| *** juiceme_ is now known as juice | 06:56 | |
| *** juice is now known as juiceme | 06:56 | |
| krnlyng | piggz, hybris based? | 07:07 |
|---|---|---|
| T4 | <adampigg> krnlyng, nope :) | 07:08 |
| T4 | <adampigg> arm with mali graphics, using stock kernel and lima driver + mesa (at least, that is the plan) | 07:09 |
| rinigus | kimmoli: any progress with the testing? for me, all works well. | 07:10 |
| T4 | <Harsh18262> mal:output from the above command https://paste.ubuntu.com/p/ZRxV822vmn/ | 08:19 |
| kimmoli | rinigus: charging... | 09:28 |
| rinigus | kimmoli: ... charged? | 15:05 |
| kimmoli | charged and updated ota. not tested. | 15:24 |
| piggz | hi, could someone cast an eye of this lipstick crash.... https://paste.gnome.org/pw1rxul7n | 16:07 |
| piggz | line 1515 is the only thing that gets printed out | 16:08 |
| piggz | maybe fixed.... | 16:10 |
| piggz | ok, jolla-startup-wizard is running, but no display.... | 16:17 |
| piggz | just the kernel output is on the screeen | 16:17 |
| mal | piggz: check https://freenode.irclog.whitequark.org/panfrost/2019-03-15 something about --device | 16:18 |
| piggz | mal: yes, i fixed the crash by using card1, not card0 | 16:19 |
| piggz | mal: any tips for running sailfish on a stock kernel? | 17:33 |
| piggz | mce is failing with noi aparent output | 17:33 |
| piggz | strace shows | 17:33 |
| piggz | https://paste.gnome.org/pp924qqhc | 17:34 |
| piggz | also, what packages contain the test_egl etc? | 17:38 |
| mal | piggz: test_* come from libhybris | 17:42 |
| piggz | mal: libhybris is no good to me! | 17:43 |
| piggz | (and i thought that was the case) | 17:43 |
| piggz | mal: whats the other test for wayland? | 17:44 |
| mal | piggz: maybe try building this https://github.com/linux-sunxi/sunxi-mali/blob/master/test/test.c | 17:45 |
| mal | piggz: is journal working, anything from mce? | 17:48 |
| piggz | mal: yes, journale is working... | 18:00 |
| piggz | mal: | 18:00 |
| piggz | May 06 18:38:27 Sailfish mce[2888]: mce.c: main(): MCE 1.100.2 (release) starting up | 18:00 |
| piggz | May 06 18:38:27 Sailfish mce[2888]: event-input.c: evin_iomon_init(): opendir() failed; No such file or directory | 18:00 |
| piggz | May 06 18:38:27 Sailfish systemd[1]: mce.service: Main process exited, code=exited, status=1/FAILURE | 18:00 |
| piggz | May 06 18:38:27 Sailfish systemd[1]: mce.service: Unit entered failed state. | 18:01 |
| piggz | mal: also of note is that lime is a 'render only' adapter, and uses sun4i to do the actual display ... im not sure how to configuer this | 18:13 |
| T4 | <DSstill01> @piggz [mal: any tips for running sailfish on a stock …], Imposs | 18:32 |
| piggz | why imposs? | 18:33 |
| piggz | thats a rather defeatist view :D | 18:33 |
| T4 | <DSstill01> You mean prebuilt or from sources? | 18:33 |
| mal | @DSstill01 I have sailfish partially running on mainline kernel on a qcom device | 18:34 |
| T4 | <DSstill01> Me too | 18:35 |
| T4 | <DSstill01> No | 18:35 |
| T4 | <DSstill01> Lol | 18:35 |
| T4 | <DSstill01> Old af | 18:35 |
| mal | piggz: so looks like mce doesn't like it there are no input devices | 18:35 |
| mal | *if | 18:35 |
| T4 | <DSstill01> Im pretty sure that if i use mainline kernel ill have problems with blobs | 18:36 |
| piggz | sure u will, thats why im using an OS friendly dev board | 18:36 |
| mal | @DSstill01 I'm trying to avoid blobs and do a native port, so no android side at all running | 18:36 |
| T4 | <DSstill01> Imposs | 18:37 |
| mal | only the driver firmwares would be used | 18:37 |
| mal | no it's not, I have modem up also | 18:37 |
| T4 | <DSstill01> Fp2? | 18:37 |
| mal | yes | 18:37 |
| T4 | <DSstill01> Aww | 18:37 |
| T4 | <DSstill01> What about camera? | 18:38 |
| T4 | <DSstill01> Thats impossible | 18:38 |
| piggz | geez, nothing is impossible | 18:38 |
| mal | I haven't done that yet, but the drivers are available for some of the modules and only needs adding support for the platform to camera controller | 18:39 |
| T4 | <DSstill01> Proprietary cameras are impossible to fix without blobs | 18:39 |
| mal | there is some old patch for that already | 18:39 |
| mal | nope | 18:39 |
| mal | on fp2 I have wlan, bt, vibrator, led, display panel+backlight, touchscreen, hw buttons, sd card etc working in mainline | 18:40 |
| T4 | <DSstill01> How can they work without android stuff? | 18:40 |
| mal | of course they can | 18:40 |
| T4 | <DSstill01> How? | 18:40 |
| mal | kernel drivers | 18:41 |
| T4 | <adampigg> likely the same way android stuff works...with drivers talking to hw.... | 18:41 |
| T4 | <DSstill01> My device has kernel drivers ofc but wont work | 18:41 |
| T4 | <adampigg> you just need different, open-source drivers | 18:41 |
| mal | modem needed some userspace daemon also but that was provided by sony | 18:41 |
| T4 | <DSstill01> Hmm, i dont know, ofc they need userspace stuff but i didnt know it could be added in sailfish | 18:42 |
| mal | it can | 18:43 |
| T4 | <DSstill01> I mean, theyre different packages that replace libhybris | 18:43 |
| T4 | <DSstill01> And all its system | 18:43 |
| mal | here is an example of nexus 5 running mainline https://www.youtube.com/watch?v=u9zw_ewkBJ0 | 18:43 |
| mal | @DSstill01 for graphics we just would use mesa for example, like piggz did in latte | 18:44 |
| mal | other stuff of course needs also some different middleware | 18:45 |
| T4 | <DSstill01> Hmm, and this id the future of sailfish? | 18:45 |
| mal | probably not for a long time, just for some devices | 18:46 |
| T4 | <DSstill01> I dont think this can be adopted for all devices, especially for the ones whose vendor is not dev friendly | 18:46 |
| mal | it takes quite a while to get things running, at least on arm devices | 18:46 |
| T4 | <DSstill01> Anyway gg if you get everything to work | 18:48 |
| mal | sensors will be painful | 18:48 |
| mal | need to reverse engineer the protocol | 18:48 |
| T4 | <DSstill01> Heh,nfc? | 18:48 |
| mal | no nfc on fp2 | 18:49 |
| T4 | <DSstill01> Ah, in case it had nfc i think that would be difficul to fix as well | 18:49 |
| mal | that depends on how nfc is added to the device | 18:50 |
| T4 | <DSstill01> And fingerprint | 18:50 |
| mal | if it's i2c then it's easier | 18:50 |
| T4 | <DSstill01> @mal [that depends on how nfc is added to the device], Proprietary stuff | 18:50 |
| mal | even if something is proprietary in android doesn't mean the chip really is | 18:51 |
| T4 | <DSstill01> If vendor makes it... | 18:51 |
| T4 | <DSstill01> Probably it will be | 18:51 |
| piggz | mal: got mce to start by creating /dev/input for now.... | 18:56 |
| piggz | user dbus session doesnt exist now | 18:56 |
| piggz | though, systemctl says it is... | 18:57 |
| piggz | mal: https://paste.gnome.org/pwty3grqi | 18:59 |
| piggz | and this is the status of user@100000 https://paste.gnome.org/pbjukuefb | 19:01 |
| mal | piggz: shouldn't XDG_RUNTIME_DIR be /run/user/100000 | 19:11 |
| mal | that's what it is on my device | 19:11 |
| piggz | mal: yes, but that was only there when hacing earlier and changing it doesnt help | 19:12 |
| mal | piggz: what does that folder contain? | 19:14 |
| piggz | /dev/input? nothing ... but it allows mce to start | 19:15 |
| mal | piggz: /run/user/100000 | 19:15 |
| piggz | oh...its empty | 19:15 |
| piggz | infact, doesnt exist | 19:15 |
| mal | how is that possible if user@100000 is running | 19:15 |
| piggz | sorry, wrong dir.... | 19:16 |
| piggz | [root@Sailfish ~]# ls -lh /run/user/100000/ | 19:16 |
| piggz | total 0 | 19:16 |
| piggz | drwxr-xr-x 2 nemo nemo 60 May 6 20:28 dbus | 19:16 |
| piggz | drwx------ 2 nemo privileged 60 May 6 20:29 dconf | 19:16 |
| piggz | drwx------ 2 nemo privileged 100 May 6 20:28 mapplauncherd | 19:16 |
| piggz | drwx------ 2 nemo nemo 80 May 6 20:28 pulse | 19:16 |
| piggz | drwxrwxrwx 3 nemo privileged 60 May 6 20:28 single-instance-locks | 19:16 |
| piggz | dr-xr-xr-x 4 nemo nemo 0 May 6 20:28 state | 19:16 |
| piggz | drwxr-xr-x 2 nemo nemo 80 May 6 20:28 systemd | 19:16 |
| mal | so why doesn't lipstick find the dbus from that, are you sure XDG was correct in latest attempts | 19:17 |
| piggz | mal: maybe it was a red herring...now doenst log that message, but segfaults non the less | 19:23 |
| mal | piggz: show new log | 19:24 |
| piggz | im getting a difference between journal log and output when run manually.... | 19:25 |
| piggz | i'll get both | 19:25 |
| mal | piggz: is the compositor config same in both cases? | 19:25 |
| piggz | mal: journal https://paste.gnome.org/pqrydgf9n | 19:26 |
| mal | so dbus is not up, check the path manually? | 19:27 |
| piggz | env https://paste.gnome.org/poqtful6v | 19:27 |
| mal | check /run/user/100000/pulse/dbus-socket | 19:27 |
| mal | also please show more of the log, difficult so see if there are some other errors | 19:28 |
| piggz | [nemo@Sailfish ~]$ ls -lh /var/run/user/100000/dbus/user_bus_socket | 19:29 |
| piggz | srw-rw-rw- 1 nemo nemo 0 May 6 20:28 /var/run/user/100000/dbus/user_bus_socket | 19:29 |
| mal | hmm, why the different name | 19:30 |
| mal | ah, sorry, not a problem | 19:30 |
| mal | misread the path | 19:30 |
| mal | piggz: does gdb say anything? | 19:31 |
| piggz | mal: https://paste.gnome.org/pjsbbo4i1 | 19:31 |
| piggz | (gdb) bt | 19:34 |
| piggz | #0 0xf6452a46 in ?? () from /lib/libc.so.6 | 19:34 |
| piggz | #1 0xf6463998 in raise () from /lib/libc.so.6 | 19:34 |
| piggz | #2 0xf6464fc0 in abort () from /lib/libc.so.6 | 19:34 |
| piggz | #3 0xf674cd4c in QMessageLogger::fatal(char const*, ...) const () | 19:34 |
| piggz | from /usr/lib/libQt5Core.so.5 | 19:34 |
| piggz | #4 0x00049fae in ?? () | 19:34 |
| piggz | #5 0x00049fae in ?? () | 19:34 |
| piggz | Backtrace stopped: previous frame identical to this frame (corrupt stack?) | 19:34 |
| piggz | mal: could EGL_PLATFORM be something else? | 19:37 |
| mal | "sun4i-drm display-engine: No panel or bridge found... RGB output disabled" | 19:37 |
| mal | hmm | 19:37 |
| piggz | mal: i have a frambufer showing kernel output on the panel | 19:42 |
| mal | piggz: but that error is from drm | 19:42 |
| mal | piggz: so fb might be ok but drm is not properly configured maybe | 19:42 |
| mal | or panel is misconfigured | 19:44 |
| T4 | <adampigg> Mal | 19:56 |
| T4 | <adampigg> (Photo, 1920x2560) https://af42kgggd3mfqvgy.tor2web.io/B2ppkWs09w.png | 19:56 |
| T4 | <adampigg> (Photo, 1920x2560) https://af42kgggd3mfqvgy.tor2web.io/Qqml0XPLek.png | 19:56 |
| piggz | this is kmscube test app ... lima devs say if this works, problem is in Qt | 19:58 |
| piggz | :/ | 19:58 |
| piggz | https://paste.gnome.org/pzbgxbdpq | 19:58 |
| mal | nice | 20:02 |
| piggz | mal: except i need to debug lipstick ! | 20:02 |
| piggz | though, it means panel is working ok | 20:03 |
| mal | so what is qt doing wrong, it would be nice if you install more debuginfo packages | 20:03 |
| mal | yes | 20:03 |
| mal | could be EGL env var issue | 20:03 |
| mal | piggz: is it eglfs_kms you want to use? | 20:04 |
| piggz | i _think_ so ??? | 20:04 |
| piggz | this is what kmscube does https://gitlab.freedesktop.org/mesa/kmscube | 20:04 |
| mal | can you try to enable QT_QPA_EGLFS_DEBUG env var and see if you get some debug | 20:06 |
| mal | piggz: pastebin that /etc/eglfs-config.json | 20:07 |
| piggz | mal: | 20:08 |
| piggz | [nemo@Sailfish ~]$ cat /etc/eglfs-config.json | 20:08 |
| piggz | { | 20:08 |
| piggz | "device" : "/dev/dri/card1" | 20:08 |
| piggz | } | 20:08 |
| mal | is that enough, no need to define output? | 20:08 |
| mal | like in the example here https://doc.qt.io/qt-5/embedded-linux.html | 20:08 |
| piggz | just seen that...i wonder what name to use | 20:09 |
| mal | piggz: but latte didn't need that | 20:16 |
| mal | piggz: is the display size correct, although shouldn't matter much probably | 20:19 |
| mal | piggz: maybe try to give it also display resolution with QT_QPA_EGLFS_WIDTH and height | 20:22 |
| piggz | mal: this board does have a hdmi out also | 20:22 |
| piggz | mal: thoughts on all the dbus errors? | 20:31 |
| mal | which ones? | 20:35 |
| qiangong2 | When using build_packages.sh, I'm getting this issue: https://paste.ubuntu.com/p/TjrNBNggjF/ | 20:37 |
| piggz | the ones during lipstick start | 20:37 |
| qiangong2 | The only problem is, I've actually installed those packages already | 20:37 |
| piggz | mal: think i'll quit for the night ... if you have thoughts, @adampigg me! | 20:38 |
| qiangong2 | Any ideas why that might be happening? | 20:38 |
| qiangong2 | It's blocking me from building | 20:38 |
| mal | qiangong2: installed how? | 20:39 |
| qiangong2 | through zypper | 20:39 |
| qiangong2 | in the platform sdk | 20:39 |
| mal | you do understand difference of sdk and sdk target? | 20:39 |
| qiangong2 | Do I have to checkout it differently? | 20:40 |
| qiangong2 | I've been following the hadk | 20:40 |
| mal | what does this tell: sb2 -t $VENDOR-$DEVICE-$PORT_ARCH -R -msdk-install zypper se qt5-qttools-kmap2qmap | 20:41 |
| mal | same for rsync as last parameter | 20:41 |
| mal | does it find those packages | 20:42 |
| qiangong2 | no matching items found for the qttools, but it does find sync | 20:42 |
| qiangong2 | *rsync | 20:42 |
| mal | run this and try the search command again sb2 -t $VENDOR-$DEVICE-$PORT_ARCH -R -msdk-install zypper ref | 20:44 |
| qiangong2 | okay | 20:44 |
| mal | that package should be in jolla repo which should be seen in output of sb2 -t $VENDOR-$DEVICE-$PORT_ARCH -R -msdk-install ssu lr | 20:45 |
| qiangong2 | https://paste.ubuntu.com/p/3f69QSj5MF/ | 20:46 |
| qiangong2 | Can't find them at all it seems | 20:46 |
| qiangong2 | same issue with adaptation0 too | 20:46 |
| qiangong2 | invalid metadata | 20:46 |
| mal | where does it get that adaptation0 | 20:47 |
| mal | that is not in target | 20:47 |
| qiangong2 | when you run rpm/dhd/helpers/build_packages.sh | 20:47 |
| mal | did you update the target and sdk or are both recently installed | 20:47 |
| qiangong2 | Fresh install | 20:48 |
| mal | does it find any other qt packages, change the search command to search for qt5 for example | 20:48 |
| qiangong2 | Yes, a whole lot of them | 20:49 |
| qiangong2 | both packages and srcpackages | 20:49 |
| mal | but not the tools package | 20:55 |
| mal | or anything similar | 20:55 |
| qiangong2 | Not the kmap2qmap, but qt5-qttools is | 20:55 |
| mal | hmm, how is that possible | 20:59 |
| qiangong2[m] | I don't know. It didn't happen when I was building 3.0.0.8 | 20:59 |
| mal | I just did a fresh install of 3.0.3.9 target and it had no issues finding that package | 21:01 |
| mal | qiangong2[m]: maybe use sdk-assistant to remove the target and install again | 21:02 |
| qiangong2[m] | Hmm, I don't know then. This is the second time I've tried a fresh install. Even removing the lineage builds and starting completely from scratch | 21:02 |
| qiangong2[m] | Okay, I'll try again | 21:02 |
| mal | android side build has nothing to do with this issue, do not remove those | 21:03 |
| mal | qiangong2[m]: is the tooling installed correctly and correct version | 21:03 |
| qiangong2[m] | I believe so, like the hadk said, I changed the URL and the name from latest to 3.0.3.9 | 21:04 |
| mal | ok | 21:05 |
| mal | very odd issue then | 21:05 |
| qiangong2[m] | Here is the output from the target creation: https://paste.ubuntu.com/p/43xbVRH3hh/ | 21:05 |
| qiangong2[m] | is the no-host-gcc an issue? | 21:06 |
| qiangong2[m] | same build dependency issue as before btw | 21:07 |
| mal | same output that I have | 21:11 |
| mal | I assume it first refreshed the repos, I only did this sb2 -t $VENDOR-$DEVICE-$PORT_ARCH -R -msdk-install zypper in qt5-qttools-kmap2qmap | 21:12 |
| mal | to test it | 21:12 |
| qiangong2[m] | When I run that, I get valid metadata not found | 21:12 |
| qiangong2[m] | "Please check if the URIs defined for this repository are pointing to a valid repository." | 21:13 |
| mal | where is that coming from, what is it trying to refresh | 21:15 |
| mal | it makes no sense | 21:15 |
| mal | which repo gives the error? | 21:15 |
| qiangong2[m] | All of them | 21:15 |
| qiangong2[m] | https://paste.ubuntu.com/p/F4NzB3thdj/ | 21:16 |
| qiangong2[m] | and yes, I am connected to the internet | 21:16 |
| mal | show output of sb2 -t $VENDOR-$DEVICE-$PORT_ARCH -R -msdk-install ssu lr | 21:16 |
| qiangong2[m] | https://paste.ubuntu.com/p/46R732hyQc/ | 21:17 |
| mal | I have same urls and no issues | 21:18 |
| mal | are you sure your internet doesn't block something? | 21:18 |
| qiangong2[m] | Nothing is blocked on my end | 21:19 |
| qiangong2[m] | It worked fine four months ago when I last built | 21:19 |
| mal | so this also gives those errors? sb2 -t $VENDOR-$DEVICE-$PORT_ARCH -R -msdk-install zypper ref | 21:20 |
| qiangong2[m] | Yes, the same repo issue | 21:21 |
| mal | then there must be something wrong with network connection or somewhere between your computer and the server the data is lost | 21:23 |
| qiangong2[m] | well, now I get this when running build_packages.sh: https://paste.ubuntu.com/p/rPTS4W8vdt/ | 21:28 |
| qiangong2[m] | back to the issue with zypp-cache that I was having before | 21:29 |
| mal | are write permission to /home ok? | 21:49 |
| qiangong2[m] | I can't edit /home/.zypp-cache without root | 22:01 |
| mal | so sudo doesn't work inside sdk? | 22:05 |
| mal | or actually is it doing it inside target | 22:06 |
| mal | I just tried to run sudo mkdir /home/foo and that worked, it shows also in target | 22:06 |
| mal | so not sure what is so different in your setup | 22:07 |
| qiangong2[m] | No, that works fine. I can't edit anything as a standard user in /home/.zypp-cache. using sudo works | 22:19 |
| qiangong2[m] | I'm on Ubuntu 16.04 if that makes any difference | 22:37 |
| mal | I don't quite understand the issue | 23:02 |
| qiangong2[m] | It says error building the cache: can't create /home/.zypp-cache | 23:04 |
| qiangong2[m] | Is that a false positive? | 23:05 |
| mal | it's probably some issue but no idea why it happens | 23:19 |
| qiangong2[m] | Oh | 23:22 |
| qiangong2[m] | Either way, I still can't build | 23:22 |
| mal | need to investigate more some other time, it's already very late | 23:32 |
| qiangong2[m] | Okay | 23:38 |
| qiangong2[m] | thanks for trying | 23:39 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!
