| T42 | <Mister_Magister> maaan, i just grepped for lpm_cpuidle_enter cause "many other people ported xiaomi maybe they encountered this issue" | 01:35 |
|---|---|---|
| T42 | <Mister_Magister> i freaking found myself mentioning that problem | 01:36 |
| T42 | <Mister_Magister> maaan xd | 01:36 |
| T42 | <Mister_Magister> and it's me from… 7 years ago holy | 01:39 |
| T42 | <Mister_Magister> but hey i guessed correctly it was xiaomi | 01:42 |
| T42 | <Mister_Magister> so back in 2018 the issue was issue with busybox not being executable, but i doubt thats the issue since i'm in ramdisk no issues | 01:56 |
| T42 | <Mister_Magister> this is going to be hard | 01:57 |
| T42 | <Mister_Magister> i shuold try and debug switch_root cause it seems to be cause of the issue | 01:59 |
| T42 | <Mister_Magister> its not segfault in kernel, but Attempted to kill init | 01:59 |
| T42 | <Mister_Magister> which would suggest userspace i think…? | 01:59 |
| T42 | <Mister_Magister> if anyone has any idea please help | 01:59 |
| T42 | <Mister_Magister> or maybe its selinux thing | 02:02 |
| T42 | <Mister_Magister> in that case the problem is that the logs are not being saved and i gotta debug userspace… i think i'll try stopping init script from disabling telnet so that i can see what happens | 02:03 |
| Mister_Magister | mal: hi did you have time to look at the repos? | 12:59 |
| mal | sorry was a bit busy yesterday | 13:01 |
| mal | what is the status of that latest build | 13:02 |
| Mister_Magister | nah no worries just asking | 13:03 |
| Mister_Magister | mal: when it switches root it kills init, before executing the init-done | 13:03 |
| mal | hmm | 13:04 |
| Mister_Magister | hmm indeed | 13:04 |
| Mister_Magister | i don't have any logs beside segfault either, currently i'm thinking of modifying init script to not close telnet when switching root and just tail the output | 13:05 |
| mal | so it never gets to second telnet | 13:05 |
| Mister_Magister | no | 13:05 |
| mal | even is you have that stop thingy there | 13:05 |
| Mister_Magister | yeap | 13:05 |
| Mister_Magister | hence i concluded it never gets to init-done | 13:05 |
| mal | and userdata is properly ext4? | 13:05 |
| Mister_Magister | yeah im in ramdisk i have it mounted its all fine | 13:06 |
| Mister_Magister | its quite nice debug setup because it enters ramdisk every time and then when it segfaults it just goes back into ramdisk and i have pstore | 13:06 |
| mal | anything useful in pstore? | 13:06 |
| Mister_Magister | the kernel panic | 13:06 |
| Mister_Magister | other than that nah | 13:06 |
| Mister_Magister | btw mal did they remove the bootctl or something? despite me rebooting way before bootctl service would run it doesn't switch slots like it used to | 13:07 |
| mal | does it hint anything where is panics? | 13:07 |
| mal | btw, do you know the way to decode kernel backtraces to more readable format? | 13:07 |
| Mister_Magister | i know addr2line | 13:08 |
| Mister_Magister | it attemts to kill init and not segfault so i concluded its in userspace | 13:08 |
| Mister_Magister | and the spot is VERY generic | 13:09 |
| Mister_Magister | so i don't think its problem with kernel | 13:09 |
| Mister_Magister | https://paste.opensuse.org/pastes/26e36af8ab10 enjoy the dmesg | 13:10 |
| Mister_Magister | the panic points to CONFIG_MSM_PM so yeah definitely not the kernel fault | 13:12 |
| Mister_Magister | if it was kernel fault it would be segfault i think | 13:12 |
| mal | Mister_Magister: does your device have kernel modules in init? i.e. in vendor_boot? btw, did you flash correct vendor_boot also | 13:18 |
| Mister_Magister | i flashed dtbo, boot and vendor boot | 13:18 |
| Mister_Magister | all 3 | 13:18 |
| Mister_Magister | and as to other question i have no idea | 13:18 |
| mal | because the issue could be "request_module: modprobe binfmt-0000 cannot be processed, kmod busy with 50 threads for more than 5 seconds now" which suggests init hangs in module loading | 13:18 |
| Mister_Magister | im not good in this android stuff i leanred about this weird new kernel stuff yesterday | 13:19 |
| mal | but that would not explain that you can get into first telent | 13:20 |
| Mister_Magister | mal: but like, in ramdisk it also loads kernel modules so why would it fail in init | 13:21 |
| Mister_Magister | yeah | 13:21 |
| mal | because it should be loading those already in first init | 13:21 |
| mal | is the first init stable for long time? | 13:21 |
| Mister_Magister | yeap | 13:21 |
| Mister_Magister | i can stay indefinitely | 13:21 |
| Mister_Magister | and it goes there every time | 13:21 |
| Mister_Magister | very solid debugging environment | 13:21 |
| mal | maybe try to check if there are some pending kernel stuff, can't remember the path in sys to check that | 13:22 |
| mal | did you have any issues extracting rootfs? | 13:23 |
| Mister_Magister | i extracted it via the ramdisk and nah | 13:23 |
| Mister_Magister | https://paste.opensuse.org/pastes/867e4fed3a04 whachu need :D | 13:29 |
| Mister_Magister | i tried chrooting target and got chroot: can't execute '/bin/bash': Exec format error | 13:35 |
| Mister_Magister | that normal? | 13:35 |
| mal | hmm | 13:35 |
| Mister_Magister | when i try running /target/bin/bash nothing happens | 13:35 |
| Mister_Magister | could be entirely normal | 13:36 |
| Mister_Magister | mal: i mostly wanted you to check repos cause i might have missed something but also this https://github.com/sailfishos-sony-nagara/droid-config-sony-nagara/tree/main/sparse/overlays/system/lib, i don't know what it is or why it is, i can't find it in any instructions so i don't have that | 13:42 |
| Mister_Magister | idk if its relevant | 13:42 |
| Mister_Magister | im just biiiig dummy xd | 13:42 |
| Mister_Magister | i left the telnet running and tried tailing /target/init-debug-stderrout but nothing ever appeared | 13:45 |
| Mister_Magister | let me try chrooting from ramdisk on another phone to see if thats supposed to work | 13:49 |
| Mister_Magister | hmm, init_enter_debug sadly doesn't freeze the boot | 13:50 |
| Mister_Magister | well at least i can chroot from recovery on other phone let me try that on the tablet | 13:56 |
| Mister_Magister | can't do it from recovery but the error is not the same | 13:58 |
| Mister_Magister | in recovery it says chroot: exec /bin/sh: No such file or directory | 14:00 |
| Mister_Magister | i should be able to chroot | 14:01 |
| Mister_Magister | mal: huh, ramdisk has 32bit executables but rootfs has 64bit executables | 14:04 |
| Mister_Magister | should it be like that? | 14:04 |
| Mister_Magister | holdon plot thickens, /usr/bin/busybox is empty file | 14:05 |
| Mister_Magister | eyyyyyyyyyyyy i re-extracted the archive and can chroot now | 14:10 |
| Mister_Magister | eyyyyy its not segfaulting now and the screen got brighter | 14:11 |
| Mister_Magister | meaning something's gotta be haappening | 14:11 |
| Mister_Magister | I don't have second telnet but thats probably because of messed up init script so let me clean it up and see how it goes | 14:12 |
| Mister_Magister | lol now it gets stuck in ramdisk | 14:19 |
| Mister_Magister | removed init_enter_debug and now i get init-debug in real rootfs but no ip | 14:21 |
| Mister_Magister | okay i'm in real rootfs | 14:21 |
| Mister_Magister | but now the screen doesn't light up :< | 14:21 |
| Mister_Magister | ah cause it frozen in script obviously | 14:22 |
| Mister_Magister | think dummy | 14:22 |
| Mister_Magister | after i continue from init_enter_debug2, it takes a while but then screen lights up, it kills telnet and rndis, | 14:23 |
| Mister_Magister | i can lock and unlock screen so i think that mce is running and probably the rest but why it kills rndis no clue | 14:24 |
| Mister_Magister | odm fails to mount thats all i know so far | 14:30 |
| Mister_Magister | ahh i've inputted wrong super partition | 14:35 |
| Mister_Magister | okidoki https://paste.opensuse.org/pastes/b16fefc081b6 this suddenly happens in journal like in an instant and in the same moment i lose telnet | 14:38 |
| mal | fix the mounts then | 14:38 |
| Mister_Magister | i've fixed the superparition dmsetup yes | 14:39 |
| mal | fixup seems broken or missing some partitions | 14:39 |
| mal | like that dsp one | 14:39 |
| Mister_Magister | what seems to be missing is entirety of /dev/block/bootdevice directory | 14:39 |
| Mister_Magister | or maybe its created later | 14:40 |
| mal | or maybe the name is different | 14:40 |
| mal | check fstab in sources | 14:40 |
| Mister_Magister | when i was checking the directory was not there | 14:40 |
| Mister_Magister | i've made sure fixup-mountpoints is correct | 14:40 |
| mal | ok | 14:40 |
| mal | but anyway mount services seem wrong, those should not have those path but the device nodes | 14:41 |
| Mister_Magister | is the /dev/block/bootdevice created by systemd? | 14:41 |
| mal | yes | 14:41 |
| Mister_Magister | vendor_a system_ext_a they're all super partitions | 14:41 |
| Mister_Magister | https://paste.opensuse.org/pastes/a8f5d2f6c45a | 14:42 |
| Mister_Magister | yeah the vendor-dsp.mount has What=/dev/block/bootdevice/by-name/dsp | 14:44 |
| Mister_Magister | that… seems wrong | 14:44 |
| Mister_Magister | ohh | 14:45 |
| Mister_Magister | i have dsp_a in the fixup mountpoints | 14:45 |
| Mister_Magister | i've grepped for bootdevice and only 3 mounts are wrong so lemme fix them | 14:45 |
| Mister_Magister | dsp/bluetooth/modem | 14:46 |
| mal | yeah, that _a issue is quite common to make | 14:47 |
| mal | I have done the same mistake myself couple of times | 14:47 |
| Mister_Magister | yeah i didn't know what parititiohns are going to be in use so i only removed _a in thigns like boot | 14:47 |
| Mister_Magister | okay now just dev-block-by\x2dname-metadata.device: Job dev-block-by\x2dname-metadata.device/start timed out. | 14:50 |
| Mister_Magister | that and ohmd seems to die | 14:51 |
| Mister_Magister | lemme check that mount | 14:51 |
| Mister_Magister | oh because it doesn't have bootdevice it has /dev/block/by-name/metadata | 14:52 |
| mal | yes, that is another common issue | 14:53 |
| mal | you can see that in some fixup entries, I think fp5 and some sonys | 14:53 |
| Mister_Magister | ok now it dies much quicker | 14:53 |
| mal | droid-hal failing? | 14:53 |
| Mister_Magister | i have no clue because telnet dies | 14:54 |
| Mister_Magister | the device is fine | 14:54 |
| Mister_Magister | gimme sec i really gotta shut up audit | 14:54 |
| mal | does it reboot or just usb fails? | 14:54 |
| mal | if usb then probably usb-moded and or connman config needs fixing | 14:54 |
| Mister_Magister | just usb is being disabled | 14:55 |
| mal | some newer devices need custom connman config for usb stuff | 14:55 |
| Mister_Magister | screen lights up because mce starts | 14:55 |
| Mister_Magister | and its just hanging around | 14:55 |
| mal | is usb inteface on device usb0 or rndis0? | 14:55 |
| Mister_Magister | give me one second | 14:56 |
| Mister_Magister | usb0 | 14:56 |
| Mister_Magister | shall i steal this? https://github.com/sailfishos-sony-nagara/droid-config-sony-nagara/blob/main/sparse/etc/connman/main-custom.conf | 14:56 |
| Mister_Magister | yeah last message in log i have is usb_moded 0.86.0+mer66 starting | 14:57 |
| mal | https://github.com/mlehtima/droid-config-fp5/blob/master/sparse/etc/connman/main-custom.conf and https://github.com/mlehtima/droid-config-fp5/blob/master/sparse/var/lib/environment/connman/override-main-config.conf | 14:58 |
| mal | check if those are the same | 14:58 |
| mal | I think 5.0 brought some changes | 14:58 |
| mal | looks like nagara has been updated also | 14:59 |
| Mister_Magister | they seem to have same config | 15:00 |
| mal | also check that usb-moded has correct config for usb | 15:00 |
| Mister_Magister | yeah the connman stuff didn't help and since its dying on usb_moded its gotta be it | 15:01 |
| mal | especially this part https://github.com/mlehtima/droid-config-fp5/blob/master/sparse/etc/usb-moded/usb-moded-fp5.ini#L8 | 15:01 |
| mal | but gadget paths etc might wrong also on your device | 15:01 |
| Mister_Magister | but hey i've shut up audit so we got that | 15:02 |
| mal | does your device have dummy udc? | 15:02 |
| Mister_Magister | i have no idea | 15:02 |
| mal | in addition to the real one | 15:02 |
| mal | ok | 15:02 |
| Mister_Magister | but it has a600000.dwc3 | 15:02 |
| Mister_Magister | i didn't have any custom usb_moded config | 15:03 |
| Mister_Magister | so imma try yours | 15:03 |
| Mister_Magister | nada | 15:03 |
| mal | note that I use different gadget_conf_directory, community devices use c.1 I think | 15:04 |
| Mister_Magister | but we're making great progress :P | 15:04 |
| Mister_Magister | i'll check it next | 15:04 |
| mal | my build is done in the official way not the community way | 15:04 |
| Mister_Magister | yeah i get you | 15:05 |
| mal | meaning that I have packaged system/vendor stuff etc | 15:06 |
| Mister_Magister | oh yeah its supposed to be c.1 | 15:06 |
| mal | to prevent licensing issues I host those rpms from a custom repo from my server | 15:07 |
| Mister_Magister | okay so it restarted rndis | 15:07 |
| Mister_Magister | but i don't think telnet is running anymore | 15:07 |
| Mister_Magister | wait lemme check ssh | 15:07 |
| Mister_Magister | no that is not running either | 15:08 |
| mal | check if it's just missing ip? | 15:08 |
| Mister_Magister | yeah i can't ping it | 15:08 |
| mal | enable persistent journal and check it | 15:08 |
| Mister_Magister | nor dhcp is running | 15:08 |
| Mister_Magister | on it | 15:08 |
| Mister_Magister | but dhcp is not even working from init-debug only from ramdisk | 15:09 |
| mal | not sure which udev things you need in that config, if any | 15:09 |
| Mister_Magister | we're going through those issues like storm, the dream team is back :P | 15:11 |
| Mister_Magister | i hope i'm not bothering you too much mal | 15:11 |
| mal | well I'm preparing changes for glibc update | 15:12 |
| Mister_Magister | hmmm the issue is now that to read the journal i have to resume init but when i do that, i cannot uhh read the journal | 15:12 |
| mal | you could make usb-moded | 15:13 |
| mal | and only start it manually for testing | 15:13 |
| mal | *mask | 15:13 |
| Mister_Magister | hm | 15:13 |
| Mister_Magister | oh hey, you could've mentioned that earlier lol | 15:14 |
| mal | hehe | 15:14 |
| mal | I thought it was obvious, sorry | 15:14 |
| Mister_Magister | no worries | 15:14 |
| Mister_Magister | binder-list: not found thats interesting, but anyway back to the issue on hand | 15:14 |
| Mister_Magister | what are you interested in from journal | 15:14 |
| mal | just check usb related messages and see if you can something that could be causing issues | 15:15 |
| mal | *can find | 15:15 |
| Mister_Magister | https://paste.opensuse.org/pastes/723b6a47c57a | 15:15 |
| Mister_Magister | i don't think there's much of anything | 15:16 |
| mal | yeah, not useful | 15:17 |
| mal | about binder-list https://github.com/mlehtima/droid-config-fp5/blob/master/rpm/droid-config-fp5.spec#L31 | 15:17 |
| Mister_Magister | yeah the nagara had that too | 15:18 |
| Mister_Magister | mal: oh btw why official devices don't have reset device option? | 15:18 |
| Mister_Magister | Provides: jolla-settings-system-reset this | 15:18 |
| mal | depends whether there is suitable partition for factory reset | 15:19 |
| Mister_Magister | but ports do factory reset without such partition? | 15:20 |
| mal | not really possible to do it if there is no good partition | 15:20 |
| Mister_Magister | huh, in my unofficial port i used reset device multiple times without having any partition and it worked just fine | 15:21 |
| Mister_Magister | weird | 15:21 |
| mal | hmm | 15:21 |
| mal | ok | 15:21 |
| mal | I need to have a break now and go for a walk | 15:22 |
| Mister_Magister | no worries | 15:22 |
| Mister_Magister | i'll rebuild the zip | 15:22 |
| Mister_Magister | make sure all the fixes are in etc | 15:22 |
| mal | recheck the usb stuff to see that the config is ok | 15:23 |
| Mister_Magister | alrigh | 15:23 |
| mal | also check the mounts, if those cause droid-hal to fail, those from super partition | 15:24 |
| Mister_Magister | mounts seem to be working fine atm | 15:25 |
| Mister_Magister | oh | 15:25 |
| Mister_Magister | odm is not mounted | 15:25 |
| Mister_Magister | and vendor | 15:25 |
| Mister_Magister | good call | 15:25 |
| Mister_Magister | dummy_netd.service also fails if thats relevant | 15:25 |
| mal | probably needs droid-hal to be running | 15:26 |
| Mister_Magister | and /usr/libexec/droid-hybris/system/bin/bootctl is missing | 15:26 |
| mal | no point in check other stuff than usb until you get droid-hal working | 15:26 |
| Mister_Magister | yeah i know | 15:26 |
| Mister_Magister | hm /dev/mapper/dynpart-vendor_a doesn't seem to have any filesystem on it | 15:27 |
| Mister_Magister | /dev/mapper/dynpart-vendor_a: block special (252/4 | 15:27 |
| Mister_Magister | same with odm, thats why they're not mounting | 15:27 |
| mal | strange | 15:28 |
| Mister_Magister | yeah | 15:28 |
| Mister_Magister | go go i'll just spam here for logs | 15:29 |
| T42 | <Mister_Magister> @b100dian yooo do you have a sec | 15:32 |
| T42 | <b100dian> What are you encountering now? | 15:38 |
| T42 | <Mister_Magister> when usb_moded starts, it yeets telnet and ssh and rndis. While we fixed rndis, the device has no ip, so not sure if telnet/ssh is running cause can't talk to it | 15:40 |
| T42 | <Mister_Magister> and i saw you fix some stuff in tucana | 15:40 |
| T42 | <b100dian> Yeah, but at this point I would mask the service and get a static IP on the PC. Let me see what was the fix for tucana though | 15:42 |
| T42 | <Mister_Magister> yeah we got that far but it would be proper to have it fixed | 15:42 |
| T42 | <Mister_Magister> if i mask usb-moded its fine | 15:42 |
| T42 | <b100dian> Pretty sure my tucana fix is inspired from some HelloVolla droid-config, don't remember which | 15:46 |
| T42 | <Mister_Magister> all i know is that tucana fix doesn't fix my ip sadly | 15:46 |
| T42 | <b100dian> actually those I mentioned ewere updated to ecm now? | 15:47 |
| T42 | <Mister_Magister> no idea | 15:47 |
| T42 | <b100dian> Maybe you need the DontBringDownAtStartup = usb as in https://github.com/HelloVolla/droid-config-halium-mimameid/commit/d8e5e7848dd0982f71469e1c041eabc9061c8b9c ? | 15:48 |
| T42 | <Mister_Magister> i could try changing to eem | 15:49 |
| Mister_Magister | or i could try ignoring it, masking usb-moded and trying to get droid-hal up | 15:51 |
| Mister_Magister | and figuring out why i can't mount vendor | 15:51 |
| Mister_Magister | ohhhhhhhhh | 15:55 |
| Mister_Magister | vendor is erofs not ext4 | 15:55 |
| Mister_Magister | ohhhh vendor mount is from other device (fp4) thats why it has worng filesystem type | 15:56 |
| T42 | <b100dian> so another device path right? | 15:57 |
| Mister_Magister | no same device path, different filesystem | 15:59 |
| Mister_Magister | and odm.mount is autogenerated so it has wrong path so i have to stop it from generating and odm_root which is also copied from fp4 also needs switch to erofs | 16:00 |
| Mister_Magister | now they mounted fine just fine | 16:00 |
| Mister_Magister | yeah makefstab_skip_entries is missing odm | 16:01 |
| T42 | <b100dian> Android has entered the chat. That is hilarious:) | 16:03 |
| Mister_Magister | WE GOT GUI | 16:03 |
| Mister_Magister | WHA | 16:03 |
| Mister_Magister | THATS TOO SIMPLE | 16:03 |
| Mister_Magister | WHAT THE HELL | 16:03 |
| Mister_Magister | even touch works and i didn't add the config file for it | 16:04 |
| Mister_Magister | what the hell | 16:04 |
| Mister_Magister | this is way too simple | 16:04 |
| Mister_Magister | i expected at least 2 weeks to get gui | 16:04 |
| T42 | <b100dian> haha.. remember the libuicompat on sake, that got me for 2 weeks. You probably have other horror stories. | 16:05 |
| Mister_Magister | it took us half a year to hack the display on g2 and its not fixed to this day | 16:07 |
| Mister_Magister | but it looks and works great super fast | 16:07 |
| Mister_Magister | now i gotta get everything to work, audio, wifi, bt, everything | 16:07 |
| Mister_Magister | everything seems to be broken | 16:07 |
| Mister_Magister | with wifi i'll be able to debug usbmoded | 16:08 |
| Mister_Magister | everything just magically clicked into place lol | 16:09 |
| Mister_Magister | mal: binder-list error just… like… went away | 16:23 |
| Mister_Magister | ah no, it didn't nvm | 16:23 |
| Mister_Magister | bruh it even works at 144 HZ | 16:25 |
| Mister_Magister | mal: you gotta fix the fps display to display 3 digits lol | 16:26 |
| T42 | <b100dian> now that's also a surprise. We didn't figure out what would sake need for that | 16:28 |
| T42 | <b100dian> Now you need to write a hz scaling for battery saving 🙂 | 16:28 |
| Mister_Magister | yeah sake was locked 60 | 16:28 |
| Mister_Magister | naah | 16:29 |
| Mister_Magister | its so smoooooooooth honestly | 16:29 |
| Mister_Magister | when you enable frame rate display detailed? the bars are pure green | 16:29 |
| Mister_Magister | nothing on them, just flat green | 16:29 |
| Mister_Magister | i've never seen device this smooth outside of huawei p8lite and jolla phone the og one | 16:29 |
| Mister_Magister | and not because of 144Hz because of no microstutter | 16:29 |
| Mister_Magister | that being said, i don't have wlan device despite wlan module being loaded so thats fun | 16:30 |
| Mister_Magister | also how do i enable tablet ui mode? | 16:31 |
| mal | Mister_Magister: is there only 2 number in frame rate display, need to check the code | 16:32 |
| Mister_Magister | yeah thats what i'm saying :P | 16:33 |
| Mister_Magister | limited at 99 | 16:33 |
| mal | not sure where that is even handled, trying to find it | 16:33 |
| Mister_Magister | mal: don't stress on it i was mostly joking :P | 16:34 |
| mal | for bluetooth check that the bluebinder script finds the mac address | 16:35 |
| mal | that is the usual cause for issue | 16:35 |
| Mister_Magister | i'm flashing newer zip | 16:35 |
| Mister_Magister | don't need bluetooth need wifi for now | 16:35 |
| mal | if wlan module or built in? | 16:35 |
| Mister_Magister | seems to have wlan module | 16:35 |
| Mister_Magister | just hold on a second i gotta flash new zip | 16:36 |
| mal | some need writing to some magic paths to start wlan | 16:36 |
| Mister_Magister | im flashing new zip to fix the binder thingy | 16:38 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!