| *** phlixi is now known as Guest356 | 03:08 | |
| *** phlixi_ is now known as phlixi | 03:09 | |
| rinigus | @b100dian: reading up, ramdisk should not have caused it. will try tonight starting with released boot.img by gki (after unpacking and repacking for sony using the manual steps as in our wiki) and then, when it works, proceed with compiled in-house kernel | 06:56 |
|---|---|---|
| rinigus | @b100dian: some progress. as gki, we can get several releases from https://source.android.com/docs/core/architecture/kernel/gki-android12-5_10-release-builds | 18:35 |
| rinigus | turns out that android12-5.10-2024-03 works, but android12-5.10-2024-05 doesn't | 18:36 |
| rinigus | android12-5.10-2024-03 is 5.10.205; on stock I have 5.10.198. so, these are OK | 18:37 |
| rinigus | yesterday I tested 2024-11 which failed too | 18:37 |
| rinigus | which makes it somewhat less attractive, as we can't use the latest gki. but I will test now building 2024-03 | 18:39 |
| rinigus | from grepping strings in kernel image of 2024-05, it looks like 5.10.209 | 18:48 |
| rinigus | @b100dian: ... and my regular built gki image boots :) . OK, hybris config next | 18:52 |
| T42 | <b100dian> rinigus: great progress! I myself am a bit time restrained these days. | 18:59 |
| T42 | <b100dian> The Linux version is in the first 2-3 lines of the root Makefile | 18:59 |
| T42 | <b100dian> 2024-03 sounds familiar, is that the latest patch level we have? | 19:01 |
| T42 | <b100dian> I have also noticed that stock had os release 12 and aosp os-releas 13 | 19:02 |
| rinigus | @b100dian: based on avb info and system settings it should be 2024-07. but then, this kernel branch was patched later as well | 19:04 |
| rinigus | @b100dian and mal: GKI failed to boot with hybris kernel config. I added the changes by @b100dian as in here - https://github.com/sailfishos-sony-nagara/android_kernel_sony_msm/commit/28b1502890ed87e1d110616c5b5305f6fa2a995f#diff-28621b58db47b758dfa412b5d9f600e3f4516f37222054fcf19fbf46e7d3abbeR474 | 19:48 |
| rinigus | which was expected, I think. so, we would have to construct the kernel and get vendor bits as well | 19:50 |
| T42 | <b100dian> Bootloop or just gui failure? | 19:52 |
| rinigus | bootloop | 19:52 |
| T42 | <b100dian> Unfortunately you have to bisect on those, I don't know if there's any docs about who's changing abi. My money is on CONFIG_SYSVIPC because something I've read before but didnt find again:) | 19:53 |
| T42 | <b100dian> @elros34 you mentioned disabling vermagic in your kernels a while back I think, was it CONFIG_LOCALVERSION related? | 19:54 |
| rinigus | @b100dian: but don't we need CONFIG_SYSVIPC? if we need it and cannot have it then there is no point :) | 19:56 |
| T42 | <b100dian> Dunno. I think I was reading some git log files from halium on xperia 10 iv that mentioned that this changes ABI and was backed out.. Succesfully? Really dunno | 19:59 |
| T42 | <elros34> I don't remember, is it even possible? I only have some chages so kernel version stays same so I don't have to worry about kernel modules location | 19:59 |
| T42 | <b100dian> Ok, then before modules, we need to know some error of bootloop. I wpukd be curious how voidanix pulled off https://github.com/sonyxperiadev/kernel/pull/2579/files :D | 20:01 |
| rinigus | @b100dian: I will think about it. will look into lineage option as well | 20:23 |
| rinigus | hopefully we will get also new aosp blobs soonish | 20:25 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!