**** BEGIN LOGGING AT Sun May 31 02:59:57 2020 May 31 06:21:23 JaMa: The webos_preferred_gfx_image_format was introduced by webos_app bclass (which we surely do need). I'm not sure what is the background of this webos_preferred_gfx_image_format class. We could probably do without it. It seems to convert some images in appinfo.json to another format, not sure why though, You could see that in some internal notes maybe, May 31 06:22:51 I guess webos_qmake5 can indeed be dropped, doesn't seem to be included anywhere just like webos_prerelease_dep May 31 06:26:13 * BT40 is wating for instructions to get out of bootloop May 31 06:28:31 BT40: Did you put your files osmewhere so we can have a look with you? May 31 06:29:08 which log files are needed? May 31 06:29:46 I mean your meta-webos-ports and meta-smartphone bits May 31 06:29:47 * BT40 is parallely reading github tutorial and will upload land files there May 31 06:29:54 Or did I miss those? May 31 06:30:05 will upload in some minutes on github May 31 06:30:21 Are you using LInux or Windows? May 31 06:30:27 ubuntu May 31 06:30:29 I use GitKraken on Linux, pretty OK May 31 06:30:44 ok, will install gitkraken May 31 06:31:06 Pretty OK GUI so you don't need to deep dive into command line git right away ;) May 31 06:45:08 meta webos ports..... https://www.dropbox.com/s/f8c7te3muzbll5n/meta-webos-ports.zip?dl=0 May 31 06:45:32 meta-xiaomi........ https://www.dropbox.com/s/189gxigimt5pd33/meta-xiaomi.zip?dl=0 May 31 06:46:05 meanwhile i'll also be uploading to github. To reduce sixe, i have removed halium image from above shared folders. May 31 06:53:33 Great May 31 07:13:57 For Github: Basically create an account, fork meta-webos-ports and meta-smartphone to your github account. Make new branches for both based on zeus and call it zeus-land for example. Then put your changes in these branches (simply copy the files, commit them and push to github) May 31 07:18:29 Let me push it quickly into a branch, give me a few minutes May 31 07:19:25 Herrie: webos_preferred_gfx_image_format was used by some TVs to convert images to hw accellerated formats like etc May 31 07:21:23 JaMa: OK we could drop it from our webos_app bbclass then. WE need webos_app for ACG auto generation of role files based upon requiredPermisisons in appinfo.json May 31 07:21:59 yes, I agree with that May 31 07:22:20 BT40: https://github.com/webOS-ports/meta-webos-ports/commit/701ec3a55a848b2a4603391b1889f72c0481ba27 please note that you need to make a change to meta-luneos/recipes-core/packagegroups/packagegroup-luneos-extended.bb at your end! May 31 07:27:10 BT40: Where did you get the 70-land.rules from? May 31 07:27:27 * BT40 is happy that Herrie made work easier by creating land github May 31 07:32:15 70_rosy.rules .... just renamed that files May 31 07:33:48 anything wrong in that? May 31 07:41:08 For kernel, you need to update this really: BT40: Nothing critical for now, you should actually do: May 31 07:41:17 Sorry, wrong paste: https://paste.ubuntu.com/p/rhnZrzN8kF/ May 31 07:41:39 I ran the kernel check config script against: https://raw.githubusercontent.com/kskarthik/android_kernel_xiaomi_msm8937/57a27d4b54a60f55346d1c08e6c0de3544008d6c/arch/arm64/configs/lineageos_land_defconfig May 31 07:43:07 For udev rules basically follow: https://raw.githubusercontent.com/kskarthik/android_kernel_xiaomi_msm8937/57a27d4b54a60f55346d1c08e6c0de3544008d6c/arch/arm64/configs/lineageos_land_defconfig May 31 07:44:13 WARNING: CONFIG_MODULES is invalid: that was why wlan compilation failed am i right? May 31 07:51:10 Yes May 31 07:56:17 For udev: http://docs.halium.org/en/latest/porting/debug-build/udev.html?highlight=udev May 31 07:56:20 Let me do that quickly May 31 08:02:39 udev rules are so much different comparing rosy and kartik one's May 31 08:02:46 This one has updated udev rules: https://github.com/Herrie82/meta-smartphone/commit/8dc6dd87ccf33778fd17d16414c6a6d2b7201640 May 31 08:03:47 They shouldn't be critical AFAIK for the kernel initially, but good to have the right ones May 31 08:05:04 but the updated ones on githum looks different from kartik guy halium reference which you gave earlier.......syntax wise May 31 08:05:15 ACTION=="add", KERNEL=="genlock"........... this is not present in that guys file May 31 08:06:37 that guys file has somewhat like this May 31 08:06:38 CONFIG_LOCALVERSION="-lineageos" May 31 08:06:39 CONFIG_AUDIT=y May 31 08:06:39 CONFIG_NO_HZ=y May 31 08:07:50 Sorry my paste was wrong above... May 31 08:08:12 For kernel I mean you need to keep defconfig as it is, just adjust the values in there to https://paste.ubuntu.com/p/rhnZrzN8kF/ May 31 08:11:43 The ones that are not there or missing May 31 08:32:08 Morning! May 31 08:34:13 Morning! May 31 08:36:58 * BT40 here 2.05pm May 31 08:37:16 It's always morning when joining in :) May 31 08:38:25 BT40: I was thinking also, one way to know where it bootloops (kernel ? initrd ? userland ?) is to add "emergency" to the kernel command line May 31 08:38:48 that way systemd will fallback right away on an emergency shell May 31 08:39:11 you won't see anything on the screen, adb won't be available... but it won't bootloop. So we would know that at least systemd is started May 31 08:40:38 If it bootloops at the kernel loading stage, it could be wrong values (like typos) for the android boot image parameters May 31 08:47:08 BT40: https://github.com/Herrie82/android_kernel_xiaomi_msm8937/commit/02fecde7b0c73d98009fb00f91cce5d75b3b5a4f May 31 08:47:13 Like that May 31 08:47:26 https://github.com/Herrie82/meta-smartphone/commit/8dc6dd87ccf33778fd17d16414c6a6d2b7201640#diff-c9be41e41d3bf21ba0c95ce9165bd582R12 doesn't match https://github.com/kskarthik/android_device_xiaomi_land/blob/halium-7.1/board/kernel.mk#L5 May 31 08:50:18 Also it should be "earlycon=msm_hsl_uart,0x78B0000" at the end of the command line https://github.com/Herrie82/meta-smartphone/commit/8dc6dd87ccf33778fd17d16414c6a6d2b7201640#diff-c9be41e41d3bf21ba0c95ce9165bd582R11 May 31 08:51:08 but the non matching ANDROID_BOOTIMG_KERNEL_RAM_BASE is sure fatal to the boot May 31 08:51:21 Tofe: Ah this I didn't check these yet in detail May 31 08:51:41 the rest of the values seem ok May 31 08:53:55 Tofe: You sure about the RAM BASE? We have same with Mido: https://github.com/Herrie82/android_device_xiaomi_mido/blob/pgz-14.1/BoardConfig.mk#L41 and https://github.com/shr-distribution/meta-smartphone/blob/zeus/meta-xiaomi/recipes-kernel/linux/linux-xiaomi-mido_git.bb#L12 May 31 08:54:42 Rosy as well: https://github.com/shr-distribution/meta-smartphone/blob/zeus/meta-xiaomi/recipes-kernel/linux/linux-xiaomi-rosy_git.bb#L12 and https://github.com/Tofee/android_device_xiaomi_rosy/blob/halium-7.1/BoardConfig.mk#L45 May 31 08:54:59 I never understood where this 8000 came from but well May 31 08:55:25 ah, mmh May 31 08:56:34 It was there in all other targets so I simply did the same May 31 08:56:54 ok, so false lead here May 31 08:57:10 The earlycon shouldn't be critical right? May 31 08:59:24 right May 31 09:00:03 userdata isn't encrypted I hope ? May 31 09:07:33 BT40 can you share your Halium .tar.gz that you excluded from upload? May 31 09:13:42 sure May 31 09:15:13 Still need to setup my halium environment May 31 09:15:23 Otherwise would have built myself May 31 09:16:09 meanwhile, can we fetch some log file from my phone? to investigate? May 31 09:17:28 will bb built all packages from scratch after changes to kernel? oe some packages and reuse the rest? May 31 09:18:42 * BT40 .uploading file to dropbox, one minute remaining May 31 09:19:52 Here is halium image.... https://www.dropbox.com/s/kmkhwopvk3c4fk3/halium7-redmi3s-x1bz2.tar.bz2?dl=0 May 31 09:20:42 gotta go, lunch outside May 31 09:20:53 (wow... "outside"...) May 31 09:21:19 * BT40 :) lunch in the morning? May 31 09:21:50 well it's noon soon May 31 09:23:31 shall i flash android and use the phone meanwhile? May 31 09:24:21 if its gonna take long time to figure out bootloop. atleast i'd be able to receive phones May 31 09:58:17 BT40: yes flash android for now May 31 09:58:28 I'll build you a new image later today May 31 09:58:36 Now going out with kids for a bit May 31 09:59:17 okay May 31 11:19:48 bt40: started a build here May 31 11:44:23 BT40: Started a build here... What did you do to get the kernel to build agian? Remove -wError from Makefile, patch /include/linux/log2.h and there should be something else for WLAN? May 31 11:45:02 I'm getting "/home/herrie/LuneOS/zeus/webos-ports/tmp-glibc/work-shared/land/kernel-source/drivers/staging/prima/CORE/VOSS/src/wlan_nv_template_builtin.c:590:10: error: initializer element is not computable at load time May 31 11:45:02 | 590 | (((char *)&(nvDefaults.tables.hwCalValues.calData.psSlpTimeOvrHdxLNA5G) +" May 31 11:45:17 yes May 31 11:45:27 i have to see... not remembering May 31 11:45:39 i disabled werror, this is sure May 31 11:45:48 regarding wifi, i neeed to check May 31 11:46:39 Ah I guess this one: https://github.com/Herrie82/android_kernel_xiaomi_msm8953/commit/6fd6c9da0c0c95de31e99d4dea80a4d5212945ce May 31 11:47:06 Seems you might need this fix in your kernel: https://github.com/Tofee/android_kernel_xiaomi_msm8953/commit/c6be5e8f275d42a6de93607868a6c384bfec1e35 May 31 11:47:40 yes... both are same May 31 11:49:26 Ok retrying May 31 11:53:03 * BT40 believes 24 core monster will rip the build in less than an hour May 31 11:53:28 It's almost done ;) May 31 11:53:40 I had sstate from other target still I guess May 31 11:53:54 It's extremely quick... i would say 10-15 mins altogether if not less May 31 11:55:16 Fixing the kernel took most time LOL May 31 11:55:41 BT40: How is your connection? You OK to download the full image? May 31 11:55:47 yeah May 31 11:55:56 50mbps May 31 11:56:20 Even after 100gb, i get throttled speed of 10mbps May 31 11:56:40 Ah ok May 31 11:56:57 Let me create an account for you on my router's FTP server, so you can pull it frm there May 31 12:01:53 patch /include/linux/log2.h............... i didn'd implement this. Or maybe i forgot if implemented. May 31 12:02:42 * BT40 's bootloop might be due to absence of that patch????? May 31 12:06:01 No it wouldn't compile without May 31 12:06:15 I sent you DM with details, file is there May 31 12:54:28 Herrie, downloading failed twice, after fetching 350Mb plus...both times May 31 12:54:46 Herrie, can u please upload that file on my dropbox account? May 31 12:55:38 i can share username and password, no sensitive info is there :) so no problem in sgaring for me May 31 12:56:13 Yes May 31 12:56:16 Can do that May 31 12:56:55 messaged you details in personal chat. May 31 13:01:00 Is there any emulator for LuneOS? For testing developed apps? May 31 13:01:45 I am thinking of app development until OS is successfully installed. I mean I can do app development in spare time May 31 13:02:48 Most probably I'd be using enyo, or some other tool to create web apps. Avoiding native apps due to QT. May 31 13:05:35 BT40: Yes we offer virtualbox ;) May 31 13:06:03 Herrie, code: 494557 May 31 13:06:30 great May 31 13:06:42 will start learning enyo from tomorrow May 31 13:08:07 Well we support Enyo 1.0, 2.0, Enact (and any other webapp framework basically ;)) May 31 13:09:14 can i use GTK with python? Although i think gtk will not wrk due to qt5, but still sometimes asking is better than assumptions :) May 31 13:33:49 Not sure May 31 13:33:55 Tofe might know May 31 14:12:02 installing luneos May 31 14:13:47 installation complete, rebooting May 31 14:14:43 again, bootloop May 31 14:18:38 neither it did fallback to emergency shell. May 31 14:44:53 SRCREV = "aad1958e0b7ffd00ca2af58ad3f25578d241f233" May 31 14:44:53 May 31 14:44:53 KV = "3.18.31" These two lines i didn't change from rossy. I might be wrong here. isn't it? May 31 14:45:30 These should be according to kernel version of my halium build, not from rossy. May 31 14:47:13 could this cause issue if halium kernel is different than this? Lets say halium version is 3.18.20 and here we mentioned 3.18.31. Wouldn't if blent two different commits or patches if there? May 31 15:08:08 i have received this: Version = 3 May 31 15:08:08 PATCHLEVEL = 18 May 31 15:08:09 SUBLEVEL = 20 ........from xiaomi open source kernel sources. Seems it is pointing to 3.18.20. May 31 15:08:40 Link for source code: https://github.com/MiCode/Xiaomi_Kernel_OpenSource/tree/land-m-oss May 31 15:17:45 BT40: Do you have a console_ramoops file this time? May 31 15:18:56 What I usually do is hold the volume up, boot twrp and see if it's there... Sometimes it takes a few tries May 31 15:20:32 ls: /sys/fs/pstore/console-ramoops: No such file or directory May 31 15:21:34 pstore folder is empty when viewed through file manager of recovery May 31 15:29:17 I might want to add these to kernel: https://github.com/Herrie82/android_kernel_xiaomi_msm8953/blob/4a6943a93484c7376b3c671a7f73e59d5fd5e548/arch/arm64/configs/mido_defconfig#L4090 May 31 15:31:00 I'll build a new one May 31 15:32:58 I think we can try just the kernel May 31 15:40:16 BT40: will have something for you soonish I guess May 31 15:41:44 A kernel which you can just boot with fastboot boot kernel.uImage for example May 31 15:41:55 So you don't need a whole image May 31 15:42:05 The KV doesn't matter really May 31 15:42:49 Technically it should be 3.18.20 but will have no effect on the outcome of testing May 31 15:45:56 back May 31 15:46:42 Did you try adding "emergency" to kernel command line ? May 31 15:52:19 also tried adb logcat while bootloop. but no output. Adb said waiting for device ad thats it. So read from a thread>>capturing a logcat during a bootloop is that the boot process has to be far enough along to have started-up May 31 15:55:57 Tofe: seems the various pstore defconfig options were missing May 31 15:56:13 Should have a kernel soon that has them enabled May 31 15:57:47 BT40: it's not surprising; adb daemon is started quite late in the boot process May 31 15:58:47 BT40: notice that you don't need to flash the whole luneos image if you just want to test a new kernel build May 31 15:58:59 okay May 31 15:59:08 eager May 31 15:59:29 Added these now: https://github.com/Herrie82/android_kernel_xiaomi_msm8937/commit/e17d32daf2ad00b9f3e170c1a92fb04d7de91bf8 May 31 15:59:39 just boot into fastboot (power + volume *Down* ), then from your pc you can do "fastboot boot path/to/kerne-fastboot-image.img" May 31 16:00:04 that should do it May 31 16:00:54 BT40: in my case, the image is tmp-glibc/deploy/images/tissot/Image.gz-dtb.fastboot May 31 16:01:10 BT40: Just added Image.gz-dtb--3.18.31+gitr1+e17d32daf2-r0.2-land-0-0.fastboot to your Dropbox May 31 16:03:45 i tried to copy kmsg file via twrp file manager to sdcard. Its now stuck at copying file May 31 16:04:32 you can kill it, it'll never succeed May 31 16:04:41 kmsg can only be an output May 31 16:04:46 shall i reboot? May 31 16:05:03 i mean ho wto kill May 31 16:05:09 Ctrl+C ? May 31 16:05:28 ah it's via the UI of TWRP ? not adb ? May 31 16:05:33 i am in twrp May 31 16:05:42 direct file manager copy in twrp May 31 16:05:48 not adb May 31 16:05:49 ok, reboot May 31 16:08:06 how to install this fastboot file? May 31 16:08:29 do i need to be in fastboot? May 31 16:10:41 yes May 31 16:11:16 you can either flash it on the "boot" partition, or just start it without flashing May 31 16:12:11 i have selected reboot option in twrp. should i further select bootloader? will it go in fastboot? May 31 16:12:22 From PC: fastboot boot Image.gz-dtb--3.18.31+gitr1+e17d32daf2-r0.2-land-0-0.fastboot May 31 16:13:50 i need to go into fastboot mode. how to do? does fastboot work without android? May 31 16:15:29 i have done factory data reset to erase luneos. Hope it is okay. May 31 16:16:27 that... wasn't needed :) You'll need to reflash luneos now May 31 16:17:08 fastboot is the bootloader, so yes rebooting into the bootloader also means going into fastboot mode May 31 16:17:30 or you can hold Volume Down key while starting the device May 31 16:17:38 booting... May 31 16:17:39 OKAY [ 0.634s] May 31 16:17:39 finished. total time: 1.145s May 31 16:17:45 again in bootloop May 31 16:18:27 will flash lune os, then in fastboot, kernel reflash. May 31 16:20:09 yes May 31 16:22:07 BT40: something else, also, before flashing luneos May 31 16:22:22 okay May 31 16:22:27 could you format /data ? I don't mean wipe, I mean a real format step May 31 16:22:29 although already flashed May 31 16:22:31 :) May 31 16:22:45 okay May 31 16:22:50 i will do May 31 16:22:53 I think twrp proposes it May 31 16:23:15 just /data though, nothing else :) May 31 16:23:26 okay May 31 16:23:30 then you'll need to reflash luneos, of course May 31 16:23:37 will first do factory reset May 31 16:23:44 then data format May 31 16:23:51 don't bother, with factory reset May 31 16:23:52 then again factory reset May 31 16:24:03 then flash luneos May 31 16:24:09 no reset after format May 31 16:24:15 will then fastboot and run command May 31 16:24:25 ok May 31 16:24:39 I'm not sure what reset does, but in case it encrypts back the partition, I don't want to risk an unknown state May 31 16:25:10 After format we know the partition is clean an unencrypted, so it's fine to go on May 31 16:25:26 you might have to reboot once into twrp after formatting, though May 31 16:25:41 so that it updates its partition table May 31 16:27:14 From luneos point of view, reset isn't that helpful; we don't use android dalvik cache or data, it's just a folder in /data. Plus the boot partition. May 31 16:29:33 Tofe: it might be good to double check in twrp if data is really unencrypted? May 31 16:29:41 4 errors are thrown while formatting data May 31 16:29:58 1. error deleting crypto block device May 31 16:30:23 then it says continuing anyway May 31 16:30:37 fformatting data using mke2fs May 31 16:30:54 Herrie: I don't know how to check that :p May 31 16:30:56 OK that should be fine May 31 16:31:09 Tofe: wouldn't "mount" show that? May 31 16:31:17 Herrie: ah, yes, right ! May 31 16:31:19 then error: process ended with error 1. Unable to wipe data. Unable to format to remove encryption May 31 16:31:31 That's not good May 31 16:32:05 3rd error: failed to mount /data. device or resource busy May 31 16:32:22 You might want to try to force unmount it in twrp May 31 16:32:41 There should be somewhere option to mount/unmount in twrp May 31 16:32:41 4th, updating partition details, again error failed to mount partition May 31 16:34:42 in main twrp menu, there is an mount button May 31 16:35:19 when i select it, various partitions are displayed. out of them, system and cache are ticked. rest are not. May 31 16:35:54 below there are two options: select storage and disable mtp (mtp is useless here i think) May 31 16:40:53 data format completed. i think issue was due to my connection with pc May 31 16:41:53 although it has removed all internal storage data: like luneos rom May 31 16:44:52 reboot and go into twrp again May 31 16:45:27 then enter the "mount" command and paste us the output to see where we are May 31 16:46:04 mount command in ubuntu or twrp terminal? May 31 16:48:44 Twrp May 31 16:48:44 in adb this is the result: https://pastebin.com/2Kxu8JWF pstore is mounted:) May 31 16:49:28 This looks good: /dev/block/mmcblk0p49 on /data type ext4 (rw,seclabel,relatime,data=ordered May 31 16:49:30 same result on twrp, outputs are similar as expected May 31 16:50:10 Which means encryption is gone May 31 16:50:23 BT40: ok, good, flash luneos and then we can retry May 31 16:51:19 luneos flashed. should i boot into fastboot now? May 31 16:52:00 or give a try with old kernal once? May 31 16:53:18 No try the new kernel right away I would say May 31 16:53:26 But doesn't hurt to try the old one May 31 16:54:46 so i will do fastboot right? lets try new one May 31 16:55:51 yes, go into fastboot May 31 16:56:29 then on pc do "fastboot flash boot Image.gz-dtb--3.18.31+gitr1+e17d32daf2-r0.2-land-0-0.fastboot" May 31 16:56:40 this is output: okay finished. total time: 1.145s May 31 16:56:51 then "fastboot boot" May 31 16:56:53 automatically it has rebooted May 31 16:57:20 rebooted ? May 31 16:57:22 and we are in bootloop again May 31 16:57:33 fastboot flash shouldn't have rebooted right away May 31 16:57:54 did you do "fastboot flash" or "fastboot boot" ? May 31 16:58:35 fastboot boot Image.gz-dtb--3.18.31+gitr1+e17d32daf2-r0.2-land-0-0.fastboot May 31 16:58:35 downloading 'boot.img'... May 31 16:59:03 ok. go into fastboot again, and this time flash the kernel with "fastboot flash boot Image.gz-dtb--3.18.31+gitr1+e17d32daf2-r0.2-land-0-0.fastboot" May 31 17:00:26 flashing finished May 31 17:00:31 shall i reboot? May 31 17:00:58 just do "fastboot boot" May 31 17:01:21 fastboot: usage: expected argument May 31 17:01:37 ah, then fastboot boot Image.gz-dtb--3.18.31+gitr1+e17d32daf2-r0.2-land-0-0.fastboot May 31 17:01:59 it'll bootloop, but at least each time it will boot the newer kernel May 31 17:02:33 how many time shall i let in bootloop? 4 times enough May 31 17:02:37 do one or two boots, then go into twrp again May 31 17:02:47 ok May 31 17:03:02 now, check again /sys/fs/pstore May 31 17:04:45 nothing is there when viewed through twrp file manager May 31 17:04:53 empty folder May 31 17:05:40 ok... May 31 17:11:26 can we check just kernel boot? without luneos? i mean kust kernel with basic commandline? May 31 17:11:31 just May 31 17:12:04 terminal type os, no gui. May 31 17:13:41 Bt40 that is what we're trying May 31 17:13:52 First kernel boots, then it loads OS May 31 17:15:37 * BT40 thinks using kernel from untrusted guy named karthik is a mistake. Self compilation would have been more reliable May 31 17:16:40 as i have original xiaomi oem image, can we extract kernel from that? May 31 17:17:52 also xiaomi is providing kernel source code, if we can use that somehow May 31 17:20:29 Well we can take LineageOS kernel run Mer's kernel script and make changes to defconfig May 31 17:22:19 But karthik did the same more or less May 31 17:22:37 It could be some of the UBports defconfig changes cause issues though May 31 17:25:17 The kernel from karthik is based on LineageOS kernel as well which is based on Xiaomi's May 31 17:26:45 karthik's kernel is used for UBports ? May 31 17:30:27 wao. @Herrie82@TofeeHerrie82 authored and Tofee committed on Mar 7, 2019 commit 1ad88b041787d8ce8407a021271ef1031e95cba6..... and the patch was signed off by linus. May 31 17:30:32 Tofe: I see AppArmor stuff in defconfig so yes May 31 17:30:52 * BT40 is glad wo work with such great guys May 31 17:32:26 BT40: We probably just cherry picked some upstream commit from Linus... I don't recall contributing to upstream kernel anything myself May 31 17:32:59 Except for some patches to wlan & raid drivers I need for my builder May 31 17:33:21 Which tend to break with new kernels all the time May 31 17:37:14 I only contributed for a missing accelerator driver in my tablet May 31 17:38:13 ah, the commit is the fstat thing... yes it's a backport May 31 17:38:30 mostly useful for very old kernels (3.4, 3.10) May 31 17:39:51 so :) what should be our next target and procedure to sort bootloop May 31 17:40:14 * BT40 is desperate to try luneos May 31 17:40:32 compare a bit the commits between a working one (like for rosy) and yours May 31 17:41:38 but we followed all the procedure for rosy. May 31 17:41:54 theoretically it should work May 31 17:43:10 that's the theory; now we're at the reality step :) May 31 17:43:25 hahaha May 31 17:43:35 each case is a bit different, there's always something going wrong May 31 17:44:26 if we can know how you fixed bootloop then maybe we can get some clue.... both redmi snapdragons running miui May 31 17:45:04 will those changes reflect in github? the ones you made during fixation of bootloop? May 31 17:45:18 BT40: Well it's usually kernel or defconfig May 31 17:45:53 I could take a clean CM14.1/LOS kernel and modify the defconfig accordingly May 31 17:49:21 wouldn't land kernel source code more easy? i mean, xiaomi would have already configured kernel for that board. Sorry for half baked knowledge, just asking for learning May 31 17:49:57 Herrie: we could also remove the "apparmort" bits from the defconfig May 31 17:50:12 defconfig might be having 1000s or 1000s of options, modifying those will be time consuming May 31 17:50:26 BT40: Yes, but we need to always modify some May 31 17:50:37 LOS = Xiaomi dump + some LOS optimizations, which are in general fine May 31 17:51:33 then why we need to again make changes if they work fine? due to webos structure? or some other reason? May 31 17:52:09 are kernels different for lune and android? May 31 17:52:24 userland is; we use systemd, for a start May 31 17:52:39 We use different features in the kernel that Android doesn't use and vice versa May 31 17:52:47 okay May 31 17:52:49 I.e. we have "proper linux" not some funny variant ;) May 31 17:52:55 hahaha May 31 17:55:23 maybe last commit for tissot is needed as well (by description)? https://github.com/Tofee/tissot/commit/a8412322dc34a81517cfdb9f4ef1c75bcf263ad0 May 31 17:56:43 "2018 Jun 27 17:55:15 Herrie: actually my rosy does boot, after adding logs into the initrd... it's just that I forgot to add the ${LIBHYBRIS_RDEPENDS} that you've added today for rosy... so I didn't have the Halium image in the rootfs..." May 31 17:57:24 nizovn: Why not... May 31 17:57:33 nizovn: that a good lead, let's try May 31 17:57:39 Tofe: Yeah I fixed that one already for BT40 ;) May 31 17:57:53 BT40: Give me a few mins ;) May 31 17:59:46 nizovn, you nailed it man. My device behaves similar. On splash logio it waits for 10 seconds and reboots May 31 18:01:05 we'll see. :) maybe there are more than 1 issue :) May 31 18:01:42 as soon as we get an adb shell it should get easier May 31 18:02:06 BT40: Kernel is uploading to dropbox, filename: Image.gz-dtb--3.18.20+gitr0+5a2c473f94-r0.2-land-0-0.fastboot May 31 18:02:11 Done May 31 18:02:12 * BT40 believes if nizovn has signalled there are more than 1 issues, there will devinitely be. May 31 18:02:46 Let's start with this, if not I'll redo clean from CM14.1 branch. It are 4 commits + defconfig changes, should be do-able May 31 18:03:03 3.18.20 May 31 18:04:10 Herrie: BT40: should be 3.18.31, it's a bit strange ?... May 31 18:07:30 Tofe: I changed it to 3.18.20 as per Makefile May 31 18:07:58 but Makefile says 31 May 31 18:08:17 Ah wait :P May 31 18:08:18 should i fastboot flash or fastboot boot? May 31 18:08:22 https://github.com/kskarthik/android_kernel_xiaomi_msm8937/blob/halium-7.1/Makefile#L3 May 31 18:08:23 Well it shouldn't matter for boot May 31 18:08:35 I'm just worried you built the wrong kernel somehow May 31 18:08:45 No I changed the value in my .bb May 31 18:08:48 Changed it back now May 31 18:08:50 Herrie: ok :) May 31 18:09:04 BT40: do a flash first, it doesn't hurt May 31 18:09:20 we know the previous doesn't work anyway May 31 18:10:21 and if you bootloop but after starting userland (mount issue, or something else, who knows) it's better to be on the newer one after the first bootloop May 31 18:11:07 fastboot boot May 31 18:11:08 < waiting for any device > May 31 18:11:17 you're not in fastboot mode ? May 31 18:11:40 BOOOOOOOOM May 31 18:11:59 ANDROID REST IN PEACE May 31 18:12:13 hmmm ? May 31 18:12:20 YEAH May 31 18:12:30 looks like a happy comment May 31 18:12:40 SELECTING LANGUAGE May 31 18:12:51 wow, did it go that far? May 31 18:12:55 very well May 31 18:13:09 nizovn: thanks ! May 31 18:14:08 now it is looping at getting started May 31 18:14:20 BT40: app is crashing? May 31 18:14:26 yes May 31 18:14:35 language selection doesn't work May 31 18:14:40 asking again and again May 31 18:14:49 touch does work then May 31 18:14:59 automatically crashing and lets get started came again May 31 18:15:13 yes May 31 18:15:16 ok, well, now we can pull logs May 31 18:15:19 scrolling works May 31 18:15:50 you should have an adb access, too May 31 18:16:55 can you do "journalctl --no-pager > /tmp/land_journal.txt" and then adb pull /tmp/land_journal.txt May 31 18:17:10 no device shown in list of attached devices May 31 18:17:23 ah May 31 18:17:43 that's unfortunate May 31 18:17:52 :) fortunately Tofe solved many kernel issues, we just need to find which solution is needed this time May 31 18:19:33 crashing on language selection stopped as soon as i connect cable to pc May 31 18:19:34 Herrie: do you have an idea for easily retrieving logs ? I only can think of a complicated way May 31 18:19:46 now screen is getting on and off automatically May 31 18:20:02 beep is comming like some notification May 31 18:20:06 BT40: wait, you didn't have the usb cable connected ? May 31 18:20:17 connected usb May 31 18:20:29 I mean, adb won't work without the cable :) May 31 18:21:04 cable is connnected, still no adb May 31 18:21:09 ok May 31 18:21:12 i will restart phone and see May 31 18:21:17 ok May 31 18:21:30 BT40 getting stuff to show on screen is a really good state to be in at this point already May 31 18:21:38 It took us way longer to get there May 31 18:21:58 Touch not working etc is usually easy to fix May 31 18:22:08 power button does not work. i cant power off phone :) May 31 18:22:17 neither screen is in my control May 31 18:22:21 hahaha May 31 18:22:23 power button for 5s always work May 31 18:22:34 it's hard-wired May 31 18:22:41 BT40 that is easy to fix eventually May 31 18:22:58 We just need to figure out how the button is recognized May 31 18:23:01 I think Halium didn't start for some reason May 31 18:23:03 yeah, restarted after 5-10 second press May 31 18:23:33 Herrie: that's easy, just test /dev/input/event* and set it up for nyx May 31 18:23:37 now stuck in bootloop May 31 18:23:45 same as before May 31 18:23:53 BT40: I thought you flashed the kernel ? May 31 18:24:08 yes i flashed May 31 18:24:17 that's a bit weird May 31 18:24:27 will attempt once again May 31 18:24:39 yes, go to fastboot, and retry May 31 18:24:46 Well you can at least pull logs now I guess from TwRp May 31 18:25:04 Herrie: hehe, no, our /var/log is volatile by default May 31 18:25:15 thus me saying there isn't an easy way May 31 18:25:55 He'll have to create /data/luneos/var/log/journal directory from twrp May 31 18:26:35 (and even so, I think we mount a tmpfs on top of /var/log, so...) May 31 18:26:54 ah, no, we have /var/log -> /run/log, and /run is a tmpfs May 31 18:27:06 something like that May 31 18:27:46 so it's "rm /data/luneos/var/log ; mkdir -p /data/luneos/var/log/journal" May 31 18:28:08 but let's get back to booting the kernel May 31 18:28:37 fastboot: usage: expected argument............this is result of fastboot boot May 31 18:28:40 strange May 31 18:28:55 although flashing is complete May 31 18:29:32 do fastboot flash boot image.fastboot May 31 18:30:19 Tofe: shouldn't the console-ramoops be there anyway? May 31 18:30:20 then fastboot boot image.fastboot May 31 18:30:33 Herrie: ah, probably yes May 31 18:31:19 let's boot once more, do the crash and retrieve pstore May 31 18:33:18 again booted May 31 18:33:28 adb devices do not return anything May 31 18:33:45 usb is connected i make sure :) May 31 18:33:54 ok, let's forget about adb for now May 31 18:34:32 make the welcome app crash again, and then force the device to reboot to twrp May 31 18:34:59 welcome app is gone May 31 18:35:12 any other app i open, say messaging or clock May 31 18:35:58 when i click on it, it dissapears. screen goes off. when i press power button, screen wakes with app open. after one second, screen gets off again and loop continues May 31 18:36:11 ok, reboot to twrp :) May 31 18:36:52 volume buttons are working May 31 18:37:02 that's surprising. good! May 31 18:37:02 booted into twrp May 31 18:37:28 ok, now in adb, check /sys/fs/pstore May 31 18:37:36 pstore is empty this time although May 31 18:37:40 ok May 31 18:37:45 well, kernel didn't crash May 31 18:37:53 seent throuch file manager of twrp May 31 18:38:00 still in adb, do "rm /data/luneos/var/log ; mkdir -p /data/luneos/var/log/journal" May 31 18:38:59 oops, you should probably mount data first, if it's not already the case May 31 18:39:37 data is mounted May 31 18:39:44 command has been run May 31 18:40:02 ok, now reboot the device normally May 31 18:41:51 when in luneos, start an app or two, then reboot again to twrp May 31 18:42:04 interesting, does fingerterm crash as well? May 31 18:42:47 nizovn: from his feedback, apps aren't all crashing, it's the screen that gets black too often May 31 18:43:16 ok May 31 18:43:25 yupp May 31 18:43:44 also a beep is comming continuously when screen goes onand off May 31 18:43:52 I had the bug with the screen, but I think I never really solved it. It disappears if we plug off the usb cable and plug it in again, iirc May 31 18:43:56 Bt40 that is charging May 31 18:44:18 Tofe: might be OSE migration related May 31 18:44:21 i have removed cable, still this behavior May 31 18:44:29 Herrie: no, I had it on rosy long ago May 31 18:44:30 I also have screen turning off May 31 18:44:45 Tofe: I mean OSE not ACG May 31 18:44:55 BT40: ok reboot to twrp, let's have a look at those logs May 31 18:45:06 There is some issue in ActvityManager etc I think May 31 18:45:06 Herrie: oh ! maybe then yes May 31 18:46:16 both folders are emply. journal as well as parent log May 31 18:46:33 BT40: nothing in /data/luneos/var/log/journal then ? May 31 18:46:42 yes, nothing May 31 18:46:55 ok, I must have forgotten a step, let me see May 31 18:49:09 * BT40 >> nizovn is like a sniper soldier. Rarery fires (speaks), and the single bullet hit the target. :) May 31 18:50:02 BT40: if you do ls -ld /data/luneos/var/log, what do you see ? is it a symlink, or a folder ? May 31 18:51:19 * BT40 feels nizoadb shell May 31 18:51:20 land:/ # ls -ld /data/luneos/var/log May 31 18:51:20 drwxrwxrwx 3 root root 4096 2016-01-02 05:15 /data/luneos/var/log May 31 18:51:29 mmmmh May 31 18:51:30 ooooooh May 31 18:52:04 it's was /data/luneos-data May 31 18:52:25 BT40: can you do "rm /data/luneos-data/var/log ; mkdir -p /data/luneos-data/var/log/journal" May 31 18:53:29 (we never touch /data/luneos, so we mount /data/luneos-data/* on top of the directories we want to write to) May 31 18:54:52 rm: /data/luneos-data/var/log: No such file or directory May 31 18:54:52 mkdir: cannot create directory ‘/data’: Permission denied May 31 18:55:21 you're in adb shell in twrp, right ?... not on your computer ? :) May 31 18:55:36 i was on computer sorry :) May 31 18:55:41 will try in twrp May 31 18:55:54 Herrie: for adb it's probably a missing gadget in defconfig May 31 18:56:09 we use "adb on /dev/usb-ffs/adb type functionfs" May 31 18:58:11 Herrie: yes, nearly no gadget is enabled May 31 18:58:53 it says no such file or directory May 31 18:59:05 for rm command May 31 18:59:43 Let me see if I can find this defconfig option May 31 18:59:53 Should be functionfs I think? May 31 19:00:20 https://github.com/Tofee/android_kernel_xiaomi_msm8953/blob/halium-7.1/arch/arm64/configs/lineageos_rosy_defconfig#L3340 the whole section here May 31 19:01:02 BT40: I guess you did "adb shell rm /data/luneos-data/var/log ; mkdir -p /data/luneos-data/var/log/journal" previously May 31 19:01:15 so only the first command was done inside adb May 31 19:01:51 yes May 31 19:02:04 I would recommand not doing like this for adb commands: do a simple "adb shell", then when on the prompt of the shell inside adb, do the commands. That way you're sure where things go May 31 19:02:44 okay got it May 31 19:02:51 anyway, I think you now do have a correct /data/luneos-data/var/log/journal May 31 19:03:43 reboot to luneos May 31 19:04:09 * Tofe bets that once we succeed getting the log, Herrie will have a kernel ready where adb works... May 31 19:04:49 boot complete... im in luneos May 31 19:04:59 Tofe: Well most relevant bits seem to be in defconfig already May 31 19:05:10 BT40: ok, do a thing or two, then reboot to twrp May 31 19:05:28 Herrie: also the G_ANDROID one ? May 31 19:06:54 Herrie: damn, you're right, I didn't spot it earlier May 31 19:08:24 done, into recovery mode May 31 19:08:41 ok, do you have content in /data/luneos-data/var/log/journal ? May 31 19:08:58 yes May 31 19:09:03 a long filename May 31 19:09:03 good ! then pull the whole folder May 31 19:10:42 on the pc, do journalctl --no-pager --directory="path of your pull directory" May 31 19:10:55 pulled* May 31 19:11:34 if you just pulled the file, there is also an option --file= which will work too May 31 19:12:50 then paste the whole thing somewhere, and we'll see if something interesting is logged in there May 31 19:13:56 terminal is out of scroll. so many lines May 31 19:14:08 can i write this directly to a text file? May 31 19:14:27 output of terminal to a text file May 31 19:14:39 with > May 31 19:15:02 journalctl --no-pager --directory="path of your pull directory" > my_log.txt May 31 19:19:57 2%] /data/luneos-data/var/log/journal/f9a531d8a2d243b8914e445edd1735e1/system@0005a53b52f4a634-e1d2e1faf06a243f.journal~: 3% May 31 19:20:19 these contents of file are different from terminal one May 31 19:20:37 no useful info in file, terminal info was useful one May 31 19:20:52 in file, only fetching progress is there May 31 19:21:44 that's weird May 31 19:23:09 BT40: you can also try with --file May 31 19:23:27 or, can you share the journal over dropbox, I can try on my side May 31 19:23:37 sure May 31 19:23:56 also, don't forget it's the middle of the night on your side :) May 31 19:25:55 https://www.dropbox.com/s/uane2jd0qysx3rn/luneos-first-log.txt?dl=0 i copied some contents from terminal and manually saved in text file. May 31 19:26:26 Ah, the journal file as-is would have been fine too May 31 19:26:45 https://www.dropbox.com/s/jhj54ha2w73qhpa/system.journal?dl=0 Journal file May 31 19:27:16 1 am ...... night :) May 31 19:27:53 whats time your side Tofe May 31 19:28:04 21:30PM May 31 19:28:07 hope i'm not disturbing others May 31 19:28:24 Everyone's free to come and go :) May 31 19:30:19 https://www.dropbox.com/s/q5znsq2pvm54i4g/land-journal.txt?dl=0 here is the text version of the whole journal May 31 19:34:32 msm-thermal: probe_vdd_rstr:Failed reading node :) May 31 19:35:05 will have a look on these tomorrow..... feeling tired now. Eye strain May 31 19:35:19 got to have a nap May 31 19:35:21 :) sure May 31 19:35:22 :) May 31 19:35:52 Thank you Herrie Tofe nizovn JaMa . May 31 19:38:39 setup[965]: /usr/bin/android-gadget-setup: line 27: /sys/class/android_usb/android0/iSerial: No such file or directory <-- well, ok. but why... May 31 19:42:41 Herrie: I think we need to activate CONFIG_USB_FUNCTIONFS May 31 19:43:13 Tofe: could be May 31 19:43:38 otherwise "grep -q functionfs /proc/filesystems" will probably fail May 31 19:44:09 I might be lucky on rosy about not needed the few lines of the android-gadget-setup script after that "if" May 31 19:51:46 We should definitely change https://github.com/kskarthik/android_kernel_xiaomi_msm8937/blob/cm-14.1/drivers/power/qpnp-smbcharger.c#L12 into just "#define pr_fmt(fmt)" May 31 19:51:59 it's totally polluting the logs May 31 19:59:07 LunaSysService[1262]: terminate called after throwing an instance of 'std::length_error' oops. May 31 19:59:18 ok, but overall it's not that bad... **** ENDING LOGGING AT Mon Jun 01 02:59:57 2020