| rinigus | @b100dian: morning! we are missing few techpack kernel modules as it is. for jolla devices, they added them as submodules. I think I will try to add them into local manifest instead, as it is done with sony aosp. | 09:03 |
|---|---|---|
| rinigus | we may end up using slightly different android manifests, so it is ok to try. | 09:04 |
| T42 | <b100dian> interesting. This is needed even for AOSP build right? Still haven't updated my build to r67, maybe I'll do that now | 10:32 |
| T42 | <b100dian> rinigus: so after checking out https://github.com/sonyxperiadev/local_manifests I just fetch that kernel-extras.xml between them 🤔? | 10:36 |
| rinigus | @b100dian: I have added local manifests step to https://github.com/sailfishos-sony-nagara/main/wiki/HADK-for-Sony-Nagara-devices | 10:38 |
| rinigus | slowly writing it as I am progressing | 10:38 |
| rinigus | but that local_maniufests would be used with regular repo sync, after you cloned into empty directory droid-hal-sony-nagara | 10:38 |
| rinigus | which has kernel as submodule | 10:39 |
| rinigus | as mal warned about r67, I am using older aosp14 (r34) - default for hybris and others. | 10:40 |
| rinigus | right now its mainly to get boot images. was updating hadk using your kernel howto description and just adding the env vars that we use | 10:41 |
| T42 | <b100dian> rinigus: thanks for documenting the steps, I will probably follow along tonight and edit as neccesary - for now I have only started an unattended r67 build because curiosity | 13:03 |
| T42 | <b100dian> I would also like to clean up the hybris-boot diff a bit more (rndis.usb0 stuff), I will just push a cleanup commit now instead of re-writing history | 13:04 |
| rinigus | @b100dian: these docs are gonna save us in future :) . good idea, let's keep history now | 13:09 |
| rinigus | I am going to setup env at obs - for now in my own repo. later, when it works, we can copy it to regular devel/testing | 13:10 |
| rinigus | good idea to try r67! | 13:10 |
| mal | I should try booting that r67 android | 15:54 |
| rinigus | I have some weird error at OBS, probably just missing some typo in packaging. I am adding nagara config in the same way as it was for tama. config repo for nagara at https://github.com/sailfishos-sony-nagara/droid-config-sony-nagara | 16:40 |
| rinigus | OBS error for package https://build.sailfishos.org/package/show/home:rinigus:nagara/droid-config-xqct54 | 16:41 |
| rinigus | pasted at: https://paste.mozilla.org/vKSgZzcT/raw | 16:41 |
| rinigus | patterns package works: https://build.sailfishos.org/package/show/home:rinigus:nagara/patterns-sailfish-device-configuration-xqct54 | 16:41 |
| rinigus | similar conf for tama works: https://build.sailfishos.org/package/show/nemo:devel:hw:sony:tama:aosp10/droid-config-h8216 | 16:42 |
| rinigus | surely something trivial, just getting blind over here | 16:42 |
| rinigus | hmm, maybe found something while building locally. no newline at the end of one file. checking ... | 16:47 |
| T42 | <b100dian> maybe _just service file missing branch or revision? | 16:47 |
| T42 | <b100dian> that sounds more like it | 16:48 |
| rinigus | nope - was a missing new line at the end of included file. fix: https://github.com/sailfishos-sony-nagara/droid-config-sony-nagara/commit/54c98038d73bde762c44aaa5e26373b4ae89aca3 | 16:50 |
| rinigus | we are building :) | 16:51 |
| rinigus | ... and failing somewhere else | 16:51 |
| T42 | <b100dian> Why am I so tempted to just build the mic image and start debugging booting after you've made obs build pass :D | 16:59 |
| mal | what header*.inc stuff seems odd, just makes things more complicated imo | 17:17 |
| mal | rinigus: also is that missing something? you can see here the patterns being included https://github.com/mer-hybris/droid-config-sony-murray/blob/master/rpm/droid-config-xqcc54.spec but I can't see that being done in your repo | 17:19 |
| mal | or am I missing something | 17:20 |
| mal | wait, you have those patterns as specs here https://github.com/sailfishos-sony-nagara/droid-config-sony-nagara/tree/main/rpm which is very odd | 17:21 |
| mal | it seems the whole thing is for some reason done in totally different way that usually so difficult to debug | 17:22 |
| rinigus | mal: I'll simplify it a bit and then see if I can find the error. It's coming from 6 tama devices. | 17:35 |
| rinigus | mal: streamlined the config and made it close to zambezi. almost there, just issue is at obs since repo location is not common | 18:53 |
| rinigus | would it be possible to get nemo:devel:hw:sony:nagara? | 18:54 |
| rinigus | and set us with b100dian as maintainers? | 18:54 |
| rinigus | also, later issue was due to absence of repo / device definitions in ssu. that had to be added for KS | 18:55 |
| rinigus | @b100dian: mic image will be close. but we need your device hal images, at least that I had for different tama devices | 18:58 |
| rinigus | let's also see what will happen without system images. currently I hope that we can just flash boot and few more to get it working and keeping system / vendor other images the same | 18:59 |
| T42 | <b100dian> rinigus: I was thinking to flash boot and untar .stowaways inside recovery and go from there. | 19:07 |
| T42 | <b100dian> Even with your hals, like an upside down HADK guide | 19:07 |
| T42 | <b100dian> Twisted HADK | 19:07 |
| rinigus | maybe it work. I'll be finishing soon for today, just want to add version packages as well | 19:08 |
| T42 | <b100dian> that is, I would not go to lvm build just yet, I know this is probably the end goal | 19:12 |
| T42 | <b100dian> and something like a github action as used here https://github.com/sailfishos-on-sake/Release/blob/main/.github/workflows/build.yml | 19:14 |
| rinigus | @b100dian: sounds good! | 19:16 |
| rinigus | @b100dian: version packaged and we are now missing boot images. and for that, parameters like these are needed: https://github.com/mer-hybris/droid-hal-img-boot-sony-zambezi/blob/master/rpm/droid-hal-pdx235-img-boot.spec#L5 | 19:23 |
| rinigus | temporarily we can probably comment out droid-system* as a requirement | 19:25 |
| T42 | <b100dian> I don't _think_ we need to build vendor_boot | 19:25 |
| T42 | <b100dian> I can look up the parameters from my unpacking last week, but I am surprised that this is needed in spec files | 19:26 |
| rinigus | @b100dian: hmm, it could be related to aosp packaging. but I don't know for sure. as for system, that is actually required by hal packages. so maybe that's needed | 19:28 |
| rinigus | as for vendor_boot, I guess it is built at obs from the same boot img package | 19:29 |
| T42 | <b100dian> no, it is another image, the vendor_boot (as is vendor_dklm) - but regarding aosp packaging I think you're right | 19:30 |
| T42 | <b100dian> it is the first time I see a repo called hybris-initrd :) | 19:30 |
| T42 | <b100dian> mal: can you shed some light? Those mkbootimg parameters can come just by echoing the mkbootimg command while building, right? say, wrapping it in .. $(warning ...) | 19:31 |
| rinigus | @b100dian: when everything will be resolved, it could be easier to do lvm in the end :) | 19:35 |
| T42 | <b100dian> rinigus: agreed. But curious, do you use lvm when debugging ? | 19:38 |
| rinigus | don't remember anymore :) | 19:38 |
| rinigus | that's where we can drop droid-system requirement: https://github.com/sailfishos-sony-nagara/droid-hal-sony-nagara/blob/main/rpm/droid-hal-common.inc#L23 | 19:47 |
| mal | rinigus: https://build.sailfishos.org/project/show/nemo:devel:hw:sony:nagara | 20:03 |
| mal | rinigus: for obs you can use project config to ignore dependencies, no need to comment them out | 20:03 |
| rinigus | mal: thank you very much! I will look into it tomorrow | 20:06 |
| mal | @b100dian easiest way to get mkbootimg params is to grab the android boot.img from the adaptation build and unpack it with https://android.googlesource.com/platform/system/tools/mkbootimg/+/refs/heads/main (clone the whole repo or just grab the .py files and gki folder) then run unpack_bootimg.py --boot_img boot.img --format=mkbootimg | 20:06 |
| mal | rinigus: something like this in project config should work: Ignore: package_you_are_building:dependency_to_ignore | 20:09 |
| T42 | <b100dian> mal: thanks | 20:23 |
| T42 | <b100dian> mal: but this have_custom_vendor_boot and other have_custom defines are needed because you're also patching those images for e.g. zambezi? https://github.com/mer-hybris/droid-hal-sony-zambezi/blob/master/rpm/droid-hal-common.inc | 21:04 |
| mal | custom_vendor_boot is for making images smaller i.e. only having needed things in vendor_boot.img | 21:17 |
| mal | and there is plan to include recovery in vendor_boot at some point | 21:18 |
| T42 | <b100dian> ah, that is a great plan indeed :) | 21:29 |
| T42 | <b100dian> but my question is - do we need it today for nagara (1/5 IV) or we can just set those have_custom to false/0? | 21:29 |
| T42 | <b100dian> wait, isn't recovery just one byte (or even bit) different from boot init? | 21:30 |
| T42 | <b100dian> I seem to remember that there's a kernel cmdline parameter when you enter recovery on these newer bootloaders | 21:30 |
| T42 | <b100dian> (I'm not seeing this now on 5IV though) | 21:33 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!