| T42 | <adampigg> Mal, so, options? R u happy i have everything mounted? | 06:44 |
|---|---|---|
| Mister_Magister | mal: just friendly reminder :P | 06:45 |
| T42 | <adampigg> Mister_magister: we appreciate your patience, you are now no2 in the queue and will get to you shortly | 06:47 |
| Mister_Magister | @adampigg staph it i'm just reminding | 06:47 |
| T42 | <DSstill01> Can one of you post the output of 'ls /etc/pulse'? | 09:58 |
| mal | how about you provide what you have? | 10:14 |
| T42 | <DSstill01> Nothing | 10:22 |
| T42 | <DSstill01> I deleted it | 10:22 |
| T42 | <DSstill01> After fm changes i had no incall audio, so i deleted the whole dir. Now i have to restore it by reinstalling packages | 10:23 |
| T42 | <birdzhang> lipstick crash random on 15.1 base with dmesg log binder_alloc: 14183: binder_alloc_buf, no vma | 10:26 |
| T42 | <birdzhang> full dmesg log https://pastebin.com/jvjJmxd5 | 10:27 |
| T42 | <birdzhang> i'm stuck on this issue | 10:28 |
| KALUBE | Hi, I'm porting to a Samsung Galaxy J3 (2016). I have telnet and the system is up, but systemctl hangs on some commands and masking/manually starting droid-hal-init causes a reboot to recovery (confirmed by strace) | 12:14 |
| KALUBE | systemctl status: https://pastebin.com/3nGQwkEW and journal log https://pastebin.com/NvK3psdH | 12:14 |
| KALUBE | I've removed a lot of spam from power_supply in the journal log. | 12:15 |
| mal | which android base? | 12:26 |
| KALUBE | cm-14.1 | 12:28 |
| mal | I assume you have added the 14.1 specific things from faq? | 12:34 |
| KALUBE | Oh shoot, missed them | 12:36 |
| KALUBE | Mal: Still the same issue, running systemctl mask droid-hal-init still hangs | 13:10 |
| mal | but does droid-hal-init fail? | 13:11 |
| KALUBE | yes | 13:16 |
| mal | with what error? | 13:17 |
| mal | the changes I mentioned from faq help with droid-hal-init, but there are various things that can be wrong | 13:18 |
| mal | did you need straggler files define in droid-hal spec? | 13:18 |
| KALUBE | I don't have a message about unpackaged files found when running build_packages so I assume not? This is the output of strace droid-hal-init https://pastebin.com/92DwNhPU | 13:19 |
| KALUBE | I'm not sure that is wasn't called by systemd before that, will make sure it's disabled and see if the output is different | 13:21 |
| KALUBE | No it was masked with that output. I've noticed in systemctl that init-done.service is hanging on wait-for-file /run/user-sessions/user-init-debug.flag, is this set by droid-hal-init? Or could it be relates | 13:25 |
| mal | KALUBE: looks like you didn't disable selinux correctly | 13:25 |
| mal | so enable selinux bootparam kernel config (keep audit enabled), then add selinux=0 to kernel commandline | 13:26 |
| mal | there should instructions to that somewhere | 13:26 |
| KALUBE | From the earlier log? I noticed that and patched the boot image. | 13:26 |
| KALUBE | I currently have CONFIG_AUDIT disabled, SELINUX enabled and both disabled with cmdline. I'll set them both enabled in kernel config and disable them with cmdline? | 13:26 |
| mal | show your kernel config | 13:31 |
| KALUBE | I've just adjusted it... https://pastebin.com/WGk0FD24 | 13:32 |
| KALUBE | Changes: CONFIG_SECURITY_SELINUX_BOOTPARAM=y CONFIG_AUDIT=y SELINUX_BOOTPARAM=y | 13:32 |
| KALUBE | And selinux=0 audit=0 is in cmdline | 13:33 |
| mal | no need to put audit=0, I never use that | 13:35 |
| KALUBE | Ah ok | 13:35 |
| mal | also line 3668 is wrong in defconfig | 13:35 |
| KALUBE | Should be unset? | 13:36 |
| mal | remove the line | 13:36 |
| mal | that value is already set on line 3460 | 13:36 |
| KALUBE | Oh, they do the same thing? | 13:36 |
| mal | there is no such thing as SELINUX_BOOTPARAM, it's CONFIG_SECURITY_SELINUX_BOOTPARAM | 13:37 |
| KALUBE | Interesting, the mer_verify_kernel_config script references it | 13:37 |
| KALUBE | Guessing it's shorthand | 13:37 |
| mal | yes | 13:37 |
| mal | same that it says AUDIT=y which is of course CONFIG_AUDIT=y | 13:38 |
| KALUBE | Makes sense. | 13:41 |
| KALUBE | Ok, it boots, droid-hal-init is masked | 13:42 |
| KALUBE | Running strace droid-hal-init still causes a reboot to recovery | 13:42 |
| mal | does it still complain about selinux in strace? | 13:44 |
| KALUBE | This is the strace https://pastebin.com/MF6RDyJW it mentions selinux but I'm not sure exactly what to look for | 13:45 |
| KALUBE | statfs64("/sys/fs/selinux", 88, 0xbe8278f0) = -1 ENOENT (No such file or directory)this line perhaps | 13:45 |
| mal | droid-hal-init: Security fail | 13:46 |
| mal | which means selinux afaik | 13:46 |
| mal | check the kernel commandline from /proc/cmdline | 13:47 |
| mal | also check selinux in kernel config using : "zcat /proc/config.gz | grep SELLINUX" | 13:48 |
| mal | *"zcat /proc/config.gz | grep SELINUX" | 13:48 |
| KALUBE | No mention of selinux=0 in the cmdline, or in config.gz https://pastebin.com/eDpdQZsJ | 13:49 |
| KALUBE | When I unpack hybris-boot.img it has selinux=0 in the cmdline so this is really weird | 13:49 |
| mal | and you are sure you correctly flashed the kernel? | 13:50 |
| KALUBE | I boot TWRP and flash hybris-boot.img to the "boot" partition | 13:50 |
| KALUBE | I haven't been able to use the standard sfe zip flash because tar runs out of memory | 13:51 |
| KALUBE | am I missing a step? | 13:51 |
| mal | I have never used twrp to flash kernel like that, I flash kernel using fastboot | 13:51 |
| KALUBE | The phone doesn't have fastboot, I will try flashing it using heimdall | 13:52 |
| mal | anyway you need to make sure the kernel is correct before you can continue figuring out if things fail otherwise | 13:59 |
| KALUBE | Yeah, I've just managed to flash with heimdall, it was having some... Difficulty. But the cmdline is the same | 14:00 |
| KALUBE | It's like it's using some other cmdline entirely?# | 14:01 |
| KALUBE | I may have found the cause of the issue... | 14:01 |
| KALUBE | There is a boot partition but also a kernel partition, I think the kernel partition is actually the boot partition, and it's been using the lineage boot image the whole time | 14:02 |
| KALUBE | It seems this was not it. There also a boot2 partition? | 14:07 |
| mal | no idea about that device, but since you did already boot to sfos it should have been the actual kernel | 14:09 |
| mal | are you sure you copied the correct kernel to recovery for flashing? | 14:09 |
| KALUBE | I was right, I checked the updater script for lineage and it flashed it's boot.img to the KERNEL partition. So it is the correct one. | 14:09 |
| KALUBE | I did have the correct kernel, I was flashing it to the wrong partition though. But having just flashed it, /proc/cmdline didn't change? | 14:10 |
| KALUBE | flashed it to KERNEL that is | 14:10 |
| KALUBE | No luck, I'm going to go and come back to this later. | 14:17 |
| mal | wondering what you did differently before | 14:18 |
| piggz | evening all | 18:01 |
| justgerd | hey piggz | 18:02 |
| Mister_Magister | piggz: no | 18:02 |
| piggz | Mister_Magister: no what? | 18:03 |
| Mister_Magister | you know what | 18:03 |
| justgerd | Mister_Magister: not evening yet in your timezone? | 18:05 |
| Mister_Magister | justgerd: it is | 18:05 |
| justgerd | Mister_Magister: where did that no come from then | 18:05 |
| Mister_Magister | you will see | 18:06 |
| justgerd | okay i guess | 18:06 |
| Mister_Magister | piggz: got wifi? | 18:07 |
| piggz | Mister_Magister: nope | 18:07 |
| piggz | u got sensors? | 18:07 |
| Mister_Magister | i'm getting 5z first | 18:09 |
| Mister_Magister | today is the day i get gui | 18:10 |
| Mister_Magister | i hope | 18:10 |
| piggz | Mister_Magister: the mount units you sent .... they dont appear to be for this device! im just cleaning it up now i have proper ones from dhd! | 18:10 |
| Mister_Magister | they are for this device | 18:11 |
| Mister_Magister | at least according to TheKit mounts | 18:11 |
| piggz | they look hand-balled to me, the mount commands didnt match the comments | 18:12 |
| piggz | anyway, i have proper ones now in dhd | 18:13 |
| Mister_Magister | they were totally hand made | 18:14 |
| mal | are sensors really not working on that device? | 18:15 |
| piggz | mal: yes, but i havnt looked into why yet.... | 18:16 |
| mal | so maybe the mounts really are a problem is almost nothing works | 18:16 |
| piggz | mal: well, mounts _should_ be good now..... lets see | 18:16 |
| piggz | mal: what can i show you, output of mount? | 18:18 |
| piggz | it should hopefully make more sense...providing it boots | 18:19 |
| piggz | which ofc it doesnt | 18:20 |
| Mister_Magister | mal: nah there is missing lib mount and stuff | 18:21 |
| piggz | mal: ok, mounts now as clean as i think, with a single hand-balled mount unit for /vendor, which I think we discussed adding to fstab anyway.... | 18:30 |
| piggz | https://paste.gnome.org/phlv9hfoo | 18:31 |
| Mister_Magister | wtf i flashed new rootfs now i got bootloop | 18:32 |
| piggz | Mister_Magister: oh fuck, mal was blooming right | 18:35 |
| Mister_Magister | hm? | 18:35 |
| Mister_Magister | mal is a flower? | 18:35 |
| Mister_Magister | since when? | 18:35 |
| T42 | <adampigg> (Photo, 1920x2560) https://irc.thaodan.de/.imgstore/08Kq40JSDT.png | 18:36 |
| Mister_Magister | it randomly stopped rebooting | 18:36 |
| Mister_Magister | lol | 18:36 |
| Mister_Magister | heh | 18:36 |
| Mister_Magister | gj | 18:36 |
| Mister_Magister | we have that now | 18:36 |
| piggz | mal: i bow again ;) | 18:36 |
| Mister_Magister | after 20 reboots | 18:36 |
| Mister_Magister | it stopped rebooting | 18:36 |
| Mister_Magister | cause? | 18:36 |
| Mister_Magister | i don't even have pstore lol | 18:37 |
| Mister_Magister | droid-hal-init ffs | 18:37 |
| Mister_Magister | mal: idea what patch i could be missing? https://paste.opensuse.org/view//26124512 | 18:43 |
| Mister_Magister | also mal reminder | 18:45 |
| mal | piggz: nice, I thought the mounts and symlinks should be enough :) | 18:48 |
| piggz | mal: youre a goddamn smart arse :P | 18:48 |
| piggz | so....radio | 18:48 |
| mal | :D | 18:48 |
| Mister_Magister | piggz: radio? | 18:49 |
| piggz | i got a message saying 'incompatible handset' when searching for a network | 18:49 |
| piggz | Mister_Magister: mobile net | 18:49 |
| Mister_Magister | hm | 18:49 |
| Mister_Magister | though thekit would do that but k | 18:49 |
| piggz | i wonder if the pro1 isnt compatible with my network | 18:49 |
| Mister_Magister | piggz: volte/ | 18:50 |
| Mister_Magister | ? | 18:50 |
| piggz | no, its regular old 3/4g on ThreeUK | 18:50 |
| Mister_Magister | https://paste.opensuse.org/view//85339088 lol test_egl | 18:51 |
| Mister_Magister | piggz: assume it should work | 18:51 |
| mal | Mister_Magister: where is that lib on device? | 18:51 |
| Mister_Magister | god knows but that's not important | 18:52 |
| Mister_Magister | why hwservicemanager and servicemanager fails is important | 18:53 |
| Mister_Magister | wait i will replace linker maybe | 18:53 |
| Mister_Magister | mal: btw this is different device (sorry for confusing) | 18:54 |
| Mister_Magister | piggz: if adding fstab.qcom and the other thing to hybris modules to build is enough shouldn't we replace vendorimage with that? | 18:55 |
| Mister_Magister | well some devices will fail to build but… | 18:55 |
| piggz | Mister_Magister: i dont want to change the hybris-boot process for all devices just yet.... | 18:58 |
| Mister_Magister | just giving ideas eh | 18:58 |
| Mister_Magister | hybris16 isn't really out | 18:58 |
| piggz | id like to think of a way to pass in a variable, so that each device could add to the list of things to buiild | 18:58 |
| Mister_Magister | porting is hard | 19:00 |
| Mister_Magister | i disabled droid-hal-init and now i don't get rndis | 19:01 |
| Mister_Magister | is that reasonable at all? no | 19:01 |
| piggz | rats, wifi stopped working | 19:01 |
| piggz | after reboot | 19:01 |
| Mister_Magister | see | 19:03 |
| Mister_Magister | it's hard | 19:03 |
| Mister_Magister | it does totally unreasonable things | 19:03 |
| Mister_Magister | i would really love to have mal's hybris16 | 19:03 |
| piggz | go figure, rebooted and wifi works | 19:05 |
| Mister_Magister | timing | 19:05 |
| Mister_Magister | easy | 19:05 |
| *** justgerd is now known as JustGerd | 19:05 | |
| JustGerd | wait mal is working on hybris16? | 19:07 |
| Mister_Magister | no | 19:19 |
| Mister_Magister | we are | 19:19 |
| mal | well I have almost all on my computer but I haven't pushed those to public | 19:19 |
| JustGerd | mal: ah okay makese sense | 19:39 |
| KALUBE | I'm still having no luck with droid-hal-init. It seems that my cmdline get's ignored by samsung's bootloader, I have tried disabling SELINUX and AUDIT in the kernel config but it still fails and causes a reboot if I mask droid-hal-init and start it manually https://pastebin.com/7gqXwjF7 | 19:45 |
| mal | KALUBE: you need to find some way to disable selinux in a way that it doesn't break anything | 20:19 |
| mal | Mister_Magister: droid-hal-init initializes configfs usb usually | 20:37 |
| Mister_Magister | ah | 20:37 |
| Mister_Magister | usually when i masked dhi i still had networking ;-; | 20:37 |
| mal | it works if first telnet make it work | 20:38 |
| mal | I think | 20:38 |
| piggz | i notice that the init script starts network, then, when dhi starts, the connection drops and comes up again | 20:43 |
| mal | yes | 20:43 |
| mal | that way it should usually do it, assuming init script works for the device | 20:43 |
| piggz | mal: where does /dev/vhci come from? | 20:44 |
| KALUBE | mal: Any suggestions on disabling selinux? Sailfish doesn't have /etc/selinux, I'm not sure where to go from here to be honest | 20:45 |
| mal | try what happens if you set all of the selinux kernel configs to =n, I really don't know any other way except patching more of android init | 20:47 |
| mal | piggz: you need the vhci kernel config and then bluebinder, after those it should work | 20:47 |
| piggz | i have part 2 of that lot.... | 20:48 |
| mal | piggz: did you make sure you have the bluetooth firmware symlink also | 20:48 |
| piggz | mal: unlikely.... | 20:48 |
| mal | piggz: https://github.com/mer-hybris/android_kernel_sony_msm/blob/hybris-sony-aosp-8.1.0_r52_20190206/arch/arm64/configs/aosp_nile_pioneer_defconfig#L812 | 20:48 |
| mal | piggz: I think it might need similar symlink you added for /firmware, check out/..../root again for the links it created there | 20:49 |
| mal | probably /bt_firmware or something like that | 20:49 |
| piggz | mal: i have something like that... | 20:50 |
| vknecht | hello :-) been trying to get systemd logs on sdcard, does that look like an honest attempt ? any obvious flaw which would prevent that to work ? https://pastebin.com/GSzMTVgr | 21:53 |
| mal | vknecht: why not just enable persistent journal log? | 21:55 |
| vknecht | because I'd then have to copy/dd partition back from fs to read it I think | 21:56 |
| vknecht | tho it's possible, just a hassle | 21:56 |
| vknecht | other way would be to mod preinit, but would have to override/sparse a jolla package, which seems a hassle too | 21:58 |
| vknecht | here's where I'm at, logs gotten from preinit late.d plugin to sdcard, no telnet available : https://pastebin.com/2aiY4J1G | 22:03 |
| vknecht | not in the setup to continue right now, but if you have ideas, I read logs :-) | 22:04 |
| vknecht | just having a systemd unit to spit all logs on sdcard would help | 22:05 |
| vknecht | my best guess atm is I miss a /vendor symlink to /system/vendor, but I'm somewhat blind (tried much of rndis config recommendations, even disabling configfs completely) | 22:08 |
| vknecht | symptoms are : stays on vendor logo for ~10 seconds, then purple led for ~3 seconds (or indefinitely on previous attempts) and device stucks on logo indefinitely (4/5 minutes) | 22:15 |
| vknecht | btw I tried the community-way extract-rootfs-manually but it fails/crashes twrp, so atm I'm doing mixed-way aka. lvm but keep /system partition (skipping fimage stuff) | 22:20 |
| vknecht | good night and sweet dreams :-) | 22:26 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!
