Tuesday, 2021-03-30

T42<mintphin> @mintphin [should i try another kernel perhaps?], i did that and still the same issue03:22
T42<mintphin> so uh i'm completely lost now03:22
T42<mintphin> should i compile the kernel separately perhaps?03:22
T42<mintphin> and then add the ramdisk?03:23
T42<zinstack625> You can try to unpack boot.img, to make sure initramfs is there (I think it should be there). Quite likely your kernel config is so broken, it can't load initramfs05:30
T42<zinstack625> @mintphin ^^05:31
T42<zinstack625> Initramfs is the first thing a kernel tries to do, apart from initializing hardware, and the first sign it gets there is device being recognized as "mer boatloader" or whatever05:34
T42<zinstack625> Also, you could try doing this, although I never tried just flipping the variable05:40
T42<zinstack625> https://github.com/sailfish-motog7/android_kernel_motorola_sdm632/commit/c9952d1d9fa602e6a710cd0968f636fd8662efdb05:40
*** techgeekster_ is now known as techgeekster05:41
*** wiktorek141 is now known as wiktorek14006:29
*** ZucZero is now known as Zucca06:57
techgeeksterelros34 So this is the journal (https://paste.ubuntu.com/p/Hsn2qXY4qd/) The weird things I've seen is the +SELINUX (even though I have audit=0 and selinux=0) and the mount failures for /vendor/dsp, /vendor/firmware_mnt, ld.config.28.txt07:32
T42<zinstack625> techgeekster: your main trouble for now is that ld.config.28.txt is not mounted. What exactly are you mounting? It seems that ld.config.28.txt you're trying to mount doesn't exist07:42
T42<zinstack625> The other two may be symlinks to some other place, so, perhaps, if your device's vendor works like this, these mounts may not be necessary07:58
techgeeksterzinstack625 In hybris/droid-configs/sparse/usr/lib/systemd/system/ I have the system-etc-ld.config.28.txt.mount that mounts the /usr/libexec/droid-hybris/system/etc/ld.config.28.txt file as the faq says. Should It be in another directory?08:20
T42<zinstack625> Does /usr/libexec/droid-hybris/system/etc/ld.config.28.txt exist?08:34
T42<zinstack625> techgeekster: AFAIK, you should be putting it there yourself in sparse08:35
techgeeksterI did and it was on the device. I dont know why it failed. By removing the mount service i have fewer errors. No the most "important" is the health.service that fails and brings down droid-hal-init08:37
T42<zinstack625> That should be in disabled_services.rc08:38
techgeeksterLemme check. Maybe it does not parse it08:38
techgeeksterThe service is not disabled in the disables_services.rc in my device, neither in the https://github.com/sailfishos-oneplus5/droid-config-cheeseburger/blob/hybris-16.0/sparse/usr/libexec/droid-hybris/system/etc/init/disabled_services.rc of the faq. Should I add it?08:41
T42<zinstack625> I think so08:41
T42<zinstack625> Strange, I recall healthd being disabled somewhere, but can't find that anywhere now08:55
T42<zinstack625> It could just be necessary08:56
T42<zinstack625> Fixing it would def be beneficial08:57
T42<zinstack625> Try stracing it, seeing how it fails in logcat and perhaps debugging it, maybe something will show up08:58
techgeeksterzinstack625 By adding the health in the disabled services now the droid-hal-init runs. Still no UI, have a lot of services failing in logcat with this error Could not get passthrough implementation for android.hardware09:28
T42<zinstack625> Are you sure the ld.config.28.txt is mounted and correct?09:29
techgeeksterLet me check. I bet no.09:31
T42<elros34> this ldconfig mount is not always needed, only when you have specific error in logcat which you don't have. Also probably you don't even have that file which you want to mount over09:35
T42<elros34> +AUDIT is from systemd so this is fine as long as you have selinux disable via bootparam but I can't see that message in your logs because skbuf warning flood. Maybe is not the best way but you could just temporary coment out skbuff.c:616 so at leas you will have cleaner logs09:40
techgeeksterOk elros34 . Will do09:48
T42<elros34> Is that full aarch64 build or did you use 32bit sailfish target?10:03
techgeeksterHow can I check the sfos target? I remember it was aarch64, but I could check...10:08
T42<elros34> no need to double check it. What about /init.qcom.rc. Do you have it both in device and in out/?10:10
techgeeksterYes. It is in the out/target/product/zenfone3/system/vendor/etc/init/hw/init.qcom.rc10:17
T42<elros34> ok so that probably not relevant10:19
techgeeksterI flashed the updated kernel with the comment. We should have a clear journalctl now10:23
techgeeksterNot crystal clear but it should be easier to follow https://paste.ubuntu.com/p/jfXjHGDsth/10:29
T42<elros34> so at least it's clear that selinux is disable correctly10:39
T42KathrineKarlsen was added by: KathrineKarlsen10:52
T42LaurieJGray was added by: LaurieJGray10:53
T42LouiseHansen was added by: LouiseHansen10:54
techgeeksterCould it be a kernel thing? Maybe I added something useless in the defconfig that causes these errors?11:09
T42<elros34> @techgeekster vendor keymaster fails, don't know how to fix it but you can at least try to disable vendor.keymaster-3-0 in disabled_services.rc to see if that change anything (especially these binder: transaction failed errors which appears just before keymaster fails)11:11
techgeeksterOh ok. elros34 will try it. Thank you!11:12
T42<elros34> there is also one more related service wait_for_keymaster.rc11:13
T42<mintphin> @zinstack625 [https://github.com/sailfish-motog7/android_ker …], i don't see anything that could be breaking it at first glance though https://github.com/Sailfish-OS-Mi-A3/android_kernel_xiaomi_laurel_sprout/blob/hybris-17.1/arch/arm64/configs/vendor/laurel_sprout-perf_defconfig11:52
T42<elros34> @mintphin I might be wrong but if you device reboots to fastboot immediately without any changes in defconfig then I don't think you should search what is wrong in config. How long  it needed to reboot after you flashed boot.img? Did you see anything in dmesg?12:03
T42<mintphin> @elros34 [@mintphin I might be wrong but if you device r …], like 3 seconds.12:08
T42<mintphin> i could see the "unlocked" text which means it did probably attempt to boot from the initramfs12:08
T42<mintphin> but failed12:08
T42<mintphin> i can't unpack the boot image either12:08
T42<mintphin> my unpacker can't detect it as a qualcomm image12:10
T42<elros34> in platform sdk there is split_bootimg tool (android-hadk-tools-bootimg or something similar)12:28
T42<mintphin> (Photo, 652x579) https://irc.thaodan.de/.imgstore/JZ0OiPZtMb.png hmmm12:44
T42<mintphin> it's weird12:44
T42<mintphin> this file weighs 15 mb12:44
T42<mintphin> but has only 7 mb worth of actual data12:44
T42<zinstack625> Probably just padding13:10
T42<zinstack625> Still, there doesn't seem to be any initramfs there13:10
T42<zinstack625> TEMPORARY_DISABLE_PATH_RESTRICTIONS should take care of that mostly. I know you've heard it so many times, but try doing "make clean", setting that flag correctly and make hybris-hal one more time13:13
T42<zinstack625> Bootimg for river weighted around 30mb, I'll try unpacking that after getting home13:14
T42* K %lastname%13:35
T42* zinstack62513:35
T42<mintphin> @zinstack625 [Still, there doesn't seem to be any initramfs …], initramfs was in another file13:36
T42<zinstack625> So it is there after all13:37
T42<mintphin> yup. it weights 1.5 mb though13:37
T42<elros34> @mintphin I still don't get where is the culprit but maybe is worth to try: CONFIG_INITRAMFS_SOURCE=$ANDROID_ROOT/out/target/product/$DEVICE/obj/ROOT/hybris-boot_intermediates/boot-initramfs.gz (expand variables) but I am not sure that will work correctly.14:25
malhey, if the boot image is not working correct are you sure the proper environmental variable is used (if that is for hybris-17.1)14:38
malyeah, that TEMPORARY_DISABLE_PATH_RESTRICTIONS mentioned earlier14:38
T42<zinstack625> My initrd is 987K, and it works fine14:41
T42<nikio> Hi! How can i debug fingerprint sensor? It sucessfully read the finger print but can not to save it21:02
piggz@nikio logcat probably ... maybe need to create some folders in /data/...21:04
T42<nikio> (Photo, 1155x175) https://irc.thaodan.de/.imgstore/05PE8RoyGv.png sadly, but logcat don't show anythink about it. so, daemon seems to be sucessfully started.. can i share the video of process with error?21:08
poetasterjeez. shot self in foot. how the hell to you create an account for the jolla store after reflashing?21:09
T42<zinstack625> @nikio best to just throw a logcat. Video wouldn't be handy or useful for anyone21:17
T42<nikio> @zinstack625 [@nikio best to just throw a logcat. Video woul …], is there any special attributes for logcat? or just reboot and try to add finger?21:28
malpoetaster: what do you mean?21:28
T42<nikio> (Document) https://irc.thaodan.de/.imgstore/qaqmmM00v0.txt21:30
T42<nikio> @nikio [<reply to media>], wrong path to (/data/vendor_de/0/fpdata), but only folder i have in vendor_de is "100000". so i must to change destination folder but i dont know how to do it21:34
T42<nikio> thank you all! simple ln -s solved the problem22:13

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