| T42 | <Srzsawdmfq> (Photo, 527x609) https://irc.thaodan.de/.imgstore/R7LowlmZVb.png 电报/纸飞机国内48小时在线用户推广,别的业务请看图,有需要私聊|private: @Huanqiu22 | 02:45 |
|---|---|---|
| T42 | Sid127 was added by: Sid127 | 03:59 |
| T42 | <Sid127> I'm extremely new to build environments and such stuff, but I'm interested in porting SFOS to my device (Nokia X6 - DRG_Sprout), mainly because it's the only OS that my computer can build at the moment, and because I'm a big fan of open source software and want to move out of Google's ecosystem. | 04:50 |
| T42 | <Sid127> Any help from you guys would be greatly appreciated | 04:50 |
| T42 | <Sid127> The thing is, my device doesn't have an official LineageOS build, and hence I was wondering how to go about that issue | 04:51 |
| T42 | 熊琳彬 %lastname% was added by: 熊琳彬 %lastname% | 04:54 |
| T42 | Gerald %lastname% was added by: Gerald %lastname% | 05:13 |
| T42 | alexisrev was added by: alexisrev | 05:37 |
| vknecht | @Sid127: official LOS is not a requirement, unofficial LOS <= 16 is ok | 09:04 |
| T42 | <Sid127> okay, thank you :) I'm also having trouble identifying the $VENDOR and $DEVICE strings. | 09:05 |
| T42 | <Sid127> And I suppose that following the HADK PDF should help with porting SFOS, and I can ask here if I get stuck, right? | 09:06 |
| vknecht | sure, also check the FAQ https://github.com/mer-hybris/hadk-faq | 09:07 |
| T42 | <Sid127> I will, thanks a lot! | 09:07 |
| vknecht | for $VENDOR & $DEVICE, search for PRODUCT_BRAND (or maybe PRODUCT_MANUFACTURER) and PRODUCT_DEVICE (or maybe PRODUCT_MODEL) in device/ subtree | 09:10 |
| T42 | <Sid127> I could also find these in my device's build.prop, right? | 09:11 |
| vknecht | could be, but iiuc the authoritative source/setting is in the LOS makefiles | 09:12 |
| T42 | <Sid127> could I also find these from the device's build fingerprint? | 09:18 |
| vknecht | same answer :-) | 09:19 |
| T42 | <Sid127> okay... alternatively, would it be possible to create a SFOS GSI for Android devices that support Project Treble? | 09:22 |
| vknecht | I don't know much about those, maybe search for irc logs about those since it's probably been discussed already | 09:27 |
| Danct12[m] | ok logged to freenode :P | 09:35 |
| Danct12[m] | so i just did sdk-assistant to create a device target | 09:36 |
| Danct12[m] | and there's a bunch of grep unknown option, then at the end i got this error | 09:36 |
| Danct12[m] | mount: /srv/mer/toolings/SailfishOS-3.2.1.20/etc/resolv.conf: special device /run/connman/resolv.conf does not exist | 09:36 |
| Danct12[m] | Toolchain 'patterns-sailfish-sb2-armv7hl' doesn't exist or is already installed - can not install. | 09:36 |
| vknecht | Danct12[m], it's unclear which steps you just followed... is it chapter 4.2 of hadk pdf ? | 09:53 |
| Danct12[m] | yep, but nevermind got that fixed | 09:57 |
| Danct12[m] | it was this file screaming errors /run/connman/resolv.conf | 09:57 |
| Danct12[m] | so i manually assigned a nameserver to it and got everything worked fine | 09:57 |
| vknecht | beware that "“Quick start” section is enough, do not install SDK Targets yet" warning | 09:57 |
| vknecht | target install comes at chapter 6 | 09:58 |
| Danct12[m] | is there anyway to make build_packages runs on all cores? | 10:48 |
| deathmist | don't think so, you don't really have to worry about it once you get a stable port & start building on OBS | 11:09 |
| deathmist | @Danct12 also are you building everything each time? you can do build_packages -d for just HAL, -c for droid-configs and -i for mic image etc | 11:11 |
| Danct12[m] | deathmist: yeah i guess | 11:14 |
| Danct12[m] | anyway i have a error | 11:14 |
| Danct12[m] | while building | 11:16 |
| * Danct12[m] sent a long message: < https://matrix.org/_matrix/media/r0/download/matrix.org/YmWMwEgmnccOIrisIqBtHzKB > | 11:16 | |
| deathmist | that indicates you have an outdated dhd submobule, it should be @ https://github.com/sailfishos-oneplus5/droid-hal-device/tree/upgrade-3.2.1 and it has a fix for that specific issue https://git.io/JvkYr | 11:17 |
| deathmist | wrong link sorry, I mean https://github.com/mer-hybris/droid-hal-device/commits/upgrade-3.2.1 | 11:17 |
| Danct12_ | @deathmist, didn't that patch got merged to master? | 11:23 |
| Danct12[m] | didn't that patch got merged to master? | 11:23 |
| Danct12[m] | i just checked | 11:23 |
| T42 | <Sid127> since I'm extremely new to OS development and porting, I'm gonna blindly follow the HADK | 11:25 |
| Danct12_ | sidenote, i'm building on oreo | 11:25 |
| T42 | <Sid127> and ask here if I get stuck | 11:25 |
| Danct12_ | s/oreo/pie | 11:25 |
| T42 | <Sid127> I hope that's okay with you guys | 11:25 |
| deathmist | you should not be using master branch of dhd, it's for >3.2.1 building and won't work for the current release properly afaik | 11:25 |
| Danct12_ | oh | 11:26 |
| Danct12_ | will be switching to that, hold on | 11:26 |
| Danct12_ | deathmist, just tried and still the same error | 11:39 |
| Danct12_ | libncicore did get build and also successful too | 11:40 |
| deathmist | on nice there is another package you need to clone and build then, you can either just downgrade nfcd-binder-plugin to 1.0.5 like I've done or rpm/dhd/helpers/build_packages.sh --mw=https://github.com/mer-hybris/libnciplugin | 11:43 |
| deathmist | after that the full build script should go through | 11:43 |
| deathmist | mal: any clue if that above package should be in common repo later or be built manually by build_packages script? | 11:48 |
| Danct12_ | you'll also need to run this first before installing things else | 11:52 |
| Danct12_ | PlatformSDK [danct12@melttower hadk]$ rpm/dhd/helpers/build_packages.sh --mw=https://git.sailfishos.org/mer-core/nfcd.git | 11:52 |
| Danct12_ | and finally i got it compiled! | 11:53 |
| Danct12_ | full rebuild time because i want to waste my time while eating ramen | 11:53 |
| T42 | <Sid127> I'm having trouble. when I try to run the command "repo init -u git://github.com/mer-hybris/android.git -b hybris-16.0" I get an error saying "bash: repo: command not found" even though I followed the steps on the google site to install repo | 12:00 |
| Danct12_ | Sid127, check your environment variable for $PATH | 12:00 |
| deathmist | @Sid127 run "export PATH=$HOME/bin:$PATH" and add it at the end of your ~/.bashrc | 12:03 |
| T42 | <Sid127> I ran that on HOME $, you think I should try running it again on HABUILD_SDK $? | 12:04 |
| deathmist | ??? what do you mean, just run it where the issue happens and do what I said, it's fixed. | 12:05 |
| T42 | <Sid127> yup, it worked. thanks! | 12:05 |
| mal | deathmist: it's missing from upgrade-3.2.1 branch, it's built in master afaik | 12:06 |
| mal | deathmist: ah, missed that there are two packages with similar name | 12:08 |
| deathmist | mal: is https://git.sailfishos.org/mer-core/nfcd also needed? apparently Danct12_ needed it too to finish building (I've just stuck with nfcd-binder 1.0.5 for now as to not deal with this) | 12:09 |
| mal | deathmist: you mean too old nfcd? | 12:11 |
| mal | deathmist: these things always happen sometimes when dependencies change | 12:13 |
| Danct12_ | now when i build droidmedia i get this error | 12:14 |
| Danct12_ | CameraService.cpp:2050:48: error: unused parameter 'kind' [-Werror,-Wunused-parameter] | 12:14 |
| mal | which android base? | 12:14 |
| Danct12_ | 16.0 | 12:15 |
| deathmist | the file is broken lol, I just did 'sed "s/Werror/Werror -Wno-unused-parameter/" -i frameworks/av/services/camera/libcameraservice/Android.mk' when I was still basing on Pie | 12:15 |
| deathmist | well I mean there is an unused parameter in CameraService.cpp and the build is set to error on those warnings by default | 12:16 |
| Danct12_ | does that means lineage-16.0 builds are also broken too? | 12:18 |
| Danct12_ | i mean, on the android side, not us doing ports | 12:19 |
| deathmist | probably not, I mean they would have fixed it by now if it was broken on their side too | 12:19 |
| mal | maybe some build parameter in droidmedia Android.mk is needed now | 12:20 |
| mal | or something | 12:20 |
| T42 | <Sid127> the command "repo init -u git://github.com/mer-hybris/androig.git -b hybris-16.0" is supposed to take a while to execute, right? | 12:26 |
| abranson | ages, you're cloning the full android source | 12:26 |
| mal | init should be quick, repo sync takes a while | 12:27 |
| T42 | <Sid127> (please don't mind all my questions, I'm just a 16 year old who's just getting started with OS building and porting and SFOS is my first project) | 12:27 |
| abranson | ah no that's sync. is it still doing anything? | 12:27 |
| T42 | <Sid127> repo init just has the blinking cursor pointer going on | 12:27 |
| T42 | <Sid127> it's been a good 20 minutes now | 12:27 |
| mal | ctrl+c and then try again | 12:28 |
| T42 | <Sid127> >fatal: cannot make /home/onebigsucc/hadk/.repo/repo directory: File exists | 12:29 |
| T42 | <Sid127> >fatal: cloning the git-repo repository failed, will remove '.repo/repo' | 12:29 |
| T42 | <Sid127> I think I should delete that folder manually and try it | 12:29 |
| Danct12_ | rm -rfv .repo | 12:29 |
| Danct12_ | and then repo init again | 12:29 |
| T42 | <Sid127> yup, it's working now | 12:29 |
| T42 | <Sid127> this is what I got https://del.dog/agniryrfav | 12:30 |
| T42 | <Sid127> I can continue, right? or has something gone wrong? | 12:31 |
| Danct12_ | add --no-repo-verify | 12:31 |
| Danct12_ | and it'll continue | 12:32 |
| T42 | <Sid127> okay | 12:34 |
| mal | deathmist: I could build those nfc things in common also, need to think a bit | 12:38 |
| T42 | <adampigg> mal: were those links any good for you | 12:39 |
| mal | @adampigg quite similar to the link I showed you, the other ble one was new, need to check that | 12:41 |
| mal | although that was on top of java which is not as nice | 12:42 |
| mal | could still be useful for seeing how to send those protocol buffer things to bluez | 12:43 |
| mal | @adampigg I could first grab the raw data from the hcidump I got from a sync and see how to parse that | 12:45 |
| T42 | <adampigg> mal: yeah, if you can get raw data over BT, you can look to parse it | 12:46 |
| T42 | <adampigg> then, add support to amazfish ;) | 12:46 |
| mal | @adampigg in the hcidump log I have from the sync there are about 6000 packets send/received via bt, so I need to make some script to assemble those into something reasonable | 12:50 |
| Danct12_ | ok so now i got this error while building version package | 12:53 |
| Danct12_ | https://paste.ubuntu.com/p/XVQFqSW99d/ | 12:53 |
| mal | you need to check the build log of that package | 12:54 |
| T42 | richie_flicker was added by: richie_flicker | 12:54 |
| Danct12_ | nothing too useful in there, except it provides two solution | 12:55 |
| Danct12_ | https://paste.ubuntu.com/p/6yNtjmWPt9/ | 12:55 |
| Danct12_ | oh | 12:56 |
| Danct12_ | Problem: droid-config-santoni-1-1.armv7hl obsoletes ofono-configs-mer provided by ofono-configs-mer-1.22+git9-1.26.1.jolla.armv7hl | 12:56 |
| deathmist | Danct12_: run the sb2 command from https://git.io/Jvk3s that my build_packages script does automatically for local builds, it also makes sure you get bluez5 packages instead of bluez4 (which my device uses, omit those packages from the list if yours doesn't) | 12:59 |
| deathmist | "sb2 -t $VENDOR-$DEVICE-$PORT_ARCH -m sdk-install -R zypper -n in droid-config-$DEVICE" alone should have worked too afaik | 13:06 |
| Danct12_ | deathmist: thanks :0 | 13:12 |
| Danct12_ | :) | 13:13 |
| deathmist | Danct12_: assuming your device has a similar amount of work to do as my port was to rebase on 16.0, you only have a few steps remaining to get GUI back etc | 13:30 |
| deathmist | you did already update init rc files for 16.0 in droid-configs locally as per FAQ 16.0 section, right? | 13:31 |
| T42 | <Sid127> last question for the day - in the local_manifests file (as given in section) would I have to change the paths if I'm using unofficial Lineage as the base? | 14:03 |
| deathmist | @Sid127 you'll find what & where everything should be cloned to from the lineage.dependencies / cm.dependencies files in root of the unofficial android_device_* repository, do the same for other projects mentioned there | 14:14 |
| T42 | <Sid127> I'm afraid I didn't understand | 14:15 |
| deathmist | here's the file for my device https://git.io/JvkGs it depends on android_device_oneplus_msm8998-common which has the following dependencies https://git.io/JvkGn and so on, gather all of them in <project> entries and show the manifest when you're done | 14:18 |
| T42 | <Sid127> strange, the repo for my device doesn't seem to have a branch for Lineage-16.0 | 14:20 |
| deathmist | ok, well what branch are you trying to build for? | 14:20 |
| deathmist | it was JUST an example | 14:21 |
| T42 | <Sid127> I'm trying to build for lineage 16, and I know it was just an example. My point is that my device (Nokia X6) doesn't have any official lineage release, and so I really don't think I'll find the required files on the official lineage repo | 14:22 |
| deathmist | found it for you https://github.com/hritikutekar/android_device_nokia_Dragon/blob/lineage-16.0/lineage.dependencies ofc you can't if it's not an official device | 14:24 |
| T42 | <Sid127> thank you very much! | 14:25 |
| deathmist | you can look at perhaps https://git.io/JvkGM for an example of how to add additional remotes which you'll need to add hritikutekar & RaghuVarma331 as and use for the projects you're cloning, you also need to clone that repo I linked just above and ignore the android packages | 14:30 |
| T42 | <Sid127> quite honestly, I'm stumped | 14:37 |
| T42 | <Sid127> does this look good? | 14:39 |
| T42 | <Sid127> https://del.dog/qinuppinus.txt | 14:39 |
| T42 | <Sid127> sorry if I'm causing any inconvenience | 14:45 |
| deathmist | nope. remotes do NOT define a project (github repository) to clone, just where to look for them (i.e. a GitHub user account), why did you put "stable/lineage-16.0" as the branch? it's not needed when you define it in the remote and it clearly said "pie" on the dependencies file which you should be looking at, you didn't capitalize Dragon in vendor path and you incorrectly put device repo name on all projects | 14:48 |
| deathmist | those aren't even your biggest issues, seems RaghuVarma removed pie branches from the kernel and vendor repositories entirely so you cannot use those sources and need to find different ones from github or elsewhere that are compatible with your device (nokia_sdm660 / Dragon) | 14:51 |
| T42 | <Sid127> I found other another source with the pie branch | 14:51 |
| T42 | <Sid127> https://github.com/hritikutekar/android_kernel_nokia_sdm660/tree/pie | 14:51 |
| T42 | <Sid127> how about this? | 14:53 |
| T42 | <Sid127> https://del.dog/qinuppinus.txt | 14:53 |
| deathmist | why are you defining "android_device_nokia_Dragon" to be cloned to "vendor/nokia/dragon"? you should read the file and try to understand what it does. you need proper <project> entries for BOTH vendor and device seperately (3 projects total), you also need to remove "hritikutekar/" from name part on all projects, that assumed your remote is "github.com/" and just got appended to the URL | 14:57 |
| T42 | <Sid127> ok, ok, I think I get it now. Allow me to find a vendor repo and get back | 15:05 |
| T42 | موسوی %lastname% was added by: موسوی %lastname% | 15:10 |
| T42 | Taufik %lastname% was added by: Taufik %lastname% | 15:12 |
| T42 | <Sid127> how about this? | 15:15 |
| T42 | <Sid127> https://del.dog/qinuppinus.txt | 15:15 |
| T42 | <Sid127> @deathmist | 15:16 |
| T42 | <Sid127> really sorry for all the trouble | 15:17 |
| Mister_Magister | can i install mic outside of sfossdk? | 15:21 |
| deathmist | @Sid127 looks fine, not sure if sdm660-common vendor blobs will be enough for your phone (mine also has a dir specific to my phone on top of the common one) but you can try, and you may want to remove the trailing slash at the end of dragon_sfos fetch attribute | 15:33 |
| T42 | <Sid127> those vendor blobs should work, because RaghuVarma used the same blobs to boot LineageOS17 on multiple nokia devices | 15:34 |
| T42 | <Sid127> also, thanks, and sorry for all the inconvenience | 15:35 |
| deathmist | oh and also change vendor path from "vendor/nokia/Dragon" to "vendor/nokia" as to match that repo name | 15:35 |
| T42 | <Sid127> okay, will do | 15:35 |
| T42 | <Sid127> I can now move ahead with the repo sync, right? | 15:37 |
| deathmist | sure | 15:39 |
| T42 | <Sid127> okay then. I'll stop for today after the repo sync and continue tomorrow | 15:45 |
| Danct12_ | got sailfishos booted.. to ralsei | 15:45 |
| Danct12_ | basically stuck at the splash screen | 15:46 |
| Danct12_ | fastboot screen* | 15:46 |
| Danct12_ | seems like the android services failed to start | 15:47 |
| T42 | <Lukapanio> дщпі, | 15:52 |
| T42 | <Lukapanio> any logs* | 15:52 |
| T42 | <Lukapanio> Danct12_ | 15:52 |
| Danct12_ | dmesg, yes | 15:53 |
| Danct12_ | i noticed there's a ton of [ 445.702763] binder: 2673:2673 transaction failed 29189/-22, size 32-0 line 3015 in dmesg | 15:55 |
| Danct12_ | and [ 445.367021] binder: 4063:4063 transaction failed 29189/-22, size 32-0 line 3015 | 15:55 |
| Danct12_ | https://paste.ubuntu.com/p/TBQ8wMYYPW/ | 15:59 |
| mal | Danct12_: did you check the android base specific things in faq? | 16:01 |
| Danct12_ | i have | 16:03 |
| Danct12_ | https://github.com/mer-hybris/hadk-faq/blob/master/README.rst#id21 | 16:03 |
| mal | Danct12_: then why doesn't the dmesg you gave show droid-hal-init loading anything from /usr/libexec/droid-hybris/system/etc/init ? | 16:06 |
| T42 | <Lukapanio> рьь дщі15 ифіу, | 16:07 |
| T42 | <Lukapanio> https://github.com/mer-hybris/hadk-faq/blob/master/README.rst#hybris-15-1 | 16:08 |
| T42 | <Lukapanio> Danct12_ | 16:08 |
| T42 | <Lukapanio> los 15? | 16:08 |
| Danct12_ | lukapanio: hybris 16 | 16:08 |
| T42 | <Lukapanio> santoni? | 16:08 |
| Danct12_ | yeah i wanted to rebase to pie | 16:08 |
| T42 | <Lukapanio> https://github.com/mer-hybris/hadk-faq/blob/master/README.rst#hybris-16-0 | 16:09 |
| T42 | <Lukapanio> Danct12_ | 16:09 |
| Danct12_ | i already have applied that | 16:09 |
| Danct12_ | because you can't run make hybris-hal without them | 16:09 |
| Danct12_ | regarding init files, mal: i think they just don't want to load, i checked the directory and everything was there | 16:10 |
| T42 | <Lukapanio> Copy files from https://github.com/sailfishos-oneplus5/droid-config-cheeseburger/tree/hybris-16.0/sparse/usr/libexec/droid-hybris/system/etc/init to your config repo (to hybris/droid-configs/sparse/usr/libexec/droid-hybris/system/etc/init) and rebuild config packages using rpm/dhd/helpers/build_packages.sh -c? | 16:10 |
| Danct12_ | https://paste.ubuntu.com/p/RKKzr9jFr8/ | 16:10 |
| T42 | <Lukapanio> but you did this step? | 16:10 |
| Danct12_ | lukapanio: have also applied that, you can see the paste above for the output | 16:11 |
| mal | Danct12_: and you are sure you ran the patching correctly? | 16:11 |
| Danct12_ | i'm pretty sure, just ran the patch again and now i got "patch does not apply" | 16:12 |
| Danct12_ | wait, does magisk also mess it up? as i haven't flashed the fresh los 15 before sailfish | 16:13 |
| Danct12_ | just reflashed fresh los 16 and sailfish, nothing changed | 16:25 |
| Danct12_ | however the binder kernel spam, i believe that's the cause | 16:26 |
| mal | which device is that? | 16:26 |
| Danct12_ | xiaomi redmi 4x | 16:26 |
| mal | is the lineage 64-bit for that? | 16:27 |
| Danct12_ | yep | 16:27 |
| mal | and you see /usr/libexec/droid-hybris/system/lib64/libselinux_stubs.so on device? | 16:28 |
| Danct12_ | yep | 16:29 |
| mal | and all *servicemanager.rc files in /usr/libexec/droid-hybris/system/etc/init have the preload for that | 16:29 |
| mal | there should be 3 of those | 16:29 |
| Danct12_ | yeah, all of them are LD_PRELOAD /usr/libexec/droid-hybris/system/lib64/libselinux_stubs.so | 16:31 |
| mal | a lot of android side services fail | 16:31 |
| mal | do you have ld.config.28.txt fix? | 16:32 |
| mal | or what ever ld.config*.txt you have in /system/etc/ | 16:32 |
| Danct12_ | i have /system/etc/ld.config.vndk_lite.txt | 16:33 |
| Danct12_ | oh i dont have that fix yet | 16:33 |
| Danct12_ | i'll apply that | 16:34 |
| deathmist | ok afaik that file hasn't needed any patching before since it hasn't specified "namespace.default.isolated = true" | 16:34 |
| deathmist | check if your file has that line Danct12_ | 16:34 |
| deathmist | there should also be plenty of lines with "permitted.paths" | 16:34 |
| Danct12_ | namespace.default.isolated = false | 16:34 |
| deathmist | and no permitted.paths lines I presume? also try to get a full "journalctl -b --no-pager" output after boot | 16:37 |
| Danct12_ | deathmist: here are all permitted.paths lines | 16:40 |
| Danct12_ | https://paste.ubuntu.com/p/xnXqFh6tT3/ | 16:40 |
| Danct12_ | and here is journalctl https://paste.ubuntu.com/p/4N7n48dttr/ | 16:41 |
| deathmist | seems servicemanager* services don't die at least so I think you're ok in terms of ld.config (no patch needed) | 16:45 |
| Danct12_ | maybe it's because of firmwares? | 16:46 |
| Danct12_ | since i noticed firmware load errors | 16:46 |
| T42 | <elros34> disable selinux according to HADK 9.3.1 | 16:46 |
| Danct12_ | elros34: selinux isn't compiled | 16:47 |
| Danct12_ | as CONFIG_AUDIT is also disabled | 16:47 |
| T42 | <elros34> yes its wrong, you need to have selinux disabled via bootparam | 16:48 |
| deathmist | perhaps give output of "mount" on device and "ls -l $ANDROID_ROOT/out/target/product/$DEVICE/root/" on platform SDK, you may be missing some straggler files | 16:50 |
| T42 | <elros34> mer-kernel-checker is misleading and people usually disable audit which | 16:50 |
| T42 | <elros34> is incorrect | 16:50 |
| mal | Danct12_: ah, so you accidentially disabled audit, that is bad | 16:51 |
| Danct12_ | blame mer kernel check >:( | 16:51 |
| mal | @elros34 PRs welcome for better descriptions to kernel checker | 16:52 |
| Danct12_ | but the question is, does that also cause android services to fail? | 16:53 |
| mal | Danct12_: could cause, android has quite strong dependency to selinux | 16:53 |
| mal | and selinux needs audit | 16:53 |
| Danct12_ | i see | 16:53 |
| mal | even if selinux is disabled | 16:53 |
| mal | I mean disabled at runtime | 16:53 |
| Danct12_ | deathmist: https://paste.ubuntu.com/p/WgQDkfwqN8/ | 16:56 |
| mal | Danct12_: what do you have in droid-hal spec? | 16:57 |
| Danct12_ | mal: https://raw.githubusercontent.com/sailfish-santoni/droid-hal-santoni/master/droid-hal-santoni.spec | 16:57 |
| Danct12_ | i haven't touched this file since the start of this pie rebase | 16:58 |
| deathmist | Danct12_: show "ls -l /" from device | 16:58 |
| Danct12_ | deathmist: https://paste.ubuntu.com/p/PVKnXvd9yk/ | 16:59 |
| deathmist | you probably need at least /persist, /firmware & /dsp as stragglers like in https://git.io/Jvk8W | 17:00 |
| mal | Danct12_: you probably should have some straggler_files in there: dsp, firmware, persist at least, do you have the custom vendor.mount which you probably need? | 17:00 |
| Danct12_ | i don't think i do | 17:01 |
| mal | deathmist: did you need the vendor.mount hack to make sure mounts are done in correct order? | 17:02 |
| Danct12_ | also regarding straggler files, they don't show up during i was building package, despite it was mentioned in the pdf file | 17:02 |
| deathmist | mal: I didn't even have /vendor in my fstab https://git.io/Jvk80 so I needed it due to that, not sure if he'd need it though | 17:02 |
| mal | Danct12_: there is a bug | 17:02 |
| mal | deathmist: he needs it | 17:03 |
| mal | deathmist: check https://paste.ubuntu.com/p/WgQDkfwqN8/ it has /vendor and then other partitions mounted to subfolders of /vendor | 17:03 |
| mal | Danct12_: so you need these https://github.com/sailfishos-oneplus5/droid-config-cheeseburger/blob/master/sparse/lib/systemd/system/vendor.mount and https://github.com/sailfishos-oneplus5/droid-config-cheeseburger/blob/master/sparse/lib/systemd/system/local-fs.target.wants/vendor.mount (note that the latter of those is a symlink not a file) | 17:05 |
| Danct12_ | replace /dev/sdf6 with the correct partition for /vendor? | 17:06 |
| deathmist | I think he also needs to remove vendor/cust line from fstab then as to not make droid-hal build pick those up and create incorrect units in droid-hal RPMs and cause conflict with droid-configs https://git.io/Jvk8B | 17:06 |
| deathmist | ofc Danct12_ | 17:06 |
| mal | deathmist: true | 17:06 |
| T42 | <Sid127> how large is the source tree after a sync with the hybris-16.0 branch? | 17:09 |
| deathmist | Danct12_: you should remove "vendor-bt_firmware.mount" from vendor.mount since you don't seem to have a mount for that | 17:09 |
| deathmist | mine was less than 50G but I also do a shallow clone, drop tags & bundles during sync etc | 17:11 |
| T42 | <Sid127> Oh okay... | 17:12 |
| T42 | <Sid127> This could be an issue | 17:12 |
| Danct12_ | and now i need to add the straggler files to droid-hal | 17:12 |
| T42 | <Sid127> I'm dual booting Linux and windows, and my dad has let me have only 100GB for the home partition on Linux | 17:13 |
| Danct12_ | and then i rebuild droid-hal, droid-configs then update the packages | 17:13 |
| deathmist | Danct12_: sounds about right | 17:14 |
| T42 | <Sid127> 100gb wouldn't be enough, right? | 17:16 |
| Danct12_ | i'm suffering with 100 gb | 17:17 |
| Danct12_ | and 53 gb is after i cloned everything | 17:17 |
| T42 | <Sid127> I'll abandon ship for now then | 17:18 |
| T42 | <Sid127> I'll give it a shot when I have more space | 17:18 |
| Danct12_ | maybe you should just buy yourself a computer, you have more freedom on that one :P | 17:18 |
| Danct12_ | and i dont think that's even needed, a portable hdd would do just fine | 17:19 |
| T42 | <Sid127> I should get a computer :) but sadly, I can't do that yet because I'll be entering college in a few years | 17:20 |
| T42 | <Sid127> but yes, portable hdd sounds like a good idea, and that's something I can save up for | 17:20 |
| mal | deathmist: btw, you don't have to edit fstab to skip generating mount services, you can add those to ignore line in droid-hal spec https://github.com/mer-hybris/droid-hal-sony-ganges-pie/blob/master/rpm/droid-hal-common.inc#L32 | 17:22 |
| deathmist | ah, right! @Sid127 you should add /vendor to your droid-hal spec then & not touch the fstab :) | 17:25 |
| Danct12_ | while i'm at it, about audit | 17:25 |
| deathmist | wrong @ oop Danct12_ | 17:25 |
| Danct12_ | for selinux, you enable it, but disable it on runtime? | 17:26 |
| Danct12_ | mal: ^ | 17:27 |
| deathmist | basically, you should leave it enabled in kernel and disable in cmdline https://git.io/Jvk87 (maybe not disable audit, it didn't have any visible regressions for me however) | 17:27 |
| Danct12_ | oh | 17:27 |
| Danct12_ | deathmist: and the change to mount vendor in systemd is still to kept, right? | 17:28 |
| deathmist | yep, it's now provided by droid-config RPM instead of droid-hal after the changes | 17:29 |
| Danct12_ | back | 17:56 |
| Danct12_ | had to restart my computer lol | 17:56 |
| T42 | <NotKit> @Sid127 [100gb wouldn't be enough, right?], can be enough, but do shallow clone | 17:59 |
| Danct12_ | and also the minimum ram amount for sailfishos builds is 8 gb :P | 18:05 |
| Danct12_ | not official minimum requirements but that's what i'm on now | 18:06 |
| Danct12_ | deathmist: file /lib/systemd/system/vendor.mount conflicts between attempted installs of droid-config-santoni-1-1.armv7hl and droid-hal-santoni-0.0.6-202001181758.armv7hl | 18:07 |
| Danct12_ | i guess that should be removed | 18:07 |
| deathmist | Danct12_: like mal linked before, just add /vendor to makefstab_skip_entries in droid-hal spec like so https://git.io/Jvk4c and build_packages -dc, should just work | 18:11 |
| deathmist | or is that on-device? | 18:19 |
| *** ahjolinna_ is now known as ahjolinna | 20:02 | |
| Danct12_ | deathmist: okay so it's still broken | 20:08 |
| deathmist | what do you mean? what is broken, boot process? show new journal then | 20:10 |
| Danct12_ | yeah the boot process is still broken :( | 20:11 |
| Danct12_ | https://paste.ubuntu.com/p/p9Zdh7zVnd/ | 20:14 |
| Danct12_ | deathmist: ^ | 20:14 |
| T42 | <elros34> Did you follow 9.3.1? Selinux is still not disabled correctly | 20:17 |
| deathmist | yeah your journal should have "selinux: Disabled at boot" or something like that | 20:24 |
| deathmist | Danct12_: you didn't flash new hybris_boot.img did you? (or even make it for that matter) | 20:26 |
| deathmist | assuming you already added "selinux=0" to boot cmdline | 20:26 |
| Danct12_ | just reflashed | 20:32 |
| Danct12_ | still the same error: https://paste.ubuntu.com/p/jP8TZXnC4d/ | 20:34 |
| Danct12_ | though, i don't think the binder spam was really a good error | 20:37 |
| T42 | <elros34> it's better, at least you added selinux=0 to boot params but you still have wrong kernel config | 20:45 |
| deathmist | Danct12_: SELinux wasn't disabled... you don't have SECURITY_SELINUX_BOOTPARAM (https://git.io/JvkRE) enabled in your kernel so right now doesn't care about the boot option | 20:45 |
| deathmist | binder spam should go away once everything starts working | 20:49 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!
