Monday, 2018-12-03

eugeniohmmmm although I supply the correct input_match for the accelerometer and als sensors it doesn't seem to find the devices... perhaps my udev is still broken somehow01:02
eugeniowill debug more tomorrow, time to go to bed01:02
eugeniogood night :)01:02
*** ChanServ sets mode: +v T402:09
mortalglitchStill struggling with everything user on my HTC 10. Gets to telnet 2323 then /sbin/preinit start choking out with "/etc/systemd/user.conf.d" - no such file but when I check there is a "/etc/systemd/user.conf" just not a conf.d    there are others like /run/systemd/user.conf.d which is missing entirely04:54
mortalglitchOk so I have realized that when I make changes to my pme_defconfig then re-run "make hybris-boot" if I check the config it acts like nothing changed06:07
T4<adampigg> Is there a description anywhere of all the different egl and qpa plugins?07:48
r0kk3rzi dont think so08:34
r0kk3rzthe usual ones were fbdev and hwcomposer08:34
r0kk3rztheres probably others08:35
*** ChanServ sets mode: +v T408:53
r0kk3rzdoes the xperia xa2 do vulkan?08:59
r0kk3rzcould be a project for someone08:59
spiiroinreview appreciated:
spiiroin^ display state woes after compositor role hand-off09:22
malr0kk3rz: I have a wip vulkan support for libhybris but still missing part of it, currently it can do calculations but no graphics10:10
r0kk3rzyeah i remember10:12
malr0kk3rz: I probably should continue doing that at some point10:15
r0kk3rzim sure you've got enough going on already :)10:15
malr0kk3rz: yep, that is not the highest priority for me10:15
r0kk3rzmal: was that pushed anywhere?10:18
malr0kk3rz: it seems no10:20
T4<adampigg> Fixed slow connection, dodgy cable12:17
T4<adampigg> Some odd permission errors starting user service, jot seen before12:18
T4<adampigg> Permissions in /run/user/...12:18
T4<adampigg> mal: current status
T4<adampigg> still some fundamental issues, like Failed to create generator directory /run/user/100000/systemd/generator: Permission denied13:02
T4<adampigg> guhl had something similar recently, was there a solution?13:02
mal@adampigg have you verified you have the proper tmpfs related kernel configs enabled13:11
T4<adampigg> mal: i dont recall any dire sounding warnings from the kernel config checker, but i will check later13:32
T4<adampigg> mal:
T4<adampigg> those look the same as i have in mido14:01
mal@adampigg how do the permission look on device?14:07
T4<adampigg> mal:
mal@adampigg but how do those look on mido14:17
T4<adampigg> mal:  … very similar
*** ChanServ sets mode: +v T414:24
T4<adampigg> mal, but on mido, /run/user is populated, latte isnt14:25
T4<adampigg> put perms look the same14:25
mal@adampigg could you pastebin journalctl output14:40
T4<adampigg> mal, that what i started with ;)
taixzomal: any news on matissewifi?14:53
maltaixzo: not right now, I should continue that again14:56
maltaixzo: or maybe I just need to give a less perfect image14:57
T4<adampigg> mal: first failure seems to be autologin@10000015:04
taixzomal: I'd be happy with a less than perfect image for now.15:04
maltaixzo: ok, I'll see if I have time to build image tonight with at least some things working15:06
mortalglitchmy build still can't seem to get anything setup related to the user during boot. "/etc/systemd/user.con.d" no such file /run/systemd/user.conf.d missing etc..   systemctl also reporting the autologin@100000 and user@100000 failed any clues on this one? I've messed around with toggling things like the fw_user_helper in the kernel and still seem to get no results15:29
T4<adampigg> mal: ^^ could be similar to my issue too15:43
mal@adampigg maybe16:08
T4<adampigg> Mal, does anything jump out in my joirnalctl?16:17
T4<adampigg> @eugenio_g7 did you have any.permission problems with sysyemd etc?16:23
T4<eugenio_g7> Well, currently I have to chmod -R  777 /dev to make UI working from the nemo user16:29
mortalglitchtrying to rebuild packages and now getting a buffer overflow when rebuilding pulseaudio16:35
T4<adampigg> @eugenio_g7 can u post a journalctl after a reboot16:43
T4<eugenio_g7> Ok, will try once I return home from work :)16:46
mal@adampigg what kind of mount services do you have?16:48
malat least booster-silica is segfaulting16:49
T4<adampigg> Mal, i can tell you when i get home, approx an want *.mount systemd services?16:49
T4<adampigg> Yes, saw that one16:49
T4<jersobh> is there a port for zenfone?17:50
*** OhYash1 is now known as ohyash18:14
mortalglitchrolled tooling/target back to and rebuilt everything now the user services start but the droid-hal-init.service "can't set property" and fails18:14
malmortalglitch: show the whole log18:33
mortalglitchjournalctl dump :
*** Gizmokid2005 is now known as Mogwai18:39
*** Mogwai is now known as Gizmokid200518:40
piggzmal: mounts
*** Gizmokid2005 is now known as Mogwai18:42
piggzso thats interesting, i thought the slow usb-net connection was down to the cable, but it seems to be my laptop...using the same cable that I used elsewhere today, and the issue persists ... also, im getting loads of kernel messages about power, which i wasnt getting on the other computer18:44
piggz[  197.715649] dwc3 otg_set_power: 500 mA18:44
piggz[  197.715732] intel-cht-otg intel-cht-otg.0: Draw 500 mA18:44
piggz[  197.715788] dwc3 otg_set_power: 500 mA18:44
piggz[  197.715841] intel-cht-otg intel-cht-otg.0: Draw 500 mA18:44
piggz[  197.717311] dwc3 otg_set_power: 2 mA18:44
piggz[  197.717375] intel-cht-otg intel-cht-otg.0: Draw 500 mA18:44
T4<elros34> mortalglitch: SELinux:  Starting in permissive mode. How did you disable selinux?18:45
mortalglitchI did the commandline parameter from the boardconfig18:59
T4<elros34> do you have CONFIG_SECURITY_SELINUX_BOOTPARAM enabled?18:59
mortalglitchright now it's showing is not set, could of swore I changed that to y19:01
mortalglitchconfig_audit is still y19:01
T4<elros34> audit can be also disabled via boot param but not sure it's really needed19:02
mortalglitchsidenote : If I run droid-hal-init on the device from telnet it reboots to recovery19:04
mortalglitchrebuilding kernel now19:04
T4<elros34> iirc: you can't run droid-hal-init twice, you must mask it and then run it manually19:05
malaudit should be enabled if bootparam is used19:23
mortalglitchgot the flags set bootparam=y and audit=y now, rebuilt and still seeing a similar behavior. How do I check on the device to see if it's truly disabled?19:28
mortalglitchcat /etc/sysconfig/selinux doesn't exist19:31
malmortalglitch: if selinux would be enabled the device would most likely reboot when droid-hal-init is starting19:37
T4<meierrom> @jersobh: I think Mister_Magister is working on port for zenphone. If I remember well it's a rather difficult port with many challenges.20:37
*** ChanServ sets mode: +v T420:58
eugeniopiggz, so this is my systemd journal just after the UI started:
piggzmal: interesting21:20
piggzmal: i moved that plugin out of the way to stop it loading....21:20
piggznow, files are created in /var/run, and lipstick tries to start, but no gfx21:20
piggzso might install eugenio's mesa bits21:20
malpiggz: anything in log why graphics fail, check also logcat21:21
piggzmal: logcat
piggzmal: me running lipstick
malpiggz: looks normal21:23
piggzi can try setting the size variables21:23
piggzbut i doubt that will fix the crash21:23
malpiggz: does minimer work? or test_hwcomposer21:23
piggzno, i tried test_hwcomposer and it segfaulted...21:24
mortalglitchhaha, I almost feel we're almost at the same point. I finally got the screen to change and then it went black21:24
mortalglitchlipstick error21:25
piggz[nemo@Sailfish feedback]$ EGL_PLATFORM=hwcomposer test_hwcomposer21:25
piggztest_hwcomposer: test_hwcomposer.cpp:204: int main(int, char**): Assertion `err == 0' failed.21:25
piggzmal: strace
piggzmal: how does the surfaceflinger plugin work?21:26
piggzwhere is it?21:27
malpiggz: in mer-hybris, it needs some additional files21:28
piggzmal: what do you think, that, or the mesa route?21:29
malpiggz: not sure21:29
malwould be interesting to know why hwc fails21:29
piggzdidnt TheKit ahve a theory?21:29
piggzTheKit: ^^21:29
TheKitthat was for EGL/GLES itself, not hwcomposer assertion fail21:31
piggz_TheKit: ok21:32
piggz_mal: probably a job for krnlyng :D21:38
krnlyngopen("/sys/class/drm/card0/power/i915_videostatus", O_WRONLY|O_CREAT|O_TRUNC, 0666) = -1 EACCES (Permission denied)21:44
krnlyngopen("/cache/hwc.reg", O_RDONLY)        = -1 EACCES (Permission denied)21:45
krnlyngopen("/proc/cmdline", O_RDONLY)         = -1 EACCES (Permission denied)21:45
krnlyngi would check these first piggz_21:45
eugenioso I'm having some trouble getting sensors to work, kernel modules now load, and I do have the devices present in /sys/bus/iio. I have also recompiled sensorfw with the other spec file, and I do have indeed the iioaccelerometer adaptor compiled21:45
eugeniosensorfwd configuration + output in test mode:
maleugenio: why did you recompile sensorfw?21:55
maleugenio: how did you test the sensors?21:55
eugeniothe iio adaptor is not bundled with sensorfw-hybris, so had to use the other spec file (which doesn't define CONFIG+=hybris)21:56
eugenioin lineageos they were working21:56
malthe other part is always built and installed21:57
maleugenio: how did you test the sensors?21:58
eugeniohm... just picked the -hybris spec file21:59
eugenioin lineageos orientation and als worked fine21:59
eugeniomal, I'm talking about this:
eugeniothere're two spec files on the sensorfw git repository22:00 picks sensorfw-qt5-hybri.spec instead, with disables the other adaptors (as per
eugenioalso tried via csd but nothing22:03
maleugenio: that other spec is build in jolla repos, so it is always available just like I said22:07
eugeniooh ok22:08
maleugenio: you need something like this
maladapt the settings based on your device features22:09
eugeniook thanks, trying now (also switching to the sensorfw in the official repositories in the meantime)22:10
piggz_krnlyng: i fixed the first to of htose errors ... but i dont have a /cache/hwc.reg22:11

Generated by 2.17.1 by Marius Gedminas - find it at!