T42 | markvesime was added by: markvesime | 10:19 |
---|---|---|
T42 | <markvesime> Hi, I have a slightly defective pinephone I'd like to give to someone who can help with the sfos port....adampigg mentioned mal in particular. Feel free to pm me for details | 10:21 |
piggz | rinigus: youre a bit of an expert with the lvm recovery image ... once booted, is there anyway to continue normal boot from it? | 10:35 |
rinigus | piggz: (brushing my new lvm expert sign) no idea, sorry. good question, would be great to know. so far I have not found any obvious way, but wasn't investigating it heavily. maybe some true expert can step in | 11:11 |
piggz | mal: ^^ ? | 11:12 |
rinigus | mighty17: in general, we have been advised against including anything extra to the builds. users can add openstore if needed. | 11:15 |
rinigus | in the end, it is up to you and probably is fine. | 11:17 |
rinigus | mighty17: looks like wonderful platform then :) | 11:19 |
T42 | <meierrom> r0kk3rz: are you there? | 14:17 |
mal | Mister_Magister: try new libhybris | 15:22 |
Mister_Magister | about that, wanna ssh to phone: | 15:22 |
Mister_Magister | ? | 15:22 |
Mister_Magister | mal: got bus error | 15:33 |
mal | ? | 15:33 |
Mister_Magister | in the output | 15:34 |
mal | show the error | 15:34 |
Mister_Magister | Bus error | 15:35 |
Mister_Magister | literally nothing else | 15:35 |
Mister_Magister | HYBRIS_TRACE=1 HYBRIS_TRACE_UNHOOKED=1 HYBRIS_TRACE_HOOKED=1 EGL_PLATFORM=hwcomposer QT_QPA_PLATFORM=hwcomposer test_hwcomposer | 15:35 |
Mister_Magister | Bus error | 15:35 |
mal | kernling: ^ | 15:36 |
kernling | Mister_Magister, and strace ? | 15:38 |
Mister_Magister | on it | 15:39 |
Mister_Magister | mal: kernling: https://paste.opensuse.org/4147d107 | 15:39 |
kernling | and gdb ? | 15:40 |
Mister_Magister | here https://paste.opensuse.org/a1f64d08 | 15:41 |
kernling | Mister_Magister, any difference if you remove HYBRIS_TRACE_HOOKED=1 or HYBRIS_TRACE_UNHOOKED=1? | 15:42 |
kernling | i.e try with only one of them | 15:42 |
kernling | and then the other | 15:42 |
Mister_Magister | nope | 15:43 |
kernling | and both removed? | 15:46 |
Mister_Magister | now its standard segfault | 15:47 |
kernling | Mister_Magister, and try with HYBRIS_TRACE=1 HYBRIS_TRACE_DYNHOOKED=1 EGL_PLATFORM=hwcomposer QT_QPA_PLATFORM=hwcomposer test_hwcomposer | 15:50 |
Mister_Magister | bus error | 15:50 |
kernling | interesting | 15:50 |
kernling | i cannot find some logs anymore | 15:54 |
kernling | could you paste the logs of this again? | 15:54 |
kernling | HYBRIS_TRACE=1 HYBRIS_LOGGING_LEVEL=debug EGL_PLATFORM=hwcomposer QT_QPA_PLATFORM=hwcomposer test_hwcomposer | 15:55 |
Mister_Magister | this big big file? | 15:56 |
kernling | ah right | 15:57 |
kernling | i still have it | 15:59 |
Mister_Magister | mal: kernling: tough problem? | 16:30 |
kernling | Mister_Magister, your device doesn't seem to like the wrappers | 16:42 |
kernling | it's working fine on mine | 16:42 |
kernling | so debugging is a bit stupider | 16:42 |
Mister_Magister | :D | 16:42 |
Mister_Magister | kernling: you are the only one working on this or are you collabing with mal? | 16:43 |
kernling | Mister_Magister, can you run gdb on the sigbus and do layout asm and then print the registers? | 16:57 |
Mister_Magister | eeeeeeee not sure how to do that | 16:57 |
kernling | Mister_Magister, HYBRIS_TRACE=1 HYBRIS_TRACE_UNHOOKED=1 HYBRIS_TRACE_HOOKED=1 EGL_PLATFORM=hwcomposer QT_QPA_PLATFORM=hwcomposer gdb test_hwcomposer | 16:59 |
kernling | then run | 16:59 |
kernling | then layout asm | 16:59 |
kernling | then print $0 | 16:59 |
kernling | then print $1 | 16:59 |
kernling | etc | 16:59 |
kernling | also paste the instruction it stopped at the the instructions after it | 17:00 |
Mister_Magister | i'm really confused | 17:01 |
Mister_Magister | https://youdieifyou.work/files/txsqaqqbchd.png | 17:01 |
kernling | ah :( | 17:16 |
kernling | Mister_Magister, and how about this | 17:20 |
kernling | Mister_Magister, use the previous command with gdb | 17:20 |
kernling | Mister_Magister, and then | 17:20 |
kernling | break trace_callback | 17:20 |
kernling | then run it | 17:20 |
Mister_Magister | you mean this command ? HYBRIS_TRACE=1 HYBRIS_LOGGING_LEVEL=debug EGL_PLATFORM=hwcomposer QT_QPA_PLATFORM=hwcomposer test_hwcomposer | 17:21 |
Mister_Magister | kernling: ^ | 17:21 |
kernling | Mister_Magister, HYBRIS_TRACE=1 HYBRIS_TRACE_UNHOOKED=1 HYBRIS_TRACE_HOOKED=1 EGL_PLATFORM=hwcomposer QT_QPA_PLATFORM=hwcomposer gdb test_hwcomposer | 17:22 |
kernling | this | 17:22 |
Mister_Magister | okay i'm at the breakpoint, what now | 17:22 |
kernling | Mister_Magister, can you paste your screen contents? | 17:23 |
Mister_Magister | https://paste.opensuse.org/87aa3e72 | 17:23 |
kernling | aha! | 17:24 |
kernling | what happens if you type step | 17:24 |
Mister_Magister | sigbus | 17:24 |
kernling | can you paste the screen contents again? | 17:25 |
Mister_Magister | https://paste.opensuse.org/95e61799 | 17:25 |
kernling | Mister_Magister, ok i will get mal to compile a new version | 17:27 |
Mister_Magister | so you make code and mal compiles it mhm mhm i see | 17:27 |
Mister_Magister | i mean, its my repo if you want simple build latest commit refresh i can do that | 17:28 |
kernling | Mister_Magister, mal is comminting the changes | 17:29 |
Mister_Magister | ah | 17:29 |
mal | Mister_Magister: test now | 17:41 |
Mister_Magister | ehh | 17:41 |
kernling | Mister_Magister, HYBRIS_TRACE=1 HYBRIS_TRACE_UNHOOKED=1 HYBRIS_TRACE_HOOKED=1 EGL_PLATFORM=hwcomposer QT_QPA_PLATFORM=hwcomposer test_hwcomposer | 17:42 |
kernling | Mister_Magister, use this command | 17:42 |
Mister_Magister | aaand bus error | 17:46 |
Mister_Magister | kernling: ^ | 17:48 |
mal | Mister_Magister: new libhybris available again | 20:04 |
Mister_Magister | again??? | 20:04 |
Mister_Magister | kernling: what logs u need? | 20:07 |
kernling | Mister_Magister, HYBRIS_TRACE_UNHOOKED=1 HYBRIS_TRACE_HOOKED=1 EGL_PLATFORM=hwcomposer QT_QPA_PLATFORM=hwcomposer test_hwcomposer | 20:08 |
Mister_Magister | bus error | 20:09 |
kernling | Mister_Magister, HYBRIS_LOGGING_LEVEL=debug HYBRIS_ANDROID_SDK_VERSION=18 HYBRIS_TRACE=1 HYBRIS_TRACE_HOOKED=1 HYBRIS_TRACE_UNHOOKED=1 HYBRIS_TRACE_DYNHOOKED=1 EGL_PLATFORM=hwcomposer test_hwcomposer | 20:11 |
Mister_Magister | oh it gonna be big log | 20:11 |
Mister_Magister | sdk version 18? u sure? | 20:11 |
kernling | shouldn't matter too much | 20:12 |
kernling | Mister_Magister, you can also use this HYBRIS_LOGGING_LEVEL=debug HYBRIS_TRACE=1 HYBRIS_TRACE_HOOKED=1 HYBRIS_TRACE_UNHOOKED=1 HYBRIS_TRACE_DYNHOOKED=1 EGL_PLATFORM=hwcomposer test_hwcomposer | 20:12 |
kernling | Mister_Magister, so lipstick isn't working? can you get me a strace from lipstick too? | 20:57 |
kernling | run it as nemo | 20:57 |
kernling | Mister_Magister, using usual options | 20:57 |
Mister_Magister | nemo is no more man | 20:58 |
kernling | defaultuser | 20:59 |
Mister_Magister | i know i'm just joking with you | 21:00 |
Mister_Magister | so serous | 21:00 |
Mister_Magister | no lipstick isn't working | 21:00 |
Mister_Magister | you want like EGL_PLATFORM=hwcomposer QT_QPA_PLATFORM=hwcomposer strace -f lipstick? | 21:00 |
kernling | i guess so | 21:01 |
kernling | i usually run it with -platform hwcomposer at the end but not sure if that matters | 21:01 |
kernling | also please use strace -f -s 1024 | 21:01 |
Mister_Magister | okay | 21:02 |
Mister_Magister | it segfaults anyway :D | 21:02 |
kernling | let it segfault away | 21:03 |
kimmoli | segway ? | 21:27 |
Mister_Magister | hm? | 21:38 |
kernling | pun | 21:38 |
kernling | Mister_Magister, i think it has built | 21:45 |
kernling | Mister_Magister, new libhybris | 21:46 |
Mister_Magister | indeed it did | 21:46 |
Mister_Magister | ohno it segfaults again | 21:48 |
kernling | can you give me some straces? | 21:49 |
Mister_Magister | sure | 21:49 |
kernling | Mister_Magister, mal hmm seems that didn't do what i wanted it to do | 21:55 |
Mister_Magister | isn't it like same with all the bugs | 21:55 |
T42 | Ljroc was added by: Ljroc | 21:59 |
T42 | <adampigg> @Ljroc [Ljroc was added by: adampigg], 👋 | 22:01 |
T42 | <Ljroc> Hey piggz | 22:02 |
T42 | <adampigg> dropping by via telegram these days? | 22:04 |
T42 | <adampigg> 'tis the future ;) | 22:04 |
T42 | <Ljroc> Nokius said someone was chasing me | 22:09 |
mal | Mister_Magister: new libhybris again | 22:41 |
Mister_Magister | aaaand segfault again | 22:43 |
mal | more important is the gralloc printing it hopefully did | 22:51 |
mal | Mister_Magister: did it print any messages about gralloc? something like "GRALLOC ALLOC->ALLOC" | 22:56 |
Mister_Magister | it did | 22:56 |
mal | show those | 22:56 |
Mister_Magister | https://paste.opensuse.org/02701b37 | 22:56 |
mal | kernling: ^ | 23:04 |
kernling | curious | 23:05 |
Mister_Magister | if we manage to fix this it gonna be pic | 23:06 |
Mister_Magister | epic | 23:06 |
Mister_Magister | couldn't we liek go to old base and test commit by commit to see which commit breaks it?/ | 23:06 |
kernling | i hope somebody already did that | 23:13 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!