Sunday, 2023-02-26

T42<vsvoid> Any idea what this error means00:30
T42<vsvoid> *Error reading from Uevent Fd: I/O error*00:30
T42<vsvoid> 00:30
T42<vsvoid> and how to fix it?00:30
malis that normal build?00:35
T42<vsvoid> Yeah my very first build00:36
malthe folder and mount service is created during droid-hal packaging based on fstab00:36
malcan you link your device repo which contains the fstab00:36
malwhere is that "Error reading from Uevent Fd: I/O error" error?00:37
T42<vsvoid> I have dynamic partition so before I figure out how to package that, I mounted system vendor and product from img files got from the running system00:37
T42<vsvoid> I saw the error as I was going through the  `journalctl`00:38
malwhat issue are you trying to fix now?00:39
malif you manually mounted the system.img then maybe you do the same as I mentioned i.e. mount it to /system_root and then bind mount /system_root/system to /system00:39
T42<vsvoid> Am trying to figure out why my device is bootlooping00:40
malhow soon is it restarting?00:40
T42<vsvoid> Yes I created /system_root mount bind /system_root/system to /system00:40
T42<vsvoid> About a minute or less, but just before it reboots the screen dims a little00:41
T42<vsvoid> It's not giving me enough time to even telnet00:41
malcan you access the rootfs from recovery? maybe disable some things like droid-hal-init.service, also make sure it actually tries to get that far in the boot, do you see anything in host dmesg about the usb device, it could even stop at first telnet which is port 2301:18
T42<vsvoid> Yes, and that's how am getting the logs from journelctl ,It is going as far as droid-hal-init, and I have checked dmesg it is rebooting after it goes into telnet 2323 which means it is booting into rootfs01:21
T42<elros34> why didn't you use parse-android-dynparts for dynamic partitions? Less chance you will forget about something critical. So what are the first errors from journal/dmesg?01:35
T42<vsvoid> I could find the script, the links I found were all saying about not found01:37
T42<elros34> ineed script is gone but still instruction from parse-android-dynparts Redme is valid. I bet you can find script in the sailfish-on-nabu or some other device which used it. That error you have is probably because something important fails is without context is not very useful01:39
T42<elros34> I will search tomorrow I have the script somewhere if nobody will provide you it01:44
T42<vsvoid> @elros34  Oh okay, I was trying to send a pastebin of journelctl but it's too big02:05
T42<vsvoid> But from what I see droid-hal-init fails and is causing a reboot02:05
mal@elros34 btw, I have trouble with parse-android-dynparts, I tried to run losetup and dmsetup in a service and the mounts in normal mount services but first service always fails for some reason02:06
mal@vsvoid did you do the needed selinux changes mentioned in faq?02:06
T42<vsvoid> Yes02:07
T42<elros34> mal: frankly I have never use it. Added script for porters who struggle to get dyn parat  working but AFAIK they mount everyting in /init-script02:08
T42<elros34> @vsvoid upload dmesg/journal to gist if you want any help02:08
mal@elros34 looks like most do the mounts in droid-hal-early-init.sh02:08
malI wanted to do the mounts earlier to get firmware in place for module probing etc02:09
T42<elros34> definitely, I am not sure it will even work in droid-hal-early-init.sh02:09
malmaybe I have some problem with timing of the initial setup service02:10
T42<vsvoid> @elros34
T42<elros34> instead rebooting to recovery and chrooting  just create /init_enter_debug2 file02:25
malapexd fails based on that log which causes reboot02:25
T42<vsvoid> Okay02:31
T42<elros34> unfortunately hybris patches are missing some changes so if apex fails then ueventd starts and everything fails02:32
mal@elros34 patches are always welcome02:33
malbut apexd should not fail02:33
T42<vsvoid> I have to do a fresh build don't I?02:33
T42<elros34> did you apply kernel patch for loop_set_status error you have?02:40
T42<vsvoid> No I did not patch that but that error just shows there down the log you will see the img actually mount02:45
T42<elros34> but as in instruction: it cause apexd failure, ueventd start and reboot02:47
Aboothahir[m]mal:  hi02:54
T42<elros34> @vsvoid link updated10:48
T42<vsvoid> Which one?10:51
T42<elros34> the one we were talking about which mounts dynamic partitions10:51
mal@elros34 I managed to get dynamic partition mounts working via systemd mount services and one extra service run before those12:33
T42<elros34> great!12:52
T42<vsvoid> how do i make parse-android-dynparts  build with system14:49
T42<elros34> like any other package with IIRC parameters clone manually and use -b switch with path to sources or full git url with --mw option14:55
HengYeDev[m]I am trying to ssh into device, but don't have defaultuser password. telnet 23 and 2323 are connection refused14:58
HengYeDev[m]btw this is forked from so should i disable encryption before going further14:59
T42<elros34> but do you have even telnet connection setup for this? Are you sure it's not disabled15:02
HengYeDev[m]is telnet provided in boot.img or in the rootfs?15:04
HengYeDev[m]reason i'm asking is that the hybris-boot is swapped with halium-boot ramdisk here15:04
T42<elros34> it's both in hybris-boot.img initramfs (23 port) and /init-debug in rootfs but can be disabled so check host dmesg15:04
HengYeDev[m]host dmesg has nothing related to telnet or mer, but i see device name and register rndis_host15:06
HengYeDev[m]the screen can be turned on and off so i think mce is on15:06
T42<elros34> I guess is not enabled then15:07
HengYeDev[m]/init is symlinked to /sbin/preinit15:08
HengYeDev[m]is telnet enabled before preinit?15:08
T42<elros34> yeah15:08
T42<elros34> you must check what is executed in halium-boot ramdisk15:08
T42<elros34> if you can chroot to rootfs then you could probable set password for defaultuser and use ssh15:09
HengYeDev[m]halium-boot runs /init (symlinked to preinit)15:11
HengYeDev[m]i will try chroot15:11
HengYeDev[m]cant chroot due to architecture mismatch15:48
T42<elros34> don't you have some halium recovery image?15:53
HengYeDev[m]not yet15:54
HengYeDev[m]and no twrp either for now15:54
HengYeDev[m]edit /etc/shadow directly?15:55
T42<elros34> I am not sure, maybe so set password after/during image creation but I can not believe there is nothing to debug15:55
T42<vsvoid> I must be missing something here, am building and am getting the following errors16:50
T42<vsvoid> 16:50
T42<vsvoid> Do I have to clone the git repo into external/droidmedia like in tutorial for android 11 here
T42<elros34> you already have it there, search this channel logs, this is known error. You need downgrade droid-hal submodule16:55
T42<vsvoid> Okay16:55
HengYeDev[m]in the ks file is it possible to set passwd? does not work since user is no longer there18:11
T42<Verevka86> revert (re @vsvoid: I must be missing so...)18:24
T42<elros34> @HengYeDev[m] probably no because user is created at first boot so you would have to add oneshot script or something like that but why don't you switch to awesome hybris-boot.img + /init-debug script.18:31
T42<HengYeDev> im trying to use the halium method since we already have ubports port with many needed hacks18:33
T42<HengYeDev> i wil see if i can connect init-drbug with halium boot18:33
T42<elros34> it's just a script18:34
T42<elros34> btw it's defaultuser not nemo unless you have changed it in some startup file18:34
T42<HengYeDev> yes i changed the username but it still does not work18:35
T42<elros34> I think starting /init-debug script instead /sbin/preinit might be enough18:37
HengYeDev[m]oh wait don't have init-debug since no droid-hal18:41
T42<elros34> not a big deal, copy it18:42
HengYeDev[m]i have copied it and symlinked /init -> /init-debug19:00
HengYeDev[m]telnet still refuses connections on both 23 and 232319:00
HengYeDev[m]also is it a sailfish thing that powered by android logo becomes purple?19:10
T42<edp_17> Is it possible to have nemo as default user? 😮 What do I need to change for that? (re @elros34: btw it's defaultuser...)19:11
T42<elros34> depends on whether you are doing it after or before first boot19:14
T42<elros34> @hengYEDev[m] so you should have now /init.log19:14
T42<HengYeDev> alright, this is why i should have got the recovery working first19:15
T42<elros34> do you have busybox-static installed?19:18
T42<elros34> removed /init_disable_telnet if you have it19:20
T42<elros34> simlink instead real file might change how script is executed, did you change DONE_SWITCH to yes?19:23

Generated by 2.17.1 by Marius Gedminas - find it at!