**** BEGIN LOGGING AT Fri Mar 22 02:59:57 2019 Mar 22 07:11:38 Morning! Mar 22 07:13:59 JaMa: I think the "[ 148.723205] init: couldn't write 1224 to /dev/cpuset/system-background/tasks: No space left on device" is considered "normal" according to bshah Mar 22 07:14:17 That worried me as well with some of the ports, but things seem to work regardless and bshah said it was normal ;) Mar 22 07:14:22 bshah: ^ correct? Mar 22 07:28:40 Morning ! Mar 22 07:29:04 Herrie|Laptop, JaMa: yes, this error has been there for a long time Mar 22 07:30:20 And that's Android specific, I think. It could be it's a special proc fs or module which isn't mounted in our case Mar 22 09:04:04 JaMa: I think the "[ 148.723205] init: couldn't write 1224 to /dev/cpuset/system-background/tasks: No space left on device" is considered "normal" according to bshah ; even i had this error a while ago on cancro Mar 22 12:19:14 morning Mar 22 12:20:39 ok, then there are those syscall notes and the strange delay when failing to load firmware Mar 22 12:25:30 I've restarted the PR server on kaylee, because last testing job seems to be stuck contacting it and previous one failed to get PRAUTO from it Mar 22 14:42:30 looks like we don't have any nyx-modules installed in qemux86 images, causes nyx-utils.service and sleepd.service to fail starting (maybe more) Mar 22 14:42:33 2019-03-22T14:41:34.534494Z [652] user.err sleepd [] NYX_OPEN_ERR {} module (nyxSystemMain.module) does not exist in /usr/lib/nyx/modules nor in /usr/lib/nyx/modules.mock Mar 22 14:42:37 2019-03-22T14:41:34.534677Z [652] user.crit sleepd [] sleepd NYX_DEVICE_OPEN_FAIL {"CAUSE":"Unable to open the nyx device system"} Mar 22 14:42:50 I guess it's since the OSE migration Mar 22 14:46:00 nevermind my fault Mar 22 14:46:02 packagegroup: temoporary drop certmgrd, pmcertificatemgr, mojomail-imap, mojomail-pop, mojomail-smtp, nyx-modules Mar 22 14:46:12 it's because of openssl upgrade Mar 22 14:46:41 will need to resolve this before upgrading to thud Mar 22 15:13:17 JaMa: FYI AMD RAID drivers are quite a royal pain in the butt ;) Seems that with every new minor Linux kernel upgrade (4.20, 5.0) stuff breaks ;) But patched it for 5.0 now, so we should be good for a while again hopefully. https://github.com/thopiekar/rcraid-dkms/commit/8167f72bdd03fc282c1148af2f1d37817d9378d8 Mar 22 15:15:15 Tofe: this is the mouse/nyx integration used in webOS OSE I've mentioned yesterday: https://github.com/webosose/meta-webosose/blob/master/meta-webos/recipes-qt/qt5/qtbase/0098-Emulator-NYX-integration-with-eglfs-plugin.patch maybe that's better option than vboxtouch if we want to support both QEmu and VirtualBox Mar 22 15:15:33 Herrie|Laptop: without raid the controller works fine? Mar 22 15:16:35 JaMa: Yes AHCI is no issue ;) Mar 22 15:16:58 hmm Append modprobe.blacklist=ahci to GRUB_CMDLINE_LINUX_DEFAULT in /etc/default/grub Mar 22 15:17:24 Just I have some RAID1 that I need to have working ;) Mar 22 15:17:33 so you cannot use rcraid for mb controller and AHCI from PCI-E controller (if you need one) Mar 22 15:17:55 did you compare the performance with mdraid? Mar 22 15:17:55 JaMa: Well my SSD is not in RAID and it works fine ;) Mar 22 15:18:44 on my old MB, mdraid was as fast as integrated raid on mb, but mdraid was easier to manage Mar 22 15:20:01 JaMa: I had existing array from Windows that I didn't want to bother to change Mar 22 15:21:19 Seeing there was 4TB of data on it ;) Mar 22 15:22:23 ok, understood Mar 22 15:22:51 after moving around 10TB last month I understand :) Mar 22 15:25:29 JaMa: that's a jolly big patch Mar 22 15:49:33 JaMa: I'm not so sure it would help, if the kernel vboxguest driver doesn't expose a standard evdev device Mar 22 15:50:52 the nyx mouse driver behaves quite like the evdevtouch qt driver Mar 22 18:06:24 Tofe: there is event device from vbox (without vboxtouch plugin) Mar 22 18:06:29 I: Bus=0003 Vendor=80ee Product=0021 Version=0110 Mar 22 18:06:31 N: Name="VirtualBox USB Tablet" Mar 22 18:06:34 P: Phys=usb-0000:00:1f.4-1/input0 Mar 22 18:06:36 S: Sysfs=/devices/pci0000:00/0000:00:1f.4/usb1/1-1/1-1:1.0/0003:80EE:0021.0001/input/input5 Mar 22 18:06:39 U: Uniq= Mar 22 18:06:42 H: Handlers=mouse1 event4 Mar 22 18:06:47 and it even looks working fine in evtest Mar 22 18:08:28 and that's what vboxtouch plugin is reading Mar 22 18:08:30 root@qemux86:~# grep VIRT /etc/luna-next/environment.conf Mar 22 18:08:30 VIRTUALBOX_TOUCH_EVDEV_MOUSE=/dev/input/event4 Mar 22 18:09:59 I'm playing with an old virtualbox image, and event4 doesn't output anything with cat Mar 22 18:10:35 (with luna-next stopped) Mar 22 18:12:52 or, should I start luna-next first ? Mar 22 18:14:21 I have newer image but without luna-next running and see it with evtest Mar 22 18:14:42 and with cat as well Mar 22 18:14:59 ok Mar 22 18:16:55 please check vbox machine config, if you have system - motherboard - pointing device - USB Tablet Mar 22 18:17:50 be aware that we also have QT_QPA_EGLFS_HIDECURSOR=1 in environment.conf Mar 22 18:18:14 and QT_QPA_EGLFS_DISABLE_INPUT=1 Mar 22 18:18:46 I have an usb tablet yes Mar 22 18:19:53 I see event4 outputs with evtest when using the VBox driver with luna-next Mar 22 18:21:21 maybe vboxtouch initialized something for me before it died with luna-next Mar 22 18:22:04 I've disabled luna-next and trying to reboot vbox Mar 22 18:22:08 I stopped luna-next, and hop, no more events Mar 22 18:22:34 I start luna-next again with VBoxPlugin, it re-works Mar 22 18:22:40 strange Mar 22 18:22:52 as if VBoxPlugin activates some event forwarding Mar 22 18:22:55 I've checked with ps that luna-next wasn't running when you mentioned it before Mar 22 18:23:47 if event4 works well for you, can you just try luna-next -plugin evdevtouch:/dev/input/event4 ? Mar 22 18:24:05 also with evdevmouse, if the former doesn't work Mar 22 18:24:29 I will but in this image I don't have DRM working Mar 22 18:24:42 oh, it's newer Mesa Mar 22 18:24:57 it's yocto master and something went wrong Mar 22 18:25:02 :) ok Mar 22 18:25:05 vbox is complaining about lack of ram Mar 22 18:25:43 JaMa: do we have a built image I could try, where DRM works and event4 is probably working ? Mar 22 18:26:13 and the previous build was suffering from SIGILL inside llvm https://paste.ubuntu.com/p/PDnnJGvFw8/ Mar 22 18:26:38 let me check with last testing-pyro Mar 22 18:26:52 that should have DRM working Mar 22 18:27:17 ok, I'll start a download Mar 22 18:27:48 I wonder, isn't "There is no EGL_WL_bind_wayland_display extension" supposed to be fatal ? Mar 22 18:29:55 where do you see it? when starting luna-next? Mar 22 18:30:52 I don't see the events from event4 after rebooting the vbox Mar 22 18:33:53 no, in your master log you just pasted Mar 22 18:34:17 ah sorry, misunderstood Mar 22 18:34:21 let me re-answer :) Mar 22 18:35:02 What I see: I start luna-next with VBoxPlugin plugin, and then I do in another shell "evtest /dev/input/event4" Mar 22 18:35:09 I'm testing luneos-dev-emulator-qemux86-20190307135517-testing-0-559 Mar 22 18:35:39 I'm still fetching ETA 2min Mar 22 18:36:35 hmm probably is, in the 2nd log where vboxtouch was crashing (but gui was working after removing vboxtouch plugin or fixing it) it isn't https://paste.ubuntu.com/p/gJktWTFKfC/ Mar 22 18:41:44 gui works, events are being received, let me try to change luna-next config Mar 22 18:42:22 events are still flowing after systemctl stop luna-next.service Mar 22 18:42:46 meh, I don't have the same results ! which options are we talking about ? Mar 22 18:43:32 the default from this image Mar 22 18:43:45 weird Mar 22 18:43:53 I haven't changed anything, even imported it as appliance to have the default VM config Mar 22 18:43:59 now I've commented out: Mar 22 18:43:59 #QT_QPA_EGLFS_HIDECURSOR=1 Mar 22 18:44:00 #QT_QPA_EGLFS_DISABLE_INPUT=1 Mar 22 18:44:07 I don't even have the mouse pointer in VirtualBox Mar 22 18:44:14 and see the cursor, but no response on it Mar 22 18:44:24 you won't see the cursor without commenting out above ^ Mar 22 18:44:31 ok Mar 22 18:44:45 because vboxtouch isn't really running (it shutdowns immediately because of those ioctls failing) Mar 22 18:45:23 ah yes of course Mar 22 18:46:06 root@qemux86:~# journalctl -u luna-next.service | grep EGL Mar 22 18:46:06 Mar 22 19:40:37 qemux86 luna-next[587]: WARNING: 19:40:37.301: QtCompositor: Failed to initialize EGL display. There is no EGL_WL_bind_wayland_display extension. Mar 22 18:46:17 I see the Ui, so it's not fatal Mar 22 18:47:50 ok, well, good to know Mar 22 18:47:57 I see no change when using -plugin evdevtouch:/dev/input/event4 or evdevmouse Mar 22 18:49:00 same Mar 22 18:49:08 let me check if qtbase is built with libinput support Mar 22 18:50:13 well we already use evdevkeyboard isn't it Mar 22 18:51:53 yes we do Mar 22 18:52:01 ah I see what you mean :) Mar 22 18:54:38 hmm there might be some events, but inverted or otherwise broken, I've just opened launcher somehow Mar 22 18:54:51 "Found matching devices ("/dev/input/event5")" Mar 22 18:54:59 but that's all we get... Mar 22 18:55:27 FWIW: someone also reported to me recently that input events stopped working for him since last qt upgrade Mar 22 18:55:32 let me check which upgrade it was Mar 22 18:56:48 it was from 5.11 but he was reporting that qt itself gets events, but web page shown in simplebrowser doesn't get the events since 5.11 Mar 22 18:57:18 on iMX6 with etnaviv Mar 22 18:57:22 ah, so maybe a webengine issue or something Mar 22 19:01:09 hmm Mar 22 19:00:51 qemux86 daemon.info LunaAppManager[585]: Failed to move cursor on screen VGA1: -13 Mar 22 19:01:36 why is that one involved ? :) Mar 22 19:02:09 but I have it too, now that I look in the journal Mar 22 19:03:50 something like this origin: https://code.woboq.org/qt5/qtbase/src/plugins/platforms/eglfs/deviceintegration/eglfs_kms/qeglfskmsgbmcursor.cpp.html#243 Mar 22 19:03:53 I was just checking what's in the messages and this one spams quite often and luna-webappmanager is failed Mar 22 19:04:38 there are more drm related errors Mar 22 19:04:52 Mar 22 19:40:40 qemux86 LunaAppManager[585]: Running on a software rasterizer (LLVMpipe), expect limited performance. Mar 22 19:04:55 Mar 22 19:40:40 qemux86 LunaAppManager[585]: Could not set DRM mode for screen VGA1 (Permission denied) Mar 22 19:04:58 Mar 22 19:40:40 qemux86 LunaAppManager[585]: Could not queue DRM page flip on screen VGA1 (Permission denied) Mar 22 19:05:01 Mar 22 19:40:40 qemux86 LunaAppManager[585]: Failed to move cursor on screen VGA1: -13 Mar 22 19:05:04 Mar 22 19:40:40 qemux86 LunaAppManager[585]: Could not set cursor on screen VGA1: -13 Mar 22 19:05:07 Mar 22 19:40:40 qemux86 LunaAppManager[585]: Failed to move cursor on screen VGA1: -13 Mar 22 19:06:08 it would be nice to go back to something simpler like luneos-example-image, to get it working first on some simple qtbase examples Mar 22 19:07:00 root@qemux86:/usr/share/examples/webengine# minimal/minimal --no-sandbox Mar 22 19:07:04 shows similar errors Mar 22 19:07:27 ok Mar 22 19:07:52 well the mouse cursor is probably just the graphical pointer itself Mar 22 19:08:52 "Could not queue DRM page flip on screen VGA1" should be resolved by the newer vboxvideo (or at least I thought it was before) Mar 22 19:09:34 and this image already has 4.19.0-yocto-standard so it should be vboxvideo from there Mar 22 19:09:58 kernel-module-vboxguest-4.19.0-yocto-standard - 4.19+git0+57b791cb9f_122d468967-r0.24 Mar 22 19:10:01 kernel-module-vboxvideo-4.19.0-yocto-standard - 4.19+git0+57b791cb9f_122d468967-r0.24 Mar 22 19:11:39 https://github.com/webOS-ports/meta-webos-ports/commit/19730134d7e108b3014b50f11329869ca5a0d200 Mar 22 19:12:22 yes it stopped showing it to me for luna-next and qtbase examples according to commit message (if I can believe myself) Mar 22 19:12:48 let me check with newer sumo based image Mar 22 19:15:34 one less :) Mar 22 19:16:56 interesting UI is shown with warrior based image Mar 22 19:18:49 but the cursor isn't rendered even after updating luna-next environment Mar 22 19:19:45 and there is also that WARNING: 19:18:07.722: Failed to move cursor on screen VGA1: -14 Mar 22 19:19:58 but already in luna-next.service journal Mar 22 20:00:44 going to check in xubuntu with qemu-native Mar 22 20:43:56 Tofe: I've imported the rest of vbox history with git svn in https://github.com/shr-project/virtualbox-org-svn-vbox-trunk Mar 22 23:00:48 thanks! **** ENDING LOGGING AT Sat Mar 23 03:00:16 2019