| Mister_Magister | mal: i'll try to cook something up this weekend | 13:51 |
|---|---|---|
| Mister_Magister | with the gstreamer thing | 13:51 |
| T42 | <Гектор> hey people can some explain to me what should i do in here all i understand is that we do want to recognize how the device is partitioned | 14:24 |
| T42 | <Гектор> im pretty much new into this | 14:24 |
| T42 | <Гектор> i wanna make a port for poco x3 nfc (surya) | 14:24 |
| T42 | <Гектор> https://hadk.sailfishos.org/android/#configure-mountpoint-information | 14:24 |
| mal | there appears to be existing but abandoned port for that so maybe you could use that as base and update needed things https://gitlab.com/groups/xiaomi-poco-x3/-/inactive | 14:26 |
| mal | the misc repo appears to have some instructions | 14:27 |
| mal | https://gitlab.com/xiaomi-poco-x3/misc | 14:27 |
| T42 | <Гектор> ive seen it but i'd like to start cleanly and learn it | 14:28 |
| T42 | <Гектор> ive seen it but i'd like to start clean and learn it (edited) | 14:29 |
| mal | to determine the mount point information you need to check first in device repo what kind of paths are used in fstab file, it should be somewhere under device/$VENDOR/, then on the device check which actual devices match the things you see in fstab | 14:31 |
| T42 | <Гектор> Is this what are you talking about ? | 14:46 |
| T42 | <Гектор> https://github.com/GloboPatro/android_device_xiaomi_surya/blob/lineage-18.1/rootdir/etc/fstab.qcom | 14:46 |
| Mister_Magister | yes that's the fstab | 14:48 |
| Mister_Magister | and now check on the device, using command "mount" what devices match | 14:48 |
| T42 | <nc1x72> Mal, I have rebuilt new image (5.0.0.77) with ngfd binder plugin. Got no vibration at all. Nothing meaningful comes up in dmesg, journalctl or logcat either. | 19:39 |
| T42 | <nc1x72> Since ngfd can work with kernel API, maybe what I need is the correct sysfs path and maybe use the native vibrator plugin instead | 19:42 |
| mal | and you have the binder plugin installed? | 19:44 |
| T42 | <nc1x72> Yes | 19:44 |
| mal | and the vibrator binder service is running? | 19:44 |
| T42 | <nc1x72> How to check? | 19:45 |
| mal | check binder-list -d /dev/binder and binder-list -d /dev/hwbinder, which one lists it? | 19:45 |
| T42 | <nc1x72> Ok a sec | 19:45 |
| mal | no need to show whole output, just check which has something related to vibrator | 19:46 |
| T42 | <nc1x72> Hmmm, I don't see anything regarding vibrator or ngfd in binders list indeed | 19:56 |
| T42 | <Гектор> mal, thats all i could find idk what else should i do to find them all 🥲 | 20:36 |
| T42 | <Гектор> https://hastebin.com/share/ecejefizut.bash | 20:36 |
| mal | @nc1x72 maybe your device uses hidl vibrator then, not sure, needs some checking | 21:28 |
| mal | @Гектор you can see there the needed information, so you need to make the fixup so it replaces the by-name paths with the real device nodes, you can get full list of those by doing "ls -l /dev/block/bootdevice/by-name/" | 21:30 |
| mal | the exception is metadata which is in /dev/block/by-name/ based on fstab | 21:30 |
| T42 | <nc1x72> mal, checks point to the fact that it's aidl, not hidl | 21:36 |
| T42 | <nc1x72> ```[root@Phone1 defaultuser]# strings /vendor/bin/hw/vendor.qti.hardware.vibrator.service | grep -i aidl | 21:37 |
| T42 | <nc1x72> _ZN4aidl7android8hardware8vibrator10BnVibratorC2Ev | 21:37 |
| T42 | <nc1x72> | 21:37 |
| T42 | <nc1x72> vendor/qcom/opensource/vibrator/aidl/service.cpp``` | 21:37 |
| mal | but why is it not running then | 21:37 |
| mal | is that process running? | 21:37 |
| T42 | <nc1x72> yes | 21:37 |
| mal | then it should be seen in binder-list -d /dev/binder | 21:38 |
| mal | afaik android.hardware.vibrator.IVibrator/default | 21:38 |
| T42 | <nc1x72> [root@Phone1 defaultuser]# ps -A | grep vibrator | 21:38 |
| T42 | <nc1x72> 5706 system /vendor/bin/hw/vendor.qti.hardware.vibrator.service | 21:38 |
| T42 | <nc1x72> | 21:38 |
| T42 | <nc1x72> Also confirmed by htop | 21:38 |
| T42 | <nc1x72> Last time I checked it didn't exist! But running the command again, it shows up : | 21:40 |
| T42 | <nc1x72> | 21:40 |
| T42 | <nc1x72> [root@Phone1 defaultuser]# binder-list -d /dev/binder | grep vib | 21:40 |
| T42 | <nc1x72> android.hardware.vibrator.IVibrator/default | 21:40 |
| T42 | <nc1x72> I did run some killall few times | 21:40 |
| T42 | <nc1x72> The service restarted automatically ofc | 21:41 |
| mal | so it's running, why is ngfd not working | 21:42 |
| mal | do you see anything in logcat or journal related to vibra or ngfd | 21:43 |
| T42 | <nc1x72> logcat and dmesg return nothing. | 21:45 |
| T42 | <nc1x72> But journalctl has something interesting : | 21:45 |
| T42 | <nc1x72> | 21:45 |
| T42 | <nc1x72> systemd[3983]: ngfd.service: Service hold-off time over, scheduling restart. | 21:45 |
| T42 | <nc1x72> | 21:45 |
| T42 | <nc1x72> Apr 03 21:45:33 Phone1 systemd[3983]: ngfd.service: Scheduled restart job, restart counter is at 1. | 21:45 |
| T42 | <nc1x72> | 21:45 |
| T42 | <nc1x72> Apr 03 21:53:51 Phone1 ngfd[5141]: [497.821] ERROR: droid-vibrator: failed to write to binder (status 113) | 21:46 |
| T42 | <nc1x72> | 21:46 |
| T42 | <nc1x72> The last line spams a lot | 21:46 |
| mal | hmm, anything in logcat about binder or vibra? | 21:47 |
| T42 | <nc1x72> Nothing | 21:48 |
| mal | does restart of ngfd help? systemctl-user restart ngfd | 21:50 |
| T42 | <nc1x72> There is no systemd unit for ngfd. And I was exactly about to ask | 21:51 |
| mal | it's in user systemd, not system systemd | 21:51 |
| T42 | <nc1x72> ah | 21:52 |
| mal | which is why that systemd-user (or systemd --user depending on the environment) | 21:52 |
| T42 | <nc1x72> It definitely did something | 21:52 |
| T42 | <nc1x72> Now getting some outputs in logcat | 21:53 |
| T42 | <nc1x72> D vendor.qti.vibrator: Starting on on another thread | 21:53 |
| T42 | <nc1x72> 04-03 21:51:53.815 5706 6780 | 21:53 |
| T42 | <nc1x72> | 21:53 |
| T42 | <nc1x72> D vendor.qti.vibrator: Notifying on complete | 21:53 |
| T42 | <nc1x72> Wait ... Let me try something ... | 21:55 |
| T42 | <nc1x72> I made systemd unit to automatically restart ngfd. Doesn't help | 22:23 |
| mal | so even restarting doesn't make it work? | 23:13 |
| T42 | <nc1x72> No luck yes. Although both ngfd and the vendor vibrator service are running | 23:16 |
| mal | and no errors in logs? | 23:41 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!