Friday, 2020-03-20

*** kimmoli_ is now known as kimmoli04:11
T42<birdzhang> popup camera on sailfishos
T42<RealDanct12> @birdzhang [popup camera on sailfishos …], 👍06:05
T42<edp_17> @elros34 : should I ask @lbt here or on tjc? Thanks07:10
T42<RealDanct12> did anyone here getting a issue where lipstick would randomly crash when touching something?08:41
T42<birdzhang> i have with hybris15.1/16.0 branch on vince08:44
T42<RealDanct12> @birdzhang [i have with hybris15.1/16.0 branch on vince], have you found any fix for it?08:46
DylanVanAssche<piggz "Dylan Van Assche: kernel PR is u"> Tonight :)(09:03
T42<birdzhang> @RealDanct12 is it working?09:53
T42<RealDanct12> i applied the binder patch, nope, still crash randomly on touch09:55
T42<RealDanct12> sometimes when i touch a thing it crashes09:55
T42<birdzhang> yup, if you keep screen always on and don't touch, it not crashes09:56
T42<birdzhang> @RealDanct12 is this commits patched?
T42<RealDanct12> i wonder how is that related10:07
T42<RealDanct12> will try that10:07
T42<birdzhang> fingers crossed10:09
T42<RealDanct12> @birdzhang [fingers crossed], nope still f10:16
T42<birdzhang> okay :(10:16
T42<RealDanct12> @birdzhang [okay :(], have you added any commits when you thought it was fixed?10:18
T42<birdzhang> @birdzhang [ …], that's it, but i have no memory about 16.0 base10:20
T42<RealDanct12> :( then10:20
T42<RealDanct12> ok10:20
T42<elros34> iirc DRM_AUTH patch is needed to allow starting compositor as a non root user10:21
T42<RealDanct12> @elros34 [iirc DRM_AUTH patch is needed to allow startin …], for me it still works without that patch10:21
T42<RealDanct12> my device doesn't use DRM btw, so I guess that patch doesn't do anything10:21
T42<elros34> I guess you don't even use mali blobs?10:22
T42<RealDanct12> adreno device10:22
T42<elros34> yeah so don't use it, it's only for mali10:23
T42<DarknessHiddenorg> How to port sailfish?10:52
T42<elros34> @DarknessHiddenorg follow instruction and also
piggzjusa: hi, availbale to discuss pinepne volume isssues?12:32
T42<RealDanct12> @birdzhang
T42<RealDanct12> do you remember that hybris-16.0 also have this crash issue?15:09
T42<RealDanct12> since i see a few commits here about the binder, not really sure atm15:09
T42<birdzhang> @RealDanct12 16.0 base crashes too15:12
T42<RealDanct12> @birdzhang [@RealDanct12 16.0 base crashes too], oh, but have you got it fixed there too?15:13
T42<birdzhang> not yet, i gave up15:13
DylanVanAsschespiiroin (IRC): On the PinePhone we have a  weird situation related to charging.15:14
DylanVanAsscheWhen plugging in the charging cable, it shows the dialogue to select charge/dev mode/MTP but when unplugging, it goes to 'charging only mode' even though the cable is removed. Any idea?15:14
piggzspiiroin: the above happens when MCE is started and the cable is connected.  if MCE is started with the cabble disconnected it looks normal. (cable detection wise)....15:58
piggzthe logs always print the correct state, even tho the UI shows wrong state15:59
spiiroinDylanVanAssche: piggz: possibly... power supply device misdetection -> might need to blacklist false positive devices16:46
spiirointhere was somewhat similar case with xa2 - there is charger like power supply device in udev. status of which does not change -> causes state evaluation faults -> needs to be ignored16:49
spiiroinyou can get more insight into what is happening by: ssh over wlan, stop mce, run manually in verbose mode: mce -Tq -l modules/battery-udev.c:*16:52
spiirointhen connect/disconnect cable and see how/if udev changes make sense16:53
piggzspiiroin: ok, will try that17:13
piggzspiiroin: dev has 2 udev devices.....17:15
piggz[root@PinePhone ~]# cat /sys/class/power_supply/axp813-ac/online17:15
piggz[root@PinePhone ~]# cat /sys/class/power_supply/axp20x-usb/online17:16
piggz[root@PinePhone ~]# cat /sys/class/power_supply/axp20x-usb/online17:16
piggz[root@PinePhone ~]# cat /sys/class/power_supply/axp813-ac/online17:16
piggzwhich seem to be in harmony, but will blacklist one of them17:16
spiiroinpiggz: is it mtk kernel?17:17
piggzspiiroin:no, its pine, fully native kernel17:18
piggzkernel 5.517:18
spiiroinlooks similar to mtk, but maybe it is qc that is different from normal17:19
spiiroinif it goes like in mtk devices, both are needed: wall charger shows up in ac, pc connection in usb17:19
piggzspiiroin: they both seem to say the same whether in wall or laptop17:20
piggzposting logs......17:20
piggzcable was plugged into wall when mce was started17:21
piggzgaps are when i unplugged/plugged17:22
spiiroinmmm... on a quick glance that looks okay17:25
spiiroinbut that mtp vs charger mode is in usb-moded, not mce17:25
piggzyeah .... its just when I unplug, UI says "Charging" !17:25
piggzok, maybe its usb-moded17:26
spiiroindoes the charging indicator in status line behave in expected manner?17:26
piggzspiiroin: no ....17:27
piggzbut, if i restart mce with it unplugged it will17:27
spiirointhe icon in top left corner? the one with "lightning bolt" + shows cable connect/disconnect17:28
piggzyes ... the cable show connected all the time17:28
DylanVanAsschespiiroin (IRC): That one behaves odd17:29
spiiroini'll take a closer look at the log17:29
piggzspiiroin: if mce is started with cable plugged in, icon shows connected all the time .... if mce is started with the cable unplugged, it behaves correctly17:29
DylanVanAsscheAlso, if set mce to full verbose in its Systemd service: `-vvvvvv` then the icon behaves correct on my PinePhone17:30
spiirointhere are no POWER_SUPPLY_STATUS events -> whatever is reported in the initial probe stays effective17:32
spiiroinwhich is: axp20x-battery.POWER_SUPPLY_STATUS : (null) -> Charging17:33
spiirointhere are ad-hoc workarounds to deal with weird reporting, but it is more the other way around: if there are chargers online, we assume state=charging even if it is not reported17:36
spiirointhis might actually require that we start polling those battery properties - there are other cases too where some devices go too long / do not report some things at all17:37
spiiroinand that usb mode thing is not usb moded ... to facilitate mtk devices, ui feigns "charging" usb-mode under some conditions17:38
spiiroin(i.e. mtk kernel and charger is ac power supply device, not usb power supply device -> usb moded does not know about dedicated chargers -> make ui behave like it would)17:40
DylanVanAsscheThe missing POWER_SUPPLY_STATUS events can also cause a frozen fuel gauge? When MCE is restarted the fuel gauge is updated but that's it.17:42
piggzspiiroin: so, the issue is no event for axp20x-battery.POWER_SUPPLY_STATUS ?17:43
spiiroinpiggz: yes. would be nice to get notifications about changes like: charging, discharging, full, etc17:45
piggzspiiroin: why does it work ok if mce started when cable is not plugged in?17:46
spiiroinDylanVanAssche: I'm not 100 percent sure but... that initial udev probe involves reading from sysfs files, which triggers kernel driver to do something to provide the requested info, and it /might/ also cause notifications to be sent17:46
spiiroinpiggz: because of is_charging = || battery.status == "charging"17:47
DylanVanAsschespirroin: Okay first tackling the cable detection then :)17:47
spiiroinnow, if the battery status is stuck in something else but "charging" -> drivers the output17:48
spiiroinwhich might kind of work, except for battery full / charging error situations17:49
spiiroinbtw, if you poll that battery status (cat /sys/class/power_supply/axp20x-battery/status ?) - does it show sane values that depend on charger connect status?17:59
piggzspiiroin: yes18:03
DylanVanAsscheIf fully charged, the pattern is Charging -> Discharging (not Full)18:19
spiiroinDylanVanAssche: IIRC that ought to be ok - triggers battery level based heuristics18:20
DylanVanAsscheYes, I saw that in the code that there are 2 patterns for the heuristics18:21
piggzspiiroin: so, we nned to implement polling?18:37
piggzspiiroin: pine devs intend to fix the lack of an event, so, probably just leave it for now19:21
walidhi'm try build sfos for lg v20 tmo19:27
walidhgot this error pkgconfig(bluez5) is needed by bluebinder-1.0.8-1.armv7hl19:27
walidhbased on libhybris15.119:27
malwhere are you building that?19:28
walidhcan any help me with this error pkgconfig(bluez5) is needed by bluebinder-1.0.8-1.armv7hl20:18
malsb2 -t $VENDOR-$DEVICE-$PORT_ARCH -R -msdk-install zypper in bluez5 and then select suitable one from the suggestions20:22
walidhmal: got this error now
malcd hybris/mw/ofono-ril-binder-plugin/ && git reset --hard 1.0.10 && cd -21:42
malafter that use " -o" so it won't update sources21:43

Generated by 2.17.1 by Marius Gedminas - find it at!