Danct12_ | deathmist: i already do have it | 02:36 |
---|---|---|
Danct12_ | https://paste.ubuntu.com/p/VZjGNs4Fpd/ | 02:37 |
Danct12_ | still doesn't work though | 02:37 |
Danct12_ | binder spam still exists | 02:37 |
Danct12_ | anyway i think i should go back to hybris-15.1 | 02:45 |
Danct12_ | since i don't think this device tree works for me | 02:45 |
T42 | <alexisrev> I started earning the day I took the courage to invest in Bitcoin with Amos Bauer , now I'm glad I took that courage. I just got $5500 profits sent directly to my PayPal account. | 04:43 |
T42 | <alexisrev> https://t.me/joinchat/AAAAAEeKU6SC5UqGZZ6lUw | 04:43 |
T42 | 章翰雅 %lastname% was added by: 章翰雅 %lastname% | 06:15 |
Danct12_ | so i gave up with pie rebase, back with oreo branch | 06:43 |
Danct12_ | and now lipstick fails to start, here's the error in journalctl | 06:43 |
Danct12_ | https://paste.ubuntu.com/p/Q5K98JPdD2/ | 06:43 |
Danct12_ | test_hwc2 runs perfectly fine, i can see the spiral | 06:43 |
Danct12_ | looking at strace, it seems like its looking for libsdm-disp-vndapis.so | 06:57 |
Danct12_ | and they don't seem to exist anywhere | 06:58 |
Danct12_ | nope that's not the case, still crashing | 07:15 |
deathmist | Danct12_: I had that libsdm-disp-vndapis.so missing too, it's not an issue | 07:26 |
Danct12_ | hm | 07:27 |
Danct12_ | still dont know why it doesn't work | 07:28 |
Danct12_ | [F] unknown:0 - QPA-HWC: hwc_module->methods->open(hwc_module, HWC_HARDWARE_COMPOSER, &hwc_device) in create returned -22 | 07:28 |
Danct12_ | could it be qt5-qpa-hwcomposer-plugin isn't compiled for libhwc2? since test_hwc2 works fine | 07:29 |
deathmist | SELinux is still not disabled as per your full journal, show a new one if you aren't sure it's disabled, it definitly supports hwc2 since my phone uses hwc 2.1. perhaps SECURITY_SELINUX_BOOTPARAM_VALUE being 1 in your defconfig is causing issues, please set it to 0 | 07:30 |
Danct12_ | sadly i already went back to oreo branch, will come back sometime soon | 07:31 |
Danct12_ | services starts perfectly fine here, just not lipstick | 07:32 |
deathmist | oh I may know why it didn't work, you never installed stragglers droid-hal RPM did you on Pie? that had the mount units from stragger_files define | 07:33 |
Danct12_ | i did, i got /persist along with others | 07:33 |
Danct12_ | i checked the symlinks and they're symlinked to vendor and it's directories | 07:34 |
deathmist | hm, hard to guess then. show new journal from oreo base, can't help but guess if I don't see anything. you probably left some pie stuff in your trees which doesn't work for oreo or something | 07:35 |
Danct12_ | i rm -rf'd hadk folder, but i'll give you journalctl | 07:36 |
Danct12_ | deathmist: https://paste.ubuntu.com/p/rWb6jNppzh/ | 07:38 |
deathmist | very strange, I don't see anything obvious myself (though I may just have missed it). have you formatted /data when jumping across droid versions? I had issues with that before iirc | 07:57 |
Danct12_ | also how can ofc | 07:57 |
Danct12_ | * ofc i did | 07:58 |
Danct12_ | anyway how can i connect to wifi from the terminal? | 07:58 |
Danct12_ | connmanctl isn't available here | 07:58 |
Danct12_ | but it's still strange that test_hwc2 works, while not lipstick | 07:59 |
Danct12__ | figured it out | 08:07 |
Danct12__ | it was a permission problem, i can run minimer as root just fine | 08:08 |
Danct12__ | but if i run minimer as nemo it doesn't work | 08:08 |
Danct12__ | just fails with -22 | 08:08 |
Danct12__ | https://paste.ubuntu.com/p/8NjYPmnqZF/ | 08:28 |
Danct12__ | here's the strace | 08:28 |
Danct12__ | just chmod'd it and now it worked | 08:32 |
Danct12__ | good news, added the nemo user to graphics group worked | 08:55 |
Danct12__ | not sure if this is what you're supposed to do.. but i got to the ui | 08:56 |
deathmist | Danct12_: running "groups" as nemo gives me "nemo video users ssu timed oneshot system bluetooth graphics input audio camera media media_rw mtp inet", not sure why you wouldn't be in that group. you sure your dhd submodule back on 15.1 is on upgrade-3.2.1 branch? | 09:35 |
Danct12__ | deathmist: oh it's not, fixed that | 09:50 |
Danct12__ | rebuilding packages and rootfs now | 09:50 |
deathmist | the group handling was changed for master and releases >3.2.1 so that would make sense that it was broken then | 09:51 |
Danct12__ | deathmist: thanks, it works! | 11:12 |
deathmist | how does one disable NFC via a command? I'm just adding a temporary fix till the next release for the NFC toggle in settings not saving state across reboots thanks to droid groups being added after nfcd chowns during mic image creation etc | 12:05 |
Danct12__ | regarding codecs, does anyone know where to get free implementation of them? | 12:13 |
mal | deathmist: why didn't you add the usual chown hack to .ks before image build? | 12:42 |
mal | I used a similar one for ofono | 12:43 |
deathmist | oh I guess I should do that instead then | 12:43 |
mal | deathmist: I have this just below ### begin 60_ssu in .ks: chown -R radio:radio /var/lib/ofono | 12:44 |
mal | so maybe something like that but for different group and folder could work? | 12:45 |
deathmist | yeah I know "chown nfc:nfc /var/lib/nfcd" is the fix, can these hacks be removed on next release that the group handling has been changing for in dhd master? | 12:48 |
mal | maybe | 12:54 |
T42 | picolopoco was added by: picolopoco | 13:01 |
T42 | <picolopoco> hello | 13:01 |
T42 | Hello %lastname% was added by: Hello %lastname% | 13:29 |
Danct12__ | can anyone help me with video decoding? | 13:33 |
T42 | <elros34> what is your issue? Did you add gstreamer1.0-droid to patterns and package droidmedia and gst-droid? | 13:46 |
Danct12__ | videos just won't play, and yes, i do have gstreamer1.0-droid installed | 13:50 |
Danct12__ | minimediaservice is also running | 13:50 |
Danct12__ | and gst-droid is of course packaged, otherwise i won't be able to take pictures | 13:50 |
Danct12__ | actually, the video plays, but only audio | 13:50 |
deathmist | what types of videos specifically? I needed https://git.io/JvkdC to be able to play a bunch of them, I may still need more too | 13:51 |
Danct12__ | the video is black screen | 13:51 |
Danct12__ | mp4 | 13:51 |
T42 | <AnonGhost_3301> Any development on poco ? | 13:54 |
Danct12__ | if i run the gallery app under a terminal this error shows when a video is played | 13:57 |
Danct12__ | https://paste.ubuntu.com/p/F9SxzqZr4m/ | 13:57 |
deathmist | did you try the file I sent? I don't know much at all about decoding | 14:00 |
Danct12__ | i've tried | 14:00 |
deathmist | you'll have to restart too afaik | 14:00 |
Danct12__ | i rebooted the device to make sure | 14:00 |
Danct12__ | still black screen with audio playing | 14:01 |
T42 | <elros34> you should check logcat why video decoder can't be found | 14:09 |
Danct12__ | oh it did output something | 14:13 |
Danct12__ | https://paste.ubuntu.com/p/zNFFDyyYTt/ | 14:13 |
Danct12__ | looking at this doesn't seem to be useful at all | 14:15 |
Danct12__ | and appearing as video decoding works without problem | 14:15 |
Danct12__ | except for ANDR-PERF-RESOURCEQS: Failed to apply optimization | 14:16 |
T42 | <elros34> pasting whole logs is good habit | 14:18 |
Danct12__ | https://paste.ubuntu.com/p/QNcQN4YYsw/ | 14:28 |
T42 | <elros34> by full log I mean: Restart device, run logcat so it include early boot messages, then try to play video maybe with GST_DEBUG=droid*:5. BTW you can disable time_daemon which restarts in your logcat | 14:45 |
Danct12__ | here is full logcat | 14:54 |
Danct12__ | https://paste.ubuntu.com/p/mrM4sNXN8R/ | 14:54 |
Danct12__ | and here is the output from the gallery app https://paste.ubuntu.com/p/hShb7XNvb6/ | 14:56 |
T42 | <elros34> Danct12__ I wonder about 2 x "ServiceManager: Waiting for service sensorservice" but maybe it's normal at the begging. Do you have all mini services running? "pgrep -a mini" | 15:49 |
Danct12__ | yes, minimedia, miniaf, minisf is shown | 15:51 |
Danct12__ | and also sensors are working fine here, tested with csd test app | 15:52 |
T42 | <elros34> assuming you started video playback somewhere around 14:50:28 I really can't find any messages in logcat from video decoder. Have you tried to get general gstreamer debug log like GST_DEBUG=4 to see which decoder it wants to load? | 16:28 |
Danct12__ | https://paste.ubuntu.com/p/WRPVhgJ63z/ | 16:34 |
Danct12__ | just did | 16:34 |
Danct12__ | tried to load video/x-h264 | 16:35 |
Danct12__ | i have a feeling that gst-droid video decoder isn't being used at all | 16:36 |
Danct12__ | since out of curiousity i tried to gst-inspect the decoder | 16:39 |
Danct12__ | here's the output | 16:39 |
Danct12__ | https://paste.ubuntu.com/p/crZHtFp48x/ | 16:39 |
T42 | <elros34> looks like it missing sink capabilities | 16:41 |
Danct12__ | yeah, i believe that's the cause | 16:42 |
Danct12__ | which video decoding probably works just fine as we've seen logcat | 16:43 |
T42 | <elros34> If you don't have better idea you could try to downgrade droidmedia/gst-droid. Both contains some dynamic_codecs changes. In sink capabilities you must have video/x-h264 otherwise droidvdec wil not be used by playbin | 16:54 |
Danct12__ | maybe reverting this commit should fix? | 16:59 |
Danct12__ | https://github.com/sailfishos/gst-droid/commit/bb0e1aca705a0d611a641fa0b702fd25b57c4bcd | 16:59 |
T42 | <elros34> I have no idea, that is question to mal | 17:00 |
mal | what is the problem? | 17:18 |
mal | which android base is that? | 17:19 |
Danct12__ | hybris 15.1 | 17:19 |
mal | do you have media_codec*.xml files in /system/etc? if so are there symlinks in /etc/ pointing to those? | 17:20 |
Danct12__ | they all are linked to /vendor/etc, yes | 17:21 |
Danct12__ | here's all of them symlinked: https://paste.ubuntu.com/p/fmwQ96vRNH/ | 17:23 |
mal | Danct12__: remove file /home/nemo/.cache/gstreamer-1.0/registry.armv7hl.bin and then grab output of "GST_DEBUG=droid*:5 jolla-gallery" when you try to play some for example h.264 video | 17:35 |
Danct12__ | mal: https://paste.ubuntu.com/p/6VbJPv5cCj/ | 17:41 |
mal | Danct12__: now show logcat | 17:43 |
mal | Danct12__: you can see those "No hardware support found for" lines in the log which tells android side is not doing what it should be | 17:43 |
Danct12__ | https://paste.ubuntu.com/p/YkYH2jfgCK/ | 17:45 |
Danct12__ | mal: ^ | 17:45 |
mal | do those time stamps match the time you tried to playback videos | 17:46 |
mal | are you sure the log is not missing something earlier | 17:46 |
Danct12__ | i apparently pressed enter to make new lines before launching jolla-gallery | 17:47 |
Danct12__ | in logcat | 17:47 |
Danct12__ | so i dont think i missed anything earlier | 17:47 |
Danct12__ | and even then, the timestamp do match | 17:49 |
mal | just to be sure, show output "ls -l /system/etc/media*.xml" and "ls -l /vendor/etc/media*.xml" | 17:52 |
Danct12__ | mal: https://paste.ubuntu.com/p/CcFy5XczJ7/ | 17:53 |
T42 | ot_kurva was added by: ot_kurva | 17:58 |
mal | Danct12__: remove the /home/nemo/.cache/gstreamer-1.0/registry.armv7hl.bin again, then reboot and try playback and grab new logcat, make sure it's the whole logcat | 18:02 |
Danct12__ | mal: https://paste.ubuntu.com/p/MSP9FFS6mM/ | 18:07 |
mal | 01-19 18:04:18.233 2849 4958 E /vendor/bin/hw/android.hardware.media.omx@1.0-service: Failed to allocate omx component 'OMX.qcom.video.decoder.avc' err=InsufficientResources(0x80001000) | 18:10 |
mal | 01-19 18:04:18.234 2831 4278 W OmxInfoBuilder: Fail to add mime video/avc to codec OMX.qcom.video.decoder.avc | 18:10 |
mal | so something is wrong in android side | 18:10 |
Danct12__ | hmm | 18:15 |
Danct12__ | yeah i guess i have no more idea about that.. | 18:27 |
Danct12__ | maybe kernel config? | 18:27 |
mal | no, probable some android initialization issue, or some missing something, or firmware issue | 18:37 |
Danct12__ | wishing i could strace the service though | 18:47 |
T42 | <elros34> OMX.qcom.video.decoder.avc fails but OMX.google.h264.decoder looks like its initialized. About firmware: I guess you are using this strange kernel config with audit disabled like previously. How about CONFIG_FW_LOADER_USER_HELPER, do you have it enabled? | 19:17 |
Danct12__ | elros34 FW_LOADER_USER_HELPER_FALLBACK (which enables the above) is disabled | 19:25 |
Danct12__ | should i turn it back on? | 19:25 |
T42 | <elros34> enable both | 19:29 |
Danct12__ | elros34 done, booted | 19:48 |
Danct12__ | but anyway that doesn't matter since we have this patch | 19:50 |
Danct12__ | https://github.com/sailfish-santoni/android_kernel_xiaomi_msm8937/blob/106ad9dd90c96c8a728ba073795d8983531f1462/drivers/base/firmware_class.c#L308-L321 | 19:50 |
T42 | Cyril %lastname% was added by: Cyril %lastname% | 20:05 |
T42 | 匡慧梅 %lastname% was added by: 匡慧梅 %lastname% | 23:38 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!