Fedora 39 Kinoite Standard Notes flatpak from flathub refuses to load

Newly installed Flatpak on newly installed and updated Fedora 39 Kinoite. Here is the output from terminal:

$ flatpak run org.standardnotes.standardnotes
[3:0207/163941.595827:ERROR:bus.cc(407)] Failed to connect to the bus: Failed to connect to socket /run/dbus/system_bus_socket: No such file or directory
LaunchProcess: failed to execvp:
xdg-settings
[3:0207/163943.215324:ERROR:bus.cc(407)] Failed to connect to the bus: Failed to connect to socket /run/dbus/system_bus_socket: No such file or directory
[3:0207/163943.215373:ERROR:bus.cc(407)] Failed to connect to the bus: Failed to connect to socket /run/dbus/system_bus_socket: No such file or directory
Checking for beta autoupdate feature for deb/rpm distributions
Found package-type: deb
16:39:43.284 › Checking for update
extServer: Server started at http://127.0.0.1:45653
[49:0207/163943.320229:ERROR:gbm_pixmap_wayland.cc(82)] Cannot create bo with format= RGBA_8888 and usage=SCANOUT
[49:0207/163943.321691:ERROR:gbm_pixmap_wayland.cc(82)] Cannot create bo with format= RGBA_8888 and usage=GPU_READ
[49:0207/163943.321767:ERROR:shared_image_factory.cc(926)] CreateSharedImage: could not create backing.
[49:0207/163943.321815:ERROR:shared_image_factory.cc(758)] DestroySharedImage: Could not find shared image mailbox
[49:0207/163943.321916:ERROR:gpu_service_impl.cc(1089)] Exiting GPU process because some drivers can't recover from errors. GPU process will restart shortly.
[3:0207/163943.348433:ERROR:gpu_process_host.cc(991)] GPU process exited unexpectedly: exit_code=8704
[91:0207/163943.423373:ERROR:gbm_pixmap_wayland.cc(82)] Cannot create bo with format= RGBA_8888 and usage=SCANOUT
[91:0207/163943.424839:ERROR:gbm_pixmap_wayland.cc(82)] Cannot create bo with format= RGBA_8888 and usage=GPU_READ
[91:0207/163943.424905:ERROR:shared_image_factory.cc(926)] CreateSharedImage: could not create backing.
[91:0207/163943.424964:ERROR:shared_image_factory.cc(758)] DestroySharedImage: Could not find shared image mailbox
[91:0207/163943.425058:ERROR:gpu_service_impl.cc(1089)] Exiting GPU process because some drivers can't recover from errors. GPU process will restart shortly.
[73:0207/163943.427895:ERROR:command_buffer_proxy_impl.cc(127)] ContextResult::kTransientFailure: Failed to send GpuControl.CreateCommandBuffer.
[3:0207/163943.446009:ERROR:gpu_process_host.cc(991)] GPU process exited unexpectedly: exit_code=8704
[104:0207/163943.521860:ERROR:gbm_pixmap_wayland.cc(82)] Cannot create bo with format= RGBA_8888 and usage=SCANOUT
[104:0207/163943.523148:ERROR:gbm_pixmap_wayland.cc(82)] Cannot create bo with format= RGBA_8888 and usage=GPU_READ
[104:0207/163943.523214:ERROR:shared_image_factory.cc(926)] CreateSharedImage: could not create backing.
[104:0207/163943.523285:ERROR:shared_image_factory.cc(758)] DestroySharedImage: Could not find shared image mailbox
[104:0207/163943.523382:ERROR:gpu_service_impl.cc(1089)] Exiting GPU process because some drivers can't recover from errors. GPU process will restart shortly.
[3:0207/163943.547500:ERROR:gpu_process_host.cc(991)] GPU process exited unexpectedly: exit_code=8704
16:39:43.974 › Checking for update (already in progress)
16:39:44.094 › Update for version 3.189.1 is not available (latest version: 3.187.3, downgrade is disallowed).

Might be better reported on Flathub. Did you make sure that all Flatpaks and Runtimes are up-to-date?

$ flatpak update
1 Like

The errors here appear to be GPU related.

Would you mind sharing the output of inxi -CMSGxx in a preformatted/code block?

Thanks for the reply. It is a new installation, but yes I ran the command. Nothing to do.

Sooo. Completely new to this.

$ inxi -CMSGxx bash: inxi: command not found

Do I run
$ sudo rpm-ostree install inxi
$ rpm-ostree install inxi
or something else?

Try: sudo rpm-ostree install inxi --apply-live

1 Like

OK, here is the output:

$ inxi -CMSGxx
12System:
  12Kernel 6.7.3-200.fc39.x86_64 12arch x86_64 12bits 64 12compiler gcc 12v 2.40-14.fc39 12Desktop KDE Plasma
    12v 5.27.10 12tk Qt 12v 5.15.12 12wm kwin_wayland 12dm SDDM 12Distro Fedora release 39 (Thirty Nine)
12Machine:
  12Type Desktop 12System LENOVO 12product 90RW005PUL 12v IdeaCentre Gaming5 14ACN6
    12serial <superuser required> 12Chassis 12type 3 12serial <superuser required>
  12Mobo LENOVO 12model 3743 12v NOK 12serial <superuser required> 12UEFI LENOVO 12v O5EKT1AA 12date 06/10/2021
12CPU:
  12Info 6-core 12model AMD Ryzen 5 5600G with Radeon Graphics 12bits 64 12type MT MCP 12arch Zen 3 12rev 0
    12cache 12L1 384 KiB 12L2 3 MiB 12L3 16 MiB
  12Speed (MHz) 12avg 929 12high 3666 12min/max 400/4464 12cores 121 400 122 400 123 400 124 400 125 400 126 400 127 3666
    128 400 129 400 1210 3482 1211 400 1212 400 12bogomips 93429
  12Flags avx avx2 ht lm nx pae sse sse2 sse3 sse4_1 sse4_2 sse4a ssse3 svm
12Graphics:
  12Device-1 NVIDIA TU116 [GeForce GTX 1660 SUPER] 12vendor Micro-Star MSI 12driver nouveau 12v kernel
    12arch Turing 12pcie 12speed 2.5 GT/s 12lanes 16 12ports 12active HDMI-A-1 12empty DP-1,DVI-D-1 12bus-ID 01:00.0
    12chip-ID 10de:21c4 12temp 27.0 C
  12Display wayland 12server X.org 12v 1.20.14 12with Xwayland 12v 23.2.4 12compositor kwin_wayland
    12driver N/A 12display-ID 0
  12Monitor-1 HDMI-A-1 12res 1920x1080 12size N/A
  12API EGL 12v 1.5 12platforms 12device 0 12drv nouveau 12device 1 12drv swrast 12gbm 12drv nouveau 12surfaceless
    12drv nouveau 12wayland 12drv nouveau 12x11 12drv nouveau
  12API OpenGL 12v 4.5 12compat-v 4.3 12vendor mesa 12v 23.3.5 12glx-v 1.4 12direct-render yes 12renderer NV168
    12device-ID 10de:21c4 12display-ID :0.0
  12API Vulkan 12v 1.3.268 12surfaces xcb,xlib,wayland 12device 0 12type cpu 12driver mesa llvmpipe
    12device-ID 10005:0000
1 Like

It looks like you have a discrete nvidia GPU and an integrated AMD GPU and the nvidia GPU is using the default nouveau driver. It’s possible that this application is starting with DRI_PRIME=1 and nouveau is crashing. If that’s the case, then installing the nvidia drivers from RPMFusion might help.

1 Like

OK, after having installed the nvidia drivers, and some pushed KDE updates, the error has some differences from first time around, but whether they are substantial or not are impossible for me to say:

$ flatpak run org.standardnotes.standardnotes
[3:0208/091944.743054:ERROR:bus.cc(407)] Failed to connect to the bus: Failed to connect to socket /run/dbus/system_bus_socket: No such file or directory
LaunchProcess: failed to execvp:
xdg-settings
[3:0208/091948.625522:ERROR:bus.cc(407)] Failed to connect to the bus: Failed to connect to socket /run/dbus/system_bus_socket: No such file or directory
[3:0208/091948.625564:ERROR:bus.cc(407)] Failed to connect to the bus: Failed to connect to socket /run/dbus/system_bus_socket: No such file or directory
Checking for beta autoupdate feature for deb/rpm distributions
Found package-type: deb
09:19:49.053 › Checking for update
extServer: Server started at http://127.0.0.1:45653
[46:0208/091949.256769:ERROR:gbm_pixmap_wayland.cc(82)] Cannot create bo with format= RGBA_8888 and usage=SCANOUT
[46:0208/091949.258115:ERROR:gbm_pixmap_wayland.cc(82)] Cannot create bo with format= RGBA_8888 and usage=GPU_READ
[46:0208/091949.258238:ERROR:shared_image_factory.cc(926)] CreateSharedImage: could not create backing.
[46:0208/091949.258371:ERROR:shared_image_factory.cc(758)] DestroySharedImage: Could not find shared image mailbox
[46:0208/091949.258501:ERROR:gpu_service_impl.cc(1089)] Exiting GPU process because some drivers can't recover from errors. GPU process will restart shortly.
[3:0208/091949.274498:ERROR:gpu_process_host.cc(991)] GPU process exited unexpectedly: exit_code=8704
[88:0208/091949.351681:ERROR:gbm_pixmap_wayland.cc(82)] Cannot create bo with format= RGBA_8888 and usage=SCANOUT
[88:0208/091949.352845:ERROR:gbm_pixmap_wayland.cc(82)] Cannot create bo with format= RGBA_8888 and usage=GPU_READ
[88:0208/091949.352896:ERROR:shared_image_factory.cc(926)] CreateSharedImage: could not create backing.
[88:0208/091949.352938:ERROR:shared_image_factory.cc(758)] DestroySharedImage: Could not find shared image mailbox
[88:0208/091949.353012:ERROR:gpu_service_impl.cc(1089)] Exiting GPU process because some drivers can't recover from errors. GPU process will restart shortly.
[3:0208/091949.374346:ERROR:gpu_process_host.cc(991)] GPU process exited unexpectedly: exit_code=8704
[100:0208/091949.457046:ERROR:gbm_pixmap_wayland.cc(82)] Cannot create bo with format= RGBA_8888 and usage=SCANOUT
[100:0208/091949.458239:ERROR:gbm_pixmap_wayland.cc(82)] Cannot create bo with format= RGBA_8888 and usage=GPU_READ
[100:0208/091949.458302:ERROR:shared_image_factory.cc(926)] CreateSharedImage: could not create backing.
[100:0208/091949.458352:ERROR:shared_image_factory.cc(758)] DestroySharedImage: Could not find shared image mailbox
[100:0208/091949.458432:ERROR:gpu_service_impl.cc(1089)] Exiting GPU process because some drivers can't recover from errors. GPU process will restart shortly.
[3:0208/091949.474900:ERROR:gpu_process_host.cc(991)] GPU process exited unexpectedly: exit_code=8704
[79:0208/091949.481495:ERROR:command_buffer_proxy_impl.cc(127)] ContextResult::kTransientFailure: Failed to send GpuControl.CreateCommandBuffer.
[79:0208/091949.598739:ERROR:command_buffer_proxy_impl.cc(127)] ContextResult::kTransientFailure: Failed to send GpuControl.CreateCommandBuffer.
09:19:50.222 › Checking for update (already in progress)
09:19:50.832 › Update for version 3.189.1 is not available (latest version: 3.187.3, downgrade is disallowed).

inxi output now looks like this:

$ inxi -CMSGxx
12System:
  12Kernel 6.7.3-200.fc39.x86_64 12arch x86_64 12bits 64 12compiler gcc 12v 2.40-14.fc39 12Desktop KDE Plasma
    12v 5.27.10 12tk Qt 12v 5.15.12 12wm kwin_wayland 12dm SDDM 12Distro Fedora release 39 (Thirty Nine)
12Machine:
  12Type Desktop 12System LENOVO 12product 90RW005PUL 12v IdeaCentre Gaming5 14ACN6
    12serial <superuser required> 12Chassis 12type 3 12serial <superuser required>
  12Mobo LENOVO 12model 3743 12v NOK 12serial <superuser required> 12UEFI LENOVO 12v O5EKT1AA 12date 06/10/2021
12CPU:
  12Info 6-core 12model AMD Ryzen 5 5600G with Radeon Graphics 12bits 64 12type MT MCP 12arch Zen 3 12rev 0
    12cache 12L1 384 KiB 12L2 3 MiB 12L3 16 MiB
  12Speed (MHz) 12avg 616 12high 3001 12min/max 400/4464 12cores 121 3001 122 400 123 400 124 400 125 400 126 400 127 400
    128 400 129 400 1210 400 1211 400 1212 400 12bogomips 93427
  12Flags avx avx2 ht lm nx pae sse sse2 sse3 sse4_1 sse4_2 sse4a ssse3 svm
12Graphics:
  12Device-1 NVIDIA TU116 [GeForce GTX 1660 SUPER] 12vendor Micro-Star MSI 12driver nouveau 12v kernel
    12arch Turing 12pcie 12speed 2.5 GT/s 12lanes 16 12ports 12active HDMI-A-1 12empty DP-1,DVI-D-1 12bus-ID 01:00.0
    12chip-ID 10de:21c4 12temp 25.0 C
  12Display wayland 12server X.org 12v 1.20.14 12with Xwayland 12v 23.2.4 12compositor kwin_wayland
    12driver N/A 12display-ID 0
  12Monitor-1 HDMI-A-1 12res 1920x1080 12size N/A
  12API EGL 12v 1.5 12platforms 12device 0 12drv nouveau 12device 1 12drv swrast 12gbm 12drv nouveau 12surfaceless
    12drv nouveau 12wayland 12drv nouveau 12x11 12drv nouveau
  12API OpenGL 12v 4.5 12compat-v 4.3 12vendor mesa 12v 23.3.5 12glx-v 1.4 12direct-render yes 12renderer NV168
    12device-ID 10de:21c4 12display-ID :0.0
  12API Vulkan 12v 1.3.268 12surfaces xcb,xlib,wayland 12device 0 12type cpu 12driver mesa llvmpipe
    12device-ID 10005:0000

Shouldn’t nouveau be blacklisted? I would have loved to check everything nvidia installed, however, there doesn’t appear to be an rpm-ostree version of dnf list installed '*nvidia*' My nvidia install command was sudo rpm-ostree install akmod-nvidia.

The default nvidia driver installed that way will not load if secure boot is enabled.
To allow it to load you must either disable secure boot or sign the module so it can be loaded.
To determine if that is a secure boot issue the command mokutil --sb-state should tell you if secure boot is enabled or disabled.

1 Like

See the docs for NVIDIA users: Troubleshooting :: Fedora Docs

I recommend looking at Universal Blue NVIDIA images: Nvidia - Universal Blue

2 Likes

Thanks to all for the helpful replies. I will sidestep the issue for now, but will return to it when I have more experience with Silverblue in its various iterations. For now, I will Install Kinoite on a computer, with Radeon graphics.

Second universal blue. The moment you do manual changes like overrides and blocklisting drivers, immutable distros get less stable.

Better share the struggle by using an image created for you.

Noted. When I take a look at the Nvidia drivers again, I will study the options. I get where you are coming from regarding the advice, but from what I consider to be the viewpoint of a fairly average Linux desktop user, while rebasing looks relatively easy, it feels like a disenfranchisement.

Sooo maybe the whole atomic/immutable thing isn’t really up to snuff - yet. On the other hand, one could argue that the disenfranchisement already takes place as soon as a consumer buys a computer system that requires proprietary drivers,. The difference is that it hasn’t been much of an issue for many years. There have been numerous work arounds…

It still comes down to the lack of readily available open source hardware for consumer electronics such as mobile phones and computers, and that isn’t something that Linux distributions can easily do something about.

I had time to read through the basic documentation, so I did a fresh install of Kinoite and did a test rebase to Onyx, as well as an rpm-ostree rollback. All was easy and without problems.

However, if I understand this correctly, to rebase to a universal blue image of Kinoite with Nvidia drivers, I need to accept installing unsigned images. I have been looking into Fedora Silverblue and it’s iterations to improve security. I suppose that I could try to roll my own, but I am definitely not there yet, so I guess Silverblue on machines with Nvidia cards is a no go for me.

1 Like

Not quite true.
The os image is signed.
It is easy to sign the nvidia modules when the system builds them so they would be signed and usable with secure boot enabled. The steps to do that are shown in detail in the file /usr/share/doc/akmods-README.secureboot which is placed on the system when the user installs the akmods package.
The procedure I recommend is

  1. install the akmods package then reboot
  2. perform the steps in the README file above, which also includes a reboot.
  3. install the nvidia drivers from rpmfusion, wait 5 minutes, then one more reboot to load the (now signed) drivers.

If the nvidia drivers were installed before completing step 2 the driver modules will be built at the time of the install and will remain unsigned so that involves a removal of the unsigned drivers then a rebuild and install of the signed drivers.
On Workstation I do that with

  1. sudo dnf remove kmod-nvidia-*
    followed by
  2. sudo akmods --force
    to rebuild the packages just removed and create the now signed driver modules

The commands are a little different on Kinoite but the process remains the same.

1 Like

I was referring to universal blue images with nvidia drivers preinstalled. I have tried installing the nvidia drivers from RPM Fusion, but that method is apparently unstable with Silverblue and as a result didn’t solve my issue with the Standard Notes Flatpak.

No comment to supposedly signed images producing the following error?

$ rpm-ostree rebase ostree-image-signed:docker://ghcr.io/ublue-os/kinoite-nvidia:39
Pulling manifest: ostree-image-signed:docker://ghcr.io/ublue-os/kinoite-nvidia:39
error: Preparing import: Fetching manifest: containers-policy.json specifies a default of insecureAcceptAnything; refusing usage

I’m curious if you have similar issues with Fedora Onyx? Enabling 3rd party repos enables RPMFusion and the nVidia driver there. I know it’s not quite the same desktop environment as Silverblue, but it seems there are some slight out of box differences between the Atomic Desktops, such as Onyx having automatic updates enabled by default, too.

1 Like

Wow. Thanks for the tip. I will give it a try and post back.