| genericlinks | https://paste.opensuse.org/pastes/ff0d8d383cfb Lastest journalctl | 00:01 |
|---|---|---|
| genericv3 | Now i need to go, i'll try to continue tomorrow | 00:01 |
| genericv3 | Thanks again everyone | 00:01 |
| T42_ | <b100dian> > Cannot find '/odm/bin/hw/android.hardware.keymaster@3.0-service-qti_HYBRIS_DISABLED | 00:03 |
| T42_ | <b100dian> Interesting, why is that disabled | 00:03 |
| T42_ | <elros34> apexd failed but could be simple issue with missing kernel patch form hadk-hot | 00:04 |
| mal | looking at the logs copying apexd.rc might have been incorrectly done droid-hal-init: /usr/libexec/droid-hybris/system/etc/init/apexd.rc: 1: Invalid section keyword found | 00:06 |
| mal | which would result in the wrong apexd to be used which will fail | 00:06 |
| T42_ | <b100dian> is anyone working on fairphone 5 already? | 00:08 |
| T42_ | <b100dian> Am asking because I search for a new device: either one of https://www.gsmarena.com/compare.php3?idPhone1=11962&idPhone3=12013&idPhone2=12540 or the Xperia 5 IV that I have got one in the family. | 00:13 |
| mal | @b100dian I have an fp5, although at the moment it has sfos on top of mainline kernel, not the usual port | 00:18 |
| T42_ | <b100dian> mal: that is very good to know. mainline however doesn't bring the camera capabilities does it? | 00:19 |
| mal | well I did implement the camera subsystem support in mainline for it but no sensor drivers yet | 00:22 |
| mal | I need to find some android base to use for normal porting, unfortunately the sources from fairphone can't be built at the moment | 00:23 |
| T42_ | <b100dian> I read about that stuff too. Are you considering that there's some reverse engineering applicable to those (probably sony sensor) blobs that could help with the mainline? | 00:24 |
| T42_ | <b100dian> really lately I've poked at the postmarket wiki and cameras are off the table for anything 2018 and beyond | 00:26 |
| T42_ | <b100dian> But I dont' know if something like waydroid can help as a workaround | 00:27 |
| mal | I have some non-working sensor drivers for some of the cameras, not sure what the issue is yet with those since I made those myself | 00:28 |
| T42_ | <b100dian> Sounds like a full-time job:) | 00:29 |
| T42_ | <b100dian> Anyway, that answers my question, thank you. I will still have to figure out if 5 IV or a xiaomi for me and will trust you for FP5 ;) camera or no camera | 00:29 |
| mal | I found some unofficial lineage device repos for fp5, need to make test build of those | 00:31 |
| T42_ | <b100dian> Maybe there is something you can reply to this guy then? https://forum.sailfishos.org/t/what-would-be-required-for-a-port-to-fairphone-5/17870 | 00:32 |
| mal | yeah, I was planning on answering | 00:32 |
| T42_ | <b100dian> cool | 00:36 |
| T42_ | <b100dian> Thank you | 00:37 |
| genericv2 | Hello, i am back. I'll try to fix errors i found, anyone have a idea about last log i sent? | 10:44 |
| genericv2 | Fixed apexd.rc now, a laptop-moment typo | 10:48 |
| genericv2 | Apexd at least runs now | 10:53 |
| genericv2 | Mar 15 13:33:21 Mi6 droid-hal-init: mkdir("/dev/dm-user", 0755) failed File exists | 11:05 |
| genericv2 | Mar 15 13:33:21 Mi6 droid-hal-init: mknod("/dev/kmsg_debug", S_IFCHR | 0622, makedev(1, 11)) failed File exists | 11:05 |
| genericv2 | Mar 15 13:33:21 Mi6 droid-hal-init: Init encountered errors starting first stage, aborting | 11:05 |
| genericv2 | is this happening cuz droid-hal-init fails for another reason and restarts? | 11:05 |
| genericv2 | Yeah it's failing for another reason and tries to restart, then dies again because these files already exist | 11:14 |
| genericv2 | Mar 15 13:50:58 Mi6 droid-hal-init: Failed to umount /debug_ramdisk: No such file or directory | 11:22 |
| genericv2 | nope this is not we are looking for, probably | 11:26 |
| genericv2 | hwservicemanager is failing, this is why droid-hal-init is exiting | 11:27 |
| genericv2 | trying to find logs related to hwservicemanager, also now i can exetue binaries is /system/bin etc, | 11:29 |
| genericv2 | Hmm, maybe these? libc: Unable to set property "ro.logd.kernel" to "true": connection failed; errno=111 (Connection refused) | 11:39 |
| genericv2 | Same happens with hwservicemanager | 11:39 |
| genericv2 | Found nothing related to these in had-faq and hadk-hot | 11:40 |
| genericv2 | Checking irc logs, looks like it's a selinux error, i'll mess with kernel | 11:41 |
| T42_ | <elros34> did you check kernel patch related to loop and apexd errors I was talking about? | 11:52 |
| genericv2 | i was about to logout when you write that, i forget it. I'll check it now | 11:54 |
| genericv2 | Yeah THE SAME THING happens. Both apexd-bootstrap and fastboot. "Leads to apexd-bootstrap failure, ueventd start and reboot to fastboot mode. Apply this: https://github.com/mer-hybris/android_kernel_sony_msm/pull/95" | 11:55 |
| T42_ | <elros34> what about mal's advice from yesterday , did you check whether added rc files are correct? | 11:56 |
| T42_ | <elros34> this kernel patch changed anything? apexd-bootstrap still exit with 1 cod? | 11:57 |
| genericv2 | Yeah i checked them, 2 of them were blank, i fixed them | 11:57 |
| genericv2 | i'll apply it noiw, backed up current kernel | 11:57 |
| genericv2 | failed to apply :fire: | 11:58 |
| genericv2 | i'll apply it manually | 11:58 |
| genericv2 | These are different files, oof | 12:00 |
| T42_ | <elros34> ah and you are booting to act dead mode (charging). Next time reboot (command) and boot normally | 12:06 |
| genericv2 | I tried to apply the patch, there is some differences between these kernels but it must be okay i guess | 12:16 |
| genericv2 | okaay | 12:16 |
| genericv2 | lots of errors, great | 12:19 |
| genericv2 | fixed the compile errors | 12:24 |
| genericv2 | kernel booted fine | 12:27 |
| T42_ | <elros34> did you manage to boot to normal mode instead act dead? | 12:28 |
| genericv2 | rebooted now, it must boot to normal mode now | 12:29 |
| genericv2 | how can i check boot mode? | 12:31 |
| T42_ | <elros34> search "welcome" word, there should be boot mode somewhere around | 12:31 |
| genericv2 | Mar 15 14:56:03 Mi6 preinit: (18.21) Welcome to Sailfish OS 4.5.0.19 (Struven ketju) | 12:31 |
| genericv2 | BOOTSTATE = USER | 12:32 |
| T42_ | <elros34> boot target | 12:32 |
| genericv2 | Booting to default.target | 12:32 |
| T42_ | <elros34> thats good | 12:32 |
| genericv2 | Apexd is okay i think. "Wait for property 'apexd.status=activated' took 5810ms" | 12:33 |
| genericv2 | hwservicemanager is still crashing | 12:33 |
| T42_ | <elros34> is droid-hal-init up? try to get logcat too | 12:34 |
| genericv2 | it's dead, but i can execute logcat but it fails to start because logd is dead too | 12:34 |
| genericv2 | / # /system/bin/logd | 12:35 |
| genericv2 | libc: Unable to set property "ro.logd.kernel" to "true": connection failed; errno=111 (Connection refused) | 12:35 |
| genericv2 | libc: Fatal signal 6 (SIGABRT), code -1 (SI_QUEUE) in tid 2613 (logd), pid 2613 (logd) | 12:35 |
| genericv2 | Aborted (core dumped) | 12:35 |
| T42_ | <elros34> you need to be faster, try it few times right after booting | 12:35 |
| genericv2 | / # /system/bin/logcat | 12:38 |
| genericv2 | logcat: Failed to wait for logd.ready to become true. logd not running? | 12:38 |
| genericv2 | trying to start logd gives the same error | 12:38 |
| T42_ | <elros34> did you run this after droid-hal-init started or after it crashed? | 12:39 |
| T42_ | <elros34> and use one in /usr/libexec/droid-hybris/ | 12:40 |
| genericv2 | i ran it before it crashed | 12:43 |
| genericv2 | Active: active (running) | 12:43 |
| genericv2 | usr/libexec/droid-hybris/system/bin/logcat is same too | 12:43 |
| genericv2 | libc: Unable to set property "ro.logd.kernel" to "true": error code: 0xb | 12:49 |
| genericv2 | ERROR CODE CHANGED | 12:49 |
| genericv2 | nope its same | 12:50 |
| genericv2 | https://paste.opensuse.org/pastes/03a23e64cbd6 | 12:52 |
| T42_ | <b100dian> 11 resource temporarily not available. Curious that logcat cannot set that property, maybe you did a release build instead of debug of the android build? | 12:52 |
| genericv2 | Also hwservicemanager can't set it | 12:53 |
| genericv2 | i probably made a release build by accident | 12:53 |
| T42_ | <elros34> I am not sure you can do it by accident. How new journal looks like? | 13:01 |
| genericv3 | i am alive again, network issues | 13:02 |
| genericv3 | i built everything like instructions, rootfs has "release" in its filename | 13:03 |
| T42_ | <b100dian> No, I mean the android build, breakfast command or what you're using, what does it show you're building? Although to be fair, logcat should work in release builds too :hmm | 13:24 |
| genericv3 | Ah okay, it's userdebug build. | 13:30 |
| T42_ | <elros34> show latest log because for now we play guess game | 13:32 |
| T42_ | <elros34> did you change anything selinux related in kernel? | 13:33 |
| genericv3 | I'll be afk for a few minutes, i'll send it when i am back. i'll keep my irc client running so you can send messages. | 13:35 |
| genericv3 | i added a few flags for selinux when i applied that apexd related patch. | 13:35 |
| T42_ | <elros34> what flags? I am asking about defconfig changes | 13:37 |
| genericv3 | i am back | 14:10 |
| genericv3 | defconfig flags, wait i'll find them | 14:10 |
| genericv3 | CONFIG_SECURITY_SELINUX=y | 14:14 |
| genericv3 | CONFIG_SECURITY_SELINUX_BOOTPARAM=y | 14:14 |
| genericv3 | CONFIG_SECURITY_SELINUX_BOOTPARAM_VALUE=0 | 14:14 |
| genericv3 | CONFIG_SECURITY_SELINUX_DEVELOP=y | 14:14 |
| genericv3 | CONFIG_DEFAULT_SECURITY_SELINUX=y | 14:14 |
| T42_ | <elros34> you may try to remove at least SECURITY_SELINUX_BOOTPARAM*. I didn't see anything related in old logs but generally selinux shouldn't be disabled | 14:24 |
| T42_ | <elros34> basically you should have not change anything selinux related in defconfig | 14:25 |
| genericv3 | Normally i only had CONFIG_SECURITY_SELINUX=y | 14:27 |
| genericv3 | i added these as a fix attempt | 14:28 |
| genericv3 | for logcat | 14:28 |
| T42_ | <elros34> its illusion:) you have probably minimized defconfig for who knows what reasons. | 14:29 |
| T42_ | <elros34> so it's harder to figure out what changes you have. For your own curiosity you should check generated .config | 14:30 |
| genericv3 | hehe, probably. | 14:33 |
| genericv3 | i am at laptop (again), i'll check that generated .config file about in a hour | 14:34 |
| genericv3 | yeah i disabled selinux by a mistake | 14:37 |
| genericv3 | is permissive enough by the way? Or we need it to enforce | 14:39 |
| T42_ | <elros34> leave what was default. I think it logs there should be some signs of wrong selinux state | 14:41 |
| genericv3 | Okay then i will revert selinux changes from my defconfig and check generated full .config file | 14:56 |
| genericv3 | What to do next | 14:56 |
| T42_ | <elros34> no idea, because you still didn't show latest journal log | 14:57 |
| genericv3 | Oops sorry, i forget about it | 14:57 |
| genericv3 | https://paste.opensuse.org/pastes/dbcecd582391 | 15:03 |
| genericv3 | i rebooted manually at the end | 15:03 |
| T42_ | <elros34> at least apexd-bootstrap exit with 0 but still have some strange errors like "DM_DEV_CREATE failed for [com.android.tzdata]: Bad file descriptor" | 15:09 |
| genericv3 | yeah these are strange, i dont have a idea what is this tbh 😅 | 15:12 |
| T42_ | <elros34> ah you disabled selinux in this log | 15:26 |
| genericv3 | yeah, it's still disabled. Still on laptop, can't rebuild kernel | 15:27 |
| genericv3 | i can in approximately 30minutes, probably | 15:27 |
| T42_ | <elros34> I am curious what instruction did you read which suggest that? | 15:28 |
| genericv3 | i found these flags in irc log and simply added them for a quick test. | 15:29 |
| genericv3 | didin't meant to disable selinux | 15:29 |
| T42_ | <elros34> this breaks servicemanager, but logd complain about something else | 15:31 |
| genericv3 | but servicemanager was crashing before this too | 15:32 |
| T42_ | <elros34> this fails currently https://github.com/LineageOS/android_system_logging/blob/lineage-20.0/logd/main.cpp#L78 | 15:50 |
| genericv3 | Finally, i'll be afk for a few more minutes. After that i can rebuild the kernel. | 15:55 |
| genericv2 | rebuilding kernel | 16:18 |
| genericv2 | selinux is Permissive now. | 16:24 |
| genericv2 | https://paste.opensuse.org/pastes/8dcc267f14cb lastest journal with permissive selinux | 16:26 |
| genericv2 | HABUILD_SDK [sagit] kita@suse:~/hadk$ cat out/target/product/sagit/obj/KERNEL_OBJ/.config | grep SELINUX | 16:28 |
| genericv2 | CONFIG_SDFAT_VIRTUAL_XATTR_SELINUX_LABEL="u:object_r:sdcard_external:s0" | 16:28 |
| genericv2 | CONFIG_SECURITY_SELINUX=y | 16:28 |
| genericv2 | # CONFIG_SECURITY_SELINUX_BOOTPARAM is not set | 16:28 |
| genericv2 | # CONFIG_SECURITY_SELINUX_DISABLE is not set | 16:28 |
| genericv2 | CONFIG_SECURITY_SELINUX_DEVELOP=y | 16:28 |
| genericv2 | CONFIG_SECURITY_SELINUX_AVC_STATS=y | 16:28 |
| genericv2 | CONFIG_SECURITY_SELINUX_CHECKREQPROT_VALUE=0 | 16:28 |
| genericv2 | # CONFIG_SECURITY_SELINUX_POLICYDB_VERSION_MAX is not set | 16:28 |
| genericv2 | CONFIG_DEFAULT_SECURITY_SELINUX=y | 16:28 |
| mal | genericv2: you need to copy task_profiles.json from /system/etc (or something like that) to /etc and also add symlink /etc/task_profiles which points to /system/etc/task_profiles | 16:30 |
| mal | this is one of the currently undocumented things needed for android 13 base :) | 16:30 |
| mal | that should fix logd | 16:31 |
| genericv2 | hehe, there is a lot needed for android 13 base docs ig | 16:32 |
| genericv2 | copied and linked | 16:32 |
| genericv2 | libc: Unable to set property "ro.logd.kernel" to "true": connection failed; errno=111 (Connection refused), again :( | 16:33 |
| genericv2 | but it didint terminated, hm | 16:33 |
| T42_ | <elros34> check logcat now | 16:33 |
| genericv2 | IT WORKED | 16:34 |
| genericv2 | i'll be quick and get logcat output | 16:36 |
| genericv2 | https://paste.opensuse.org/pastes/6455949f2079 | 16:38 |
| T42_ | <elros34> search for selinux stubs in hadk-hot | 16:40 |
| genericv2 | i know that, i already added that line but it can't load it | 16:40 |
| genericv2 | WAIT | 16:44 |
| genericv2 | droid-hal-init didin;t exit this timew | 16:44 |
| genericv2 | Only dev-binderfs.mount and systemd-tmpfiles-setup.service failed to start | 16:45 |
| genericv2 | hwservicemanager is throwing some could not find errors rn | 16:46 |
| T42_ | <elros34> so you added selinux stubs lines before you had working logcat? Then remove them and see. Lineage was started in selinux permissive mode too? | 16:46 |
| genericv2 | i already had that selinux stub thing in hwservicemanager.rc, now i made it /data/.stowaways/sailfishos/usr/libexec/droid-hybris/system/lib64/libselinux_stubs.so instead of /usr/libexec/droid-hybris/system/lib64/libselinux_stubs.so and it worked i think | 16:47 |
| genericv2 | SELinux was enforcing in lineage | 16:48 |
| T42_ | <elros34> so what did you change so is permissive? | 16:49 |
| T42_ | <b100dian> mal: so the libprocessgroup spam is not a non-error in A13, interesting, since it was present in A10 too | 16:49 |
| genericv2 | I don't know why but i added "selinux=0" to bc.mk, that made selinux disable xD. | 16:50 |
| genericv2 | removing that made selinux permissive again, selinux is set to permissive in config file. | 16:50 |
| T42_ | <elros34> https://github.com/mer-hybris/hybris-patches/blob/hybris-20.0/system/core/0023-hybris-disable-the-usage-of-libprocessgroup-since-it.patch | 16:51 |
| T42_ | <elros34> @genericv2 from where this hwservicemanager.rc with selinux stubs comes from? I could be wrong but I think this is not something added by default | 16:56 |
| genericv2 | i may added it sometime, i don't remember it well tbh | 16:57 |
| T42_ | <elros34> maybe it's good idea to upload your repos in case there are more unusual changes there | 16:59 |
| genericv2 | i am a bit confused right now, what is the current issue? | 17:01 |
| T42_ | <elros34> not working hw/vnd servicemanager | 17:03 |
| genericv2 | Oh right. it's running but spamming logs, thats the issue | 17:04 |
| T42_ | <elros34> I don't get how changing path from / to /data/.stowaways/sailfishos could fix anything | 17:06 |
| genericv2 | i can restore its previous path to confirm it | 17:07 |
| T42_ | <elros34> is it because of permitted_paths /data? | 17:08 |
| mal | selinux_stubs is in manifest | 17:09 |
| mal | @b100dian yes, seems it's critical in a13 | 17:09 |
| genericv2 | yeah it reallly fixes things. When its /usr/libexec/droid-hybris/system/lib64/libselinux_stubs.so logcat brokes after a few seconds., droid-hal-init also dies | 17:10 |
| genericv2 | these are okay with /data/.stowaways/sailfishos/usr/libexec/droid-hybris/system/lib64/libselinux_stubs.so | 17:10 |
| T42_ | <elros34> what a security hole:) | 17:10 |
| genericv2 | yeah xD | 17:10 |
| genericv2 | linux lore | 17:10 |
| mal | @b100dian maybe there is some way to disable processgroup stuff more | 17:11 |
| mal | from logd, but that needs then some more changes to rc files etc | 17:11 |
| genericv2 | nothing to do?:( | 17:18 |
| T42_ | <b100dian> Well it's not as security hole in a lvm build, but loading things from /data is.. the way we roll | 17:19 |
| T42_ | <elros34> I was wondering how to fix this permited_paths in android13 but this quirk is great:) | 17:20 |
| mal | linker config has become a bit more difficult now because it's generated at runtime | 17:21 |
| T42_ | <b100dian> ah, that two-step madness / mounting of APEX? Does it need changing or just waiting for it to be read? | 17:28 |
| T42_ | <b100dian> I had a bunch of wait-for-linkerconfig sprinkled around here https://github.com/search?q=repo%3Asailfishos-on-sake%2Fdroid-config-sake+linkerconfig&type=code | 17:28 |
| mal | oh, never seen that | 17:30 |
| T42_ | <b100dian> that is A11 but the asus way may have incorporated some things earlier than others | 17:32 |
| genericv2 | was trying to sign in to github, can i try these> | 17:39 |
| T42_ | <b100dian> Not yet, you should fix hwcomposer first | 17:40 |
| T42_ | <b100dian> These appear in boot racr conditions you may not have | 17:40 |
| T42_ | <b100dian> *race | 17:41 |
| T42_ | <b100dian> Like: I make lipatick work but not at first boot, needs a restart to display | 17:41 |
| genericv2 | i don't know what to do but i'll try | 17:44 |
| T42_ | <b100dian> You mentioned it doesnt find some stuff | 17:51 |
| T42_ | <b100dian> I can't access the logcat paste | 17:52 |
| genericv2 | wait a minute please i'll get logcat again | 17:52 |
| genericv2 | https://paste.opensuse.org/ec5a817230fc | 17:55 |
| genericv2 | must be enough | 17:55 |
| genericv2 | i only can understand one thing from this logcat right now: everything is messed up😠| 17:57 |
| T42_ | <b100dian> What's this "Expecting header 0x53595354 but found 0x564e4452. Mixing copies of libbinder?" | 17:58 |
| genericv2 | There is | 18:00 |
| genericv2 | ./system/lib/libbinder.so | 18:00 |
| genericv2 | ./system/lib64/libbinder.so | 18:00 |
| genericv2 | And so much more in phone rn | 18:00 |
| genericv2 | Also ./usr/libexec/droid-hybris/system/lib64/libbinder.so etc. | 18:01 |
| T42_ | <b100dian> Sounds like some kind of version mismatch. | 18:01 |
| genericv2 | Like copies of libbinder.so is different than each other? | 18:04 |
| mal | check gbinder.conf in /etc you probably have wrong version, it should have value 31 | 18:06 |
| mal | https://github.com/mer-hybris/droid-hal-configs/blob/android13/sparse-13/etc/gbinder.conf | 18:06 |
| genericv2 | it does not exist | 18:08 |
| genericv2 | created now | 18:08 |
| T42_ | <b100dian> do you have `android_version_major` defined in droid config spec? | 18:09 |
| mal | @b100dian the root of the issue is incorrect submodule version | 18:09 |
| genericv2 | yeah, i have. but its set to 33 | 18:10 |
| genericv2 | mal i can rebuild everything right now/ | 18:10 |
| mal | also build newer mce so you don't have to hack things | 18:11 |
| mal | build_packages.sh --mw=https://github.com/sailfishos/mce.git | 18:11 |
| genericv2 | first of all where was the submodule | 18:12 |
| genericv2 | compiling mce | 18:13 |
| T42_ | <b100dian> I think it's nice to have branches for android versions in the submodule.. maybe a bit of headache to refactor existing functionality, but right now the "get latest" is also confusing a lot | 18:13 |
| genericv2 | * Building of mce finished successfully | 18:13 |
| mal | @b100dian I should get one thing fixed in that android13 branch and merge it | 18:14 |
| T42_ | <b100dian> The submodule of hybris/droid-configs | 18:15 |
| genericv2 | i am afk, i'll do everything when i am back. | 18:18 |
| genericv2 | i am back | 18:52 |
| genericv2 | anything? | 18:52 |
| genericv2 | i'll try to add our previous modifications to sparsa while waiting | 18:58 |
| T42_ | <b100dian> You need to change rhe submodule to point to the android13 branch | 19:17 |
| genericv2 | on which directory, i forget everything | 19:21 |
| genericv2 | okay got it | 19:23 |
| genericv2 | no didint got it | 19:25 |
| genericv2 | like git submodule add -b android13 https://github.com/mer-hybris/droid-hal-configs droid-configs-device ? | 19:28 |
| genericv2 | gives an error nvm | 19:28 |
| genericv2 | hm, looks like it's okay now | 19:30 |
| mal | genericv2: don't you have submodule already | 19:42 |
| mal | in which case you just go to submodule folder and do git fetch, then git checkout android13 | 19:43 |
| genericv2 | i didin't knew that, this is my first time messing with git submodules | 19:43 |
| genericv2 | okay correct the submodule branch is enough for now? | 19:46 |
| genericv2 | what did i just write | 19:46 |
| mal | add also the task profile things | 19:46 |
| mal | to sparse and other things you might need | 19:46 |
| genericv2 | There was some rc files, i'll add them | 19:48 |
| genericv2 | But i'll continue tomorrow, thanks everyone! | 19:49 |
| mal | those are in submodule already | 19:49 |
| genericv2 | a | 19:49 |
| genericv2 | true | 19:49 |
| genericv2 | right* | 19:49 |
| mal | as you remember you copied the things from the submodule repo | 19:49 |
| genericv2 | yeah yeah, i remembered it now. | 19:50 |
| genericv2 | bye for now :) | 19:51 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!