Saturday, 2022-06-11

UmeaboyThaodan: Without --fetch-submodules I got this in the end:00:06
Umeaboyrepo sync has finished successfully.00:06
UmeaboySo something is wrong with the flag to fetch the submodules.00:07
ThaodanThat sounds nice, as long as none of the package are actually using code from the submodules00:07
ThaodanThe issue is that some of the packages have submodules with wrong commit refs.00:07
UmeaboyFor some reason I only have v2.22 of the repo binary.00:10
UmeaboyI thought 2.26 is the latest.00:10
UmeaboyAt least it'00:10
Umeaboys the latest version from Google.00:11
UmeaboyDo I really need the vendor repo? I thought they only contained non essential stuff that Sailfish OS doesn't use.00:18
UmeaboyHmmmmmmm. fatal: duplicate path device/samsung/beyond2lte in /home/kristoffer/hadk/.repo/manifest.xml00:19
UmeaboyI have never created it or changed the content of it.00:20
UmeaboyNo matter how I try I just can't get passed the vendor/samsung/exynos9820-common/ issue.00:28
UmeaboySomething isn't correct and I can't see where the error is from my side.-00:28
UmeaboyI see what it says, but adding that repo to $DEVICE.xml didn't help.00:28
ThaodanUmeaboy: did you extract blobs for your device? I'd first try to build lineage and then try to build hybris-hal00:55
UmeaboyThaodan: Building previous versions of LOS has never been an issue.00:56
UmeaboyBut I'll try.00:56
*** Daaanct12 is now known as Daanct1203:36
UmeaboyThaodan: Whenever you read this....... It fails to fully sync Lineage 18.1 as well.05:49
UmeaboyAnd I even did a fresh installation of Ubuntu 22.10 Kinetic (Daily build from yesterday)05:50
UmeaboyNo other issue so far so I suspect that either the repo binary is the problem or git itself.05:51
UmeaboyI could be wrong though.05:51
UmeaboyIf so, I'm sorry.05:51
UmeaboySeems like the repo binary is unreliable.06:31
UmeaboyThe second time the sync finished successfully.06:31
UmeaboyBuilding LOS as we speak. I'll be back later..... Have to sleep.08:59
T42<elros34> you do not need to do this09:00
T42<Verevka86> Revert this commit, build droidmedia ok πŸ‘ (re @SailfishFreenodeIRCBridgeBot: <Thaodan>https://git...)09:08
ThaodanVerevka86: Good to hear, I hope it fixed your issue.11:19
T42<Verevka86> 11:48
T42<Verevka86> Video recording not work πŸ˜” (re @SailfishFreenodeIRCBridgeBot: <Thaodan>Verevka86: ...)11:48
malcan you show logcat (make sure you remove gst cache first)?11:56
T42<Verevka86> (re @SailfishFreenodeIRCBridgeBot: <mal>can you show lo...)12:15
malI see messages "ServiceManager: Waiting for service 'media.audio_policy' on '/dev/binder'..." which suggests you don't have audio policy service enabled in droidmedia, which android base is that?12:25
T42<Verevka86> Android 1112:27
T42<Verevka86> (re @SailfishFreenodeIRCBridgeBot: <mal>I see messages ...)12:27
malso do you have that?12:27
T42<Verevka86> Yes, miniaudiopolicy bin Ρ‚ΠΎΠΆΠ΅ Π΅ΡΡ‚ΡŒ Π² систСмС12:28
T42<Verevka86> Yes, miniaudiopolicy bin is also in the system (edited)12:29
malis it running?12:29
T42<Verevka86> how can i check this?12:31
malps aux | grep mini12:31
T42<Verevka86> only minimediaservice and minisfservice (re @SailfishFreenodeIRCBridgeBot: <mal>ps aux | grep m...)12:37
piggzThaodan: im not doing it as a diff to the source, running the build with12:39
piggzi mean12:39
piggz%define kcflags "KCFLAGS=-Wno-misleading-indentation -Wno-format -Wno-bool-operation"12:39
piggzmake -C %{source_directory} O=out %{kcflags} defconfig KBUILD_DEFCONFIG=$defconfig_name12:39
mal@Verevka86 so why is the miniaudiopolicy not running, are you missing the entry in some .rc file or what?12:50
malor is starting it failing12:50
malI see nothing about that in logcat so probably not enabled to start12:50
malcan you see miniaudiopolicy.rc anywhere on the device?12:51
T42<Verevka86> I don't find it, maybe it doesn't start?12:53
maltry adding it12:54
T42<adampigg> +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++12:55
maladd this file to /usr/libexec/droid-hybris/system/etc/init/12:55
T42<TheVancedGamer> um, what12:56
T42<Verevka86> miniaudiopolicy start13:10
maldoes video recording work now?13:30
T42<Verevka86> No, the picture slows down when recording, and at the end of the recording, the camera freezes and prompts you to complete the process (re @SailfishFreenodeIRCBridgeBot: <mal>does video reco...)13:35
ThaodanYou can if it is active by running dumpsys13:35
Thaodanfrom android13:35
malthere seems to be some permission failures from servicemanager related to camera13:36
Thaodan /system/bin/dumpsys13:36
T42<Verevka86> but now the camera can start again after an error, without miniaudiopolicy the camera would not start again until the device was rebooted13:36
malbut those permissions issues might not be critical13:37
Thaodanthey aren't in SELinux permissive mode13:38
T42<Verevka86> 9 mb file, can I send it as a file? (re @SailfishFreenodeIRCBridgeBot: <Thaodan> /system/bi...)13:45
piggzThaodan: now added  -Wno-unused-variable into the mix13:50
T42<kayveepr> Please help - I am stumped by the following error:14:52
T42<kayveepr> d-pre-user-session[3443]: [F] unknown:0 - QPA-HWC: hwc_module->methods->open(hwc_module, HWC_HARDWARE_COMPOSER, &hwc_device) in create returned -2214:52
T42<kayveepr> Please help. I am stumped by the following error:-14:52
T42<kayveepr> Please help. I have no clue about the following error. This apparently seems to result in no GUI. test_hwcomposer works fine with a running video pattern until test exit. The error as per journal is   -  d-pre-user-session[3443]: [F] unknown:0 - QPA-HWC: hwc_module->methods->open(hwc_module, HWC_HARDWARE_COMPOSER, &hwc_device) in create returned -22. Seems to be called from hwcomposer_backend.cpp. GUI was working until May 25th on 414:59
T42<kayveepr> Sorry but that was not external link but from journalctl output. Never mind.15:00
T42<kayveepr> Lipstick fails with the following error resulting in no GUI. QPA_HWC: hwc_module->methods->open(hwc_module, HWC_HARDWARE_COMPOSER, &hwc_device) in create returned -22. test_hwcompser works fine with correct width/height on OpenGL ES 3.2 V@415.0 (GIT@7331a27, Ieeb4a86f7b, 1573670678) (Date:11/13/19). I have no clue about the above error which seems to be from hwcomposer_backend. Please help and thanks in advance15:11
malwhich android base?15:22
T42<kayveepr> Lipstick fails with the following error resulting in no GUI. QPA_HWC: hwc_module->methods->open(hwc_module, HWC_HARDWARE_COMPOSER, &hwc_device) in create returned -22. test_hwcomposer works fine with correct width/height on OpenGL ES 3.2 V@415.0 (GIT@7331a27, Ieeb4a86f7b, 1573670678) (Date:11/13/19). I have no clue about the above error which seems to be from hwcomposer_backend. Please help and thanks in advance (edited)15:27
malwhich android version did you use as base, the hybris-* branch15:30
T42<kayveepr> 10. I'm not sure if the above query was addressed to me. Sorry if it's not.15:39
Thaodan@kayveepr: Android 10 AOSP?15:42
T42<kayveepr> No, Lineage OS.15:42
T42<kayveepr> This started after I made some changes to see in config to see why wlan(qca_cld) was having limited connectivity though MAC address was seen by router and listed under ifconfig. No changes were made to kernel. I still have the tarball(May 25) which works the same way as it did today too.15:46
T42<elros34> did you see any other errors (permissions) in logcat?15:47
T42<kayveepr> No. I followed faq and most importantly followed many devices how it's done. Still had to use execstart.conf and all services loaded/active/running or exited as the case may be.15:49
T42<kayveepr> Got GUI on on May 25 after which it's failing on and
T42<elros34> execstart.conf?15:50
T42<kayveepr> Yes for tmpfiles service15:51
T42<kayveepr> Else it fails15:51
T42<elros34> not sure what excatly do you have in mind, did you enable all cgroup config recommended by mer-kernel-check?15:52
T42<elros34> what permissions do you have in /dev/*?15:53
T42<kayveepr> Android kernel configuration 3.18.140. does res_ctrl settings really matter? However have tried with cg all controllers enabled and disabled with no success. I'll post the logs once I'm allowed to post external links...! No idea if I can post pastebin links here without them getting deleted.15:58
T42<elros34> So is the systemd-tmpfiles-setup.service which fails for you?16:00
T42<kayveepr> It doesn't fail if I use the /sys as mentioned here earlier for exclude_prefix. Else with default option tmpfiles service fails16:02
T42<kayveepr> Like exclude_prefix=/sys/fs/cgroup/memory16:03
T42<kayveepr> Like exclude_16:04
T42<kayveepr> -prefix=/sys/fs/cgroup/memory (edited)16:04
T42<kayveepr> Like exclude16:04
T42<kayveepr> -prefix=/sys/fs/cgroup/memory (edited)16:04
ThaodanLogs or nothing happened16:04
T42<elros34> so usually this mean you have no cgroup controllers enabled? MEMCG_SWAP/KMEM16:04
T42<kayveepr> Tried with both. I'll present the logs soon.16:05
ThaodanI wonder if such kernels will even work with Cgroups v216:05
T42<elros34> Were they really enabled? but this is quite irrelevant for your current issue16:05
ThaodanSailfishOS currently boots in hybrid/legacy mode but for restricting users properly unified is needed, I'm just thinking out loud because 3.18 is really old.16:07
T42<kayveepr> You are right. They were not when I had the GUI anyway. Probably as mentioned cg v2 may be required. Besides I've only custom ROMs to check. Device: LeMax 2 - msm8996. Last mfr fw M. Year 2016. Could get UB ports going except for camera. Wanted to try this.16:09
T42<elros34> like previously check permissions, ueventd*.rc -> udev rules generation, logcat, params in droid-hal-device.conf16:11
ThaodanThere's halium for that device maybe they faced similar issues as you:
Thaodanah lol16:11
Thaodanthat's you16:11
T42<kayveepr> That Halium is from me only16:12
T42<kayveepr> That was on GSI. I moved to UBP CI where I could get it to work except for cam. I was working on it when I thought of SFOS go too.16:13
T42<kayveepr> Anyway I'll post all info so that someone can dig up hwc error -22. You guys are doing wonderful job but I'm very slow compared to you guys. Thank you all.16:16
T42<b100dian> I have the same issue on my xiaomi mi note 10. Also, miniaudiopolicy only helps with 'dont need to reboot'. I really hope you find the cause:) fingerscrossed (re @Verevka86: No, the picture slow...)20:53
T42<b100dian> My intuition was that either someone is not consuming the frames, so the slowdown is because more memory is needed for the new ones. Or that the sound stream is 'wrong' at 48khz (or 44?) and is locking up the in sound source. I dont have proof at hand and these were not fruitful intuitions for me. Any device reported as acces denied from dev/pcm or snd?20:56
T42<b100dian> Search the logs for to find some of mines, largish too to upload20:57
T42<b100dian> @Verevka86 ^^20:58
T42<Verevka86> 403 (re @b100dian: Search the logs for ...)21:11
T42<b100dian> for example21:13
Thaodanb100dian: Try the new miniaudiopolicy it should work better21:33
Thaodanthe old one spawned a full version of audiopolicy21:33
T42<Gwise> Forwarded from VE: Please keep up the good work , It was mind blowing when I opened my wallet and I saw the profits of $8,400 from the investment of $1000. Join and contact admin now @scott_Alvin123:20
T42<Gwise> πŸ‘‰πŸ‘‰πŸ‘‰πŸ‘‰βœβœβœβœ23:20

Generated by 2.17.1 by Marius Gedminas - find it at!