T42_ | <dgmarie2> actually after unpacking my boot img it looks like there is no initramfs | 01:10 |
---|---|---|
T42_ | <dgmarie2> what are some common reasons why hybris might not build the initramfs into the boot img? im using an android 10 base | 01:11 |
T42_ | <dgmarie2> /bin/bash: cpio: command not found | 01:25 |
T42_ | <dgmarie2> i think this is the problem | 01:25 |
T42_ | <dgmarie2> do i just do sudo apt install cpio? or do i have to do something else | 01:25 |
T42_ | <dgmarie2> ok i fixed it | 01:28 |
mal | you probably didn't do "export TEMPORARY_DISABLE_PATH_RESTRICTIONS=true" before running make hybris-hal | 01:44 |
mal | android build doesn't allow cpio anymore by default and that export is needed to make it allow it | 01:45 |
mal | did you read the faq about hybris-17.1 builds https://github.com/mer-hybris/hadk-faq#hybris-17-1 | 01:45 |
mal | @dgmarie2 ^ | 01:46 |
T42_ | <dgmarie2> yes i did read it, i fixed the problem though | 01:46 |
T42_ | <dgmarie2> still rebooting to fastboot though, made some changes to the kernel and im gonna rebuild and try again | 01:47 |
T42_ | <dgmarie2> guessing something is wrong with selinux. I have it enabled and set to permissive because im building android 10, not sure if i made a mistake somewhere though | 01:50 |
mal | just to be sure, what did you set for CONFIG_AUDIT in defconfig? | 01:50 |
T42_ | <dgmarie2> i have it enabled | 01:51 |
mal | good | 01:52 |
mal | reboot to fastboot sounds like a bad issue, missing initramfs can cause that, and that can be missing because "export TEMPORARY_DISABLE_PATH_RESTRICTIONS=true" was not done | 01:53 |
mal | how soon does it reboot? | 01:53 |
T42_ | <dgmarie2> almost instantly | 01:53 |
T42_ | <dgmarie2> the boot img has the initramfs | 01:54 |
T42_ | <dgmarie2> i unpacked it and checked | 01:54 |
mal | hmm, did you check the content? how big was the initramfs? | 01:54 |
T42_ | <dgmarie2> ramdisk folder is 1.5mb | 01:55 |
T42_ | <dgmarie2> is this good for the kernel cmdline? androidboot.selinux=permissive audit=0 selinux=1 | 01:55 |
mal | seems ok | 01:55 |
mal | hmm, on one of my builds the ramdisk.cpio.gz is 4.2 MB | 01:56 |
mal | wait, that was a different build, let me check again | 01:56 |
T42_ | <dgmarie2> should i rebuild the sfos package after changing the kernel source? or should I just rebuild and flash hybris-boot.img | 01:57 |
mal | hybris-boot.img is enough for testing | 01:58 |
mal | just a moment I'll check the size of the ramdisk | 01:58 |
mal | that 1.5 MB seems ok size based on some checking | 02:03 |
mal | one possible reason for failure to boot is that your device could be using verity for verifying boot images | 02:04 |
mal | or using one kernel commandline option to skip ramdisk | 02:05 |
mal | @dgmarie2 you could try something like this https://github.com/mer-hybris/android_kernel_sony_msm/commit/e3648beea7a581e1671a1ac9c59cdb993475c95d note that the change revert a commit so you could try to find a similar commit in your kernel commit history and then try to revert it | 02:09 |
mal | so look for skip_initramfs in git history of your kernel | 02:09 |
T42_ | <dgmarie2> ok thanks for showing me that, time to build again | 02:18 |
T42_ | <dgmarie2> still reboots to fastboot | 02:42 |
T42_ | <dgmarie2> could it be a problem with my defconfig? | 02:44 |
mal | it's possible, if possible push your kernel changes to github for example in your own project so it's easier to see what you have done so far | 02:51 |
mal | I have these in one of my android 10 based devices | 02:54 |
mal | CONFIG_SYSTEM_DATA_VERIFICATION=n | 02:54 |
mal | CONFIG_FORCE_INITRAMFS=y | 02:54 |
T42_ | <dgmarie2> https://github.com/dgmarie/android_kernel_xiaomi_sm6150/blob/lineage-17.1/arch/arm64/configs/davinci_defconfig heres my defconfig | 02:54 |
mal | you seem to have quite many changes to defconfig and some of those don't seem directly related to sailfish changes | 02:59 |
mal | how did you edit the defconfig? | 03:00 |
T42_ | <dgmarie2> make ARCH=arm64 davinci_defconfig menuconfig savedefconfig | 03:00 |
mal | ok, usually we edit the defconfig manually | 03:01 |
mal | often just by adding any needed changes to the end of file | 03:01 |
T42_ | <dgmarie2> ok I can try doing that | 03:02 |
mal | as an example https://github.com/mlehtima/android_kernel_fairphone_sdm632/commit/e410f2565b951a56764cd71f576659308529a481 | 03:02 |
mal | some of those changes are device specific of course | 03:02 |
mal | so I suggest you go back to original davinci_defconfig and then run make hybris-boot, then fix the errors of kernel checker (and the other relevant changes) | 03:03 |
mal | I suggest trying that CONFIG_SYSTEM_DATA_VERIFICATION=n | 03:08 |
T42_ | <dgmarie2> yeah im doing that now, thanks for the advice! | 03:09 |
mal | hopefully that help, I can help again tomorrow | 03:10 |
mal | well actually today since it's quite late already but need some sleep | 03:10 |
*** karry_ is now known as karry | 07:21 | |
T42_ | <elros34> @dgmarie2 not sure if that will help but have you tried flashing vbmeta image? You should be able to find one for your device | 09:59 |
rinigus | poetaster, piggz : -simplecrop turned out to still have some libs in the source as a part of PIL. we would have to get rid of those, ideally by packaging PIL separately | 12:20 |
rinigus | I wonder whether anyone has seen Fedora or Suse packages for PIL? couldn't find them during short search at fedora sources | 12:20 |
rinigus | ... wrong channel, will move to #sailfsishos | 12:21 |
poetaster | rinigus: the whole point of that build is to into harbour, hence the libs. I thkink the solution is a subdir build and the python lib being built insitu. | 12:56 |
poetaster | the pil libs are on openrepos (I maintain them). | 12:57 |
poetaster | https://openrepos.net/content/planetosstore/python3-pillow-simd | 12:57 |
poetaster | But that's not a proper build either. | 12:58 |
rinigus | poetaster: I think it should be possible to have python3-pillow-simd as build dependency and then copy it into your package shared libs, as you do for other libs. then it will get to harbour without any issues | 12:58 |
rinigus | but "proper" PIL lib is needed for it | 12:59 |
poetaster | Hmmm. I was told 'not', by vige, I believe. But maybe I was going about it wrong. There's a thread on the forum. I'll revisit. | 13:01 |
T42 | <dgmarie2> when editing my defconfig manually i get a lot of errors while building | 14:53 |
T42 | <dgmarie2> https://github.com/dgmarie/android_kernel_xiaomi_sm6150/blob/hybris-17.1/arch/arm64/configs/davinci_defconfig | 15:06 |
T42 | <dgmarie2> https://pastebin.com/3Utn0SKr https://pastebin.com/vwdhBnHg the errors | 15:09 |
T42 | <dgmarie2> yes I did try that (re @elros34: @dgmarie2 not sure i...) | 15:10 |
T42 | <elros34> how fast does it reboot to fastboot? | 15:22 |
T42 | <dgmarie2> shows xiaomi logo for about 1 second and then boot into fastboot | 15:26 |
T42 | <elros34> quite fast so probably it didnt even start booting, did you see anything in host dmesg? Anyway, the easy way to check whether /init-debug started is existence of /init.log | 15:31 |
mal | @dgmarie2 could you squash the two commits into one so it's more clear what you have done | 15:33 |
T42 | <elros34> that change in kernel boot args is wrong: https://github.com/dgmarie/android_device_xiaomi_davinci/commit/90c89cf9dedd12e74338e5dc2b34ae80a68e367f. ,115200n8 is a part of console=ttyMSM0, do not remove comma. Same for earlycon | 15:53 |
T42 | <dgmarie2> https://github.com/dgmarie/android_kernel_xiaomi_sm6150 I fixed up my kernel source and I'm building it now | 20:01 |
T42 | <b100dian> @dgmarie2 are you aware there was a davinci port that stopped at 3.4? https://forum.xda-developers.com/t/sailfishos-on-redmi-k20-mi-9t-beta.4062983/ | 21:29 |
T42 | <dgmarie2> yes im aware but i want to upgrade to an android 10 base | 21:30 |
T42 | <b100dian> ok, just checking that you have the background - did the kernel change between the android versions? | 21:31 |
T42 | <dgmarie2> how do I change password in HABUILD_SDK | 22:07 |
T42 | <dgmarie2> or access root user | 22:07 |
T42 | <dgmarie2> actually i figured it out | 22:11 |
mal | what are you trying to do? | 22:12 |
T42 | <dgmarie2> I was making sure the default shell in chroot was bash | 22:14 |
T42 | <dgmarie2> https://pastebin.ubuntu.com/p/kVZNm68Vcq/ trying to figure out what is causing this error when building | 22:15 |
mal | just curious which version of habuild_sdk do you have? | 22:17 |
mal | did you follow the hadk pdf version 4.1.0.0? | 22:18 |
T42 | <dgmarie2> ubuntu-focal-20210531 | 22:21 |
mal | yes, that is the correct one | 22:21 |
T42 | <dgmarie2> looks like the problem is that /bin/sh is a symlink to dash | 22:24 |
T42 | <dgmarie2> I think thats the default but it might be better if it was pointing to bash' | 22:25 |
T42 | <dgmarie2> changed it and am still getting the error. I guess i did something wrong updating the kernel source | 22:27 |
T42 | <elros34> did you clean up kernel build files somewhere from out/ when you update kernel, just to be sure? Did you fix BOARD_KERNEL_CMDLINE? What about my other questions? | 22:46 |
T42 | <dgmarie2> yes i ran make clobber and I fixed the cmdline (re @elros34: did you clean up ker...) | 22:48 |
T42 | <elros34> on github it's still wrong | 22:49 |
T42 | <dgmarie2> this isnt right? https://github.com/dgmarie/android_device_xiaomi_davinci/blob/hybris-17.1/BoardConfig.mk#L127 | 22:51 |
T42 | <elros34> no: https://github.com/ArianK16a/android_device_xiaomi_davinci/compare/lineage-17.1...dgmarie:hybris-17.1 | 22:52 |
T42 | <dgmarie2> oh thanks for letting me know | 22:54 |
T42 | <elros34> I know this is unlikely based on how fast it reboots but have you checked whether /init.log exists? | 22:56 |
T42 | <dgmarie2> it doesnt exist | 22:57 |
T42 | <dgmarie2> I've made a lot of changes since I last built hybris-boot.img though. real problem right now is that I can't build it | 22:58 |
T42 | <elros34> why don't you use lineage sources, I guess they should build fine? | 22:59 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!