Wednesday, 2025-07-23

T42<b100dian> how to debug flight mode indicator being stuck? It doesn't log anything when I try to click it and subsequent clicks say "in progress"..02:17
rinigusto libgbinder experts: I am confised regarding reading data from AIDL message. for volte, I am trying to parse reginfo response and original implementation of qti_radio_ext_read_ims_reg_status_info is using info = gbinder_reader_read_hidl_struct(reader, QtiRadioRegInfo) which fails with error, probably as sizes don't match (looking at gbinder code).06:40
rinigusthen I tested another implementation and tried to read each structure element separately (commented out code at https://github.com/rinigus/ofono-binder-plugin-ext-qti/blob/aidl/src/qti_radio_ext.c#L324). not sure whether it works correctly06:40
rinigusfinally, calling gbinder_reader_read_parcelable(reader, &size) looks to be giving some data, but then I will probably have to get offsets and parsing right manually, including parsing strings.06:40
rinigusso, questions: how I am expected to read the data given by AIDL, possibly via struct? how I am expected to handle memory allocation, i.e. do I need to free strings, for example?06:40
rinigusmemory allocation seems to be an issue in the case if the message arrived via QTI_RADIO_IND_REG_STATE_INDICATION. in this case, any touching via reader leads to core dump :(. whether it is an error on sender side or anything else - no clue06:40
rinigusmal: if you have time, would you mind to look into the above and help? or maybe point towards someone who can help06:40
*** Mister_Magister_ is now known as Mister_Magister09:26
malrinigus: in aidl things are not read as structs but as the values in it, one by one, also some of the content is parcelables which have a header (two int32 values), some even nested parcelable which is fun10:04
rinigusmal: thanks! and how do you determine whether its parcelable or not? or nested? do you happen to know how IND_REG_STATE_INDICATION and REG_STATUS_CHANGED?10:12
rinigusit looks like we have ability to read in gbinder_reader_read_parcelable... is it the one then we have to use?10:14
malfor example in the java code "public class RegistrationInfo implements Parcelable"10:16
malI didn't use that read_parcelable since it needs the size10:17
maljust look at the code from my other aidl PRs10:18
rinigusmal: OK, so from java/smali code I can figure it out. now do we have any mechanism inside gbinder? btw, size is an output of that function10:18
rinigusmal: nfc? as an example?10:18
malgeoclue hybris, sensorfw or nfc10:21
rinigusthanks!10:21
malhmm, maybe that could be used in some cases but not sure10:21
rinigusmal: looks like geoclue has the details, thanks!10:24
T42<edp_17> @elros34 : Any suggestions on this? https://paste.ubuntu.com/p/vHX9hfyQt4/10:37
T42<elros34> @edp_17 did you run depmod -a?11:21
T42<elros34> check also dmesg, all errors from kernel will be there11:21
T42<edp_17> No, I haven't run depmod -a. Now I did and got a different error: Unknown symbol in module, or unknown parameter11:33
T42<edp_17> dmesg: https://paste.ubuntu.com/p/GhnRCNj3b3/11:33
T42<edp_17> Got an error when building droidmedia 0.20211101.0: https://paste.ubuntu.com/p/KW2z8q9W7C/13:53
T42<elros34> did you flash the kernel?15:37
T42<edp_17> Sorry, I am working on two devices atm. I've built backports, kernel and packages for the S4 and installed kernel on it. That's where I am trying to load the module.15:41
T42<edp_17> On the other hand, I built backports and kernel for Nexus 5 and trying to build packages, including droidmedia.15:41
T42<elros34> installed ok but is it really flashed/used? uname -r really matches what you have built?15:48
T42<elros34> are you sure you have built and copy all modules? I am not sure but shouldn't there be for example rfcomm module15:52
T42<elros34> even dates for modles are different in your paste15:54
T42<elros34> modules*15:54
rinigusmal: after reading it in with those extra int fields, I am getting registration info data parsed. what looks like successful registration leads to state=1. while error is some huge number, uri is provided and has phone number in it17:27
rinigushow is it going for you?17:28
rinigushmm, based on RegState 1=REGISTERED. not 0 as in ofono-binder-plugin-ext-qti17:54
rinigus... and registration works with updated state values17:58
*** user25644[m] is now known as SatoshiNakamoto[m]18:39
T42<edp_17> @elros34 If you remember I asked how those .ko files will go into rpm. I think they should be in kernel-module rpm. However, they were not there. So, what I did was: flashed hybris-boot.img and COPIED those .ko files onto device. I also copied all other kernel modules on device from older folder to that new one (ends with *-dirty)19:33
T42<elros34> what modules do you have enabled now in kernel?19:36
T42<elros34> or even better show whole .config/config.gz19:37
T42<edp_17> Those what the instruction mentions. Config_bt is set to be a module.19:37
T42<edp_17> Will do when I get back to my pc. Thanks!19:38
T42<elros34> but is it really set in config or just in defconfig?19:38
T42<elros34> I do not remember what modules should be there but I have impression that you are missing some of them19:39
T42<elros34> for example "Unknown symbol backport_dependency_symbol" is defined here https://github.com/search?q=repo%3Amlehtima%2Fbackports-bluetooth%20backport_dependency_symbol%20&type=code19:42
T42<elros34> look at fp2 https://build.sailfishos.org/package/binary/nemo:testing:hw:fairphone:fp2-sibon/droid-hal-fp2-sibon?arch=armv8el&filename=droid-hal-fp2-sibon-kernel-modules-0.0.6-202204022246.armv7hl.rpm&repository=sailfishos_4.6.0.15: compat.so, rfcomm.so, bluetooth.so and so on19:57
T42<adampigg> Furiphone : https://irc.thaodan.de/.imgstore/068e90f7/file_7530.jpg19:58
T42<edp_17> Thanks! I'll check .config for those flags.20:55

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