13:08:10 ok, removed the folder. i hope it will be created automatically on next boot :) Jun 28 13:08:36 18.38 start Jun 28 13:12:11 yes, got full log this time, posting in a minute Jun 28 13:13:03 Here it is: https://www.dropbox.com/s/h2cp7nidgwb3wgn/log28jun5.txt?dl=0 Jun 28 13:15:16 look cropped again... Jun 28 13:19:41 Started Android Debug Bridge. in line 289.......................... Unable to open '/sys/class/android_usb/android0/f_ffs/aliases': No such file or directory in line 933, .......... nable to open '/config/usb_gadget/g in 919 line Jun 28 13:19:53 can these give some clue? Jun 28 13:20:49 even removed the sim, deleted the logs. still cropped logs. Jun 28 13:25:20 not sure why /sys/class/android_usb/android0/f_ffs/aliases doesn't exist Jun 28 13:25:38 do you have something in /run/log/journal? Jun 28 13:27:18 no /run folder in main directory Jun 28 13:27:31 ok Jun 28 13:32:52 in spare time, please have a look at https://github.com/boundarydevices/linux-imx6/commit/fcc22b2d061b3f525dcb6c86ea5cf0f1ec04dc98 If it could help Jun 28 13:42:47 In order for this to work configfs must be available, so CONFIGFS_FS must be ‘y’ from https://www.kernel.org/doc/html/v5.4-preprc-cpu/usb/gadget_configfs.html. Is this relevant to us nizovn ? Jun 28 13:45:21 configfs is another way to get adb, but in luneos android usb gadet is assumed, it works for all targets for now Jun 28 13:48:32 ok Jun 28 14:34:26 so how to proceed next Jun 28 14:36:06 do you remember why we are looking journald logs, and not /sys/fs/pstore/console-ramoops Jun 28 14:36:10 ? Jun 28 14:36:19 it's empty? Jun 28 14:54:31 nizovn: Well they should be the same normally in an ideal situation Jun 28 14:54:41 console-ramoops is sometimes a bit garbled Jun 28 14:54:54 i'm starting to hate journald :P Jun 28 14:55:26 it hides useful messages Jun 28 15:00:42 nizovn, yes it is empty. just checked now even with updated kernel Jun 28 15:02:46 i dont know if it is okay to ask in this channel LOL, is there any other OS which i can run instead of android? I just need basic stuff. Maybe some other webos port, or something else :) Just for the time being we figure out luneos Jun 28 15:05:07 does usb_gadgets files reside in memory? or they are physically created? analysing >write_file: Unable to open '/config/usb_gadget/g1/os_desc/b_vendor_code': No such file or directory Jun 28 15:05:35 i mean in virtual file system in ram? Jun 28 15:07:50 list of some distibutions: https://wiki.pine64.org/index.php/PinePhone_Software_Releases Jun 28 15:08:22 usb_gadget seems not related, i had those warnings with working adb iirc Jun 28 15:12:00 okay Jun 28 15:15:40 qcom-usb-sh' (pid 55) exited with status 0 is this one suspicious? Jun 28 15:16:26 and last one is sys/class/android_usb/android0/f_ffs/aliases No other usb related error found in current log. Jun 28 15:16:51 qcom-usb-sh not sure what it is, probably not needed Jun 28 15:16:57 If we can try to tackle suspicious among above two? Jun 28 15:17:26 the second we are trying to solve, yes Jun 28 15:17:57 i found the difference between land and tissot regarding pstore Jun 28 15:18:37 it should have something like this in early log: Reserved memory: reserved region for node 'pstore_reserve_mem_region@0': Jun 28 15:26:23 can we manually trigger adb, ffs through writing an itit file? assuming if something that init these services might not get invoked in our current case which could be causing problem Jun 28 15:26:37 itit=init typo Jun 28 15:27:42 it's being done by android-gadget-setup script Jun 28 15:38:55 Find nizovn 's commit for xiaomi :) https://gitlab.com/nizovn/meta-smartphone/commit/f3b819d6ed85e5e5a09924f1f2775077f9ffb672 Jun 28 15:39:44 :) Jun 28 15:39:48 yes, u did in android gadget script regarding adb. Jun 28 15:40:26 no, i just modified some lines, the script was created before Jun 28 15:55:25 Tofe: After reading the notes and redoing some steps I did for 8.1 I can now get a bit further in Halium 9.0 it just seems to fail to start LXC container telling me it cannot find /sbin/init while this exists as a symlink lrwxrwxrwx 1 root root 22 Feb 14 1970 /sbin/init -> ../lib/systemd/systemd Jun 28 15:55:42 This is my log from LXC: https://paste.ubuntu.com/p/FG8JpR8BMR/ Jun 28 19:36:40 Hello all! I made a post on the WebOS nation forums a few days ago about getting mainline kernels working on the HP Touchpad and was invited here. Don't know if it's been done yet, though I've managed to finally get framebuffer output! I wrote up a short DTB (can be seen here: https://gist.github.com/PieGuy128/dc973ade2746b5589ab19869d2b31ca8) to get simplefb to work. Of course I'm getting some oops messages Jun 28 19:36:40 but I think this is a good step! Jun 28 19:38:17 Hi pieguy128, thanks for dropping by! Jun 28 19:44:01 pieguy128 Seems there's quite some work done for getting the dragonboard with the same SOC supported in mainline kernel as per: https://dflund.se/~triad/krad/dragonboard/ Jun 28 19:54:34 pieguy128: There's also quite some good detailed steps on https://wiki.postmarketos.org/wiki/The_Mainline_Kernel Jun 28 19:55:53 There's this one for Dragonboard: https://elixir.bootlin.com/linux/latest/source/arch/arm/boot/dts/qcom-apq8060-dragonboard.dts Jun 28 19:56:26 And this one for the SOC: https://elixir.bootlin.com/linux/latest/source/arch/arm/boot/dts/qcom-msm8660.dtsi Jun 28 20:32:11 Aye, if we do need to write/modify DTs for everything that could be a bit of work! Postmarket's qualcomm porting page suggests we need a DT-based downstream kernel, though to my knowledge the 3.4 QCom kernel doesn't use any DTBs? Wonder if the relevant values are hardcoded or if the bootloader supplies them Jun 28 21:25:34 pieguy128: I don't know that much about kernels to be honest, I can build and modify them a bit but that's about it Jun 28 21:25:49 Tofe knows a lot more but he's away for a couple of days Jun 28 21:26:02 I guess he'll be back Tuesday or Wednesday Jun 28 21:26:17 He has been looking at mainline Hammerhead and Pinephone Jun 29 02:18:53 So, I think I found the 3.4 kernel's device tree alternative (https://github.com/invisiblek/android_kernel_hp_tenderloin/tree/lineage-15.0/arch/arm/mach-msm and subdirs) Jun 29 02:20:27 Whenever I try to boot with an initrd I do run into a panic. Since it scrolls through the log quite quickly I'm not able to see the whole message. Anyone know of a UART port on the touchpad? Jun 29 02:22:52 Post on XDA Devs suggests there's one in the headphone jack, but no schematic **** ENDING LOGGING AT Mon Jun 29 02:59:57 2020