SUMMARY On Fedora 39 + KDE Plasma 5.27.10 + Wayland on native notebook hardware including * NVIDIA 3060 RTX mobile GPU * Intel Tiger Lake 11800H iGPU I seem to start getting random system hands on boot to a graphical desktop; the log then contains ``` Dec 15 18:25:24 dell-7610.home sddm-helper-start-wayland[2560]: "kwin_core: Failed to open /dev/dri/card0 device (Did not receive a reply. Possible causes include: the remote application did not send a reply, the message bus security policy blocked the reply, the reply timeout expired, or the network connection was broken.)\nkwin_wayland_drm: failed to open drm device at \"/dev/dri/card0\"\n" ``` In such a state, the system simply hangs dead with a blank screen for a handful of seconds (15? 25?). It will then show on screen the notebook vendor's boot logo. And stay there. The logs for this are ``` Dec 16 20:19:03 dell.home sddm-helper-start-wayland[2515]: "No backend specified, automatically choosing drm\n" ... Dec 16 20:19:30 dell.home sddm-helper-start-wayland[2515]: "kwin_core: Failed to open /dev/dri/card0 device (Did not receive a reply. Possible causes include: the remote application did not send a reply, the message bus security policy blocked the reply, the reply timeout expired, or the network connection was broken.)\nkwin_wayland_drm: failed to open drm device at \"/dev/dri/card0\"\nkwin_wayland_drm: No suitable DRM devices have been found\n" Dec 16 20:19:30 dell.home sddm-greeter[2518]: Creating a fake screen in order for Qt not to crash Dec 16 20:19:30 dell.home sddm-greeter[2518]: The Wayland connection broke. Did the Wayland compositor die? Dec 16 20:19:30 dell.home sddm-helper-start-wayland[2515]: Stopping... "/usr/bin/sddm-greeter" Dec 16 20:19:30 dell.home sddm-helper[2420]: pam_unix(sddm-greeter:session): session closed for user sddm ``` The "native" GPU sequence (which I cannot change in the BIOS) is, * /dev/dri/card0 == Intel * /dev/dri/card1 == NVIDIA Now, SDDM has peculiar configuration to force on Wayland and to exclude the NVIDIA card in /etc/sddm.conf.d/someconfig.conf ``` [General] DisplayServer=wayland GreeterEnvironment=QT_WAYLAND_SHELL_INTEGRATION=layer-shell,KWIN_DRM_DEVICES=/dev/dri/card0 ``` Physically connected are * the notebooks's internal display and an HDMI screen are alive and alight on Intel GPU (card0) * a 2.5K screen is on Thunderbolt -> DisplayPort on the NVIDIA GPU (card1) STEPS TO REPRODUCE 1. have setup as described above 2. boot OBSERVED RESULT Every once in a while, the system hangs dead with the above symptoms. EXPECTED RESULT Reliable start to SDDM greeter.
I am reporting this against kwin because of the implementation in `int LogindSession::openRestricted(const QString &fileName)` with the origin of the log entry evidently being ``` const QDBusMessage reply = QDBusConnection::systemBus().call(message); if (reply.type() == QDBusMessage::ErrorMessage) { qCWarning(KWIN_CORE, "Failed to open %s device (%s)", qPrintable(fileName), qPrintable(reply.errorMessage())); return -1; } ``` From context it is clear that the _device_ exists (10 lines earlier, the `stat` succeeds). From a user perspective, it may be beneficial to establish * a _shorter_ timeout on the dbus call * a mechanism for exponential back-off retries, capped to, say 30 seconds in kwin_core, in trying to acquire the device but I have no idea whether that is feasible at all.
Based on comments in https://bbs.archlinux.org/viewtopic.php?id=250684 I looked at there being a chance of racing between drm and systemd-logind (which apparently is responsible for handing out the file handle to card0). In the two blocks below based on `journalctl -b<the boot> | grep -E '\[drm\]|systemd-login'`, I do not see much of a difference in ordering, and none that would seem to matter (not that I would be competent in assessing that): ``` +++ broken fedora kernel: i915 0000:00:02.0: [drm] VT-d active for gfx access fedora kernel: i915 0000:00:02.0: [drm] Using Transparent Hugepages fedora kernel: i915 0000:00:02.0: [drm] Finished loading DMC firmware i915/tgl_dmc_ver2_12.bin (v2.12) fedora kernel: i915 0000:00:02.0: [drm] GT0: GuC firmware i915/tgl_guc_70.1.1.bin version 70.1.1 fedora kernel: i915 0000:00:02.0: [drm] GT0: HuC firmware i915/tgl_huc_7.9.3.bin version 7.9.3 fedora kernel: i915 0000:00:02.0: [drm] GT0: HuC: authenticated for all workloads fedora kernel: i915 0000:00:02.0: [drm] GT0: GUC: submission enabled fedora kernel: i915 0000:00:02.0: [drm] GT0: GUC: SLPC enabled fedora kernel: i915 0000:00:02.0: [drm] GT0: GUC: RC enabled fedora kernel: i915 0000:00:02.0: [drm] Protected Xe Path (PXP) protected content support initialized fedora kernel: [drm] Initialized i915 1.6.0 20201103 for 0000:00:02.0 on minor 0 fedora kernel: i915 0000:00:02.0: [drm] fb0: i915drmfb frame buffer device fedora kernel: i915 0000:00:02.0: [drm] Selective fetch area calculation failed in pipe A fedora systemd[1]: Starting systemd-logind.service - User Login Management... fedora systemd-logind[1520]: New seat seat0. fedora systemd-logind[1520]: Watching system buttons on /dev/input/event1 (Power Button) fedora systemd-logind[1520]: Watching system buttons on /dev/input/event0 (Lid Switch) fedora systemd-logind[1520]: Watching system buttons on /dev/input/event2 (Sleep Button) fedora systemd-logind[1520]: Watching system buttons on /dev/input/event15 (Intel HID events) fedora systemd-logind[1520]: Watching system buttons on /dev/input/event16 (Intel HID 5 button array) fedora systemd-logind[1520]: Watching system buttons on /dev/input/event3 (AT Translated Set 2 keyboard) fedora audit[1]: SERVICE_START pid=1 uid=0 auid=4294967295 ses=4294967295 subj=system_u:system_r:init_t:s0 msg='unit=systemd-logind comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success' fedora systemd[1]: Started systemd-logind.service - User Login Management. fedora kernel: [drm] [nvidia-drm] [GPU ID 0x00000100] Loading driver dell.home systemd-logind[1520]: Watching system buttons on /dev/input/event8 (Cherry GmbH CHERRY Corded Device) dell.home kernel: [drm] Initialized nvidia-drm 0.0.0 20160202 for 0000:01:00.0 on minor 1 dell.home kernel: nvidia 0000:01:00.0: [drm] Cannot find any crtc or sizes dell.home systemd-logind[1520]: Watching system buttons on /dev/input/event10 (Generic USB Audio Consumer Control) dell.home kernel: nvidia 0000:01:00.0: [drm] fb1: nvidia-drmdrmfb frame buffer device dell.home systemd-logind[1520]: New session c1 of user sddm. dell.home systemd-logind[1520]: Session c1 logged out. Waiting for processes to exit. dell.home systemd-logind[1520]: Removed session c1. dell.home systemd-logind[1520]: New session 2 of user stefan. dell.home systemd-logind[1520]: The system will reboot now! dell.home systemd-logind[1520]: System is rebooting. dell.home systemd-logind[1520]: Session 2 logged out. Waiting for processes to exit. dell.home systemd[1]: Stopping systemd-logind.service - User Login Management... dell.home systemd-logind[1520]: Removed session 2. dell.home systemd[1]: systemd-logind.service: Deactivated successfully. dell.home audit[1]: SERVICE_STOP pid=1 uid=0 auid=4294967295 ses=4294967295 subj=system_u:system_r:init_t:s0 msg='unit=systemd-logind comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success' dell.home systemd[1]: Stopped systemd-logind.service - User Login Management. ``` ``` +++ working fedora kernel: i915 0000:00:02.0: [drm] VT-d active for gfx access fedora kernel: i915 0000:00:02.0: [drm] Using Transparent Hugepages fedora kernel: i915 0000:00:02.0: [drm] Finished loading DMC firmware i915/tgl_dmc_ver2_12.bin (v2.12) fedora kernel: i915 0000:00:02.0: [drm] GT0: GuC firmware i915/tgl_guc_70.1.1.bin version 70.1.1 fedora kernel: i915 0000:00:02.0: [drm] GT0: HuC firmware i915/tgl_huc_7.9.3.bin version 7.9.3 fedora kernel: i915 0000:00:02.0: [drm] GT0: HuC: authenticated for all workloads fedora kernel: i915 0000:00:02.0: [drm] GT0: GUC: submission enabled fedora kernel: i915 0000:00:02.0: [drm] GT0: GUC: SLPC enabled fedora kernel: i915 0000:00:02.0: [drm] GT0: GUC: RC enabled fedora kernel: i915 0000:00:02.0: [drm] Protected Xe Path (PXP) protected content support initialized fedora kernel: [drm] Initialized i915 1.6.0 20201103 for 0000:00:02.0 on minor 0 fedora kernel: i915 0000:00:02.0: [drm] fb0: i915drmfb frame buffer device fedora kernel: i915 0000:00:02.0: [drm] Selective fetch area calculation failed in pipe A fedora systemd[1]: Starting systemd-logind.service - User Login Management... fedora systemd-logind[1526]: New seat seat0. fedora systemd-logind[1526]: Watching system buttons on /dev/input/event1 (Power Button) fedora systemd-logind[1526]: Watching system buttons on /dev/input/event0 (Lid Switch) fedora systemd-logind[1526]: Watching system buttons on /dev/input/event2 (Sleep Button) fedora systemd-logind[1526]: Watching system buttons on /dev/input/event15 (Intel HID events) fedora systemd-logind[1526]: Watching system buttons on /dev/input/event17 (Intel HID 5 button array) fedora systemd-logind[1526]: Watching system buttons on /dev/input/event3 (AT Translated Set 2 keyboard) fedora systemd[1]: Started systemd-logind.service - User Login Management. fedora audit[1]: SERVICE_START pid=1 uid=0 auid=4294967295 ses=4294967295 subj=system_u:system_r:init_t:s0 msg='unit=systemd-logind comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success' fedora kernel: [drm] [nvidia-drm] [GPU ID 0x00000100] Loading driver dell.home systemd-logind[1526]: Watching system buttons on /dev/input/event8 (Cherry GmbH CHERRY Corded Device) dell.home kernel: [drm] Initialized nvidia-drm 0.0.0 20160202 for 0000:01:00.0 on minor 1 dell.home kernel: nvidia 0000:01:00.0: [drm] Cannot find any crtc or sizes dell.home systemd-logind[1526]: New session c1 of user sddm. dell.home systemd-logind[1526]: Watching system buttons on /dev/input/event10 (Generic USB Audio Consumer Control) dell.home kernel: nvidia 0000:01:00.0: [drm] fb1: nvidia-drmdrmfb frame buffer device dell.home systemd-logind[1526]: New session 2 of user stefan. dell.home systemd-logind[1526]: Session c1 logged out. Waiting for processes to exit. dell.home systemd-logind[1526]: Removed session c1. ```
I got the reject on the Intel card ("/dev/dri/card0") again this morning, on a modified system configuration. I had tried to mitigate this issue by modifying the following two configuration parts on my system: * modified sddm configuration to remove KWIN_DRM_DEVICES (i.e. sddm may now use primary card0 == Intel and card1 == nvidia) - it is still force-configured to use Wayland * added all nvidia modules to `/etc/modules-load.d/nvidia.conf` and rebuilt initramfs via `dracut --regenerate-all --force` (cf `lsinitrd`) Additionally, other than before, the NVIDIA GPU did not have any output connected. In the logs (excerpt below) this is what I see: * the session for user sddm comes up fine * dbus-broker.service seems to have successfully started * there is SELinux denial for sys_admin on systemd-modules (triggered by AVC `avc: denied { sys_admin } for pid=1483 comm="nv_queue"`), but that would be NVIDIA matching the nv_queue kernel thread, if at all (and frankly, I suspect that this is a defect elsewere, because the whole point of systemd-modules _is_ loading modules as an admin?) * kwin_wayland terminates with an error while trying to get hold of the Intel card "/dev/dri/card0" * sddm? kwin? tries again on the NVIDIA card (`sddm-helper-start-wayland[2526]: "OpenGL vendor string: NVIDIA Corporation`) * sddm-greeter does not find a screen (correct, the NVIDIA GPU does not have a screen attached, and I don't think it would be reasonable for the NVIDIA GPU to be able to talk to the apparently not-so-happy Intel GPU at this stage) So, right now I am mostly interested in maximizing relevant log output from * sddm * dbus (traffic) * systemd-logind (accessing /dev/dri/card0) in the hope to get more context. FWIW, I recall one more peculiarity on this system: I am forcing the non-default Intel GPU firmware loading for guc and huc (`options i915 enable_guc=3`), see https://wiki.archlinux.org/title/intel_graphics . Turning this off for now. ``` Dec 17 08:24:37 dell.home systemd-logind[1684]: New session c1 of user sddm. Dec 17 08:24:37 dell.home systemd[1]: Started uresourced.service - User resource assignment daemon. Dec 17 08:24:37 dell.home audit[1]: SERVICE_START pid=1 uid=0 auid=4294967295 ses=4294967295 subj=system_u:system_r:init_t:s0 msg='unit=uresourced comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success' Dec 17 08:24:37 dell.home systemd[1]: Finished user-runtime-dir@988.service - User Runtime Directory /run/user/988. Dec 17 08:24:37 dell.home audit[1]: SERVICE_START pid=1 uid=0 auid=4294967295 ses=4294967295 subj=system_u:system_r:init_t:s0 msg='unit=user-runtime-dir@988 comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success' Dec 17 08:24:37 dell.home systemd[1]: Starting user@988.service - User Manager for UID 988... Dec 17 08:24:37 dell.home dbus-broker-launch[1599]: Activation request for 'org.freedesktop.home1' failed: The systemd unit 'dbus-org.freedesktop.home1.service' could not be found. Dec 17 08:24:37 dell.home audit[2508]: USER_ACCT pid=2508 uid=0 auid=4294967295 ses=4294967295 subj=system_u:system_r:init_t:s0 msg='op=PAM:accounting grantors=pam_unix acct="sddm" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success' Dec 17 08:24:37 dell.home audit[2508]: CRED_ACQ pid=2508 uid=0 auid=4294967295 ses=4294967295 subj=system_u:system_r:init_t:s0 msg='op=PAM:setcred grantors=? acct="sddm" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=failed' Dec 17 08:24:37 dell.home audit[2508]: USER_ROLE_CHANGE pid=2508 uid=0 auid=4294967295 ses=4294967295 subj=system_u:system_r:init_t:s0 msg='pam: default-context=unconfined_u:unconfined_r:unconfined_t:s0-s0:c0.c1023 selected-context=unconfined_u:unconfined_r:unconfined_t:s0-s0:c0.c1023 exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success' Dec 17 08:24:37 dell.home (systemd)[2508]: pam_unix(systemd-user:session): session opened for user sddm(uid=988) by sddm(uid=0) Dec 17 08:24:37 dell.home audit[2508]: USER_START pid=2508 uid=0 auid=988 ses=1 subj=system_u:system_r:init_t:s0 msg='op=PAM:session_open grantors=pam_selinux,pam_selinux,pam_loginuid,pam_keyinit,pam_namespace,pam_systemd_home,pam_umask,pam_keyinit,pam_limits,pam_systemd,pam_unix acct="sddm" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success' Dec 17 08:24:37 dell.home uresourced[2484]: Setting resources on user.slice (MemoryMin: 262144000, MemoryLow: 0, CPUWeight: -, IOWeight: -) Dec 17 08:24:37 dell.home uresourced[2484]: Setting resources on user-988.slice (MemoryMin: 262144000, MemoryLow: 0, CPUWeight: 500, IOWeight: 500) Dec 17 08:24:37 dell.home uresourced[2484]: Setting resources on user@988.service (MemoryMin: 0, MemoryLow: 0, CPUWeight: 100, IOWeight: 100) Dec 17 08:24:37 dell.home systemd[2508]: Queued start job for default target default.target. Dec 17 08:24:37 dell.home systemd[2508]: Created slice app.slice - User Application Slice. Dec 17 08:24:37 dell.home systemd[2508]: grub-boot-success.timer - Mark boot as successful after the user session has run 2 minutes was skipped because of an unmet condition check (ConditionUser=!@system). Dec 17 08:24:37 dell.home systemd[2508]: Started systemd-tmpfiles-clean.timer - Daily Cleanup of User's Temporary Directories. Dec 17 08:24:37 dell.home systemd[2508]: Reached target paths.target - Paths. Dec 17 08:24:37 dell.home systemd[2508]: Reached target timers.target - Timers. Dec 17 08:24:37 dell.home systemd[2508]: Starting dbus.socket - D-Bus User Message Bus Socket... Dec 17 08:24:37 dell.home systemd[2508]: Listening on pipewire-pulse.socket - PipeWire PulseAudio. Dec 17 08:24:37 dell.home systemd[2508]: Listening on pipewire.socket - PipeWire Multimedia System Sockets. Dec 17 08:24:37 dell.home systemd[2508]: Starting systemd-tmpfiles-setup.service - Create User's Volatile Files and Directories... Dec 17 08:24:37 dell.home systemd[2508]: Listening on dbus.socket - D-Bus User Message Bus Socket. Dec 17 08:24:37 dell.home systemd[2508]: Reached target sockets.target - Sockets. Dec 17 08:24:37 dell.home systemd[2508]: Finished systemd-tmpfiles-setup.service - Create User's Volatile Files and Directories. Dec 17 08:24:37 dell.home systemd[2508]: Reached target basic.target - Basic System. Dec 17 08:24:37 dell.home systemd[2508]: Reached target default.target - Main User Target. Dec 17 08:24:37 dell.home systemd[2508]: Startup finished in 125ms. Dec 17 08:24:37 dell.home systemd[1]: Started user@988.service - User Manager for UID 988. Dec 17 08:24:37 dell.home audit[1]: SERVICE_START pid=1 uid=0 auid=4294967295 ses=4294967295 subj=system_u:system_r:init_t:s0 msg='unit=user@988 comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success' Dec 17 08:24:37 dell.home systemd[1]: Started session-c1.scope - Session c1 of User sddm. Dec 17 08:24:37 dell.home systemd[1]: Startup finished in 8.663s (firmware) + 7.186s (loader) + 1.020s (kernel) + 4.798s (initrd) + 8.032s (userspace) = 29.701s. Dec 17 08:24:37 dell.home audit[2443]: USER_START pid=2443 uid=0 auid=4294967295 ses=4294967295 subj=system_u:system_r:xdm_t:s0-s0:c0.c1023 msg='op=PAM:session_open grantors=pam_unix,pam_systemd acct="sddm" exe="/usr/libexec/sddm-helper" hostname=dell.home addr=? terminal=/dev/tty1 res=success' Dec 17 08:24:37 dell.home sddm-helper-start-wayland[2526]: "No backend specified, automatically choosing drm\n" Dec 17 08:24:37 dell.home systemd[2508]: Created slice session.slice - User Core Session Slice. Dec 17 08:24:37 dell.home systemd[2508]: Starting dbus-broker.service - D-Bus User Message Bus... Dec 17 08:24:37 dell.home dbus-broker-launch[2533]: Service file '/usr/share//dbus-1/services/imsettings-daemon.service' is not named after the D-Bus name 'com.redhat.imsettings'. Dec 17 08:24:37 dell.home dbus-broker-launch[2533]: Service file '/usr/share//dbus-1/services/org.kde.kscreen.service' is not named after the D-Bus name 'org.kde.KScreen'. Dec 17 08:24:37 dell.home dbus-broker-launch[2533]: Service file '/usr/share//dbus-1/services/org.kde.dolphin.FileManager1.service' is not named after the D-Bus name 'org.freedesktop.FileManager1'. Dec 17 08:24:37 dell.home dbus-broker-launch[2533]: Service file '/usr/share//dbus-1/services/org.kde.plasma.Notifications.service' is not named after the D-Bus name 'org.freedesktop.Notifications'. Dec 17 08:24:37 dell.home dbus-broker-launch[2533]: Policy to allow eavesdropping in /usr/share/dbus-1/session.conf +31: Eavesdropping is deprecated and ignored Dec 17 08:24:37 dell.home dbus-broker-launch[2533]: Policy to allow eavesdropping in /usr/share/dbus-1/session.conf +33: Eavesdropping is deprecated and ignored Dec 17 08:24:37 dell.home systemd[2508]: Started dbus-broker.service - D-Bus User Message Bus. Dec 17 08:24:37 dell.home dbus-broker-launch[2533]: Ready Dec 17 08:24:38 dell.home kernel: thunderbolt 0-1: new device found, vendor=0x16b device=0x9010 Dec 17 08:24:38 dell.home kernel: thunderbolt 0-1: i-tec TB3CDUALDPDOCKPD Dec 17 08:24:38 dell.home boltd[1931]: [004158c8-bd26-TB3CDUALDPDOCKPD ] parent is 9096ea89-ad77... Dec 17 08:24:38 dell.home boltd[1931]: [004158c8-bd26-TB3CDUALDPDOCKPD ] connected: authorized (/sys/devices/pci0000:00/0000:00:0d.2/domain0/0-0/0-1) Dec 17 08:24:38 dell.home boltd[1931]: [004158c8-bd26-TB3CDUALDPDOCKPD ] udev: device changed: authorized -> authorized Dec 17 08:24:38 dell.home chronyd[1962]: Selected source 85.195.224.28 (2.fedora.pool.ntp.org) Dec 17 08:24:38 dell.home chronyd[1962]: System clock TAI offset set to 37 seconds Dec 17 08:24:44 dell.home systemd[1]: dbus-:1.3-org.fedoraproject.SetroubleshootPrivileged@0.service: Deactivated successfully. Dec 17 08:24:44 dell.home audit[1]: SERVICE_STOP pid=1 uid=0 auid=4294967295 ses=4294967295 subj=system_u:system_r:init_t:s0 msg='unit=dbus-:1.3-org.fedoraproject.SetroubleshootPrivileged@0 comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success' Dec 17 08:24:45 dell.home systemd[1]: setroubleshootd.service: Deactivated successfully. Dec 17 08:24:45 dell.home audit[1]: SERVICE_STOP pid=1 uid=0 auid=4294967295 ses=4294967295 subj=system_u:system_r:init_t:s0 msg='unit=setroubleshootd comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success' Dec 17 08:24:45 dell.home systemd[1]: NetworkManager-dispatcher.service: Deactivated successfully. Dec 17 08:24:45 dell.home audit[1]: SERVICE_STOP pid=1 uid=0 auid=4294967295 ses=4294967295 subj=system_u:system_r:init_t:s0 msg='unit=NetworkManager-dispatcher comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success' Dec 17 08:24:53 dell.home audit[1483]: AVC avc: denied { sys_admin } for pid=1483 comm="nv_queue" capability=21 scontext=system_u:system_r:systemd_modules_load_t:s0 tcontext=system_u:system_r:systemd_modules_load_t:s0 tclass=capability permissive=0 Dec 17 08:24:55 dell.home systemd[1]: Starting setroubleshootd.service - SETroubleshoot daemon for processing new SELinux denial logs... Dec 17 08:24:55 dell.home systemd[1]: Started setroubleshootd.service - SETroubleshoot daemon for processing new SELinux denial logs. Dec 17 08:24:55 dell.home audit[1]: SERVICE_START pid=1 uid=0 auid=4294967295 ses=4294967295 subj=system_u:system_r:init_t:s0 msg='unit=setroubleshootd comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success' Dec 17 08:24:55 dell.home systemd[1]: Started dbus-:1.3-org.fedoraproject.SetroubleshootPrivileged@1.service. Dec 17 08:24:55 dell.home audit[1]: SERVICE_START pid=1 uid=0 auid=4294967295 ses=4294967295 subj=system_u:system_r:init_t:s0 msg='unit=dbus-:1.3-org.fedoraproject.SetroubleshootPrivileged@1 comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success' Dec 17 08:24:56 dell.home setroubleshoot[2542]: SELinux is preventing systemd-modules from using the sys_admin capability. For complete SELinux messages run: sealert -l 8cabac7f-032b-4161-8b08-01cbd2cfef2d Dec 17 08:24:56 dell.home setroubleshoot[2542]: SELinux is preventing systemd-modules from using the sys_admin capability. ***** Plugin catchall (100. confidence) suggests ************************** If you believe that systemd-modules should have the sys_admin capability by default. Then you should report this as a bug. You can generate a local policy module to allow this access. Do allow this access for now by executing: # ausearch -c 'systemd-modules' --raw | audit2allow -M my-systemdmodules # semodule -X 300 -i my-systemdmodules.pp Dec 17 08:25:02 dell.home sddm-helper-start-wayland[2526]: "kwin_core: Failed to open /dev/dri/card0 device (Did not receive a reply. Possible causes include: the remote application did not send a reply, the message bus security policy blocked the reply, the reply timeout expired, or the network connection was broken.)\nkwin_wayland_drm: failed to open drm device at \"/dev/dri/card0\"\n" Dec 17 08:25:03 dell.home systemd[1]: systemd-hostnamed.service: Deactivated successfully. Dec 17 08:25:03 dell.home audit[1]: SERVICE_STOP pid=1 uid=0 auid=4294967295 ses=4294967295 subj=system_u:system_r:init_t:s0 msg='unit=systemd-hostnamed comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success' Dec 17 08:25:03 dell.home audit: BPF prog-id=77 op=UNLOAD Dec 17 08:25:03 dell.home audit: BPF prog-id=76 op=UNLOAD Dec 17 08:25:03 dell.home audit: BPF prog-id=75 op=UNLOAD Dec 17 08:25:05 dell.home systemd[1]: dbus-:1.3-org.fedoraproject.SetroubleshootPrivileged@1.service: Deactivated successfully. Dec 17 08:25:05 dell.home audit[1]: SERVICE_STOP pid=1 uid=0 auid=4294967295 ses=4294967295 subj=system_u:system_r:init_t:s0 msg='unit=dbus-:1.3-org.fedoraproject.SetroubleshootPrivileged@1 comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success' Dec 17 08:25:06 dell.home systemd[1]: setroubleshootd.service: Deactivated successfully. Dec 17 08:25:06 dell.home audit[1]: SERVICE_STOP pid=1 uid=0 auid=4294967295 ses=4294967295 subj=system_u:system_r:init_t:s0 msg='unit=setroubleshootd comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success' Dec 17 08:25:11 dell.home sddm-helper-start-wayland[2526]: "OpenGL vendor string: NVIDIA Corporation\nOpenGL renderer string: NVIDIA GeForce RTX 3060 Laptop GPU/PCIe/SSE2\nOpenGL version string: 3.1.0 NVIDIA 545.29.06\nOpenGL shading language version string: 1.40 NVIDIA via Cg compiler\nDriver: NVIDIA\nDriver version: 545.29.6\nGPU class: Unknown\nOpenGL version: 3.1\nGLSL version: 1.40\nLinux kernel version: 6.6.6\nRequires strict binding: no\nGLSL shaders: yes\nTexture NPOT support: yes\nVirtual Machine: no\n" Dec 17 08:25:12 dell.home sddm-helper-start-wayland[2526]: "kf.service.services: The desktop entry file \"/usr/share/applications/qemu.desktop\" has Type= \"Application\" but has no Exec field.\n" Dec 17 08:25:12 dell.home sddm-helper-start-wayland[2526]: "kf.service.services: The desktop entry file \"/usr/share/applications/org.freedesktop.Xwayland.desktop\" has Type= \"Application\" but has no Exec field.\n" Dec 17 08:25:12 dell.home sddm-helper-start-wayland[2526]: "kf.service.services: The desktop entry file \"/usr/share/applications/logmgr.desktop\" has Type= \"Application\" but has no Exec field.\n" Dec 17 08:25:12 dell.home sddm-greeter[2529]: Creating a fake screen in order for Qt not to crash Dec 17 08:25:22 dell.home audit[1483]: AVC avc: denied { sys_admin } for pid=1483 comm="nv_queue" capability=21 scontext=system_u:system_r:systemd_modules_load_t:s0 tcontext=system_u:system_r:systemd_modules_load_t:s0 tclass=capability permissive=0 Dec 17 08:25:24 dell.home systemd[1]: Starting setroubleshootd.service - SETroubleshoot daemon for processing new SELinux denial logs... Dec 17 08:25:24 dell.home systemd[1]: Started setroubleshootd.service - SETroubleshoot daemon for processing new SELinux denial logs. Dec 17 08:25:24 dell.home audit[1]: SERVICE_START pid=1 uid=0 auid=4294967295 ses=4294967295 subj=system_u:system_r:init_t:s0 msg='unit=setroubleshootd comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success' Dec 17 08:25:24 dell.home systemd[1]: Started dbus-:1.3-org.fedoraproject.SetroubleshootPrivileged@2.service. Dec 17 08:25:24 dell.home audit[1]: SERVICE_START pid=1 uid=0 auid=4294967295 ses=4294967295 subj=system_u:system_r:init_t:s0 msg='unit=dbus-:1.3-org.fedoraproject.SetroubleshootPrivileged@2 comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success' Dec 17 08:25:25 dell.home setroubleshoot[2577]: SELinux is preventing systemd-modules from using the sys_admin capability. For complete SELinux messages run: sealert -l 8cabac7f-032b-4161-8b08-01cbd2cfef2d Dec 17 08:25:25 dell.home setroubleshoot[2577]: SELinux is preventing systemd-modules from using the sys_admin capability. ***** Plugin catchall (100. confidence) suggests ************************** If you believe that systemd-modules should have the sys_admin capability by default. Then you should report this as a bug. You can generate a local policy module to allow this access. Do allow this access for now by executing: # ausearch -c 'systemd-modules' --raw | audit2allow -M my-systemdmodules # semodule -X 300 -i my-systemdmodules.pp Dec 17 08:25:34 dell.home systemd[1]: dbus-:1.3-org.fedoraproject.SetroubleshootPrivileged@2.service: Deactivated successfully. Dec 17 08:25:34 dell.home audit[1]: SERVICE_STOP pid=1 uid=0 auid=4294967295 ses=4294967295 subj=system_u:system_r:init_t:s0 msg='unit=dbus-:1.3-org.fedoraproject.SetroubleshootPrivileged@2 comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success' Dec 17 08:25:35 dell.home systemd[1]: setroubleshootd.service: Deactivated successfully. Dec 17 08:25:35 dell.home audit[1]: SERVICE_STOP pid=1 uid=0 auid=4294967295 ses=4294967295 subj=system_u:system_r:init_t:s0 msg='unit=setroubleshootd comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success' Dec 17 08:25:44 dell.home chronyd[1962]: Selected source 192.33.214.47 (2.fedora.pool.ntp.org) Dec 17 08:26:37 dell.home systemd[1]: virtlxcd.service: Deactivated successfully. Dec 17 08:26:37 dell.home audit[1]: SERVICE_STOP pid=1 uid=0 auid=4294967295 ses=4294967295 subj=system_u:system_r:init_t:s0 msg='unit=virtlxcd comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success' Dec 17 08:26:37 dell.home systemd[1]: virtqemud.service: Deactivated successfully. Dec 17 08:26:37 dell.home audit[1]: SERVICE_STOP pid=1 uid=0 auid=4294967295 ses=4294967295 subj=system_u:system_r:init_t:s0 msg='unit=virtqemud comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success' Dec 17 08:26:52 dell.home chronyd[1962]: Selected source 2a02:1210:249e:f00:1eb0:44ff:fefa:ec90 ... it just goes downhill from here ... Dec 17 08:29:40 dell.home systemd-logind[1684]: Power key pressed short. Dec 17 08:29:40 dell.home systemd-logind[1684]: Powering off... Dec 17 08:29:40 dell.home systemd-logind[1684]: System is powering down. Dec 17 08:29:40 dell.home systemd[1]: Stopping session-c1.scope - Session c1 of User sddm... Dec 17 08:29:40 dell.home sddm[2368]: Error from greeter session: "Process crashed" Dec 17 08:29:40 dell.home sddm[2368]: Auth: sddm-helper (--socket /tmp/sddm-auth-fb32114e-e536-4e4f-a8ad-e6000f61baa8 --id 2 --start /usr/bin/sddm-greeter --socket /tmp/sddm--CAFWSv --theme /usr/share/sddm/themes/01-breeze-fedora --user sddm --display-server kwin_wayland --no-lockscreen --greeter) crashed (exit code 1) Dec 17 08:29:40 dell.home sddm[2368]: Error from greeter session: "Process crashed" Dec 17 08:29:40 dell.home sddm[2368]: Auth: sddm-helper exited with 1 ```
FWIW, I am concerned that this issue here has been sent into the direction of "NVIDIA" - while there certainly is NVIDIA hardware in this system, at least some failures were with KWIN_DRM_DEVICES=/dev/drm/card0 which ties this straight to Intel.
Right now it is a lottery game whether I get to the sddm prompt, without ""kwin_core: Failed to open /dev/dri/card0 device (Did not receive a reply. " Some findings: * Intel firmware (removing the non-default `guc=3`) does not make a difference * only with KWIN_DRM_DEVICES=/dev/drm/card0 for sddm do I get the change to switch to a console (Ctrl+Alt+F4), so this option is back on, as on this notebook I never want to have the NVIDIA GPU (card1) drive the graphical login * `reboot` from the console ends up in the graphics stack again, so never reboots successfully; I have to use ACPI power down (hold power button for 10 seconds) to force shutdown Just for laughs, I have now set selinux policy from enforcing to permissive ("/etc/selinux/config"). The boot after doing that worked (and I am writing this from a working session).
It appears as if - somehow - this is indeed related to the NVIDIA driver "being around". With all prior changes reverted and only the following changes on place, I now again get reliable boots into a KDE Plasma graphical user interface: * blacklist modules nvidia, nvidia-drm, nvidia-modeset, nvidia-uvm ("/etc/modprobe.d/nvidia-blacklist.conf") * dracut blacklist the above modules ("/lib/dracut/dracut.conf.d/99-nvidia-dracut.conf") * set module parameter "options nvidia_drm fbdev=0" ("/etc/modprobe.d/nvidia_drm.conf") - this is somewhat pointless given that the module is blacklisted, but on my system I will never need an NVIDIA framebuffer * mask falling back to the nouveau driver ("systemctl mask nvidia-fallback.service") The net effect of this is that "lsmod | grep -E '^(nvidia|nouveau)'" will return only ``` nvidia_uvm 3522560 0 nvidia 62394368 9 nvidia_uvm ``` which _disables_ the complete graphics stack (including NVIDIA Prime with OpenGL / Vulkan), but which _retains_ the ability run NVIDIA CUDA processing loads (I cannot explain at this time where nvidia_uvm actually gets loaded). There also seems to have been some kind of interaction with multiple screens attached (all of them attached to the Intel GPU, though!) - booting with a single screen seems to have always worked. FWIW, I do not have a GNOME installation on this box, so I am unable to tell whether this would be affected as well; by the look-and-feel of everything, I would suspect that GNOME would suffer from exactly the same problems.
This does not seem to be related to NVIDIA. Right now I suspect multi-monitor support on the Intel 11800H iGPU (there is a 4K screen attached to that GPU, but the monitor has been working very fine going through all of the Plymouth "nice animated system boot" graphics fun) Reason: I have now had a cold boot on this notebook (power on after having been turned off for six hours) and got the same problem, again. With more logging enabled, it is now clear that systemd_logind simply does not respond in time: * 09:35:52 - kwin_wayland (apparently) sends the TakeDevice request via dbus * 09:35:52 - systemd_logind acks request [25 secs after request] * 09:36:17 - kwin_wayland gives up waiting, "Failed to open /dev/dri/card0 device"; sddm is toast [34 secs after request] * 09:36:26 - systemd_logind sends response, apparently a success ``` Dec 20 09:35:52 dell.home systemd[2417]: Started dbus-broker.service - D-Bus User Message Bus. Dec 20 09:35:52 dell.home dbus-broker-launch[2448]: Ready Dec 20 09:35:52 dell.home systemd-logind[1531]: Got message type=method_call sender=:1.36 destination=org.freedesktop.login1 path=/org/freedesktop/login1/session/c1 interface=org.freedesktop.login1.Session member=TakeDevice cookie=15 reply_cookie=0 signature=uu error-name=n/a error-message=n/a Dec 20 09:35:54 dell.home kernel: thunderbolt 0-1: new device found, vendor=0x16b device=0x9010 Dec 20 09:35:54 dell.home kernel: thunderbolt 0-1: i-tec TB3CDUALDPDOCKPD Dec 20 09:35:54 dell.home boltd[1725]: [004158c8-bd26-TB3CDUALDPDOCKPD ] parent is f07211eb-5a97... Dec 20 09:35:54 dell.home boltd[1725]: [004158c8-bd26-TB3CDUALDPDOCKPD ] connected: authorized (/sys/devices/pci0000:00/0000:00:0d.2/domain0/0-0/0-1) Dec 20 09:35:54 dell.home boltd[1725]: [004158c8-bd26-TB3CDUALDPDOCKPD ] udev: device changed: authorized -> authorized Dec 20 09:36:01 dell.home systemd[1]: NetworkManager-dispatcher.service: Deactivated successfully. Dec 20 09:36:01 dell.home audit[1]: SERVICE_STOP pid=1 uid=0 auid=4294967295 ses=4294967295 subj=system_u:system_r:init_t:s0 msg='unit=NetworkManager-dispatcher comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success' Dec 20 09:36:17 dell.home sddm-helper-start-wayland[2433]: "kwin_core: Failed to open /dev/dri/card0 device (Did not receive a reply. Possible causes include: the remote application did not send a reply, the message bus security policy blocked the reply, the reply timeout expired, or the network connection was broken.)\nkwin_waylan> Dec 20 09:36:17 dell.home sddm-greeter[2444]: Creating a fake screen in order for Qt not to crash Dec 20 09:36:17 dell.home sddm-greeter[2444]: The Wayland connection broke. Did the Wayland compositor die? Dec 20 09:36:17 dell.home sddm-helper-start-wayland[2433]: Stopping... "/usr/bin/sddm-greeter" Dec 20 09:36:17 dell.home sddm-helper[2323]: pam_unix(sddm-greeter:session): session closed for user sddm ... Dec 20 09:36:26 dell.home systemd-logind[1531]: Sent message type=method_return sender=n/a destination=:1.36 path=n/a interface=n/a member=n/a cookie=73 reply_cookie=15 signature=hb error-name=n/a error-message=n/a ```
(In reply to Stefan Hoffmeister from comment #5) > Just for laughs, I have now set selinux policy from enforcing to permissive > ("/etc/selinux/config"). The boot after doing that worked (and I am writing > this from a working session). This sounds like selinux is misconfigured or has issues with NVidia then - and this is definitely not something we can do anything about tbh. If you still have this issue, please report it to Fedora.