Thursday, 2021-05-27

T42<A_T_R> @elros34 [@A_T_R did it help at least for permission err …], No actually03:09
piggzmy system.mount looks ok:08:01
piggzbut /system doesnt look like a /system ... but /system/system does08:02
deathmist_piggz: welcome to system-as-root :)08:03
piggzdeathmist_: cool, known problem .. whats the workaround? :)08:05
deathmist_piggz: system-as-root should be automagically handled by droid-hal when it generates mount units, but only as long as fstab defines the system partition mountpoint as /08:07
piggzdeathmist_: it doesnt :/09:42
piggz /dev/block/platform/bootdevice/by-name/system       /system            ext4        ro                                                    wait,avb=boot,first_stage_mount09:42
piggzsafe to change and rebuild?09:42
T42<b100dian> mine says `/` and it mounts correctly (as system_root then overlayed/bound `/system_root/system` as `/system`)09:51
T42<adampigg> yeah, i edited the fstab and its built correctly now10:27
T42<adampigg> now onto d-h-i errors/failing services10:28
T42<adampigg> probably some linker fixes needed10:28
piggzwhen i try and run logcat, it says "no such file or directory" .. although, there certainly is!10:47
T42<elros34> droid-hal-init is probably dead10:58
deathmist_piggz: is this a 17.1 port? if so check
piggzdeathmist_: it is, thx11:06
piggzdeathmist_: mal: ive not got logcat at least!11:24
piggzi guess this is bad May 04 00:02:21 VollaPhone unknown:  (5)[982:logd.auditd]type=1701 audit(1620075741.604:675): auid=4294967295 uid=1000 gid=1003 ses=4294967295 subj=kernel comm="android.hardwar" exe="/vendor/bin/hw/" sig=611:33
T42<b100dian> piggz what is the error in /tombstones or logcat11:56
malpiggz: as usual full logs might help debugging11:56
piggzdeathmist_: mal: @b100dian: do i need to do anything with ld.config.2?.txt ?12:38
T42<b100dian> I think vold shoild have been disabled by default on 17.1 can you also show your rc files? What is the owner/group on loop* devices?12:46
piggz@b100dian root:disk 66012:51
piggzvold isnt disabled in the services file12:54
piggzhmm, ome moment12:55
piggzvold fixed ... logs now look like
piggzi had a bunch of rc files from the pre-17 port13:07
T42<A_T_R> @elros34 hi , after changing permission :input now permission error one ..but still issue persist13:17
T42<elros34> create udev rule for that then if it work  as usual grab fresh logs13:18
T42<A_T_R> sorry , how can i create udev ?13:24
T42<A_T_R> udev rule*13:25
T42<elros34> use any guide you can find or read some examples from /lib/udev/rules.d/. Have you checked what set such a permission on your device? If it's et in init*.rc files then udeve rule will not work13:31
T42<A_T_R> my init.rc
T42<elros34> check yourself: ctrl+f then 'uevent' or 'input'13:36
T42<elros34> sorry 'event'13:36
T42<elros34> be aware that there are more init*rc files on your device13:37
T42<b100dian> piggz: you don't still have selinux_stubs in the rest of the rc files do you13:38
T42<b100dian> >  SELinux : Unknown class hwservice_manager13:39
T42<b100dian> Can you also share journalctl --boot after setting it persistent?13:39
T42<A_T_R> searched for event and input nothin similar with /dev/input13:39
piggzb100dian: no, ive only got the 2 rc files from sparse-10 now13:39
T42<b100dian> selinux=1 in kernel cmdline?13:40
T42<b100dian> you need to actually do enable selinux on hybris-17.1 if I understand correctly13:42
T42<b100dian> permissiive is good though13:42
piggzi didnt actually change that from the device tree, thats how it was13:43
T42<elros34> I can't find any sign of system_root in your log13:46
piggz /dev/mmcblk0p31 on /system_root type ext4 (ro,relatime)13:49
T42<elros34> maybe if you check mount but still this should be printed in journal but it's not in your latest log. Do you use some custom The one from submodule should bind mount modified ld.config.29.txt13:55
T42<adampigg> @elros34 the descrition will say /system becuase I hadnd balled it13:57
T42<adampigg> and yes, that file is bind mounted13:58
piggz/ # cat  /system/etc/ld.config.29.txt | grep libexec13:58
piggznamespace.default.permitted.paths += /usr/libexec/droid-hybris/system/${LIB}13:58
T42<b100dian> hmm.. I had specific sphal services crashing because of the wideness of system_overlay.mount, but I see it does not start in your log14:00
T42<b100dian> However the early systemd errors are new to me14:00
T42<b100dian> (like systemd-gpt-auto-generator[731]: Failed to dissect: Input/output error)14:01
T42<A_T_R> @elros34 /lib/udev/rules.d/999-android-system.rules -14:03
T42<A_T_R> here it mentioned14:03
T42<A_T_R> already*14:04
T42<elros34> @A_T_R then it must be overwritten to system:system by something else, most likely by some rc file14:30
T42<A_T_R> droid-hal-init: Could not find 'android.hardware.keymaster@3.0::IKeymasterDevice/default' for ctl.interface_start14:31
T42<A_T_R> this important ?14:32
T42<A_T_R> dmesg14:32
T42<elros34> I think keymaster is used for fpd but no idea whether is required for gui14:34
T42<A_T_R> i will share full log here now14:34
T42<A_T_R> dmesg -
T42<A_T_R> journal -
T42<A_T_R> logcat -
T42<ankaos> hı. how fix this error?
malin which environment are you trying to build that and which android base version?20:46
T42<b100dian> Hi, I'm trying to replace ngfd-plugin-native-vibrator with the droid version because the kernel "drv260x only accept custom waveforms" message regarding periodic effects21:36
T42<b100dian> The trouble is that altough I've built it, added it to patterns and installed it with `sb2 -t $VENDOR-$DEVICE-$PORT_ARCH -R -msdk-install zypper in ngfd-plugin-droid-vibrator`, the image creation command fails21:36
T42<b100dian> "patterns-sailfish-device-configuration-tucana-1-202105272125.aarch64 requires ngfd-plugin-droid-vibrator, but this requirement cannot be provided, not installable providers: ngfd-plugin-droid-vibrator-1.2.2-1.aarch64"21:37
malhow did you build that package?21:41
mal@b100dian there is no need to have it installed in target because image build doesn't care about that the same way21:42
T42<b100dian> with `rpm/dhd/helpers/  --mw=`21:47
T42<b100dian> ok, there is something I'm missing about image building indeed21:48
mal@b100dian there are two of those vibrator packages, also remember to change have_native_vibrator in droid-hal-version spec to have_vibrator21:53
maland then rebuild droid-hal-version rpms using rpm/dhd/helpers/ -v21:53
mal@b100dian are you sure which type of vibrator you want to use, can you show output of "evdev_trace -i" from mce-tools package21:55
T42<b100dian> this is the output
T42<b100dian> My supposition is based on the fact that ngfd dies after the above drv2604 error21:57
T42<b100dian> that it "only accepts custom waveforms"21:57
mal@b100dian based on that you could be using ffmemless vibrator21:57
T42<b100dian> but how do you define a custom waveform in the ini file?21:58
malthat is what memless driver supports21:58
mal you can ignore the kernel driver part of that21:59
malmaybe droid vibrator would also work but memless would support more advanced vibrator effects22:00
T42<b100dian> I don't know if this is expanded to a custom pattern `_WAVEFORM = [sine|triangle|square], defaults to sine`22:04
T42<b100dian> the message here implies it's not
T42<b100dian> Hmm rebuilt —configs, same error (and yes, I made the change for the qt package too)22:08
T42<b100dian> should have_memless be in drod-hal-version instead?22:12
T42<b100dian> (instead of me replacing have_hibrator_native with have_vibrator)22:12
malwell based on that driver maybe it's better to try droid first22:13
T42<b100dian> mal: ok, continuing, I still have the —mic error after building —configs22:14
malthe memless vibrator code in sailfish seems to only support sine22:14
mal@b100dian did you do the droid-hal-version changes and rebuild?22:14
T42<b100dian> maybe pbkac, re-checking. . Hmm I have native vibrator in adaptation too22:17
T42<b100dian> yes, that was it - I'm not sure why both pattern inc files were including the # Vibra section22:20
T42<b100dian> thanks mal for pointing me out the have_vibrator_native version spec22:21
T42<b100dian> still ngfd "Vibra upload effect: Invalid argument" and crashes..22:31
malwait, that message suggest it's trying to use memless22:48
mal@b100dian try editing /usr/share/ngfd/ngfd.ini and remove the mention of memless from there22:52
mal@b100dian the line with memless in it now should then be "plugins-optional = droid-vibrator;devicelock"22:52
T42<b100dian> ok, at least it is now running, but I don't get vibrations yet23:01
T42<b100dian> I only have this message that was there before `ngfd[6860]: libprofile: session bus connection requested while blocked`23:08
T42<elros34> that message is not critical, I have it also in logs23:16
T42<b100dian> there's a vibrator service disabled, but I don't find that path, let me comment it and reboot23:21
T42<b100dian> this one's started but still no vibration: /vendor/bin/hw/vendor.qti.hardware.vibrator@1.2-service23:27
T42<b100dian> going to bed now, thanks23:27

Generated by 2.17.1 by Marius Gedminas - find it at!