**** BEGIN LOGGING AT Thu Oct 01 02:59:57 2020 Oct 01 07:39:36 Seems there are some differences between Erfan's GSI device and the fork in our Halium repo. Mobile data doesn't work anymore compared to the v9 installer, on the Pixel 3a. I'm not sure what's missing though. ip tools etc are in there for sure. Oct 01 07:41:46 stub_netd is also running. Oct 01 07:44:59 @fredldotme [Seems there are some differences between Erfan's GSI device and the fork in our …], I've built them on my GSI but they were not needed to get mobile data so didn't include to final tree Oct 01 07:46:36 @erfanoabdi [I've built them on my GSI but they were not needed to get mobile data so didn't …], which ones were that? Oct 01 07:47:23 @fredldotme [Seems there are some differences between Erfan's GSI device and the fork in our …], @RealDanct12 does it work on Pro1? Oct 01 07:51:15 @fredldotme [which ones were that?], `netutils-wrapper-1.0 ndc ip6tables iptables` packges Oct 01 07:54:44 ndc is netd cli tools and it's unused Oct 01 07:55:08 Cool, thanks! Will try that with a sargo/bonito build which doesn't have working mobile data either. Oct 01 08:05:38 Yup, that was it, mobile data works now! Oct 01 08:05:47 wew Oct 01 08:06:30 PR please (@erfanoabdi should I still rebase as you asked?) Oct 01 08:07:00 @NotKit [PR please (@erfanoabdi should I still rebase as you asked?)], hybris-patches/halium.mk or generic device repo? Oct 01 08:07:27 better halium.mk I think for those Oct 01 08:07:40 just ammend my patch so we have easier time rebasing to 11 Oct 01 08:07:55 [Edit] just amend my patch so we have easier time rebasing to 11 Oct 01 08:09:03 sure! after work Oct 01 08:09:29 @fredldotme [sure! after work], thanks Oct 01 08:10:00 @NotKit [PR please (@erfanoabdi should I still rebase as you asked?)], yeah it's cleaner Oct 01 08:31:52 @NotKit [@RealDanct12 does it work on Pro1?], I'll check Oct 01 09:01:41 Also, mic is muffled in a big way on the receivers end. I can hear the other person fine though. Oct 01 09:06:08 Almost as if it used a wrong audio codec in the voicecall profile. Oct 01 09:07:51 [Edit] Almost as if it used a wrong audio codec in the voicecall profile. Silence is distinguishable from speaking. Oct 01 11:12:06 Tobias was added by: Tobias Oct 01 11:26:43 Hello together, is there any way to get ubuntu touch to Huawei Mate 10 pro? If yes, I would love to get some help on this... Oct 01 12:00:36 @Tobias [Hello together, is there any way to get ubuntu touch to Huawei Mate 10 pro? If y …], Is BL unlocked? Oct 01 12:00:49 What android version? Oct 01 12:01:06 Kernel sources are avalible? Oct 01 12:01:16 @Lukapanio [What android version?], 10 Oct 01 12:01:22 @Tobias [10], Ouch Oct 01 12:01:38 (Photo, 640x1280) https://irc.ubports.com/QNKwkrR1.png Oct 01 12:01:56 @Tobias [10], If there was no 9 or 7 then no way yet Oct 01 12:02:00 Script for porting halium to A10 mtk device Oct 01 12:02:07 (Sticker, 512x152) https://irc.ubports.com/01HbBSXl.webp Oct 01 12:02:32 …it was 9 yesterday, but I had to update for some reason… Oct 01 12:03:01 @Tobias […it was 9 yesterday, but I had to update for some reason…], Oh if there was 9 then ok Oct 01 12:03:13 Is it treble compatible? Oct 01 12:03:25 I think so, should be... Oct 01 12:03:46 I’m no expert in this, that’s why I’m asking Oct 01 12:16:15 @itexpert120 [Script for porting halium to A10 mtk device], no script, otherwise there would be no need for this group Oct 01 12:16:24 does it have kernel source code? Oct 01 12:46:53 @NotKit [does it have kernel source code?], How do I check? Oct 01 12:47:38 @Lukapanio [Is it treble compatible?], Yes Oct 01 12:47:44 that was response to @itexpert120's message. Huawei normally provides kernel source code according to GPLv2 Oct 01 12:48:30 @NotKit [that was response to @itexpert120's message. Huawei normally provides kernel sou …], I have Redmi Note 9 Oct 01 12:48:34 No kernel source as of now Oct 01 12:49:25 @itexpert120 [I have Redmi Note 9], that's too new, get an older device then Oct 01 12:49:49 there is no Android 10-based Halium tree yet Oct 01 12:49:59 @NotKit [that's too new, get an older device then], :/ Oct 01 12:58:05 @fredldotme [Almost as if it used a wrong audio codec in the voicecall profile. Silence is di …], Ok, so mic quality is okay when routed through the headphones. Oct 01 12:58:16 @fredldotme [Almost as if it used a wrong audio codec in the voicecall profile. Silence is di …], [Edit] Ok, so mic quality is okay when routed through the headphones/headset mic. Oct 01 13:18:28 I've been getting this error while building the boot image, could anyone help me out? Oct 01 13:18:28 https://del.dog/gingillere.txt Oct 01 13:24:20 @Lukapanio [Oh if there was 9 then ok], So you think it might be possible? … Is there anybody did it with some treble Huawei device? Oct 01 13:24:53 @Tobias [So you think it might be possible? … Is there anybody did it with some treble Huaw …], It must be possible Oct 01 13:25:08 Just need to flash android 9 Oct 01 13:25:16 Test gsi Oct 01 13:25:18 @Lukapanio [Just need to flash android 9], How? Oct 01 13:25:42 @Tobias [How?], It is device specific google is you're friend Oct 01 13:26:03 @Lukapanio [Test gsi], Android one, to make sure treble workes Oct 01 13:26:19 Compile kernel with needed options/patches Oct 01 13:26:31 Flash ubports gsi and new kernel Oct 01 13:26:42 @Sid127 [I've been getting this error while building the boot image, could anyone help me …], did it actually compile kernel? Oct 01 13:26:43 And it must work Oct 01 13:26:50 it seems something is wrong with device tree Oct 01 13:26:59 what is different from device you previously ported? Oct 01 13:27:11 [Edit] @Sid127 what is different from device you previously ported and booted? Oct 01 13:27:30 it's the same device, I'm just building using a different tree Oct 01 13:27:41 let me compare to check what's missing Oct 01 13:27:59 then diff BoardConfig.mk, yes Oct 01 13:28:19 I'm just modifying the gsi tree to allow it to build the boot image for my device(s) Oct 01 13:28:32 the same halium-boot is working on 3 devices by my oem Oct 01 13:28:57 ...why, there are lot of guides in the internet on how to compile Linux kernel out of tree Oct 01 13:29:12 because I also wish to build a systemimage Oct 01 13:29:27 overlays and stuff Oct 01 13:29:43 I do know how to compile kernel out of tree, I've done it before too Oct 01 13:29:46 I'd really suggest to do it the way we can integrate later on Oct 01 13:29:56 hold on Oct 01 13:30:17 it isn't very elaborate … https://github.com/ubports-nokia-sdm660 Oct 01 13:30:26 so if it works with GSI, preferred approach is to have device-specific tarball without systemimage Oct 01 13:30:43 [Edit] so if it works with GSI, preferred approach is to have device-specific tarball with overlays and boot.img but without systemimage Oct 01 13:31:00 huh, okay Oct 01 13:31:22 just to make it easier to integrate, right? Oct 01 13:31:40 got you, I'll switch to the tarball Oct 01 13:32:55 basically not to have CI server compile 1.5 hour systemimage for each specific device if it doesn't contain import changes from generic tree Oct 01 13:33:03 yup, got you Oct 01 13:33:03 [Edit] basically not to have CI server compile 1.5 hour systemimage for each specific device if it doesn't contain important changes from generic tree Oct 01 13:33:26 I'll use the fxtec tarball as reference and modify it as needed Oct 01 13:36:34 I'm not able to understand the overlay directory in that repo though Oct 01 13:37:07 I see there's udev rules, ofono configs, etc Oct 01 13:37:17 but the rest..? Oct 01 13:37:44 [Edit] but the rest..? and I'm not sure if I can straight up use the ofono config for my devices Oct 01 13:37:49 what do you mean as the rest? Oct 01 13:38:30 ubuntu-touch-session.d, modules-load.d Oct 01 13:38:41 I gather that they're pretty generic too Oct 01 13:39:23 modiles-load.d is specific for Pro1 to load Wi-Fi Oct 01 13:39:33 ubuntu-touch-session.d has configuration for scaling Oct 01 13:40:05 hm, I'm not sure how to adapt these... Oct 01 13:40:22 do not include what you do not need yet Oct 01 13:40:39 you will probably need scaling according to your device DPI though Oct 01 13:41:02 yup Oct 01 13:41:10 what about the ofono config Oct 01 13:42:21 @fredldotme [Seems there are some differences between Erfan's GSI device and the fork in our …], mobile data doesn't seem to work here too.. Oct 01 13:42:24 on fxtec pro1 Oct 01 13:43:10 @RealDanct12 [mobile data doesn't seem to work here too..], https://t.me/halium/163276 Oct 01 13:43:21 Will send a PR later today Oct 01 13:44:07 How is your microphone during calls? Does it sound distorted to the other call participant? Oct 01 13:44:17 audio doesn't work yet Oct 01 13:44:23 so can't tell Oct 01 13:44:35 adsp firmware is already loaded, just need to figure out why it doesn't work Oct 01 13:49:45 what's the display density of the fxtec pro 1? Oct 01 13:49:57 420? Oct 01 14:09:17 @NotKit figured out why audio crashed Oct 01 14:09:25 ```10-01 15:09:00.570 21635 21635 D vndksupport: Loading /vendor/lib64/hw/audio.primary.msm8998.so from current namespace instead of sphal namespace. … 10-01 15:09:00.575 21635 21635 E HAL : load: module=/vendor/lib64/hw/audio.primary.msm8998.so … 10-01 15:09:00.575 21635 21635 E HAL : dlopen failed:``` Oct 01 14:09:57 so why it failed? Oct 01 14:10:15 well it cut off there Oct 01 14:10:32 but we do know that the audio lib seems to fail Oct 01 14:10:55 start pulseaudio manually, it should print why Oct 01 14:16:11 https://paste.ubuntu.com/p/VZ5vRFFtXc/ Oct 01 14:18:17 "/android/vendor/lib64/libstdc++.so" is 32-bit instead of 64-bit - em what... Oct 01 14:24:18 lolwtf Oct 01 14:25:53 who's to blame? 🤔 Oct 01 14:26:44 try mount -o bind /system/lib64/libstdc++.so /vendor/lib64/libstdc++.so Oct 01 14:27:29 @NotKit [try mount -o bind /system/lib64/libstdc++.so /vendor/lib64/libstdc++.so], fwiw Android's mount totally ignores "-o bind", if source and dist are file, then it's bind :P Oct 01 14:27:54 @NotKit [try mount -o bind /system/lib64/libstdc++.so /vendor/lib64/libstdc++.so], and audio works! Oct 01 14:29:52 @phhusson [fwiw Android's mount totally ignores "-o bind", if source and dist are file, the …], that's neat, but I think coreutils mount still needs it Oct 01 14:30:09 @NotKit [that's neat, but I think coreutils mount still needs it], yup yup, but i assumed the command was in android lxc? Oct 01 14:31:51 no, it was (mistakenly?) loaded by PulseAudio through libhybris on the "other side" Oct 01 14:32:23 oh ok Oct 01 14:32:26 but I wonder why vendor partition has 32-bit library in lib64 Oct 01 14:32:46 I'd bet user modded vendr to boot my gsi Oct 01 14:32:57 (i have a longstanding issue that i don't have libstdc++ in vndk-spà Oct 01 14:32:58 [Edit] (i have a longstanding issue that i don't have libstdc++ in vndk-sp) Oct 01 14:33:03 it's stock Pro1 vendor :) Oct 01 14:33:08 what Oct 01 14:33:23 i never actually touched the vendor partition when flashing ut Oct 01 14:33:39 flashed boot, system then wiped userdata Oct 01 14:33:40 i don't have mine handy, i'll look Oct 01 14:34:16 @NotKit [it's stock Pro1 vendor :)], speaking of pro1 vendor, have the vendor tree updated? Oct 01 14:34:46 yes, it should have been, especially if you updated Android with OTA before Oct 01 14:34:50 ah Oct 01 14:35:04 but there are fastboot images if needed Oct 01 14:35:05 then i haven't done any OTA yet Oct 01 14:35:21 the vendor image was pretty much RTM Oct 01 14:47:45 @NotKit just flashed the latest vendor.. Oct 01 14:47:51 still same thing Oct 01 14:48:29 you can probably put it to /usr/sbin/mount-android.sh after mounting vendor and include that in overlay Oct 01 15:06:47 @NotKit does it make sense to globally disable the `persist.audio.fluence.voicecall` property for all devices using the generic GSI device? Or do we want some sort of configuration for that, via a script for example? Oct 01 15:06:59 setting it to `false` fixed the audio issues during calls. Oct 01 15:07:32 is that QCOM-specific? Oct 01 15:08:14 I've seen that property on multiple qcom devices, so yeah, probably Oct 01 15:11:16 OR, I just add an upstart session script which sets the property before pulseaudio starts. Oct 01 15:12:22 Pulseaudio should be 32 … /Runz Oct 01 15:13:06 @fredldotme [@NotKit does it make sense to globally disable the persist.audio.fluence.voiceca …], Make sense to disable it for all Oct 01 15:13:44 Ok, then I'll just add it in on-post-fs and go on with my life then Oct 01 15:14:30 @fredldotme [Ok, then I'll just add it in on-post-fs and go on with my life then], I'd say add it to system/prop.default Oct 01 15:14:40 @fredldotme [Ok, then I'll just add it in on-post-fs and go on with my life then], [Edit] I'd say add it to system/etc/prop.default Oct 01 15:14:56 @erfanoabdi [I'd say add it to system/etc/prop.default], Is that surviving overrides from the vendor build.prop? Oct 01 15:15:17 or am I backwards with my thinking what overrides what? Oct 01 15:15:41 Persists are read-only props iirc Oct 01 15:16:01 well, it survived the reboot to test it Oct 01 15:20:51 via setprop, that is Oct 01 19:06:44 Another thing I'd like to change for the generic device: remove the fingerprint "arrow keys" hack that allows you to scroll through lists and grids but also wakes up the device when touched slightly. Fine for every device? Oct 01 19:09:14 Also: MTP script for every device via halium-overrides or provide it as a sort of overrides "template" for the device tarball? Oct 01 19:10:50 probably template, there can be more tricky ones Oct 01 19:11:43 I wouldn't like it by default because it would complicate even more life for people debugging boot Oct 01 19:12:03 Fine with me, just expected every device to be ConfigFS now, even the 4.4 MTK ones. Oct 01 19:13:02 Volla is, yes Oct 01 20:04:19 @fredldotme [Fine with me, just expected every device to be ConfigFS now, even the 4.4 MTK on …], out of curiosity, how would one know if USB uses configfs? Oct 01 20:10:22 @Shakendo [out of curiosity, how would one know if USB uses configfs?], Most probably it uses ConfigFS when `/sys/devices/virtual/android_usb/android0/enable` does not exist Oct 01 20:13:30 @fredldotme [Most probably it uses ConfigFS when /sys/devices/virtual/android_usb/android0/en …], ok, what if it does exist? Oct 01 20:16:06 Then it uses the old, Android specific gadget driver: https://osmocom.org/projects/quectel-modems/wiki/Android_USB_Gadget Oct 01 20:16:48 *function driver Oct 01 20:17:21 ok, is it possible to switch to ConfigFS? Oct 01 20:18:25 Don't know, depends on whether your kernel even has ConfigFS, whether USB setup is too tightly integrated into the old driver, etc. Oct 01 20:20:33 ok, I actually didnt know that, so thank you, the link will probably help me set a RNDIS MAC address and maybe even get mtp and adb working correctly Oct 01 21:55:39 I have a question, which one would be the persist partition? PERSDATA or PERSISTENT? Oct 01 22:49:03 persistent if I had to guess Oct 01 23:41:36 Got in Oct 01 23:41:37 * Got it **** ENDING LOGGING AT Fri Oct 02 02:59:57 2020