| mal | so it's tracker probably, or libtinysparql which tracker uses | 00:00 |
|---|---|---|
| Mister_Magister | well it is running thats all i can tell really | 00:00 |
| mal | is tracker still stuck in background? | 00:00 |
| Mister_Magister | yes | 00:00 |
| mal | what happens if you kill it when trying to start gallery? | 00:01 |
| Mister_Magister | nothing | 00:01 |
| Mister_Magister | already tried | 00:01 |
| mal | maybe install debug symbols from libtinysparql and libgio to see what those are trying to do when it hangs | 00:02 |
| Mister_Magister | yessir | 00:02 |
| mal | and maybe gdb also tracker | 00:02 |
| mal | I'll go offline now | 00:02 |
| Mister_Magister | me too | 00:02 |
| Mister_Magister | eepy time | 00:02 |
| Mister_Magister | it's in tracker_bus_connection_new | 00:06 |
| Mister_Magister | nothing interesting in either gallery or localsearch gdb | 00:09 |
| Mister_Magister | maybe i should try with kernel from newer android that hopefully has all that stuff correctly backported | 00:15 |
| Mister_Magister | yeah latest one has changes from 9 years ago not 15 | 00:30 |
| Mister_Magister | naah it doesn't have seccomp bpf backported at all :/ | 00:31 |
| Mister_Magister | okay there's android 17 kernel with BPF implemented | 00:33 |
| Mister_Magister | android 10 sorry lineage 17 | 00:33 |
| Mister_Magister | and has getrandom | 00:34 |
| *** n2klb_ is now known as n2klb | 02:00 | |
| T42 | <nightishaman> Sharks_ I need dual sim tho since I only use eSIMs and it’s the second one | 08:27 |
| Sharks_ | Has anyone come across this before? On 240 or 480p in Youtube, stock browser, I get weird visual artifacts. All other resolutions play fine. Captured some examples: | 13:01 |
| Sharks_ | https://ibb.co/album/ZBC9Dt | 13:01 |
| Mister_Magister | mal: it might be stupid idea but i'm gonna try using kernel from android 10 instead of android 5 because it has all that stuff backported and is relatively newer | 16:44 |
| Mister_Magister | should work fiiiiine…? | 16:45 |
| mal | Mister_Magister: usually it works, I use android 16 kernel on fp4 and fp5 with a14 base | 17:08 |
| Mister_Magister | well thats only 2 android versions not 5 xd | 17:09 |
| mal | I think on fp2 my kernel is from some newer android version | 17:09 |
| Mister_Magister | it rebooted straight into recovery xd | 17:09 |
| mal | oh | 17:09 |
| Mister_Magister | i'll try with the android 5 defconfig | 17:10 |
| Mister_Magister | ayy with original defconfig i got to the ramdisk | 17:16 |
| Mister_Magister | ayy it's booting! | 17:16 |
| Mister_Magister | the new one had some changes in the cpu and also no modules lmao | 17:16 |
| Mister_Magister | it's not all sunshine and rainbows though because it reboots after booting to gui :( | 17:18 |
| Mister_Magister | I just wanna know if gallery will start | 17:19 |
| Mister_Magister | mal: I don't think gallery is starting on this kernel either :/ | 17:24 |
| Mister_Magister | i'll try once again this time with modules but i don't have any hope | 17:25 |
| Mister_Magister | i can also try and find some other kernels but idk | 17:27 |
| Mister_Magister | mal: uhhhh | 17:36 |
| Mister_Magister | i packaged rpm with modules and installed the rpms instead of flashing just kernel | 17:36 |
| Mister_Magister | and you won't imagine, gallery works and it no longer reboots | 17:36 |
| Mister_Magister | can i get a hurray?! | 17:37 |
| Mister_Magister | :D | 17:37 |
| Mister_Magister | bluetooth still doesn't work but lemme test everything | 17:38 |
| Mister_Magister | dang camera seems broken | 17:41 |
| mal | did camera work with older kernel? | 17:42 |
| Mister_Magister | yeah | 17:43 |
| Mister_Magister | everything was working on 5.0 and i mean everything | 17:44 |
| Mister_Magister | GStreamer encountered a general supporting library error. | 17:44 |
| Mister_Magister | ah /system/bin/mm-qcamera-daemon is segfaulting | 17:45 |
| Mister_Magister | but ye gallery works, bt still doesn't, fm still redirects microphone | 17:45 |
| Mister_Magister | so one thing got fixed other got broken | 17:46 |
| mal | I meant did camera work in the earlier 5.1 attempt? | 17:46 |
| Mister_Magister | yes | 17:46 |
| mal | ok | 17:46 |
| Mister_Magister | let me see if something changed in defconfig | 17:47 |
| Mister_Magister | nope nothing that I can see | 17:49 |
| Mister_Magister | mal: can you deduce anything from this? https://paste.opensuse.org/pastes/6d01dd5c4eff | 17:50 |
| mal | does HYBRIS_ENABLE_LINKER_DEBUG_MAP=1 help with getting backtrace from gdb? | 17:58 |
| mal | anything in logcat? | 17:58 |
| Mister_Magister | i mean its android binary so I won't get much of a backtrace no? | 17:59 |
| mal | hmm, yeah, could be | 18:00 |
| Mister_Magister | in gdb it just… exits | 18:00 |
| Mister_Magister | no crash no nothing | 18:00 |
| Mister_Magister | Inferior 1 (process 8381) exited normally | 18:01 |
| mal | are you sure the process is not already running? | 18:03 |
| Mister_Magister | shouldn't be if its segfaulting | 18:07 |
| Mister_Magister | it is huh | 18:07 |
| Mister_Magister | but once i launch camera its segfaulting periodically | 18:10 |
| Mister_Magister | got it to run in gdb, it just periodically (every second) loads set of libraries does some stuff | 18:13 |
| rinigus | @b100dian: after update to 5.1.0.11, it gets stuck on LED with blank screen after entering home encryption key | 18:14 |
| rinigus | you mentioned that it took longer for you to boot. what duration are we talking about? | 18:15 |
| Mister_Magister | mal: logcat https://paste.opensuse.org/pastes/8dbca75d237b | 18:15 |
| T42 | <b100dian> ringius: did you update the submodule of hybris and rebuilt the lineage hybris-hal then the droid-hal rpms? | 18:16 |
| Mister_Magister | maybe android 10 is bit too new | 18:16 |
| T42 | <b100dian> there's a package that needs that update, I think its qt5scenegraph-hwcomposer ? | 18:17 |
| rinigus | @b100dian: I used repo sync - that should have pilled it all | 18:17 |
| T42 | <b100dian> does that work for submodules though? | 18:17 |
| rinigus | and used OTA from OBS, so all updates there should work | 18:18 |
| rinigus | re submodule. let me check | 18:18 |
| rinigus | but boot for you - its not minutes, right? | 18:18 |
| Mister_Magister | mal: i'll try yet another kernel | 18:19 |
| T42 | <b100dian> rinigus: Not minutes, but I do have the feeling it is way longer. You mean that was it, a very long boot? O_O | 18:24 |
| rinigus | @b100dian: switched it off after 10 minutes or so | 18:24 |
| T42 | <b100dian> nono it is not thaaat longer | 18:25 |
| rinigus | never got to GUI, never got via wifi/ssh into it. no reply to ping | 18:25 |
| rinigus | libhynris looks to be updated | 18:25 |
| rinigus | and droid-hal is all new | 18:25 |
| rinigus | same commit ID for libhybris and with its submodule https://github.com/sailfishos-sony-nagara/libhybris/tree/385778eeaa06b762f2c1ed71d10ab3609dc4f8ca | 18:26 |
| T42 | <b100dian> and you don't have telnet before the password entry by any chance :-S | 18:27 |
| rinigus | nope. I don't think so. | 18:28 |
| T42 | <b100dian> You would if you'd flash the hybris recovery ? I think I'm daily driving that | 18:28 |
| rinigus | which means I will have to reflash LOS recovery first, no? | 18:29 |
| T42 | <b100dian> no, just swap the boot image from hybris-boot to hybris-recovery.img in flashing mode | 18:29 |
| rinigus | what does that do? | 18:30 |
| T42 | <b100dian> fastboot flash hybris-recovery.img ? it just differs from a debug parameter, let me look it up | 18:31 |
| rinigus | boot with telnet? | 18:31 |
| T42 | <b100dian> hopefully, yes | 18:31 |
| T42 | <b100dian> there it is, it _should_ set this alwaysdebug to true https://github.com/mer-hybris/hybris-boot/blob/master/Android.mk#L218C26-L218C46 | 18:34 |
| T42 | <b100dian> Elros or mal correct me if I'm wrong | 18:34 |
| T42 | <b100dian> rinigus: did it? or you need to boot into lineage recovery..? | 18:41 |
| rinigus | just flashed, let's see | 18:42 |
| rinigus | maybe, let's see if I can get in | 18:43 |
| rinigus | @b100dian: nope. will try with LOS recovery | 18:46 |
| T42 | <b100dian> rinigus: you don't get anything in dmesg, that would be fixed by e.g. a static IP? What port did you try 2323? | 18:49 |
| rinigus | I have dmesg with device. added address to my PC (10.42.66.65) and tried to get into 10.42.66.66. | 18:50 |
| rinigus | @b100dian: no ping. didn't try 2323. or is IP different? | 18:51 |
| T42 | <b100dian> 192.168.2.15 should pe phone | 18:52 |
| T42 | <elros34> if you use hybris-recovery then it's 23 port | 18:52 |
| rinigus | hmm. so I was using wrong IP. I wonder where do we set IP address | 18:52 |
| Mister_Magister | rinigus: you don't | 18:54 |
| Mister_Magister | its in init script | 18:54 |
| Mister_Magister | after device is fully booted you can change it in settings | 18:54 |
| rinigus | thanks | 18:55 |
| Mister_Magister | but there's no point in changing it unless it conflicts | 18:56 |
| rinigus | @b100dian and @elros34: port is 2323 and IP 192.168.2.15 | 18:56 |
| rinigus | let's see | 18:56 |
| *** xmn_ is now known as xmn | 19:08 | |
| rinigus | @b100dian: looks like some android linker issue: CANNOT LINK EXECUTABLE "/apex/com.android.os.statsd/bin/statsd": library "libstatssocket.so" | 19:10 |
| rinigus | same with "/apex/com.android.media.swcodec/bin/mediaswcodec": library "libmedia_codecserviceregistrant.so" not found: | 19:11 |
| rinigus | "/system/bin/mediaserver": library "libicu.so" | 19:11 |
| rinigus | hmm. it used all to work before update | 19:11 |
| Mister_Magister | story of my life | 19:11 |
| rinigus | and overlays are mounted before droid-hal | 19:12 |
| T42 | <b100dian> hmm libicu is the android one hopefully? | 19:12 |
| Mister_Magister | sfos updates love to break my apps and my devices and now i'm deep in kernel to fix one of the deices | 19:12 |
| rinigus | I presume so | 19:13 |
| rinigus | @b100dian ^ | 19:13 |
| rinigus | @b100dian: are we expected to run /system/bin/mediaserver ? | 19:17 |
| rinigus | from android side? | 19:18 |
| T42 | <b100dian> good question - let me check | 19:18 |
| T42 | <b100dian> I have it running | 19:19 |
| rinigus | OK, I'll continue tomorrow. will compare with older logs... | 19:20 |
| rinigus | getting late and its clear that this will take a bit of time | 19:21 |
| T42 | <b100dian> feel free to share logs maybe something pops up | 19:21 |
| Mister_Magister | found yet another kernal | 19:31 |
| Mister_Magister | lol now i'm recognizing from where they took the source from becaues they exhibit exact same issues xd | 19:36 |
| Mister_Magister | *gasp* this kernel has new error | 19:42 |
| Mister_Magister | and i got nowhere | 20:24 |
| Mister_Magister | mal: i'm quite confident it's something to do with the memfd syscall because without it tracker woudl just crash and now it just hangs so like i'm quite confident. I could use the patches from the android 10 kernel instead of your fairphone ones but thats also a looot of work | 20:31 |
| T42 | <b100dian> rinigus: maybe we need to apply the linkerconfig patch from march https://github.com/mer-hybris/hybris-patches/commits/hybris-21.0/ ? | 21:28 |
| T42 | <b100dian> on lineage 19.1 I used a hack to "wait" from either system or user services, for correct linkerconfig / apex mount https://github.com/sailfishos-on-sake/droid-config-sake/commit/735f6a73c349b59a9ed161ebce5b2d3d55c1eeed | 21:29 |
| mal | @b100dian do you see in that device messages about waiting for vold? | 22:32 |
| T42 | <b100dian> Not exactly. "Xperia1IV apexd: Could not retrieve vold service: Skipping retrieving vold service" | 22:33 |
| T42 | <b100dian> I think we have a vold patch, let me find it | 22:34 |
| T42 | <b100dian> Ah yes, that's why the message is different https://github.com/sailfishos-sony-nagara/main/blob/main/patches/system/apex/0001-Don-t-wait-for-vold.patch | 22:34 |
| mal | https://github.com/mer-hybris/hybris-patches/blob/hybris-23.2/system/apex/0003-hybris-Don-t-look-for-vold-in-apexd.patch | 22:36 |
| mal | so a bit different way | 22:36 |
| mal | what speeds up boot usually and sometimes solves some other issues also | 22:36 |
| T42 | <b100dian> Thanks - yeah, I should try that, slow boot seems to be a problem, not sure if because of vold | 22:46 |
| T42 | <b100dian> however the linker errors seem to be just that, linkerconfig not done mounting - although I would expect services started from .rc scripts such as mediaserver to not be buggy :hmm: | 22:47 |
| Mister_Magister | mal: I fixed camera! (with help of android guy) | 22:48 |
| Mister_Magister | and if you wonder how it will become painfully obvious | 22:48 |
| Mister_Magister | "I wonder what this commit that says "Requires updated camera binaries" might be doing…" https://github.com/RenanQueiroz/android_kernel_motorola_msm8226/commit/c0e538b1499299bcb1caa3c1375cf75bb8cf7d2b | 22:49 |
| Mister_Magister | simple revert and it works | 22:49 |
| Mister_Magister | that means only fm radio and bt are left broken | 22:50 |
| Mister_Magister | wait hold on a second localsearch is segfaulting | 23:24 |
| Mister_Magister | oh, my oopsie | 23:26 |
| Mister_Magister | fixed | 23:36 |
| Mister_Magister | wait, bluetooth | 23:38 |
| Mister_Magister | WAIT | 23:38 |
| Mister_Magister | I just had random memory unlocked aboout having to build mal's backports for bluez5 | 23:38 |
| Mister_Magister | ayay | 23:43 |
| Mister_Magister | it won't build becaues it uses random32 and both me and fp2 has backported rename changes | 23:44 |
| Mister_Magister | yes yes i found your random comment mal about replacing random32 with prandom_u32 i got it xd | 23:49 |
| Mister_Magister | dang, that wasn't it sadly :( | 23:50 |
| Mister_Magister | hmm faq states to have CONFIG_CRYPTO_CMAC, except i don't have that one :( | 23:51 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!