lbt | PSA: git.sailfishos.org (git.merproject.org) is down for upgrades | 07:22 |
---|---|---|
rinigus | morning! as I have `make -j$(nproc --all) systemimage vendorimage` failing (missing libMpeg4SwEncoder), I have question regarding these targets | 07:53 |
rinigus | is the -syspart build expected to be just a "shadow" clean AOSP9 build? if it is, can I use sony instructions directly with the same android base as used for generation of tagged manifest? | 07:53 |
rinigus | ^ sony aosp9 based port | 07:53 |
vknecht | can't really say, but there's also the possibility to skip droid-{src,system,vendor} and only generate boot and userdata images (aka. community build way), but then you need to flash aosp beforehand (and disable userdata encryption) | 08:25 |
vknecht | but it's probably better to get it working the way it's intended :-) | 08:27 |
vknecht | iirc droid-system is a stripped down version of aosp, and it's fitted with vendor with lvm in a single partition so the vendor one can be used for factory reset files | 08:31 |
rinigus | vknecht: but I guess I can just use plain AOSP9 instructions for those images. which means that prebuilt kernel is in the repo tree. but since we boot into hybris anyway, it shouldn't matter | 08:33 |
rinigus | (I'll try with plain aosp9 and get the systemimage with vendorimage from that) | 08:44 |
rinigus | next issue: I am getting error while rpm/dhd/helpers/build_packages.sh --droid-hal . File not found: ... installroot/verity_key | 08:45 |
rinigus | ... make verity_key to the rescue. I do wonder why its not built with hybris-hal | 09:01 |
vknecht | check you have needed fstab and ueventd.rc files in out/, you might have to run additional make commands for those | 09:02 |
vknecht | won't prevent building, but cause problems for booting ;-) | 09:03 |
vknecht | eg. for suzu, had to "make fstab.suzu", and check I had those two files : | 09:06 |
vknecht | out/target/product/suzu/system/vendor/etc/fstab.suzu | 09:06 |
vknecht | out/target/product/suzu/root/fstab.loire | 09:06 |
vknecht | and had to manually copy vendor/ueventd.rc from device repo so that I had : | 09:07 |
vknecht | out/target/product/suzu/vendor/ueventd.rc | 09:07 |
vknecht | out/target/product/suzu/root/ueventd.rc | 09:07 |
rinigus | vknecht: thank you! ueventd.rc is there (vendor & root). there is no fstab in root, but is in vendor/etc; system/vendor is absent. btw, looks like your fstab in root is for loire | 09:19 |
vknecht | yes, no separate vendor partition for X / suzu, which is from Loire device family | 09:20 |
vknecht | you could check with "make modules" if there's a device-specific fstab target ; but maybe it's already ok with your device | 09:23 |
rinigus | vknecht: hard to say - I guess I'll see when I get to the flashing and booting stage :) | 09:28 |
vknecht | I'd expect a fstab.tama and fstab.akari somewhere in out/, in your case | 09:30 |
vknecht | indeed :-) | 09:31 |
vknecht | the droid-hal build log should give indications about what's found (but not necessarily about what's not found) | 09:32 |
rinigus | Have to check for .tama, forgot about that... | 09:38 |
rinigus | vknecht: interesting, I don't have any fstab.tama for aosp build either | 10:44 |
rinigus | looks like xperia 10 is mounting vendor_b as opt (https://github.com/mer-hybris/droid-config-sony-ganges-pie/blob/master/sparse/lib/systemd/system/opt.mount). that seems to be missing in https://github.com/vknecht/droid-config-sony-loire | 10:58 |
rinigus | there is also magical https://github.com/mer-hybris/droid-config-sony-ganges-pie/tree/master/sparse/usr/libexec/droid-hybris/system/etc with ld.config28.txt . no idea how device specific is that... | 11:15 |
vknecht | rinigus, there's simply no separate partition for /opt (nor on nile it seems) ; I guess a/b devices have more partitions to spare | 11:32 |
rinigus | vknecht: ok, I see. that does make sense | 11:33 |
rinigus | during libhybris build, I've got "Could not remove mesa-llvmpipe to install libhybris, assuming new mesa>=v19 and attempting alternative libhybris installation". this is something new as before we had to replace mesa by hybris | 11:59 |
rinigus | should I get it still replaced somehow or its automatic now? | 11:59 |
rinigus | when looking on whats installed with sb2 -t $VENDOR-$DEVICE-$PORT_ARCH -m sdk-install -R zypper se llvmpipe seems like only mesa-llvmpipe-libgbm and mesa-llvmpipe-libglapi are there | 12:02 |
vknecht | gitlab is back, but not yet ready for cloning ? | 12:12 |
vknecht | (with https, at least) | 12:13 |
Thaodan | Is it save to use an older version of bionic on android? | 12:35 |
Thaodan | I want to use bionic from android 9 on 10 because the tls handling was changed and I cant rebase the current patches. | 12:36 |
vknecht | ah, cloning worked now :-) | 12:39 |
g_k | I got this error https://pastebin.com/0JCGZ0C5 when im building hybris/mw/qt5-qpa-surfaceflinger-plugin | 14:26 |
g_k | how to fix it ? | 14:26 |
vknecht | just skip it, generally it's not needed | 14:27 |
g_k | ok ill try | 14:27 |
g_k | i builded the img using mic after that I flashed it and now its stuck on logo :( | 16:25 |
g_k | only telnet is working | 16:29 |
g_k | here is my init.log https://pastebin.com/7AuyS3XB | 16:32 |
g_k | help :C | 16:32 |
T42 | <NotKit> @Thaodan [Is it save to use an older version of bionic o …], likely won't work at all | 16:38 |
Thaodan | that's not good. do you have an idea how to rebase the patches for hybris on newer bionic? | 16:40 |
vknecht | g_k, is init.log complete ? also show "dmesg" and "journalctl -alb --no-pager" | 17:07 |
*** agander_ is now known as agander | 17:10 | |
g_k | https://pastebin.com/qrJ0ZLnd journalctl -alb --no-pager | 17:19 |
g_k | https://pastebin.com/cyHg7WJQ dmesg | 17:20 |
g_k | https://pastebin.com/USbUyiCn dmesg** | 17:28 |
vknecht | hmm, you don't seem to set selinux=0 on kernel commandline ? also, which android base is it ? | 17:34 |
g_k | lineage 15.1 | 17:39 |
g_k | i didn't set this | 17:40 |
g_k | so i need to set it ? | 17:40 |
vknecht | all devices need it, afaik | 17:41 |
g_k | ok | 17:41 |
vknecht | wonder if you're missing some fstab and/or ueventd.rc, afaik with 15.1+ there's some manual copying to do | 17:43 |
vknecht | showing the droid-hal rpm build log would be good too | 17:43 |
vknecht | and your fixup-mountpoint mod too | 17:44 |
g_k | ok | 17:46 |
g_k | droid-hal https://pastebin.com/Pb6Ncmb7 | 17:51 |
g_k | fixup-mountpoint https://pastebin.com/Lry0gtbQ | 17:58 |
vknecht | the '-' in mmcblk-1p.. looks strange | 17:59 |
vknecht | shouldn't all the '-1' be '0' ? | 18:01 |
g_k | https://pastebin.com/TdmCfULm here is the correct one | 18:03 |
g_k | sry | 18:03 |
vknecht | if you just fixed it, you should rebuild droid-hal, make new image and reflash | 18:09 |
g_k | no i did it before i just sent you the wrong file | 18:11 |
vknecht | ok ; well I can't pinpoint what's wrong exactly, hopefuly someone else will have some ideas | 18:12 |
g_k | ok ty | 18:14 |
vknecht | you can also read/search irc logs about 15.1 issues | 18:14 |
vknecht | yw | 18:14 |
*** agander_ is now known as agander | 18:19 | |
xreactx[m]__ | hello | 22:26 |
xreactx[m]__ | https://pastebin.com/UEwDVaEK I cannot figure out this error "libxz_intermediates/export_includes'" | 22:28 |
xreactx[m]__ | can anyone point me in the right direction? | 22:28 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!