T42 | <foxclore> Good morning everyone! A question: successfully built Sailfishos, trying to boot it | 09:30 |
---|---|---|
T42 | <foxclore> Flashed on phone via TWRP (it said everything is alright), /data/.stowaways/ has a lot of files | 09:30 |
T42 | <foxclore> | 09:30 |
T42 | <foxclore> Yet after flashing phone doesn't boot into system, stuck on Samsung Galaxy S8 logo, reboot after a minute or so. | 09:30 |
T42 | <foxclore> Dmesg doesn't show any new device and neither sudo lsusb -v | grep iSerial does. Any help please? | 09:30 |
T42 | <Ayman_rgab> Force reboot to recovery and check /proc/last_kmsg | 09:33 |
T42 | <foxclore> Anything specific to look for? (re @Ayman_rgab: Force reboot to reco...) | 09:38 |
T42 | <Ayman_rgab> No | 09:42 |
T42 | <elros34> if usb rndis connection cannot be created then hybris initramfs will reboot device after 60s: https://github.com/mer-hybris/hybris-boot/blob/master/init-script#L273. Check usb/configfs/gadget/rndis related configs enabled in kernel | 10:19 |
SebastianoGaleazzo[m] | Hi everyone... As per hadk-hot after "make hybris-hal droidmedia" under out/.../$DEVICE/root/ folder there should be fstab and ueventd.rc files... If they are under out/.../$DEVICE/vendor/etc/ folder is it the same or they must be copied to out/.../$DEVICE/root/? | 10:43 |
mal | SebastianoGaleazzo[m]: these are the paths that it searches https://github.com/mer-hybris/droid-hal-device/blob/master/droid-hal-device.inc#L484 similar for fstab a few lines later in the file | 11:31 |
SebastianoGaleazzo[m] | Great so nothing to do since they are already there... Thank you! | 11:35 |
T42 | <foxclore> So, I added some configs in defconfig, after rebuilding kernel was rebuilding packages | 14:33 |
T42 | <foxclore> and build_packages.sh —gg failed with error about missing disabled_services.rc | 14:33 |
T42 | <foxclore> | 14:33 |
T42 | <foxclore> Here is the log: https://pastebin.com/EmXjR7kE | 14:33 |
T42 | <foxclore> | 14:33 |
T42 | <foxclore> Any help please? (re @elros34: if usb rndis connect...) | 14:33 |
mal | you didn't use the suggested version of droidmedia | 14:50 |
T42 | <foxclore> mal: suggested version is 0.20211101.0 (im building for 4.4.0.x) | 15:08 |
T42 | <foxclore> I did: | 15:08 |
T42 | <foxclore> [HADK] | 15:08 |
T42 | <foxclore> cd $ANDROID_ROOT/external/droidmedia | 15:08 |
T42 | <foxclore> git pull github 0.20211101.0 | 15:08 |
T42 | <foxclore> cd $ANDROID_ROOT | 15:08 |
T42 | <foxclore> make droidmedia | 15:08 |
T42 | <foxclore> | 15:08 |
T42 | <foxclore> then went to PlatformSDK | 15:08 |
T42 | <foxclore> rpm/dhd/helpers/build_packages.sh --gg | 15:08 |
T42 | <foxclore> | 15:08 |
T42 | <foxclore> it failed again with the same file missing | 15:08 |
mal | did you check the droidmedia repo is really in correct revision | 16:06 |
mal | the earlier build log had "Processing files: droidmedia-0.20220929.0-1.aarch64" | 16:06 |
T42 | <elros34> @foxclore you don't need to build everything again to test kernel. Just make hybris-hal and then flash hybris-boot.img | 18:29 |
T42 | <foxclore> @elros34 | 19:41 |
T42 | <foxclore> In source files in that folder there are a lot #ifdef CONFIG_..._RNDIS_... | 19:41 |
T42 | <foxclore> (for instance CONFIG_USB_RNDIS_MULTIPACKET), those #ifdefs are related to the values in my defconfig? (CONFIG_USB_RNDIS_MULTIPACKET=y => CONFIG_USB_RNDIS_MULTIPACKET is defined?) (re @elros34: if usb rndis connect...) | 19:41 |
T42 | <elros34> what source file are you talking about? No idea what_multipacket enables but unlikely you need it for telent | 19:43 |
T42 | <foxclore> kernel/samsung/universal8895/drivers/usb/gadget/function | 19:45 |
T42 | <foxclore> I did grep for all config*, trying to find which ones are related (re @elros34: what source file are...) | 19:45 |
T42 | <elros34> first check what your device use in device/*/*/*/*usb.rc configs or android0 | 19:46 |
T42 | <elros34> configfs* | 19:46 |
T42 | <elros34> you can also boot recovery and see whether you have /sys/class/android_usb/android0 | 19:50 |
T42 | <elros34> then check existing defconfig, do not look into kernel sources, this will not help you | 19:50 |
T42 | <foxclore> so, there are many lines like | 20:27 |
T42 | <foxclore> on property:sys.usb.config=ncm,adb && property:sys.usb.configfs=1 | 20:27 |
T42 | <foxclore> | 20:27 |
T42 | <foxclore> can you tell me please what do i do with them? (re @elros34: first check what you...) | 20:27 |
T42 | <elros34> can you show whole file? | 20:28 |
T42 | <foxclore> https://pastebin.com/EX0CYAzN | 20:32 |
T42 | <elros34> you didn't provide github url so I would not show you exact line but you can see that configfs is used but there is also android0 involved. So first check what options related to configfs you have in kernel .config. You may need to modify init-script, I thing there were quite few ports with similar configfs and android0 | 20:37 |
T42 | <foxclore> @elros34 | 20:49 |
T42 | <foxclore> these are the only options related to configfs in .config: | 20:49 |
T42 | <foxclore> https://gist.github.com/foxclore/57da2f2895e01fd8daa47f51a6ba8c30 | 20:49 |
T42 | <elros34> so I guess you need to make some modification to init script, search in channel logs, for sure there were something about it, maybe even for some samsung so you would not have to reinvent the wheel. For now probably this line is executed: https://github.com/mer-hybris/hybris-boot/blob/master/init-script#L244 but you may need usb_setup_configfs instead or some kind of mix. All of this is true only if you really have problem with usb/rndis not | 20:55 |
T42 | <foxclore> Well, I measured time before reboots, it's literally 60+-2 seconds (re @elros34: so I guess you need ...) | 20:58 |
T42 | <elros34> so most likely init script mess with you but you can always modify it to be sure before modifying usb part | 21:00 |
T42 | <elros34> it -> sleep 60 | 21:00 |
T42 | <foxclore> @elros34 can confirm, 1) Hybrid-Boot flashes correctly | 21:30 |
T42 | <foxclore> The problem is indeed with usb part, changing sleep time confirms | 21:30 |
T42 | <foxclore> @elros34 can confirm, 1) Hybris-Boot flashes correctly | 21:31 |
T42 | <foxclore> The problem is indeed with usb part, changing sleep time confirms (edited) | 21:31 |
SebastianoGaleazzo[m] | Another question... While building droid-hal I noticed these errors:... (full message at <https://matrix.org/_matrix/media/r0/download/matrix.org/gfyzPoyrkJJjlfQzdteyfxeZ>) | 21:59 |
T42 | <elros34> most likely | 22:05 |
T42 | <elros34> there is one alternative path already, 3th is probably needed | 22:06 |
T42 | <elros34> I wonder if your device use flattened apexes because this path sounds familiar to my issue | 22:07 |
SebastianoGaleazzo[m] | How can I check if it uses it? | 22:08 |
T42 | <elros34> not sure about any certain way, I think there were so prop value but for sure you will know if boringssl will fail | 22:09 |
SebastianoGaleazzo[m] | Ok so I'll continue the build with the above fix and then I'll see and let you know | 22:10 |
T42 | <elros34> could be this: android::sysprop::ApexProperties::updatable, you need to find what is hidden behind this | 22:14 |
T42 | <elros34> or grep getprop in android | 22:14 |
SebastianoGaleazzo[m] | Ok just found it (I think) "ro.apex.updatable=true", is it? | 22:17 |
T42 | <elros34> if it's true then you should be good | 22:18 |
SebastianoGaleazzo[m] | Yes it's true | 22:18 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!