| genericv3 | there were more libs missing, i only linked libandroid.icu for now and that strange library error came for it too. | 05:18 |
|---|---|---|
| genericv3 | gone for 10 hours* | 05:19 |
| genericv2 | helo | 14:05 |
| genericv2 | i'll also link these now: libcutils.so libz.so libc++.so libhardware.so | 14:08 |
| genericv2 | But i am worried about that not accessible error | 14:08 |
| mal | those should not be needed | 14:13 |
| mal | it's normal to have some messages which appear as errors but really are not | 14:14 |
| mal | about libraries | 14:14 |
| genericv2wonetwork | Okay then, i'll generate new logs now | 14:17 |
| mal | just in case remove those new symlinks you added and keep just the libandroidicu.so | 14:17 |
| mal | so we don't have any unexpected behavior | 14:17 |
| genericv2wonetwork | removed other libraries then | 14:20 |
| genericv2wonetwork | https://paste.opensuse.org/8a68fbc4391b journal | 14:23 |
| genericv2wonetwork | https://paste.opensuse.org/84ccfea5d1e0 logcat | 14:23 |
| mal | is this new build or same as before? | 14:29 |
| genericv2wonetwork | it is | 14:29 |
| mal | just making sure all of the fixes we have done before are there | 14:29 |
| genericv2wonetwork | i'll check sepolicy again | 14:29 |
| genericv2wonetwork | also systemctl output, may be interesting: https://paste.opensuse.org/f41b44f842e7 | 14:29 |
| genericv2wonetwork | looks okay, i made no modifications except that odm library thing, but i reverted that | 14:30 |
| mal | check output of binder-list, "binder-list -d /dev/vndbinder" and "binder-list -d /dev/binder" | 14:31 |
| genericv2wonetwork | two command, two empty output | 14:32 |
| mal | those are 3 commands | 14:33 |
| genericv2wonetwork | i am blind + empty output again | 14:33 |
| mal | all give empty output? | 14:34 |
| genericv2wonetwork | yeah :( | 14:34 |
| mal | is droid-hal even running then? | 14:35 |
| genericv2wonetwork | it is running | 14:35 |
| genericv2wonetwork | Loaded: loaded (/usr/lib/systemd/system/droid-hal-init.service; disabled; vendor preset: enabled) | 14:35 |
| genericv2wonetwork | Active: active (running) | 14:35 |
| genericv2wonetwork | ah and | 14:35 |
| genericv2wonetwork | forget to mention | 14:35 |
| genericv2wonetwork | '/usr/bin/droid/droid-hal-early-init.sh does not exist | 14:36 |
| genericv2wonetwork | i created it as blank to make droid-hal-init work | 14:36 |
| genericv2wonetwork | poweroff\ | 14:41 |
| genericv2wonetwork | oops | 14:41 |
| T42_ | <elros34> droid-hal-early-init.sh is optional, remove what you created | 14:51 |
| genericv2 | okay np then. it's just blank as i said so no need to remove | 14:51 |
| genericv2 | just wondered if it does something for binder etc. | 14:52 |
| T42_ | <elros34> what is output of ls -al /dev/binder* | 14:53 |
| genericv2 | crw-rw-rw- 1 root root 10, 55 Mar 16 12:26 /dev/binder | 14:55 |
| genericv2 | /dev/binderfs: | 14:55 |
| genericv2 | total 0 | 14:55 |
| genericv2 | just /dev/binder and binderfs directory exist | 14:56 |
| T42_ | <elros34> that looks incorrect | 14:57 |
| genericv2 | yeah i think so | 14:57 |
| genericv2 | dev-binderfs.mount this dude is failing to mount tho | 14:57 |
| T42_ | <elros34> does adding selinux stubs helps? | 14:58 |
| genericv2 | It always failed | 14:59 |
| genericv2 | With selinux stubs too | 14:59 |
| mal | wondering why binderfs mount fails | 15:00 |
| genericv2 | [Mount] | 15:00 |
| genericv2 | What=binder | 15:00 |
| genericv2 | HM | 15:00 |
| mal | "unknown filesystem type 'binder'" | 15:01 |
| genericv2 | kernel config moment then | 15:01 |
| genericv2 | my mistake again | 15:01 |
| genericv2 | what defconfig flag to add? | 15:01 |
| mal | why would that be missing from defconfig, unless you for some reason removed it | 15:04 |
| genericv2 | i didin't probably. Stock defconfig only has CONFIG_ANDROID_BINDER_IPC. because of minimized defconfig maybe | 15:06 |
| mal | wondering what android device repos and kernel you are really using | 15:08 |
| genericv2 | All stock lineage | 15:10 |
| genericv2 | https://github.com/LineageOS/android_kernel_xiaomi_msm8998 | 15:10 |
| genericv2 | https://github.com/LineageOS/android_device_xiaomi_msm8998-common | 15:10 |
| genericv2 | https://github.com/LineageOS/android_device_xiaomi_sagit | 15:10 |
| genericv2 | and lineage-20 branches | 15:10 |
| genericv2 | i'll upload my kernel changes | 15:12 |
| genericv2 | mal: https://github.com/LineageOS/android_kernel_xiaomi_msm8998 | 15:40 |
| genericv2 | how can i forget to copy new one | 15:40 |
| genericv2 | https://github.com/ttwoxa/android_kernel_xiaomi_msm8998 | 15:40 |
| mal | but if android didn't use binderfs then you shouldn't need it either | 15:42 |
| genericv2 | then what we need to do for binder | 15:42 |
| genericv2 | does it use another fs etc? | 15:42 |
| mal | still not sure why the other binder are missing | 15:43 |
| mal | @elros34 any ideas why other binder devices could be missing? when bindefs was not enabled in original defconfig | 15:44 |
| genericv2 | away for a few minutes | 15:47 |
| genericv2 | any news? | 15:58 |
| T42_ | <elros34> mal I dont have any experience in that area | 16:03 |
| T42_ | <elros34> genericv2 maybe you could flash lineage kernel, boot to lineage and compare logs/dev/binder*, I guess /proc/config.gz will be missing | 16:05 |
| T42_ | <elros34> btw can you show 'zcat /proc/config.gz' from device? | 16:06 |
| T42_ | <elros34> mal remember this is old device released with much older android version so maybe something is not done as usual | 16:07 |
| genericv2 | i'll boot it up now | 16:07 |
| genericv2 | it's released with android 7 😄 | 16:07 |
| T42_ | <elros34> be aware booting to los might encrypt your /data | 16:08 |
| genericv2 | Also, oneplus 5 uses the same chipset with this. If it helps with anything | 16:08 |
| genericv2 | i'll backup .stowaways, no worries | 16:08 |
| genericv2 | https://paste.opensuse.org/03316f6b2ecb | 16:14 |
| mal | CONFIG_ANDROID_BINDER_DEVICES="binder,hwbinder,vndbinder" is there | 16:15 |
| mal | afaik that is the default one | 16:15 |
| mal | so it's possible the binderfs issue is not an issue, last time you checked ls -al /dev/binder* which of course wouldn't show vndbinder and hwbinder | 16:17 |
| mal | also binderfs is only availble after certain kernel versions | 16:17 |
| genericv2 | so i can disable that service | 16:17 |
| mal | correct way is to ignore it in droid-hal spec | 16:18 |
| mal | there it is possible add line like "%define makefstab_skip_entries /some_path /dev/some_other_path" where paths are the mount points of the mount services | 16:19 |
| genericv2 | hmm, i'll check it out on next package | 16:20 |
| genericv2 | If binder is not a problem, any idea on what is current problem | 16:22 |
| T42_ | <elros34> right, I get you wrong command: ls -alR /dev/*binder*. Adding selinux stubs helps with hwcomposer error? | 16:23 |
| genericv2 | /dev/binder /dev/hwbinder /dev/vndbinder | 16:25 |
| T42_ | <elros34> is you latest paste from sailfish or from los? | 16:25 |
| genericv2 | this is from sfos | 16:25 |
| genericv2 | i remembler like there is no problem with hwservicemanager right now | 16:26 |
| mal | still odd how binder-list would return nothing if droid-hal is running | 16:26 |
| T42_ | <elros34> what about hwcomposer (no hwservicemanager) | 16:27 |
| genericv2 | its looping | 16:28 |
| genericv2 | Service 'vendor.hwcomposer-2-1' (pid 6703) exited with status 1 | 16:28 |
| genericv2 | Sending signal 9 to service 'vendor.hwcomposer-2-1' (pid 6703) process group... HYBRIS: killing PID instead of process group. | 16:28 |
| genericv2 | starting service 'vendor.hwcomposer-2-1'... | 16:28 |
| genericv2 | lipstick[2588]: failed to get hwcomposer service | 16:28 |
| genericv2 | https://paste.opensuse.org/pastes/bf19b4c6d0bc | 16:29 |
| T42_ | <elros34> why did you enable CONFIG_SECURITY_SELINUX_DEVELOP? | 16:30 |
| genericv2 | i don't have a idea tbh. I added it while trying to fix se | 16:30 |
| T42_ | <elros34> this makes you boot to permissive mode by default | 16:31 |
| genericv2 | CONFIG_SECURITY_SELINUX=y | 16:31 |
| genericv2 | CONFIG_DEFAULT_SECURITY_SELINUX=y | 16:31 |
| genericv2 | only leaving these them | 16:31 |
| genericv2 | then | 16:31 |
| genericv2 | do i need to remove DEFAULT thingy? | 16:32 |
| T42_ | <elros34> I told you this already more than once, you should not change anything selinux related. btw you dont start test_hwcomposer like that, you must use EGL_PLATFORM=hwcomposer test_hwcomposer | 16:33 |
| T42_ | <elros34> you can start test_sensors without any additional commands | 16:34 |
| genericv2 | Yeah sorry, i made all se things to their defaults now | 16:34 |
| genericv2 | okayt rebuilt kernel and flashed | 16:38 |
| genericv2 | test_hwcomposer it gives the same output like this too | 16:38 |
| genericv2 | every test binary gives "Aborted", really good build | 16:39 |
| T42_ | <elros34> usually worth to try are only test_hwcomposer/egl_configs/sensors | 16:42 |
| T42_ | <elros34> so anything changed in logs now? binder-list? | 16:47 |
| genericv2 | Nope, still same :( | 16:49 |
| genericv2 | test_hwcomposer: test_common.cpp:373: HWComposer* create_hwcomposer_window(): Assertion `err == 0' failed. | 17:23 |
| genericv2 | OHH | 17:27 |
| genericv2 | ServiceManager: Waiting for service 'display.qservice' on '/dev/vndbinder'.. | 17:27 |
| genericv2 | mal: any idea about this? Binder issue right | 17:28 |
| T42_ | <elros34> did you get rid of "could not self register *servicemanager"? | 17:30 |
| genericv2 | yes ig | 17:34 |
| genericv2 | i can't see anything like that | 17:34 |
| genericv2 | i realized simply nothing works | 17:36 |
| genericv2 | except usb networking :fire: | 17:36 |
| T42_ | <elros34> could you show latest logs + binder-list -d /dev/hwbinder, /system/bin/lshal | 17:39 |
| genericv2 | https://paste.opensuse.org/pastes/451ed5fd6037 | 17:41 |
| genericv2 | all binder list thingies are still empty | 17:42 |
| genericv2 | nevermind i don't need gui, i'll take my laptop to everywhere and interact with the phone via telnet | 17:42 |
| T42_ | <elros34> mind its /dev/hwbinder not /dev/binder | 17:46 |
| genericv2 | / # binder-list -d /dev/hwbinder | 17:47 |
| genericv2 | / # binder-list -d /dev/binder | 17:47 |
| genericv2 | / # binder-list -d /dev/vndbinder | 17:47 |
| genericv2 | i am trying all of them | 17:47 |
| T42_ | <elros34> what about latest logs? | 17:52 |
| genericv2 | i'll take them now | 17:52 |
| genericv2 | things are waiting for binder but binder is dead | 17:59 |
| genericv2 | i'll kill binder | 17:59 |
| genericv2 | https://paste.opensuse.org/898da94febc5 | 18:07 |
| genericv2 | and https://paste.opensuse.org/ea6b0c6cea8a | 18:07 |
| T42_ | <elros34> first error: servicemanager: Could not self register servicemanager | 18:10 |
| genericv2 | which is great, i think like i fixed it for ages | 18:11 |
| T42_ | <elros34> this one might be also important "Missing permissions to add IBase" | 18:11 |
| T42_ | <elros34> something at least try to start now: Name: QTI Hardware Composer Module | 18:14 |
| genericv2 | there is no service like that i think | 18:17 |
| T42_ | <elros34> do you have droid-config/device uploaded somewhere? | 18:22 |
| T42_ | <elros34> /linkerconfig/ld.config.txt is missing right? | 18:23 |
| genericv2 | not yet, i can upload everything | 18:23 |
| genericv2 | no it's not, it exist | 18:24 |
| genericv2 | / # ls /linkerconfig/ | 18:25 |
| genericv2 | ld.config.txt | 18:25 |
| genericv2 | no luck ig | 18:56 |
| genericv2 | bye again, i am not sure if i'll continue or not tomorrow. If you have a suggestion please mention my nick and write it, at least i'll check irc logs. | 19:11 |
| genericv2 | thanks again for helping me | 19:11 |
| T42_ | <elros34> suggestions are still same: fix "servicemanager: Could not self register servicemanager", and maybe "Missing permissions to add IBase", at least find where its defined in source code | 19:12 |
| genericv2 | okay, booting the thing for the last time today. | 19:13 |
| genericv2 | and source code? | 19:13 |
| genericv2 | hwservicemanager: hwservicemanager is ready now. ah yeah its totally ready | 19:14 |
| T42_ | <elros34> What do you mean? Mentioned errors are gone for all *servicemanagers? | 19:16 |
| genericv2 | nope, it's just saying these after servicemanager errors | 19:16 |
| genericv2 | also what do you think about keymaster? ()wait_for_keymaster[1447]: Waiting for keymaster) | 19:16 |
| T42_ | <elros34> I am not 100% sure but wouldn't worry about it | 19:17 |
| genericv2 | any idea about permission thingy | 19:24 |
| genericv2 | because i dont have a idea about both | 19:24 |
| genericv2 | i thinked about maybe there is a conflict with vendor etc. but no | 19:24 |
| T42_ | <elros34> mistake in /etc/selinux? Anway I would focus on "Could not self register " first, because IBase error might fix itself or be not important | 19:27 |
| genericv3 | Probaby. I am not sure about anything i made from now on, i'll check everything again and again | 19:32 |
| genericv3 | Also i'll add every known-working thingy to sparse and continue like that | 19:33 |
| genericv3 | I remember like i added task_profiles.json, i'll add seplocy etc. | 19:33 |
| genericv3 | And a little bit of hope appears again, gonna continue tomorrow | 19:35 |
| mal | one thing to check is that the those two selinux files are copies of the files and not symlinks | 19:50 |
| genericv3 | already copied them. And that fixed a lot as long as i remember | 19:50 |
| genericv3 | bye | 20:00 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!