**** BEGIN LOGGING AT Sat Feb 10 03:00:02 2018 Feb 10 03:00:17 `# CONFIG_ANDROID_PARANOID_NETWORK is not set` Feb 10 03:00:33 That's not going to fly if you want to make rild happy Feb 10 03:01:15 But let's wait till we see result of repo sync Feb 10 03:01:59 ssh, and... still dead Feb 10 03:03:04 Alright please toggle the paranoid network and other option I've added in the kernel.. config_uid_gid.... something Feb 10 03:06:54 Oh hey Feb 10 03:07:00 I was touching the wrong kernel config the whole time Feb 10 03:07:10 That's neat Feb 10 03:07:51 Guess I'll, uh, try to enable those configs now Feb 10 03:12:17 `note: expected 'uid_t' but argument is of type 'kuid_t'` … I was wondering when this was going to show up Feb 10 03:13:21 @UniversalSuperBox, So I don't need to rebuild systemimage to test the kernel? Feb 10 03:13:28 just hybris-boot? Feb 10 03:13:33 Correct Feb 10 03:15:19 (Document) https://irc.ubports.com/WJCGXnOq/file_4151.mp4 Feb 10 03:15:27 basically, yeah Feb 10 03:17:07 Ah, it was user namespaces Feb 10 03:17:12 ain't nobody got time for that Feb 10 03:18:14 And I still can't log in Feb 10 03:18:57 Hm Feb 10 03:19:04 (Photo, 735x501) https://irc.ubports.com/4ehOqx0x/file_4153.jpg Feb 10 03:19:06 here it sits for eternity Feb 10 03:19:21 Do you have `logcat -b radio` with ofono disabled? Feb 10 03:19:43 I can't log in 😓\ Feb 10 03:20:13 Disabling ofono did not actually fix it, it just prolonged the time until it happened again Feb 10 03:20:14 But you said with ofono disabled it works? Feb 10 03:20:24 Heh Feb 10 03:20:33 I can move reference back into place Feb 10 03:20:50 Can you fetch e.g. syslog from the rootfs? Feb 10 03:21:00 Well reference works so ignore it Feb 10 03:21:13 That's the thing... Feb 10 03:21:17 syslog doesn't get written Feb 10 03:21:20 Plasma mobile works so ignore it... Let's debig what's broke? Feb 10 03:21:42 There is something majorly wrong with how this thing runs software, it seems. Feb 10 03:22:21 ah.. so your rootfs gets remounted ro Feb 10 03:22:29 Ah ah Feb 10 03:22:42 *boots laptop* Feb 10 03:22:51 No, because `/data/system-data/var/log` exists. Unless `/data` is being remounted Feb 10 03:23:16 Which is possible... Can I write `mount` into the motd? Feb 10 03:23:32 Yeah it essentially ro remounts everything we pass to Android container Feb 10 03:23:55 (Document) https://irc.ubports.com/UMpBndPB/file_4154.mp4 Feb 10 03:24:04 I'm not giving the container `/data` though... Feb 10 03:24:21 It gets `/data/android-data` to play in Feb 10 03:26:10 ` … - exec - root root system -- /system/bin/init.bullhead.qseecomd.sh … + # exec - root root system -- /system/bin/init.bullhead.qseecomd.sh … ` … have you tried this already? Feb 10 03:26:22 Yep, because that script is a pain Feb 10 03:27:04 I don't understand its purpose Feb 10 03:27:10 Does it block init? Feb 10 03:27:58 ```- init_user0 … + #init_user0``` … try this in rootdir/init.rc Feb 10 03:28:10 In `out`? Feb 10 03:28:13 @UniversalSuperBox, tbf, me neither Feb 10 03:28:28 in `system/core/rootdir/init.rc` Feb 10 03:29:15 Okay, rebuilding Feb 10 03:29:43 Replacing... Feb 10 03:30:18 Here we go Feb 10 03:30:41 Still no dice Feb 10 03:31:37 okay let's disable rebooting into recovery shite Feb 10 03:31:44 https://ptpb.pw/Vm6E.diff Feb 10 03:32:00 But it doesn't reboot? Feb 10 03:32:05 (just to debug what's going on) Feb 10 03:32:09 well container does Feb 10 03:32:13 oic Feb 10 03:32:19 and it put system in super weird state Feb 10 03:33:31 Replace dat system image Feb 10 03:34:51 Still nothing :( Feb 10 03:35:07 :/ Feb 10 03:35:57 Wait. Feb 10 03:36:02 Feb 10 03:33 obj/STATIC_LIBRARIES/libinit_intermediates/service.o Feb 10 03:36:13 Clean build it is Feb 10 03:36:34 (It's 9:36 now) Feb 10 03:37:32 Hm Feb 10 03:42:21 Still nothing. Feb 10 03:42:50 Do I need to provide any kernel commandline arguments? I have audit=0 selinux=0 right now. Feb 10 03:47:28 You shouldn't, no. Maybe try removing a `console` argument Feb 10 03:47:43 Since a nonexistent console will cause the kernel to panic in my experience Feb 10 03:48:59 @UniversalSuperBox I er... um, let's try to disable container, ssh in and start container manually to capture dmesg Feb 10 03:49:06 You bet Feb 10 03:49:39 Uh... where did my rootfs go Feb 10 03:49:40 great Feb 10 03:49:47 It was there on the last boot Feb 10 03:49:51 @UniversalSuperBox, Is a console argument implied anywhere? I just have those two args set, audit and selinux Feb 10 03:49:57 it explains why my change didn't work :P Feb 10 03:50:05 @rockybulwinkle, `abootimg -i hybris-boot.img` Feb 10 03:50:13 @bhushanshah, No, it doesn't. Feb 10 03:50:22 Because it wouldn't have given me the motd without the rootfs. :P Feb 10 03:50:39 well Feb 10 03:50:46 it rm'ed after that Feb 10 03:50:55 so it rebooted container neverthless Feb 10 03:51:03 Oh damn, I didn't delete .writable_image Feb 10 03:51:39 Android: The only place where `rm -rf /` is the only sane thing to do when a service fails. Feb 10 03:54:10 Now waiting for `halium-install` to do its thing Feb 10 03:56:58 *yawns* Feb 10 03:57:02 all done Feb 10 03:57:27 Alright, we have a rootfs. Feb 10 03:57:37 @UniversalSuperBox in other world : https://i.imgur.com/SsuFuFp.jpg I enabled gst-droid gst-omx stuff for PlaMo Feb 10 03:57:49 oh look a camera Feb 10 03:58:03 nice laptop Feb 10 03:58:16 Dell Inspiron Feb 10 03:58:21 nice powerline Feb 10 03:58:47 Moment of truth. Will the thing be done? Feb 10 03:59:30 Gues who didn't enable ssh Feb 10 03:59:31 What the hell. Connection refused? Feb 10 04:03:32 Fwd from Flohack: porting is a strange sadism Feb 10 04:03:50 At least dropbear is okay with the rootfs being read-only Feb 10 04:04:24 So.. Back to where we started. No logs Feb 10 04:04:34 @UniversalSuperBox, https://gist.github.com/rockybulwinkle/2ac7f8e28348fccdaa0a0f9acde5fae5 Feb 10 04:04:45 Your cmdline is there Feb 10 04:04:59 You need to pass all that in addition to your custom parameters when using `fastboot boot -c` Feb 10 04:05:03 Ok, I'm back, here is my manifest wothout android_policy, I'm syncing the repos Feb 10 04:05:05 pastebin.com/uUM7Z2iv Feb 10 04:05:41 Oh, right! Disabling Android container. Feb 10 04:05:45 I'm posting the error when in a few minutes Feb 10 04:07:11 $ ssh phablet@10.15.19.82 … ssh: connect to host 10.15.19.82 port 22: Connection refused Feb 10 04:07:14 At this point it's just mocking me. Feb 10 04:07:44 FYI dropbear and ssh can't co-exist Feb 10 04:07:51 And maybe it's missing host key Feb 10 04:08:00 If rootfs was never rw Feb 10 04:08:11 That's in the writable section. let me see if it made it. Feb 10 04:08:40 Yep, `/etc/ssh` is filled with stuff. Feb 10 04:09:17 @UniversalSuperBox, fastboot boot -c is that a different way then to boot the kernel without flashing? Feb 10 04:09:23 Yes Feb 10 04:11:02 @UniversalSuperBox, can you give me instructions to test stuff? I might give it a whirl Feb 10 04:12:09 on halium-boot side I am stuck on vendor symlinks Feb 10 04:12:36 I made a tweak to the pre-start.sh script now Feb 10 04:12:39 It's on the gist. Feb 10 04:12:44 Replace yours with that one, should do it right up Feb 10 04:14:21 okay Feb 10 04:14:36 and I am bit out of date on halium-boot instructions.. Feb 10 04:14:46 is instructions in the ticket up-to-date? Feb 10 04:14:51 Well, if Travis would just deploy to reelases like it's supposed to Feb 10 04:15:16 I used a dash where I should have used an underscore. Feb 10 04:15:37 Well, at least it was a simple fix. kicks self Feb 10 04:17:23 https://github.com/Halium/projectmanagement/issues/40 so s/ubports-boot/halium-boot/g ? Feb 10 04:17:35 Basically yeah Feb 10 04:17:50 Wait, the releases provider isn't working Feb 10 04:18:03 So the initrd isn't on GitHub where it should be. Let's see, I have a local copy here... Feb 10 04:19:01 (Document) https://irc.ubports.com/4S0221Wz/file_4155.gz Feb 10 04:19:13 Toss this file into `/device/lge/bullhead/initramfs.gz`, add `BOARD_USE_LOCAL_INITRD := true` to BoardConfig.mk, then you're good to go Feb 10 04:19:51 ok Feb 10 04:20:55 I'm still thinking I'm going to remove the `ubports` boot mode. Feb 10 04:21:06 It has no place if the system.img is updated independently of the rest of the system. Feb 10 04:21:46 oh also what rootfs you're building? Feb 10 04:22:05 https://ci.ubports.com/job/xenial-rootfs-armhf/lastSuccessfulBuild/artifact/out/ubports-touch.rootfs-xenial-armhf.tar.gz Feb 10 04:22:17 okay Feb 10 04:22:19 535MB, jeez those are getting weighty Feb 10 04:22:43 I'd recommend using `wget` to download that file. Feb 10 04:22:50 Jenkins doesn't like serving files very much. Feb 10 04:23:22 I use aria2c with jenkins Feb 10 04:23:38 TIL Feb 10 04:24:02 also maybe add build step to zsyncmake so if download is corrupt it can be fixed Feb 10 04:24:13 does the build output where the symbols are? Feb 10 04:24:34 Okay, updated #40 Feb 10 04:25:03 `[#767488 19MiB/535MiB(3%) CN:4 DL:2.1MiB ETA:4m]` … wee Feb 10 04:25:23 That's faster than I can do it Feb 10 04:25:25 I'm jealous Feb 10 04:29:36 so... it's crashing at qcom_ice_setup_ice_hw Feb 10 04:34:58 https://gist.github.com/rockybulwinkle/b2cdfa892f1a313c01647cc7a1dab53d Feb 10 04:46:25 Bah, ssh just won't behave. I give for now. Feb 10 05:29:49 Im going to try something else for now. Does anyone know here I can get systemimages for nexus 5? Feb 10 10:15:55 So when I'm modifying the kernel , Feb 10 10:16:00 I run the mer kernel check Feb 10 10:16:18 and then I fix all of the errors it throws out right ? Feb 10 10:16:37 Most of the lines don't exist so I just add them to the defconfig ? Feb 10 11:11:55 (Photo, 1280x960) https://irc.ubports.com/zRCvd5so/file_4159.jpg Systemimage finally started building. Turns out I was using legacy vendor drivers Feb 10 11:16:16 copy and paste is seriously broken on the ubuntu desktop? Feb 10 11:46:08 after many debugs lxccontaine crash and this is error … Unable to open '/dev/cpuset using monitor socket name "lxc/ad055575fe28ddd5//var/lib/lxc" (length of socket name 33 must be <= 105) Feb 10 11:46:44 i enablebed CONFIG_CGROUP_CPUSET=y, but no impact Feb 10 12:15:34 I cannot install rootfs :/ Feb 10 12:16:12 what's the issue Feb 10 12:19:07 I use TWRP Feb 10 12:19:34 and the halium install needs adb in recovery mode Feb 10 12:20:16 ugh so? Feb 10 12:20:22 twrp has adb in recovery mode Feb 10 12:21:03 twrp has adb in sideload mode Feb 10 12:21:09 you are wrong Feb 10 12:21:15 I'm looking for other options Feb 10 12:21:20 TWRP has adb in recovery mode and all of us use it Feb 10 12:21:32 i use twrp daily and it has freaking adb Feb 10 12:21:33 The repo sync took more time that I expected, this is the error, thanks Feb 10 12:21:35 pastebin.com/q0EVqnYq Feb 10 12:21:36 It appears as unauthorized Feb 10 12:21:48 Im not really sure why Feb 10 12:22:04 it's more of a local issue Feb 10 12:22:07 fix your workspace Feb 10 12:22:09 I have not downloaded android_policy repo and I didn't get the error of last time Feb 10 12:22:23 alright Feb 10 12:22:25 I mean, the error I got without that repo Feb 10 12:22:28 and don't blame twrp -_- Feb 10 12:22:37 we all use twrp here. myself included Feb 10 12:22:48 you misunderstood me Feb 10 12:22:54 I wasn't blaming twrp Feb 10 12:23:17 I didn'nt know it had a normal adb recovery option ... because it was unauthorized for me Feb 10 12:23:20 Im sorry :) Feb 10 12:23:26 @vanyasem, Are you talking to me? How do I do that? Feb 10 12:23:45 @Pablo Alejandro, no, i am not Feb 10 12:23:53 you got in the middle of my conversation with Panzerox Feb 10 12:23:56 Ah, ok Feb 10 12:24:04 I'll así later then Feb 10 12:24:07 Ask* Feb 10 12:24:33 @Pablo Alejandro, what's the rror Feb 10 12:24:49 the output is normal Feb 10 12:24:54 it's totally fine Feb 10 12:24:59 what's wrong exactly? Feb 10 12:34:06 OKay. So i got it working with an older version of twrp Feb 10 12:34:15 But now I'm faced with this issue Feb 10 12:34:22 error: ext4_allocate_best_fit_partial: failed to allocate 837 blocks, out of space? Feb 10 12:37:45 Isn't this an error "Trying dependencies-only mode on a non-existing device tree?" Feb 10 12:38:26 Nope Feb 10 12:38:31 well it is error, yes, but that alone doesn't say anything Feb 10 12:43:39 @bshah, Do you mean without the resto of the log? Here is it pastebin.com/q0EVqnYq Feb 10 12:43:57 @Pablo Alejandro, you can ignore that in most cases Feb 10 12:44:41 also, that log looks perfectly fine Feb 10 12:44:46 what is your current issue? Feb 10 12:45:55 Only that, I thought it would be a problem, thanks Feb 10 12:46:13 I'll continue at the evening Feb 10 12:46:15 actually i was scared by that as well at first bshah Feb 10 12:46:23 we should probably mention that in docs Feb 10 12:50:52 @KunalBhat, anyone ? Feb 10 12:51:02 are you out of space? Feb 10 12:51:12 if not, then try JBB's install script Feb 10 12:51:17 https://github.com/JBBgameich/halium-install Feb 10 13:22:34 Can someone review & merge this please: https://github.com/Halium/hybris-boot/pull/41 Feb 10 13:30:20 Thanks @bhushanshah Feb 10 14:16:54 I can access telnet Feb 10 14:17:01 But I have no idea what to do next Feb 10 14:24:49 I'm tryign to add udev rules , but I get this error Feb 10 14:25:03 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//' >/etc/udev/rules.d/70-z3c.rules Feb 10 14:25:06 In telnet you don't havee to add udev rules Feb 10 14:25:17 Oh... Feb 10 14:25:18 cat /diagnosis.log in telnet Feb 10 14:25:20 I'm sorry Feb 10 14:25:26 new to this Feb 10 14:25:33 you are still far from adding udev rules :) Feb 10 15:20:35 I get "`/mnt/foo/halium-dev/out/target/product/axon7/obj/KERNEL_OBJ/drivers/crypto/msm/ice.o': can't read symbols: File format not recognized." when I try to open the offending object file with arm-none-eabi-gdb from the AUR. Feb 10 15:22:15 nvm, need to use aarch64 gdb, not arm Feb 10 15:58:10 @KunalBhat, Add this in from ssh terminal Feb 10 17:20:02 Ratattuile was added by: Ratattuile Feb 10 18:18:14 I've updated my manifest.xml and downloaded the additional code again using force sync Feb 10 18:18:26 Maybe it's now working Feb 10 18:20:08 Output here: https://pastebin.com/Dzqe2NWs Feb 10 18:22:00 (Voice, 0s)https://irc.ubports.com/AGPyYtNv/file_4166 Feb 10 18:46:58 @manu you should disable the doze target, it's not needed. Feb 10 18:47:20 Okay how to do that? Any command? Feb 10 18:47:38 You have to find where it's being included in the .mk files Feb 10 18:48:51 I may be misremembering Feb 10 18:50:47 Fwd from vince1171: In /vendor/your/device/Android.mk … Remove all code with .apk our .jar in Feb 10 18:51:20 Okay Feb 10 18:57:11 So just replace .apk with .jar? Feb 10 18:57:47 no, remove all the apk and jar packages listed there. Feb 10 18:57:55 Ah okay Feb 10 19:08:16 Well, I'm back and I'm stuck here docs.halium.org/en/latest/porting/build-sources.html#modify-the-kernel-configuration, I don't find the kernel configuration nor the folders descibed in there Feb 10 19:08:43 I had some trouble finding my kernel config too Feb 10 19:08:59 and did you find it? Feb 10 19:09:21 well.. I didn't see the line in my boardconfig, but did find it over in the kernel Feb 10 19:10:40 kernel////arch/arm64/lineageos__defconfig Feb 10 19:10:46 executing grep ""TARGET_KERNEL_CONFIG" device/leeco/zl1/BoardConfig.mk" gives me "TARGET_KERNEL_CONFIG := lineage_zl1_defconfig" Feb 10 19:10:56 ok Feb 10 19:11:19 let me see Feb 10 19:11:23 the kernel configs are in arch//yconfigs in the kernel director Feb 10 19:11:31 in my PC, i'll search Feb 10 19:11:51 @JBBgameich Hi jbb. I've tried to apply your 3 patches but during compilation i got some errors. I've tried to resolve them but nothing... Can you help me? Feb 10 19:12:25 Sure we can help, without the errors Feb 10 19:12:39 is the kernel folder inside the buildir? Feb 10 19:13:01 yes... look in your manifest, you'll find where you put the kernel repo there. Feb 10 19:13:49 ok, thanks Feb 10 19:13:54 @Mattia990, Please send the errors :) Feb 10 19:15:42 https://pastebin.com/NZgVPsbP Feb 10 19:15:51 Have I to remove the whole line? Feb 10 19:15:57 Or what is to do? Feb 10 19:17:29 @Ellenjott, Ok 1 sec... Thanks for help! Feb 10 19:18:19 @Ellenjott, https://paste.ubuntu.com/=MC8XcdQqjC/ … Here you are! Feb 10 19:18:36 @manu you should remove the whole sections. I'll give you a before and after with mine if you like Feb 10 19:18:47 Okay thanks Feb 10 19:19:09 @Mattia990, Link to kernel repository or manifest? Feb 10 19:20:35 @manu my before: https://github.com/TheMuppets/proprietary_vendor_zte/blob/cm-14.1/axon7/Android.mk Feb 10 19:21:04 @Ellenjott, Ehm i haven't got repository on github... only on my pc. If you want i can send you the file that you need Feb 10 19:21:10 after: https://gist.github.com/rockybulwinkle/b987f33acf7ea3072c959edb289735cc Feb 10 19:23:01 lol, looking back at it again looks like I missed a jar. Feb 10 19:23:18 @Mattia990, You can search for the definition of `'struct mnt_namespace' has no member named 'seq'` and see if it's inside an `#ifdef`; if so, then enable the option (e.g. there is `#ifdef CONFIG_ABC`, then enable `CONFIG_ABC` in your defconfig). Feb 10 19:23:51 @Mattia990, And you should publish the kernel, including the git history Feb 10 19:25:14 Hey! I'm telnetted in to my device! I decided to try the stock lineageos config for the kernel. Feb 10 19:25:21 @Ellenjott, I can't do more since you haven't published the kernel Feb 10 19:25:34 @rockybulwinkle, It won't work, you have to adapt it for halium Feb 10 19:25:41 I'm sure I do. Feb 10 19:25:58 If you use the default lineageos config, you won't get farer than telnet Feb 10 19:26:12 @Ellenjott, Thanks a lot! Tomorrow i Will try to publish all on github Feb 10 19:26:21 @Mattia990, 👍 Feb 10 19:26:31 @Mattia990, And publish the manifest at halium-devices Feb 10 19:26:36 I understand. But with a working (telnet) starting point it makes it easier for me to figure out where it goes wrong.... Feb 10 19:27:06 @rockybulwinkle, No I mean this … device/samsung/qcom-common/Android.mk:36: *** missing separator Feb 10 19:27:08 @rockybulwinkle, mer check kernel config will also tell you what's wrong Feb 10 19:27:13 But thanks Feb 10 19:27:41 some of the things mer check kernel wanted me to set weren't available in menuconfig. The kernel is new enough Feb 10 19:28:23 the crashes were within qualcom's hardware UFS crypto engine but I couldn't figure out how to disable it Feb 10 19:29:23 @rockybulwinkle, Just enable them manually in the defconfig file Feb 10 19:29:51 ok. Even though no one is using those configs in the kernel source? Does hybris-boot use the .config in the kernel fs to read those settings or something? Feb 10 19:30:58 @rockybulwinkle, You compile your kernel with a certain set of features, if you miss something, hybris-boot, systemd, lxc container, etc. will fail Feb 10 19:31:34 @rockybulwinkle, and edit the defconfig in `arch/arm/configs/`, not your local .config file! (important) Feb 10 19:31:42 Look, I'm not new to compiling a kernel. But the features _aren't there_. Feb 10 19:32:30 @rockybulwinkle, Ok, then you can omit them Feb 10 19:33:11 @rockybulwinkle, What are the errors? Feb 10 19:33:30 give me a sec, I'll give you the logs Feb 10 19:33:35 +1 Feb 10 19:34:11 @rockybulwinkle, And please also the kernel repository or manifest Feb 10 19:34:18 ok Feb 10 19:35:31 I may take a little bit to get back to where I was. Feb 10 19:35:53 Ok Feb 10 19:38:25 @Manu, Can you give me a link to the exact repository you're using? If you use halium-7.1, you should replace the repository in your manifest with this one: https://github.com/Halium/android_device_samsung_qcom-common Feb 10 19:39:52 I used it from LineageOS, wait Feb 10 19:40:45 @Manu, https://github.com/LineageOS/android_device_samsung_qcom-common Feb 10 19:41:59 @Manu, Have you modified anything, because you get an error in a line, that doesn't exists on the cm-14.1 branch? Feb 10 19:42:22 Ahm no Feb 10 19:42:58 @Manu, Compare your file to this: https://github.com/LineageOS/android_device_samsung_qcom-common/blob/cm-14.1/Android.mk Feb 10 19:44:37 Okay that's slightly different Feb 10 19:45:57 @Manu, enter the directory of the repository and execute `git checkout -- .` Feb 10 19:46:13 Okay Feb 10 19:47:08 Or better just replace the repository in your manifest with this one: https://github.com/Halium/android_device_samsung_qcom-common (hal remote) Feb 10 19:51:20 @Ellenjott Sorry... JBB said me to apply this patches: https://github.com/LineageOS/android_kernel_sony_apq8064/commit/b3bbbf6ce0bd76076e4e02bc7a2de90ea523535e … https://github.com/LineageOS/android_kernel_sony_apq8064/commit/7c9f141d37b833a99f2bf27f9793cd061c55c2e6 … https://github.com/LineageOS/android_kernel_sony_apq8064/commit/9e920c6371fb5aefd35b45cbb58d5c307db6ebda … But the first and the latest are in contrast... which of them i must Feb 10 19:54:04 @Mattia990, I'm not sure, but try it without any patches. You've got a 3.18 kernel, right? (then the patches aren't needed anyway) Feb 10 19:54:24 @Ellenjott, Sorry, I was wrong on one count, with CONFIG_NETPRIO_CGROUP, it does appear in source, but not in the Kconfig's. There are many config options that don't show up anywhere though. Feb 10 19:54:50 @rockybulwinkle, np Feb 10 19:54:51 @Ellenjott, No, my kernel is 3.4 Feb 10 19:55:10 @Mattia990, Oh sry ... Feb 10 19:56:57 @Mattia990, Do you use the apq8064 kernel or another? If you use another, you probably don't need the patches Feb 10 19:57:53 (That are reverts of partial back ports and only specific to the lineageos version of the sony apq8064 kernel) Feb 10 19:59:50 @Ellenjott, My device is xiaomi mi 3 (Cancro) … I got ssh but i have got a problem with rsyslog and some kernel configuration which are missing with lxc-checkconfig and on my defconfig are present Feb 10 20:00:54 @Mattia990, If lxc-checkconfig isn't satisfied, you're still missing something Feb 10 20:01:17 So I added a remote to the halium github because the compiler didn't recognize the "hal" remote Feb 10 20:01:34 @Mattia990, (But you don't need everything of lxc-checkconfig) Feb 10 20:02:16 @Manu, Use this in your manifest: … `` Feb 10 20:06:07 @Mattia990, Which are exactly missing? Feb 10 20:07:33 @Ellenjott, Ehm... All this … CONFIG_NF_NAT_IPV4: missing … CONFIG_NF_NAT_IPV6: missing … CONFIG_IP6_NF_TARGET_MASQUERADE: missing … CONFIG_PACKET_DIAG: missing … CONFIG_NETLINK_DIAG: missing Feb 10 20:07:46 @Mattia990, But on defconfig are "y" Feb 10 20:08:57 @Mattia990, These are not important (only necessary if you want your drivers to be able to call their home and spy you) Feb 10 20:09:26 Just remove this repo in the first line? Feb 10 20:09:29 https://pastebin.com/ShmAeMMs Feb 10 20:09:35 @Ellenjott, binary blobs, you never know 👀 Feb 10 20:10:18 @Ellenjott, Ah ok... The last problem that i've got is this... Feb 10 20:10:30 @Manu, I don't think you need that repo (vendor/qcom) Feb 10 20:10:51 @Ellenjott, use vendor/sony or vendor/samsung or what ever Feb 10 20:10:53 Okay Feb 10 20:12:00 @Ellenjott, Ahh ok... Another problem that i've got is this... … Everytime i boot my phone, rsyslog starts up and it occupies 200% of cpu and my rootfs fills up... so i haven't got space (Sorry for my bad english) Feb 10 20:13:00 @Mattia990, Which rootfs do you use? If you use the ubports rootfs, you'll need to resize, to get more space than 2 GB Feb 10 20:13:17 @Ellenjott, No no, i use halium rootfs Feb 10 20:13:33 @Mattia990, np, my English isn't very good, too ... Feb 10 20:14:02 @Mattia990, And there's no space left? Feb 10 20:16:35 @Ellenjott, No, i've increased to 3Gb but nothing... Everytime rsyslog 180% - 200% cpu and it creates two files in /var/log called syslog and kern.log ... In both file there is a looping dmesg (i don't know if you understand) from 0.00000 to 800.00000 and then loop Feb 10 20:17:28 @Mattia990, ah ok, so your dmesg gets so big instantly ... Feb 10 20:18:06 You could theoretically disable some kernel configs to make it less verbose, but then you proably also can't find the error Feb 10 20:19:20 @Ellenjott, So it can be a kernel problem or halium(rootfs) problem? Feb 10 20:19:41 Shouldn't be a problem caused by the rootfs Feb 10 20:19:45 halium/devices/manifests/[manufacturer]_[device].xml Feb 10 20:19:47 What do we do in this file? Feb 10 20:20:22 @furylions, #readthedocs http://docs.halium.org/en/latest/porting/get-sources.html#adding-your-device-specific-source Feb 10 20:47:28 @furylions, In short: you tell repo which repositories you need and it has to clone Feb 10 20:47:28 @Ellenjott, Nice! It works now! Feb 10 20:47:28 I ask because I do not understand. Feb 10 20:47:28 @Manu, Cool! :) Feb 10 20:47:28 I'm understand, thank you ! Feb 10 20:47:28 I didn't understand it some weeks ago and now I'm Building halium for my device😁 Feb 10 20:47:28 @Manu, The best part still follows .... DEBUGGING! 🙈 **** BEGIN LOGGING AT Sat Feb 10 20:48:42 2018 Feb 10 20:48:42 the ip will be the same as with telnet? What's the user, root? Feb 10 20:49:30 I think I forgot the kernel modification😅 Feb 10 20:49:36 @rockybulwinkle, No, the ip is different. That's also described in the docs Feb 10 20:49:42 How to use this command Feb 10 20:50:43 afk, taking my dog to the park :3 Feb 10 20:52:32 Thanks again for the help, you all are awesome. Feb 10 20:53:01 Okay got an output lineageos_jf_defconfig is the file called Feb 10 20:54:05 But where is the file located? Feb 10 20:56:51 I'm un the dame situation as you Feb 10 20:57:36 Ya seems so Feb 10 20:58:42 I did a search on the builddir and I found anything Feb 10 20:58:51 Okay Feb 10 20:58:56 (Photo, 1920x1080) https://irc.ubports.com/66w0zZSU/file_4169.jpg Feb 10 20:59:20 Okay Feb 10 21:00:11 Is the file located in one of the downloaded repos? Feb 10 21:00:36 or it should be generated? Feb 10 21:00:54 the defconfig is located in the kernel repository Feb 10 21:01:42 and that repository is from los or them? Feb 10 21:02:03 and thanks for the help, you guys are awesome Feb 10 21:05:20 kernel repositories are from los Feb 10 21:05:54 Ok, I think I found them, I'll sync them Feb 10 21:06:03 if you know the name of the kernel config you can find in arch/arm/configs in your kernel repo Feb 10 21:07:33 I had not downloaded it yet :p Feb 10 21:08:27 is not on the lineage.dependences file of zl1 repo, is on the android_device_leeco_msm8996-common's lineage.dependences file Feb 10 21:08:46 Manu Did you found the kernel config? Feb 10 21:08:58 Yep Feb 10 21:09:33 @Pablo Alejandro, well all of these common devices use the same kernel ... Feb 10 21:10:08 yep, i wasn't concerned of that, is my first time doing something like this xd Feb 10 21:10:33 i didn't kow about that* Feb 10 21:11:26 @Pablo Alejandro, :) Feb 10 21:12:38 i didn't mean concerned, I trying to mean that I didn't knew about that, sorry for bad english Feb 10 21:13:20 I'm also adding some repos from lineage.dependences of leeco_msm8996_common repo Feb 10 21:13:46 @Pablo Alejandro, No problem as far as we can understand each other :D Feb 10 21:14:09 @Pablo Alejandro, 👍 all dependencies should be added :) Feb 10 21:15:27 agreed :) Feb 10 21:23:39 What is if some values aren't in the config file? Feb 10 21:24:28 Just add them? Feb 10 21:27:54 @Manu, Yeah Feb 10 21:28:05 Okay Feb 10 21:35:38 @bhushanshah you can now test halium-boot in all its glory Feb 10 21:36:03 And the 5X is still terrible. That is all. Feb 10 21:36:15 Hm Feb 10 21:36:19 @UniversalSuperBox, Hm Feb 10 21:36:33 What? :P Feb 10 21:38:32 mer_kernel_verify_config gives me this output https://pastebin.com/2CA6NB9P, is this the same as Manu? should I add them? Feb 10 21:38:48 Set all the options the way it wants Feb 10 21:39:16 I think so, I'll begin to add them Feb 10 21:39:45 (or wait some time and use the ubports script with halium-boot ....) Feb 10 21:40:16 Hmm... @bhushanshah did we have some config conflicts between halium and ubports, Feb 10 21:40:29 We did iirc Feb 10 21:40:36 @Ellenjott, ? Feb 10 21:40:38 Well ok systemd things are missing Feb 10 21:41:12 Well, I'll run the mer script after I run ours and see what happens Feb 10 21:41:28 Once I'm back in front of the tree Feb 10 21:41:55 @UniversalSuperBox, I once did that with my yuga, resulting in that the lxc didn't want to start ... Feb 10 21:42:21 That seems wrong Feb 10 21:43:11 I think the ubports script is setting more than the mer verifier will check ... Feb 10 21:43:34 @Ellenjott, So better forget that Feb 10 21:59:23 https://pastebin.com/FtkKn9am Feb 10 21:59:31 What should I do now? Feb 10 22:03:06 @Manu, You made a syntax mistake in fixup mountpoints Feb 10 22:03:31 Okay where do you think? Feb 10 22:04:12 @Manu, good question Feb 10 22:04:22 @Ellenjott, Ya Feb 10 22:04:40 Hello, I got the following error while running `mka ubports-boot` (that should be related with fstab.qcom modification of /data, as you can see in https://pastebin.com/Ermw5E5S): https://pastebin.com/bdXAWGLM Feb 10 22:05:42 I can confirm, reading previous case of the same error, that /data was not duplicated. I was wondering how to manage fixup-mountspoints to make it congruent Feb 10 22:06:22 @stefanomelchior, You probably have multiple fstab files in your device tree, just remove all others than qcom.fstab Feb 10 22:06:33 hybris-boot, system.img and rootfs.img from halium works pretty well Feb 10 22:07:18 It's the same code that will throw the error in ubports-boot and hybris-boot … `*** There should be a one and only one device entry for HYBRIS_BOOT_PART and HYBRIS_DATA_PART.` Feb 10 22:08:35 you mean I should delete or overwrite those in out/.../fstab.qcom : … `guf@Guf-Book:/media/HALIUM/halium-new$ find . -iname fstab.qcom … ./device/bq/tenshi/rootdir/etc/fstab.qcom … ./out/target/product/tenshi/root/fstab.qcom … ./out/target/product/tenshi/obj/ETC/fstab.qcom_intermediates/fstab.qcom … ./out/target/product/tenshi/recovery/root/fstab.qcom` Feb 10 22:08:44 Maybe do `find -iname "*fstab*" device/` to locate all fstabs Feb 10 22:08:55 @stefanomelchior, oh ok Feb 10 22:09:03 you only got one fstab ... Feb 10 22:09:12 to do this I imagine I should perform `mka clean` Feb 10 22:10:17 Just a guess, but maybe there are mutiple partitions mounted as data in the fstab file Feb 10 22:10:43 no, there is only that on mcc*46 Feb 10 22:10:47 sry, no not the case Feb 10 22:11:45 I just noticed this error comes from hybris-boot, not ubports-boot ... Feb 10 22:12:03 `halium/hybris-boot/Android.mk:71: *** There should be a one and only one device entry for HYBRIS_BOOT_PART and HYBRIS_DATA_PART.` Feb 10 22:14:02 Is the tenshi manifest in halium-devices up to date? Feb 10 22:14:24 It should be, I just updated it today Feb 10 22:14:28 yep, I followed the steps on `https://docs.ubports.com/en/latest/porting/building-ubports-boot.html` Feb 10 22:15:04 as I said halium's hybris-boot, system and rootfs are working Feb 10 22:15:48 now, I am supposed to create ubports-boot to replace hybris-boot to work with a ubports's rootfs Feb 10 22:16:08 that was the output of mka ubports-boot Feb 10 22:16:35 Just to be sure, have you included ubports boot in build/main.mk Feb 10 22:16:46 I suppose it is almost based on hybris-boot Feb 10 22:17:24 yep, done `subdirs += \ … halium/hybris-boot \ … halium/ubports-boot` Feb 10 22:18:33 @stefanomelchior, it is Feb 10 22:20:50 For me it just works: https://pastebin.com/raw/BfQnaFWT Feb 10 22:21:09 Maybe try `mka clean` Feb 10 22:21:29 After changing that, re-source and do breakfast again Feb 10 22:21:44 A change to main.mk requires the environment get set up again Feb 10 22:22:00 Same failure I can't find any syntax in this file Feb 10 22:22:18 Rerun breakfast Feb 10 22:22:26 Manu @stefanomelchior Feb 10 22:25:23 my device has a src of swap, should I add it to fix-mountpoints? Feb 10 22:25:42 done: `source build/envsetup.sh` … `breakfast tenshi` Feb 10 22:25:58 @Pablo Alejandro, No need Feb 10 22:26:06 ok, thanks Feb 10 22:26:13 @stefanomelchior, does it work now? Feb 10 22:26:17 and `mka clean` then ka ubports-boot` -> same error Feb 10 22:27:18 Do this: … * `mka clean` … * `./halium/devices/setup tenshi` … * `source build/envsetup.sh` … * `breakfast tenshi` … * `mka ubports-boot` Feb 10 22:37:51 I get this error executing make mkbootimg pastebin.com/gVjFiZLF should I care about that? Feb 10 22:38:39 ok, panzerox had the same error, I'll read Feb 10 22:41:07 @Pablo Alejandro, You can comment out the complete `doze` thing in your common device tree Feb 10 22:42:24 ok, thanks :) Feb 10 22:43:04 Or only comment out this line … https://github.com/LineageOS/android_device_leeco_msm8996-common/blob/cm-14.1/doze/Android.mk#L33 Feb 10 22:46:30 Have I to be root to build Halium? Feb 10 22:47:34 No, don't use root when not neccesary. It would destroy the permissions of files inside your Halium tree Feb 10 22:47:47 Okay Feb 10 22:48:47 @Ellenjott, Thanks Feb 10 22:50:47 now it output vendor/nxp-nfc/opensource/frameworks/Android.mk:23: build/core/java_library.mk: No such file or directory Feb 10 22:51:11 pastebin.com/SBmD0d55 Feb 10 22:51:25 should I comment too? Feb 10 22:51:42 Yes, you can comment out everything that's apk or java related Feb 10 22:52:48 Excellent, thanks :D Feb 10 22:54:16 @Ellenjott same error even though I ran all the command you suggested to do Feb 10 22:55:22 @stefanomelchior, Delete the whole `out` folder and run the commands again or completely recheckout your halium tree but this is the very last thing I can think of ... Feb 10 22:55:47 I get some errors running mka hybris-boot Feb 10 22:55:48 https://pastebin.com/x1veQmKK Feb 10 22:56:09 @Ellenjott, Or run `repo diff` first so you see what your local changes that @Ellenjott doesn't have are Feb 10 22:56:33 @stefanomelchior, Or does halium/devices/setup give any errors? Feb 10 22:56:48 @Manu, Still the same fixup-mountpoints file? Feb 10 22:57:28 Yes/No I've tried to make some changes but the same content Feb 10 22:57:45 @Manu, Can you send the new one (to be sure) Feb 10 22:58:06 (Document) https://irc.ubports.com/KBtCn10M/file_4170 Feb 10 22:58:10 Here u are Feb 10 22:59:01 @Manu, Oh, you used two spaces, instead of one Feb 10 22:59:26 between partition name and partition id (mmcblk0pXX) Feb 10 23:01:25 Ah okay Feb 10 23:02:10 @Ellenjott, I've also tried to modify rsyslog.conf but nothing Feb 10 23:03:16 @Ellenjott yep, a lot, so I needed to do a `repo sync --force-sync kernel/bq/msm8937` that makes my repository a mess Feb 10 23:03:36 That's because I moved the kernel source to a new repository Feb 10 23:04:04 The error won't appear again after syncing once Feb 10 23:04:08 @stefanomelchior, a force sync on the kernel could delete your changes to the defconfig, if they aren't uploaded yet Feb 10 23:04:27 I imported the defconfig from the issue into the kernel already Feb 10 23:04:31 @Mattia990, Wait, I'll send you the configs you can disable .... Feb 10 23:05:09 no problem, everything backed up! so, I am working to order the things in place and create 2trees, one for pure halium and the other for halium+ubports Feb 10 23:05:32 @Mattia990, This might make the log doesn't spam your disk ... … `CONFIG_ANDROID_RAM_CONSOLE=n` … `CONFIG_ANDROID_RAM_CONSOLE_ENABLE_VERBOSE=n` Feb 10 23:06:15 meanwhile I got working hybris-boot and hybris-recovery lost, but this is only a matter of reconstruct things done Feb 10 23:07:59 Executing "make hybris-boot" is normal that Building kernel takes more than 8 minutes? I have a intel core i7 4510U Feb 10 23:08:08 So I'm syncing all the repositories in a 2nd directory and then add the additional files because I get some errors like don't have permissions Feb 10 23:08:15 @Ellenjott, Thanks a lot!!! Feb 10 23:09:03 @Pablo Alejandro, Yeah, I've got an i7-7567U and it also takes about 10 mins Feb 10 23:09:54 @Mattia990, You're welcome :) (I hope this helps at all, I'm just kinda trying some things out) Feb 10 23:09:55 Oh, okay then, thanks Feb 10 23:11:50 @Ellenjott, I know but only few people are like you (helpful and kind). So i must thank you! Feb 10 23:12:46 ☺️ Feb 10 23:15:48 ugh, now I'm getting this error pastebin.com/ASDZr38d Feb 10 23:18:06 Should I add "CONFIG_OF=y" to kernel config? Feb 10 23:18:16 @Pablo Alejandro, Try out! :) Feb 10 23:21:59 It gave me I think the same error :c pastebin.com/E006Dr1v I'll see what else can I do Feb 10 23:24:46 So, I've installed the halium rootfs and changed the password like in the guide. I don't see a new network interface though in "ip addr". Looking at dmesg: https://gist.github.com/rockybulwinkle/c43fda80d84fc55cb7583544f3da5016 Feb 10 23:24:58 I have a usb modem attatched according to dmesg Feb 10 23:27:34 @rockybulwinkle, That's not the normal system .... and also not telnet, so better reboot into recovery and try to fetch the last kmsg Feb 10 23:27:43 ok Feb 10 23:28:14 Could be the kernel panic handler Feb 10 23:29:10 Does the device normally restart on panic? Mine didn't. Feb 10 23:29:52 Some devices have a debug software built-in that catches kernel panics and allows you to debug Feb 10 23:33:15 https://gist.github.com/rockybulwinkle/7e6c04904d04c316fe9ca42876a4f83c Feb 10 23:36:12 @Pablo Alejandro, Try to add this to your defconfig `CONFIG_PFK=y` Feb 10 23:36:44 I think I already fixed the issue adding CONFIG_MODULES=y Feb 10 23:36:47 thanks Feb 10 23:37:01 Ah, ok Feb 10 23:37:47 Is building the kernel and is taking more time than last time, is that why I think that part is fixed :) Feb 10 23:38:12 @Pablo Alejandro, we will see Feb 10 23:40:41 @rockybulwinkle, The only thing I can read from that, is that your init system got killed ... :/ Feb 10 23:42:34 yeah... Feb 10 23:42:45 It didn't show anything before that. It died when the cpu entered idle... Feb 10 23:44:05 is there an argument I can pass to init to not try to boot into the rootfs? I want to make sure telnet is still working Feb 10 23:45:43 @rockybulwinkle, `mka hybris-recovery` … and then flash hybris-recovery.img to boot partition Feb 10 23:47:54 I got another error pastebin.com/vSfmZGfG :/ … I already added CONFIG_PFK=y Feb 10 23:48:33 ok, telnet still works Feb 10 23:48:36 I'll be looking for solutions Feb 10 23:50:21 I forgot to mention, the error is very similar to the others Feb 10 23:53:34 @Pablo Alejandro, Maybe try to add `#include ` at the top in `fs/ext4/crypto_key.c` Feb 10 23:54:29 @Ellenjott, GitHub can only find the definition of that struct in that file (linux/pfk.h) ... Feb 10 23:54:49 @UniversalSuperBox, do you know how to utilize that? Feb 10 23:54:57 Nope Feb 10 23:55:01 haha ok Feb 11 00:01:38 Didn't work, Im triyng changing CONFIG_EXT4_FS_ENCRYPTION from y to n Feb 11 00:01:54 What does "echo "continue" >/init-ctl/stdin" actually do? Who is listening for that? Feb 11 00:02:12 I'm trying to understand the boot process in halium. Feb 11 00:02:22 It writes 'continue' to init's stdin Feb 11 00:03:01 systemd is our init at this point? Feb 11 00:03:04 No Feb 11 00:03:11 Busybox is still getting the system set up Feb 11 00:03:17 ok Feb 11 00:03:48 @Pablo Alejandro, Apparently is working Feb 11 00:04:57 it looks like the init is just a shell script at this point Feb 11 00:05:53 It's not even a real init, it's started before systemd and just mounts the rootfs so systemd / upstart can be started from there Feb 11 00:06:36 I'm familiar with that. At work I did the same thing bringing up debian. u-boot->busy box /init script mounting filesystems -> switchroot and exec debian's init Feb 11 00:11:18 @rockybulwinkle, May I ask where you work? Just curious Feb 11 00:13:13 NHanced Semiconductor Feb 11 00:19:21 you? Feb 11 00:19:24 @rockybulwinkle, Interesting. But you can tell your SysAdmins they shall update the SSL certificate :D Feb 11 00:19:34 Yeah. Um. yeah. Feb 11 00:19:34 @rockybulwinkle, I'm still a student Feb 11 00:19:36 ok, I have a new error pastebin.com/3n2NA8ML … At least all related to EXT4 encryption is gone Feb 11 00:24:19 @Pablo Alejandro, The missing function is defined here: … https://github.com/LineageOS/android_kernel_leeco_msm8996/blob/352390421ab207b194810f5c7450e377af478836/drivers/video/backlight/lm3697_bl.c#L159 Feb 11 00:24:53 Ok, thanks :D Feb 11 00:27:27 @Pablo Alejandro, But I don't understand why it's not finding it :/ Feb 11 00:27:50 I'll manually check the file Feb 11 00:29:04 Are there any more logs output by the init system I can look at to figure out more problem? I can see from the kernel log it's starting to set up the ethernet gadget, so I would think if it's doing that it's probably already mounted a filesystem? Feb 11 00:31:05 @rockybulwinkle, Sry, no idea Feb 11 00:31:11 I'm poking around in rootfs. faillog is 23.5k but all zeros... Feb 11 00:32:20 same with lastlog.... Feb 11 00:32:37 is the bootstrap from debootstrap (or multistrap or whichever you're using) or from boot? Feb 11 00:33:31 yeah probably looking at it, lol Feb 11 00:42:25 Well, I'm searching the error in internet and it seems is a problem with functions in the file, or something like that, I'll try to fix it Feb 11 00:42:54 @Pablo Alejandro, Good luck Feb 11 00:43:02 I'll got to bed now Feb 11 00:43:12 good night Feb 11 00:43:24 Night :) Feb 11 00:43:38 @Ellenjott, Haha thanks and good night c: Feb 11 00:46:11 @pablo what do you get if you do "find . -name lm3697_bl.o" in your build directory? Feb 11 00:46:56 based on what I see in your build log, lm3697_bl.c wasn't built Feb 11 00:48:18 you should enable CONFIG_BACKLIGHT_LM3697 in your config Feb 11 00:48:23 is that set? Feb 11 00:52:27 Am I interpreting the kernel log correctly, is that init printing these messages? https://gist.github.com/rockybulwinkle/b30e5883dfa285e962ea7717ea0ee230 Feb 11 00:52:41 If so, I'm trying to find where those are printed, it's right before the kernel panics Feb 11 00:53:03 oh my god I finally got a root shell Feb 11 00:53:10 *applause* Feb 11 00:53:11 What a pain in the ass Feb 11 00:53:34 @rockybulwinkle, Oh, sorry, I didn't see your message, I'll see, thanks Feb 11 00:53:36 Hmm, those messages kinda look like they're coming from Android's init. Feb 11 00:53:54 Want the full log? Feb 11 00:54:05 Try `systemctl disable lxc@android.service` from a chroot into the rootfs then boot again Feb 11 00:54:18 ok Feb 11 00:54:30 Android does weird things sometimes. Feb 11 00:55:36 btw, what is system.img exactly? Is it the next part of the boot after the bit with telnet? Feb 11 00:55:50 system.img is the Android image, rootfs.img is the linux image Feb 11 00:56:01 well, GNU/Linux image Feb 11 00:56:13 so rootfs has systemd? Feb 11 00:56:18 Yep! Feb 11 00:56:26 Oh, you're not getting there are you Feb 11 00:56:34 and the busybox shell script exec/switchroots to it? Feb 11 00:56:38 Yep Feb 11 00:57:03 I'm not 100% sure if I'm getting there. Somebody is doing something that crashes the system. I am able to get to telnet, but if I try booting further it crashes Feb 11 00:57:35 Alright. `ls /rootmnt` if I remember correctly Feb 11 00:57:46 while in telnet? Feb 11 00:57:50 Yep Feb 11 00:58:29 @rockybulwinkle, CONFIG_BACKLIGHT_LM3697 is in m, I change it to y and when I try to build it changes again to m, I'm editing KERNEL_OBJ/.config file, in lineage_zl1_deconf is already set to y Feb 11 00:58:44 It may not support being built in to the kernel Feb 11 00:58:48 @UniversalSuperBox, Congratulations :) Feb 11 00:59:03 was it set to 'm' before? Feb 11 00:59:14 yes, it was Feb 11 00:59:40 @UniversalSuperBox I don't have a /rootmnt. I'm in hybris-recovery Feb 11 00:59:58 Erm, can you show me /? Feb 11 01:00:07 I don't remember the file layout exactly Feb 11 01:00:28 there's something mounted on target/ Feb 11 01:00:50 Ah, right. There. Feb 11 01:00:55 What's in there? Feb 11 01:01:01 https://gist.github.com/rockybulwinkle/be690c9e09889beae053ed347b3cbd89 Feb 11 01:01:13 looks like a rootfs to me. Feb 11 01:01:41 systemd is in there. Must be rootfs.img Feb 11 01:01:44 Where was your full dmesg log again? Feb 11 01:01:50 When it crashes Feb 11 01:03:00 https://gist.github.com/rockybulwinkle/8a77c298ee920fa7e92af21159ea81f5 Feb 11 01:03:21 sorry about the corruption, I don't seem to be able to reboot straight into twrp without the device stopping dram refresh Feb 11 01:04:46 huh, it doesn't show on gist... Feb 11 01:05:21 Attempted to kill init... so switch_root seems to be failing. Is the file `/data/systemd_stdouterr` around? Feb 11 01:05:44 in which filesystem? oh ok Feb 11 01:06:08 it is not Feb 11 01:06:18 Systemd never starts, then Feb 11 01:06:37 Aha, why is /target read-only Feb 11 01:06:45 Can you remount it read-write? Feb 11 01:06:49 sure Feb 11 01:08:00 mount -o remount,rw /dev/block/loop0 /target Feb 11 01:08:08 that's the right command, right? Feb 11 01:08:12 Just `mount -o remount,rw /target` Feb 11 01:08:21 I tried that too, it's still ro Feb 11 01:08:23 This is a weird situation. I am allowed a maximum of one ssh session before the system locks up Feb 11 01:08:33 Ah, the loop device is read-only then Feb 11 01:08:34 is pts mounted? Feb 11 01:08:41 Yep Feb 11 01:08:56 I guess you probably wouldn't be able to get a single ssh session if it weren't.... Feb 11 01:09:10 I can use that ssh session to my heart's content, but some programs will fail to run and then the system is dead Feb 11 01:09:27 which programs fail? Feb 11 01:10:17 Hmm, really nothing I can test with... the container isn't running. Feb 11 01:10:29 Let's see if it'll let me in again Feb 11 01:11:06 Indeed it will not Feb 11 01:12:09 The session stops either after `sshd` sends the 'authentication success' packet, or after that whenever I try to start a session Feb 11 01:12:17 https://gist.github.com/rockybulwinkle/50d5795ec40928707f85f6720fa21bf6 Feb 11 01:12:29 I can't seem to mount it rw Feb 11 01:12:44 `dmesg`? Feb 11 01:12:47 Strangely enough, the ssh error is usually `broken pipe` rather than `timeout` Feb 11 01:13:55 the dmesg is flooded with IP messages Feb 11 01:14:28 There should be a complaint about the mount somewhere Feb 11 01:14:48 ~ # dmesg | grep -i mount … <6>[01-01 00:00:00.000] [0][0: swapper/0]Mount-cache hash table entries: 8192 (order: 4, 65536 bytes) … <6>[01-01 00:00:00.000] [0][0: swapper/0]Mountpoint-cache hash table entries: 8192 (order: 4, 65536 bytes) Feb 11 01:15:17 What the heck? So, the simple act of mounting the Android system.img and the things it contains means that I can't ssh in. Feb 11 01:15:36 wat. Feb 11 01:15:59 can you set up a telnet server to do a post mortem on ssh? Feb 11 01:16:01 It's like there's a limited number of something, and each program uses some something Feb 11 01:16:08 ulimit -a? Feb 11 01:16:25 Well this is neat... what am I looking for? Feb 11 01:16:36 in ulimit? Feb 11 01:16:48 https://paste.ubuntu.com/=x9HH8TZMtn/ Feb 11 01:17:49 @rockybulwinkle, After ssh stops working, the bash session for the user won't start through telnet either Feb 11 01:17:56 grr Feb 11 01:18:37 reverse shell? Feb 11 01:18:58 maybe set up a netcat of dmesg? Feb 11 01:19:05 I'm spitballing Feb 11 01:19:09 Nothing interesting in dmesg Feb 11 01:19:42 I don't see anything weird in your ulimit Feb 11 01:20:03 actually wait Feb 11 01:20:07 nvm Feb 11 01:20:30 I thought pending signals maybe shouldn't be the same size as max user processes, but that's how it is on my laptop too Feb 11 01:20:44 https://paste.ubuntu.com/=z7r4VmwnNY/ is dmesg. Feb 11 01:20:50 It's a riveting log, lemme tell you Feb 11 01:21:24 keygen terminated? Feb 11 01:21:27 The only fun is all of the `start on startup` services exiting with status 1. Which is actually not much fun at all Feb 11 01:21:34 @rockybulwinkle, It does, but the keys are generated Feb 11 01:21:55 yup. That is the most exciting log ever! Feb 11 01:22:30 jeeze where was I... right, rootfs mounted read-only Feb 11 01:22:45 so systemd may have started, but couldn't write it's errors because ro Feb 11 01:22:48 Sounds like the filesystem on it is unhappy Feb 11 01:23:00 Let's see, was that thing merged in... Feb 11 01:23:16 It was not. So you don't have e2fsck Feb 11 01:23:23 Reboot to recovery and run e2fsck on the image maybe Feb 11 01:23:31 I'm in recovery Feb 11 01:23:37 I have been this whole time Feb 11 01:23:47 No, like TWRP recovery Feb 11 01:23:59 oh ok Feb 11 01:24:18 Is there some secret kernel memory that I'm running out of? Feb 11 01:24:31 That doesn't even make sense, but neither does this error Feb 11 01:24:40 what's your device anyway? Feb 11 01:24:45 Nexus 5X, bullhead Feb 11 01:25:31 https://gist.github.com/rockybulwinkle/ffa51032d7354e32151ae36dfa557277 Feb 11 01:25:37 looks like it did *something* Feb 11 01:25:46 Great! Now can you mount it read-write again? Feb 11 01:26:10 huzzah! and there was much rejoicing! Feb 11 01:26:59 *forgot to unmount before rebooting* Feb 11 01:27:21 I wonder if `loglevel=7` will help Feb 11 01:28:04 On my end or yours? Feb 11 01:28:11 Mine Feb 11 01:28:18 I think out loud a lot Feb 11 01:28:27 me too, don't worry about it Feb 11 01:28:39 yay, the log is the exact same. Feb 11 01:28:45 I *try* not to get too spammy about it Feb 11 01:29:23 Alas, I still can't mount it rw within hybris-recovery. It does that perfectly fine in twrp though! Feb 11 01:29:43 What kernel version is it again? Feb 11 01:30:00 I havn't found a solution yet for my problem, I don't know why it can't find lm3697_bl_set :( Feb 11 01:30:01 3.18.31 Feb 11 01:30:26 Do you find the .o file for it in your build dir? Feb 11 01:31:06 and yes, ext2,3,4 support is built in to my kernel from /proc/filesystems. Feb 11 01:31:19 Oh, sorry, I'll search it Feb 11 01:31:46 `dmesg |grep EXT4` Feb 11 01:32:02 And maybe `dmesg |grep loop0` for good measure Feb 11 01:32:44 that doesn't print anything. Tried case insensitive too. … as for searching for loop0: … <6>[01-01 00:00:00.893] [3][1: swapper/0]loop: module loaded Feb 11 01:33:01 That's a fuzzy search Feb 11 01:33:28 Anyway, you had all the options from the mer kernel check enabled? Feb 11 01:34:30 @rockybulwinkle, I found it Feb 11 01:34:45 Not... quite.... A handful of options weren't available involving cgroups stuff. I didn't enable the rtc because the kernel doesn't support it for some reason on aarch64, I doubt that's the problem right now though. Feb 11 01:35:04 I didn't see anything thought would readily explain not being able to mount this read-write Feb 11 01:35:20 Unless it's just busybox being busybox Feb 11 01:35:24 is `mount -v` a thing? Feb 11 01:36:25 It didn't seem to do anything Feb 11 01:36:36 It is in the --help though Feb 11 01:36:49 OH Feb 11 01:36:56 busybox handles mount options differently Feb 11 01:37:19 oh nvm it also supports the rw option Feb 11 01:39:30 it just doesn't like that file. I created a 50MB file and made an ext2 partition on it, mounted, and it's rw. Feb 11 01:39:45 Weird. Feb 11 01:39:46 Guess I'll transfer the contents over then? Feb 11 01:39:50 Can you use halium-install instead? Feb 11 01:40:07 https://github.com/JBBgameich/halium-install Feb 11 01:40:13 Takes a little longer, but it's less dumb than Android busybox ports Feb 11 01:40:36 that reminds me... I got my make_ext4 from some random source on the internet. I'm used to just using dd to make my files and mkfs and loopback to create my filesystem images. Feb 11 01:41:00 Oh. Very possible that it's got an option in it that isn't supported by your kernel Feb 11 01:41:19 both mdss_dsi_panel and lm3697_bl has their .o files Feb 11 01:41:42 so they did build, right? Feb 11 01:41:49 yeah, that means they built Feb 11 01:41:58 I'm perplexed at your error now.... Feb 11 01:42:21 Haha, sh*t ;-; Feb 11 01:42:22 So wait... rsyslogd is running but not writing Feb 11 01:42:42 Initscripts are exiting with `1` even though they work correctly Feb 11 01:42:51 What is this upside-down world!?!? Feb 11 01:42:57 I think is enough for today, I'm tired Feb 11 01:43:34 Ok. The next thing I'd try Pablo Alejandro is a clean build... Feb 11 01:44:06 Do you mean start again? Feb 11 01:44:18 `make clean` then build again Feb 11 01:44:21 ^^^ Feb 11 01:44:30 ah, ok ok, I'll try Feb 11 01:45:11 Sorry, I'm not used to use make Feb 11 01:45:55 ugh it's paining me letting my OLED sit there at the bootloader Feb 11 01:46:17 it didn't clear the screen when it booted into hybris-recovery Feb 11 01:46:25 Definitely give `halium-install` a try Feb 11 01:46:31 your version? Feb 11 01:46:39 Nah, that's jbb's stuff Feb 11 01:46:40 I'll try it in a moment. Feb 11 01:46:49 s/your/jbb's/ Feb 11 01:47:40 And now rsyslogd won't stop! Feb 11 01:47:46 What is this!? Feb 11 01:48:03 https://www.youtube.com/watch?v=HgzGwKwLmgM Feb 11 01:48:22 And it still didn't write a syslog! Feb 11 01:51:20 So, you meant jbb's halium-install right? Feb 11 01:51:29 Yerp Feb 11 01:52:00 I did manage to continue the boot with that filesystem rw but don't see a systemd file in root Feb 11 01:52:34 that is, by creating a new partition manually and transfering the contents Feb 11 01:53:17 Should be at `/sbin/init` Feb 11 01:53:31 hm? Feb 11 01:53:51 The system is starting? Feb 11 01:53:58 no Feb 11 01:54:09 I was looking for that systemd error file you mentioned a while back Feb 11 01:54:14 Ah Feb 11 01:55:22 where can I get update-binfmts? Feb 11 01:57:12 uh Feb 11 01:57:29 not in my package manager or AUR as far as I can tell Feb 11 01:58:11 qemu-user-static Feb 11 02:00:08 that doesn't have it.... Feb 11 02:00:43 It depends on binfmt-support, no? Feb 11 02:01:57 optional dependency :) Feb 11 02:04:22 I found the source, I'll just build it Feb 11 02:04:52 I did make clean, I get exactly the same error Feb 11 02:06:28 haha, enough for me for today Feb 11 02:06:46 yeah I'm just about at that point too Feb 11 02:07:51 I refuse to be beaten by the LGE Nexus 5X Feb 11 02:07:54 well, good night to everyone :) Feb 11 02:08:18 I'm running jbb's install script Feb 11 02:08:56 'night Feb 11 02:12:36 https://gist.github.com/rockybulwinkle/c0a841acd23312ea67de23dbc4212628 Feb 11 02:12:47 Should I be concerned about those "short read" messages? Feb 11 02:13:07 I have no idea Feb 11 02:13:30 the passwd error is also worrying. What architecture was jbb working on? Feb 11 02:14:04 I think it should do any, but he's a Debian guy Feb 11 02:14:25 But, y'know, all the build tools expect you're using Ubuntu. Feb 11 02:15:10 that should only be a problem if they are including precompiled dynamically linked programs I would think Feb 11 02:15:31 the passwd error is something running on the phone probably? Did he prebuild a passwd executable? Feb 11 02:15:56 Nope, it's running on your computer Feb 11 02:16:07 Through `qemu-user-static`, which is syscall emulation in userspace Feb 11 02:16:15 ah ok Feb 11 02:16:32 so perhaps he's expecting a different architecture... Feb 11 02:16:57 Than armhf? Feb 11 02:17:03 idk! Feb 11 02:17:14 Nah, that script works on Ubuntu and Debian Feb 11 02:17:33 I do have a debian computer on hand, perhaps I should try that... Feb 11 02:18:02 I'm a little iffy though because it's running my nextcloud, I'd rather leave it dedicated to that Feb 11 02:18:20 It mocks me. Feb 11 02:18:50 :/ Feb 11 02:19:32 wow, on TWRP: … /data # ls --help … Segmentation fault Feb 11 02:20:53 Oh, yeah Feb 11 02:20:56 Everything segfaults Feb 11 02:21:26 well anyway, I mounted the fs and changed the password via TWRP Feb 11 02:24:07 ugh. Feb 11 02:25:18 it is still mounted ro Feb 11 02:25:24 waaat Feb 11 02:25:29 why Feb 11 02:25:53 `dmesg |grep ext`? Feb 11 02:26:06 case insensitive Feb 11 02:26:34 halium-install doesn't touch the boot partition does it? Feb 11 02:26:39 Nope Feb 11 02:27:40 <6>[05-23 05:47:18.913] [3][550: mount]EXT4-fs (loop0): mounted filesystem with ordered data mode. Opts: (null) Feb 11 02:27:56 Not even a complaint Feb 11 02:28:06 it's type ext4, it used to be ext2 I think? Feb 11 02:28:38 ~ # mount -o remount,rw /target Feb 11 02:28:45 and dmesg: <6>[05-23 05:48:24.943] [3][678: mount]EXT4-fs (loop0): re-mounted. Opts: data=ordered Feb 11 02:30:26 what. Feb 11 02:30:40 Welcome to porting Feb 11 02:30:58 Oh, I'm no stranger. Feb 11 02:31:22 We had to port u-boot and the kernel for the chip at work too. Feb 11 02:32:10 where's the busybox config for hybris-boot? Feb 11 02:32:33 That is a great question. Horribly complex to find Feb 11 02:32:37 Hey do you want to test a new thing Feb 11 02:32:43 maybe? Feb 11 02:32:47 depends. Feb 11 02:32:51 I never figured out which config is ours though Feb 11 02:33:10 https://github.com/Halium/projectmanagement/issues/40 Feb 11 02:33:23 It's an initramfs based on Debian Feb 11 02:34:27 like, you pulled in the programs from multstrap/debootstrap and wrote your own init script? Feb 11 02:34:46 https://github.com/Halium/initramfs-tools-halium Feb 11 02:34:54 Nah, it's Debian's initrd with a couple extra scripts and hooks Feb 11 02:35:02 Full glibc busybox Feb 11 02:35:03 ok Feb 11 02:37:14 OH I see, you're trying to replace the initrd? Feb 11 02:37:19 Yep Feb 11 02:37:44 how much stuff from debian's initrd is actually being used? Feb 11 02:38:11 It's a Debian initrd with two extra scripts... so... Feb 11 02:38:20 lol Feb 11 02:38:34 I may take a look at it tomorrow Feb 11 02:38:41 The real fun happens in mount_root() here: https://github.com/Halium/initramfs-tools-halium/blob/halium/scripts/halium Feb 11 02:39:07 y u no run software, 5X!? Feb 11 02:39:31 "bootmode" does this thing support dual booting? Feb 11 02:40:02 Maybe it could hack it, but I wouldn't recommend it Feb 11 02:40:10 That's actually for booting into Charger or Factory mode Feb 11 02:40:24 So you get the pretty charging animations Feb 11 02:40:40 oh. Feb 11 02:42:26 @UniversalSuperBox, Time to get serial working Feb 11 02:42:27 that's... a tad bit more complicated than what boots my board at work Feb 11 02:42:33 :) Feb 11 02:42:42 hahaha Feb 11 02:42:43 I have a shell! Feb 11 02:42:49 I am root Feb 11 02:42:51 Hey, open it up, I'm sure you'll find a serial port Feb 11 02:42:59 There's one off the headphone jack Feb 11 02:43:03 oh ok Feb 11 02:43:04 @rockybulwinkle, No it's more easier Feb 11 02:43:13 I don't have a FDDI though Feb 11 02:44:01 I'm done for the night methinks Feb 11 02:44:14 oke Feb 11 02:44:32 I'll still be lurking though Feb 11 02:45:44 g'night! Feb 11 02:45:48 I saw something about an android container at some point. What's that used for? Feb 11 02:45:58 Android drivers Feb 11 02:46:03 We didn't have you build Android for nothing. :P Feb 11 02:46:14 :( Feb 11 02:46:49 They aren't built into the kernel, then? Feb 11 02:46:59 No, that'd be way too easy Feb 11 02:47:18 yeah. Some of the drivers are though. Like freedreno. Feb 11 02:47:29 Freedreno is still... early. Feb 11 02:47:36 oh. Feb 11 02:47:46 I'm looking forward to the librem 5 Feb 11 02:47:58 And that's ignoring Bluetooth, WLAN, WWAN, LEDs, vibrator, sensors... Feb 11 02:48:12 Radios? Who needs those on a phone! Feb 11 02:48:43 Here radio means audio video calls Feb 11 02:48:58 Err *audio calls Feb 11 02:49:14 ... I was being sarcastic Feb 11 02:59:56 lxc-start: cgroups/cgfsng.c: recursive_destroy: 1274 Error destroying /sys/fs/cgroup/debug//lxc/android **** ENDING LOGGING AT Sun Feb 11 03:00:00 2018