| T42 | <edp_17> No, I couldn't try to connect as my whole family is sleeping. 😊 The jolla demo video is played fine in the Gallery app. Does it say the video playback with gst-droid work? | 00:04 |
|---|---|---|
| T42 | <elros34> I guess. You can always confirm with logcat or GST_DEBUG. There should be messages from omx if hardware decoder is used. Is that same device for which you had wrong framerate in dconf? Maybe you have tried different values or removed it previously and that is why it worked? | 00:05 |
| T42 | <edp_17> I have not found any mini* service in either 'systemctl list-units -t service --all' nor 'systemctl --user list-units -t service --all' | 00:06 |
| T42 | <elros34> pgrep -af mini | 00:06 |
| T42 | <edp_17> No, it is a different device. It is the hammerhead. | 00:06 |
| T42 | <elros34> ah cm12? | 00:07 |
| T42 | <edp_17> @elros34 [pgrep -af mini], https://paste.ubuntu.com/p/t5sgzCkSjB/ | 00:07 |
| T42 | <edp_17> @elros34 [ah cm12?], Yes. 12.1 | 00:07 |
| T42 | <edp_17> I gave up the n7000 for a while because I don't know what I changed in my local repo that caused the camera stopped working on fresh builds. So, I moved on to hammerhead to try fixing the video recording. (If it is possible.) | 00:08 |
| T42 | <elros34> just idea I wonder whether this warning is important: https://github.com/sailfishos/gst-droid/blob/0.20201104.0/gst/droidcamsrc/gstdroidcamsrcdev.c#L588 | 00:13 |
| T42 | <edp_17> Could be. mal mentioned last time to check the media_codecs* for the n7000. | 00:16 |
| T42 | <edp_17> This is what I got on the hammerhead: https://paste.ubuntu.com/p/8J3wk2jdCV/ | 00:18 |
| T42 | <edp_17> The /etc/media_codecs_performance.xml is missing on the n7000 and that symlink is broken on the hammerhead. Should I remove that broken symlink? | 00:19 |
| T42 | <edp_17> This is the jolla-camera-hw.txt: https://paste.ubuntu.com/p/FFvQDKsxJB/ | 00:21 |
| T42 | <edp_17> and the output of GST_DEBUG=6 mk-cam-conf 0 & 1 /dev/null 2>&1 | grep params_parse | sed -e 's/.*param\s//': https://paste.ubuntu.com/p/wxDqY8gTKx/ | 00:23 |
| T42 | <elros34> sure you can remove broken symlinks but unlikely this will change anything just make sure all xml files in system and vendor are symlinked in /etc | 00:24 |
| T42 | <edp_17> Some of them are missing: https://paste.ubuntu.com/p/GxZPnNP6Ks/ | 00:27 |
| T42 | <edp_17> I'll remove this one too. | 00:27 |
| T42 | <elros34> there are few default symlinks in droid-config sparse I wouldn't worry about them but you can remove them like that: https://github.com/elros34/droid-config-moto_msm8960_jbbl/blob/master/delete_file.list | 00:29 |
| T42 | <edp_17> So, these are not needed? Okay, thanks. | 00:32 |
| T42 | <elros34> https://github.com/mer-hybris/droid-hal-configs/commit/0087626a1d7ef52e192fb83791c99889b6af0a18 | 00:33 |
| T42 | <edp_17> @elros34 [there are few default symlinks in droid-config …], Oh, you meant droid-config-device sparse. I've found them in there. | 00:34 |
| T42 | <edp_17> @elros34 [https://github.com/mer-hybris/droid-hal-config …], Thanks! I'll double check them and will add to all my ports. | 00:35 |
| T42 | <edp_17> These two urls do have contrary info. E.g. one of them says to keep '/system/etc/media_codecs_google_video.xml' while the other says to remove it. | 00:38 |
| T42 | <elros34> Unknown HAL color format 0x7f000789: video-frame-format = android-opaque = > OMX_COLOR_FormatAndroidOpaque = 0x7F000789 that is from you logs. Maybe you could ask some smart guy whether this is important | 00:39 |
| T42 | <edp_17> I am asking him. 😃 | 00:39 |
| T42 | <elros34> sure I am just a porter like you:) but seriously maybe mal or abranson would know something more about this | 00:40 |
| T42 | <edp_17> Is abranson on this channel? | 00:40 |
| T42 | <edp_17> I know he is on the fan club one. | 00:41 |
| T42 | <edp_17> Yep, he is not in here. | 00:41 |
| T42 | <edp_17> Not with tg at least. | 00:42 |
| T42 | <elros34> not sure if he is still available here but for sure he was | 00:42 |
| T42 | <elros34> btw that delete_file.list correct for my device, do not use same for yours | 00:42 |
| T42 | <edp_17> Okay, I see. | 00:43 |
| T42 | <edp_17> Right before the camera service dies, the logcat says: "E/DroidMediaCodec( 2109): Received a buffer without a timestamp!" | 00:47 |
| T42 | <elros34> you have a lot "Replacing invalid timestamp: 99:99:9" so that could be related | 00:52 |
| T42 | <edp_17> Whether is this time stamp used when generating the file name? | 00:52 |
| T42 | <edp_17> Interestingly, there are .mp4 files in $HOME/Videos/Camera. All are empty but their file name is correct. | 00:54 |
| T42 | <edp_17> Okay, I leave this for now. I'll ask Andrew tomorrow. Now I install the fresh image on the n7000 to see whether camera works there. Annoying, as I only have the camera and bluetooth to fix and the device works fine. (and the camera is fixed on the first build on the device but doesn't work on fresh builds.) | 00:57 |
| techgeekster | Hi to everyone! When building the mw i get this error: Problem: libhybris-libGLESv2-0.0.5.42+master.20210311093239.d16e02c-1.aarch64 conflicts with mesa-llvmpipe-libGLESv2 provided by mesa-llvmpipe-libGLESv2-19.3.3+git2-1.5.2.jolla.aarch64 | 09:36 |
| techgeekster | Solution 1: deinstallation of mesa-llvmpipe-libGLESv2-19.3.3+git2-1.5.2.jolla.aarch64 | 09:36 |
| techgeekster | Solution 2: do not install libhybris-libGLESv2-0.0.5.42+master.20210311093239.d16e02c-1.aarch64 | 09:36 |
| techgeekster | I tryied removing the mesa-llv... package but the problem remains | 09:36 |
| *** techgeekster_ is now known as techgeekster | 09:46 | |
| srxl | Hello all! Looking to start work on a port for a device I have. The HADK talks about using libhybris to make a HAL for my device, but I'd rather avoid that if possible - I already have a (very close to) mainline kernel that works very well for my device. Are there any resources for building a port without libhybris? | 10:07 |
| techgeekster | Problem solved. I was dumb and did the FAQ Solution in PlatformSDK. | 11:06 |
| rinigus | I have a rare issue of mce starting panic leds after flashing 4.0.1.48 and installing jolla default apps. doesn't seem to occur after reboot. summarized https://github.com/sailfishos-sony-tama/main/issues/130 . couldn't see anything special in the logs either | 11:23 |
| rinigus | in the meeting, sledges mentioned something regarding xperia 10 reboots. any idea if it could be related? a fix that I can look into? | 11:24 |
| *** mal__ is now known as mal | 13:07 | |
| T42 | <lackingcheesepuffs> I am building droid-hal but it saying this while the build goes on... Will this cause an issue later? | 13:29 |
| T42 | <lackingcheesepuffs> https://paste.ubuntu.com/p/P45g8FJQJF/ | 13:29 |
| T42 | <elros34> yes and even it points you to solution in next line | 13:32 |
| mal | run "export TEMPORARY_DISABLE_PATH_RESTRICTIONS=true" before make hybris-hal | 13:33 |
| T42 | <swanux> Just wondering if anyone has any ideas for it (still that lipstick situation) - https://pastebin.ubuntu.com/p/ZbzppWmSJf | 13:37 |
| T42 | <lackingcheesepuffs> Alright | 13:39 |
| T42 | <lackingcheesepuffs> I will try again | 13:39 |
| mal | @swanux which android base? which device? | 13:40 |
| T42 | <swanux> mal: dipper (Mi 8) los 16.0 | 13:41 |
| mal | does test_hwcomposer work? | 13:41 |
| T42 | <swanux> No. It's in the paste (with link to full logcat). Both test_hwcomposer and lipstick logs | 13:42 |
| mal | and you run it using EGL_PLATFORM=hwcomposer test_hwcomposer | 13:43 |
| swanux | Yes of course. | 13:43 |
| mal | swanux: can you find your generated android-config.h from droid-hal devel rpm? | 13:46 |
| swanux | mal: Yes, here it is (from droid-hal-dipper-devel.rpm) https://pastebin.ubuntu.com/p/hJH7CZ9pgd/ | 13:48 |
| mal | ok, so that rules out one possible reason | 13:50 |
| mal | is that aarch64 build? | 13:51 |
| swanux | Yes it is. | 13:51 |
| mal | do any other tests work? | 13:51 |
| swanux | test_gps, test_dload, test_opencl seem to work | 13:52 |
| mal | hopefully you have strace installed, could you run this: EGL_PLATFORM=hwcomposer strace test_hwcomposer &> log.txt | 13:52 |
| mal | and then show that log.txt | 13:53 |
| swanux | Ok | 13:53 |
| swanux | mal: Well, I had to stop it manually (as you can see) because the last part loops infinitely otherwise. https://paste.ee/p/gVRXt | 14:03 |
| mal | swanux: can you make a quick test, run this on device: ln -s /usr/libexec/droid-hybris/system/lib64/libGLESv2.so /usr/libexec/droid-hybris/system/lib64/libGLESv3.so | 14:08 |
| swanux | I ran it. | 14:09 |
| mal | any difference with test_hwcomposer? | 14:10 |
| swanux | Yeah. Now it isn't running infinitely but throwing segfault instead (like lipstick before) | 14:11 |
| mal | new strace please | 14:12 |
| swanux | Here's the output : https://pastebin.ubuntu.com/p/9rtghQYr47/ (doing strace now...) | 14:12 |
| mal | hmm, wait | 14:12 |
| mal | swanux: another attempt, remove that /usr/libexec/droid-hybris/system/lib64/libGLESv3.so and then make a copy of the /usr/libexec/droid-hybris/system/lib64/libGLESv2.so with name /usr/libexec/droid-hybris/system/lib64/libGLESv3.so | 14:14 |
| swanux | Done. Should I reboot or can I run the test now? | 14:15 |
| mal | just run it again | 14:15 |
| mal | maybe you could reboot also at some point just to be sure | 14:16 |
| swanux | Same output | 14:16 |
| swanux | Again after reboot | 14:17 |
| swanux | Here's the strace (from the first time after creating the symlink, but as I can see nothing changed) - https://pastebin.ubuntu.com/p/vXS6pqDD7f/ | 14:18 |
| mal | swanux: do you have your config repo in github? | 14:20 |
| swanux | Yes (it's private though). I'll share it in a minute | 14:20 |
| swanux | mal: https://github.com/swanux/droid-config-dipper-16/settings | 14:22 |
| swanux | * https://github.com/swanux/droid-config-dipper-16 | 14:23 |
| mal | ok, you are missing something | 14:23 |
| techgeekster | Small Question. In the build_packages.sh --gg, it produces this error (meson.build:36:0: ERROR: Git program not found.) Were does it want to install git? | 14:24 |
| mal | does this help: sudo zypper in git | 14:24 |
| swanux | mal: What am I missing? | 14:25 |
| mal | swanux: I can't see the ld.config.28.txt things in your repo mentioned here https://github.com/mer-hybris/hadk-faq#hybris-16-0 | 14:26 |
| mal | the last 3 parts in there | 14:27 |
| techgeekster | 'git' is already installed. Maybe it wants git in the Target build? What else could it be? | 14:27 |
| mal | then try: sb2 -t $VENDOR-$DEVICE-$PORT_ARCH -m sdk-install -R zypper in git | 14:27 |
| swanux | mal: Thanks. I'll push them now manually just to check what they change (will put them in my repo later) | 14:28 |
| techgeekster | Installing git in the target solved the previous fault. Now a new one. 'gstreamer-photography-1.0' not found in package names. Trying capabilities. No provider of 'gstreamer-photography-1.0' found. | 14:42 |
| techgeekster | The exact error is: Looking for a fallback subproject for the dependency gstreamer-photography-1.0 | 14:44 |
| techgeekster | meson.build:36:0: ERROR: Subproject directory not found and gstreamer-plugins-bad.wrap file not found | 14:44 |
| swanux | mal: I placed those files according to the guide and rebooted, but test_hwcomposer outputs the same as before. | 14:44 |
| mal | techgeekster: ah, sorry, porting is a bit complicated now because latest gst-droid shouldn't be used | 14:46 |
| techgeekster | Oh ok. How can I fix this? Or there is nothing I can do? | 14:47 |
| mal | techgeekster: go to hybris/mw/gst-droid and run "git clean -x -f -d" and then "git reset --hard 0.20201104.0" and then check that the common/ folder in there has content, then back to $ANDROID_ROOT and run rpm/dhd/helpers/build_packages.sh -b hybris/mw/gst-droid | 14:52 |
| techgeekster | mai: got it. trying it now | 14:53 |
| mal | swanux: is the ld.config.28.txt mounted, check output of mount | 14:53 |
| mal | swanux: also make sure you copied the file correctly (in raw mode) | 14:53 |
| swanux | mal: Here's the cat from the file - https://pastebin.ubuntu.com/p/n3f7ZGJ32D/ and here's the output of mount - https://pastebin.ubuntu.com/p/f6BdZNbBQT/ - I don't see it as mounted | 14:56 |
| mal | swanux: I don't see it in mount | 14:57 |
| mal | did you reboot? | 14:57 |
| swanux | Yes. | 14:58 |
| mal | did you symlink the service file correctly? | 14:58 |
| T42 | <lackingcheesepuffs> @elros34 [yes and even it points you to solution in next …], When building droid-hal in platform sdk I face this | 14:58 |
| T42 | <lackingcheesepuffs> https://paste.ubuntu.com/p/cQkjjr7pV3 | 14:58 |
| T42 | <lackingcheesepuffs> Is this something to do with my defconfig? | 14:58 |
| T42 | <lackingcheesepuffs> https://github.com/SailfishOS-MI-A2/android_kernel_xiaomi_sdm660/blob/hybris-17.1/arch/arm64/configs/wayne_defconfig | 14:58 |
| T42 | <elros34> probably, mer-kernel-check finished without errors? | 14:59 |
| swanux | mal: I think so. I checked and it doesn't seem to be broken, and rebooted again (just in case) but no change in mount | 15:00 |
| mal | swanux: meaning /usr/lib/systemd/system/system-etc-ld.config.28.txt.mount is a file and /usr/lib/systemd/system/local-fs.target.wants/system-etc-ld.config.28.txt.mount is a symlink pointing to that earlier file (ln -s ../system-etc-ld.config.28.txt.mount /usr/lib/systemd/system/local-fs.target.wants/system-etc-ld.config.28.txt.mount) | 15:00 |
| mal | swanux: note the paths in faq used the old /lib location | 15:00 |
| mal | I should fix that | 15:00 |
| techgeekster | mal: The git reset solution worked. Thanks! | 15:10 |
| swanux | mal: I changed the paths, here's the mount output now ( still can't see it) and the commands after that (which show that the files - I think - are at proper locations) https://pastebin.ubuntu.com/p/6Y33yjBxmJ/ | 15:11 |
| mal | swanux: show content of /usr/lib/systemd/system/system-etc-ld.config.28.txt.mount | 15:16 |
| mal | swanux: btw, do you even have file /system/etc/ld.config.28.txt | 15:17 |
| swanux | I have that file under usr/libexec/droid-hybris/system/etc/ld.config.28.txt ... I guess new location here too (?) | 15:18 |
| mal | swanux: I asked if there is a file /system/etc/ld.config.28.txt | 15:20 |
| swanux | mal: here's the contetn: https://pastebin.ubuntu.com/p/n2JDJ6wb2W/ | 15:20 |
| swanux | No, there isn't | 15:20 |
| mal | so then the service won't work, can you show the content of any ld.config file in /system/etc/ | 15:21 |
| mal | if there are any | 15:21 |
| swanux | https://pastebin.ubuntu.com/p/QGH44FnkKQ/ There is one | 15:22 |
| swanux | mal: Soo, how to proceed? | 15:31 |
| mal | not sure yet | 15:36 |
| T42 | <edp_17> @elros34 : Is it possible to install your mouse tracker without patchmanager? (On sfos 4.0.1.48 I cannot install it with patchmanager.) Thanks. | 15:46 |
| techgeekster | When building with --mic, this error appeared (Error <creator>[03/11 15:41:13] : Unable to find pattern: Jolla Configuration raphael). I tried upgrading tooling and target. Everything is on 4.0.1.48. Still the error stands | 15:46 |
| mal | swanux: just to verify, you have fully aarch64 based build? | 15:48 |
| T42 | <elros34> @edp_17 should be possible if you ignore dependencies and then apply patch manually. You can also use pm2: https://build.merproject.org/package/binaries/home:elros34:sailfishapps/patchmanager?repository=sailfishos_4.0.1.48 | 15:48 |
| mal | swanux: noticed that one patch is missing from hybris-16.0 base for aarch64 | 15:48 |
| T42 | <edp_17> Thanks @elros34 ! | 15:49 |
| swanux | mal: Yeah, I think so. | 15:50 |
| T42 | <elros34> @techgeeksterlooks looks you have outdated target/tooling | 15:51 |
| mal | swanux: so PORT_ARCH is aarch64? | 15:51 |
| swanux | Yes | 15:52 |
| techgeekster | This is the result of sdk-foreach-su -ly ssu re (https://paste.gnome.org/p1i7jgvwd) | 15:52 |
| mal | I will push the needed fixes to github soon | 15:52 |
| mal | techgeekster: is that port using PORT_ARCH aarch64? | 15:53 |
| T42 | <elros34> @techgeekster do you really updated it or just set release with ssu? I ask because " Jolla Configuration" is not used anymore on 4.0.1.x | 15:53 |
| mal | @elros34 probably the issue that aarch64 mic builds have issues because of broken sed | 15:53 |
| mal | in build_packages.sh | 15:54 |
| techgeekster | I've started with outdated tooling but did this ( run "sdk-foreach-su -ly ssu re 4.0.1.48" then "sdk-foreach-su -ly zypper ref" and finally "sdk-foreach-su -ly zypper dup") | 15:54 |
| T42 | <elros34> mal: no, that is different error. It's because latest target/tooling points to 3.4.0.x | 15:54 |
| mal | swanux: looks like I forgot to push the changes before, I have had the needed changes locally since november | 15:55 |
| swanux | Which repo? | 15:55 |
| mal | swanux: hybris-patches, I will check that my commit is ok before pushing | 15:55 |
| T42 | <elros34> @techgeekster check with sb2 /etc/os-release in target | 15:56 |
| mal | @elros34 are you sure? the ssu re looked ok | 15:58 |
| T42 | <elros34> mal: I thinks so. "Unable to find pattern: Jolla Configuration raphael" shouldn't happen on 4.0.1. FOr sed issue he should have 'Unable to find patterns-sailfish-device-configuration-raphae' | 15:59 |
| T42 | <elros34> @techgeekster do you use meta packages for patterns? Do you in droid-config patterns/*.inc files? | 16:01 |
| techgeekster | @T42 Not sure, let me check | 16:03 |
| mal | techgeekster: check .ks file in $ANDROID_ROOT does it have any adaptation* repos in the repo list | 16:05 |
| T42 | <elros34> and whether you have pattern or package in %packages section of ks file | 16:07 |
| techgeekster | %packages | 16:07 |
| techgeekster | @Jolla Configuration raphael | 16:07 |
| techgeekster | %end | 16:07 |
| T42 | <elros34> rebuild droid-config and then check that file in droid-local-repo/$DEVICE/droid-config/droid-config-$DEVICE-ssu*.rpm | 16:09 |
| techgeekster | Ok. Should I delete the .ks file? | 16:09 |
| techgeekster | In order to start "clean"? | 16:10 |
| T42 | <elros34> you can but it should be generated every time --mic is used | 16:10 |
| techgeekster | Ok! | 16:11 |
| mal | swanux: download (as raw file) and apply this https://github.com/mer-hybris/hybris-patches/blob/rebase-16.0/frameworks/native/0007-hybris-Fix-loading-64-bit-EGL-libraries.patch to frameworks/native | 16:38 |
| swanux | mal: I downloaded it. How do I apply it? (sorry for the stupid question) | 16:43 |
| mal | swanux: copy it to that frameworks/native folder and run git am 0007-hybris-Fix-loading-64-bit-EGL-libraries.patch | 16:47 |
| mal | of course in that folder | 16:47 |
| T42 | <lackingcheesepuffs> @mal [swanux: download (as raw file) and apply this …], is this required for android 10? | 16:48 |
| mal | android 10 has that already, this is for android 9 | 16:48 |
| mal | hmm, actually no | 16:49 |
| mal | wait | 16:49 |
| swanux | mal: https://pastebin.ubuntu.com/p/YyHF7yztz6/ | 16:49 |
| mal | @lackingcheesepuffs I need to check that | 16:50 |
| mal | @lackingcheesepuffs looks like that was also missing from hybris-17.1 | 16:53 |
| T42 | <lackingcheesepuffs> oof, so do I need to apply patch before building android hal? | 16:54 |
| mal | yes, there are two patches I added to hybris-17.1 | 16:55 |
| mal | not yet merged | 16:56 |
| swanux | How to resolve that error? | 16:58 |
| mal | swanux: I have no idea how that is possible | 17:00 |
| mal | swanux: is there file opengl/libs/EGL/Loader.cpp in frameworks/native? | 17:01 |
| swanux | mal: Yes | 17:02 |
| mal | try running "git fetch github" first | 17:05 |
| swanux | Done. Same output on apply attempt | 17:07 |
| mal | what does git log say | 17:10 |
| swanux | Here's the first 3 commit from log: https://pastebin.ubuntu.com/p/rmbdSZzzNK/ | 17:12 |
| mal | swanux: wait that is from your frameworks/native? does git diff show something? | 17:18 |
| mal | swanux: run git am --abort | 17:18 |
| swanux | Sure it does https://pastebin.ubuntu.com/p/RSX7kGdhXk/ | 17:19 |
| swanux | OK | 17:19 |
| mal | swanux: so did you accidentially run hybris-patches/apply-patches.sh instead of hybris-patches/apply-patches.sh --mb? | 17:20 |
| mal | anyway after the git am --abort run patch -p1 < 0007-hybris-Fix-loading-64-bit-EGL-libraries.patch | 17:21 |
| deathmist_ | mal: any idea about https://irclogs.sailfishos.org/logs/%23sailfishos-porters/%23sailfishos-porters.2021-03-10.log.html#t2021-03-10T13:02:20? I can't seem to boot SFOS 4.0 on hybris-16.0/arm32 after updating submodules and such | 17:24 |
| mal | deathmist_: did you move systemd service to /usr/lib from /lib also which kernel version does the device? | 17:25 |
| mal | so many typos in my comment | 17:26 |
| deathmist_ | mal: 4.4 kernel and yes I did that back when SFOS 3.4 was current as per your previous suggestion: https://github.com/sailfishos-oneplus5/droid-config-cheeseburger/commit/6e447a9 | 17:26 |
| swanux | mal: To be honest I built it so many times in the previous weeks and tried so many variations that I have no idea at all | 17:27 |
| swanux | Ok, I patched it. | 17:27 |
| mal | swanux: then make hybris-hal as usual and package droid-hal | 17:27 |
| mal | deathmist_: can you show output of mer kernel checker for you kernel, make sure you have latest kernel checker first | 17:28 |
| swanux | So I need only the rootfs to be repackaged, not the boot.img? (after rebuilding of course | 17:28 |
| mal | swanux: or you can manually install the needed droid-hal rpms for testing directly to device | 17:30 |
| swanux | Good to know. So every rpm from droid-hal? | 17:30 |
| mal | no, only those that have been installed before | 17:31 |
| deathmist_ | mal: I suppose droid-hal-$DEVICE.log works too since it contains that output: https://pastebin.com/raw/0K6zQEid | 17:31 |
| swanux | mal: So not everything is installed from droid-local-repo/droid-hal-DEVICE? | 17:32 |
| mal | deathmist_: yes, assuming you had updated to latest kernel checker before that | 17:32 |
| mal | deathmist_: should be easy to rerun it if you have the out/ still in your build env | 17:32 |
| mal | deathmist_: did you do the changes in the init.rc patch mention in the chat log? | 17:33 |
| deathmist_ | mal: seems to be the same output https://pastebin.com/raw/Vk9ZWTCW and if you mean https://github.com/mer-hybris/hybris-patches/commit/54008fe yes I manually commented out the remaining lines mentioning cg2_bpf (mount was already commented during droid-hal build) | 17:38 |
| mal | ok | 17:39 |
| deathmist_ | ah, because of droid-hal that mount was ofc turned into a systemd unit, let me try masking it | 17:40 |
| mal | deathmist_: what do you mean? | 17:41 |
| mal | ah, yes, it might be running on device in theory | 17:41 |
| deathmist_ | hmm that didn't seem to help still | 17:44 |
| deathmist_ | still same "systemd[1]: System is tainted: var-run-bad" with "API bus" and "system bus" errors after D-Bus is started | 17:46 |
| mal | deathmist_: show output of mount | 17:49 |
| deathmist_ | mal: https://pastebin.com/raw/EMHqvFiD | 17:50 |
| mal | deathmist_: is /var/run a symlink | 17:51 |
| deathmist_ | nope, directory | 17:51 |
| deathmist_ | seems it should point to /run? | 17:51 |
| mal | yes | 17:52 |
| mal | wondering why it's like that | 17:52 |
| mal | deathmist_: on my xa2 with 4.4 kernel also I have /var/run -> ../run | 17:52 |
| deathmist_ | I'll remove the dir from recovery and create the symlink | 17:53 |
| deathmist_ | or is there some info you'd like from me before I do that? | 17:53 |
| mal | deathmist_: the hint for that was "System is tainted: var-run-bad" | 17:53 |
| T42 | <elros34> check whether systemd-tmpfiles-setup.service create that directory | 17:53 |
| deathmist_ | mal: @elros34 on my 4.0.1.48 rootfs inside the flashable zip /var/run exists as a directory containing connman, console, faillock & vpnc subdirs | 17:55 |
| deathmist_ | (those subdirs are all empty) | 17:57 |
| deathmist_ | I assume this isn't normal? | 17:58 |
| mal | deathmist_: I check some image I made, those have just the symlink | 17:58 |
| mal | deathmist_: how did you build the image, using manual mic command or using build_packages.sh --mic? | 17:59 |
| deathmist_ | build_packages.sh --mic | 17:59 |
| mal | deathmist_: one thing to run if you still can, rpm -qf /var/run | 17:59 |
| deathmist_ | file /var/run is not owned by any package | 18:00 |
| mal | are the folders under it owned by something? | 18:01 |
| deathmist_ | /var/run/{console,faillock} are both owned by pam-1.3.1+git1-1.2.1.jolla.armv7hl, vpnc /var/run/{connman,vpnc} aren't owned by anything however | 18:03 |
| mal | those are normal | 18:05 |
| mal | wondering what goes wrong during mic build | 18:06 |
| deathmist_ | I can send a build log if you'd like | 18:06 |
| mal | deathmist_: why do you have this custom version https://github.com/sailfishos-oneplus5/droid-config-cheeseburger/blob/master/kickstart/pack_package-droid-updater | 18:07 |
| deathmist_ | mal: the default hybris-boot installer (https://github.com/mer-hybris/hybris-boot/blob/master/updater-script + updater-unpack.sh) didn't work for me due to broken TWRP tar, so I went an extra mile and created a custom one @ https://github.com/sailfishos-oneplus5/hybris-installer (I also do a few other smaller things on top of just working around broken tar) | 18:13 |
| T42 | <elros34> maybe running something like: 'rpm -qp --scripts *-4.0.1.48/* | grep /var/run' in /srv/mer/sdks/sfossdk/var/tmp/mic/cache/packages could reveal if there is some package which create that directory | 18:13 |
| deathmist_ | I did check if that file needed updating before updating to 4.0 however and it didn't seem like it was touched in years | 18:13 |
| mal | deathmist_: how long have you had that custom installer? | 18:13 |
| deathmist_ | since the beginning pretty much :p | 18:14 |
| deathmist_ | Jul 2019, when 3.0.3.9 was current | 18:14 |
| mal | deathmist_: just wondering that the difference there is the --numeric-owner option, maybe it's not critical then | 18:14 |
| mal | deathmist_: so the zip already has the folder instead of symlink? | 18:15 |
| mal | so then installation is then not the problem | 18:15 |
| deathmist_ | the rootfs .tar.bz2 yes, it has /var/run directory as I described earlier | 18:15 |
| mal | deathmist_: can you check output of "sdk-foreach-su -ly ssu re" | 18:17 |
| deathmist_ | mal: https://pastebin.com/raw/WnCp1gLE | 18:18 |
| mal | deathmist_: in your 3.4.0 build in ci the symlink is correct | 18:19 |
| deathmist_ | really odd issue, removing the dir and adding the proper symlink in manually does fix the issue for now though | 18:34 |
| mal | deathmist_: does a new image build result in the same folder? | 18:36 |
| deathmist_ | let me try | 18:36 |
| techgeekster | Where in the rpm/dhd/helpers/build_packages.sh script is the kernel_check called? | 18:39 |
| mal | techgeekster: https://github.com/mer-hybris/droid-hal-device/blob/master/droid-hal-device.inc#L410 | 18:40 |
| mal | it's not in the script, it's in the inc file used in droid-hal spec | 18:40 |
| techgeekster | Oh okk. Thank you! | 18:42 |
| deathmist_ | mal: another mic image build still resulted in /var/run directory with the 4 empty subdirs getting created inside rootfs.tar.bz2 | 18:43 |
| mal | deathmist_: anything obvious in build log? | 18:45 |
| deathmist_ | mal: searching for "/run" at least didn't yield anything in https://pastebin.com/raw/fumung84. here's also "rpm -qp --scripts *-4.0.1.48/* | grep /var/run" as per @elros34's suggestion https://pastebin.com/raw/G7hK4BEn | 18:50 |
| T42 | <Chrisme69> @Miklos have you got any idea about nexus 5 not starting after a Sfos 3.2 reflash cover Cm12.1? | 18:52 |
| T42 | <Chrisme69> I tried the Cm 12.1 unofficial and even this version by itself doesn't start | 18:53 |
| T42 | <Chrisme69> Then I tried with a Cm12.1 yoga: it starts but then the overwritten sfos doesn't boot | 18:54 |
| T42 | <elros34> @deathmist: error: filesystem-3.1+git4-1.2.1.jolla.noarch: install failed. filesystem provides that symlink | 18:59 |
| deathmist_ | @elros34 aha, makes more sense now as that was indeed not installed on my device, BUT a "zypper in filesystem" just worked strangely enough on device after the fact | 19:03 |
| deathmist_ | installing the package also fixed practically all apps crashing, seems it's pretty critical ;) (hint: mic should fail if that package fails to install for whatever reason) | 19:16 |
| deathmist_ | I suppose now that my hybris-16.0/arm32 port is working again I can attempt going over to arm64, and then finally attempt migrating to hybris-17.1 once again | 19:21 |
| T42 | <elros34> Did you figure out why it fail? Maybe findmnt will show some unusual permissions on /sys. At least error suggest sysfs issue | 19:22 |
| T42 | <elros34> on my host /sys have dr-xr-xr-x root:root | 19:27 |
| mal | what could cause that filesystem package failure | 19:30 |
| deathmist_ | @elros34 I've got the same, for now I'm assuming it's due to the fact that I was running platform SDK (chroot) inside a systemd-nspawn container so I'll do a re-test soon with platform SDK setup on my host instead | 19:37 |
| deathmist_ | I was building all of SFOS + HAL stuff inside the container so I had a clean env and so I could write personal instructions on what deps were needed as the Ubuntu 14.04 HA BUILD chroot is practically useless now that repo needs Python 3.5+ | 19:38 |
| T42 | <elros34> repo for python2 is still available at android.com | 19:40 |
| deathmist_ | oh well, I notiecd HAL build times were noticeably faster on my host as well and I'd rather not deal with building Android 10 under Ubuntu 14.04 soon anyway ;) | 19:44 |
| mal | deathmist_: I do my hybris-hal builds directly on ubuntu 20.04, just need some small things usually | 19:57 |
| mal | except fp2 for which I do those still in ubu-chroot | 19:58 |
| rinigus | after OTA (3.4 -> 4.0.1) is finished (version --dup) and screen is off, pressing power button shows GUI for a sec, then blank screen, and then leads to reboot. after automatic OTA induced boot image update, reboots into GUI. is such reboot after OTA expected now? | 20:01 |
| rinigus | issue with log https://github.com/sailfishos-sony-tama/main/issues/131 | 20:02 |
| T42 | <Chrisme69> An info. I don't know if this is the right group. I'm trying to install Sfos 3.2 on a nexus 5 (hammerhead) without results | 20:28 |
| mal | where did you get the image for that? | 20:28 |
| mal | what problem do you have? | 20:29 |
| T42 | <Chrisme69> I'm flashing it over a CM12.1 unofficial but It doesn't work | 20:29 |
| mal | you mean installation fails? | 20:29 |
| mal | or booting it fails? | 20:29 |
| T42 | <Chrisme69> Cm12.1 doesn't boot even. | 20:30 |
| T42 | <Chrisme69> I tried the yoga one and that's starts | 20:30 |
| T42 | <Chrisme69> Then I tried to flash sfos over it 3.2 | 20:30 |
| T42 | <Chrisme69> And doesn't start | 20:31 |
| T42 | <Chrisme69> Boot | 20:31 |
| T42 | <Chrisme69> The only sfos that worked over a Cm 13 was a sfos 3.0.3 that hasn't got any sfos opportunity | 20:31 |
| T42 | <Chrisme69> https://talk.maemo.org/showpost.php?p=1562106&postcount=785 | 20:32 |
| T42 | <Chrisme69> That's the post where I took some instructions | 20:33 |
| mal | why did you install sfos 3.2? | 20:34 |
| T42 | <Chrisme69> Cause it had android support | 20:35 |
| T42 | <Chrisme69> I tried the 3.2.0 | 20:37 |
| T42 | <Chrisme69> And the 3.4 | 20:37 |
| T42 | <Chrisme69> Now I'm downloading 4.0.1 | 20:38 |
| T42 | <Chrisme69> But has some issues with android support | 20:38 |
| mal | so installation works fine? | 20:45 |
| T42 | <Chrisme69> Not at all | 20:46 |
| T42 | <Chrisme69> The only one that works is the 3.0.3 from cm 13 | 20:46 |
| T42 | <Chrisme69> Then I do not manage to.upgrade to other versions | 20:47 |
| T42 | <Chrisme69> Tried 3.3 sfos and installation came to the language choice but touch screen wasn't working. It booted in half an hour. It's a lot | 20:59 |
| mal | @edp_17 are you still doing hammerhead, maybe you could help ^ | 21:06 |
| T42 | <Chrisme69> Yess | 21:08 |
| T42 | <Chrisme69> Now trying with 4.0.1 | 21:14 |
| T42 | <Chrisme69> Thank's a lot. Hope @edp_17 can help me. He already helped me finding the files hosted on talk.maemo | 21:23 |
| deathmist_ | any ideas where this "bash: Pictures: unknown operand" line could be coming from twice when opening a shell session for "defaultuser"? I'll go digging for this if no one else knows | 22:12 |
| deathmist_ | oh it comes from one of my extra additions to /etc/profile.d/ xD I'll need to fix that | 22:23 |
| deathmist_ | /bin/bash: symbolic link to ../usr/bin/busybox well that explains things | 22:26 |
| mal | yeah, sounded like something you had done to environment | 22:26 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!