**** BEGIN LOGGING AT Sun Jan 14 03:00:04 2018 Jan 14 09:53:13 I'm having trouble building the systemimage: … it says the recovery image is too large Jan 14 09:53:44 https://paste.kde.org/pijrpxna5 Jan 14 10:06:36 @Ellenjott, you can tune the recovery size in the device BoardConfig or so Jan 14 10:06:46 But then dont ever try to flash that one ^^ Jan 14 10:07:04 Ok :) Jan 14 10:07:11 I'll try that Jan 14 10:07:19 We dont use the android recovery normally Jan 14 10:08:11 @Walid, Hmm I think your script has an infinite loop for me Jan 14 10:10:14 maybe you have pb inn your wncss driver Jan 14 10:10:34 I was online with WiFi in Halium rootfs, so somehow it worked already Jan 14 10:10:48 But I dont remember what I did, tried a few things ;) Jan 14 10:11:13 Also in my /etc is already init.qcom.bt.sh which should enable bluetooth the same like you do Jan 14 10:11:20 @Flohack, There is only a BOARD_USERDATAIMAGE_PARTITION_SIZE var ... Jan 14 10:11:21 shall I disable this maybe later Jan 14 10:12:20 its called service hciattach Jan 14 10:12:52 @Ellenjott, Wait I take a look. Basically if you run the build with make -n it will show you where the warning comes from Jan 14 10:13:41 ah forget you must disable any other wcnss / wifi service in init.qcom.rc Jan 14 10:13:43 yes for me the size is in BoardConfig hmmm Jan 14 10:13:49 @Walid, :D Jan 14 10:14:23 @Flohack, make? not mka? Jan 14 10:14:47 @Ellenjott, same probably. Dunno if mka has options ^^ Jan 14 10:15:04 with mka -n, makes no difference Jan 14 10:15:20 and make has no targets Jan 14 10:15:33 make works only in croot from top ^^ Jan 14 10:16:00 upss .. I was in the wrong dir ... Jan 14 10:16:03 @Walid, this one too? … service wcnss-service /system/bin/wcnss_service … class late_start … user root … group system wifi … oneshot Jan 14 10:16:48 Ahhh! class main Jan 14 10:16:55 Its late_start here lol Jan 14 10:17:47 @Flohack, Still no difference in the output ... Jan 14 10:21:37 @Ellenjott, hmm Jan 14 10:21:59 there is a size check at the end of this make run. The constants are either hardcoded or derived. grep for mkbootimg Jan 14 10:22:05 somewhere it must be ^^ Jan 14 10:26:58 what rootfs did you use? Jan 14 10:37:55 @Walid, you mean mine? Jan 14 10:42:27 @doniks, Here now Jan 14 10:42:40 hallo Jan 14 10:43:20 Hello there Jan 14 10:43:39 Is there a reason to try to change kernel or device tree? Jan 14 10:44:17 if one doesn't work and the other looks more promising Jan 14 10:44:24 maybe I misunderstood the question Jan 14 10:45:31 Well I was thinking maybe my trees have a problem or something Jan 14 10:48:26 at what state is it now? did bhushansha's tips help? Jan 14 10:50:03 @doniks, Trying now Jan 14 10:50:43 which kernel and device tree are you using? Jan 14 10:51:29 Want link on git? Jan 14 10:55:20 prajjwallaad was added by: prajjwallaad Jan 14 10:57:45 @Flohack, yes Jan 14 10:58:01 @KuranKaname, yes Jan 14 10:58:57 @doniks, device: https://github.com/slim-x2/android_device_leeco_x2-old … kernel: https://github.com/celtare21/android_kernel_leeco_msm8996/tree/blue_bsi_test Jan 14 10:59:13 Also removed the sepolicy folder from device tree Jan 14 11:03:06 Александр Якубенко was added by: Александр Якубенко Jan 14 11:03:18 @Александр Якубенко, welcome Александр Jan 14 11:04:04 <Александр Якубенко> @vanyasem, Hello 😊 Jan 14 11:04:38 @doniks, dmesg still has recovery stuff? Jan 14 11:06:40 the github names look a bit random. how did you pick them? are there no lineageos repositories? Jan 14 11:07:08 @KuranKaname, Actually sepolicy is required without it, it would show error Jan 14 11:07:46 @doniks, Kernel is mine Jan 14 11:07:49 @Walid, started with 16.-04 but now back on 15.04 - problem is booting stops, but no reboot. Also nothing in last_kmsg. Somehow though it loads android init.rc bc I get a small error msg about it Jan 14 11:07:52 Device tree is from slim roms Jan 14 11:07:57 Mine is based on that too Jan 14 11:08:11 @doniks, Lineage os ones are old and out of date Jan 14 11:09:08 ok Jan 14 11:09:35 For a fact I know they work on a normal rom Jan 14 11:09:44 you can always keep it as a last resort to try other trees, but doesn't sound like its the next step Jan 14 11:09:49 that's good Jan 14 11:09:55 then I'd keep them for now Jan 14 11:10:02 Oke Jan 14 11:10:22 did you say YOUR last_kmsg is still showing you recovery logs Jan 14 11:10:30 or did you ASK ME about MINE? Jan 14 11:10:39 @doniks, dmesg still has recovery stuff? Jan 14 11:10:53 Mine Jan 14 11:11:15 [ 0.000000] Linux version 3.18.20-lineage (root@instance-1) (gcc version 4.9 20150123 (prerelease) (GCC) ) #1 SMP PREEMPT Sat Jan 13 23:18:14 UTC 2018 Jan 14 11:11:43 @KuranKaname check if your kernel configs have this … CONFIG_ANDROID_RAM_CONSOLE=y … CONFIG_ANDROID_RAM_CONSOLE_ENABLE_VERBOSE=y Jan 14 11:11:50 Yes I do Jan 14 11:11:56 mhm Jan 14 11:12:15 but Jan 14 11:12:24 Sat Jan 13 23:18:14 UTC 2018 Jan 14 11:12:29 it was build yesterday Jan 14 11:12:44 I think that IS your kernel Jan 14 11:12:51 Yeah yesterday it's from last night Jan 14 11:13:06 so, then it's not recovery stuff right? Jan 14 11:13:11 @doniks, Can't be. My kernel is 3.18.86 and built with uber 6 Jan 14 11:13:43 @KuranKaname, But the build date... Jan 14 11:13:51 I first tried a 3.18.20 kernel and it wouldn't build Jan 14 11:13:57 display-caf was failing Jan 14 11:14:03 After kernel change it works Jan 14 11:14:11 And my kernel doesn't compile if it doesn't have uber Jan 14 11:14:19 what is uber? Jan 14 11:14:26 uber gcc toolchain Jan 14 11:15:40 Also, [ 99% 2035/2038] Making initramfs : /root/halium2/out/target/product/x2/obj/ROOT/hybris-boot_intermediates/boot-initramfs.gz … 2990 blocks … It says 2990 blocks. I have it synced 2 times, first time it wouldn't compile but it was saying 3110 blocks Jan 14 11:16:03 why ANDROID_PARANOID_NETWORK=n doesn't work? Jan 14 11:16:11 New android kernel limit? Jan 14 11:16:36 @doniks, https://forum.xda-developers.com/android/software/toolchain-uber-toolchains-t3527997 Jan 14 11:17:48 lets do this differently Jan 14 11:18:14 set your CONFIG_LOCALVERSION="-bladiblub" Jan 14 11:18:18 rebuild Jan 14 11:18:20 flash Jan 14 11:18:23 boot into telnet Jan 14 11:18:28 and check with dmesg Jan 14 11:18:44 that you are actually running that kernel now Jan 14 11:18:55 @doniks, CONFIG_LOCALVERSION="-BlueScreen" Jan 14 11:18:59 Already had that changed Jan 14 11:19:00 sure Jan 14 11:19:12 I mean, it was like that already Jan 14 11:19:33 ok, but it is like that now? Jan 14 11:19:45 Yes it is and always been Jan 14 11:19:50 It was never -lineage Jan 14 11:19:52 ok boot into telnet Jan 14 11:20:04 Ok lemme flash Jan 14 11:20:25 then post the dmesg Jan 14 11:39:25 @Walid, Reverting now apparmor patches to see if this helps :) Jan 14 11:40:07 yaserpathan was added by: yaserpathan Jan 14 11:42:14 Doesn't boot anymore Jan 14 11:42:21 @doniks, After adding that param Jan 14 11:42:48 which one? Jan 14 11:43:58 @doniks, CONFIG_DEVPTS_MULTIPLE_INSTANCES=y Jan 14 11:44:40 ok, then lets see whether we can establish reliable last_kmsg debugging Jan 14 11:45:02 so far it still seemed unclear to me which log file were looking at Jan 14 11:45:16 cd .. Jan 14 11:45:21 meh Jan 14 11:45:28 so remove the config, back to telnet, confirm that log and kernel match Jan 14 11:47:58 why does the rootfs get corrupted when restarted? Jan 14 11:54:13 @ErnyTech, the lxc rootfs? Jan 14 11:54:41 I think pre-start isn't built to run multiple times Jan 14 11:55:17 but maybe even if it where the android stuff inside wouldn't behave sane if restarted?! Jan 14 12:02:44 @doniks, Ok it doesn't boot no matter what Jan 14 12:04:29 Why now mount system.img read only?? Jan 14 12:08:07 @KuranKaname, well, I guess you have to backtrack until you get it back to telnet Jan 14 12:08:28 Not sure how much you kept track of all changes you made Jan 14 12:08:37 Maybe starting over is easiest Jan 14 12:19:47 @Flohack, you can disable apparmor add apparmor=0 in cmdline Jan 14 12:20:44 @Walid, I disabled already in kernel. So now situation is: … - boot and telnet to 192.168.2.15 is possible with debug appended … - normal boot: no ssh anymore, eventually sometimes get a black screen and also a kernel panic. Smth is wrong with ubports-boot I think Jan 14 12:21:01 ah wait now I got smth Jan 14 12:21:34 I got RNDIS up but no sshd Jan 14 12:21:37 damn Jan 14 12:21:50 how can this be? Jan 14 12:22:43 ah look: Jan 14 12:22:45 Starting Nmap 7.01 ( https://nmap.org ) at 2018-01-14 13:22 CET … Nmap scan report for 10.15.19.82 … Host is up (0.0096s latency). … Not shown: 999 closed ports … PORT STATE SERVICE … 22/tcp open ssh Jan 14 12:23:00 its just not using my credentials or what Jan 14 12:23:12 ssh -v Jan 14 12:23:12 use @UniversalSuperBox script to install it make ssh work fine Jan 14 12:23:17 or -vv Jan 14 12:23:19 @Walid, I used Jan 14 12:23:42 debug2: fd 3 setting O_NONBLOCK … debug1: Authenticating to 10.15.19.82:22 as 'phablet' … debug1: SSH2_MSG_KEXINIT sent … Connection reset by 10.15.19.82 port 22 Jan 14 12:23:53 did you change ip address Jan 14 12:23:58 its definitely not configured correctly Jan 14 12:24:15 no otherwise I would not get to this point Jan 14 12:24:19 10.15.19.100 Jan 14 12:24:28 not sure how the ubports rootfs likes their ssh connections Jan 14 12:24:35 password Jan 14 12:24:36 key Jan 14 12:24:37 PING 10.15.19.82 (10.15.19.82) 56(84) bytes of data. … 64 bytes from 10.15.19.82: icmp_seq=1 ttl=64 time=0.377 ms Jan 14 12:24:40 root Jan 14 12:24:42 phablet Jan 14 12:25:01 no its user phablet, and the rootstock-ng script injects all settings Jan 14 12:25:07 I check manually.... Jan 14 12:25:47 with -vv there should be way more output can you post it? Jan 14 12:26:29 @doniks, Yes but its not interesting. It shows the banner of the server, so its up ^^ Jan 14 12:26:36 I need to manually inject my PW Jan 14 12:31:06 I can't remount!! Jan 14 12:31:30 @ErnyTech, what, where Jan 14 12:39:19 I'm trying to get ubports-boot to work Jan 14 12:39:36 but I'm stuck at "I hit a nail" Jan 14 12:39:39 then telnet Jan 14 12:39:48 I think I tracked it down to https://github.com/ubports/ubports-boot/blob/master/initramfs/scripts/touch#L377 Jan 14 12:40:03 this line failing with status 255 Jan 14 12:40:12 here is my dmesg https://paste.ubuntu.com/26384482/ Jan 14 12:40:39 with some extra logging I put in prefixed with UT.init Jan 14 12:40:49 anybody has a clue? Jan 14 12:47:57 @Walid, can it be that ssh works only in xenial rootfs? Jan 14 12:48:07 I tried now vivid, and I get no RNDIS Jan 14 12:49:37 I also destroyed my kernel log again, I get only warnings from some drivers. Its impossible with this kernel config to find out when it logs correctly and when not. Its heavily customized it seems Jan 14 12:51:17 @Flohack, Vivid probably doesn't have hybris-usb, but ADB could work there Jan 14 12:51:36 @flohack, maybe you can already look at all other kinds of logs in /var/log/syslog Jan 14 12:51:58 if the rootfs boots and ssh starts there should be normal logging Jan 14 12:53:17 @doniks, Seems empty Jan 14 12:53:26 maybe its not mounted r/w? Jan 14 12:53:33 that would also explain crash of sshd Jan 14 12:53:47 but sshd is talking to you? Jan 14 12:54:16 no this time its not started Jan 14 12:54:37 disable lxc? Jan 14 12:54:43 i definitely wiped /data before to be sure its fresh Jan 14 12:55:05 hmm Jan 14 12:55:15 how was this on upstart Jan 14 12:55:26 oh upstart Jan 14 12:55:30 dunno Jan 14 12:56:10 rename the android folder :) Jan 14 12:56:46 hmm can it be just dead slow? Jan 14 12:56:48 look at this: Jan 14 12:57:14 Until I killed it by buttons: https://pastebin.com/A1kLUnkV Jan 14 12:57:36 second 11: its in android init.rc Jan 14 12:58:20 systemd-logind[1204]: Failed to start user service, ignoring: Unknown unit: user@32011.service Jan 14 12:58:23 yep its slow Jan 14 12:58:28 135 seconds after boot Jan 14 12:58:40 doesn't ubports porting guide say you need to disable all kind of "radio" stuff Jan 14 12:59:45 @doniks, hmm dunno. Did not see it now Jan 14 13:00:04 lets disable lxc Jan 14 13:00:57 https://docs.ubports.com/en/latest/porting/building-ubports-boot.html#edit-init-rc Jan 14 13:01:17 yes I think disabling lxc is a good next step Jan 14 13:01:58 damn where is this script again - those things have gone bad since upstart & system became real Jan 14 13:02:41 upstart uses initctl or something righ? Jan 14 13:02:52 oh but I'm not sure it works in chroot Jan 14 13:03:58 btw, I started a list :) Jan 14 13:04:00 https://gist.github.com/doniks/2ac14a1bb6750165bf782f62139c8aab Jan 14 13:05:59 Ok now got a bootloop finally ^^ Jan 14 13:12:14 Hmm can I somehow create the udev rules offline? Jan 14 13:13:01 @Flohack, define offline Jan 14 13:13:05 bootloop with lxc disabled? Jan 14 13:13:10 in adb. there are some files in / Jan 14 13:13:19 no bootloop but also no ssh ^^ Jan 14 13:14:13 you are in adb? Jan 14 13:14:18 yep Jan 14 13:14:19 how is / mounted? Jan 14 13:14:22 rw/ro? Jan 14 13:14:45 rw Jan 14 13:14:54 oh and to be clear you are in adb in the ubports rootfs? Jan 14 13:15:13 no currently in the adb rootfs ^^ what you want me to do? Jan 14 13:15:45 https://paste.ubuntu.com/26384691/ Jan 14 13:15:51 @SALPHA8506, any solution? Jan 14 13:16:27 @flohack, I have a hard time following. so lxc is disabled? Jan 14 13:16:46 @doniks, I made an lxc.override file with manual inside, yes Jan 14 13:17:05 ok and if you boot into the ubports rootfs what does it look like? Jan 14 13:17:11 sshd talking to you? Jan 14 13:18:05 @doniks, He talsk but hangs up. Its running but it cant give me entry. I dunno why, thats all the same with @UniversalSuperBox special script that enabled ssh. I checked the config, its password auth and a passowrwd has been set. I would love to see the error output of sshd Jan 14 13:18:41 can you try with ssh key? Jan 14 13:18:43 is password set Jan 14 13:18:59 yes Jan 14 13:19:00 i was the one that wrote the part of the script regarding setting the password Jan 14 13:19:05 it might not work correctly Jan 14 13:19:13 Ok I try again manually with chroot Jan 14 13:21:19 i think yes ssh work only in xenial, for vivid you need hack adbd binary file to start adb shell, and remove any init.usb*.rc import Jan 14 13:21:38 there is also no syslog in /var/log/ so I still doubt that it is r/o mounted Jan 14 13:22:32 for vivid some rootfs can't create syslog, strange i have this issue before Jan 14 13:23:13 or maybe it goes into some tmpfs? Jan 14 13:24:24 im on xenial now again Jan 14 13:24:49 which vivid rootfs are you using anyway> Jan 14 13:24:53 ? Jan 14 13:25:00 latest from ci.ubports Jan 14 13:25:12 i thougth ubportsonly works with xenial on halium Jan 14 13:27:14 https://github.com/ubports/docs.ubports.com/pull/67/files#diff-53e04d5e6e3e8482d46bc84fb6e51cf2R24 Jan 14 13:27:26 Right now, we only have one available. Simply download ``ubports-touch.rootfs-xenial-armhf.tar.gz`` from `our CI Jan 14 13:30:03 Ok I have to give up for the moment. Going to the airport soon, off for some days. Jan 14 13:30:25 Its a pity - I was already booting into graphical UI and now after changing a few small things its totally broken again ;) Jan 14 13:30:43 graphical in vivid? Jan 14 13:30:47 @Flohack, don't forget that porting is a strange sadism Jan 14 13:30:54 I know Jan 14 13:31:10