**** BEGIN LOGGING AT Mon Feb 26 03:00:02 2018 Feb 26 09:59:38 @rockybulwinkle, reboot to recovery, chroot to rootfs, and setup phablet passwd Feb 26 12:16:51 I used halium-boot + ubports-xenial rootfs to boot a Bq U plus (tenshi) device (halium porting and plasma mobile work well), but I can't connect either via telnet and ssh (using JBB's halium-install script ``launch.sh connect -p telnet -u phablet`` for instance). On top my laptop ``dmesg -w`` didn't report any msg when devic Feb 26 12:16:51 e started booting Feb 26 12:17:41 Any suggestion? I installed with JBB's halium-install script (``halium-install rootfs.tar.gz system.img ut``) Feb 26 12:19:37 I also disabled rsyslogd (``echo 'manual' |tee /etc/init/rsyslog.override``) and I left KERNEL_CMDLINE ``console=tty0`` Feb 26 12:20:14 mer script and check_kernel_config passed without warning Feb 26 13:01:46 @Walid, I'd already done that Feb 26 13:06:30 @rockybulwinkle, and I forgot to say, it didn't flash Feb 26 14:06:33 why in halium-5.1 i have this error (in halium-7.1 there is not) … halium-boot_intermediates/halium-initramfs.gz Error 23 Feb 26 14:07:41 i copy ob/ROOT folder from halium-7.1(out/target/product/titan/obj/ROOT) into 5.1, its work Feb 26 14:08:02 For whatever reason cURL decides it can't put a file there... Feb 26 14:08:03 But why Feb 26 14:10:23 Can you remove `--silent` from the call to `curl` in get-initrd.sh? Feb 26 14:13:21 https://pastebin.com/fwyDGU5e Feb 26 14:15:39 this is full log … https://pastebin.com/QYiPSXdU Feb 26 14:16:30 What happens if you do `mkdir halium-boot_intermediates/halium-initramfs.gz: No such file or directory` Feb 26 14:16:37 [Edit] What happens if you do `mkdir halium-boot_intermediates/halium-initramfs.gz` Feb 26 14:16:38 edited Feb 26 14:18:40 [Edit] What happens if you do `mkdir out/target/product/titan/obj/ROOT/halium-boot_intermediates/halium-initramfs.gz` Feb 26 14:18:42 edited again Feb 26 14:19:09 just create folder ROOT/halium-boot_intermediates, work Feb 26 14:19:29 ... the build system should really do that for us Feb 26 14:20:11 missing some modules in Android.mk ? Feb 26 14:20:29 ...am I? Feb 26 14:22:11 in halium-7.1 work fine, Feb 26 14:22:46 Missing halium-boot in build/core/main.mk? Feb 26 14:23:55 # Specific projects for halium … subdirs += \ … halium \ … hybris-boot \ … halium/halium-boot Feb 26 14:24:50 why … halium \ … hybris-boot \ <=== this why ? should be halium/hybris-boot ?r right ? Feb 26 14:25:08 I think yes Feb 26 15:56:51 i noticed with halium-boot, phone get correct date and time, not 1970 Feb 26 17:39:09 @Walid, Maybe is using bootfix for time i the boot scripts Feb 26 17:41:03 Like ubports boot have https://github.com/ubports/ubports-boot/blob/master/initramfs/scripts/local-premount/fixrtc Feb 26 19:32:55 Is UBports the only OS that currently supports Halium? Feb 26 19:33:05 No, Plasma Mobile does as well Feb 26 19:33:13 And also LuneOS partially Feb 26 19:34:35 Was there ever any word back from Sailfish? Last I heard, it was mentioned I'm not sure if ever went anywhere. Feb 26 19:34:51 [Edit] Was there ever any word back from Sailfish? Last I heard, it was mentioned but I'm not sure if ever went anywhere. Feb 26 22:25:22 ssh is definitely starting... Feb 26 22:25:57 Hmm... does it take a while before it closes the connection? Feb 26 22:26:12 I telnet'd into ssh's port just to see if it send me any text Feb 26 22:26:14 https://hastebin.com/kepinizegi.sql Feb 26 22:26:19 it dies immediately Feb 26 22:26:26 hm Feb 26 22:26:28 `ssh -v`/ Feb 26 22:27:07 https://hastebin.com/logulunoze.pas Feb 26 22:28:41 Reports say to do `dpkg-reconfigure openssh-server` to regenerate the keys Feb 26 22:28:43 But that's kinda hard with no console Feb 26 22:28:47 ... Feb 26 22:28:53 I can chroot Feb 26 22:29:11 Though I did do this already: https://paste.ubuntu.com/p/b3CXnHBJTm/ Feb 26 22:29:23 the first key it told me already existed Feb 26 22:29:28 I'm not saying that this isn't strange Feb 26 22:29:32 But it's really strange Feb 26 22:29:42 Wait, those sentences didn't make sense Feb 26 22:31:49 https://hastebin.com/xuxacumoje.vbs Feb 26 22:32:13 Looks ok Feb 26 22:33:30 maybe something wrong in my kernel like paranoid network or something? Feb 26 22:33:49 no, then it wouldn't be communicating at all Feb 26 22:40:29 I don't see logs appearing in /var/log. I do have a .writable_image in /data Feb 26 22:42:43 They'd be in /data/system-data Feb 26 22:44:23 I'm in. Feb 26 22:44:37 I removed ssh.override Feb 26 22:44:54 Oh, duh Feb 26 22:44:55 I forgot about that Feb 26 22:45:25 hah. I was thinking, I don't know what this is, how it works, but it seems to be stopping the normal sshd_config from doing its thing. Feb 26 22:45:36 The halium-scripts post_inst does modify the ssh.override though. Feb 26 22:46:34 https://hastebin.com/lonelupoji.rb Feb 26 22:47:02 I guess I need to do the udev stuff? Feb 26 22:47:02 You've been juked! Feb 26 22:47:02 Welcome to Upstart Feb 26 22:47:19 *sigh* I sysvinit was so simple.... Feb 26 22:47:31 Just leave off `systemctl` Feb 26 22:47:38 When managing services Feb 26 22:47:55 oh. Feb 26 22:47:56 right, it's not systemd (facepalm) Feb 26 22:48:06 Right, it's Upstart Feb 26 22:48:25 phablet@ubuntu-phablet:/$ status lxc … lxc start/running Feb 26 22:48:53 The Android container is called lxc-android-config Feb 26 22:48:55 It should be in post-start Feb 26 22:49:45 I'm not sure what to do with that info. I may need to read up on upstart lol Feb 26 22:49:57 `status lxc-android-config` Feb 26 22:50:14 phablet@ubuntu-phablet:/$ status lxc-android-config … lxc-android-config stop/post-start, (post-start) process 1016 Feb 26 22:50:22 So I see it's stopped I think? Do I have to start post-start? Feb 26 22:50:23 Your container is running! Feb 26 22:50:27 :D Feb 26 22:50:43 why does that tell you that. the process ID? Feb 26 22:51:13 Uh... Feb 26 22:51:29 I mean, it says "stop" Feb 26 22:51:37 Well, because it's post-start. I don't have the job config in front of me Feb 26 22:52:28 The first part is the target, the second is the current status Feb 26 22:52:30 I think Feb 26 22:52:39 ok Feb 26 22:53:02 Anyway Feb 26 22:53:13 Let's see if we can get this show on the road Feb 26 22:53:18 So... I guess I should add CAF? Feb 26 22:55:16 In the frequent porting problems gist: https://hastebin.com/sijepedasu.vbs … Do the same on ubuntu touch rootfs? Feb 26 22:55:47 No, we should be able to work this out... Feb 26 22:55:56 `sudo unity-system-compositor` Feb 26 22:56:31 https://hastebin.com/qojihecima.rb Feb 26 22:56:55 Okay, `ls /android` Feb 26 22:57:22 https://hastebin.com/qiqinenami.rb Feb 26 22:57:56 Wow, seems rather sparse for Android Feb 26 22:58:02 But `/firmware` is in there Feb 26 22:58:15 it's empty. Feb 26 22:58:19 :o Feb 26 22:58:26 oh. Feb 26 22:58:29 Well that'll do it Feb 26 22:58:38 The mount of `/firmware` should be noted in `dmesg` Feb 26 22:59:13 https://hastebin.com/ixolajifel.rb Feb 26 22:59:35 config_fw_somethingsomething_loader? Feb 26 22:59:57 Hmm, no... Feb 26 23:00:07 Can we see a full dmesg? Feb 26 23:00:12 sure Feb 26 23:01:24 I'm gonna do a grep -v TCP if that's alright Feb 26 23:02:08 https://hastebin.com/igufekalej.cs Feb 26 23:02:32 It's missing the beginning >_< Feb 26 23:02:42 oh Feb 26 23:02:47 kmsg got full Feb 26 23:04:52 I don't have that command. On reboot it's short enough to catch it all with dmesg Feb 26 23:05:42 https://hastebin.com/gamenecele.scala Feb 26 23:06:15 mounting firmware failed. Feb 26 23:06:19 Invalid argument Feb 26 23:06:25 Wonder what filesystem it is? Feb 26 23:08:56 vfat Feb 26 23:09:50 root@ubuntu-phablet:/android/firmware# ls … image verinfo Feb 26 23:10:16 I'm not sure if I needed to mount it with any special options so less priviledge users can access it Feb 26 23:10:31 no, phablet can access it Feb 26 23:10:44 No... but I don't know what halium-boot is trying to do Feb 26 23:11:18 have you made significant changes in the past week? Feb 26 23:11:26 Nope Feb 26 23:12:09 wow rild and adsprpcd is using all my CPU Feb 26 23:12:14 phone's getting warm Feb 26 23:12:37 Does usc run now? Feb 26 23:12:47 no Feb 26 23:13:09 https://hastebin.com/xedigitexu.rb Feb 26 23:13:17 I don't think anything changed Feb 26 23:21:45 not even test_vibrator is working anymore :/ Feb 26 23:21:57 yeah, the libhybris tests don't really work Feb 26 23:22:20 ok Feb 26 23:22:27 https://hastebin.com/oducuvacuv.coffeescript Feb 26 23:22:46 Oh! Feb 26 23:23:03 `export LD_LIBRARY_PATH=/android/system/lib64:$LD_LIBRARY_PATH` Feb 26 23:23:03 oh? Feb 26 23:23:17 and do an ldconfig after? Feb 26 23:23:33 No need Feb 26 23:24:00 logcat works now. usc Feb 26 23:24:01 doesn't Feb 26 23:24:22 Okay, what appears in logcat when usc dies? Feb 26 23:28:34 12-19 22:00:08.574 0 21195 E Adreno-GSL: : open(/dev/kgsl-3d0) failed: errno 2. No such file or directo … ry … 12-19 22:00:08.574 0 21195 W libEGL : eglInitialize(0xabf60c0) failed (EGL_BAD_ALLOC) Feb 26 23:28:47 https://hastebin.com/uzoqesugaq.pl Feb 26 23:29:55 https://hastebin.com/avokosecoj.coffeescript Feb 26 23:30:47 Ah, udev rules? Feb 26 23:30:52 They go somewhere else, one moment Feb 26 23:31:28 https://paste.ubuntu.com/p/KM2qgfs8yd/ Feb 26 23:31:50 can I disable rild? it's making my phone hot Feb 26 23:31:59 Yes, but you'd have to rebuild Android Feb 26 23:32:12 I always set the 5X on a spare heatsink. :P Feb 26 23:32:15 heh Feb 26 23:32:32 ....so... how's battery life on ubuntu touch? :p Feb 26 23:32:44 Poor when the port isn't working Feb 26 23:32:57 Oh, whoosh Feb 26 23:33:52 there's a missing quote or something in that paste Feb 26 23:34:26 There's a ` at the end Feb 26 23:35:20 https://hastebin.com/uzebahekir.hs Feb 26 23:35:49 ughhh, the hastebin logo gets in the way Feb 26 23:35:53 @UniversalSuperBox, I don't have one. I guess a frozen bag of peas will do Feb 26 23:36:32 https://hastebin.com/ecacekikom.http Feb 26 23:36:38 https://paste.ubuntu.com/p/NqQDqnQJwC/ Feb 26 23:36:45 Don't forget to replace `[CODENAME]` Feb 26 23:37:16 Oh Feb 26 23:37:17 I thought I did? Feb 26 23:37:23 `sudo mount / -o remount,rw` first Feb 26 23:37:34 @rockybulwinkle, I can't see Feb 26 23:37:46 (Photo, 571x104) https://irc.ubports.com/82qhs7xo.png Feb 26 23:37:48 https://hastebin.com/raw/ecacekikom Feb 26 23:38:03 Yeah, backtick at the end of the line Feb 26 23:38:13 I missed it when I copied my markdown Feb 26 23:38:53 There is a backtick at the end of the line. I am confused. Feb 26 23:39:02 oh Feb 26 23:39:29 Destroy it Feb 26 23:40:04 https://hastebin.com/raw/nuqoteyeru Feb 26 23:40:26 Oh! Feb 26 23:40:32 Bag of frozen onions is doing the trick for the heat :p Feb 26 23:40:43 Why is everything getting html quoted for you? Feb 26 23:40:49 uhhhh Feb 26 23:41:17 oh, derp Feb 26 23:41:47 I had wget'd it because the ubuntu's paste is a bit annoying in that it has no white margin on the right side of the long lines Feb 26 23:42:12 ok, we're good now, it ran. Feb 26 23:42:16 I'd paste it here but then I'd get mad at myself Feb 26 23:42:25 hastebin? or gist? Feb 26 23:42:29 So, reboot that phone and let's see what fails Feb 26 23:42:35 Hastebin makes me more upset. :P Feb 26 23:42:40 lol Feb 26 23:42:43 and gist? Feb 26 23:42:46 @UniversalSuperBox, This is part of the reason Feb 26 23:42:53 I could do gists Feb 26 23:42:58 yeah that's pretty dumb. Feb 26 23:43:12 someone needs to knock some sense into them Feb 26 23:43:24 Like who thought that was a good idea? Feb 26 23:43:27 right? Feb 26 23:43:38 "Works on my 5K monitor" Feb 26 23:43:38 great now I'm worried my phone is getting too COLD Feb 26 23:43:57 I'd be more worried about condensation Feb 26 23:43:57 you're right. Feb 26 23:43:59 And you know it will thermal throttle correctly, right? :P Feb 26 23:44:26 I'm more worried about heat+fully charged batter. That and OLED burn in is faster when it's warm Feb 26 23:44:33 ah Feb 26 23:44:49 I may have tux burned in by the time the gui starts. Feb 26 23:44:59 You get a framebuffer console? Feb 26 23:45:02 Lucky Feb 26 23:45:20 Be back in a bit Feb 26 23:46:32 (Photo, 723x1280) https://irc.ubports.com/L8QCVFsC.png Feb 26 23:47:36 still getting same errors Feb 26 23:48:36 it just display's that. Nothing else. I'm not sure if it's the bootloader or linux doing it. Feb 26 23:51:32 https://hastebin.com/vuhukevehi.pl Feb 26 23:51:57 @rockybulwinkle, to be clear, that's not new. It's done that even when I had LoS installed. Feb 27 00:05:16 [ 29.293952] kgsl kgsl-3d0: |_load_firmware| request_firmware(a530_pm4.fw) failed: -2 Feb 27 00:09:34 @univ Feb 27 00:09:53 [Edit] @UniversalSuperBox methinks stuff isn't getting mounted right at boot. Feb 27 00:17:29 oh wait Feb 27 00:17:46 Did you replace the pre-start.sh script like on halium reference? Feb 27 00:17:52 uhhhhh Feb 27 00:17:54 nope! Feb 27 00:18:03 I meant to ask if any of that stuff was needed. Feb 27 00:18:12 Might help Feb 27 00:20:39 most of that stuff is systemd related. Do I need to do the upstart equivalent of any of that? https://github.com/Halium/projectmanagement/issues/40 Feb 27 00:21:45 Nope Feb 27 00:22:18 Is /firmware mounted? If so, can you give me your kernel config? Feb 27 00:22:43 It wasn't before, `invalid argument` Feb 27 00:22:45 Which is strange Feb 27 00:23:17 selinux option in fstab? Feb 27 00:23:26 That'll do it Feb 27 00:23:36 @UniversalSuperBox it's not letting my connect over ssh after replacing that script Feb 27 00:24:09 In your fstab, remove all `context=` options Feb 27 00:24:10 connection refused :/ Feb 27 00:24:18 ok Feb 27 00:24:19 As for ssh, maybe the logs are up this time Feb 27 00:24:57 To be clear, that's your Android fstab... you'll need to rebuild your systemimage Feb 27 00:24:59 For ssh.. kill all import init*usb*rc lines from init.*.rc Feb 27 00:27:13 @bhushanshah https://github.com/rockybulwinkle/android_kernel_zte_msm8996/blob/halium_axon7/arch/arm64/configs/lineageos_axon7_defconfig Feb 27 00:27:34 Taking out context will fix that Feb 27 00:27:45 Wait... @bhushanshah how did hybris-boot get around that? Feb 27 00:27:56 Was it just selinux-aware? Feb 27 00:28:23 I'm using halium-boot remember? Feb 27 00:28:43 Yeah I know Feb 27 00:28:45 ok Feb 27 00:28:53 If you were using hybris-boot you'd have way more problems. :) Feb 27 00:28:59 @UniversalSuperBox, It just didn't use Android fstab to mount data Feb 27 00:29:06 Right' Feb 27 00:29:09 [Edit] Right Feb 27 00:30:17 But I do want to pick your brain a bit @bhushanshah Feb 27 00:31:05 We've got those crazy segfaults continuing on the 5X under UT with a few applications. The Unity spinner, unity8 itself, `maliit`... Feb 27 00:31:06 Wrong time or pretty much wrong day for it Feb 27 00:31:12 Oh, that's okay Feb 27 00:31:28 Short version, we get `syscall 283` followed by a CPU register dump in dmesg Feb 27 00:32:13 removed context= from stab, the *usb* stuff in *.rc was already commented out Feb 27 00:32:14 [Edit] removed context= from fstab, the *usb* stuff in *.rc was already commented out Feb 27 00:32:17 Without the UT config options, it says unhandled input address fault: https://paste.ubuntu.com/p/YvVQQsZnHW/ Feb 27 00:32:28 rebuilding Feb 27 00:32:39 Which was later changed to level 0 translation fault in newer kernels Feb 27 00:32:46 Sorry.. I am off today fo for Germany Visa application appointment for upcoming plasma sprint but I'll take a look tomorrow or maybe evening Feb 27 00:32:51 Ooooh! Feb 27 00:32:55 have fun! Feb 27 00:34:50 Also you might want to use a utility called ...something-which-i-dont-remember... to convert that stack trace to human readable output. (I.e actual symbols instead of addresses) Feb 27 00:35:03 I mean... gdb does that Feb 27 00:35:15 But a CPU register dump? Feb 27 00:35:43 @mariogrip, might I hazard a guess that you know what that utility would be called? Feb 27 00:37:13 @UniversalSuperBox, It's.. not exactly fun thing btw.. 😭 Feb 27 00:37:34 btw that adreno firmware wasn't in /firmware when I mounted it manually Feb 27 00:37:42 Ever changing rules and regulations annoys hell out of me Feb 27 00:37:57 @rockybulwinkle, Hopefully the device will fix itself Feb 27 00:38:04 the `/dev` part Feb 27 00:38:27 @rockybulwinkle, Adreno firmware is in /vendor/firmware generally Feb 27 00:38:30 Might be created by something that comes up before adreno. Feb 27 00:38:38 ok Feb 27 00:38:49 @UniversalSuperBox, No.. 😛 Feb 27 00:39:02 @bhushanshah, Not the firmware, the specific device that it's trying to open Feb 27 00:39:15 the device it's trying to open needs the firmware Feb 27 00:39:33 that may be the real source of the 404 error Feb 27 00:40:03 Mmh, yeah Feb 27 00:40:53 Does halium-install have an option for just updating system.img? or should I just do it manually Feb 27 00:41:27 There's replace-android-image script Feb 27 00:41:32 There's a `replace-android-system` in github.com/halium/halium-scripts Feb 27 00:41:42 got the url right, yes! Feb 27 00:42:51 *does it manually anyway* Feb 27 00:43:58 is it possible to cross compile the test_* for arm on a desktop? Feb 27 00:44:07 Is your rootfs with weird crashes arm64 btw @UniversalSuperBox ? Feb 27 00:44:25 Both armhf and arm64 do it Feb 27 00:44:25 The only thing in common is the good ol' 5X Feb 27 00:44:54 The syscall is different on armhf, though. Feb 27 00:45:01 Which makes sense Feb 27 00:45:28 In either case, my googling led me to find that the syscall is supposed to be `membarrier`, which didn't come around until kernel 4.3 Feb 27 00:46:28 @UniversalSuperBox, ehm, no clue Feb 27 00:46:29 Well thanks for the answer! Feb 27 00:46:34 :D Feb 27 00:46:51 throws 5X out window Feb 27 00:47:07 * mariogrip has a feeling dalton does not like ths 5X Feb 27 00:47:24 It was better with LineageOS on it. Feb 27 00:47:24 * mariogrip has a feeling dalton does not like his 5X Feb 27 00:47:32 well, fix it :) then it will be better with ut Feb 27 00:48:00 Just fix it, he says Feb 27 00:48:00 It'll be easy, he says Feb 27 00:48:07 @UniversalSuperBox, Umm in 3.10 it's just sys_connect? Feb 27 00:48:25 It's 283 on arm64, 389 on armhf Feb 27 00:48:54 @bhushanshah, I im 80% sure i will be there :) Feb 27 00:49:03 Ah Feb 27 00:49:22 *is pushing new system.img* Feb 27 00:49:34 Y'know, the script would have done it in about 30 seconds. :P Feb 27 00:49:57 me too Feb 27 00:49:57 A lot of that time was it building :p Feb 27 00:50:24 @UniversalSuperBox, Um wat.. https://elixir.bootlin.com/linux/v3.10/source/arch/arm64/include/asm/unistd32.h#L402 Feb 27 00:51:06 Oh good Feb 27 00:51:08 It's impossible Feb 27 00:51:50 ssh connection refused still... Feb 27 00:52:07 Got journalctl? Or last_kmsg? Feb 27 00:52:20 Ubuntu Touch :/ Feb 27 00:52:50 _syslog_? Feb 27 00:53:00 [Edit] syslog? Feb 27 00:53:00 yeah Feb 27 00:53:17 :P Feb 27 00:53:39 I don't have that comand in my chroot Feb 27 00:54:20 Right, that's at `/var/log/syslog` Feb 27 00:54:27 so `/data/system-data/var/log/syslog` Feb 27 00:55:17 I don't have that file. I have rootfs.img mounted at /data/a, so I look in /data/a/var/log/syslog right? Feb 27 00:55:28 @UniversalSuperBox, Nope, here Feb 27 00:55:38 o Feb 27 00:55:44 The overlays for system files are in `/data/system-data` Feb 27 00:55:50 And user data goes in `/data/user-data` Feb 27 00:55:51 @mariogrip, I hope it's 100% soon.. Feb 27 00:56:03 Me too! Feb 27 00:56:34 no such file Feb 27 00:56:46 @mariogrip have I ever told you how terrible rsyslogd is Feb 27 00:57:03 -_- Feb 27 00:57:12 https://gist.github.com/rockybulwinkle/b1b1c77382871365e8bdcc2bac5cd703 Feb 27 00:57:19 @UniversalSuperBox, You have Feb 27 00:57:22 lol Feb 27 00:57:36 Hmm... There may be something in upstart/ssh.log Feb 27 00:57:49 @bhushanshah, Well I'm wrong, because it's most likely our config or build of it Feb 27 00:59:12 https://gist.github.com/rockybulwinkle/eb6ee63a13b876bb74851ed2750863bc Feb 27 00:59:19 @UniversalSuperBox, well good news that systemd is working nice on bonic, so we soon may do tha switch Feb 27 00:59:29 @UniversalSuperBox, [Edit] well good news that systemd is working nice on bionic, so we soon may do tha switch Feb 27 00:59:29 @mariogrip, RIP legacy devices. Feb 27 00:59:42 @rockybulwinkle, That darn race condition again Feb 27 00:59:43 maybe Feb 27 01:00:13 This is the PR that fixes it btw: https://github.com/ubports/lxc-android-config/pull/1 Feb 27 01:00:50 Though, once the keyfiles are in place you should be good t ogo Feb 27 01:01:26 Either way, try installing this deb but telling it to leave your ssh.override in place: https://ci.ubports.com/job/ubports/job/lxc-android-config/view/change-requests/job/PR-1/ Feb 27 01:01:57 ok Feb 27 01:02:26 "leave my ssh.override in place" I had moved the original ssh.override, doing that was what originally let me ssh in Feb 27 01:02:40 So... Feb 27 01:02:50 This is kinda screwy Feb 27 01:03:13 If the file in `/data/system-data` doesn't exist, `halium-boot` copies it from the rootfs to the overlay Feb 27 01:03:23 But if the file gets changed in the rootfs, it doesn't get copied again Feb 27 01:03:32 So that's why writable paths are used sparingly Feb 27 01:04:25 https://gist.github.com/rockybulwinkle/2986aa535fe7a7754993c2dff5416971 Feb 27 01:04:37 Looks ok Feb 27 01:05:05 still connection refused. Feb 27 01:05:25 Wonder what the heck it's doing... Feb 27 01:06:41 What was the original instructions you used to get it mounted and moved? Feb 27 01:06:43 That paste Feb 27 01:07:09 get which thing mounted and moved? Feb 27 01:07:43 The rootfs Feb 27 01:07:51 The paste I sent you while you were running rootstock, or soon after Feb 27 01:08:19 blarg. I don't remember. I've been using JBB's halium-install to install my rootfs Feb 27 01:09:04 Found it! Feb 27 01:09:12 Well, let's try killing rsyslogd for the moment. Feb 27 01:10:02 https://paste.ubuntu.com/p/zz5NvnkFgs/ Feb 27 01:11:20 ok, don't disable it with upstart then Feb 27 01:11:31 Well, you can't really Feb 27 01:11:39 That's how you disable services with upstart Feb 27 01:11:53 wat Feb 27 01:11:53 Good luck if, uh, there's already an override in place Feb 27 01:12:11 Yeah Feb 27 01:12:20 The override file can change any option in the job config Feb 27 01:13:27 I'm in! Feb 27 01:13:27 Pretty neat, until I learned that `systemctl edit` exists Feb 27 01:16:19 usc doesn't work still. Also still see load_firmware failing for a530_pm4.fw. However I see it does exist now in /android/system/etc/firmware/a530_pm4.fw Feb 27 01:17:14 @rockybulwinkle, Kernel config please Feb 27 01:17:53 https://github.com/rockybulwinkle/android_kernel_zte_msm8996/blob/halium_axon7/arch/arm64/configs/lineageos_axon7_defconfig Feb 27 01:19:13 maybe not related but at some point I lost Cgroup namespace according to lxc-checkconfig Feb 27 01:19:25 or maybe ubuntu's checks for more things? Feb 27 01:20:19 Toggle FW_LOADER_USER_HELPER_FALLBACK and retry Feb 27 01:20:26 ok Feb 27 01:20:49 @rockybulwinkle, Also is container running? Feb 27 01:21:38 it appears so phablet@ubuntu-phablet:~$ status lxc-android-config … lxc-android-config stop/post-start, (post-start) process 1009 Feb 27 01:22:53 Umm Feb 27 01:23:21 ? Feb 27 01:23:35 @UniversalSuperBox is it correct way to verify container? ^ Feb 27 01:23:44 that's what he had me do before Feb 27 01:26:24 Okay Feb 27 01:29:17 yeah, actually Feb 27 01:29:24 You could also do `lxc-info -nandroid` Feb 27 01:29:53 (which is `-n android` but I'm too lazy to hit the spacebar) Feb 27 01:30:30 it's running according to that too Feb 27 01:33:00 hey, that firmware request isn't failing Feb 27 01:33:12 Hey, cool Feb 27 01:33:13 What's new? Feb 27 01:33:27 usc isn't crashing yet? It's doing nothing so far? Feb 27 01:33:39 "to and from FDs are required for display manager" Feb 27 01:33:40 Oh... Feb 27 01:33:45 `sudo` Feb 27 01:34:17 `sudo stop apport` will make things crash faster since it doesn't try to process their crash dumps Feb 27 01:34:20 I did sudo -s, then usc Feb 27 01:34:55 hmm Feb 27 01:35:01 `stop lightdm` then do it again Feb 27 01:35:45 I had tried it again before stopping lightdm but after stopping apport. It's not crashing but not displaying anything... Feb 27 01:36:14 I'l reboot Feb 27 01:36:15 progress. :P Feb 27 01:36:27 Maybe backlight is off?? Feb 27 01:36:35 Definitely not off. Feb 27 01:36:46 It's an OLED, it's still showing bootloader stuff Feb 27 01:37:28 The reason I use usc is because it's a relatively simple mir server Feb 27 01:37:49 There's one simpler, though... You'll need to bring wi-fi up and install `mir-demos` to get to it, though Feb 27 01:38:28 https://gist.github.com/rockybulwinkle/9199a5176878ad2f3967dcac4848beac Feb 27 01:38:52 My wifi isn't working yet Feb 27 01:38:55 Maybe `unity-system-compositor --debug-without-dm` Feb 27 01:39:19 Ideally you get the Ubuntu Touch spinner Feb 27 01:39:19 segfault Feb 27 01:39:30 (Photo, 504x414) https://irc.ubports.com/1aKtGDA8.png Feb 27 01:39:51 (Sticker, 494x512) https://irc.ubports.com/GazWBvy0.webp Feb 27 01:40:13 (Sticker, 512x512) https://irc.ubports.com/oZOPClOV.webp Feb 27 01:40:23 https://gist.github.com/rockybulwinkle/a4ab92f1b88cee550ee656292035baeb Feb 27 01:40:24 Well... Feb 27 01:41:07 That's quite a ways into initializatoin Feb 27 01:41:09 [Edit] That's quite a ways into initialization Feb 27 01:41:45 I think it's working actually... But I also think Marius reverted the CAF unification changes because our Exynos device just can't keep it together Feb 27 01:42:58 Let's see if I can dig up a usc with caf changes Feb 27 01:43:30 Bahhhh, that's before we brought up arm64. :( Feb 27 01:43:38 grrr Feb 27 01:43:51 #BlameMarius Feb 27 01:44:05 Wonder if I could just tweak a Jenkinsfile... Feb 27 01:44:05 Who needs samsung SoC's anyway? Feb 27 01:44:19 brb, possibly breaking all UBports devices again Feb 27 01:46:48 yay Feb 27 01:47:32 https://ci.ubports.com/blue/organizations/jenkins/ubports%2Fmir/detail/xenial-7.1-caf-sudoku/16/pipeline Feb 27 01:47:51 Does wi-fi not work in the halium rootfs either? Feb 27 01:48:20 It's very possible that toggling the fw loader fixes it. Feb 27 01:48:23 Try this: Feb 27 01:48:29 ```echo 1 > /dev/wcnss_wlan … echo sta > /sys/module/wlan/parameters/fwpath``` Feb 27 01:48:59 That device have broadcom WiFi btw Feb 27 01:49:05 mhm Feb 27 01:49:13 No? Feb 27 01:49:14 oh. Feb 27 01:49:15 wait. Feb 27 01:49:28 I think it had some atheros chipset? Feb 27 01:49:43 that's what I found in lspci Feb 27 01:50:25 Eh, try it Feb 27 01:50:31 Worst comes to worst, it doesn't work Feb 27 01:50:56 the build failed Feb 27 01:51:00 That it did Feb 27 01:51:24 Oh boy, Blue Ocean kills Firefox Feb 27 01:52:34 the "echo sta" stalls Feb 27 01:53:02 which is different from when I was on the reference Feb 27 01:53:28 Could be because it's trying to bring up the firmware Feb 27 01:53:34 The driver will yell at you from `dmesg` if it is Feb 27 01:54:06 I can't actually find the build error in that job's output... Feb 27 01:54:44 failed to build with cowbuilder? Feb 27 01:55:19 Something far before that Feb 27 01:55:30 It's `-j64`, it tends to bury its failures Feb 27 01:56:22 wow, what cpu do you have? Feb 27 01:56:32 s/you/ubports build server/ Feb 27 01:56:43 It's an arm64 server... actually not sure what it's running Feb 27 01:57:04 Donated by Packet as part of the Works on ARM campaign Feb 27 01:57:12 cool Feb 27 01:57:44 `build/mir-0.24.20ubuntu11+ubports+0~20180227014643.16~1.gbp195b21/src/platforms/mesa/server/kms/kms_page_flipper.cpp:80:5: error: missing initializer for member '_drmEventContext::page_flip_handler2' [-Werror=missing-field-initializers]` Feb 27 01:58:01 Just found it, but thank you Feb 27 01:58:05 I don't think I see any errors in dmesg, rebooting so I can check right away Feb 27 01:58:28 @UniversalSuperBox, pro-tip: always search for 'error:' Feb 27 01:58:45 I normally search for 'unfinished', but that sounds like a better idea Feb 27 01:59:42 hah, ok. Feb 27 02:00:08 btw I'm getting different output from usc: https://gist.github.com/rockybulwinkle/1e93681e0399571da34a872cf8d5766b Feb 27 02:00:25 that's on the first run after a reboot. on second run it segfaults Feb 27 02:00:25 `stop lightdm` first Feb 27 02:00:35 oh right Feb 27 02:00:53 You're lucky the phone doesn't reboot Feb 27 02:01:12 If lightdm crashes enough, there's a script that reboots the device Feb 27 02:02:35 https://gist.github.com/rockybulwinkle/6437452abf3809384a468df4320d7f47 Feb 27 02:02:44 It paused for a while before printing that it couldn't find that lib Feb 27 02:03:13 Yeah, I'm still not too concerned there Feb 27 02:03:25 why do I need wifi for the other mir test? Feb 27 02:03:28 ` mirserver: Mir version 0.24.2` makes me think that it's 100% ready to go Feb 27 02:03:33 So you can download it Feb 27 02:03:38 I can do usb tether Feb 27 02:03:46 But if the problem is that you need caf, they'll fail too Feb 27 02:03:49 They all rely on the same lib Feb 27 02:03:59 I'm tethered to usb right now? Feb 27 02:04:38 I gots da internetz Feb 27 02:05:06 Yeah, you can try `apt update && apt install mir-demos` Feb 27 02:05:15 And then `mir_demo_standalone_render_to_fb` Feb 27 02:05:22 But I don't think they'll work due to your caf-ness Feb 27 02:05:35 ok Feb 27 02:06:50 segfault Feb 27 02:06:58 Yeah Feb 27 02:07:12 oh well Feb 27 02:07:21 I don't think this build is going to go well.. Feb 27 02:07:29 :( Feb 27 02:08:06 you're sure I don't need to add the CAF repo? Feb 27 02:08:21 unfortunately it's not built for Ubuntu Touch Feb 27 02:08:27 ah Feb 27 02:08:43 oh hey, test_vibrator is working again Feb 27 02:08:52 Progress! Feb 27 02:09:13 I'm not really sure why that broke the build Feb 27 02:09:16 test_audio get's further than I remember Feb 27 02:09:21 The build step is the exact same Feb 27 02:09:28 GPS WORKS! Feb 27 02:09:35 Oh... wow Feb 27 02:09:38 that's definitely new Feb 27 02:10:05 test_sensors still doesn't find any sensors Feb 27 02:10:07 Na, build failed again Feb 27 02:12:31 well... I'm getting nmea sentences but not coordinates Feb 27 02:12:58 It's something! Feb 27 02:13:04 But no wi-fi, huh? Feb 27 02:13:07 also my phone doesn't have a fever anymore Feb 27 02:13:18 no wifi.... Feb 27 02:13:25 cnss something somethign errors. Feb 27 02:14:36 12-20 00:34:48.004 778 778 E cnss-daemon: bind: Operation not permitted(1) … 12-20 00:34:48.004 778 778 E cnss-daemon: Failed to init nl_loop Feb 27 02:15:00 pretty dang sure that's part of why wifi isn't working Feb 27 02:15:10 Operation not permitted Feb 27 02:15:17 thousand-mile stare Feb 27 02:15:21 yeah. Not sure what to do about it. Feb 27 02:15:26 Got your most recent dmesg snippet? Feb 27 02:16:02 huh. dmesg is spammed with i2c errors Feb 27 02:17:17 dmesg incoming Feb 27 02:17:32 ducks Feb 27 02:18:49 https://gist.github.com/rockybulwinkle/99ae58c87d9ce4172a5efcaf260b27e2 Feb 27 02:18:51 *pew* Feb 27 02:19:15 You're getting exactly what I was hoping you wouldn't Feb 27 02:19:27 yay Feb 27 02:19:42 https://gist.github.com/rockybulwinkle/99ae58c87d9ce4172a5efcaf260b27e2#file-gistfile1-txt-L1773 Feb 27 02:19:51 XD Feb 27 02:20:03 So it's not just the 5X Feb 27 02:20:09 That's... great Feb 27 02:20:09 oh dear. Feb 27 02:20:19 Just think of all the devices you'll fix Feb 27 02:20:29 I did Feb 27 02:20:33 But it broke all the other ones. :P Feb 27 02:20:42 oh, well, that's their problem then ;) Feb 27 02:21:08 what do I need to do to apply your fix? Feb 27 02:21:24 It was actually enabling link-time optimization Feb 27 02:21:35 really? Feb 27 02:21:37 Really Feb 27 02:21:44 that..... doesn't make any sense. Feb 27 02:21:44 Things don't make sense any more. Feb 27 02:22:01 Hmm... maybe you can get this without libc6-dbg Feb 27 02:22:06 we accidentally bork the universe. I just don't know what went wrong Feb 27 02:22:12 Can you do `gdb unity-system-compositor` Feb 27 02:22:14 And `run` Feb 27 02:22:21 I just want the last few lines after it segfaults Feb 27 02:22:29 I've used gdb before :) Feb 27 02:22:48 Well that's good... most haven't. :) Feb 27 02:23:20 I hadn't until recently, debugging QEMU Feb 27 02:23:29 and u-boot Feb 27 02:23:38 Suddenly I don't envy you any more Feb 27 02:23:42 oh, right, disable lightdm Feb 27 02:25:04 specifically, getting a chip's flash to appear at both 0x0 and 0x20000000 AND reflect r/w changes in both regions. Only one other device had attempted it, some Xiling part. after a couple weeks we decided we didn't /REALLY/ need it XD Feb 27 02:25:46 some bug in their TLB emulation. Feb 27 02:28:36 should I run it without the dm in gdb? Feb 27 02:28:42 Yeah Feb 27 02:29:55 https://gist.github.com/rockybulwinkle/1aff5cece2d4f4dc731068af580f8b54 Feb 27 02:30:29 want me to do anything else? Feb 27 02:30:29 Okay, so it's probably the caf problem. Feb 27 02:30:54 No, that's fine Feb 27 02:30:55 Thanks Feb 27 02:31:47 it's trying to load 0x10000000024 when it crashes. So helpful. Feb 27 02:32:13 Well... what's that mapped into? :P Feb 27 02:32:26 I don't even remember what sysfile that map is in Feb 27 02:32:42 It's probably adreno or hwcomposer Feb 27 02:32:46 I am not familiar with this one's memory map Feb 27 02:33:07 wait, that's a virtual address at this point right? Feb 27 02:33:22 >_< Feb 27 02:34:01 what symbol is at 0x7fb5730828 ? Feb 27 02:34:20 it was going to do a bl to there Feb 27 02:34:29 probably was setting up args. Feb 27 02:34:31 ¯\_(ツ)_/¯ Feb 27 02:34:43 So, the reason for the CAF split Feb 27 02:34:59 can you build debug info into it? Feb 27 02:35:23 build debug info / look at the map? Feb 27 02:35:25 Is someone added a value in the middle of a `struct` Feb 27 02:35:47 Actually installing `libc6-dbg` from the repository would probably give the best hint Feb 27 02:35:48 ... and broke binary compatibility with a library? Feb 27 02:36:00 @rockybulwinkle, yep Feb 27 02:36:17 *sigh* Feb 27 02:36:59 And that library is in Android, where anything can be shipped as a binary or source depending on how the manufacturer feels that day Feb 27 02:40:12 gdb's info sharedlibrary doesn't show ANYTHING mapped into that address. So I guess it's unity-system-compositor Feb 27 02:40:21 really, do you have a map or symbol list or anything? Feb 27 02:40:25 Let me pull that up Feb 27 02:40:51 Should be this: https://ci.ubports.com/job/ubports/job/unity-system-compositor/job/xenial/ Feb 27 02:40:57 The ddeb Feb 27 02:41:01 For your architecture Feb 27 02:41:42 what now? is the debug info built into the deb package? Feb 27 02:42:07 You need to install the `.deb` and `.ddeb` for your architecture Feb 27 02:42:17 And... then you've got debug symbols! Feb 27 02:42:22 yay Feb 27 02:42:39 there isn't a ddeb in there Feb 27 02:42:53 am I missing something? Feb 27 02:43:15 Oh great, I'm blind Feb 27 02:43:49 The commit was reverted Feb 27 02:44:49 my phone has a fever again >_< Feb 27 02:45:15 ...and gps stopped working.. what did I do??? Feb 27 02:45:42 Container got sick of you maybe Feb 27 02:46:12 ...it's not running Feb 27 02:46:12 Well, this is the commit that fixed the 5X but broke everything else: https://ci.ubports.com/blue/organizations/jenkins/ubports%2Funity-system-compositor/detail/xenial/4/artifacts Feb 27 02:46:12 It has ddebs Feb 27 02:46:19 hheheh, the container got sick of you. :P Feb 27 02:46:43 >_< and it won't start again! blarg! Feb 27 02:48:03 Reboot... But either way, we need that CAF fix Feb 27 02:48:06 it's started again on a reboot, but gps still doesn't work Feb 27 02:48:14 I'll have to ask Marius about it Feb 27 02:48:17 oh well, come back to that. Feb 27 02:48:56 oh there we go it just needed to get a fix Feb 27 02:49:12 ....and android container died again. Feb 27 02:51:32 Me suggest taking this to ubports (porting) group 😛 Feb 27 02:51:54 bleh. yeah maybe. I should be sleeping soon. **** ENDING LOGGING AT Tue Feb 27 03:00:01 2018