| T42 | <k1gen> mal: vibration is max on brightness 1, but I can turn it off by writing 0 to brightness | 09:01 |
|---|---|---|
| T42 | <k1gen> should I replace activate with brightness in /usr/libexec/droid-hybris/system/etc/init/vibrator.rc and reboot? | 09:08 |
| dodoradio[m] | hi, I'm trying to debug gps on a casio wsd-f20. test_gps gets this far, then just hangs:... (full message at <https://matrix.org/_matrix/media/v3/download/matrix.org/HyyfKkEjzIiuIfIANjYETpjj>) | 10:33 |
| T42 | <edp_17> mal : Do you remember the missing audio issue on my moto G7 port? I've built sfos 4.5 and test audio with that. Audio is properly working if routed to: a BT device or jack. However, still no sound when using the built-in speaker. Also tried all options in "Audio Output" app, but no luck. Any idea? | 11:02 |
| mal | @edp_17 hmm, would need logs, sounds like some audio routing issue, have you checked the quirks? | 11:27 |
| T42 | <k1gen> mal: I think you missed my question: should I replace activate with brightness in /usr/libexec/droid-hybris/system/etc/init/vibrator.rc and reboot? | 12:04 |
| mal | @k1gen does the duration work correctly? | 12:15 |
| mal | @k1gen to support brightness you need to add this file and change the paths https://github.com/mlehtima/droid-config-fp4/blob/devel/sparse/usr/share/ngfd/plugins.d/60-droid-vibrator.ini | 12:16 |
| mal | and of course make sure the .rc changes correct permissions | 12:17 |
| mal | so your paths would be "native.path = /sys/class/leds/vibrator/duration" and "native.activate_path = /sys/class/leds/vibrator/brightness" | 12:17 |
| mal | I think that /sys/class/leds/vibrator linked to the actual path so you can use it like that | 12:18 |
| T42 | <k1gen> if duration is in miliseconds - then it doesn't work | 12:18 |
| mal | hmm | 12:18 |
| T42 | <k1gen> it's max again, but unlike with activator can be turned off | 12:19 |
| T42 | <k1gen> mal: https://vomitb.in/raw/e4SHqf5hpd | 12:24 |
| T42 | <k1gen> let me reboot and try again | 12:24 |
| T42 | <k1gen> vibro doesn't work at all now, wait for logs | 12:31 |
| T42 | <k1gen> mal: https://vomitb.in/raw/FmZmxvu6Le; journal: http://ix.io/4s2Z | 12:37 |
| T42 | <k1gen> sorry it took long | 12:37 |
| mal | permissions are clearly wrong | 12:38 |
| mal | a bit confused why the content of that looks so different from the other path | 12:38 |
| T42 | <k1gen> does it: https://vomitb.in/raw/XmmKk7Gkzd? | 12:40 |
| mal | in this paste it did https://vomitb.in/raw/e4SHqf5hpd | 12:41 |
| T42 | <k1gen> oh, yeah, weird | 12:42 |
| T42 | <edp_17> mal : what logs do you need? So far I've tried these quirks: https://paste.ubuntu.com/p/vC3cKGSsCM/ | 12:44 |
| T42 | <edp_17> mal : and currently I have this: https://paste.ubuntu.com/p/3nkvyD7pPt/ | 12:46 |
| mal | @k1gen so if the duration does not do anything then you could change the config file I mentioned to have ""native.path = /sys/class/leds/vibrator/brightness" and ""native.activate_path =" | 12:48 |
| T42 | <k1gen> should I remove duration from .rc too? | 12:49 |
| mal | no | 12:49 |
| mal | well you can but it doesn't matter | 12:49 |
| mal | @edp_17 which android base do you use? | 12:49 |
| mal | @edp_17 does that set_parameters quirk have any effect? logcat might be useful, also verbose output from pulseaudio | 12:50 |
| T42 | <edp_17> mal : Lineage 17.1. What effect I should expect from that set_parameters quirk? Now the device doesn't have audio, unless I connect a headphone or a bt speaker. Although for calls there is no sound in that case either, only music's sound is routed to bt or jack. (I'll get the logs.) | 13:02 |
| T42 | <edp_17> mal : logcat: https://paste.ubuntu.com/p/wQ7NvwkBGW/ | 13:12 |
| T42 | <edp_17> mal : Pulseaudio verbose log : https://paste.ubuntu.com/p/VqHqsNSvkM/ | 13:13 |
| mal | @edp_17 also grab output "pactl list" as defaultuser | 13:18 |
| T42 | <edp_17> mal : pactl list as defaultuser : https://paste.ubuntu.com/p/wQsKnnNthF/ | 13:21 |
| mal | @edp_17 did that pulse log contain some playback attempt? | 13:27 |
| T42 | <edp_17> Hmm, in the log there is a crash about wifi driver. I've noticed it yesterday, that out of the blue (after a reboot) the wifi was not available and I couldn't turn on. When was trying to load the module I got : "modprobe: ERROR: could not insert 'wlan': Operation not permitted". I had to restore a backup to fix it but I have no idea what's causing. | 13:27 |
| T42 | <edp_17> mal : no, I haven't tried to play anything. I'll get another log. | 13:28 |
| T42 | <edp_17> mal : pulseaudio log with playback : https://paste.ubuntu.com/p/rrZk9s4ztG/ | 13:29 |
| mal | it segfaults? | 13:30 |
| mal | ah, you exited manually | 13:30 |
| T42 | <edp_17> No. I ended with ctrl+c | 13:30 |
| T42 | <edp_17> Yes I did. | 13:30 |
| mal | also grab logcat which contains that playback attempt | 13:31 |
| T42 | <edp_17> here it is : https://paste.ubuntu.com/p/tDtmbyZMRW/ | 13:32 |
| T42 | <edp_17> Hmm, should I start pulseaudio before getting that logcat? | 13:32 |
| T42 | <edp_17> logcat with play attempt and running pulseaudio service: https://paste.ubuntu.com/p/PfnKpDYr3V/ | 13:34 |
| mal | I don't see anything obvious in those, needs some more checking | 13:42 |
| mal | does disabling deep_buffer change logs much? | 13:43 |
| T42 | <edp_17> Disabling deep_buffer where? | 14:13 |
| T42 | <edp_17> Yeah, last time we came to the same conclusion, there is nothing obvious in there. | 14:14 |
| mal | ok | 14:22 |
| T42 | <edp_17> mal : what else should I check/try? | 14:26 |
| mal | not sure | 14:29 |
| T42 | <TheVancedGamer> ``` | 14:34 |
| T42 | <TheVancedGamer> [ 0.784450] -(6)[1:swapper/0]Kernel panic - not syncing: VFS: Unable to mount root fs on unknown-block(1,0)``` | 14:34 |
| T42 | <TheVancedGamer> | 14:34 |
| T42 | <TheVancedGamer> getting this, installed cpio, unpacking hybris-boot.img reveals a gzip compressed initramfs, I did flash hybris-boot.img, not sure what fails | 14:34 |
| dodoradio[m] | Could anyone please help me with the GPS issue I posted above? | 16:53 |
| T42 | <k1gen> mal: can we leave the vibro for now and get back to sound? I tried different combinations of activate, brightness and duration, and it doesn't work properly: activate isn't even there anymore. you said we could probably get the sound working with more testing | 16:55 |
| T42 | <k1gen> dodoradio: my GPS works in CSD and Fernschreiber, but test_gps can't even find interface ¯\_(ツ)_/¯ | 16:57 |
| dodoradio[m] | Hmm, I'll have to do some more debugging then. I'm porting for asteroidos (also hybris based) and it's not working in any of our UIs either | 16:58 |
| dodoradio[m] | Does your device use hal or binder backend? I don't think test_gps supports binder at all | 16:59 |
| T42 | <k1gen> hal, I guess. I didn't touch GPS at all, it just worked straight up. I wish sound would work like that :P | 17:01 |
| dodoradio[m] | Has anyone seen GPS being initialised by an android service written in dalvik? I'm really hoping that's not the issue here | 17:04 |
| *** Daanct12 is now known as Danct12 | 18:04 | |
| T42 | <edp_17> @k1gen : what's the problem with sound? | 18:09 |
| T42 | <k1gen> @edp_17 there is no sound, lol | 18:09 |
| T42 | <edp_17> I have the same problem. Funny because sound worked out of the box on all my other ports, except this one. :) | 18:11 |
| T42 | <edp_17> Does it work when a BT is connected? And via jack? | 18:12 |
| T42 | <edp_17> Also, only music, calls or both has no sound? | 18:13 |
| T42 | <k1gen> no sound whatsoever. also, no jack for pixel owners | 18:15 |
| T42 | <k1gen> if you can help - check channel logs, I haven't moved on since mal helped me last time. still on that pulseaudio segfault | 18:16 |
| mal | @k1gen does bluetooth work on the device yet? what other things are broken? | 18:23 |
| mal | did the latest vibrator test work? | 18:23 |
| T42 | <k1gen> mal: bluetooth works just fine, iirc you helped me fix it. vibro doesn't work, but it's fine, I can live without it. what I can't live without is sound. if I get it working - I will be able to use the device normally | 18:25 |
| T42 | <k1gen> I don't have sound, camera, vibro, flashlight and sensors | 18:26 |
| mal | maybe sensors could be one thing to check now | 18:26 |
| mal | not sure what went wrong with the latest vibrator test, was there anything in logs from ngfd? | 18:26 |
| T42 | <edp_17> @k1gen : Have you tried play music via BT? | 18:27 |
| T42 | <k1gen> mal: couple thousand of ERROR: tonegen-ausrv: ausrv: server connection failure: Connection refused errors | 18:27 |
| mal | that is because of pulseaudio maybe | 18:28 |
| T42 | <k1gen> @edp_17 dude, pulseaudio segfaults | 18:28 |
| T42 | <edp_17> Oh, yes. I forgot that. | 18:28 |
| T42 | <k1gen> yeah, that's the sad part | 18:28 |
| mal | @k1gen did you test pulseaudio-modules-droid-jb2q yet, just to be sure | 18:29 |
| T42 | <k1gen> no, what should I do after installing them? | 18:29 |
| T42 | <k1gen> should I first delete pulseaudio-modules-droid-*? | 18:29 |
| mal | if you have any quirks enabled then those are defined in different format, other than no difference | 18:30 |
| mal | you can install the new ones and uninstall the old ones at the same time | 18:30 |
| mal | like this zypper in some.rpm -some_package | 18:30 |
| T42 | <k1gen> oh, nice | 18:31 |
| T42 | <k1gen> mal: I'm sorry, where's the -jb2q package should be? it's not in local repo, and not in remotes | 18:32 |
| mal | you need to build it, using build_packages.sh --mw=https://github.com/mer-hybris/pulseaudio-modules-droid-jb2q.git | 18:33 |
| T42 | <k1gen> got it, thanks | 18:33 |
| T42 | <k1gen> Problem: the installed patterns-sailfish-device-adaptation-flame-1-202303232045.aarch64 requires 'pulseaudio-modules-droid-hidl', but this requirement cannot be provided | 18:37 |
| T42 | <k1gen> deleted providers: pulseaudio-modules-droid-hidl-1.5.0-1.aarch64 | 18:37 |
| mal | don't remove that | 18:37 |
| T42 | <k1gen> The following 3 NEW packages are going to be installed: | 18:37 |
| T42 | <k1gen> pulseaudio-modules-droid-jb2q pulseaudio-modules-droid-jb2q-common pulseaudio-modules-droid-jb2q-devel | 18:37 |
| T42 | <k1gen> | 18:38 |
| T42 | <k1gen> The following package is going to be REMOVED: | 18:38 |
| T42 | <k1gen> pulseaudio-modules-droid | 18:38 |
| T42 | <k1gen> this looks fine? | 18:38 |
| mal | yes | 18:38 |
| mal | well no need to install devel package | 18:38 |
| mal | also uninstall the pulseaudio-modules-droid-common | 18:38 |
| T42 | <k1gen> The following 2 NEW packages are going to be installed: | 18:39 |
| T42 | <k1gen> pulseaudio-modules-droid-jb2q pulseaudio-modules-droid-jb2q-common | 18:39 |
| T42 | <k1gen> | 18:39 |
| T42 | <k1gen> The following 2 packages are going to be REMOVED: | 18:39 |
| T42 | <k1gen> pulseaudio-modules-droid pulseaudio-modules-droid-common | 18:39 |
| mal | yes | 18:39 |
| T42 | <k1gen> did that | 18:39 |
| T42 | <k1gen> what should I run after reboot? gdb, strace or just pulse? | 18:41 |
| mal | try to start pulseaudio first | 18:41 |
| T42 | <k1gen> failure. journal: http://ix.io/4s4M; stdout of pa: https://vomitb.in/raw/1DVdlLic0B | 18:46 |
| T42 | <k1gen> forgot the logcat... | 18:46 |
| mal | I told you that if you have quirks then you need to either remove those or change to the different format | 18:47 |
| mal | you have one quirk in the that custom .pa file | 18:47 |
| T42 | <k1gen> yeah, sorry, forgot about that. let me remove it | 18:47 |
| T42 | <k1gen> mal: https://vomitb.in/raw/RaATVpE3IM wihtout /etc/pulse/arm_droid_card_custom.pa | 18:49 |
| mal | so gdb of course if it crashes | 18:49 |
| T42 | <k1gen> mal: sure: https://vomitb.in/raw/ZjwROpKUQv | 18:50 |
| mal | so similar the way non-jb2q crashed | 19:14 |
| mal | did you ever get strace of pulseaudio? | 19:14 |
| mal | it will result in quite a lot of output though | 19:14 |
| T42 | <k1gen> no, never. I don't understand strace syntax. if only you could give me command which also can log to file... :) | 19:19 |
| T42 | <TheVancedGamer> `strace -s 128 <command here> &> log.txt` | 19:21 |
| T42 | <TheVancedGamer> works for me | 19:21 |
| T42 | <k1gen> mal: sorry, I searched in chat history and it appears I've ran strace on pa | 19:21 |
| T42 | <k1gen> let me get another one | 19:21 |
| T42 | <k1gen> HYBRIS_USE_VENDOR_NAMESPACE=yes strace -o outfile.txt -f -s2048 pulseaudio --daemonize=no -n --file=/etc/pulse/arm_droid_default.pa -vvvvvvvvvv like that? | 19:22 |
| T42 | <k1gen> any pastebin to curl 2.7M strace file to? | 19:25 |
| T42 | <TheVancedGamer> transfer.sh maybe | 19:26 |
| T42 | <k1gen> mal: https://transfer.sh/bapEbt/strace.txt | 19:27 |
| T42 | <k1gen> thanks @TheVancedGamer | 19:28 |
| T42 | <k1gen> mal: you still here? any ideas with strace I sent? | 19:58 |
| mal | not much ideas based on that, I do wonder if that last message before the crash is relevant or not | 20:13 |
| T42 | <k1gen> mal: well, I have 40 minutes to try anything you ask :D | 20:18 |
| T42 | <elros34> @dodoradio[m] have search in *rc files? If there is some service or initialization then it will be in rc. I can see some logs in logcat when starting test_gps, did you check it? | 20:22 |
| mal | @k1gen need to think and check the logcat more carefully | 20:30 |
| T42 | <elros34> could be related to wrong /linkerconfig? I remember he had some error about not accessible media library somewhere | 20:32 |
| T42 | <k1gen> mal: I haven't sent you logcat yet, have I? what do I have to while logcat is running? | 20:33 |
| T42 | <k1gen> *have to do | 20:33 |
| T42 | <elros34> you are debuging pulse audio so? get full logcat after rebot with marked time (in logs) when you start pulseaudio and when it crash. This way you increase the chance somebody else will really look in to your logs and find something | 20:35 |
| T42 | <k1gen> can I pipe logcat to file > like this? | 20:37 |
| mal | yes, this should work: logcat &> file | 20:40 |
| mal | don't remember if logcat prints to stdout or stderr so write both to file | 20:41 |
| T42 | <k1gen> mal: journalctl: http://ix.io/4s5e; logcat: http://ix.io/4s5f. I executed HYBRIS_USE_VENDOR_NAMESPACE=yes pulseaudio --daemonize=no -n --file=/etc/pulse/arm_droid_default.pa around 20:41-42 | 20:45 |
| mal | just to be sure, have you masked pulseaudio so it doesn't try to start, and you are sure it's not running | 20:49 |
| mal | @k1gen can you check where on device you have libaudioroute.so? | 20:51 |
| T42 | <k1gen> find / -name libaudioroute.so? | 20:51 |
| T42 | <k1gen> also, I'm about to send journal and logcat with pulse masked, sorry | 20:52 |
| T42 | <k1gen> /apex/com.android.vndk.v30/lib64/libaudioroute.so | 20:53 |
| T42 | <k1gen> mal ^ | 20:53 |
| T42 | <k1gen> there's also 32-bit one in ../lib | 20:53 |
| mal | ok, based on strace it did find that libaudioroute.so | 20:55 |
| T42 | <k1gen> I can see that in stdout: library "libaudioroute.so" not found | 20:55 |
| mal | yes, but the strace you sent before showed it getting loaded | 20:56 |
| T42 | <k1gen> oh, so what should I do? it can't find 6 libraries in total | 20:57 |
| mal | those are probably not critical | 21:00 |
| mal | @k1gen do you have any files in /data/vendor/audio/ | 21:04 |
| mal | or anything with extension .cal in /data | 21:05 |
| T42 | <k1gen> a minute, let me check | 21:06 |
| T42 | <k1gen> I'm also dealing with freezes on lockscreen password entering stage, so sometimes a reboot means 3-4 reboots | 21:08 |
| T42 | <k1gen> mal: https://vomitb.in/raw/anvrgJTrYF | 21:08 |
| T42 | <k1gen> do you still need journal & logcat with masked pa? | 21:09 |
| T42 | <k1gen> mal: masked pulseaudio: journal: http://ix.io/4s5o; logcat: http://ix.io/4s5p; started HYBRIS_USE_VENDOR_NAMESPACE=yes pulseaudio --daemonize=no -n --file=/etc/pulse/arm_droid_default.pa at ~00:10:35 | 21:12 |
| mal | @k1gen one thing to check one thing, grep for trigger_restart_framework in /vendor/etc/init and maybe also /system_ext/etc/init and /system/etc/init | 21:34 |
| T42 | <k1gen> mal: https://vomitb.in/raw/zFz9CnPqXi | 21:41 |
| T42 | <k1gen> sorry for cut text, sshing from iphone in bed | 21:41 |
| mal | that seems ok | 21:42 |
| dodoradio[m] | <T42> "<elros34> @dodoradio[m] have..." <- So all the references in the rc files are about setting permissions on the gps serial device | 21:52 |
| dodoradio[m] | There is a log from the gps library: it logs the firmware version of the gps module, which means that it is successfully communicating with it over that serial port (/dev/ttySAC1) (yes this thing is a casio smartwatch with two screens, an exynos soc and a custom gps module that runs freertos attached via uart) | 21:54 |
| dodoradio[m] | I have also seen it fail to get the firmware version from the gps module (eg. When permissions were wrong of the port driver was missing so there was no such device). It logs 'give up' and exits. Since I am currently getting a correct FWVER log in logcat, that means that I have currently got it set up for correct communication between library and uart port. | 21:57 |
| dodoradio[m] | Sorry, ask if clarification is needed. That last message is a masterpiece of word spaghetti | 21:57 |
| T42 | <elros34> I was expecting some errors in logcat, in sailfish I do not remember to use test_gps ever and this works automatically with geoclue plugin. Is this gps library even open source? | 22:01 |
| dodoradio[m] | The gps library is closed source. | 22:02 |
| T42 | <elros34> I assume comparing uart communication against working android is not an option. I wouldn't rely so much on test_gps. Have you tried geoclue hybris plugin? | 22:05 |
| dodoradio[m] | Comparing uart against android is really problematic. Firstly, the android on this thing decided to kill itself in a really strange way. I have gotten it booting back into android but it's very unstable. Additionally, I still haven't managed to write a patch for the samsung uart driver that sniffs all data correctly. My current patch logs the stream from the system console port correctly, but sees nothing on the gps port, | 22:09 |
| dodoradio[m] | despite the fact that there's obviously data going through. I'm a bit stumped. | 22:09 |
| dodoradio[m] | I have worked withe hybris plugin before. I was just trying test_gps because I was trying to eliminate all unknown variables. Currently rebuilding an image with a (hopefully working) geoclue-hybris plugin | 22:10 |
| T42 | <elros34> could it be it waits for signal to really give some callback?:) | 22:13 |
| dodoradio[m] | I'm messing with some geoclue-hybris patches written by one of our porters which should log any nmea messages. And nmea messages should be coming through even when the module is seeking. | 22:14 |
| dodoradio[m] | And I have gone outside (haha) a few times with this thing, and had no luck with gps. And there wasn't anything new in logcat afterwards either if I recall | 22:15 |
| T42 | <elros34> yeah I see them straightaway when starting test_gps | 22:15 |
| T42 | <elros34> quite offtopic but I have tried asteroid on some no name watch mt6580 long time ago but couldn't figure out how enabled uart. I got only few random lines probably from bootloader | 22:20 |
| dodoradio[m] | Most smartwatches don't seem to expose uart, at least not in any obvious way. The casio's system console port is somewhere inside, I haven't messed with it. | 22:22 |
| dodoradio[m] | In any case, the asteroid experience is significantly better on watches that originally ran wearos. Those mediateks are horrible. | 22:23 |
| T42 | <elros34> I managed to get it booting with but it was unstable/unusuable | 22:24 |
| dodoradio[m] | Hmm. I'll test geoclue-hybris again, and then, if it doesn't magically start working, I might try decompiling that library. | 22:30 |
| dodoradio[m] | If it was looking for some files or directories in the fs, it would show up in strace, so I'm betting that's not the issue. I do also have the gps module firmware, but I'm not sure what arch it's for. I could try attacking that as well, but that's really not preferable... | 22:31 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!