liux | jusa, mal, the breacktrace of pulseaudio crash was pasted here: https://paste.ubuntu.com/p/Zcxg4BNXNJ/ | 00:46 |
---|---|---|
liux | i've installed debuginfo for glibc and pulseaudio but stiil some "?" there | 00:55 |
birdzhang | liux: mal will ask you for full log :p | 02:10 |
liux | birdzhang, full log before pulseaudio crashed was pasted here: https://paste.ubuntu.com/p/6FRvZVhNvk/ | 04:36 |
Mister_Magister | mal: do you need to replace linker in | 09:08 |
Mister_Magister | lineage 15? | 09:08 |
Mister_Magister | ports | 09:08 |
birdzhang | !ghosalmartin | 09:14 |
merbot | birdzhang: Error: "ghosalmartin" is not a valid command. | 09:14 |
birdzhang | okay, it's not working now | 09:15 |
birdzhang | !seen ghosalmartin | 09:16 |
merbot | birdzhang: ghosalmartin was last seen in #sailfishos-porters 3 weeks, 6 days, 11 hours, 40 minutes, and 1 second ago: <ghosalmartin> before | 09:16 |
birdzhang | seems only ghosalmartin and i have UI crash issue on 15.1 | 09:17 |
mal | Mister_Magister: what? | 09:18 |
T42 | <adampigg> mal, on androd 9 we had to replace the linker in /system with a custom one because it refused to load the selinux_stubs lib from /usr/libexec | 09:19 |
T42 | <adampigg> my idea is to overlay or bind mount over /system ;) | 09:20 |
mal | why did it refuse? | 09:20 |
mal | there should be no need for such things | 09:20 |
mal | maybe you are missing something | 09:20 |
T42 | <adampigg> we think it was due to the path.... | 09:20 |
T42 | <adampigg> mal: | 09:26 |
T42 | <adampigg> 05-06 13:30:55.356 3907 3907 E linker : library "/usr/libexec/droid-hybris/system/lib64/libselinux_stubs.so" ("/usr/libexec/droid-hybris/system/lib64/libselinux_stubs.so") needed or dlopened by "/system/bin/hwservicemanager" is not accessible for the namespace: [name="(default)", ld_library_paths="", default_library_paths="/system | 09:26 |
T42 | /lib64", permitted_paths="/system/lib64/drm:/system/lib64/extractors:/system/lib64/hw:/system/product/lib64:/system/framework:/system/app:/system/priv-app:/vendor/framework:/vendor/app:/vendor/priv-app:/odm/framework:/odm/app:/odm/priv-app:/oem/app:/system/product/framework:/system/product/app:/system/product/priv-app:/data:/mnt/expand"] | 09:26 |
mal | and you have the patch in system/core? | 09:27 |
T42 | <adampigg> not sure, which /patch? | 09:28 |
mal | the one I gave yesterday | 09:28 |
T42 | <adampigg> i will look later, m1ster_magister will know likely | 09:29 |
Mister_Magister | what me | 09:33 |
T42 | <adampigg> you built it! | 09:34 |
T42 | <adampigg> :D | 09:34 |
Mister_Magister | @adampigg and u typed wrong nick | 09:38 |
Mister_Magister | mal: tbh i applied ur patch after replacing | 09:40 |
Mister_Magister | mal: this patch? https://pastebin.com/7sDP5NZC | 09:47 |
Mister_Magister | mal: if this patch you mean then i will check today probably if it works without replacing | 09:52 |
mal | no need for any extra patches, so whatever random patch you are talking about when talking about replacing then don't use the old random patch | 10:03 |
Mister_Magister | mal: it's not random patch u gave it to me | 10:04 |
mal | I have no idea what you are talking about anymore | 10:05 |
mal | what patch are you replacing with what? | 10:05 |
Mister_Magister | not patch | 10:06 |
Mister_Magister | linker | 10:06 |
Mister_Magister | mal: you mean that https://pastebin.com/7sDP5NZC this patch fixes the problem so i won't have to replace linker? | 10:06 |
deathmist | I have pulseaudio working after adapting audio_policy.conf \o/ still no in-call audio or mic tho (mic works outside calls fine); 15.1 base, journalctl https://pastebin.com/nhvGJN13 dmesg https://pastebin.com/u732hJ7H | 10:32 |
mal | Mister_Magister: no idea, I just know that replacing linker should not be needed | 10:38 |
mal | deathmist: do you have miniaf running and the matching pulseaudio plugin installed as told in hadk pdf | 10:38 |
Mister_Magister | mal: i'll replace it back and report on problems there | 10:45 |
Mister_Magister | then* | 10:45 |
birdzhang | deathmist: try link /vendor/etc/*audio*.[xml|conf] to /etc/ | 10:52 |
birdzhang | i fixed sound by make a symbol link /vendor/etc/r_submix_audio_policy_configuration.xml to /etc/ on vince | 10:55 |
deathmist | birdzhang: I have that same file, will try with it alone and afterwards with all the other ones :) | 10:56 |
T42 | <adampigg> @eugenio_g7 im sure latte has non linear charging! | 10:56 |
r0kk3rz | lol t42 | 10:57 |
birdzhang | r0kk3rz: T1000 is coming :P | 10:57 |
r0kk3rz | missed opportunity for T1000 | 10:58 |
T42 | <eugenio_g7> @adampigg yes agreed :) | 11:08 |
Mister_Magister | t420 like thinkpad | 11:08 |
r0kk3rz | most of the people on there are smoking something so t420 fits | 11:15 |
Mister_Magister | see ima genius | 11:18 |
deathmist | birdzhang: tried the symlinks and it still doesn't work :/ mal miniafservice is running and here's installed pulse packages https://pastebin.com/r5pW538W which section was that exactly in the hadk? | 12:54 |
T42 | <birdzhang> @deathmist paste the dmesg log when playing sound, like trigger the pulley menu | 12:56 |
T42 | <birdzhang> btw, does headphone works ? | 12:57 |
deathmist | yes headphones work in stereo mode (haven't tested in call yet), but it seems I'm getting spammed by time_daemon and qti services >:( it seems tfa98xx 9-0036: state changes between 1 and 0, nothing much in dmesg | 13:02 |
T42 | <birdzhang> dmesg|grep tfa98xx_load_container | 13:05 |
T42 | <birdzhang> and paste your /lib/firmware files | 13:05 |
Mister_Magister | something wrong with tfa? | 13:06 |
Mister_Magister | what is the problem? | 13:06 |
T42 | <birdzhang> i'm wondering the sound card firmware not loaded | 13:07 |
deathmist | https://pastebin.com/SmDwMBT8 and uuuh I have no /lib/firmware, Mister_Magister I have no audio during a call but it works everywhere else | 13:07 |
Mister_Magister | oh | 13:07 |
T42 | <birdzhang> okay, Firmware init complete | 13:08 |
Mister_Magister | i had problem once with tfa that caused me reboots | 13:08 |
mal | some android 9 devices need some extra things to find firmware https://pastebin.com/khcCKLEr | 13:09 |
T42 | <birdzhang> @deathmist do you have /vendor/lib/firmware ? | 13:09 |
mal | because android has split the firmware to too many places | 13:09 |
Mister_Magister | andrdoid9? someone other than us are porting android9? | 13:09 |
deathmist | birdzhang: I don't have that, but I do have /vendor/firmware (which I've also symlinked to /etc/firmware) | 13:10 |
deathmist | I'm on 15.1 base :p | 13:11 |
T42 | <birdzhang> @deathmist i know, me too | 13:12 |
Mister_Magister | dunno what mal is talking about :P | 13:13 |
mal | deathmist: find where the firmware files are, many devices need something to be symlinked to /lib/firmware | 13:14 |
mal | Mister_Magister: I can't remember who is using what base | 13:15 |
Mister_Magister | mal: yeah i'm just joking :P no hard feelings | 13:15 |
mal | I used to remember but not anymore | 13:18 |
Mister_Magister | yeah :P | 13:18 |
* Mister_Magister pats mal | 13:18 | |
mal | too many people and devices | 13:18 |
Mister_Magister | mal: back in the day there weren't many ports really | 13:18 |
mal | also my memory is quite much worse than it used to be, that's what happens when you don't sleep well | 13:18 |
Mister_Magister | yaehish | 13:19 |
Mister_Magister | poor poor mal | 13:19 |
deathmist | time_daemon and qti services aren't required on SFOS right? they just keep starting and getting killed over and over again | 13:19 |
Mister_Magister | time_daemon isn't | 13:19 |
Mister_Magister | dunno about second one | 13:19 |
deathmist | they both spam dmesg with messages like these https://pastebin.com/w5jhraxp | 13:22 |
Mister_Magister | i know, you can disable time_daemon | 13:24 |
Mister_Magister | but dunno about second one fix it or disable it | 13:24 |
deathmist | where would I disable it? running grep -rniw . -e "*time_daemon*" yields nothing on $ANDROID_ROOT | 13:48 |
Mister_Magister | /init.rc or /init.qcom.rc | 13:49 |
T42 | <birdzhang> @deathmist /vendor/etc/init/ | 13:54 |
T42 | <birdzhang> remount /vendor is needed | 13:55 |
Mister_Magister | you shouldn't modify them in /vendor | 13:55 |
mal | in 15.1 the .rc files can be overridden in /usr/libexec/droid-hybris/system/etc/init/ | 13:56 |
Mister_Magister | yeah thats what i mean | 13:56 |
mal | or actually individual services can be | 13:56 |
mal | the files already there can used as an example to see how a service can be disabled | 13:57 |
Mister_Magister | mal: btw i had to add LD_PRELOAD with selinux_stubs to servicemanager to get it working. you know something about that? | 13:57 |
T42 | <birdzhang> if i want disable only one service, it's not working just create a new rc file and put service in it with _HYBRIS_DISABLED | 14:00 |
Mister_Magister | why not add line disabled? | 14:00 |
T42 | <adampigg> I think disabled servicrs can still be started | 14:03 |
deathmist | I just disabled it via a /usr/libexec/droid-hybris/system/etc/init/time_daemon.rc file and after a reboot it's still there https://pastebin.com/EX3uV7b0 | 14:06 |
mal | Mister_Magister: of course it's needed https://github.com/mer-hybris/android_frameworks_native/commit/531f3f97be9d25ab5a6072148909e70984f9161c | 14:07 |
mal | Mister_Magister: so you didn't apply all patches | 14:07 |
mal | deathmist: where was the original time_daemon.rc? | 14:24 |
mal | or the service | 14:25 |
T42 | FranzKavka was added by: FranzKavka | 14:43 |
Thaodan | Hey the bot will be down at 24.07.2019 9:40 - 11:40 CEST. The downtime of your server will approximately last for 30 minutes. A fix will applied to the hypervisor of the server. If you want to read more ask or look at: https://forum.netcup.de/administration-eines-server-vserver/vserver-server-kvm-server/p121256-umgang-mit-sicherheitsl%C3%BCcken-zombieload-mds | 16:35 |
deathmist | mal: I'm finally back, there was no original time_daemon.rc file, just the /vendor/bin/time_daemon binary | 17:17 |
deathmist | find / -name "time_daemon*" only returned that binary | 17:17 |
Mister_Magister | mal: so, without replacing linker i'm getting this https://paste.opensuse.org/view//53280473 | 17:23 |
mal | deathmist: it's in some other .rc then | 17:28 |
mal | Mister_Magister: and that is with preload? | 17:29 |
Mister_Magister | with what preload | 17:29 |
Mister_Magister | i didn't preload stubs to droid-hal-init | 17:29 |
Mister_Magister | servicemanager has preload tho | 17:29 |
mal | Mister_Magister: hmm, what libhybris did you use? | 17:43 |
Mister_Magister | mal: fae0a52632893cbf3bb7ce0762147b7067db176b | 17:45 |
Mister_Magister | slightly patched | 17:45 |
mal | I assume patched in the hwc2 part | 17:46 |
Mister_Magister | maybe | 17:47 |
Mister_Magister | and some stuff to make it compile but nothing related | 17:47 |
mal | yes, android 9 needed some patching | 17:48 |
mal | but only to hwc2 code | 17:48 |
Mister_Magister | okay okay so what about that error? | 17:48 |
mal | not sure yet | 17:48 |
T42 | <DSstill01> I know it | 17:52 |
mal | Mister_Magister: also maybe don't reinvent the wheel for the other things related to android 9 (like droidmedia etc), I have patches for most already on my computer | 17:52 |
Mister_Magister | mal: you have patches ready since how long huh | 17:53 |
T42 | <DSstill01> If you have 32 bit device you dont have to set 'lib64' | 17:53 |
Mister_Magister | mal: i'm gonna patch some lineage16 sources now | 17:53 |
T42 | <DSstill01> It's a mistake that i also did when copy pasting things from the device in the tutorial | 17:53 |
mal | I would assume that is 64-bit device but not sure | 17:53 |
Mister_Magister | ofcourse it is | 17:54 |
T42 | <DSstill01> Try to check if that lib exists in that place | 17:54 |
Mister_Magister | mal: do you have some lineage16 patches that i could apply? | 17:56 |
deathmist | mal: found it, time_daemon defined in /vendor/etc/init/hw/init.qcom.rc, so how should I go about keeping it disabled for new flashes? | 17:59 |
T42 | <DSstill01> Add a "disabled" line in the service | 17:59 |
Mister_Magister | no | 18:00 |
T42 | <DSstill01> Yes, unless something runs it, it wont start | 18:01 |
Mister_Magister | you can just put that file in /usr/libexec/droid-hybris/system/etc/init/ and add disabled to service there and put that in your build (i don't know how to disable single service instead of putting whole file tho) | 18:01 |
Mister_Magister | you shouldn't modify /vendor partition | 18:01 |
mal | it should ignore the services with same name | 18:02 |
mal | file doesn't matter | 18:02 |
Mister_Magister | then just put file with random name and that service with addition of disabled | 18:03 |
Mister_Magister | mal: if u get any idea please ping me | 18:03 |
Mister_Magister | let's prelaod it from /data/.stowaways/sailfishos/usr/lib… as /data is allowed | 18:04 |
* Mister_Magister why it doesn't sound stupid it my head | 18:05 | |
mal | heh, that is quite hacky | 18:05 |
Mister_Magister | quite :P | 18:05 |
Mister_Magister | and won't work if we wanna lvm it | 18:05 |
Mister_Magister | well there is mount --bind :p | 18:06 |
Mister_Magister | i don't see any other way really as linker comes from vendor and only those paths are allowed | 18:08 |
T42 | <DSstill01> Have you tried what i told you to do? | 18:10 |
Mister_Magister | @DSstill01 your ideas are without any kind of sense so no, i didn't | 18:10 |
T42 | <DSstill01> WHAAAT?! I had the same error. Always listen to what the squirrel says! | 18:11 |
Mister_Magister | eh | 18:11 |
T42 | <DSstill01> Retired dev doesn't mean retarded dev! | 18:12 |
Mister_Magister | meh let's have some fun huh (sry for spam) | 18:13 |
Mister_Magister | @DSstill01 then yes it exists | 18:13 |
Mister_Magister | what now? | 18:13 |
T42 | <DSstill01> You copy pasted the not-found path from the error and it exists? | 18:13 |
Mister_Magister | yes ofcourse | 18:13 |
T42 | <DSstill01> o.O | 18:14 |
Mister_Magister | oh i'm not sure if you read link | 18:14 |
Mister_Magister | there is no "not-found" | 18:14 |
Mister_Magister | "<DSstill01> Retired dev doesn't mean retarded dev!" can you repeat that? | 18:14 |
T42 | <DSstill01> Ahhh | 18:15 |
T42 | <DSstill01> *_* | 18:15 |
T42 | <DSstill01> I didnt read the full line lol | 18:15 |
Mister_Magister | that was fun :P | 18:15 |
Mister_Magister | spammy tho | 18:15 |
T42 | <DSstill01> No, it was a potential fiz | 18:16 |
T42 | <DSstill01> Future reference :P | 18:16 |
Mister_Magister | mccreary: ping | 18:17 |
Mister_Magister | mal: or maybe we can prelaod something to allow /usr path? | 18:23 |
mal | Mister_Magister: there is another way, wait a moment | 18:23 |
Mister_Magister | okay i'm waiting | 18:24 |
Mister_Magister | gotta make some pierogi's | 18:24 |
mal | what kind of ld.config* files do you have on the device? | 18:24 |
Mister_Magister | in what path | 18:24 |
mal | anywhere | 18:24 |
mal | probably in /system or /vendor | 18:25 |
Mister_Magister | find / -name ld.config\* i'm on it | 18:25 |
Mister_Magister | o7 | 18:25 |
Mister_Magister | mal: single file /system/etc/ld.config.28.txt | 18:32 |
mal | Mister_Magister: I need go for a moment but I'll then give you a fix, not the best but should work | 18:39 |
Mister_Magister | mal: okay! no hurry! | 18:39 |
mal | some bind mount magic after editing a copy of that file | 18:40 |
Mister_Magister | so it's mount bind after all heh xd but not whole linker this time | 18:41 |
Mister_Magister | mal: i'm going to sleep hit me up in pm (i'm always online and pm | 20:32 |
Mister_Magister | s are saved) | 20:32 |
Mister_Magister | why do i missclick | 20:32 |
mccreary | Mister_Magister, I see we've missed each other. I'll try to catch you later, but you could just ask your question | 23:24 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!