**** BEGIN LOGGING AT Sun Feb 25 03:00:05 2018 Feb 25 08:16:50 I finally got taoshan (Sony Xperia L) to boot with halium. With hybris-boot it did not boot at all, but with halium-boot it does. I can login via telnet. Feb 25 08:17:07 Now I tried to follow these instructions: https://github.com/Halium/projectmanagement/issues/40 Feb 25 08:17:43 But the command systemctl is not found. Feb 25 08:17:56 Here is the dmesg output: https://pastebin.com/LfjPN47U Feb 25 08:19:18 JBBgameich: I wonder, would it be possible to add a third layout to halium-boot ? Feb 25 08:20:23 @jonny, I just recognized that it looks at `/bin/sh` but it is located at `/root/bin` Feb 25 08:20:33 @jonny, It's not yet booted far enough. Currently it 's still in the initrd, not in the real system. There's something to fix before that Feb 25 08:21:04 Should I simply use that other one? Feb 25 08:21:09 JBBgameich: currently on LuneOS we install our rootfs in /data/luneos, as a bare file tree. To make it short, we unpack our rootfs in /data/luneos. Feb 25 08:22:55 @JBBgameich, So it does not make sense to run those commands befor booting the Ubuntu based system? Feb 25 08:23:01 @JBBgameich, [Edit] So it does not make sense to run those commands before booting the Ubuntu based system? Feb 25 08:24:21 @Tofe, I think that should be possible. But this boot layout is quite different from Helium's and could require more changes Feb 25 08:24:53 JBBgameich: doesn't it just sum up to a bind-mount instead of a loop-mount ? Feb 25 08:24:59 @Tofe, [Edit] I think that should be possible. But this boot layout is quite different from Halium's and could require more changes Feb 25 08:26:26 There's also some e2fsck running on the rootfs image in halium-boot before booting iirc, which we would need to disable for the luneOS boot mode Feb 25 08:26:26 JBBgameich: we could also switch to a rw image loop-mount; I'd have to discuss a bit the pro-con with our team Feb 25 08:27:11 JBBgameich: oh ok, I misunderstood that part, I thought it was the whole userdata partition that was checked Feb 25 08:27:41 But checking rootfs' image is a good idea of course Feb 25 08:28:13 JBBgameich: isn't there a performance difference, especially in rw mode, when using an loop mount of an image? Feb 25 08:28:50 @Tofe, If you compress the rootfs directory as tar.gz, it should install fine with one of the halium-install scripts Feb 25 08:29:34 @Tofe, I'm not sure, but I will research on it Feb 25 08:30:21 You could also flash luneOS to the system partition btw Feb 25 08:30:36 JBBgameich: you may say we shouldn't care that much about performance right now, this kind of deployment is only for devs; the final one should be to wipe the system partit--- ah, you said it :) Feb 25 08:31:35 got to go, I'll continue my analyze later; thanks :) Feb 25 08:43:42 I'm trying to boot using hybris-boot but it reboots after a few seconds Feb 25 08:43:52 Here's console-ramoops … https://paste.kde.org/p7fdbtsuu Feb 25 13:25:13 Thanks everyone for the help. I got “make failed to build some targets” on building system image. Where do I look to see which parts failed specifically? Feb 25 13:29:30 @Igor Dikiy, depend of error, sometime kernel, somtime device tree and so on Feb 25 15:27:43 hey guys, so I got halium-boot installed, I have ssh (although it drops the connection a few times) but I cant get it to boot the ubports rootfs Feb 25 15:28:08 well more acurately I cant get it to boot the gui portion Feb 25 15:28:28 Today's task: Nuke the `ubports` boot mode from Halium-boot Feb 25 15:29:07 do you think it could just be something in the kernel config? Feb 25 15:30:01 Unlikely. Do you have the full boot log? Feb 25 15:30:10 Well, just `dmesg` Feb 25 15:30:36 from my developer machine or the device itself Feb 25 15:30:42 The phone Feb 25 15:30:58 gimme a sec to get it Feb 25 15:32:11 its just spamming "usb 1-1: new high-speed USB device number 99 using xhci_msm_hsic Feb 25 15:32:12 [ 462.305465] usb 1-1: Device not responding to set address. Feb 25 15:32:12 [ 462.515412] usb 1-1: Device not responding to set address. Feb 25 15:32:12 [ 462.725368] usb 1-1: device not accepting address 99, error -71 Feb 25 15:32:12 " atm Feb 25 15:32:32 Use pastebin Feb 25 15:32:33 Careful pasting multiple messages Feb 25 15:32:59 yea, I forgot IRC does that, ive gotten used to telegram Feb 25 15:33:46 Well, looks like it's booted but USB tethering didn't come up (or adb is being stupid) Feb 25 15:33:49 What rootfs are you using? Feb 25 15:34:20 ubports, I can ssh in so wouldnt that be USB tethering in itself? Feb 25 15:34:45 You can ssh in? Feb 25 15:34:48 Well, you're booted Feb 25 15:34:51 ...congrats? Feb 25 15:35:00 yea, but im trying to get the gui to boot on the device Feb 25 15:35:36 Damn.. he got ssh Feb 25 15:35:42 haha Feb 25 15:35:54 i got ssh'd son :P Feb 25 15:37:07 im thinking its likely the console=tty0 in my kernel config stuff, I know samsing galaxy devices are ... interesting when it comes to that Feb 25 15:37:21 If you have SSH, the rootfs is booted Feb 25 15:37:38 No need to do anything else Feb 25 15:38:15 but I would like to oh i dunno use the pretty GUI on the devices screen, thats kinda the point of having one isnt it? Feb 25 15:38:35 Well, troubleshoot why the container isn't starting then Feb 25 15:38:54 any advice on where to start? Feb 25 15:39:14 The logs in `/var/log/lxc` Feb 25 15:39:23 Or running `lxc-start -nandroid -F` Feb 25 15:42:35 ok, that gave me a bunch of "no such file or directory/file exists" errors as well as a permission denied error Feb 25 15:44:12 ok forgot i was logged in as root, exited down to phablet got lxc-start: utils.c: mkdir_p: 254 Permission denied - failed to create directory '/run/user/32011/' as the only error Feb 25 15:48:15 making the directories is showing progress i guess, but I feel like it might not be the right path Feb 25 15:49:14 lxc-start: tools/lxc_start.c: main: 317 Executing '/sbin/init' with no configuration file may crash the host Feb 25 15:49:20 thats now my error Feb 25 16:00:05 guys what can I do when my dimming of the screen when turning the device on/off is very slow? Feb 25 16:52:26 And what can I do when everything works in 16.04 root fs, but in 15.04 I get only spinner animation Feb 25 16:52:39 Put your udev rules in place Feb 25 16:52:40 I backported myself the ssh access to 15.04 btw 😆 Feb 25 16:52:47 they are in place Feb 25 16:52:58 maybe they are different from 16.04? lol Feb 25 16:53:03 I took the 16.04 ones Feb 25 16:55:52 No they are identical Feb 25 16:56:27 It's possible that changes between halium and ubp-5.1 broke compatibility Feb 25 16:56:35 logcat complains: … cat /var/lib/lxc/android/rootfs/ueventd*.rc|grep ^/dev|sed -e 's/^\/dev\///'|awk '{printf "ACTION==\"add\", KERNEL==\"%s\", OWNER=\"%s\", GROUP=\"%s\", MODE=\"%s\"\n",$1,$3,$4,$2}' | sed -e 's/\r//' >/usr/lib/lxc-android-config/70-[codename].rules Feb 25 16:56:54 logcat complains: … W/unity8-dash(18879): type=1701 audit(0.0:1586): auid=4294967295 uid=32011 gid=32011 ses=4294967295 reason="memory violation" sig=6 … W/media-hub-serve(18884): type=1701 audit(0.0:1587): auid=4294967295 uid=32011 gid=32011 ses=4294967295 reason="memory violation" sig=11 … W/maliit-server(18885): type=1701 audit(0 Feb 25 16:56:55 .0:1588): auid=4294967295 uid=32011 gid=32011 ses=4294967295 reason="memory violation" sig=6 Feb 25 16:56:55 this is new Feb 25 16:57:01 Read the pin, Flo! XD Feb 25 16:57:05 😆 Feb 25 16:57:10 here its not mire than 4 lines Feb 25 16:57:11 😆 Feb 25 16:57:17 [Edit] here its not more than 4 lines Feb 25 16:59:47 Ok so on 16.04 I cant tell if my port has issues or if it is the root fs, and on 15.04 I got no unity at all. What next 😆 Feb 25 17:00:34 But WiFi works lol Feb 25 17:06:44 If someone wants to put his/her EGl expertise in this: https://pastebin.com/aiFCwtsG Feb 25 17:08:33 And do I need to update the system image when I change the kernel? I see that system-image target also builds a new kernel Feb 25 17:09:10 No, you don't generally Feb 25 17:11:59 hmm ok. So it seems that for some reason 15.04 is outdated. I already pushed a reasonable selection of device-files into the image (as Walid recommended) Feb 25 17:15:31 Can I safely push libhybris and libhybris-egl folders in full? Feb 25 17:21:00 So I put android back on my device for now, gotta figure out why I'm not getting a GUI Feb 25 17:44:49 When I was trying to start wifi... Feb 25 17:44:50 Fwd from rockybulwinkle: it can't find librmnetctl.so and libandroid_runtime.so Feb 25 17:44:50 Fwd from rockybulwinkle: Who is supposed to supply these libs? Feb 25 17:45:41 They weren't in the vendor folder Feb 25 17:52:59 If this is for WiFi, you can prob fi d them on the manufacturers website (assuming you know the WiFi chip you ha e) Feb 25 17:55:40 No, that's not quite how mobile drivers work Feb 25 18:08:15 My apparmor-parser sucks 100% CPU during startup, and this delays everthing for quite some time, any recommendations? Feb 25 18:09:52 ZeroPointEnergy: Thanks for that fix! It works perfectly. Feb 25 18:11:01 @UniversalSuperBox yw, needed it for debugging, so I knew where it was breaking :-) Feb 25 18:12:13 Something for ubports TODO list: Decide what to do with device buttons. My Samsung phone wakes up on the middle home button (ok that could be), but also both touch buttons on th bottom wake up the device. Means you can accidently switch on the device again Feb 25 18:13:03 Sounds like a problem in Unity's button handler Feb 25 18:33:10 @Tygerpro Tygerpro, I have the firmware built in to the kernel and the driver built in. Feb 25 18:34:05 Does anyone know where the libandroid_runtime.so and librmnetctl.so libs come from? Am I missing a repo or something? Feb 25 18:35:09 I don't have either of those files on my filesystem Feb 25 18:35:48 ok. So those missing libs is a red herring I guess. Feb 25 18:39:03 @notkit the issue with resize2fs not being available in halium-boot is fixed Feb 25 18:42:17 @UniversalSuperBox, I think its just assignment of what to do when a button is pressed. The same butztons cannot turn off the phone for example, this only works with the power button. So there must be a difference Feb 25 19:07:37 Do I have to fix the missing groups? … https://www.pic-upload.de/view-34685187/Linz.jpg.html Feb 25 19:08:01 Do I have to fix the missing groups? … starting version 219 … specified group 'qcom_diag' unknown … specified user 'bluetooth' unknown … specified group 'camera' unknown Feb 25 19:08:08 udev complains Feb 25 19:08:29 Ideally you'd change the owners in the udev configuration Feb 25 19:14:36 hmmm Feb 25 19:14:50 the list is very long apparently my device needs a lot of groups Feb 25 19:15:02 I can set them all to root lol Feb 25 19:35:03 where is the source code for the halium reference's test_* stuff? Particularly I'm looking for test_lights. It doesn't work for me, but I can access the notification LED through /sys/class. Feb 25 19:35:56 @rockybulwinkle, https://github.com/libhybris/libhybris/tree/master/hybris/tests Feb 25 19:36:20 thanks Feb 25 19:53:38 https://hastebin.com/ibakumatef.rb … My libsdmextension.so is 64 bit, but test_hwcomposer is 32 bit... Feb 25 20:01:22 I'm so confused, why can't anything find any of their libraries and in the few cases were they were, why weren't the right bitness libraries built? I don't know what to do. Feb 25 20:04:49 Should I be using a 64 bit halium reference rootfs? Feb 25 20:38:51 halium-boot work well with halium-5.1 Feb 25 20:39:09 That's good Feb 25 21:15:00 I can confirm too Feb 25 21:15:09 no issues, drop-in replacement for the others Feb 25 21:15:34 @UniversalSuperBox I'm using Halium-boot now btw Feb 25 22:50:02 JRepin was added by: JRepin Feb 25 23:23:39 Does anyone who has graphics working on their port have a " libsdmextension.so" ? Feb 25 23:24:14 None here Feb 25 23:25:09 does your test_hwcomposer compain about it missing? Feb 25 23:25:26 No, mine wants `libmm-qdcm.so`. Feb 25 23:25:31 No matter. What's the actual error? Feb 25 23:25:40 https://hastebin.com/ibakumatef.rb Feb 25 23:26:03 What new stuff appears in `logcat` when you run it? Feb 25 23:29:31 my halium installation is a bit broken on my phone from experimenting. I'm installing it now but it'll take a bit. Feb 25 23:38:41 @UniversalSuperBox, Yours wants that, did it find it? Feb 25 23:38:51 No, but it doesn't matter Feb 25 23:38:59 Those errors are generally red herrings Feb 25 23:39:02 ok, ok Feb 25 23:39:12 I'm more interested in how adreno or hwcomposer are crashing in Android Feb 26 00:02:32 I'm getting a lot of output every second to logcat, I'm not sure which part is relevant. Feb 26 00:03:08 https://hastebin.com/aqurociqip.rb Feb 26 00:03:40 Hmm. Feb 26 00:03:50 Not seeing what I want there... I think there's an exclude option in logcat Feb 26 00:04:00 Try excluding adsprpc and you should be able to do a bit better Feb 26 00:05:24 I think I got it Feb 26 00:05:32 https://hastebin.com/adasetutav.sql Feb 26 00:06:05 Display died Feb 26 00:06:06 Neet Feb 26 00:06:34 Is that file possibly in /firmware? Feb 26 00:06:54 that lib? Feb 26 00:07:00 Yeah Feb 26 00:08:02 it's in /var/lib/lxc/android/rootfs/system/vendor/lib64/libsdmextension.so. I added it to ld.so.conf.d and rebuilt the cache but it can't find it. Feb 26 00:08:12 A 32 bit executable can't use a 64 bit library right? Feb 26 00:15:01 I bring that up because libsdmextension.so is 64 bit. Feb 26 00:15:12 Ah Feb 26 00:23:12 I'm not sure where it's pulling in that library, it doesn't show up in ldd https://hastebin.com/gumulemavo.rb Feb 26 00:23:30 No, that would normally be inside Android Feb 26 00:26:36 any ideas? Feb 26 00:26:44 ...no... Feb 26 00:28:14 systemctl says "degraded" https://hastebin.com/edabubulov.pl Feb 26 01:08:03 idk if it matters but this is an OLED Feb 26 01:14:28 @UniversalSuperBox strace shows that it's looking for the lib in /android/system/lib/libsdmextension.so Feb 26 01:25:36 is there a way to quickly increase the size of an ext4 system in twrp? Feb 26 01:25:48 Yes, `resize2fs` Feb 26 01:27:18 thanks Feb 26 01:28:11 WELL WHO COULD HAVE SEEN THIS COMING? :p https://hastebin.com/imesuwezix.coffeescript Feb 26 01:28:27 hmm Feb 26 01:28:31 Well that's interesting Feb 26 01:28:40 I just stuck the lib where it was expecting Feb 26 01:28:45 UBports is tossing out an arm64 rootfs... but I'm fixing it right now Feb 26 01:29:00 I think PM has one, too? Feb 26 01:29:06 PM? Feb 26 01:29:12 Plasma Mobile Feb 26 01:29:18 ok Feb 26 01:29:52 how broken is the rootfs you're working on right now? Feb 26 01:30:07 unity8-dash may or may not segfault Feb 26 01:30:12 So you don't get a home screen Feb 26 01:30:24 Ok. That's no big deal if I'm not launching unity yet. Feb 26 01:30:41 [Edit] Ok. That's no big deal since not launching unity yet. Feb 26 01:30:55 Can you give me a link? Feb 26 01:31:07 https://ci.ubports.com/job/xenial-rootfs-arm64/ Feb 26 01:31:18 __Realizes that Marius did indeed build a new one like I asked him to_ Feb 26 01:31:20 [Edit] Realizes that Marius did indeed build a new one like I asked him to Feb 26 01:32:10 Do I need to build ubports-boot to boot this fs? Feb 26 01:32:18 Nope Feb 26 01:32:23 halium-boot will do it Feb 26 01:32:32 cool Feb 26 01:35:06 should I use JBB's halium-install with "ut" release? Feb 26 01:35:24 That should work Feb 26 01:35:32 Assuming it's okay with arm64 Feb 26 01:38:51 it doesn't look very happy.... Feb 26 01:39:08 https://hastebin.com/esovefuqeh.sql Feb 26 01:39:42 Well that's different Feb 26 01:40:22 "nonspare.img" is the system.img, uns...sparsified, with that lib inserted. Perhaps it's expecting a sparse image and can't handle nonsparse Feb 26 01:40:33 That seems likely Feb 26 01:42:16 Failed to run /bin/bash no such file or directory. qemu probably isn't set up for arm64. Do you know if it does anything more than run passwd? Feb 26 01:42:40 It also enables ssh by editing `/etc/init/ssh.override` Feb 26 01:42:40 Both things you can do after the image is pushed Feb 26 01:44:32 Ok. So, if/when this image works for me, it's the same image that boots into ubuntu touch? Feb 26 01:44:43 Yes? Feb 26 01:44:56 :D Feb 26 01:58:33 Can someone explain why I am getting "Trying dependencies-only mode on a non-existing device tree?" after lunch command, and what that means? Thanks Feb 26 01:59:04 rocky@zero ~ $ ssh phablet@10.15.19.82 … Connection closed by 10.15.19.82 port 22 Feb 26 02:00:23 @Igor Dikiy, I'm not sure what that means but I've been told it isn't really an error Feb 26 02:02:23 @UniversalSuperBox my ssh.override looks right. Feb 26 02:03:03 Does it use the same 10.15.19.82 IP? Feb 26 02:03:09 Yup Feb 26 02:03:25 Wonder if it's doing that because of the silly race condition Feb 26 02:03:39 which silly race condition Feb 26 02:04:28 Here, try this from recovery: https://paste.ubuntu.com/p/b3CXnHBJTm/ Feb 26 02:04:38 Yeah, I fixed a race condition earlier today Feb 26 02:04:53 @UniversalSuperBox, This one? Feb 26 02:05:13 Yep. Feb 26 02:05:25 ssh-keygen starts before `/etc/ssh` is writable Feb 26 02:05:35 I have an open PR, should get reviewed tomorrow Feb 26 02:05:40 ok Feb 26 02:06:09 @rockybulwinkle, Thank you Feb 26 02:07:30 @Igor Dikiy, Since your using lunch, your vendorsetup.sh is wrong in its naming Feb 26 02:07:52 If you don't have a vendorsetup.sh in your device tree, use breakfast instead Feb 26 02:08:07 @UniversalSuperBox https://hastebin.com/osuwusufag.rb Feb 26 02:08:23 one time it says it closed it. weird. Feb 26 02:08:50 Mhmm, that's what it does when it can't find its keys... but now you generated them Feb 26 02:10:38 hmmmm Feb 26 02:10:55 aw telnet isn't installed. Feb 26 02:11:09 I could do a reverse shell with netcat Feb 26 02:11:22 You could Feb 26 02:11:28 lol Feb 26 02:11:29 And if you do, please pastebin what you do. :) Feb 26 02:24:16 @Qiangong2, Thanks, but I do have vendorsetup.sh in my device folder, and it seems to be populating the lunch menu correctly (at least the names are correct). Also breakfast has the same message Feb 26 02:25:55 is your device tree codename the same as what is in your vendorsetup.sh? Feb 26 02:29:01 yes, the only question I have is why the lunch options are cm_acclaim... (acclaim is my device). Am I installing CM, I am pretty sure my device does not have it Feb 26 02:29:18 No, you're building a CM tree for your device Feb 26 02:50:21 @UniversalSuperBox I'm starting to get the feeling that upstart isn't running. I added some stuff to /etc/init.d/ssh and /etc/init.d/network to flash the LEDs when the script runs. Feb 26 02:53:13 Now I'm getting the following errors while trying to build systemimage: https://pastebin.com/DNWBRytA **** ENDING LOGGING AT Mon Feb 26 03:00:02 2018