**** BEGIN LOGGING AT Tue Jan 09 03:00:02 2018 Jan 09 03:16:41 doniks: did the test of upgrading libhybris work? Jan 09 06:53:59 @bhushanshah https://github.com/Halium/docs/pull/49 Jan 09 06:54:12 Why the check is failed? Jan 09 06:55:15 nvm, I figured it. I closed it and doing another PR. Jan 09 07:12:51 https://github.com/Halium/docs/pull/50 Jan 09 07:12:58 Check this now Jan 09 08:44:09 @Mr_Doge let's continue Jan 09 08:45:08 After you have made the kernel and systemimage stuff... We will manually change the mountpoint for the user data Jan 09 08:45:23 For that you will have to install abootimg Jan 09 08:46:13 @vanyasem do you have android on your bullhead? Jan 09 08:46:28 plasma mobile Jan 09 08:46:38 that's a no Jan 09 08:47:15 right I need android Jan 09 08:47:29 I got LOS rn. Jan 09 08:47:34 i can install it in an hour if you want me to Jan 09 08:48:18 @theimpulson can you give me `logcat -b radio` in PM? (Don't paste it here, as it generally known to contain IMEI number) Jan 09 08:48:35 @bhushanshah, Sure. Just a sec Jan 09 08:52:11 @bhushanshah, Done Jan 09 08:52:51 thanks Jan 09 08:53:15 👍🏻 Jan 09 09:05:01 @Mr_Doge, means I will be forced to change entries anyhow? Jan 09 09:07:03 @theimpulson, You can go docs route too... But that's confusing and messes with the stock stuff... Jan 09 09:08:00 @theimpulson, I'll suggest you to leave that as it is.. Jan 09 10:20:47 https://paste.ubuntu.com/26352305/ Jan 09 10:20:50 ?? Jan 09 10:21:34 @Mattia, not working Jan 09 10:21:37 @Mr_Doge, @SALPHA8506 Jan 09 10:21:53 no Jan 09 10:22:22 You forgot && Jan 09 10:22:58 EGL_PLATFORM=hwcomposer && test_hwcomposer Jan 09 10:23:26 🙈 Jan 09 10:23:31 @theimpulson, Does this work for you without root privileges Jan 09 10:24:31 I guess it was meant to run as root 🤷‍♂ Jan 09 10:24:32 @Satyam2345, https://paste.ubuntu.com/26352317/ Jan 09 10:24:57 Anything show up on phone ? Jan 09 10:25:02 Is your device CAF? Jan 09 10:25:09 @Satyam2345, Never worked for me on halium Jan 09 10:25:26 @bhushanshah, yes Jan 09 10:27:13 https://paste.ubuntu.com/26352322/ Jan 09 10:27:18 @SALPHA8506, mount Jan 09 10:28:35 no /vendor mounted, but when i mount system.img with twrp /vendor is existing... Jan 09 10:28:53 @vanyasem, @SALPHA8506 Jan 09 10:29:40 @theimpulson, its done.. Jan 09 10:29:49 @SALPHA8506, Upgraded ? Jan 09 10:29:57 @Satyam2345, yes Jan 09 10:30:59 any way to mount /vendor...? Jan 09 10:36:05 It should be mounted or symlinked already Jan 09 10:37:06 @SALPHA8506 check your fstab Jan 09 10:37:24 it was incorrect for my device so I had to manually fix it Jan 09 10:37:35 @vanyasem, where..? Jan 09 10:37:39 you can locate it with grep Jan 09 10:37:51 or actually any other searching tool Jan 09 10:39:11 @vanyasem, how to use ?? Jan 09 10:39:32 how to use Google/DuckDuckGo?? Jan 09 10:39:59 open the folder in your graphical file manager and use it's built-in search function Jan 09 10:40:09 usually activated by Ctrl+F Jan 09 10:40:21 it will do, but it's slower than grep, so be patient Jan 09 10:41:17 look for `fstab`. the file is usually called `fstab.qcom` Jan 09 10:42:20 this is in device/zuk/msm8996/rootdir/etc👈 Jan 09 10:43:30 https://paste.ubuntu.com/26352375/ Jan 09 10:43:35 @SALPHA8506, ?? Jan 09 10:44:21 How build halium with GNU male? Jan 09 10:44:54 @SALPHA8506, run `ls -ld /vendor` on your device Jan 09 10:45:18 root@localhost:~# ls -ld /vendor … lrwxrwxrwx 1 root root 14 Feb 11 2016 /vendor -> /system/vendor … root@localhost:~# Jan 09 10:46:10 well so /vendor is not issue Jan 09 10:46:26 can you run `apt-cache policy libhybris-tests` Jan 09 10:46:51 root@localhost:~# apt-cache policy libhybris-tests … N: Unable to locate package libhybris-tests Jan 09 10:47:03 err Jan 09 10:47:04 wait Jan 09 10:47:52 make it libhybris-test insstead of libhybris-tests Jan 09 10:49:29 https://paste.ubuntu.com/26352395/ Jan 09 10:49:31 @ErnyTech, Anybody can teach me? Jan 09 10:49:49 @SALPHA8506, ? Jan 09 10:50:36 @vanyasem, . Jan 09 10:50:41 You haven't done this yet Jan 09 10:51:01 @SALPHA8506, Also, don't send another message with just question mark, it serves no purpose Jan 09 10:51:20 @ErnyTech, Patience, ;) … What issues you're having? Jan 09 10:53:39 @bhushanshah, it's already done Jan 09 10:54:32 Well I'm telling you it's not done yet Jan 09 10:54:49 https://paste.ubuntu.com/26352411/ Jan 09 10:55:11 Ah Jan 09 10:55:27 You missed `>>` in the command Jan 09 10:55:28 @ErnyTech, Have you read the docs? .. follow the steps.. and tell us where you are having the problems Jan 09 10:56:09 @Mr_Doge I love the fact I told you this a week ago and now you're teaching newcomers. you're a quick learner Jan 09 10:56:13 Ah the irony at its best Jan 09 10:56:22 » where i use Jan 09 10:56:26 😂 Jan 09 10:56:53 @SALPHA8506, it should be `>>` not » Jan 09 10:58:34 @SALPHA8506, For future : >> is used when you want the write a string to the end of the file .. where the string which is defined earlier in … ```echo "*"``` Jan 09 11:06:28 https://paste.ubuntu.com/26352448/ Jan 09 11:06:55 EGL_PLATFORM=hwcomposer && is wrong Jan 09 11:07:24 you should be using `EGL_PLATFORM=hwcomposer test_hwcomposer` Jan 09 11:08:00 same Jan 09 11:08:31 well, fetch logcat while running tests then Jan 09 11:09:20 @bhushanshah, logcat how? Jan 09 11:10:39 not stopping in terminal Jan 09 11:10:56 Well, that is point Jan 09 11:11:03 let it run, and form another terminal run tests Jan 09 11:11:07 copy new output Jan 09 11:11:49 @bhushanshah, ? Jan 09 11:13:21 don't you really know how to copu? Jan 09 11:13:22 @ShrugBot Jan 09 11:13:43 ¯\_(ツ)_/¯ Jan 09 11:14:05 wait Jan 09 11:17:32 https://paste.ubuntu.com/26352609/ Jan 09 11:19:51 Can you strace test_hwcomposer? `EGL_PLATFORM=hwcomposer strace test_hwcomposer` ? Jan 09 11:20:16 with logout?? Jan 09 11:20:26 ? Jan 09 11:20:31 oh actually Jan 09 11:20:50 run it like `EGL_PLATFORM=hwcomposer strace -o logfile test_hwcomposer` Jan 09 11:21:22 logfile Jan 09 11:22:30 ? Jan 09 11:23:09 where is logfile? Jan 09 11:37:41 @Mr_Doge, Yep but they do not explain how to compile with gnu make, they only explain mkbooting Jan 09 11:38:26 They explain everything you need to get started Jan 09 11:38:42 @ErnyTech, you should probably message Google and ask why is android compiled that way and not another Jan 09 11:39:00 @ErnyTech, I think you misunderstood the project halium.. Jan 09 11:39:33 @Mr_Doge, Why? I prefer use GNU make for build Jan 09 11:39:47 well Jan 09 11:39:57 mka is 1:1 alternative to make Jan 09 11:40:09 so instead of mka you can just run make bootimage and it will work Jan 09 11:40:22 @bhushanshah, mka is optimized for faster compile times Jan 09 11:40:28 just to clarify that Jan 09 11:40:33 well sure Jan 09 11:40:48 @ErnyTech, But ... … ^^^ Jan 09 11:41:05 @bhushanshah, An for hybris-boot just make hybris-boot ? Jan 09 11:41:11 yes Jan 09 11:41:17 Fwd from bhushanshah: mka is 1:1 alternative to make Jan 09 11:41:19 Thanks Jan 09 11:41:35 you will need to manually specify an amount of threads though `-j [NUM]` Jan 09 11:41:42 well i assume you know how make works Jan 09 11:41:44 @vanyasem, I know Jan 09 11:41:48 you don't need to do that with `mka` Jan 09 11:41:51 it's optimized already Jan 09 11:42:13 so there's really no point in using make over mka Jan 09 11:43:09 @vanyasem, +1 Jan 09 11:43:34 @vanyasem, +2 Jan 09 11:44:27 @vanyasem, I can't use in my build server Jan 09 11:44:41 oh, makes sense. you should've started with that Jan 09 12:53:04 Hi! Have i to enable all missing(yellow) kernel feature that i found with lxc-checkconfig? Jan 09 12:55:03 @Mattia, Except user namespace enable everything. Jan 09 12:57:00 And where i can find the acronymous to insert in kernel conf? Jan 09 12:58:20 @Mattia, What flags are missing? Jan 09 12:59:27 @theimpulson https://pastebin.com/stVUwqVD Jan 09 13:00:40 Fwd from theimpulson: https://github.com/Halium/android_kernel_asus_msm8916/blob/halium/arch/arm64/configs/Z00L_defconfig Jan 09 13:01:40 @mariogrip are you going to continue halium for oneplus3? … https://github.com/Halium/android_device_oneplus_oneplus3?files=1 Jan 09 13:02:42 @theimpulson, Mattia I had everything enabled. See my defconfig for missing flags. Jan 09 13:03:29 @theimpulson, Ok, thanks! Jan 09 13:12:37 @Mattia, Done? Jan 09 13:15:34 No, my defconfig is long 2000+ lines and yours is 719... i used a text comparator but there are a lot of missing voices @theimpulson Jan 09 13:19:10 Ok i used this "tool" to see what i have to enable https://raw.githubusercontent.com/dotcloud/docker/master/contrib/check-config.sh Jan 09 13:23:58 @Mattia, Wait a min. I will give specific flags then. Jan 09 13:26:45 @Mattia, https://pastebin.com/514jAFAz Jan 09 13:28:41 @theimpulson, Thanks a lot! I'm trying with yours Jan 09 13:34:46 👍 Jan 09 13:39:56 @SALPHA8506, https://paste.ubuntu.com/26353149/ Jan 09 13:40:02 2 solved Jan 09 13:40:16 libsdmextension.so ?? Jan 09 13:41:57 @bhushanshah, Still need this? Jan 09 13:42:23 no Jan 09 13:42:28 I fixed rild \o/ Jan 09 13:44:03 @bhushanshah, From my log? Jan 09 13:44:17 no it was totally different issue Jan 09 13:44:41 Oh! Well, if ril starts working I will use PM on a daily basis. Jan 09 13:44:51 Though GUI needs fix too. Jan 09 13:50:52 @theimpulson some other configs are missing but now android container is active but with some errors … https://pastebin.com/9zRxCqqe Jan 09 13:52:09 @Mattia, Did you added all flags I provided and checked your defconfig if they are not repeated? Jan 09 13:55:52 Yess, they aren't repeated @theimpulson Jan 09 13:56:46 @theimpulson, its laggy right ? Jan 09 13:56:59 @Mr_Doge, And distorted Jan 09 13:57:43 @Mattia, I guess you need some extra flags then. Happend with me for memory controller and I got it enabled using some additional flags. Jan 09 13:58:33 @theimpulson, So how i can check it?? Jan 09 14:07:35 @Mattia, Can you share the lnk of your current source? Jan 09 14:10:45 @theimpulson Here you are: https://pastebin.com/3iB0uByY Jan 09 14:13:58 @Mattia, I can actually see that some flags are missing. Jan 09 14:14:21 You sure you added all? Jan 09 14:16:28 they are in the end of file Jan 09 14:17:03 @Mattia, I guess the file didn't get copied in a proper way when you pasting. Jan 09 14:19:44 Firstable i checked if i got that configs in defconfig and then i added one by one manually @theimpulson Jan 09 14:28:36 @Mattia, Have you enabled kernel modules? Jan 09 14:39:28 @Mattia, try enabling CONFIG_NETLINK_DIAG=y Jan 09 14:45:32 @theimpulson, When i'll got home i'll try, thanks! Jan 09 15:02:59 root@localhost:~# EGL_PLATFORM=hwcomposer test_hwcomposer … cannot locate symbol "_ZN3sdm3Sys7access_E" referenced by "/system/vendor/lib/libsdmextension.so"... … cannot locate symbol "_ZN3sdm3Sys7access_E" referenced by "/system/vendor/lib/libsdmextension.so"... … cannot locate symbol "_ZN3sdm3Sys7access_E" referenced by "/system/vendor/lib/libsdmextension.so"... … cannot locate symbol "_ZN3sdm3Sys7access_E" referenced by Jan 09 15:03:00 "/system/vendor/lib/libsdmextension.so"... Jan 09 15:03:07 any solution for this..? Jan 09 15:30:09 @theimpulson, @theimpulson it's already active (it's the last config) Jan 09 15:40:51 @Mattia, I am out of ideas right now also I am hungry. Maybe @vanyasem might help you. Jan 09 15:42:51 paste `systemctl` output @Mattia Jan 09 15:43:04 not lxc's status, but just plain systemctl Jan 09 15:43:13 i need to see if lxc was fully started Jan 09 15:45:29 Thanks Aayush. … @vanyasem Here you are: https://pastebin.com/GuBcKnn3 … Thanks for help! Jan 09 15:46:25 i gave you the wrong command Jan 09 15:46:32 i actually need `systemctl status` Jan 09 15:46:41 eh, too sleepy as well Jan 09 15:48:21 @vanyasem Ahahah no problem! … Here you are: https://pastebin.com/4fb3MB0H Jan 09 15:50:44 and what is your last defconfig? Jan 09 15:52:56 # lxc-info -nandroid … Name: android … State: RUNNING Jan 09 15:53:08 That was so much easier making ubports-boot do it Jan 09 15:53:09 Kill healthd from your init*.rc Jan 09 15:53:24 @UniversalSuperBox :O \o/ Jan 09 15:53:41 Mounts are all screwed, of course. systemd tries to mess with them. Jan 09 15:56:20 @vanyasem Part1: https://pastebin.com/wjuqpp1y … Part2: https://pastebin.com/097cwCJS Jan 09 15:57:47 @bshah, Mattia do what Bhushan said. don't listen to me, i'm stupid Jan 09 16:00:08 I'm missing an important step where everything in `/android` gets bind-mounted to the lxc rootfs Jan 09 16:00:57 @bshah, How i can do it? Jan 09 16:02:53 well just open your init*.rc with a text editor and comment the part about healthd Jan 09 16:03:08 (from your device tree) Jan 09 16:07:34 i've tried with init.qcom.rc , init.target.rc but there isn't the word healtd Jan 09 16:07:54 healthd Jan 09 16:07:56 try running a grep search on your tree Jan 09 16:07:58 not healtd Jan 09 16:08:10 (unless you made typo just here) Jan 09 16:08:27 i assume that's a typo in a message, not in the search Jan 09 16:13:41 No i have mistyped. Also with grep, i got nothing Jan 09 16:16:37 @saidinesh5 said this: … Yes, I've had that issue.. could be a bad build directory or build failing somewhere and it's using the older configuration or maybe some depending kernel configurations are not available to enable a feature … 3.4 kernel is not good for lxc Jan 09 16:17:19 Pretty much sure your issue is not kernel Jan 09 16:17:27 Can you search for charger service? Jan 09 16:18:48 in "device/xiaomi/cancro" ? or in halium rootdir?? Jan 09 16:19:02 cancro Jan 09 16:20:30 Here you are: https://pastebin.com/YfwHrLbE Jan 09 16:24:02 kill /charger and friends service from init.qcom.rc Jan 09 16:29:30 @bshah it is correct now? https://pastebin.com/3e8qWb6F Jan 09 16:30:24 @Mattia also disable disable qbcharger just in case Jan 09 16:30:47 ```130|android:/ # whoami … /system/bin/sh: whoami: not found``` Jan 09 16:30:53 We're in! Jan 09 16:31:09 yay Jan 09 16:31:35 There's no PATH on the Android commandline, so I'm using `source ` as a makeshift ls Jan 09 16:32:39 @bshah It is ok now? https://pastebin.com/mMmwaHtV Jan 09 16:33:14 L912 and L913 needs commenting Jan 09 16:33:17 but looks good Jan 09 16:34:25 @bshah sorry, my mistake.. now i can build it? Jan 09 16:34:46 yeah sure Jan 09 16:34:48 go ahead Jan 09 16:37:30 Thanks @bshah. I will keep you updated Jan 09 16:48:24 @bshah same problems Jan 09 16:48:38 systemctl status please Jan 09 16:50:36 @bshah https://pastebin.com/4YRQhheE Jan 09 16:51:23 @Mattia: can you upload your /system/android-ramdisk.img somewhere? Jan 09 16:51:26 err Jan 09 16:51:36 /system/boot/ is correct path Jan 09 16:52:10 @Mattia, btw, if you find a fix for android not starting fully, I would like to try it on my device too :) Jan 09 16:52:49 JBBgameich: your issue is something-binder related btw Jan 09 16:53:47 On the phone, in that folder i've got only android.ramdisk.img @bshah Jan 09 16:53:57 That's right Jan 09 16:54:17 @bshah, oh. The symptoms I saw was just android init crashing at some point, but it's good to know that it was a binder issue triggering it. Jan 09 16:54:34 yeah, that, can you upload it somewhere? (on your host it would be on $OUT/system/boot/) Jan 09 16:54:57 One second @bshah Jan 09 16:58:06 Okay, so it does require a couple of changes to userland bshah. Removing system.mount, disabling android-mount.service, making udev start after lxc@android, and a bit of a change to pre-start.sh Jan 09 16:58:36 Can you document it on projectmanagement? Jan 09 16:58:43 Sure yep Jan 09 17:04:41 @bshah Here you are: https://drive.google.com/open?id=1RWKJUdoOqYTZTZel4vhPxct9fBKv-mfd Jan 09 17:06:50 @Mattia: question, what exactly you did rebuild/install? Jan 09 17:07:01 (Photo, 723x168) https://irc.ubports.com/0jaNX8LV/file_3671.jpg nmcli is mocking me. ._. Jan 09 17:07:44 Ah, wcnss is dying as usual Jan 09 17:07:47 @bshah, I get that image from the phone because i haven't rebuilded the system, only hybris-boot Jan 09 17:09:09 Well, you should be rebuilding system.img and use replace-android-image script to install it Jan 09 17:09:21 do it and all your issues are fixed Jan 09 17:11:54 anyway I will go to sleep early today.. because I think I finally got rid of my insomnia today Jan 09 17:12:20 Go for it! Jan 09 17:24:57 @bshah, Same problems...😭 Jan 09 17:25:06 @bshah, Good Night! Jan 09 17:27:22 lxc-start: start.c: preserve_ns: 153 No such file or directory - Kernel does not support attaching to mnt namespaces. … What config is missing??? Jan 09 18:14:00 BottomNotch was added by: BottomNotch Jan 09 19:08:22 @bhushanshah, haven't tried yet, got distracted running after another butterfly https://github.com/JBBgameich/halium-install/pull/9 Jan 09 19:08:32 I'll report back when I did Jan 09 19:16:54 root@localhost:~# EGL_PLATFORM=hwcomposer test_hwcomposer … cannot locate symbol "_ZN3sdm3Sys7access_E" referenced by "/system/vendor/lib/libsdmextension.so"... … cannot locate symbol "_ZN3sdm3Sys7access_E" referenced by "/system/vendor/lib/libsdmextension.so"... … cannot locate symbol "_ZN3sdm3Sys7access_E" referenced by "/system/vendor/lib/libsdmextension.so"... … cannot locate symbol "_ZN3sdm3Sys7access_E" referenced by Jan 09 19:16:55 "/system/vendor/lib/libsdmextension.so"... … width: 1080 height: 1920 … OpenGL ES 2.0 (OpenGL ES 3.2 V@145.0 (GIT@I9c435c2712)) … ?? Jan 09 19:40:39 https://paste.ubuntu.com/26354991/ Jan 09 19:53:19 You may need to either find out what lib that lib needs, or make a shim for it. Jan 09 19:57:13 @r3pwn, And how to say to an binary t o look at an other path for an library ? as lib64 and not lib ;) Jan 09 19:58:02 @BlueKenny, They should both be in the library path, so it shouldn't have to look in a specific location for them. Jan 09 20:00:58 @r3pwn, oh you are right, it wasn't in lib64 :/ Jan 09 20:04:23 @r3pwn, How to find there are many? Jan 09 20:06:10 @JBBgameich, regarding your UT support PR for halium-install, I do indeed get the same error with or without that PR Jan 09 20:08:16 Maybe your device doesn't like ext4 (at least not the version my system uses) Jan 09 20:08:53 Hopefully I can fix this somehow, maybe just using ext2 again, but ext2 breaks more often Jan 09 20:11:00 ext4 for the ubuntu touch rootfs is a waste of time Jan 09 20:11:16 yeah, looks like it doesn't like SOMETHING https://paste.ubuntu.com/26355170/ Jan 09 20:11:25 EXT4-fs (loop0): error loading journal Jan 09 20:11:33 EXT4-fs (loop0): couldn't mount RDWR because of unsupported optional features (400) Jan 09 20:11:53 @UniversalSuperBox, because it's read-only? Jan 09 20:11:57 Yep Jan 09 20:12:16 Plasma Mobile and reference are rw though, so I used ext4 there Jan 09 20:12:40 Sounds like that's causing issues too, which in hindsight is probably why ext2 was kept in the first place. :( Jan 09 20:13:54 Maybe I can create a more minimal ext4 filesystem without optional features, but I need to find out how optional they are first Jan 09 20:20:57 Would be interesting to know which feature actually is unsupported. I can remove some of them in the script, so if you find out which one I have to remove I can fix that Jan 09 20:21:23 mayyyybee it's also just a kernel config issue ... Jan 09 20:23:23 https://tosiek.pl/ext4-fs-sda-couldnt-mount-rdwr-because-of-unsupported-optional-features-400/ Jan 09 20:23:39 I put it on the list, but at the bottom tbh :) Jan 09 20:23:52 I found this one already. Thanks anyway :) Jan 09 20:24:05 trying it with ext2 now Jan 09 20:25:23 The optional features used by a modern debian sid system are: … ```has_journal ext_attr resize_inode dir_index filetype extent 64bit flex_bg sparse_super large_file huge_file dir_nlink extra_isize metadata_csum``` Jan 09 20:26:10 one or multiple of them can be / should be removed Jan 09 20:27:37 doniks: Which kernel version does your device use? Jan 09 20:28:55 mkfs.ext4 test.img 100M ; tune2fs -l test.img | grep feature Jan 09 20:29:01 gives me the same features Jan 09 20:29:08 on U 17.10 Jan 09 20:29:13 3.4 kernel Jan 09 20:29:20 ^ on the device Jan 09 20:29:31 Can you try to create an image on the device (in recovery or something) and send me the features it supports? Jan 09 20:29:48 jup Jan 09 20:29:53 So we can compare which ones it doesn't support that debian and ubuntu 17.10 have Jan 09 20:30:14 (booting the ext2 version ....) Jan 09 20:35:22 @JBBgameich, there is no mkfs.ext4 on the device :( Jan 09 20:36:53 is something like this command available? `mke4fs -t ext4` Jan 09 20:38:07 mkdir mkdosfs mke2fs mkfifo mkfs.ext2 mkfs.vfat mknod mkswap mktemp Jan 09 20:38:22 ah but now I can chroot into the ext2 rootfs Jan 09 20:38:36 INSIDE I have have mkfs.ext4 Jan 09 20:39:24 @JBBgameich does your script mess with system.img at all? Jan 09 20:40:04 it only uses simg2img to make it mountable Jan 09 20:40:20 Huh. I can't mount mine. Let's see if I can use the replace script on it. Jan 09 20:41:24 Same error as doniks, actually Jan 09 20:41:36 has_journal ext_attr resize_inode dir_index filetype extent flex_bg sparse_super large_file huge_file uninit_bg dir_nlink extra_isize Jan 09 20:41:37 but with the system.img? Jan 09 20:41:43 thank you Jan 09 20:41:44 But with the system.img Jan 09 20:41:56 features INSIDE ext2 rootfs chroot Jan 09 20:42:22 https://paste.ubuntu.com/26355366/ Jan 09 20:42:27 It ends up getting real upset Jan 09 20:42:59 Actually sounds like it got corrupted and wanted to replay a journal or something... Jan 09 20:43:38 device misses: 64bit metadata_csum, adds: uninit_bg Jan 09 20:44:01 Ok, found out the same Jan 09 20:44:10 will try to remove both features from the rootfs.img Jan 09 20:44:13 and system.img Jan 09 20:44:32 Pffft Jan 09 20:44:51 Of course I spend 30 minutes troubleshooting why it won't mount, just to find out that the image was corrupt. -_- Jan 09 20:45:26 @UniversalSuperBox, THIS! I think we should teach initrd to detect when that fails and go into telnet Jan 09 20:45:53 FWIW, you don't seem to have that EXACT same error as I did Jan 09 20:46:02 Would be easy with ubports-boot. ;P Jan 09 20:46:07 I'll see myself out Jan 09 20:46:52 @UniversalSuperBox, I see no reason why it woudn't be easy with hybris-boot :) Jan 09 20:46:56 And we're back. Happy rootfs, happy container, ubports-boot Jan 09 20:47:35 Now, to wipe one more time and test the steps again Jan 09 20:53:56 So, @JBBgameich, interesting development... Pushing system.img is near-instant Jan 09 20:54:15 Which indicates failure to me Jan 09 20:54:29 Important question: is it empty or not :/ ? Jan 09 20:54:37 Still pushing rootfs Jan 09 20:54:51 -rw-rw-rw- 1 root root 0 Jan 9 2018 system.img Jan 09 20:54:53 whelp Jan 09 20:55:06 does your device use sparse system images at all? If not, it's the first I see not doing that, but I could probably fix the script Jan 09 20:55:23 It does use sparse Jan 09 20:57:30 @SALPHA8506, EGL version string is a nice thing to have :) Jan 09 21:19:09 doniks: I pushed a patch for the ext4 problem, please try it out if you have some free time: https://github.com/JBBgameich/halium-install/pull/11 Jan 09 21:25:41 Guys can someone explain how these 2 builds have so much difference in load offsets? Jan 09 21:25:43 (Photo, 1229x366) https://irc.ubports.com/H7WcD0Na/file_3677.jpg Jan 09 21:26:02 7.1 works, 5.1 doesnt Jan 09 21:27:09 You could compare the values of both device trees in BoardConfig.mk Jan 09 21:27:23 true. damn Jan 09 21:29:37 Alright, anyone who wants to test ubports-boot can and should: https://github.com/Halium/projectmanagement/issues/40 Jan 09 21:32:07 @JBBgameich, Seems identical Jan 09 21:33:18 Haha its running with hybris-boot from 7.1 Jan 09 21:34:02 container is up... Jan 09 21:34:55 test_egl_config works Jan 09 21:37:14 Ha test_hwcomposer does not crash, but also no output: … E/qdoverlay( 0): Bad ov dump: mdp_overlay z=0 fg=0 alpha=255 mask=-1 flags=0x220000 id=-1 … E/qdoverlay( 0): src msmfb_img w=736 h=1280 format=13 MDP_RGBA_8888 Jan 09 21:37:18 Any ideas ^^ Jan 09 21:37:19 Pro Jan 09 21:37:46 Not so pro Jan 09 21:39:48 @Flohack try to use mm.so (device-files/common) from ubports Jan 09 21:40:18 thats work for me Jan 09 21:40:56 @Walid, where is that located 😆 Jan 09 21:41:13 device-files/common/system/usr/lib/arm-linux-gnueabihf/libhybris/linker Jan 09 21:41:45 copy device-files/common/system/usr/lib/arm-linux-gnueabihf/libhybris/linker folder in your rootfs Jan 09 21:41:53 GitHub.com/ubports/device-files Jan 09 21:42:09 dont have device files folder at all ^^ Jan 09 21:42:18 No, you don't Jan 09 21:42:23 @UniversalSuperBox, It's here Jan 09 21:42:27 ajhhhhh Jan 09 21:42:49 Just put the mm.so from there at /use/lib/arm-whatever walid said Jan 09 21:43:01 the magic one Jan 09 21:45:10 scp would be grand inside rootfs Jan 09 21:45:30 neither ftp available hmm Jan 09 21:45:57 use adb Jan 09 21:46:17 mount your rootfs, and adb push Jan 09 21:46:37 I know but involves reboot ;) Jan 09 21:46:49 You can start FTP from busybox, but you need to compile it first Jan 09 21:47:56 You could try my Debian rootfs, I tried to include an sftp server, but sftp is untested. But it's probably only working on 7.1 Jan 09 21:50:22 @Walid, done, lets see Jan 09 21:50:56 could a few of you plz tell me the output of abootimg -i hybris-boot.img Jan 09 21:52:45 Android Boot Image Info: … * file name = out/target/product/titan/hybris-boot.img … * image size = 6981632 bytes (6.66 MB) … page size = 2048 bytes … * Boot Name = "" … * kernel size = 5014680 bytes (4.78 MB) … ramdisk size = 977215 bytes (0.93 MB) … * load addresses: … kernel: 0x00008000 … ramdisk: 0x01000000 … tags: 0x00000100 … * cmdline = androidboot.bootdevice=msm_sdcc.1 androidboot. Jan 09 21:52:46 utags.blkdev=/dev/block/platform/msm_sdcc.1/by-name/utags selinux=0 androidboot.selinux=permissive audit=0 … * id = 0xe51d404a 0xddd4a15e 0xc0f1f2af 0x808e5db8 0x5fd53f79 0x00000000 0x00000000 0x00000000 Jan 09 21:56:01 @Walid, Thats from 5.1? Jan 09 21:56:13 7.1 Jan 09 21:57:56 Ok interesting, would need a 5.1 Jan 09 21:58:01 and thats for ubports-boot 5.1 … https://paste.ubuntu.com/26355745/ Jan 09 21:59:09 so offsets are the same Jan 09 21:59:17 ubports-boot.img work with xenial rootfs Jan 09 21:59:18 Ok will check later what ruinied my tree Jan 09 21:59:57 Hmm now width is correct but still same error in logcat: … root@localhost:~# EGL_PLATFORM=hwcomposer test_hwcomposer … width: 720 height: 1280 … library "/vendor/lib/egl/libGLESv2S3D_adreno.so" not found … OpenGL ES 2.0 (OpenGL ES 3.0 V@84.0 AU@05.00.00.046.002 (CL@)) Jan 09 22:00:10 no segfault ^^ thats good Jan 09 22:00:57 Oh: … E/FramebufferNativeWindow( 0): couldn't open framebuffer HAL (Operation not permitted) Jan 09 22:01:03 whats this 😆 Jan 09 22:01:57 udev rules file did you added it ? Jan 09 22:02:06 I think so but maybe not loaded? Jan 09 22:06:23 can you give me this file in your kernel? /drivers/video/msm/mdss/mdss_mdp_overlay.c Jan 09 22:07:02 to make test_hwcomposer work you need change function mdss_mdp_overlay_kickoff Jan 09 22:07:15 omg ^^ Jan 09 22:07:25 just a sec Jan 09 22:08:13 found the file Jan 09 22:08:29 have you got the patch? Jan 09 22:09:21 hhh after many test and search , i found that Jan 09 22:10:13 never test_hwcomposer work but after change function its work Jan 09 22:11:03 (Document) https://irc.ubports.com/pN516Qrs/file_3678.c Jan 09 22:11:20 this is my local file Jan 09 22:12:11 function mdss_mdp_overlay_kickoff line 999-1017 Jan 09 22:13:10 ok let me test... My file looks now very different Jan 09 22:13:17 ^^ Jan 09 22:13:26 No way to fix mdss stuff in dmesg Jan 09 22:13:44 I had them in 12.1 Jan 09 22:14:20 DeadSquirrel01 was added by: DeadSquirrel01 Jan 09 22:17:05 I have this file already in the CM13.0 kernel I am using now ^^ Jan 09 22:17:45 Can I see if udev rules are loaded? Jan 09 22:19:26 No good Jan 09 22:24:20 Go to sleep. not much happened today. Meh. Jan 09 22:25:38 besure that u have this i your system.prop file in device tree … ro.product.device=[device_code_name] Jan 09 22:26:30 Another day thrown in the bin, wow Jan 09 23:08:48 @ErnyTech, I do, but slowly from time to time Jan 09 23:17:06 @mariogrip, Ready for libinput fun? Jan 09 23:17:24 Libjardware Jan 09 23:17:40 pushin libhardware patch in 1.2.3 Jan 09 23:17:49 I'll be home in about 20 minutes Jan 09 23:17:51 but also been setting up communety channels Jan 09 23:18:07 so need someone to testdrive that :) Walid Jan 09 23:27:13 @mariogrip, The one we talked earlier? Jan 09 23:27:40 CAF/non-CAF? Jan 09 23:27:44 yes Jan 09 23:28:37 as a pr of course so we can test it first Jan 09 23:29:37 Okay.. I guess we can ask all the porters to test-drive it.. Jan 09 23:33:25 https://github.com/Halium/android_hardware_libhardware/pull/1 Jan 09 23:33:35 @bhushanshah ^ Jan 09 23:34:34 we should also fork this into halium, https://github.com/ubports/android-headers/tree/xenial that has many android versions in one Jan 09 23:37:30 Okay Jan 09 23:38:48 @bhushanshah should i just -f push over master and live halium-7.1 as it is? Jan 09 23:38:57 leave** Jan 09 23:39:30 or move master to halium-5.1? Jan 09 23:39:36 For? Jan 09 23:39:43 android-headers? Jan 09 23:39:46 yeah Jan 09 23:41:10 Umm give me little bit time for it, I need to kill older packages from repo.halium (as I added a bloody epoch), change libhybris packaging.. and also see how CAF is handled Jan 09 23:41:37 I'll add this in today's Todo list Jan 09 23:41:53 ok, will push it to a temp branch then Jan 09 23:46:11 Or I'll just re-fork Jan 09 23:46:58 @UniversalSuperBox if you want to try the caf-compat stuff, you just need to use https://github.com/Halium/android_hardware_libhardware/tree/halium-7.1-mariogrip for libhardware (might need to do a clean build) then add apt repo `xenial_-_qcom-compat` and remove the `xenial-7.1-caf-sudoku` Jan 09 23:47:26 @mariogrip, but im not sure, do you still use the special mir build? Jan 09 23:47:56 @bhushanshah, nah, easier to just do -f push Jan 09 23:48:41 but maybe it would be an idea, that way we can do pull requests beteen ubp and halium Jan 09 23:49:15 ^^ Jan 09 23:54:28 @mariogrip, ok, do you have any particular problem that blocks you? or it's just a problem of time? Jan 09 23:54:54 @ErnyTech, it's time Jan 09 23:55:45 @mariogrip, OK I saw that for op5 the work is ahead Jan 10 00:03:58 @mariogrip, yes i can help to test it Jan 10 00:06:07 doniks_: does old rootfs still work for you? I'm really out of ideas on this one now Jan 10 00:10:11 @Walid, Ok, can you create a tallball for your device like this? https://papad.org/p/ubports-communety-channel Jan 10 00:10:35 #pastefail Jan 10 00:11:00 @bhushanshah, huh? Jan 10 00:11:02 0 days since last paste buffer accident 😛 Jan 10 00:11:10 @mariogrip, . Jan 10 00:11:26 Oh Jan 10 00:11:29 Interesting Jan 10 00:11:51 I thought that etherpad was about different ubports telegram channel Jan 10 00:12:03 (due to e.g title) Jan 10 00:12:06 nope :P Jan 10 00:12:13 Now I look like idiot Jan 10 00:12:17 😶 Jan 10 00:12:26 haha it's fine :) Jan 10 00:14:54 Walid then i also need you to fill in the "blanks" here, this will be the channel name `16.04/communety/[username]/devel/[device code name]` Jan 10 00:15:42 username is your username Jan 10 00:22:27 @bhushanshah might it be an idea to rename system.img to android.img or maybe halium.img at some point? would make it simplet to understand what it is Jan 10 00:40:16 @mariogrip, 16.04/community/titan/devel/titan Jan 10 00:41:05 for tarball file i upload it her ? Jan 10 00:42:06 yeah, you can do that here that's fine Jan 10 00:43:02 @Walid, username is your username btw, so something like `16.04/community/walid/devel/titan` Jan 10 00:43:24 ah ok Jan 10 00:44:54 you can also use a different uploading service, this might be a little slow. but later on we will have a tool that will do the uploading automatic for you Jan 10 00:45:04 @mariogrip... if he uploads here I think the bot will fall to its knees Jan 10 00:45:14 (maybe it would make sense to send file over Dropbox/Google drive, as telegram is insanely slow for large files) Jan 10 00:45:29 @UniversalSuperBox, Yeah that too Jan 10 00:45:48 @UniversalSuperBox, will be a good test :P hey what's the worst that could happen Jan 10 00:46:03 (Document) https://irc.ubports.com/CMXOaUDc/file_3679.mp4 Jan 10 00:52:56 @mariogrip worse that can happen is is it would break some clients Jan 10 00:54:27 Test Jan 10 00:54:35 see it's fiine Jan 10 00:54:36 Not dead Jan 10 00:54:44 it's node, it's smart Jan 10 00:54:46 I was thinking it was a larger image Jan 10 00:55:13 Walid have you made any changes to mir and libhybris? Jan 10 00:55:30 no Jan 10 00:55:40 mm.so Jan 10 00:55:41 if these are changes that was on the device-files, they are no longer needed Jan 10 00:55:53 they are now in the base image Jan 10 00:56:15 so if you have no modified any of those, should i just remove them from the device tarball? Jan 10 00:56:59 @mariogrip any libcompat update I can try today before meeting? Jan 10 00:57:33 @bhushanshah, hopefully, depends on how fast i'll work tonight Jan 10 00:57:34 just for sound , libhybris-common.so.1.0.0 this file from device-files/common don't work , so i use orignal file from rootfs Jan 10 00:57:52 @mariogrip, Oh well okay Jan 10 00:58:32 Shoot Jan 10 00:58:43 @mariogrip I have to get the stock rootfs don't I Jan 10 00:58:45 non-caf Jan 10 00:58:59 @UniversalSuperBox, yes Jan 10 00:59:09 prepares for download failures Jan 10 00:59:10 but do you still use that hack for mir? Jan 10 00:59:19 Naw Jan 10 00:59:24 Ever since we got /firmware to mount Jan 10 00:59:33 ah! nice Jan 10 00:59:38 le me remebers now Jan 10 00:59:54 Okay, 6 minutes to download Jan 10 01:00:05 make that 10 Jan 10 01:00:46 Is your internet speed block chain based? Jan 10 01:00:50 (Photo, 714x75) https://irc.ubports.com/nCv5j9CS/file_3681.jpg Jan 10 01:00:51 oh boy Jan 10 01:00:53 "ba dum tss* Jan 10 01:01:17 what's up there Jan 10 01:01:27 @UniversalSuperBox, Use aria Jan 10 01:01:29 ci doesn't like america Jan 10 01:01:39 upload the rootfs to cdimage yo Jan 10 01:01:46 When you have time Jan 10 01:01:47 will to m8 Jan 10 01:01:51 do** Jan 10 01:03:50 @mariogrip tell when i can test the new channel Jan 10 01:04:03 @Walid, Yes, will do Jan 10 01:15:54 @UniversalSuperBox, https://ubports.nyc3.digitaloceanspaces.com/rootfs/xenial/ubports-touch.rootfs-xenial-armhf.tar.gz Jan 10 01:16:57 I Jan 10 01:17:03 It finally finished ._. Jan 10 01:17:34 well could you check the difference in speed? just so we know what to use later? Jan 10 01:17:46 Don't we have a cdimage server? Jan 10 01:18:09 we do, but that's in ams3 Jan 10 01:18:29 Not much better speed, but it's more reliable Jan 10 01:18:54 Spaces gets super expensive if you go over the transit cap, so monitor that Jan 10 01:19:19 also jenkins has auto artifacts that support amazon s2 (do uses that protocoll) Jan 10 01:19:30 @UniversalSuperBox, it's free for 2 months Jan 10 01:19:39 lol Jan 10 01:20:15 Additional storage: $0.02/GB, Additional outbound transfer from Spaces to the Internet: $0.01/GB that's not expensive at all Jan 10 01:41:45 Bah. @mariogrip do you have the pin handy Jan 10 01:42:00 @UniversalSuperBox, https://github.com/ubports/livecd-rootfs/blob/xenial/live-build/ubuntu-touch/archives/ubports.pref.chroot Jan 10 01:42:22 oh I had it right Jan 10 01:42:26 I should trust myself Jan 10 02:01:18 Walid the channel is ready https://system-image.ubports.com/16.04/community/walid/devel/titan/ Jan 10 02:19:58 Youhou ! **** ENDING LOGGING AT Wed Jan 10 03:00:02 2018