budrz | Hi, I am trying to build Sailfish OS for Google Pixel 3a XL and I got stuck in section 7.2.1 of the HADK porting guide | 16:52 |
---|---|---|
budrz | My question was posted here: https://forum.sailfishos.org/t/trying-to-adapt-google-pixel-3a-xl-but-netfilter-xt-match-qtaguid-module-is-missing-from-kernel/13070 | 16:52 |
mal | budrz: there is no support for lineage 19.1 base yet, did you use 18.1 hybris manifest for that or what? | 17:51 |
budrz | mal: no I more or less followed this in conjunction with the guide: https://wiki.lineageos.org/devices/bonito/build | 18:37 |
budrz | essentially their default manifest | 18:38 |
T42 | <elros34> so what url you used in 5.1 point of HADK: repo init? | 18:44 |
budrz | I used https://github.com/LineageOS/android.git | 18:46 |
T42 | <elros34> you must use mer-hybris/android | 18:47 |
deathmist1 | budrz: if you didn't know SailfishOS is NOT an Android ROM ;) | 18:47 |
budrz | T42: which branch would you recommend? 18.1? | 18:49 |
T42 | <elros34> it's latest supported | 18:50 |
budrz | deathmist1: I know, I thought I could at least use it to build the kernel though | 18:50 |
budrz | T42: OK. I will try that | 18:51 |
T42 | <elros34> probably mal have WIP hybris-19 in some personal repo:) | 18:51 |
budrz | that would be nice, I guess I have to start over and try again though either way | 18:57 |
budrz | I didn't know I needed to use mer-hybris/android | 18:57 |
T42 | <elros34> like in instruction, read also https://etherpad.wikimedia.org/p/hadk-hot and github.com/mer-hybris/hadk-faq | 18:59 |
budrz | T42: thanks for the links. I'll take a look at them | 19:04 |
budrz | I appreciate everyone's help. thanks! | 19:06 |
deathmist1 | budrz: are you even following https://sailfishos.org/develop/hadk/? I don't know how you would come to the conclusion to sync and build LineageOS source instead of mer-hybris/android if you were reading HADK | 20:24 |
T42 | <elros34> Anybody had such an issue that writting/reading /sys/class/leds/led_w/ cause kernel panic. led_r/g/b works at the same time and all are from same driver | 22:16 |
mal | @elros34 do you need to use that led? | 23:00 |
T42 | <elros34> probably no, for now I commented out line which writes to that path in some init*.rc file. Just trying to understand this issue | 23:01 |
mal | ah, so .rc was causing that, strange, have you checked the kernel driver if there is anything obvious there? | 23:01 |
T42 | <elros34> so far I couldn't found anything, for some reason dev_get_drvdata returns null: https://github.com/Exynos7420/android_kernel_samsung_exynos7420/blob/nx-10.0/drivers/leds/led-class.c#L64 | 23:03 |
T42 | <elros34> and it must have worked in android... | 23:04 |
T42 | <elros34> this kernel is disaster, it prints all pr_info messages no matter what log level I set, even logcat contains kernel messages | 23:07 |
mal | well you could just patch it to check the return value and return 0 or empty string or something | 23:09 |
mal | still strange though | 23:10 |
T42 | <elros34> yes probably but not just this function, whole led-class | 23:11 |
mal | not sure how that can return null | 23:16 |
T42 | <elros34> btw figured out flattened apexes issue. Do you think this kind of patch could be included to hybris-patches? https://paste.opensuse.org/02fb752e | 23:19 |
mal | probably, I should test that on fp3 to see if it solves the issue I have | 23:20 |
T42 | <elros34> basically it should help if boringssl_self_test fails or some other services with reboot_on_failure option which cause uevent start and usually device reboot | 23:25 |
mal | yeah, I had issues that the apex stuff was not initialized properly and thing broke terribly | 23:26 |
mal | quite sure those tests were failing | 23:27 |
T42 | <elros34> there is still something wrong because I had to symlink many libs from /apex to /odm but that is minor issue | 23:27 |
mal | yeah, that is a known issue, for which processes did you need those? | 23:28 |
T42 | <elros34> usual libandroidicu but also few more for some android services which I am not sure I really need so just write script which symlinked everything:) | 23:30 |
mal | I have one idea for libhybris to fix some linker issues, I need to test it one some device with such issues | 23:30 |
mal | *on some device | 23:30 |
mal | fp3 had those issues as well | 23:30 |
mal | so I just need to rebuild it to test | 23:30 |
T42 | <elros34> I can test (probably not today) if you have something | 23:30 |
mal | I have only half of the fix so far, need to make the other half | 23:32 |
T42 | <elros34> ok | 23:32 |
mal | can you strace one of the problematic processes, I think camera was one, check which ld.config*.txt it tries to load and from where | 23:34 |
mal | or was it minimediaservice, something like that | 23:34 |
T42 | <elros34> looks like /system/etc/ld.config.arm.txt, /linkerconfig/ld.config.txt and /system/etc/ld.config.txt | 23:37 |
T42 | <elros34> I have some ldconfig but in different namespace: /linkerconfig/bootstrap/ and /linkerconfig/default/ | 23:38 |
mal | is any of those files there? | 23:38 |
mal | linkerconfig probably is | 23:38 |
T42 | <elros34> no | 23:38 |
mal | or is that also missing | 23:38 |
mal | hmm | 23:38 |
T42 | <elros34> only /linkerconfig/default/ld.config.txt and /linkerconfig/bootstrap/ld.config.txt | 23:39 |
mal | so it could still be missing something from init | 23:39 |
mal | I remember investigating that issue | 23:39 |
mal | I have a fix to make to load ld.config.txt from linkerconfig but not sure if it would help because the file is in the subfolders | 23:40 |
mal | @elros34 https://github.com/LineageOS/android_system_core/blob/lineage-18.1/init/mount_namespace.cpp#L189 | 23:50 |
mal | so if you figure out how to make that happen then my libhybris fix should help | 23:51 |
mal | wait, but you said it was already looking from linkerconfig so then just fixing init should help | 23:53 |
mal | at least for some processes | 23:53 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!