| nk2IsHere | it seems to stay longer until reboot, so i'll get here once i grab my old laptop to see dmesg | 00:11 |
|---|---|---|
| *** wolfe.freenode.net sets mode: +v T4 | 02:45 | |
| Thaodan_ | is the current sensorfw compatible with the latest sfos? | 05:13 |
| T4 | <Harsh18262> i am getting some errors i think it is due to kernel but what is it exactly https://paste.ubuntu.com/p/3QTjy7NkK5/ | 05:24 |
| Thaodan_ | no idea try to compare the file to the file from the sony tree | 05:45 |
| Thaodan_ | that works | 05:45 |
| wdehoog | mal: when I use the non stripped init: http://susepaste.org/view//7279084 still no idea what change could have caused this to happen | 09:31 |
| T4 | <elros34> anybody have problem with invoker in 3.0.3?: "mapplauncherd[1406]: Daemon: dlopening single-instance failed: /usr/bin/single-instance: cannot open shared object file: No such file or directory" | 10:11 |
| mal | wdehoog: not sure why it happens now but based on that backtrace it fails here https://github.com/mer-hybris/android_bionic/blob/hybris-12.1/libc/bionic/system_properties.cpp#L667 | 11:23 |
| mal | wdehoog: so the only reason that can fail is if variable pi is null | 11:24 |
| mal | but in system_core there is null check for that | 11:26 |
| mal | and backtrace also says it's not null | 11:26 |
| wdehoog | mal: the serial field seems defined 'volatile'. maybe serial is causing the segfault? | 11:37 |
| wdehoog | do you know when/how the bionic source get compiled? | 11:38 |
| mal | bionic is built when hybris-hal is built | 11:38 |
| sensefusion | Hi guys, I can't build the rpm packages via running rpm/dhd/helpers/build_packages.sh according to guide: 'make[1]: "*** No rule to make target `mkbootimg.o', needed by `mkbootimg'. Stop."'. Any ideas why is mkbootimg.o missing? | 11:57 |
| sensefusion | https://pastebin.com/s5dhazsj | 12:14 |
| Nebrassy[m] | mal: so when I'm making fixup-mountpoints for an a/b device, I need to select a slot and point to only partitions on that slot or will slotselect flasg still work? | 16:44 |
| mal | Nebrassy[m]: yes, you need use one slot, usually a | 17:14 |
| mal | or basically you can use whatever partitions you want but usually it's nicer to use those from same slot | 17:15 |
| mal | sailfish doesn't handle slotselect | 17:15 |
| Nebrassy[m] | cool, I get this while building, should I keep it boot_a instead? | 17:15 |
| Nebrassy[m] | warning: ********************* /boot appears to live on /dev/block/bootdevice/by-name/boot | 17:15 |
| mal | you use that block/bootdevice/by-name/boot in fixup and just replace it with the real boot_a partition (e.g. mmcblk0pX) | 17:16 |
| Nebrassy[m] | yeah that's what I did | 17:17 |
| mal | same for other partitions in your fstab | 17:17 |
| Nebrassy[m] | yup, done that,I forgot how to solve this, how do i get image magick? | 17:18 |
| mal | quite often people add all partitions to fixup | 17:18 |
| mal | what? | 17:18 |
| Nebrassy[m] | yeah I added all except _b partitions | 17:18 |
| * Nebrassy[m] sent a long message: < https://matrix.org/_matrix/media/v1/download/matrix.org/vpfwMgDUwOyeEGDRElBptJas > | 17:19 | |
| mal | maybe sudo apt-get install imagemagick | 17:20 |
| mal | in HABUILD_SDK | 17:20 |
| mal | it's a ubuntu chroot so use normal way there to install if something is missing | 17:21 |
| Nebrassy[m] | oh forgot it's in the lists | 17:21 |
| Nebrassy[m] | is it possbile to build without sudo access? | 17:22 |
| mal | afaik sudo should work in that environment, or was it even needed, try | 17:26 |
| Nebrassy[m] | I mean to get the environment and ubuntu chroot working | 17:26 |
| Nebrassy[m] | it needs to access /asrv/mer | 17:27 |
| mal | but you are in there already if you are building hybris-hal | 17:27 |
| Nebrassy[m] | /srv/mer | 17:27 |
| Nebrassy[m] | I'm using another server | 17:27 |
| mal | not sure if I understand, you have that chroot installed? | 17:27 |
| mal | am I missing something | 17:28 |
| Nebrassy[m] | I'm currently using another server that I have sudo access to, this is not the server I usually use, the one I usually use doesn't have sudo access | 17:28 |
| mal | but you run that in chroot | 17:30 |
| mal | if you look at chapter 5.1 you see you already used sudo in that environment | 17:31 |
| Nebrassy[m] | sudo mkdir -p $PLATFORM_SDK_ROOT/sdks/sfossdk ; | 17:34 |
| Nebrassy[m] | sudo tar --numeric-owner -p -xjf Jolla-latest-SailfishOS_Platform_SDK_Chroot-i486.tar.bz2 -C $PLATFORM_SDK_ROOT/sdks/sfossdk ; | 17:34 |
| Nebrassy[m] | this is what I stopped at on the other server | 17:34 |
| mal | well nothing really forces you to extract that to that folder, you can extract it to somewhere under your home folder, you just need to give the correct path when entering ubu chroot | 17:37 |
| Nebrassy[m] | doesn't chroot itself need sudo? | 17:37 |
| mal | not sure really | 17:38 |
| Nebrassy[m] | I'll try later | 17:38 |
| mal | check what the ubu-chroot command does | 17:38 |
| Nebrassy[m] | ninja: error: 'libsf_compat_layer', needed by 'hybris-hal', missing and no known rule to make it | 17:42 |
| Nebrassy[m] | can't seem to find that | 17:42 |
| mal | go to $ANDROID_ROOT/external/ and run git clone --recurse-submodules https://github.com/mer-hybris/libhybris.git | 17:43 |
| Nebrassy[m] | thanks | 17:48 |
| *** b0n0 is now known as sfaccimm | 17:54 | |
| sensefusion | mal: Could you help me with my issue? | 18:08 |
| mal | not sure what is causing that, but I need to get some food now so maybe later | 18:14 |
| T4 | adarshkushwah was added by: adarshkushwah | 18:16 |
| T4 | <DSstill01> @sensefusion [mal: Could you help me with my issue?], I can help you | 18:27 |
| T4 | <DSstill01> What issue? | 18:27 |
| sensefusion | https://pastebin.com/s5dhazsj | 18:28 |
| T4 | <DSstill01> Did you verify that mkbootimg dir exists? | 18:32 |
| sensefusion | yep | 18:32 |
| T4 | <DSstill01> And it contains mkbooimg.c or such | 18:32 |
| sensefusion | no, it was somewhere else | 18:33 |
| sensefusion | there're 3 files in system/core/mkbootimg: Android.mk, bootimg.h and a python file mkbootimg withou a .py ending | 18:35 |
| T4 | <DSstill01> Hmm, do you have any bootimg location definition in your device tree? I dunno where the rpm takes it from | 18:37 |
| T4 | <DSstill01> But lemme see in system/core repo | 18:38 |
| *** OhYash1 is now known as ohyash | 18:39 | |
| T4 | <DSstill01> What is your android base? | 18:39 |
| sensefusion | well, there's no repo folder | 18:39 |
| T4 | <DSstill01> [Edit] What is your android base version? | 18:39 |
| sensefusion | I'm using lineage 15.1 | 18:39 |
| T4 | <DSstill01> Well, in 15.1 it's a prebuilt executable, so i dunno why it wants to compile some c file :/ | 18:40 |
| sensefusion | what is this executable like? Might it be this python file? | 18:54 |
| T4 | <adampigg> mal, media playback isnt working here, is droidmedia udate needed? | 18:57 |
| T4 | <DSstill01> @sensefusion [what is this executable like? Might it be this …], Yes, the one without extension | 18:57 |
| sensefusion | so, i should disable its building? | 18:58 |
| sensefusion | in rmp/dhd/helpers/build_packages.sh there're 7 options which of them is responsible for mkbootimg? | 19:01 |
| sensefusion | https://pastebin.com/yb8nL9gq | 19:02 |
| T4 | <DSstill01> You just have to check what wants to compile mkbootimg instead of using the already existing one | 19:04 |
| T4 | <DSstill01> I saw that mkbootimg.c is present in 13.0 and earlier | 19:05 |
| sensefusion | haven't found mkbootimg.c in my device tree | 19:06 |
| T4 | <DSstill01> Maybe it's in rpm/dhd | 19:07 |
| sensefusion | only a mkbootimg.mk in rpm/dhd/helpers | 19:08 |
| T4 | <DSstill01> Just search lol | 19:10 |
| T4 | <DSstill01> Somewhere it's present | 19:11 |
| sensefusion | did find hadk -name mkbootimg* | 19:11 |
| sensefusion | only these 2 files | 19:11 |
| T4 | <DSstill01> Use grep on suspicious paths | 19:11 |
| T4 | <DSstill01> It's in the makefile.mk in nthe dhd | 19:13 |
| T4 | <DSstill01> Now find out what calls it | 19:14 |
| sensefusion | the mkbootimg.mk file: https://pastebin.com/HjTZ9cpq | 19:17 |
| sensefusion | haven't found anything | 19:17 |
| mal | @adampigg should not be, it works for me | 19:19 |
| T4 | <DSstill01> Something must call it | 19:19 |
| mal | sensefusion: about how it should be handled https://github.com/mer-hybris/droid-hal-device/blob/master/droid-hal-device.inc#L353 | 19:23 |
| Nebrassy[m] | how do I do this | 19:23 |
| Nebrassy[m] | Determine which node is your touchscreen by checking /dev/input/event* | 19:23 |
| mal | sensefusion: wondering how it tries to build it for you | 19:23 |
| mal | Nebrassy[m]: the ugly way is to run cat /dev/input/event0 and same for other ones and then touch the screen and see which one outputs something matching your touch | 19:24 |
| mal | or install mce-tools package and use evdev_trace -i | 19:25 |
| Nebrassy[m] | that's what I assumed, but I'm not getting anything on any of them | 19:25 |
| Nebrassy[m] | on android that is | 19:25 |
| mal | hmm | 19:25 |
| Nebrassy[m] | I'll roll with event0 and see | 19:36 |
| sensefusion | <DSstill01>, Cheers for pointing this line out, after commenting the condition everything's fine | 19:38 |
| T4 | <adampigg> mal: other odd issue, some emails dont open | 19:38 |
| mal | @adampigg odd | 19:38 |
| T4 | <adampigg> just sits with a busy indicator | 19:38 |
| mal | check logs | 19:39 |
| T4 | <adampigg> yeah will do, im just fixing up amazfish | 19:39 |
| mal | @adampigg I still haven't figured out mobile data issue | 19:40 |
| T4 | <DSstill01> Sgaang | 19:46 |
| wdehoog | mal: droid-hal-init failed due to opening /dev/__properties__ using flags 'O_CREAT | O_EXCL' (bionic/libc/bionic/system_properties.cpp map_prop_area_rw()) | 19:57 |
| wdehoog | mal: this fails when the file already exists. do you know if something has changed in this area? | 19:57 |
| mal | wdehoog: what creates the file originally? | 20:05 |
| wdehoog | mal: I have no idea | 20:05 |
| mal | wdehoog: you have 12.1 base? | 20:08 |
| wdehoog | mal: yes 12.1 | 20:09 |
| mal | it uses O_RDWR | O_CREAT | O_NOFOLLOW | O_CLOEXEC | O_EXCL | 20:09 |
| mal | I see no reason why that would fail with existing file | 20:09 |
| mal | ah | 20:10 |
| mal | now I understand | 20:10 |
| mal | but no idea why that now started to be a problem | 20:14 |
| wdehoog | mal: the flags remain the same in the other branches so it is not that I should patch bionic | 20:15 |
| mal | yep, and it seems to work on other devices | 20:16 |
| mal | wdehoog: you also have 3.4 kernel on the device? | 20:16 |
| wdehoog | mal: yes 3.4.67 | 20:16 |
| mal | wdehoog: so if you mask droid-hal-init does the /dev/__properties__ appear before you run it | 20:20 |
| wdehoog | mal: I'll try | 20:21 |
| wdehoog | mal: when masking droid-hal-init the file does not exists | 20:25 |
| mal | wdehoog: and now if you try to run droid-hal-init manually does it work? | 20:26 |
| * Nebrassy[m] sent a long message: < https://matrix.org/_matrix/media/v1/download/matrix.org/dmuQFrvaPPvBqZBuJCvoyXbC > | 20:27 | |
| mal | wdehoog: just wondering, are you sure that failure you noticed didn't come from second attempt of droid-hal-init | 20:27 |
| Nebrassy[m] | followed by more errors | 20:27 |
| wdehoog | mal: I lost my original droid-hal-init, have to rebuild and test, probably some other time, thanks for helping | 20:28 |
| mal | wdehoog: how did you manage to lose that? | 20:28 |
| wdehoog | yeah rub it in... | 20:28 |
| mal | wdehoog: you don't have the droid-hal rpms on OBS or locl build system | 20:29 |
| mal | ? | 20:29 |
| wdehoog | I copy init from the out dir | 20:29 |
| wdehoog | scp from remote to home and the scp from home to phone | 20:29 |
| mal | so you don't have the original rpms on local repo or OBS? | 20:30 |
| wdehoog | I probably have but I don;t know which package contains which file, aha pkcon should know | 20:31 |
| mal | droid-hal-$DEVICE.rpm | 20:33 |
| wdehoog | mal: indeed it seems to work. have UI now, camera crashes though | 20:36 |
| wdehoog | mal: if I unmask droid-hal-init it segfaults at boot, when deleting /dev/__properties__ I can start it manually | 20:44 |
| mal | wdehoog: does log show that it only tried to run droid-hal-init once? | 20:50 |
| mal | and it segfaulted the first time also? | 20:50 |
| wdehoog | mal: logs show it fails to start and is restarted all the time | 20:51 |
| wdehoog | mal: If I start it with systemctl start droid-hal-init it seems to fail | 20:51 |
| mal | hmm | 20:53 |
| mal | very odd | 20:53 |
| mal | wdehoog: I also have some odd issues with 3.0.3, many shell script segfaults in journal | 20:53 |
| wdehoog | libc has changed didn;t it? so stuff might behave different but bionic stuff is not using it I assume | 20:56 |
| mal | yes, unless libhybris misbehaves | 20:57 |
| wdehoog | mal: I changed the Type in the service file from notify to simple. seems to help. camera and sensors now work | 21:16 |
| wdehoog | still using libhybris 0.0.5.25 | 21:17 |
| wdehoog | mal: so maybe systemd starts droid-hal-init and kills it since there is no notify, then I start it and it fails since __properties_ already exists | 21:19 |
| mal | wdehoog: hmm, droid-hal-init should notify the service, wondering why it's failing | 21:21 |
| wdehoog | mal: with libhybris 0.0.5.31 it also seems to work but the display is very very choppy | 21:21 |
| wdehoog | mal: how does that notify work? any idea where to look? | 21:22 |
| mal | wdehoog: actually the notify goes to https://github.com/mer-hybris/droid-hal-configs/blob/master/sparse/usr/bin/droid/droid-init-done.sh | 21:22 |
| mal | wdehoog: btw, the first attempt of droid-hal-init to start, does it fail at start or does it try to go further? | 21:23 |
| mal | wdehoog: so the android init notifies when it's done, and that droid-init-done tells android init is done | 21:24 |
| mal | so other systemd services which depend on android side know that they can now run | 21:24 |
| mal | wdehoog: just an idea, in the past there were issues droid-hal-init sending HUP signal which caused failure, but we fixed it with exec nohup, wondering if that still works always https://github.com/mer-hybris/droid-hal-configs/blob/master/sparse/usr/bin/droid/droid-hal-startup.sh#L19 | 21:28 |
| wdehoog | mal: I don't understand, who/what calls droid-init-done.sh? | 21:29 |
| mal | wdehoog: droid-init-done is a separate systemd service, it listens to certain type of notifications from android init (i.e. droid-hal-init) | 21:31 |
| wdehoog | mal: aha. thanks. | 21:32 |
| wdehoog | mal: so if it is not running there will be no notify | 21:32 |
| mal | wdehoog: or actually not sure exactly if it was like that | 21:33 |
| mal | but anyway android init has this https://github.com/mer-hybris/android_system_core/commit/6a22dcb18094a76fc3e1482458b398c7880637f2#diff-a6b80436b1afea26b3590052262b2a9b | 21:34 |
| mal | wdehoog: looks like the droid-init-done only monitors the droid-hal-init systemd service | 21:35 |
| wdehoog | mal: I can't find anything calling this script on my phone | 21:36 |
| mal | wdehoog: this is the way the informing happen https://github.com/mer-hybris/android_system_core/blob/hybris-12.1/rootdir/init.rc#L667 | 21:36 |
| mal | wdehoog: sorry for giving incorrect information, I forgot how that works | 21:36 |
| wdehoog | mal: now I see it | 21:37 |
| mal | wdehoog: did you try making the droid-hal-init service back to notify and then removing the exec nohup (i.e. leave only /sbin/droid-hal-init to last line of /usr/bin/droid/droid-hal-startup.sh) | 21:40 |
| wdehoog | mal: I'll try, still having difficulties to understand how it is supposed to work | 21:40 |
| wdehoog | mal: reverting to notify and removing exec nohup gives a UI but no sensors, camera etc. | 21:44 |
| mal | wdehoog: how is that possible | 21:49 |
| wdehoog | no services | 21:49 |
| mal | so droid-hal-init doesn't really run properly? | 21:50 |
| wdehoog | it is killed I think, timeout | 21:50 |
| wdehoog | so no mini thingies etc | 21:50 |
| mal | why doesn't it get far enough, show dmesg | 21:51 |
| wdehoog | mce sets the clock, could that cause systemd to timeout ? | 21:52 |
| mal | no idea | 21:53 |
| wdehoog | dmesg log : http://susepaste.org/view//9504676 | 21:53 |
| wdehoog | journal http://susepaste.org/view//59651193 droid-hal-init starts at 78 and is killed at 134, mce sets the clock at 98 | 21:55 |
| wdehoog | mal: ^ | 21:56 |
| wdehoog | mal: I am off to sleep, thanks a lot for helping and explaining | 21:56 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!