| xreactx[m]__ | https://pastebin.com/PjmxXZXy qt5-qpa-surfaceflinger-plugin fails to build... buy why? | 00:26 |
|---|---|---|
| T42 | <birdzhang> @xreactx[m] you can skip it, it's no longer needed | 00:56 |
| T42 | <birdzhang> https://github.com/mer-hybris/droid-hal-device/commit/17e564bd6187c7c206f37b150cfef74b54664694 | 00:56 |
| xreactx[m]__ | ok, whats replacing it? | 00:57 |
| T42 | <birdzhang> qt5-qpa-hwcomposer-plugin | 01:13 |
| Thaodan | mal: I have no issue with -b. I prefer it and was annoyed by the asking to build the pkg I've asked for it to build. | 01:56 |
| xreactx[m]__ | what would cause an "updater process ended with ERROR: 7" when installing SFOS in twrp? | 03:16 |
| xreactx[m]__ | before that it says "failed to extract filesystem!" | 03:17 |
| T42 | <Akatsu %lastname%> SSH uses same ip like telnet or different? | 06:27 |
| T42 | <Akatsu %lastname%> i have init-degug in real rootfs and have no idea how to enter ssh/telnet now | 06:33 |
| lbt | PSA: git.saifishos.org / git.merproject.org going down for maintenance | 07:49 |
| deathmist | xreactx[m]__: TWRP's tar binary is broken and has a memory leak on many devices now, I've just opted to use "/system/bin/toybox tar" instead like so: https://git.io/JeBEp | 08:50 |
| deathmist | you can replace the tar line in hybris-boot's updater-unpack.sh https://git.io/JeBuf if you want to go that route, another way is by using tar from SFOS' busybox binary from out/ | 08:53 |
| deathmist | sometimes even downgrading TWRP a bit helps so you can try that first if you want | 08:54 |
| vknecht | grrr, QImage loadData not able to handle NV12 QVideoFrame data... is that expected ? | 09:02 |
| vknecht | ^ abranson ? :-) | 09:03 |
| deathmist | xreactx[m]__: if you opt to use my solution obviously use "/system/bin/toybox tar -xj -f $FS_ARC -C $FS_DST" instead for the line as I've made a custom installer solution and changed A LOT of things :p | 09:06 |
| deathmist | mal: I'll try minimer too as I get no UI during boot either and test_hwcomposer still segfaults but now the 32-bit libsdm-disp-vndapis.so lib is present and it at least seems happy about that | 09:10 |
| deathmist | mal: it seems I've also forgotten how to enable networking over telnet as I've not had to do this for many months at this point xD I need it to install "qt5-qtdeclarative-qmlscene" and I don't want to make new build just to get a single package installed | 09:45 |
| vknecht | multimedia-private & qt_imageFromVideoFrame() works :-) | 10:35 |
| T42 | <adampigg> vknecht: does that pass harbour ? | 11:20 |
| vknecht | seems so | 11:20 |
| vknecht | still have to battle to get orientation information from metadata :-/ | 11:21 |
| vknecht | hopefuly should be easier and more complete from c++ side than from qml... | 11:23 |
| mal | Thaodan: I tried -b and it didn't ask anything | 11:24 |
| deathmist | mal: look like I also can't package the 32-bit libsdm-disp-vndapis.so to my dcd sparse files https://git.io/JeBzo as it causes "nothing provides libbinder.so needed by droid-config-cheeseburger" during mic build :/ | 11:41 |
| mal | deathmist: did it actually help? | 11:43 |
| deathmist | lipstick still didn't start but test_hwcomposer didn't say anything about it missing anymore, still segfaulted tho. now gonna try out minimer or sohuld I give strace for stuff? | 11:47 |
| mal | maybe strace | 11:48 |
| mal | not sure if that lib is compatible or not | 11:49 |
| deathmist | mal: here's strace again for test_hwcomposer https://pastebin.com/8CSpfQgQ I also modified lipstick.service to output strace to a file in a similar fashion https://dpaste.de/e4vd/raw | 12:04 |
| deathmist | more specifically, I changed ExecStart line to "ExecStart=/usr/bin/strace -f -o /sdcard/lipstick-strace.log /usr/bin/lipstick $LIPSTICK_OPTIONS --systemd" and grabbed log after rebooting to TWRP | 12:06 |
| mal | deathmist: looks like it crashes at the same place as before | 12:29 |
| deathmist | damn. could it be something to do with the EGL libraries "not existing" as in they're just named differently on my device? | 12:47 |
| deathmist | e.g. it's looking for "/system/lib/egl/libGLESv1_CM.so" but it's "/system/lib/libGLESv1_CM.so" on my phone and also stuff like "/vendor/lib/egl/libGLESv1_CM.so" vs "/vendor/lib/egl/libGLESv1_CM_adreno.so" | 12:48 |
| mal | deathmist: hmm, does it find the egl libs or not? if not then add some symlink | 13:03 |
| piggz | mal: did you find that silia config? | 13:05 |
| mal | piggz: sorry, not yet, I out most of today | 13:09 |
| T42 | <Akatsu %lastname%> how can i add libminsf to build? | 13:26 |
| T42 | <Akatsu %lastname%> i executed ` make -j4 libdroidmedia minimediaservice minisfservice libminisf` and it shows that it builds something | 13:26 |
| nitin03 | hello | 13:41 |
| nitin03 | device xiaomi poco f1 | 13:41 |
| nitin03 | base lineage 16.0....already build sfos zip without any critical errors and flashed it but device not booting... checked dmesg on host | 13:42 |
| nitin03 | but it is showing up same as android | 13:43 |
| nitin03 | https://pastebin.com/LR9fVGQD | 13:47 |
| nitin03 | and check lsusb android shows up iSerial 3 d4df6fef | 13:47 |
| nitin03 | but on sfos nothing shows up | 13:48 |
| nitin03 | device is getting stuck on boot logo and stays there for 1 to 3 minutes then reboots and gets stuck at boot log again and it is going on and on | 13:50 |
| deathmist | nitin03: check if "/data/.stowaways/sailfishos/init.log" exists from TWRP and adb shell | 13:51 |
| nitin03 | ok wait a minute | 13:51 |
| deathmist | it's probably a misconfigured defconfig or boot cmdline | 13:51 |
| nitin03 | yes there is a init.log file | 13:51 |
| deathmist | can you put up a paste of it? | 13:52 |
| nitin03 | should i pull it and share it | 13:52 |
| nitin03 | https://pastebin.com/x8Wekwyv init.log | 13:54 |
| T42 | <Akatsu %lastname%> (Photo, 1280x471) https://irc.thaodan.de/.imgstore/0ConxnXBMh.png Do i have toi add every missing file under straggler_files ? I mean ones which arent dirs | 13:55 |
| nitin03 | deathmist https://pastebin.com/x8Wekwyv init.log | 13:56 |
| deathmist | ah, it can't setup rndis on interface usb0 or rndis0 which are the very common ones for devices (mine is usb0) | 13:57 |
| deathmist | can you link your defconfig? it doesn't look like it even has a USB interface (probably missing options) | 13:58 |
| nitin03 | any way to tackle this | 13:58 |
| nitin03 | deathmist https://pastebin.com/Mt7nYVsf | 14:01 |
| deathmist | nitin03: look like your device uses configfs. you can attempt to use my defconfig changes for telnet (SD835 SOC on hybris-16.0 too): https://git.io/JeB2i while making sure they stick in .config too (find out/ -name .config) | 14:06 |
| nitin03 | thank u..i will get back to u | 14:07 |
| deathmist | then "mka hybris-boot" habuild env obviously, flash and boot new image => get init.log again unless you can telnet in with "telnet 192.168.2.15 2323" | 14:08 |
| nitin03 | can i send u direct msg next time? | 14:08 |
| deathmist | sure | 14:09 |
| nitin03 | thanks , I will check it...and i should check .config file in KERNEL_OBJ of my target device | 14:10 |
| nitin03 | right? | 14:11 |
| nitin03 | if these changes are there or not? | 14:11 |
| deathmist | yes, then you know if the changes are actually in the new hybris-boot.img you're going to flash | 14:11 |
| nitin03 | thank you so much...i will get back to u | 14:12 |
| T42 | <Akatsu %lastname%> any idea why it does not detect tools after build packages run? i get zypper/sb2/mb2 not found even if i had it working yday | 14:28 |
| T42 | <Akatsu %lastname%> probably im idiot trying execute something from platform sdk inside of habuild | 14:29 |
| xreactx[m]__ | now I'm getting ended with ERROR 6 | 14:45 |
| deathmist | xreactx[m]__: check /tmp/recovery.log, it'll tell what's actually going wrong | 14:46 |
| xreactx[m]__ | one place i failed to look | 14:46 |
| xreactx[m]__ | https://pastebin.com/E3ieqUcq | 14:54 |
| xreactx[m]__ | i think I know whats going on... i'm going to try something and report it later | 15:00 |
| T42 | <Akatsu %lastname%> why i get these sailfishOS binaries like minisf compiled under codename/system instead of codename/root/system? | 15:26 |
| T42 | <Akatsu %lastname%> It cannot find them | 15:26 |
| vknecht | @Akatsu %lastname%: that's part of droidmedia, did you build it ? it's section 13.2 of hadk pdf | 15:29 |
| T42 | <Akatsu %lastname%> yes | 15:35 |
| T42 | <Akatsu %lastname%> i see that its part of system android build | 15:35 |
| T42 | <Akatsu %lastname%> but we arent flashing new base right? only sailfish rootfs which sits on /data | 15:36 |
| T42 | <Akatsu %lastname%> (Photo, 1280x486) https://irc.thaodan.de/.imgstore/cA7o2CoYjR.png | 15:37 |
| vknecht | not sure what you mean ; when you build droidmedia package, its rpm put in droid-local-repo and integrated in sfos image when you build it | 15:39 |
| T42 | <Akatsu %lastname%> ok, wrong i didnt do 13.2 part i have finished on 9 or so | 15:39 |
| T42 | <Akatsu %lastname%> now i probably have it die on droid-hal-init | 15:42 |
| T42 | <Akatsu %lastname%> after masking init phone somehow reboot into fastboot from telnet | 15:46 |
| deathmist | @Akatsu does the device reboot with it unmasked too? if not then you really don't need to mask because you don't get any android side stuff up e.g. graphics and whatnot | 15:53 |
| deathmist | I say try to get a "journalctl -b" log, maybe dmesg | 15:53 |
| deathmist | with DHI unmasked if it doesn't make the device unstable | 15:53 |
| T42 | <Akatsu %lastname%> i tried to run init by hand and im not sure but it looks like its not mounting partitions well | 15:55 |
| deathmist | show your device fstab, it's in your android_device_* repo somewhere | 15:55 |
| deathmist | and output of mount after init has ran | 15:56 |
| T42 | <Akatsu %lastname%> https://github.com/LineageOS/android_device_leeco_s2/blob/lineage-15.1/rootdir/etc/fstab.qcom | 15:57 |
| T42 | <Akatsu %lastname%> dmesg | grep "mount" is enought? | 15:57 |
| deathmist | make sure droid-hal-init has executed, then output of just "mount" command, nothing else | 15:58 |
| T42 | <Akatsu %lastname%> how can i bring it back after masking? | 15:58 |
| deathmist | unmask => systemctl unmask droid-hal-init (or if you can't then remove the service symlink from /etc/systemd/*) | 15:59 |
| T42 | <Akatsu %lastname%> i made it via ln -s | 16:00 |
| deathmist | but you'll have to reboot too obviously to make it run | 16:00 |
| deathmist | you can run DHI only once | 16:00 |
| T42 | <Akatsu %lastname%> mount only shows me mount on /data | 16:01 |
| T42 | <Akatsu %lastname%> https://paste.ubuntu.com/p/Ywxh3McXsv/ | 16:01 |
| T42 | <Akatsu %lastname%> I do have mounts done in fixup-mountpoints | 16:02 |
| deathmist | this was on 15.1 base right? you can add a system.mount unit for systemd so that gets taken care of: https://git.io/JeBnQ (NOTE: the ../ lines aren't content, it's a symlink to the file outside .wants dir and obv change device path) | 16:05 |
| T42 | <Akatsu %lastname%> yes 15.1 base | 16:06 |
| T42 | <Akatsu %lastname%> i will try soon. what about these libminisf which i pointed before? where do they need to be standard /system or inside rootfs system? | 16:09 |
| deathmist | about the other patitions, you'll likely need to add some straggler files in dhd spec file for /dsp /firmware /persist dirs at least https://git.io/JeBV4 and "-detritus" pkg in dcd patterns https://git.io/JeBVB (change cheeseburger to your codename obviously) | 16:10 |
| deathmist | SFOS doesn't touch your /system, idk what you are trying to do with libminisf right now | 16:12 |
| T42 | <BusterBg_18> deathmist: did you got GUI to start on 16.0? | 16:23 |
| deathmist | BusterBg_18: nah, both test_hwcomposer and lipstick fail still | 16:23 |
| deathmist | droid-hal-init is fine tho | 16:24 |
| deathmist | there's also minimer in faq-hadk, but I've not tried the new image I built yet | 16:25 |
| deathmist | I forgot how to get networking via rndis on the device so I just made a new image in the end lol | 16:26 |
| T42 | <BusterBg_18> I didnt know that you could get networking on a device using rndis:0 | 16:27 |
| deathmist | xreactx[m]__: have you figured out the "set_perm" function issue yet on the script? | 16:30 |
| deathmist | no clue how ET_PERMISSIONS% got expanded into that, it was an issue back in 2015 and the line should be something like set_metadata("/tmp/updater-unpack.sh", "uid", 0, "gid", 0, "mode", 00755); instead now | 16:37 |
| deathmist | https://git.io/JeBVF oh it does an Android version check and should ONLY use set_perm if < 4.4, no clue how that got into your script | 16:42 |
| T42 | <BusterBg_18> I took logs once more, I dont see any changes from yesterday https://pastebin.com/gDVnwiki dmesg https://pastebin.com/FGt7b754 journalctl | 16:44 |
| deathmist | xreactx[m]__: you sure you have that "ANDROID_VERSION_MINOR := 0" line in your hybris/hybris-boot/Android.mk ? I can only think of it failing if that was not set for 9.0 | 16:46 |
| deathmist | @BusterBg_18 do a "find / -name plat_property_contexts", for me I had to do "ln -s /system/etc/selinux/plat_property_contexts /plat_property_contexts" | 16:49 |
| deathmist | then reboot, get new journal | 16:51 |
| T42 | <BusterBg_18> I got find: /sys/kernel/slab/L2TP/IPv6: No such file or directory | 16:53 |
| deathmist | ok, do "find / -name plat_property_contexts 2>/dev/null" instead to hide all the useless stuff | 16:54 |
| deathmist | wait, nothing else then? | 16:54 |
| T42 | <BusterBg_18> I also gotfind: /sys/kernel/slab/L2TP/IPv: No such file or directory | 16:55 |
| T42 | <BusterBg_18> running 2>/dev/null returns nothing | 16:56 |
| deathmist | right. well, in that case run "find out/ -name *property_context*" on your hadk source tree, it should return some file in the root/ dir | 16:57 |
| deathmist | @BusterBg_18 looks like you'll also need straggler files too + mount units for /system and /vendor just like I did. check out https://git.io/JeBV4 https://git.io/JeBVB & https://git.io/JeBnQ while ofc adapting for your partitions, device codename in files etc | 17:01 |
| T42 | <BusterBg_18> Well It didnt fount anything either | 17:03 |
| T42 | <BusterBg_18> I do have straggler files tho | 17:03 |
| deathmist | once you've done the changes in the last msg I sent rebuild hybris-hal, build_packages -d & build_packages -c, then new mic build and flash the new build | 17:10 |
| deathmist | @BusterBg_18 you may also need to regenerate kickstarts, check hadk pdf for regular build steps :p | 17:25 |
| T42 | <BusterBg_18> No prob :p | 17:25 |
| T42 | <DylanVanAssche> My magnetometer is working now in sensorfwd (scaling overide was needed for the IIO sensor). | 17:36 |
| T42 | <DylanVanAssche> However, the X value is a factor 10 too small (in comparison to the Xperia X) and the Z value is a factor 60 too big (in comparison to the Xperia X). Is that 'normal' that this kind of sensor is way off between devices? | 17:36 |
| deathmist | @BusterBg_18 I'm looking at the changes and I think you already made a mistake on the mount units: the "What=" like should have a full valid path to your partition in it (e.g. "/dev/mmcblk0p79" or something), not something like "/mmcblk0p79" as that doesn't exist on your device | 17:37 |
| T42 | <BusterBg_18> You are right, I was missing /dev | 17:38 |
| T42 | <BusterBg_18> I forgot about how it was shown on init | 17:39 |
| deathmist | you can check that those paths exist on the device as it's booted currently still right? I think those are the same even in TWRP | 17:40 |
| T42 | <BusterBg_18> Yep I just checked on telnet and they exist | 17:43 |
| deathmist | once you boot the new image get output of "mount", it should now show your /system and /vendor etc nicely mounted | 18:03 |
| T42 | <BusterBg_18> indeed: /dev/mmcblk0p79 on /system type ext4 (rw,relatime) | 18:09 |
| T42 | <BusterBg_18> /dev/mmcblk0p80 on /vendor type ext4 (rw,relatime,data=ordered) | 18:09 |
| deathmist | alright, nice. now check systemctl status droid-hal-init | 18:11 |
| deathmist | it'll probably still fail, you can give updated journal again | 18:11 |
| T42 | <BusterBg_18> https://pastebin.com/Q8VCpPKU systemctl | 18:12 |
| deathmist | DHI definitely failed, probably since it's still missing /plat_property_contexts. try "find / -name *property_context* 2>/dev/null" on the device this time | 18:13 |
| T42 | <BusterBg_18> Yay, it now shows something: /vendor/etc/selinux/vendor_property_contexts | 18:14 |
| T42 | <BusterBg_18> /system/etc/selinux/plat_property_contexts | 18:14 |
| deathmist | progress :) now ln -s /system/etc/selinux/plat_property_contexts /plat_property_contexts | 18:14 |
| deathmist | and reboot | 18:14 |
| T42 | <BusterBg_18> Should I ignore the vendor one? | 18:15 |
| deathmist | the filename is not "plat_property_contexts" which is what it's looking for in last journal so yes | 18:16 |
| T42 | <BusterBg_18> Well so now im running journalctl -alb --no-pager but the output is bigger than what putty shows | 18:20 |
| deathmist | check DHI status again | 18:20 |
| T42 | <BusterBg_18> Ok, here is it https://pastebin.com/EfAhbCPE | 18:23 |
| T42 | <Akatsu %lastname%> @deathmist how im supposed to write mount for vendor if i have it inside /system ? | 18:25 |
| deathmist | well you don't need it then | 18:25 |
| deathmist | you only need vendor mount if you have a newer device with treble support | 18:25 |
| T42 | <Akatsu %lastname%> well i might get treble on it in future | 18:26 |
| deathmist | unlikely | 18:26 |
| T42 | <Akatsu %lastname%> (Photo, 353x105) https://irc.thaodan.de/.imgstore/GPoTE4bsGi.png very likely | 18:27 |
| T42 | <BusterBg_18> If people managed to do it for xiaomi_kenzo I think its possible for a lot of devices | 18:28 |
| deathmist | @BusterBg_18 how about journalctl? systemctl is quite useless especially if stuff fails | 18:30 |
| T42 | <Akatsu %lastname%> its just tons of work for them | 18:31 |
| jellypony | Hello everyone. I'm building SFOS for nitrogen (mi max 3). Hybris is 16. Got fail in droid-hal-init. I figured out that it is because /system cant be mounted. It seems like it is because fstab.qcom of my device don't have /system in it. What is an appropriate way to fix this? Just manually add it there and rebuild hybris-hal? | 18:34 |
| jellypony | my fstab.qcom: https://pastebin.com/raw/NqTYSrpG | 18:35 |
| T42 | <Akatsu %lastname%> @deathmist can i create these two mount files inside rootfs or do i need to create image with them? | 18:35 |
| deathmist | jellypony: check previous channel logs from me today, I listed a mount unit for /system which you'll need too | 18:35 |
| deathmist | @Akatsu sure, you can make them on the phone (omit sparse/ from paths in that case) but changes to droid-hal, spec or pattern files needs an image rebuild | 18:36 |
| deathmist | you only needed the one for /system and the ../ one is a symlink | 18:37 |
| T42 | <BusterBg_18> https://pastebin.com/ZjY3Uyz9 | 18:37 |
| T42 | <BusterBg_18> I had to use microsofts telnet client for me to get the full output | 18:37 |
| T42 | <Akatsu %lastname%> i did that and still i see no result sadly | 18:40 |
| deathmist | @Akatsu no /system mounted after restarting or systemctl start system.mount? did you change the device path to match your /system partition? | 18:44 |
| T42 | <Akatsu %lastname%> https://paste.ubuntu.com/p/KP6CgXVQ9B/ | 18:46 |
| deathmist | didn't I say ../ is a symlink? | 18:47 |
| T42 | <Akatsu %lastname%> uh im idiot once again | 18:47 |
| deathmist | the whole entry is a symlink to the file outside the .wants directory, this is how systemd works on every machine | 18:47 |
| T42 | <Akatsu %lastname%> i have to symlink from system to local-xxx rite? | 18:47 |
| deathmist | remove the old file with literal "../" in it and "ln -s ../system.mount /lib/systemd/system/local-fs.target.wants/system.mount" | 18:48 |
| T42 | <BusterBg_18> deathmist: do you dual boot sailfish and lineage? | 18:54 |
| deathmist | I do, for now even if my SFOS side is broken on 16.0 I can just switch boot target to LineageOS and at least have alarms etc working properly without having to wipe everything and set something else up | 18:55 |
| T42 | <Akatsu %lastname%> and still i get no system mounted | 18:56 |
| T42 | <BusterBg_18> I saw your boot switcher on your git that why I ask, much prettier than just flashing boot backup (thats what I do) | 18:57 |
| deathmist | @BusterBg_18 you should start checking channel logs between me and mal from yesterday as you seem to be in the same situation, specifically the ld.config.28.txt thing etc and see if they help you get further and have DHI not fail | 18:58 |
| deathmist | it's basically just a glorified dd command switching between boot.img backups in a flashable zip :) (tho I do other fixups for SFOS side too which I revert/apply depending on target) | 19:00 |
| T42 | <Akatsu %lastname%> https://paste.ubuntu.com/p/ZzTXDmjKxJ/ DMESG from mounting | 19:03 |
| T42 | <Akatsu %lastname%> function not implemented looks weird | 19:04 |
| T42 | <Akatsu %lastname%> it doesnt even want to mount by hand | 19:07 |
| T42 | <Akatsu %lastname%> mount -t ext4 /dev/block/mmcblk0p27 /system | 19:08 |
| deathmist | @Akatsu cuz you don't have the filesystem enabled in kernel https://git.io/JeBot also do "chmod 644 /lib/systemd/system/system.mount", not sure how you created that file but you definitely did it in a weird fashion. also does /dev/block/mmcblk0p27 exist or is it /dev/mmcblk0p27 ? | 19:08 |
| jellypony | deathmist: so for me line "what=dev/sde21" should look like "what=/dev/block/mmcblk0p77" | 19:10 |
| jellypony | Am i right? | 19:10 |
| T42 | <Akatsu %lastname%> again you are right, /dev/block/mmc* is symlink to /dev/mmc | 19:10 |
| T42 | <Akatsu %lastname%> just ran ls -l | 19:10 |
| deathmist | do you even have a /system directory then? what is the error it gives when you mount manually? | 19:10 |
| deathmist | jellypony: look about right, tho I think /dev/mmcblk0p77 would also work fine :) (check on device with ls /dev/mmcblk0p77 to make sure it exists) | 19:11 |
| T42 | <BusterBg_18> Will do, Im just going to first setup a clean lineage to use, afterall sirius its my only device atm | 19:11 |
| T42 | <Akatsu %lastname%> made system with mkdir, used `mount -t ext4 /dev/mmcblk0p27 /system` no single msg appear, and its still not mounted | 19:11 |
| deathmist | so you didn't have /system then? what does "mount" command output now? how about echo $? after you mount /system by hand? | 19:12 |
| T42 | <Akatsu %lastname%> when i have no system inside rootfs and i do command mentioned before i get `mount: /system: mount point does not exist.` | 19:13 |
| deathmist | well obviously since it needs a directory to exist so it can mount it | 19:14 |
| T42 | <Akatsu %lastname%> thats why i made system directory with mkdir | 19:15 |
| T42 | <Akatsu %lastname%> and it gave no error but not mounted at same time | 19:15 |
| deathmist | how are you checking if it's not mounted? ls /system? and you still haven't give me output of "mount" | 19:15 |
| T42 | <Akatsu %lastname%> i check by mount and ls -l, secundo | 19:16 |
| T42 | <Akatsu %lastname%> https://paste.ubuntu.com/p/kVJVPnQr3x/ | 19:18 |
| deathmist | keep /system created and reboot now, the system.mount unit should do it's work unless you're missing something else maybe in kernel defconfig | 19:21 |
| T42 | <Akatsu %lastname%> i miss autofs4 | 19:21 |
| T42 | <Akatsu %lastname%> do you recommend to get all warnings done? | 19:22 |
| deathmist | sure | 19:22 |
| T42 | <Akatsu %lastname%> halium have nice auto fixing script | 19:22 |
| T42 | <Akatsu %lastname%> mka hybris-boot will be enought? | 19:23 |
| deathmist | yep and flash hybris-boot.img to boot partition after to check, .config in out/ is also useful to see if your options actually get set in boot image | 19:24 |
| deathmist | I reckon the halium config option fixup script could easily be adapted to also work with mer-hybris defconfig changes. some of them rely on specific kernel version etc too tho which could be more difficult to detect in some cases. all the data is here tho https://git.io/JeBoR | 19:26 |
| deathmist | also all of those still don't even guarantee working telnet because of configfs vs old android USB method of doing stuff etc, a bunch of stuff is very kernel and device specific | 19:29 |
| T42 | <Akatsu %lastname%> thats same for halium, even with their script we have to look at specific parameters like config_dummy not being modified | 19:37 |
| T42 | <BusterBg_18> I want to give halium a try just for ubports, but being stuck with either 5.1 or 7.1 is a bummer | 19:39 |
| T42 | <Akatsu %lastname%> i got like 2 devices to boot ubports on 7.1 | 19:40 |
| T42 | <Akatsu %lastname%> having it working is different story | 19:40 |
| T42 | <BusterBg_18> So it also requires more modifications? I thought it was build and good to go like GSIs on Android | 19:42 |
| T42 | <Akatsu %lastname%> not really | 19:42 |
| T42 | <Akatsu %lastname%> for example audio is issue sometimes | 19:42 |
| T42 | <Akatsu %lastname%> ril/modem | 19:42 |
| T42 | <Akatsu %lastname%> however getting device to boot os is much easier | 19:43 |
| T42 | <Akatsu %lastname%> i have phone on msm8916 and with newest device tree i have issues with modem so i get audio ripped too, mer-hybris init changes even blocked executing init files because ro.hardware was definied as unknown, i had to make dirty-hacks on it. | 19:44 |
| T42 | <BusterBg_18> That explains why I don't hear about any device having usable ubports | 19:48 |
| T42 | <Akatsu %lastname%> one Xperia seems to be in pretty good shape | 19:48 |
| T42 | <BusterBg_18> I bet that Xperia is suzu/X | 19:51 |
| T42 | <Akatsu %lastname%> yas | 19:52 |
| T42 | <Akatsu %lastname%> `[ 412.866692] EXT4-fs (mmcblk0p27): mounted filesystem with ordered data mode. Opts: (null)` and still nothing, is it even normal? | 19:54 |
| deathmist | normal message, I don't understand how /system still could not be mounted, check systemctl status system.mount | 19:56 |
| T42 | <Akatsu %lastname%> `system.mount: Unit is bound to inactive unit dev-block-mmcblk0p27.device. Stopping, too.` | 19:57 |
| mal | oh, that should not have block in it, show your fixup-mountpoints file | 19:58 |
| mal | also fstab from your device repo | 19:58 |
| T42 | <Akatsu %lastname%> i have pasted all possible mountpoints in it | 19:59 |
| T42 | <Akatsu %lastname%> https://paste.ubuntu.com/p/XxFZK4zZWz/ | 19:59 |
| T42 | <Akatsu %lastname%> https://github.com/LineageOS/android_device_leeco_s2/blob/lineage-15.1/rootdir/etc/fstab.qcom | 20:00 |
| mal | show the file /lib/systemd/system/system.mount from your device | 20:01 |
| T42 | <Akatsu %lastname%> https://paste.ubuntu.com/p/KP6CgXVQ9B/ | 20:02 |
| T42 | <Akatsu %lastname%> ignore first i made it symlink already | 20:02 |
| T42 | <Akatsu %lastname%> wait | 20:02 |
| mal | and that was made automatically? | 20:03 |
| mal | or did you somehow manually edit that? | 20:04 |
| mal | because that shouldn't have block/ in the "What" part | 20:04 |
| T42 | <Akatsu %lastname%> https://paste.ubuntu.com/p/cr6YQSrPyp/ | 20:06 |
| T42 | <Akatsu %lastname%> i changed it to /dev | 20:06 |
| T42 | <Akatsu %lastname%> because one inside block was just symlink | 20:06 |
| T42 | <Akatsu %lastname%> ok | 20:07 |
| T42 | <Akatsu %lastname%> something is wrong | 20:07 |
| T42 | <Akatsu %lastname%> gonna change again | 20:07 |
| T42 | <Akatsu %lastname%> seems like it works | 20:10 |
| T42 | <Akatsu %lastname%> at the end i needed to fix autoFS4 for it anyway | 20:10 |
| T42 | <Akatsu %lastname%> for what do we have fixmountpoints if we have to deal with it inside sailfishOS? | 20:15 |
| deathmist | hadk pdf chapter 5.3 explains this | 20:18 |
| mal | you have data partition as something else than ext4? | 20:19 |
| deathmist | mal: hist last mount output shows it at least as ext4 https://paste.ubuntu.com/p/kVJVPnQr3x/ | 20:20 |
| T42 | <Akatsu %lastname%> is it normal to have plat and nonplat_property_contexts instead of property_contexts ? | 20:29 |
| mal | yes, new android bases have different names for contexts files | 20:30 |
| T42 | <Akatsu %lastname%> and this should work then? | 20:30 |
| T42 | <Akatsu %lastname%> do i have to do manual booting on rest partitions on which android-hal-init failed to get it working? | 20:31 |
| mal | what do you mean? | 20:31 |
| mal | you can check if those mount services have correct device node | 20:32 |
| T42 | <Akatsu %lastname%> i get errors on odm, android-hal-init strace shows me errors about not being unable to open or find file | 20:32 |
| mal | if not then something is wrong in your build, based on the fstab and fixup those should have been ok, unless it find wrong fstab, can you check output of make hybris-hal what it says after "/data appears to live on" | 20:35 |
| T42 | <Akatsu %lastname%> I will look on that tomorrow | 20:37 |
| T42 | <BusterBg_18> mal: I dont have a ld.config.28.txt on my device instead its called ld.config.vndk_lite.txt, should I just change droid-hal-early-init.sh? | 21:33 |
| T42 | <BusterBg_18> The contents of that are similar to the ones in ganges and cheeseburger | 21:42 |
| mal | I think the droid-hal-early-init.sh didn't mount it properly so for now try remounting /system manually as rw, backup the original file and the edit it as mentioned earlier (add the lines with libexec in ganges repo to your file) in /system and the remount as ro again and it hopefully helps | 21:57 |
| T42 | perrsona1234 was added by: perrsona1234 | 22:10 |
| T42 | <perrsona1234> Hi all | 22:10 |
| T42 | <perrsona1234> I have a question. In Your opinion: is Sailfish more mature than Ubuntu Touch? | 22:12 |
| T42 | <perrsona1234> Because I'm currently thinking about buying a "new" phone, because I badly damaged my current one. | 22:12 |
| T42 | <BusterBg_18> would this work out? https://pastebin.com/t2irU4ph | 22:24 |
| T42 | <BusterBg_18> It seems that ld.config28 and vndk_lite are more different after all | 22:26 |
| mal | @BusterBg_18 not sure if you need that or not, depends on what the actual error you have is | 22:29 |
| T42 | <BusterBg_18> well, DHI is failing https://pastebin.com/ZjY3Uyz9 | 22:33 |
| T42 | <BusterBg_18> I was trying th ld.config stuff to see if it would help | 22:33 |
| mal | @BusterBg_18 can you check the /usr/libexec/droid-hybris/system/etc/init/servicemanager.rc file for a line with selinux_stubs | 22:58 |
| mal | also did you check if out/ has any files related to fstab and ueventd*.rc | 23:00 |
| mal | there are several things that can be wrong | 23:01 |
| T42 | <BusterBg_18> I have setenv LD_PRELOAD /usr/libexec/droid-hybris/system/lib64/libselinux_stubs.so | 23:02 |
| *** jellypony1 is now known as jellypony | 23:03 | |
| T42 | <BusterBg_18> And ueventd.rc is present under out/*/sirius/root | 23:07 |
| jellypony | deathmist: oh shi~ its booted | 23:10 |
| jellypony | but no gui and no open ports on 192.168.2.1 | 23:11 |
| jellypony | D: | 23:11 |
| jellypony | mal: Hi! Can you help me? What exactly can cause this? Device probably booted (it vibrates!!!) after few changes in systemd via telnet (mounting /system). But now there is no gui and i cant get any access to it. There is only 192.168.2.1 IP i can ping and all it ports are closed. | 23:16 |
| mal | @BusterBg_18 check output of "make modules" (same environment where you run make hybris-hal) and see the output for fstab or ueventd* | 23:19 |
| mal | jellypony: the ip is 192.168.2.15 and port 2323 | 23:19 |
| jellypony | it was... | 23:21 |
| T42 | <BusterBg_18> I dont see fstab or ueventd when running make modules | 23:22 |
| mal | just to be sure, what kind of changes did you do to kernel defconfig, go to the kernel source directory (somewhere under kernel/ in ANDROID_ROOT) and run "git diff" and show the output of that | 23:25 |
| T42 | <BusterBg_18> https://pastebin.com/G7wsB9rn | 23:28 |
| T42 | <BusterBg_18> But thats just the most recent change isnt it? | 23:30 |
| mal | you had some older changes you commited to git already? | 23:33 |
| T42 | <BusterBg_18> Yes | 23:34 |
| mal | what device was that again? | 23:34 |
| mal | I used to remember what device everyone had but not anymore | 23:34 |
| jellypony | mal:I used nmap. The only IP is 192.168.2.1. No open ports D: | 23:35 |
| T42 | <BusterBg_18> xiaomi_sirius | 23:36 |
| mal | jellypony: so .15 doesn't work? | 23:43 |
| jellypony | no :c | 23:45 |
| mal | does the device show anything in host dmesg, some usb related messages? | 23:49 |
| jellypony | mal: yes! https://pastebin.com/9aAG3Uxm | 23:52 |
| mal | to me that suggests it rebooted after 30 seconds | 23:55 |
| mal | hmm, actually the timestamp is missing from the disconnect but I assume it's close to the next line | 23:56 |
| jellypony | [66231.813345] | 23:57 |
| jellypony | my bad | 23:57 |
| mal | jellypony: which android base? | 23:58 |
| jellypony | pie | 23:58 |
| jellypony | hybris 16 | 23:58 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!
