Thursday, 2018-11-29

OrokuSakimy drm service status is restarting as well00:29
OrokuSakihybris 11.0 running
OrokuSakilineageos 11 rom00:30
OrokuSakiwhen I try to use the camera logcat complains my camera service isn't running although codecs are working. I figure if I can get android to start the services automatically instead of manually with systemd and a start script, maybe my camera service will work00:33
OrokuSakiI recompiled droidmedia with HAL100:34
*** ChanServ sets mode: +v T402:10
*** doniks_ is now known as doniks04:37
abransonOrokuSaki: what do your init.rc additions look like? is the android camera service disabled? do you see anything in logcat or the journal from when they were supposed to start?07:24
Mister_Magisterspiiroin: ping07:26
abransonMister_Magister: what's up?07:40
Mister_Magisterabranson: u no spiiroin :C07:40
Mister_Magisteris he on vacation or smth?07:40
abransoneveryone's really busy atm07:41
abransonif you have a question, best to just submit that rather than pings07:41
abransonthen they might get answered by someone else07:41
Mister_Magisterabranson: no no there is ongoing thing between us so i need to talk with him07:45
abransonok, probably best still to lead with more than just a ping though :)07:48
T4<adampigg> Check out @adampigg’s Tweet:
T4<adampigg> £20! For my app!07:52
spiiroinMister_Magister: pong07:53
Mister_Magisterspiiroin: hiii remember led thingy? i got it to break07:54
Mister_Magisterits always on so proximity sensor is either locked in close or closing07:55
Mister_Magisterspiiroin: so now i finally got it broke and i have no idea what next xd07:56
spiiroinMister_Magister: "or closing"? ... if that would be "or opening", it t would indicate the sensor very briefly reports "not covered" states too as in so short time it gets eaten away by filtering07:57
spiiroinnow if you have journal, that might be interesting to see07:58
Mister_Magisterspiiroin: there was nothing in journal about led07:58
spiiroinMister_Magister: if it has p-sensor logging, it does not matter that much07:58
* spiiroin tries to remember whether it is raw or filtered p-sensor state that is logged ...07:59
Mister_Magisterspiiroin: ima kinda lost on what you are saying07:59
spiiroinled = just a tool to discover there is something odd with the sensor -> having led log = nice, but not necessary as we want the sensor logs; now the question is: do you have p-sensor state changes in journal (as the led activity points towards: raw sensor state does change, but filtered sensor state might not)08:01
spiiroinif there is nothing about proximity in logs, then the default logging probably is about filtered state -> need to increase verbosity so that we get raw data too in the log08:02
Mister_Magisterspiiroin: i checked journal when the lock happened and there was nothing in journal08:02
spiiroinMister_Magister: ok, I'll check the sources a bit08:03
*** ChanServ sets mode: +v T408:07
spiiroinMister_Magister: edit /lib/systemd/system/mce.service so that it says: "-ExecStart=/usr/sbin/mce --systemd -lmce-sensorfw.c:sfw_backend_ps_sample_cb -ltklock.c:tklock_datapipe_proximity_*"08:10
spiirointhen you should have a) raw sensor data as received from evdev/sensorfwd b) how it gets filtered & applied08:11
Mister_Magisterspiiroin: now we have to wait again08:15
Mister_Magisterspiiroin: because restarting mce fixes proximity sensor08:16
Mister_Magisterspiiroin: just some control log when proximity is working fine
spiiroinMister_Magister: can you throw "evdev_trace -I" output in there too08:21
spiiroin... what sensorfwd is reporting looks odd compared to what mce ends up using from evdev08:22
spiiroinaround line 92: mce, sensorfwd, **and** sensorservice have the input device open08:25
Mister_Magisterwhat about it?08:25
Mister_Magisteris it bad?08:25
spiiroinwhich could indicate that sfos sensorfwd & android sensorservice are using android sensor hal simultaneously08:26
Mister_Magisterspiiroin: am i supposed to kill sensorservice08:26
spiiroinif android side turns p-sensor off, sfos does not know anything about that -> in sfos side the sensor is enabled all the time -> never gets toggled -> it could be disabled despite sfos thinking it is enabled08:26
spiiroinMister_Magister: depends, but probably it should not be started at all08:27
spiiroinsledges: how is that ^ android sensor service thing usually handled?08:27
Mister_Magisterspiiroin: i can't really remember but i have sensorservice because sensors wasn't working08:28
spiiroinit might be ok, if it is providing data for sensors hal; if it is just using sensors hal -> that is going to be a problem08:29
Mister_Magisterspiiroin: i tried disabling it if i  won't get sensors then we will know08:33
spiirointhere are also things like        sensord(pid=891,fd=8)08:34
spiiroinwhat is that?08:34
* spiiroin also wonders why lipstick ends up having bunch of input devices open...08:35
Mister_Magisterspiiroin: i dunno sorry08:37
Mister_Magisterspiiroin: lol sensors seem to work without sensorservice08:41
spiiroinevdev_trace -I | grep mce -> gives pid=10105; readlink /proc/10105/exe -> gives /usr/sbin/mce08:43
spiiroin; rpm -qf /usr/sbin/mce -> mce-1.98.2-1.38.1.jolla.armv7hl08:43
spiiroinyou can repeat that, but for "sensord" instead of "mce"08:43
spiiroincould be "sensord" = something that provides data for sensor hal; "sensorservice" = something that uses sensor hal to provide higher level android sensor api for other services / java apps / whatnot08:45
Mister_Magisterspiiroin: im not sure what you wan't me to do08:47
spiirointhere is a possible architectural sore point if something like this happens: sfos --needs--> android camera --needs--> android sensors --conflicts-with--> sfos sensorfwd08:47
spiirointry disabling that sensor service from android service configs08:48
spiiroincheck if something breaks or not08:48
spiiroinif nothing breaks, leave it disabled08:48
Mister_Magisterspiiroin: i just told you i tried disabling it and sensors are still working :)08:49
spiiroinyeah, and I just told you that for example camera might not work anymore08:50
spiiroinyou have sensors, but might have lost something else08:50
Mister_Magistercamera seems to work08:51
Mister_Magisteri mean thats ofcurse after reboot after disabling sensorservice08:51
Mister_Magistereven two cause it bootlooped08:51
spiiroinwell, then you're done ;-)08:51
Mister_Magisterspiiroin: We will see08:51
spiiroinexactly ;-)08:52
Mister_Magistermaybe sensors didn't work 3 years ago without sensorservice but now they do08:52
spiirointhis device port is 3 years old?08:52
Mister_Magisterspiiroin: crazy huh?08:53
spiiroinwell, that is one way to put it ;-)08:54
Mister_Magisterspiiroin: whats other way?08:54
spiiroinof course there have been bunch of changes in startup in general, sensor access, etc so there just might be a rational explanation too, but probably not worth while to dig it, so ...08:55
spiiroincrazy stuff happens is just fine by me ;-)08:55
Mister_Magisterspiiroin: yeee thanks for help08:56
Mister_Magisterthis port is half of sfos age :P08:56
sledgesPSA: SFOS collab. @ #mer-meeting starting now-ish09:00
sledgesspiiroin: we disable many android services that conflict with hybris areas of influence. probably the same fate could be for sensorservice, if brings usefulness09:08
r0kk3rzsomething to watch out for11:18
r0kk3rzMister_Magister: what device was that about?11:22
Mister_Magisterr0kk3rz: ?11:24
T4<abhishek_0> works fine with latest twrp here on wt8804712:30
*** ChanServ sets mode: +v T414:10
T4<adampigg> So, there is another open source Finnish phone company.... But they're not using sailfish..... Don't you guys talk???16:00
r0kk3rzobviously, they're finnish16:07
r0kk3rzthey avoid each other at all costs16:08
UmeaboyI managed to build AOSP 7.1.2_r36 for maple now.17:23
UmeaboyI've started to create the local_manifest to build hybris and this is how it looks now:
UmeaboyThe hadk doesn't mention which other I need to use, have I missed something?17:26
UmeaboyThe device and kernel repo for AOSP is there.17:26
UmeaboyDon't I need fingerglue and droidmedia?17:27
T4<meierrom> @adampigg: Will it be possible to dual boot windows/sfos on Xiaomi MI Pad 2?17:29
T4<adampigg> Well, atm I haven't a working sfos, so.. TBD17:30
T4<adampigg> And, I won't personally be doing that, I have the baby mitab217:30
T4<eugenio_g7> It's possible theoretically17:31
T4<benzandsBot> @meierrom if it’s possible to install grub on mi pad 2, then probably yes 😄17:32
*** doniks_ is now known as doniks17:37
T4<adampigg> @eugenio_g7 [It's possible theoretically], How's your progress17:41
T4<eugenio_g7> @adampigg UI works, brightness works but still a long way to go. I guess this evening I'll push the updated mesa and libdrm so that you can try too17:43
UmeaboyI'm trying to find a similar up to date manifest to adapt from.17:50
piggzmal: abranson: rebasing gst-droid now18:16
piggzabranson: any plans to enable the AudioPoilcyService in master gst-droid?18:32
piggzi mean droidmedia18:32
malpiggz: it would be interesting to know why it's needed18:34
malpiggz: and only one some devices or bases18:34
malpiggz: we already disabled the usual sounds in frameworks/av but for some reason that is not enough18:35
piggz@eugenio_g7 have you had charging problems with your mitab2? atm mine seems daed with just a red flashing LED ... it was on last night!18:49
malpiggz: how is porting going?18:52
T4<eugenio_g7> piggz: it does that when the battery is completely drained, keep it attached to a charger, it should eventually boot up (it could take hours). Try to boot to TWRP and keep it charging there.18:53
piggzmal: not started the porting ... doing the gst0droid rebase for you first!18:55
piggzbut i have to first fix my droidmedia18:55
piggzwhich is now done18:55
T4<eugenio_g7> piggz: you could somewhat speed-up this operation by opening the tablet, removing the battery and reattaching it. Just be careful in handling the back cover as it scratches easily18:56
piggz@eugenio_g7 no i will just leave it ... i just read about the battery disconnect hack and thought it was fake!! :D18:57
piggzi think i probably left it at the boot confirmation screen last night18:57
T4<eugenio_g7> Heh :D got bitten by this thing many times, now I'm always careful18:58
piggzphew, atleast its not bricked!19:01
piggzmal: anyway, i think @eugenio_g7 has done the most porting so far ... hopefully i can fixup the finer details19:07
maldoes it boot to UI yet?19:08
piggzyes apparently19:08
piggzusing a new Mesa i think19:08
piggzi was gonna go down the regular hybris GL route, but i think he got native working .... but on an older kernel? @eugenio_g7 can you explain exactly how it is working?19:09
malpiggz: rebased gst-droid seems to work20:10
piggzmal: just about to test!20:11
piggzmal: yep, seems to work as before20:12
malpiggz: I will go through both qtmm and gst-droid changes sometime in the following days to see if there are any issues remaining, but looks ok so far20:14
piggzmal: fingers crossed!20:16
eugeniopiggz: yup, updated mesa + libdrm and recompiled Qt to enable the KMS eglfs integration20:37
eugeniostill on lineage (cm12 actually)'s kernel, but eventually I'd like to update that20:38
piggzeugenio: great, get them on our obs project! ;)20:39
eugenioyup! need to polish the specs a bit but I will soon(tm). Opened this project up to avoid polluting the main namespace20:40
piggzoff out now, laterz everyone20:41
*** ChanServ sets mode: +v T420:48
abransonpiggz: most devices don't need the audio policy thing, so it's better not to make everything run it. we have too many media related android services running as it is21:00
T4<adampigg> Abranson,  yeah  ,its fine, its a small patch now i sorted out my droidmedia repo21:08
malabranson: it would be interesting to know how many devices need it and which bases those use etc21:09
abransonmal: yeah, there must be some other instantiation of a mediaplayer that is only in those21:11
abransongdb would probably tell you, cos it hangs doesn't it?21:11
malabranson: yep, fp2 needs that so maybe could try21:12
abransonmight be in a blob though, but we could always turn the mediaplayer into a dummy21:12
malyes, dummy audio policy could work21:13
r0kk3rzandroid hals just being a house of cards could do it too21:21
T4<adampigg> Hoping the tab os alive again by the time i bag from the pub!21:44
abranson@adampigg everythings better after the pub21:50
T4<adampigg> Abransob, esp after a pool win :D21:51
eugeniohm, is the "wh" tool to configure webhooks via CLI not available anymore?22:04
eugenioI do have the sdk-utils package installed but it isn't there22:04
r0kk3rzyou can do webhooks without it22:07
eugenioyup, created the _service file manually, but that tool seemed handy22:07
eugeniohm, builds doesn't seem to fire up though: CI Bot commits fine but the build doesn't start (as an example:
eugeniowell, I'll retry tomorrow. Goodnight :)23:33

Generated by 2.17.1 by Marius Gedminas - find it at!