Sunday, 2025-07-20

rinigusb100dian: you have tested camera2 droidmedia. does it work for nagara? can we choose cameras with it?08:21
rinigusI guess it is the only functionality that's missing with cameras08:22
rinigusapart of changing optical zoom on 1iv08:22
rinigus... and motion compensation by camera, probably due to fake sensor manager in droidmedia08:29
T42<b100dian> It doesn't work yet, need to see what the new logs say, after the fake package_native development.09:01
T42<b100dian> No viewfinder, no list of cams so early failure probably09:01
rinigus@b100dian: which branch did you use? camera2 in droidmedia? that one is missing few fake services that were added recently. I am trying now to check whether merging camera2 with master branch would give better results.09:08
*** Daanct12 is now known as Danct1209:30
T42<elros34> I use 'setprop ctl.stop vendor.fps_hal' to stop android service. 'getprop | grep fps_hal'' shows it "stopping" but nothing happens, service is running and status is never set to 'stopped'. Is there anything I am missing here?09:44
rinigusb100dian: indeed, with camera2 something else is going on. getting logs similar to yours10:40
riniguswill revert for now for camera1 and start working on ofono volte plugin (aidl option)10:42
T42<elros34> @edp_17 did try something like that? https://paste.opensuse.org/pastes/1261585c7e1411:35
T42<b100dian> rinigus: it wasnt merged with latest changes was gonna test that now again.11:43
rinigusb100dian: I tried with the merged changes and it didn't work for me. logcat was the same as yours at https://github.com/sailfishos-sony-nagara/main/issues/34#issuecomment-309125559411:44
rinigusbut give it a try - maybe you are luckier :)11:45
T42<edp_17> @elros34 : I've tried this: https://paste.ubuntu.com/p/BFSNRDQVVX/12:52
T42<edp_17> But I'll try that you suggested12:52
T42<elros34> first you should build defconfig12:53
T42<edp_17> Oh, okay. I've executed both and got this: https://paste.ubuntu.com/p/nGNzmDVcJn/12:54
malrinigus: I'll probably rebase camera2 and push it, need to see if I can fix some issues in it also12:59
malrinigus: btw, I think the packagen_native issue is exactly the same that I had, and it's doable also via property, not sure yet which is better way13:06
rinigusmal: looking forward to your rebase with camera2.13:08
rinigusas for which way is better - I think it makes sense to be as close to android as possible. but its close to more philosophical discussion and maybe its better not to go there :)13:10
malrinigus: in case you are interested check frameworks/av/media/libstagefright/MediaCodec.cpp, package_native use requested by connectFormatShaper which is called from createMediaMetrics via shapeMediaFormat and controlled via "debug.stagefright.enableshaping" property13:11
malbut since that is a debug.* property not sure if it's a good thing to use13:12
T42<elros34> @edp_17 is this right after running make hybris-hal?13:16
T42<elros34> ah typo, see missing space13:17
T42<elros34> more than once13:17
T42<elros34> and clean backports directory first, I remember building defconfig without proper ARCH can cause issues13:19
rinigusmal: interesting - those features in that source are exactly the ones that are called during video recording. as seen from the other side in package_native debug logs :)13:30
malyes, so the issue is not limited to camera2, same is also in older camera13:34
malI will do some investigation a bit later today13:35
rinigusindeed. I had it with older camera API13:36
malif we go with fake package_native it would go into droidmedia like the others13:39
malI have been thinking a bit or reworking the fake services in droidmedia, it's a bit stupid now13:40
mal*of reworking13:40
rinigusmal: then it has to be written using android classes. I have such version too which I did before libgbinder version. although, it wasn't processing any requests and this will have to be added (trivial to do)13:41
riniguslet me push it somewhere13:42
malyeah, those are usually not difficult to do with android classes, I have done so many of those that it shouldn't take long13:42
rinigusmal: here you go - https://github.com/rinigus/fake-package-native-android13:47
malthanks15:10
T42<A_T_R> HI , my sfos port version 4.6.0.15 is working fine now , i wish to test my obs and for testing latest sfOS version 5.0.0.62 or any version ,what are the steps to follow ?15:30
T42<A_T_R> here https://build.sailfishos.org/user/show/iAboothahir15:30
malwhich device?15:34
T42<A_T_R> asus max pro m1 ,codename x00td (X00TD)15:34
maldo the needed changes to https://build.sailfishos.org/project/show/nemo:devel:hw:asus:x00td15:35
malso update the packages to correct versions and make sure the config packages etc are ok15:36
T42<A_T_R> update droid-confifs to a specific sfos version and compile —droid-configs and upload droid-config-*.rpm ?15:37
maldroid-config package is not uploaded, it's built on obs, make sure your droid-config github has correct version and the just trigger build on obs15:38
malI assume this is what you use https://github.com/SailfishOS-sdm660/droid-config-X00TD15:41
T42<A_T_R> yes ,master branch15:41
malis it missing some changes?15:42
malseems to be rather old15:42
T42<A_T_R> yes ,bit old - i am stuck at 4.6.0.1515:42
malso push the changes there15:42
T42<A_T_R> about the submodule ?15:43
T42<A_T_R> https://sailfishos.wiki/link/20#bkmrk-5.0.0.x use this recommended revision?15:43
malyes15:44
maland any other changes you might have locally15:44
T42<A_T_R> as of now droid-config-X00TD locally has no changes15:45
malit might work with current config submodule, not sure, maybe better to update the submodule to correct version15:45
riniguswhen looking at available binder interfaces (aidl and hidl), nagara doesn't have IImsRadioResponse nor IImsRadioIndication. looks like ofono-binder-plugin-ext-qti is trying to use it though15:50
rinigusin logs, I get "BpBinder: Cannot do a user transaction on a system stability binder (vendor.qti.hardware.radio.ims.IImsRadioResponse) in a vendor stability context."15:50
T42<A_T_R> yes bit old15:50
T42<A_T_R> locally 4.6.0.1515:50
T42<A_T_R> droid-config -remote 4.5.0.21 (re @SailfishFreenodeIRCBridgeBot: <mal>seems to be rat...)15:50
rinigushmm, will have to read code a bit more. maybe AIDL API is different...15:51
malrinigus: gbinder_local_object_set_stability(callback_pointer, GBINDER_STABILITY_VINTF); before doing the call to which that callback pointer is given16:01
malso in function doing the "open" binder call16:02
malthat is needed when using certain aidl interfaces16:02
rinigusmal: thanks!!! problem is that I don't see any vendor.qti.hardware.radio.ims.IImsRadioResponse in binder-list (/dev/binder or hwbinder). shouldn't it be there for such calls?16:03
malyes16:03
rinigusbut I can surely try16:03
malwhat do you have in those?16:03
malanything with ims?16:03
rinigusmal: https://dpaste.org/xDcUR/raw16:04
rinigusis there more in xperia 10v/iv?16:05
malrinigus: you have those vendor.qti.hardware.radio.ims.IImsRadio/imsradio* in /dev/binder which are what you are looking for16:07
malthose response things don't show in binder-list, those are part of that vendor.qti.hardware.radio.ims.IImsRadio16:08
malI haven't yet managed to get aidl ims working, I have been able to communicate with it and get responses and indications but no sign of it registering to volte16:09
rinigusthanks! so, I have to find AIDL for vendor.qti.hardware.radio.ims.IImsRadio and then see what would correspond to response...16:09
rinigushmm, if mal hasn't done it - what chance do we have...16:10
rinigusOK, will try to dig into it.16:10
rinigushave you seen by any chance aidl definition for vendor.qti.hardware.radio.ims.IImsRadio?16:10
rinigusnot that I looked for it16:11
malit's just vendor.qti.hardware.radio.ims.IImsRadioResponse and similar for Indication16:11
malI haven't seen any aidl files anywhere for that interface, I figured it out the ugly way16:12
T42<edp_17> @elros34 , you were right, I didn't spot those typos. As I fixed them, the first one has built, but the second failed: https://paste.ubuntu.com/p/Gp2GWFm2nW/16:46
T42<elros34> python2 issue?16:49
T42<edp_17> In HABUILD, the python —version gives me Python 3.8.516:52
T42<elros34> too new16:55
T42<edp_17> Hahaha. :)16:56
rinigusmal: just to make sure that I follow. I have AIDL service vendor.qti.hardware.radio.ims.IImsRadio/imsradio0 . under that, in LOS21, I can see interface vendor.qti.hardware.radio.ims.IImsRadio. so, if I understand you correctly, it also has interface vendor.qti.hardware.radio.ims.IImsRadio and maybe few others that are just not registered with binder. right?17:00
malthere is only that one interface, which inside contain the response and indication stuff17:46
malyou don't see the response and indication stuff in binder-list for any other interfaces either17:47
rinigusno, I don't see indeed.18:06
rinigusmal: in my understanding, what we see listed in binder-list are services. under each service, we can have interfaces. is that correct?18:15
malwell usually the names in binder-list are of the main interface18:21
malafaik18:21
rinigusyes, true. but in just recent case - service was package_native and interface was android.content.pm.IPackageManagerNative18:23
malok18:32
malyeah, I was thinking more of those hw side services18:32
rinigusmal: your tip with GBINDER_STABILITY_VINTF helped. looks to be going bit further. will study logs and read plugin code tomorrow18:47
riniguswe don't happen to have any libgbinder docs?18:47
maljust the api headers on github18:51
maland test code in there also gives some hints about usage18:51
maland of course other projects using it18:51
rinigusyep, was using those. will ask if I get too confused.19:00

Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!