Thursday, 2020-06-04

riniguspiggz: congratulations! I'll still have to finish the code, probably 1-2 h tonight should do it with testing05:28
T42<KernelPanix> hello07:05
T42<KernelPanix> this log07:05
T42<KernelPanix> please help me07:05
T42<KernelPanix> ☺️☺️☺️07:05
T42<adampigg> Rinigug gr8, i wonder what we tackle next!07:46
rinigus@adampigg: qt update, obviously08:10
T42<adampigg> rinigus: lol :D08:11
T42<naive17> rinigus : where would you start to update qt?08:13
riniguswill see. have to look through few stalled issues on maps.08:13
rinigusr0kk3rz: nemomobile maybe a path to new qt, indeed08:13
r0kk3rzyou can test things that require it on it08:15
rinigusr0kk3rz: did they managed to make the update? I think I missed it if they did08:16
piggzi think they did08:16
r0kk3rzi think so08:16
piggzneochapay was posting about it08:16
rinigusinteresting. maybe we should look into helping them and work through those parts which are closed source in sfos.08:19
rinigusprobably would need some extra device for nemo. something easy to find/buy, decently supported08:20
r0kk3rzyou can build an image from a sfos port08:21
r0kk3rzor at least you used to08:21
r0kk3rzneochapay has been building for pinephone i think08:22
T42<neochapay> Who say nemomobile ? ^_^08:23
riniguswell, let's get FP over the line first. and I can ask then on nemo channel for update on devices/qt/how to help.08:23
rinigus@neochapay : see above. would love to continue the discussion out-of-working hours, though08:25
piggzrinigus: PP is nice n cheap08:26
riniguspiggz: hard-to-get?08:27
T42<naive17> rinigus : not too much08:28
rinigusfunny enough, they still have xperia x in stock. 1/2 price for pp08:29
rinigusno, make it the similar price to pp08:29
rinigusdoes pinephone have nemo edition? :). will look into it bit later08:31
T42<neochapay> rinigus yeap. see nemomobile.net08:31
rinigusneochapay: I'll come back to this subject. thanks for the link08:33
piggzrinigus: can 10 be closed?
riniguspiggz: sadly, no. I still get reboots / freezed device. mainly waiting if someone else will observe the same09:02
r0kk3rzpiggz: so whats this finger whatsit?09:03
piggzr0kk3rz: wdym? its an open source implementation of sailfish-fpd09:03
piggzso, allows fingerprint support on communit devices09:03
r0kk3rzusing hybris?09:04
piggzdoesnt use a slave package, talks to android hidl over hybris09:04
r0kk3rzah hild thingie09:04
piggzr0kk3rz: this is the android side
piggzvery small09:05
piggzthose parts were taken from biometryd for UT09:06
piggzthen weve implemented a new daemon onto that mimics the sailfish-fpd dbus api09:06
piggzreverse engineered using dbus-monitor, strings, and mce source09:06
piggzand it works well for volla and rinigus sony09:06
T42<birdzhang> and vince, hybris16.0 base ;)09:08
r0kk3rzsounds cool09:08
piggzr0kk3rz: yeah, im pretty happy with it :))09:12
deathmist@piggz I may test this today as biometryd works perfectly for me on Halium 9 based UT, would I need to build an entirely new mic image or can I just grab some updated RPMs from droid-local-repo?10:23
T42<birdzhang> deathmist:
T42<birdzhang> copy out/target/product/$DEVICE/system/lib/ to your phone /usr/libexec/droid-hybris/system/lib/10:26
T42<birdzhang> then install sailfish-fpd-community and sailfish-fpd-community-test from
T42<birdzhang> this is the easiest way10:28
piggzdeathmist: what he said :D10:37
rinigusnote that the builds at obs are OLD. VERY OLD11:07
rinigusfor FPD. Deathmist ^11:08
piggzrinigus: deathmist: where "old, very old" is in the context that the project has existed for 7 days11:15
T42<ankaos> the call is going. but it turns off after 1 minute. sms is going. but sms is not coming to me.11:37
T42<ankaos> how can i solve this?11:37
deathmistpiggz: rinigus: this keeps happening always towards the last few remaining and I can never enroll a fingerprint using the latest version of the test app, any suggestions?12:26
deathmistlatest version as in built from source & commit da183c90c7742c0faec201a9ff4e57c3c9f90fa512:28
piggzdeathmist: tried enrolling in sailfish settings?12:28
piggzas usual, logs and logcat12:28
deathmistpiggz: attempting to install "sailfish-devicelock-fpd" results in either way here is journal and logcat after coming across the issue once right after rebooting12:37
rinigusDeathmist: did you build it against your own hybris? also the android bits? not sure it matters, though12:46
deathmistye I built this on my local 16.0 tree with HAL stuff etc in place12:47
rinigusDeathmist: looks like qseecomd is rather unhappy. I wonder if its an issue with the paths that are not freely selectable in practice12:48
piggzqseecomd is unhappy on my device too12:49
piggzpro1, not volla12:49
piggzdeathmist: you woulld have to remove jolla-devicelock-plugin-encsfa from you pattern, or, be brave and use rpm -e :D12:49
riniguspiggz , Deathmist: you don't need to switch to Jolla's lock implementation as enroll is not working12:50
rinigusbut the path is rather odd there (.bak)12:50
piggzneed to fix that first12:50
rinigusin logcat12:50
rinigusDeathmist and piggz: issue is in journal: `sailfish-fpd-community[6017]: setActiveGroup failed: SYS_EINVAL`12:51
piggzrinigus: folder permissions12:51
piggzdeathmist: ^^12:51
rinigusthat has to be fixed. you can wait for my code tonight, or try to make folder manually. but folder may change as well12:52
deathmistpiggz: I can add fingerprints with that in place & identification also works :)12:58
piggzgreat, hopefully it will continue to work when we make some folder structure changes12:58
T42<birdzhang> this should be add to README12:59
piggzbut, we have another user rinigus!12:59
piggzbirdzhang: it may not be needed after tonight12:59
T42<birdzhang> waiting for the magic xD12:59
riniguspiggz: nice! let's see if different folders would work as well.13:03
T42<birdzhang> ohmd.service failed to start after reboot sometimes, journalctl log: hybris16.0 base, selinux disabled via cmdline and CONFIG_SECURITY_SELINUX_BOOTPARAM_VALUE=0, also SELINUX=disabled in /etc/selinux/config13:16
T42<birdzhang> seems it run in sandbox, i'm confused13:17
deathmist enrolling in settings app works after "rpm -e jolla-devicelock-daemon-encsfa jolla-devicelock-plugin-encsfa && zypper in sailfish-devicelock-fpd" ^^13:17
deathmistand unlocking works too :D13:19
rinigusDeathmist: as a test (and possibly leading to reboot), try to enroll multiple fingers in sailfish settings, one after another13:29
rinigusDeathmist ^13:30
zinstackI'm back. And now with a revelation. Graphics are still dead, but now there're some logs to back it. Somebody willing to look? Journal: Logcat: 1) 2) I think there's a problem with android part, specifically disable_services.rc, could someone confirm?13:31
deathmistrinigus: I just re-added the 1st fingerprint and authenticated to add the 2nd one using the PIN as well, no reboot after a minute or so, how long should it take if I'm affected=13:32
rinigusDeathmist: its adding fingers one after another without leaving settings and using pin. if you are affected, you will not be able to enroll, device will freeze and reboot. if it enrolls, you are fine13:33
rinigusDeathmist: try to add 3,4,513:33
deathmistcan't add more than 5 it seems13:35
deathmistno reboot still though even while staying in settings and adding new ones while authentication with PIN13:35
rinigusDeathmist: very good. let's hope most (all) will be the same13:36
piggzso, at what point do we support more devices than sailfishx? :D13:37
riniguspiggz: I suspect we support all sailfishx devices as well13:39
zinstackCan someone guide me on what android services to disable? Or what should I do, if disabling stuff won't help me?13:42
deathmist ^^13:42
vakkovIs this channel still alive :P sledges, where are you13:44
piggzdeathmist: very nice13:45
piggzvakkov: very much so13:45
vakkovawesome, what devices have you guys been working on lately13:45
piggzme: mido, pinephone, volla, pro1 and latte13:46
piggz5 is a bit of a stretch tbh!13:46
zinstack@piggz latte? That thing lays on my table rn. It's a dead horse, but I can be a tester if that's needed13:47
piggzzinstack: no need for testers, but feel free to use it, i find it quite nice, and its a native port13:49
zinstack@piggz Can you point m to it? The one I know is very old, like 3.0 or something13:50
deathmistpiggz: feel free to ping me if you need to test other things with FP, I'd like to release an update to my port when this is somewhat "stable" and ready at OBS :D13:53
piggzzinstack: you can just OTA to 3.313:58
zinstack@piggz Never considered it could do that. Will flash it this evening, perhaps.14:00
piggzzinstack: CI builds are on porters ci
*** redj_ is now known as redj14:02
T42<elros34> zinstack: this FAILED_TRANSACTION in logcat could be because of some binder kernel changes which you need to revert. There were other devices which had to do this but I have no example, you need to search in logs14:13
T42Coltor %lastname% was added by: Coltor %lastname%16:46
T42<ankaos> whatsup work?17:12
T42<ankaos> Do you have WhatsApp?17:12
riniguspiggz: PR submitted. please read it through, test, and see if it is what expected17:17
piggzrinigus: ok, proper review of this one!17:32
riniguspiggz: yes, please. No rush as its relatively large17:34
piggzrinigus: building now17:36
riniguspiggz: for test, I removed all the corresponding folders and rebooted. Also cleaned init17:39
piggzrinigus: functionally, it still works18:09
riniguspiggz: that's a good start :)18:16
piggzmal: around?19:05
piggzmonich: any experience with ofono on mediatek ?19:16
monichpiggz: it may be tricky, there are different flavors of mediatek ril interfaces, but it's doable19:51
monichwith some luck, socket-based ril could work (almost) out of the box19:52
piggzmonich: i have a problem ill show u the radio log....19:53
piggzthekit believed this commimt fixed it, but there may be other cases
monichpiggz: do you have vendorDriver=mtk in ril_subscriptions.conf?19:55
monichif not, add it, that would enable some mtk specific magic19:56
monichlike this initial attach apn and not only that19:56
piggzmonich: i dont...will do19:56
T42<NotKit> @monich [piggz: do you have vendorDriver=mtk in ril_sub …], does that work with binder interface as well?19:56
T42<NotKit> I tried to enable it before but it seemed to break things for me19:57
monichmtk always requires some luck19:57
piggzmonich: im sure i can count on you to fix it for me if i supply the logs ;)19:58
monichfor binder, an additional component is probably required19:58
monicha plugin into a plugin19:58
monichint oa plugin19:58
T42<NotKit> the problem with MT6763/MT6771 is that they use setDataProfiles to store APN information on modem and will refuse to activate non-configured in advance APN19:59
T42<NotKit> setInitialAttachApn seems to reset APN configuration from setDataProfiles by creating a new APN with ia-only type19:59
monichyeah, setDataProfiles rings a bell...20:00
piggzmonich: so, some firther coding is required for use with binder?20:02
T42<NotKit> I personally think we encounter a bit different problem from the one previously solved by mtk plugin20:03
monichNotKit: did you try useDataProfiles=true?20:03
T42<NotKit> yes, that helps it, but my commit tries to hack it to issue setDataProfiles after setInitialAttachApn20:04
T42<ankaos> I cannot see incoming calls.20:04
T42<ankaos> how fix this problem?20:05
monichpiggz: I'm looking at mtk specific binder component, it's not in a public repo, but it has bsd license, I don't see any reason why I can't push it to the open :/20:06
piggzmonich: thx, happy to test against this device, and provice any feedback to improve it20:06
piggzrinigus: merged, can i close 17 and 5 ?20:11
riniguspiggz: excellent! 5 for sure; 17 we can close and later reopen when multiuser support will arrive to sfos.20:13
monichNotKit ^20:14
monichthis is basically a replacement for ofono-ril-binder-plugin20:14
riniguspiggz: can you make the releases then? any version you decide to put is fine by me (0.1 or 1.0, for example). releases would be great for the daemon and test software. I have included dependency on test into sony tama as well20:16
piggzrinigus: sure ... are you still getting reboots?20:16
monichpiggz: that plugin requires transport=binder-mtk in /etc/ofono/ril_subscription.conf20:17
monichand useDataProfiles=true20:17
piggzmonich: building :)20:17
riniguspiggz: yes, I am. and I am ignoring them too. will just suggest users to use -test app which works perfectly (no reboots)20:17
piggzrinigus: so, test app can enroll > 1 finger?20:17
riniguspiggz: yes, test app can do as many as it supports on device20:18
monichpiggz: actually, I'm not sure about useDataProfiles=true, it may depend on your ril20:18
monichanyway, that's what I saw working on at least one mtk phone20:19
monichover binder20:19
piggzmonich: i got:20:27
piggzJun 04 21:26:03 Phone ofonod[1010]: Initializing MediaTek RIL binder transport plugin.20:27
piggzbut, no signal icons20:27
piggzmonich: yes, thats what i have20:29
piggzradio log implies radio is off20:29
monichwell, then i need to see ofono logs20:29
monichfrom the console, with -d20:30
monichpiggz: I would also suggest that ^20:31
monichand that:20:31
monichand that too:20:31
monichall those in [Settings]20:32
monichi.e. for all slots if you have more than one20:32
piggzyes, 220:32
piggzmonich: i already have:20:34
monichlooks ril backend is not there20:34
monichpiggz: build this app:
monichand see what's there:20:36
monichbinder-list -d /dev/hwbinder20:36
monichiirc that plugin is looking for
monichthe last part (slot1 and slot2) is coming from ril_subscription.conf20:38
piggzmonich: i think my version is 3.020:39
monichpiggz: well, then you're out of luck20:40
monichthen you better try to get the standard ofono-ril-binder-plugin working20:41
monichat least is there20:41
monichmediatek seems to rewrite everything for each phone20:42
monichqcom is so much more stable as a platform20:42
piggzmonich: ok, so, we are back tothe original problem, of it causes APMproblems20:42
piggzrinigus: so, maybe changes are required to the binderplugin? which add the vendor=mtk fixes?21:00
piggzs/rinigus/monich :D21:01
monichpiggz: iirc vendor=mtk is only useful with socket-based mtk rils21:04
piggzmonich: yes, thats what i mean .... changes would be needed to binder plugin?21:04
monichpiggz: I'm afraid that each android 8+ (binder based) mtk phone would require its own adaptation like this ofono-ril-binder-mtk-plugin thing21:06
piggzmonich: how best to fx the issues we have?21:07
monichif makes you guys happy and it doesn't break anything21:09
monichI could apply it21:10
monichsubmit a patch to
monichI would need to test it first against other devices21:11
monichit's not like I had any particular reason to set profiles before initial attach apn21:12
monichit could well be a wrong order21:12
monichso, go ahead and submit a patch21:12
monichpiggz, NotKit ^21:13
T42<NotKit> piggz still seems to have issues even with that, so I'd prefer to figure out first what else may be wrong21:15
piggzmonich: yeah, even with that patch im still having issues, so, @notkit can i work with you to debug? ofono/ril isnt something im used to debugging21:25
T42<NotKit> ok, I need to find patched lib for more verbose AT commands log21:36
T42S %lastname% was added by: S %lastname%22:08
T42<ankaos> dmesg -w output23:08
T42<edp_17> Somebody can take a look to my "logcat -b radio", please? The ofono service is masked. I am trying to fix rild, any help is appreciated.
T42<ankaos> @edp_17 [Somebody can take a look to my "logcat -b radi …], rm -rf /etc/systemd/system/ofono.service and ofono.service enable23:12
T42<ankaos> Is there a compilation instruction for cm-12.1?23:31

Generated by 2.17.1 by Marius Gedminas - find it at!