Tuesday, 2026-05-26

nightishamanmal: The file that the build script expects at out/target/product/panther/kernel is that supposed to be the lz4 Image or vmlinux?00:17
nightishamanOkay, that is weird. Then I go into init_enter_debug2, then pinging 192.168.2.15 doesnt work. When I remove, it does work.01:58
T42<nightishaman> Also apparently it doesn’t call the mountpoints at /system and co anymore02:26
T42<nightishaman> When booting into systemd02:26
T42<nightishaman> If someone can take a look at it, here: https://github.com/maltesermailo/droid-config-panther02:27
*** ggabriel is now known as Guest1012404:12
*** ggabriel is now known as Guest1013005:14
T42<elros34> check journal to see what fails, maybe dmsetup06:19
T42<elros34> you have wrong descriptions in mount units so they will be confusing when reading journal06:21
*** amccarthy is now known as Guest1013809:20
*** amccarthy_ is now known as amccarthy09:20
T42<nightishaman> @elros34 I can’t look since telnet still doesn’t work. I only have kernel.log. I copied those mountpoints from FP4 repo and edited them to suit my mountpoints but the rest is the same13:12
T42<nightishaman> Later when I’m home, I’ll take a look at telnet again, I’d appreciate if you help me13:13
Mister_Magistermal: Do you have a second to talk about Lord and Savoiur ofono-binder-plugin-ext-qti15:15
Mister_Magisterbasically I would love it if you could explain to me what exactly ofono-binder-plugin connects to (i take it it connects to rild over binder) and what in the world does ext-qti do15:16
Mister_Magisterbecause i feel like if i am to debug it I first need to understand it15:16
malext plugin connects to binder also, just different interface15:26
Mister_Magisterbut connects to ril?15:28
Mister_MagisterI also don't see any binder connecting in the ext-qti at all15:29
rinigusb100dian: we could build new images soon. maybe we can squeeze in also usb changes? I'll test them a bit15:39
T42<b100dian> rinigus: please do. There are changes in nagara config main already, I made a pull request only to switch from extcon to android detection17:14
T42<elros34> @nightishaman I am not sure I am able to help you with that but anway I do not see masked usb-moded in your repo. Did you mask it only locally? If you enabled persistent journal (hadk-faq) then you can grab journal log via recovery17:26
rinigusb100dian: so far it works nicely (charging and dev mode).17:31
rinigusmtp - have to check if issue is mainly on my pc17:32
T42<elros34> @nightishaman btw you have added Provides: usb-moded-configs but without adding any config files17:39
rinigusb100dian: mtp works as well :)17:43
rinigusb100dian: I will cleanup droid-hal-nagara at OBS (remove old droidmedia) and make sure devel is copied to testing17:48
T42<b100dian> rinigus: awesome, tagged17:48
T42<b100dian> yes, some of those droidmedias were not successful ;)17:49
rinigusso, we could ship it all to testing. will start with OBS17:49
malMister_Magister: which interface does your device use? hidl or aidl?17:54
malyou can check that from binder-list -d /dev/binder and same for hwbinder17:54
Mister_Magisterhidl17:55
rinigusb100dian: devel and testing configs triggered, droidmedia only latest. should be ready soon17:55
nightishaman@elros34 im home, i dont have a device specific config, should i just remove it and let the auto detection do it?18:05
nightishamanalso usb-moded isnt masked currently, but was before and it didnt work18:05
T42<b100dian> what's the latest release, 5.0.0.76?18:12
nightishaman5.0.0.71 should be18:12
T42<adampigg> i thought it was 77 :D18:14
T42<b100dian> theres a 77 , I was only off by one ;)18:14
T42<adampigg> its common in computing (re @b100dian: theres a 77 , I was ...)18:15
nightishamanin christianity 77 is also a holy number18:15
nightishamanthe holy sailfishos release18:15
Mister_Magisteris it?18:15
Mister_Magisterwe already had pope sfos release18:15
nightishamanyeah 7 and 77 are regarded as holy numbers, the second being the amount of times you should forgive someone according to jesus18:16
T42<b100dian> the next realease will be dope, though18:16
Mister_Magisternever heard about it18:17
nightishamanwell guess why a week has 7 days18:17
T42<b100dian> then there's 777, the number of the AAA battery18:18
T42<b100dian> (sorry, getting offtopic too much)18:18
T42<elros34> @nightishaman 5.0.0.71 is ancient release form last year. If you don't know what config you need then yes better remove that provides and use default18:18
T42<b100dian> rinigus: started to build 5.0.0.77 testing image18:19
T42<b100dian> hmm now I reckon I should've added the github tag + script that piggz uses18:21
Mister_Magister777 and sfos does kaching sound on boot18:24
rinigusb100dian: great! I guess I should make one for 1iv as well18:34
T42<b100dian> rinigus: I created the release on github18:40
T42<b100dian> feel free to edit18:40
rinigusvery nice - now we just need to add ZIP files and we are good to go.18:41
Mister_MagisterI've put on my big boy pants and loaded entire vendor into ghidra, wish me luck19:18
T42<adampigg> you can automate that :) (re @SailfishFreenodeIRCBridgeBot: <rinigus>very nice -...)19:33
T42<adampigg> is zgrep command available anywhere?19:34
rinigusNo need, our releases are so rare that there is no point19:36
Mister_Magisteri'm trying to figure the ID's that vendor is using19:47
Mister_Magisteri'm getting nowhere19:54
T42<b100dian> rinigus: the point is to be able to make releases without you PlatformSDK running computer ;)20:00
nightishaman@elros34 removing usb-moded configs and upgrading to release 5.0.0.77 gave me telnet access now but only when running init_enter_debug220:01
nightishamanok after running echo "contine">/init-ctl/stdin the telnet dropped20:04
Mister_Magisternightishaman: try masking usb-moded20:05
nightishamanokay, telnet works now even booting past with continue20:17
nightishamantho i get these errors20:17
nightishamanhttps://pastebin.com/TBTRGcCL20:17
nightishamanmainly its this: CANNOT LINK EXECUTABLE "/sbin/droid-hal-init": library "libkeyutils.so" not found: needed by main executable20:19
nightishamanand this: May 26 23:15:29 Pixel7 sh[2301]: /bin/sh: can't open '/usr/bin/droid/droid-hal-early-init.sh': No such file or directory20:19
nightishamandmsetup works and the mounts are mounted20:21
Mister_Magisterthe can't open you can ignore, it's optional20:23
nightishamanokay but the missing shared library is a real problem20:26
Mister_Magisterdid you try grepping the logs20:29
Mister_Magisternightishaman: are you sure every partition is mounted? check if the lib exists anywhere on the device, and if it doesn't perhaps something is not mounted like /odm or something20:30
Mister_Magisterhm ril apparenlty has some logging, how i wish i could enable it20:32
Mister_Magisteri'll try stracing qcrilNrd20:39
nightishaman~ # find . -name "libkeyutils.so"20:40
nightishaman./usr/libexec/droid-hybris/system/lib64/libkeyutils.so20:40
nightishamanthe file exists20:40
nightishamanit also says in the log: linker: Warning: failed to find generated linker configuration from "/linkerconfig/ld.config.txt"20:41
Mister_Magisterls /linkerconfig?20:41
nightishamanits empty20:41
Mister_Magisterhere's your issue20:41
Mister_Magistergrep logs for linkerconfig and find my name i was fixing that20:42
Mister_Magisterit was 10th april 202620:43
Mister_Magisterpiggz's logs have no 2026 data :(20:43
T42<Mister_Magister> @nightishaman (re @SailfishFreenodeIRCBridgeBot: <Mister_Magister>it'...)20:44
Mister_Magisterhere is what you need20:44
nightishamanMister_Magister, it isnt linked to any message in Telegram for me20:45
Mister_Magisterlol weird, then go to that date simply20:45
Mister_Magisteror ctrl+f in telergram20:45
Mister_Magisterlong story short do you have /usr/libexec/droid-hybris/system/bin/linkerconfig20:46
Mister_Magisteralso which android base is this20:46
nightishamanandroid 16 or hybris 23.220:47
nightishamanlinkerconfig exists20:47
Mister_Magisteris hybris 23 even a thing20:47
Mister_Magisterit must be lol20:47
Mister_Magistercheck logs for linkerconfig20:48
nightishamani did cat linkerconfig and now i have an apple logo in front of my command line lmao20:48
Mister_Magisterlol20:48
nightishamanok but commands in computese i didnt want20:48
Mister_Magisterin my case linkeconfig program was missing20:48
nightishamanhttps://ibb.co/mFNmX9Bc this is journalctl in computese20:50
Mister_Magistermaybe just disconnect and reconnect?20:51
nightishamani had to reopen the terminal20:51
nightishamanit fucked the entire terminal20:51
nightishamanokay, now back to linkerconfig20:51
nightishamanthe only things related to linkerconfig are these21:00
nightishamanMay 26 23:54:44 Pixel7 systemd[1]: linkerconfig.mount: Directory /linkerconfig to mount over is not empty, mounting anyway.21:00
nightishamanMay 26 23:54:44 Pixel7 systemd[1]: Mounting Droid mount for /linkerconfig...21:00
Mister_Magisteryou can verify with mount command if its mounted21:02
Mister_Magisterwhat happens if you just run linkerconfig program manually then21:02
nightishamanbin/sh: /usr/libexec/droid-hybris/system/bin/linkerconfig: not found21:04
nightishamanfile /usr/libexec/droid-hybris/system/bin/linkerconfig21:04
nightishamanusr/libexec/droid-hybris/system/bin/linkerconfig: ELF 64-bit LSB pie executable, ARM aarch64, version 1 (SYSV), dynamically linked, interpreter /system/bin/linker64, BuildID[md5/uuid]=3d3c8681cc6b475e82c70f443fd4c2b3, stripped21:04
nightishamanlinkerconfig's interpreter is not found because /apex isnt found21:05
Mister_Magisterwell time to go into rabbithole and figure why /apex is missing21:07
nightishaman-/system/bin/linker64 resolves to /apex/.../linker6421:07
nightishamani dont know even where apex is from21:08
Mister_Magisterhere /apex is tmpfs21:08
Mister_Magisterparadoxically I think it's droid-hal-init that mounts it21:09
nightishamanlinker64 -> /apex/com.android.runtime/bin/linker6421:09
Mister_Magisteryeah if its not mounted then dhi mounts it21:10
Mister_Magisterso that's not it21:10
Mister_Magisterso go back to figuring out why dhi doesn't run21:10
Mister_Magistercan you strace it? you need to do some stuff before running it manually but its doable21:10
Mister_Magisterdo the things in /usr/bin/droid/droid-hal-startup.sh and then strace dhi21:11
nightishamanhttps://pastebin.com/sryiyF4u21:13
Mister_Magisteryou can see the paths it's chekcing21:15
Mister_Magisterhmm21:15
Mister_Magisteryou sure you don't have it anywhere else on the device? the .so21:16
nightishamanits in ./usr/libexec/droid-hybris/system/lib64/libkeyutils.so21:16
nightishamanbut it doesnt check there21:16
Mister_Magisteryes but anywhere else?21:16
nightishamanno21:16
Mister_Magisterdo you have /odm mount/partition?21:17
nightishamanstock has no odm21:17
Mister_Magisterdang21:17
Mister_MagisterLD_PRELOAD=/usr/libexec/droid-hybris/system/lib64/libkeyutils.so droid-hal-init lmao21:17
nightishamanmaybe ask mal?21:17
Mister_Magisteryou just did21:17
Mister_Magisteri'm way out of my depth21:17
nightishamanalso is it normal that my device gets hot while running telnet21:18
nightishamanits about 50 degrees celsius hot21:18
Mister_Magistersome stuff might be bootlooping or smth21:18
Mister_Magisteryou can check top21:18
Mister_Magisterbut i'd say until anything works, expect nothing21:18
nightishamanweird, libkeyutils.so exists on the vendor boot at the location where it would look21:19
nightishamanproblem is, the system mount overrides /system/lib6421:20
Mister_Magistertime to do overlayfs with vendor boot lmao21:20
Mister_Magisteror just for testing you can put it in odm location since your device doesn't use it21:21
Mister_Magisterjust as a temporary going forward thing21:21
Mister_Magisteryou can figure it out later (when m.a.l wakes up)21:21
nightishamanhttps://pastebin.com/WszELp0q21:49
nightishamannext error21:49
Mister_Magisterwhat in da world21:50
maltoo old dhd submodule to not copy linkerconfig?21:53
Mister_Magisterbut the executable is there21:54
mal?21:54
Mister_Magisteron device21:54
Mister_Magister /usr/libexec/droid-hybris/system/bin/linkerconfig exists21:54
malso apexd is failing then?21:54
Mister_Magisteryes21:54
Mister_Magisteri thought dhi mounts /apex21:54
Mister_Magisterdoes something else mount it?21:54
maldoes /apex folder exist?21:55
Mister_Magister[23:05] <nightishaman> linkerconfig's interpreter is not found because /apex isnt found21:55
Mister_Magisterthink not21:55
malthe build should have created that21:56
Mister_Magisterbuild?21:56
maldroid-hal packaging21:57
Mister_Magisterah21:57
Mister_Magisterbut what does mount and populate /apex?21:57
malhttps://github.com/mer-hybris/droid-hal-device/blob/master/droid-hal-device.inc#L77421:57
malapexd should be handling it21:57
Mister_Magisterand apexd is ran by systemd?21:57
malno, dhi21:57
Mister_Magisterand dhi is not starting21:58
malwhich android base if that again?21:58
Mister_Magistera1621:58
Mister_Magistercurrently you can forget about apex because dhi is not starting but if dhi is not starting because lack of apex now that's real chicken-egg problem xd21:58
Mister_Magisterdhi linking fails21:58
Mister_Magisterfirst there was missing libkeyutils.so so i told nightishaman to copy it from /usr/libexec/droid-hybris/system/lib64/libkeyutils.so to odm just as a temporary fix but now something else fails to link21:59
Mister_Magisterwhich he posted here [23:49] <nightishaman> https://pastebin.com/WszELp0q22:00
malthose linker64 links are not the issue, those are like that on working device also22:00
Mister_Magisterwell yeha22:01
Mister_Magisterbut those are not the issue xd22:01
malwhat has those symbols?22:01
malwhich are missing22:01
Mister_Magisteryou have them in the pastebin link22:02
Mister_Magisterohh22:02
Mister_Magisterwhat has them22:02
Mister_Magisterwell that's a good question now22:02
Mister_Magisternightishaman: i think you could grep -R / for them?22:02
malare all partitions mounted correctly?22:02
Mister_Magisterthink so?22:02
malI mean dynamic ones22:02
Mister_Magisterthats what he stated22:02
Mister_Magister[22:21] <nightishaman> dmsetup works and the mounts are mounted22:03
maland system is mounted correctly?22:03
malwould be good to check content of /system/22:03
Mister_Magisteri'm just stating what i've managed to learn xd22:04
Mister_Magisterit is mounted i think but the libkeyutils.so apparently is in /system location but on vendor_boot he said22:04
mallibkeyutils.so should exist in /system/lib64/ at least it does on an a16 device22:04
Mister_Magisterso /system mount overwrites it22:04
Mister_Magisterand then its missing22:04
Mister_Magisterthat's what i heard22:05
malyeah, something broken in mount setup22:05
Mister_Magisterwhich then i proposed jokingly to overlayfs vendor boot xd22:05
Mister_Magisteryou're mostly up-to-date now22:06
Mister_Magisterbut our guy is seemingly gone22:06
Mister_Magisterbtw mal since you're here, any tips on how in the world did you/jolla extract the specification of vendor.qti.hardware.radio.ims interfaces as in on what number what indication is being sent?22:10
Mister_MagisterI've thrown entire vendor into ghidra trying to figure stuff out and i figured exactly nothing22:14
Mister_Magisterguess not then22:16
nightishamanim back in 10-20 minutes22:17
Mister_Magisterone interesting thing I noticed is that notifyIncomingImsSms is only on interfaces 1.0 1.4 1.5 1.6 1.7 1.8 1.9. 1.1022:19
Mister_Magisterit is not present on 1.222:19
Mister_Magisterwhich would make sense as to why receiving sms only works on 1.0 i guess22:19
Mister_Magisterand might also explain why user in america cannot receive sms (because 1.0 implementation is way diff from the ones higher)22:19
Mister_MagisterI'll try to tell qti to register with 1.4 version and maybe handle that singal22:20
Mister_MagistersendImsSms is only on 1.0 and 1.522:21
Mister_Magisterbut it doesn't work for either of us22:21
Mister_Magisterso maybe i should bump it straight to 1.5, but idk how to figure out what int i should use for the signals and stuff22:22
Mister_Magisterthere's some qti ril logging i'd love to see enabled22:23
malMister_Magister: would newer interface help if you already had mismatch in older one? wouldn't the new ones have some offsets again22:25
Mister_Magistermal: sure but it also might use some different implementation and actually work22:26
malstill very odd why they have some patched interface22:26
Mister_Magisterwell it's qualcomm vendor there could be anything in there22:27
nightishamanshould i mount system.img on host device or look at telnet?22:32
nightishamanmal: The stock system.img has no libkeyutils.so22:39
nightishamanim gonna check lineageos what it has and if it has more, then i flash the partitions with lineageos22:41
malyou were using super partition from stock android and not from lineage?22:43
nightishamanmal: not super, the stock update contains a system.img, system_ext.img and co22:44
nightishamani flashed all of those manually22:44
nightishamanim gonna flash lineageos then22:46
nightishamanit has libkeyutils22:47
nightishamanno idea why stock doesnt22:47
Mister_Magisterbecause you're running on top of lineage and not stock22:50
nightishamanu mean the port is running on top of lineage22:51
Mister_Magisteryes22:51
Mister_Magisterif it was using stock sources it wouldn't need it22:51
nightishamancant, the android repository that u use to initialize hybris is using everything from lineageos22:52
Mister_MagisterI didn't suggest you do22:52
Mister_Magisteri'm just saying don't use stock partitions22:52
Mister_Magisterif port uses lineage = you need lineage22:53
Mister_Magisteryou actually can use AOSP sony does that and fxtec also does22:53
nightishamanI added product.mount to the mountpoints, do I only need to do build_packages.sh --configs or more?22:53
Mister_Magisteryou ned to rebuild droid-hal22:54
Mister_Magisteror do you mean you added product.mount to sparse in droid-configs22:54
nightishamanin sparse, yes22:54
Mister_Magisterthen yes only configs22:54
nightishamani also modified all sources to use tar.gz instead of tar.bz222:55
Mister_Magisteri just build .img22:56
Mister_Magisterno need to worry about bz2 or gz or encryption22:57
nightishamanwell, i just extract tar.gz too but when releasing it later to community, it would be cool to have the update process working22:57
Mister_Magisteryou misunderstand22:57
nightishamani dont understand what you mean, i mean the image that is created by mic22:58
Mister_Magisteryes and i said insted of creating .bz2 or .gz archive i create .img22:59
Mister_Magisterand just flash the partition22:59
Mister_Magisterno need to worry about extracting or encryption22:59
nightishamanyou mean you flash directly to userdata?22:59
Mister_Magisteryes22:59
nightishamanhow22:59
Mister_Magistermy innovative solution https://github.com/VerdandiTeam/droid-config-pipa/blob/master/kickstart/pack_package-droid-updater23:00
Mister_Magisteri extract .tar on host into image and then just flash image and resize filesystem xd23:00
nightishamanhmm, the only downside of that is, that it deletes existing userdata when someone wants to upgrade OS23:00
Mister_Magisternormal zip also does that so what's the difference23:01
Mister_Magisterhttps://github.com/mer-hybris/hybris-boot/blob/master/updater-unpack.sh#L623:01
Mister_Magisteryou don't upgrade sfos by flashing zip23:01
nightishamanMister_Magister: https://pastebin.com/Q4KMK0vN do you know how to fix or this i should i ping ma23:13
nightishamanah nvm, i know why that happens23:13
nightishamangotta fix something real quick23:14
nightishamanokay, made sure that my load_kernel_modules doesnt do those 2 anymore, so droid-hal-init can do it23:18
nightishamanMister_Magister: this is the most ugliest method i used to load the kernel modules, since google only tested them with the specific order that android init does and loading them using the modules.load from stock caused the kernel modules to hang mid load. so i went and compiled android init with just the code to load the kernel modules in the right23:19
nightishamanorder23:19
Mister_Magisteryou could just make simple shell script that modprobes them in your desired order no?23:25
nightishamanwell, then id have to extract the order from the log files and maybe it would've worked, but i like freaky solutions23:27
Mister_Magisteron my first port, the fix for gui not working was to first start surfaceflinger, then kill it, then start lipstick, so yeah (it still kinda is)23:27
Mister_Magisterwe never figured why it didn't work23:28
Mister_Magisterto this day (11ish years later and yes port is still supported)23:28
nightishamanwhen my usb cereal arrives on the end of the month, im gonna go debug and fix the mess google made23:28
Mister_Magistercereal?23:29
Mister_Magisterlike a snack with milk?23:29
nightishamanits a usb c tool that routes the uart console from the sbu contacts on the usb c port23:29
Mister_Magisteror did you mean S in USB23:29
Mister_Magisterso you did mean S in USB23:29
Mister_Magisterit's Serial, not cereal23:29
nightishamannope23:30
nightishamanits called usb cereal https://github.com/google/usb-cereal23:30
Mister_Magistercereal is food23:30
Mister_Magisterbecause someone made that exact joke23:30
nightishamanyeah well android 5 was called lollipop23:30
Mister_Magisterbecause they sound the same23:30
Mister_Magisterit's the joke on serial xd23:30
Mister_Magisterit's actually pretty funny23:31
nightishamanat lost the coders at google are funny, because they certainly cant code kernel modules23:33
nightishamanokay, maybe i read the log wrong and i just found it out23:42
nightishamani cleared my build cache since i thought my tool was still doing the stuff droid-hal-init is supposed to do, but then i thought that droid-hal-init does mknod which can only be done once per system init so looking at the log for the 3rd restart attempt was not a good idea23:43
nightishamani should've looked at the first one and why it failed there23:43
nightishamanalso means i can get bonus points for fixing the droid-hal-init to allow it to do multiple times23:44
nightishamanhttps://pastebin.com/ywUFt7BL23:53
nightishamanthis is the original error23:54
nightishamanim really confused23:55
nightishamanhttps://pastebin.com/pg6gYJFr23:55
nightishamanvendor didnt mount23:55
nightishamanno idea why23:55
nightishamanit says dependency23:56
nightishamandmsetup fails23:58
nightishamanwith this23:58
nightishamanMay 27 02:58:16 Pixel7 bash[1809]: device-mapper: reload ioctl on dynpart-product_b  failed: Invalid argument23:58
nightishamanMay 27 02:58:16 Pixel7 bash[1809]: Command failed23:58

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