| T42 | <vsvoid> Okay I think am a little confused, so my port is stuck in telnet port 2323 and the init.log is not providing enough information as to why it's stuch here, how can i get more logs? | 02:06 |
|---|---|---|
| T42 | <birdzhang> use Persistent journalctl https://github.com/mer-hybris/hadk-faq#persistent-journalctl | 02:09 |
| T42 | <vsvoid> Thanks, | 02:14 |
| T42 | <vsvoid> Any idea how I can run a script to mount img files on early init just after switching to realfs? | 02:14 |
| Mister_Magister | https://github.com/VerdandiTeam/droid-config-titan/blob/master/sparse/usr/lib/systemd/system/pds.service | 02:19 |
| Mister_Magister | so basically on my backported seccomp-bpf which is same as kernel 3.5 implementation, there's issue with sshd because sshd's seccomp-bpf's program loads the syscall number and then compares it with some syscalls to block/allow some. problem is that loading the syscall number fails (without failing) so it compares everything to 0 hence it goes to the bottom of the program where it returns fail (fail by default) which in turn | 03:59 |
| Mister_Magister | kills sshd | 03:59 |
| Mister_Magister | yes it's 5am and i'm insanely tired, but i got that far. when i wake up ill have to debug further to find why the syscall number isn't loading properly | 04:00 |
| Mister_Magister | i think the super early seccomp-bpf implementation is messing something up, not my patching | 04:02 |
| Mister_Magister | why seccomp used bpf for filtering is freaking beyond me | 04:20 |
| T42 | <TheVancedGamer> ah look, we're both (I think) doing the same thing :D | 05:41 |
| T42 | <TheVancedGamer> i'm backporting (e)BPF to kernel 3.4 | 05:41 |
| T42 | <elros34> @Mister_Magister make sure you really use latest seecomp-bpf patches. I was using one outdated patch without all changes and this caused some issues, see hadk-faq | 08:02 |
| T42 | <elros34> sry hadk-hot: seecomp section | 08:03 |
| Mister_Magister | @elros34 lol didn'trealise it's already been done | 09:04 |
| T42 | <elros34> it's just note with some patches I couldn't confirm any works but for sure one cause kernel crash | 09:10 |
| Mister_Magister | @elros34 i cannot find those patches :/ | 10:59 |
| HengYeDev[m] | how do you setup the development environment for a halium-based build? | 14:20 |
| HengYeDev[m] | is it mostly similar to hadk? | 14:25 |
| HengYeDev[m] | for example, where should droid config halium go in the directory structure | 15:04 |
| HengYeDev[m] | and how do i build without using droid-hal-device? | 15:54 |
| mal | probably piggz can help when he is back here | 16:02 |
| piggz | HengYeDev[m]: im here now, let me look back through your questions | 19:45 |
| piggz | HengYeDev[m]: so, the first thing you want to do is get your kernel/boot building, and get that on OBS ... set up your kernel pacakge like https://github.com/HelloVolla/kernel-adaptation-halium-mimameid | 19:55 |
| piggz | and modify kernel-adaption-simplified to suit ... | 19:55 |
| piggz | I also have a version for qcom devices in kernel-adaptation-halium-qx1050 | 19:55 |
| piggz | perhaps a good job to collaborate on would be making that fully generic so i can be included as a submodule | 19:56 |
| HengYeDev[m] | i'm trying to follow the ubports unihertz titan port (prebuild stock kernel + modules and userspace hacks) and build halium | 19:57 |
| HengYeDev[m] | so i would assume i make the package for prebuilt and make it Provide that | 19:58 |
| HengYeDev[m] | but i think we can build the userdata first and use the ubports boot.img? | 19:59 |
| HengYeDev[m] | piggz: also how are your dirs setup? | 19:59 |
| piggz | i do everything on OBS now, and build the image on gitlab | 20:01 |
| piggz | you could potentially boot one of my images on your device ... the only changes would be in the config package | 20:02 |
| piggz | notkit did that on a galaxy fold | 20:02 |
| piggz | all you need to do is start on your config package, then build the image using the kickstarts from either mimameid or qx1050 on gitlab | 20:03 |
| piggz | i guess its better to build locally first, gitlab has a tendancy to run out of build minutes! | 20:03 |
| HengYeDev[m] | ok, i will try | 20:03 |
| HengYeDev[m] | i just have to replace the halium-11 android-rootfs.img with 10, i don't see any other significant changes. correct me if im wrong | 20:03 |
| piggz | should be good | 20:04 |
| HengYeDev[m] | so i have the config pkg forked from yours, should i put it on obs home-project? | 20:04 |
| piggz | if you need a halium:10 project we can make one | 20:04 |
| piggz | you can create a subproject on your home for building | 20:04 |
| HengYeDev[m] | would subproject work with the gitlab ci tho? | 20:05 |
| piggz | if you build locally using a m odified kickstart, you can point the repos wherever | 20:05 |
| piggz | if you have a proper devel: project on OBS, then sure, use that | 20:05 |
| HengYeDev[m] | where is docs on kickstart files? | 20:06 |
| piggz | not sure there are any! most people following hadk dont really change them ... the changes I made are quite monir though. | 20:07 |
| piggz | kicksstart file is pretty easy to understand and follow | 20:07 |
| HengYeDev[m] | also, i think for halium-10, to make it on obs, copy everything from :11 and replace the halium-android-rootfs | 20:07 |
| piggz | https://gitlab.com/sailfishos-porters-ci/halium-qx1050-ci/-/blob/main/Jolla-%40RELEASE%40-halium-qx1050-%40ARCH%40.ks | 20:07 |
| HengYeDev[m] | so i assume i should just change the repo lines | 20:09 |
| piggz | yes | 20:20 |
| piggz | HengYeDev[m]: towards the end of the file, starting with the "group setup" is where all the halium changes are | 20:21 |
| piggz | we have to add groups/users here, and then do some stuff with the folders | 20:21 |
| HengYeDev[m] | thanks, i am getting started now | 20:27 |
| HengYeDev[m] | piggz: must community-adaptation-devel be used because of halium? i am looking on the qx1050 obs | 20:29 |
| piggz | not nescessarily | 20:30 |
| piggz | the reason I use a custom build of that (with rinigus) is to have each version in testing: as a subproject, instead of a repository, we think it makes versions easier to handle | 20:30 |
| piggz | see eg https://build.merproject.org/project/subprojects/nemo:testing:hw:fxtec:halium-qx1050 | 20:31 |
| piggz | without that, each version is a repository in the main repo ... if you add/remove packages from a version, its becomes difficult to manage. if you have each version as a seperate project, you can make such changes easily | 20:32 |
| HengYeDev[m] | I see, ok | 20:35 |
| HengYeDev[m] | should I branch that to my repo? | 20:36 |
| HengYeDev[m] | piggz: whats dtc for? | 21:06 |
| Mister_Magister | @elros34 https://github.com/torvalds/linux/commit/5ea62181f0366d29b373b714c203bfcb9127048d#diff-f7ac9251cb93f236335e6801eb8a2c2d8744e61c06770731b562e069eaaee940R912 exactly this was missing | 22:35 |
| Mister_Magister | so you were right | 22:35 |
| Mister_Magister | it was in later commit | 22:35 |
| Mister_Magister | tho it supposedly sets samestuff lower but eh | 22:36 |
| Mister_Magister | megic | 22:36 |
| Mister_Magister | lets see if this works | 22:39 |
| Mister_Magister | ayyy it works | 22:43 |
| *** amccarthy is now known as Guest5774 | 22:51 | |
| *** amccarthy_ is now known as amccarthy | 22:51 | |
| *** amccarthy is now known as Guest5775 | 22:57 | |
| *** amccarthy_ is now known as amccarthy | 22:57 | |
| *** amccarthy is now known as Guest5777 | 23:10 | |
| *** amccarthy_ is now known as amccarthy | 23:10 | |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!