**** BEGIN LOGGING AT Tue Oct 15 02:59:57 2019 Oct 15 03:50:55 JENNINGS anita was added by: JENNINGS anita Oct 15 04:41:58 Alluwwa was added by: Alluwwa Oct 15 05:14:47 harvey186 was added by: harvey186 Oct 15 05:15:19 Hi, i'm trying to port UT on my Xiaomi natrium but i'm getting no access via ssh … can me someone pls help ? I stuck here with ssh connection https://docs.ubports.com/en/latest/porting/ … phablet@10.15.19.82: Permission denied (publickey,password). … I'm trying since days. Oct 15 05:16:35 hmmm Oct 15 05:16:47 is password ok Oct 15 05:17:17 is it same as password you entedrd in halium-install Oct 15 05:28:35 @Luka [is it same as password you entedrd in halium-install], Yepp Oct 15 05:29:40 For me the error message isn't pointing to wrong pw. It seems a permission error Oct 15 05:30:58 That's the output of ssh -vvv https://del.dog/ssh Oct 15 06:34:31 try redownloading rootfs Oct 15 06:36:33 try deleting known hosts file Oct 15 06:41:52 @Martensite [try deleting known hosts file], nothing changed Oct 15 06:44:43 @Luka [try redownloading rootfs], are you talking about this ? ubports-touch.rootfs-xenial-armhf.tar.gz Oct 15 06:44:47 Roker2 was added by: Roker2 Oct 15 06:45:24 Hi. Where can I download Halium for mido? Oct 15 06:46:46 changing pw works, but still same error … root@localhost:/# passwd phablet … Enter new UNIX password: … Retype new UNIX password: … passwd: password updated successfully Oct 15 06:51:09 @harvey186 [changing pw works, but still same error … root@localhost:/# passwd phablet … Enter n …], How did you mount the system image and how did you do the chroot Oct 15 06:52:35 as discribed in documentation Oct 15 06:52:39 mkdir /a … mount /data/rootfs.img /a … chroot /a /bin/bash … . /etc/environment … passwd Oct 15 06:53:34 but chroot is giving a message: bash: passwd: command not found Oct 15 06:56:32 @harvey186 [but chroot is giving a message: bash: passwd: command not found], Well then changing pw did not work as you said above ^^ Oct 15 06:57:04 @Flohack [Well then changing pw did not work as you said above ^^], root@localhost:/# passwd phablet … Enter new UNIX password: … Retype new UNIX password: … passwd: password updated successfully Oct 15 06:57:15 @harvey186 [but chroot is giving a message: bash: passwd: command not found], ^^ Oct 15 06:57:30 @Flohack [^^], passwd: password updated successfully Oct 15 06:57:32 if chroot gives you an error you did not change the pw inside the container but elsewhere Oct 15 06:57:45 idk which password you are changing ^^ Oct 15 06:57:56 but you must be in the chroot otherwise it wont find it Oct 15 06:58:27 As far as I see, I'm in chroot ... root@localhost:/# passwd phablet Oct 15 06:59:37 so when does passwd not found message appear Oct 15 07:00:09 Also when you exit the container you can try to issue sync; umount /a; sync or so to make double sure the change is written to flash Oct 15 07:00:22 sometimes its still in cache when you reboot Oct 15 07:00:25 I have tried Oct 15 07:00:43 [Edit] I have tried umount and sync Oct 15 07:01:27 but anyway, I think it has nothing to do with pw. It#s a permission error, not wrong pw Oct 15 07:01:42 as far as I read the error message Oct 15 07:02:41 [Edit] as far as I read the error message phablet@10.15.19.82: Permission denied (publickey,password). Oct 15 07:05:36 @harvey186 [as far as I read the error message phablet@10.15.19.82: Permission denied (publ …], Thats a generic message from sshd telling you that you are not allowed to log in, while trying publickey and password method. BTW you can try it with a public key, why not? Oct 15 07:05:54 And this method worked for me all the time so far, so I wonder why it doesnt work for you Oct 15 07:06:09 But really, generate yourself an ssh key and transfer it on the device Oct 15 07:06:55 Also you can check /etc/shadow to see if the new password hash appears there. there are plenty of options left to work around or analyze the problem, you are giving up much too early ;) Oct 15 07:10:25 @Flohack [Also you can check /etc/shadow to see if the new password hash appears there. th …], I don't give up. I onyl can't understand why the documentation isn't working. And why noone really knows whats the root cause is. I'm going step by step the documention. So why isn't it working ?? Oct 15 07:12:39 I'm alos not understanding why this ssh is needed. Why is the udev role not created while building or installing. all needed information s are available. … And why is UT not booting after install. That all makes no sence to me Oct 15 07:14:23 @harvey186 [I'm alos not understanding why this ssh is needed. Why is the udev role not crea …], Welcome to porting. Makes no sense until it does. Every device is different. And you think you will see a GUI immediately after the first try? SOme people worked months only on the GUI part ^^ Oct 15 07:14:51 We want to leave adb shell behind as it is not available on non-Android devices. Thats the reason why ssh is the preferred debugging way in Halium Oct 15 07:16:08 OK, I'm a mainframe guy with a little linux and a liitle more android knowledge. Could be the my knowledge for this kind of work is to less Oct 15 07:16:51 @Flohack [We want to leave adb shell behind as it is not available on non-Android devices. …], I don't want debugging ;) Oct 15 07:17:02 I want have a running GUI on my phone Oct 15 07:17:57 @harvey186 [OK, I'm a mainframe guy with a little linux and a liitle more android knowledge. …], No, your expectation is too high: Porting needs a lot of patience. I have restarted my Samsung S3 Neo build a couple of times and started from scratch only because the kernel build was not ok. Thats happening. Nothing is documented from the ve Oct 15 07:17:57 ndors, so you need a lot of try and error. But if something works finally, you feel great! ;) Oct 15 07:18:04 but that's all out of topic. I will try with a new downloaded rootfs and when I still get the same error, I will go back to android development Oct 15 07:18:25 You should try the reference rootfs by Halium first, I told you yesterday already ^^ Oct 15 07:18:31 that one is easier to get running Oct 15 07:18:39 but it needs a different kernel config Oct 15 07:18:49 Yes, so wyh Oct 15 07:19:16 Its a similar guid here: http://docs.halium.org/en/latest/ Oct 15 07:19:34 But still important differences. Essentially it uses systemd instead of upstart Oct 15 07:20:05 Ok, anyway, I think we are leaving the topic ;) Oct 15 07:20:14 No why, thats Halium channel Oct 15 07:20:20 Its exactly here for that kind of discussion Oct 15 07:20:43 Also, you can be lucky, your device already has booted and waits for you with ssh, others are stuck with bootloops and just get emergency telnet hehe Oct 15 07:20:49 You think, I think it's more about my personal meaning ;) Oct 15 07:20:52 Thats already an achievement Oct 15 07:21:31 Well yes. That can be. There is no cooking recipe for Halium that will work on all devices out of box. Its more like a blueprint that you need to adapt Oct 15 07:22:41 Yes, that's right. But with small linux knowledge you can reslove an issue like permissin error Oct 15 07:24:07 OK, I will leave now. with new rootfs I'm getting the same error. So I will have a look into this next year or so. Hope something will changed til than Oct 15 07:43:21 Hey guys, I have a slight problem trying to port halium for my device. The repo sync command gives me an obscure error. Is this the right place to ask for help or is there somewhere else? Oct 15 07:43:56 CodeLog001 was added by: CodeLog001 Oct 15 07:53:16 nvm, I think I've fixed it Oct 15 08:12:46 Hm, I can't initialize sources for mido. … So, TheMuppets Xiaomi vendor doesn't exist. Oct 15 08:15:58 Roker2: Yeah that's a known problem for now Oct 15 08:16:14 With all Xiaomi devices more or less Oct 15 08:19:04 So, maybe use vendor by TheScarastic? … I can clone manifest and edit it. Oct 15 08:24:05 In general yes if he has CM-14.1 branch for Mido Oct 15 09:05:08 One moment... Oct 15 09:06:57 Yes, he has it. … https://github.com/TheScarastic/proprietary_vendor_xiaomi/tree/cm-14.1?files=1 Oct 15 10:17:02 @harvey186 [but that's all out of topic. I will try with a new downloaded rootfs and when I …], i had this error during initial days when porting halium , i suspect all kernel configs suggested are not applied Oct 15 15:00:58 Hi Oct 15 15:01:04 can i port halium Oct 15 15:01:08 kernel: https://github.com/LinerSRT/n370b_n1 Oct 15 15:02:22 is it mt6737 device? Oct 15 15:02:45 is there LineageOs tree for it? Oct 15 15:02:52 [Edit] is there LineageOS tree for it? Oct 15 15:02:59 @NotKit [is it mt6737 device?], no mt6737m Oct 15 15:03:06 device: https://github.com/LinerSRT/android_device_n370b Oct 15 15:03:25 vendor: https://github.com/LinerSRT/android_vendor_n370b Oct 15 15:05:07 ouch there is pre-built kernel there Oct 15 15:05:38 so i must recomile kernel and change pre-built in device tree to my new kernel? Oct 15 15:05:54 yes, or fix device tree to compile it Oct 15 15:06:28 no idea why it was done this way, probably the author had some troubles compiling it inside LineageOS Oct 15 15:06:32 [Edit] no idea why it was done this way, probably the author had some troubles compiling it inside LineageOS tree Oct 15 15:06:42 yes seems so Oct 15 15:20:32 Guys, I am getting an error when I do breakfast... It says my device wasn't found, but that's impossible, I set up the manifest and ran the device setup script and everything... Any advice? Oct 15 15:28:45 @CodeLog001 is it an officially supported device? the build guides from https://wiki.lineageos.org/devices/ should help if so Oct 15 15:31:03 It is not officially supported, no... Oct 15 15:31:27 I'm attempting a port Oct 15 15:35:46 which is the device? Oct 15 15:40:10 @CodeLog001 should've phrased my question better I guess, I meant maintained by the LineageOS community (official), if your manifest uses sources from the LineageOS (maybe even CyanogenMod) GitHub the device likely will be on that devices list Oct 15 15:42:20 The device is not officially supported by LineageOS, but there is an unofficial, LOS 16 port by SGCMarkus. I pulled his device sources. Oct 15 15:42:28 and created my manifest that way Oct 15 15:42:45 is there a way to manually add the spec? Oct 15 15:44:52 It seems I can't get any officially supported devices to build either, maybe I should just delete the source tree and start over? Oct 15 15:46:14 did you pull down official `halium-7.1` sources? Oct 15 15:47:11 yes Oct 15 15:47:50 the thing is `halium-7.1` sources are for Android 7.1 (Nougat) ports and are afaik incompatible with the 8.1 (Oreo) sources for your device kernel etc Oct 15 15:48:17 would it be possible if I created a cm-14.1 port? Oct 15 15:48:28 It wouldn't take too much effort, I think Oct 15 15:49:51 you could try Herrie's `halium-8.1` as a base too (https://github.com/Herrie82/android/tree/halium-8.1) since people have gotten SSH with it apparently, but it's not official by any means and under heavy development right now Oct 15 15:50:38 if possible a `cm-14.1` port would be highly recommended for now Oct 15 15:52:07 Okay, that seems like the best route for now. Thx Oct 15 15:52:47 what was the device again? I can try finding online sources for Nougat as well Oct 15 15:56:35 @CodeLog001 [would it be possible if I created a cm-14.1 port?], what is device stock Android version? Oct 15 15:57:14 you can't go below that Oct 15 15:57:31 I think the device is stock Oreo Oct 15 15:57:40 so that's a problem 😂 Oct 15 15:59:09 you can still start over with the previously mentioned Herrie's sources for `halium-8.1` but I can't guarantee a fully working port by any means at the end sadly :/ Oct 15 16:00:33 save your manifest if you decide to start over tho so you don't have to rewrite that again Oct 15 17:16:29 Arunnidhi was added by: Arunnidhi Oct 15 17:17:53 How to install Ubuntu touch is on my Android device Oct 15 17:20:24 model? Oct 15 17:20:33 Vivo y 90 Oct 15 17:21:00 @Luka [model?], Vivo y90 Oct 15 17:22:03 it starts with android 8.1 Oct 15 17:22:19 so you need halium 8.1, not yet ready Oct 15 17:22:45 My device Android 8.1.0 Oct 15 17:22:48 but you can always try to fix all its problems Oct 15 17:25:30 LD arch/arm/boot/compressed/vmlinux … /home/hello/doogeek/arm-eabi-4.8/arm-eabi/bin/ld:--defsym:1: syntax error … make[3]: *** [arch/arm/boot/compressed/vmlinux] Error 1 Oct 15 17:32:30 @Arunnidhi [My device Android 8.1.0], currently halium is available for 7.1 Oct 15 17:32:36 @Arunnidhi [My device Android 8.1.0], [Edit] currently halium is available for android 7.1 Oct 15 17:32:53 Ok Oct 15 17:32:57 Thank you Oct 15 17:34:43 emo321456dcvcvb.aternos.me Oct 15 17:34:44 mc Oct 15 17:39:17 @Luka [LD arch/arm/boot/compressed/vmlinux … /home/hello/doogeek/arm-eabi-4.8/arm-ea …], @NotKit is it compiler problem? Oct 15 17:40:13 seems to be an option compiler does not recognize Oct 15 17:40:16 linking issue Oct 15 17:40:23 or? Oct 15 17:40:39 seems so Oct 15 17:41:24 with vmlinuz Oct 15 17:43:00 have you used commands from readme? Oct 15 17:43:03 hmmm @NotKit using toolchain from halium, ended up with /home/hello/halium7/prebuilts/gcc/linux-x86/arm/arm-eabi-4.8/arm-eabi/bin/size: not found Oct 15 18:01:23 (Photo, 586x163) https://irc.ubports.com/06x1E064.png @Luka first error Oct 15 18:01:23 KBSS_SZ error Oct 15 18:01:23 you should look how it works and why throws error Oct 15 18:01:23 it is looking for prebuilt kernel from what i see Oct 15 18:01:23 alexandrejacarvalho was added by: alexandrejacarvalho **** BEGIN LOGGING AT Tue Oct 15 18:02:35 2019 Oct 15 18:03:49 compiling with new defconfig Oct 15 18:11:14 sorry @NotKit Akatsu Oct 15 18:11:17 not end Oct 15 18:11:19 ../drivers/misc/mediatek/mlog/mlog_logger.c: In function 'mlog_procinfo': … ../drivers/misc/mediatek/mlog/mlog_logger.c:690:28: error: 'struct signal_struct' has no member named 'oom_adj' … oom_score_adj = p->signal->oom_adj; Oct 15 18:11:29 with ut defconfig Oct 15 18:12:16 https://github.com/LinerSRT/n370b_n1 Oct 15 18:16:07 https://github.com/LinerSRT/n370b_n1/blob/nougat/drivers/misc/mediatek/mlog/mlog_logger.c#L690 Oct 15 18:16:11 error is here Oct 15 18:16:31 ifdef is commented Oct 15 18:16:52 imo u should uncomment ifdef else and endif Oct 15 18:18:23 or wait Oct 15 18:18:28 im looking inside code rn Oct 15 18:18:49 @Akatsu . [ifdef is commented], i dont think so Oct 15 18:18:56 at least this one Oct 15 18:29:05 in c // is coment not # i think Oct 15 18:29:12 or i am not right Oct 15 18:29:26 @NotKit Akatsu Oct 15 18:33:11 what if you set `CONFIG_ANDROID_LOW_MEMORY_KILLER_AUTODETECT_OOM_ADJ_VALUES=y` that's what I have on my defconfigs anyways Oct 15 18:33:58 [Edit] Luka what if you set `CONFIG_ANDROID_LOW_MEMORY_KILLER_AUTODETECT_OOM_ADJ_VALUES=y` that's what I have on my defconfigs anyways Oct 15 18:34:00 @Luka [or i am not right], yes, `//` for single line, or `/* */` as a block Oct 15 18:35:08 @Luka [i dont think so], from what i see error is legit AF Oct 15 18:35:28 idk why but mediatek wants to call for something what is not existing Oct 15 18:35:33 https://docs.huihoo.com/doxygen/linux/kernel/3.7/structsignal__struct.html Oct 15 18:35:50 int oom_score_adj … int oom_score_adj_min Oct 15 18:35:58 only these are there Oct 15 18:36:20 in that case setting the config option to y should fix it Oct 15 18:47:14 cant we just somehow disable MediatekLogger from building? Oct 15 18:54:13 lul Oct 15 18:54:20 its so easy to figure out Oct 15 18:54:26 im so stupid xD Oct 15 18:56:54 @Deathmist1337 figured out already? Oct 15 18:59:33 https://github.com/LinerSRT/n370b_n1/blob/f2063c3db04b2341e46c21bb9a2cfcbedee1e89c/drivers/staging/android/lowmemorykiller.c Oct 15 18:59:45 this driver is "source" for mtklogger Oct 15 18:59:53 however they messed up with that Oct 15 19:00:11 not reverted structs in proper way Oct 15 19:07:17 @Deathmist1337 >_> Oct 15 19:07:26 https://pastebin.com/g3XtKc8D Oct 15 19:17:52 Why is MTK so good in Mes Oct 15 19:18:00 Messing up things Oct 15 19:18:10 because mediacrap Oct 15 19:18:30 i cannot really understand their coding Oct 15 19:18:34 I would expect that it's fixed in some other targets such a Planet Gemini kernel? Oct 15 19:18:39 even 4.4 kernel have this bug Oct 15 19:19:42 @Herrie [I would expect that it's fixed in some other targets such a Planet Gemini kernel …], well, it compiles Oct 15 19:21:27 they call for oom_adj on signal struct from sched.h and there is only oom_score_adjust inside it, converter to make oom_adj is inside same file called lowmem_oom_score_adj_to_oom_adj. at the end i have no idea where they want to use oom_adj and where oom_score_adj Oct 15 19:24:28 Wasn't one of the oom_adj or oom_score_adj deprecated? At least some of the webOS / LuneOS bits give us this warning in logs Oct 15 19:25:52 (Photo, 1280x218) https://irc.ubports.com/d8cZ1RGr.png Oct 15 19:26:51 thats what i think Oct 15 19:27:19 but i have no idea what is this oom_adj at all Oct 15 19:28:03 i mean oom_score_adj too, and oom_adj is result of its recalc Oct 15 19:28:20 newer kernel not even have such thing Oct 15 19:28:35 i mean newer branch of this from same guy Oct 15 19:29:50 oh i get it by a bit Oct 15 19:30:18 OOM is used as some kind of aggresive memory management, it decided if app should be killed by some kind of score priority Oct 15 19:32:45 we use oom_score_adj in UT too Oct 15 19:33:31 i would just change second one from oom_score_adj = p->signal->oom_adj to oom_score_adj = p->signal->oom_score_adj Oct 15 19:33:44 [Edit] i would just change second one from `oom_score_adj = p->signal->oom_adj` to `oom_score_adj = p->signal->oom_score_adj` Oct 15 19:34:33 it won;t probably matter anyway if we won;t use this app Oct 15 19:34:45 but we will solve compilation errors Oct 16 01:03:47 mana_aku_tahu was added by: mana_aku_tahu **** ENDING LOGGING AT Wed Oct 16 02:59:57 2019