gnome-shell crashes (fatal error logged in create_fallback_offscreen) when attempting to enable second monitor on second GPU (Nvidia >= 495) in a Wayland session

Yeah having 'meta_monitor_mode_foreach_crtc' in the stack trace supports your bug description well, but we need a more detailed stack trace yet. Maybe try installing the libmutter debug symbols and then crashing it again?

Also keep in mind the configuration you want is known to not work: bug 1959888, bug 1963701.

Still, we also need to avoid crashes.

StacktraceSource:
 #0 0x00007f06fc89ea7c in ?? ()
 #1 0x000000000000022e in ?? ()
 #2 0x000000000000022f in ?? ()
 #3 0x0000000000000000 in ?? ()
StacktraceTop:
 ?? ()
 ?? ()
 ?? ()
 ?? ()

Thank you for your report!

However, processing it in order to get sufficient information for the
developers failed (it does not generate a useful symbolic stack trace). This
might be caused by some outdated packages which were installed on your system
at the time of the report:

no debug symbol package found for libnvidia-gl-495
no debug symbol package found for libjpeg-turbo8
no debug symbol package found for libnvidia-extra-495
no debug symbol package found for libcrypt1
no debug symbol package found for libgdk-pixbuf-2.0-0

Please upgrade your system to the latest package versions. If you still
encounter the crash, please file a new report.

Thank you for your understanding, and sorry for the inconvenience!

That's unfortunate. Sounds like it might be bug 1964120 though.

Let's try again. Same issue in bug 1965246.

From your ShellJournal.txt:

[ 137.444673] Extravaganza gnome-shell[4141]: Created gbm renderer for '/dev/dri/card1'
[ 137.444965] Extravaganza gnome-shell[4141]: Failed to allocate onscreen framebuffer for /dev/dri/card1: Failed to create gbm_surface: Operation not permitted
[ 137.445019] Extravaganza gnome-shell[4141]: GNOME Shell crashed with signal 5
[ 137.445019] Extravaganza gnome-shell[4141]: == Stack trace for context 0x55f3af3d0300 ==
[ 137.445096] Extravaganza gnome-shell[4141]: Failed to create fallback offscreen framebuffer: Failed to create texture 2d due to size/format constraints

Would you like a full backtrace? I can install debug symbols and trigger this.

Using the main branches:

#0 0x00007f3255a9fccf in g_log_structured_array ()
    at /lib/x86_64-linux-gnu/libglib-2.0.so.0
#1 0x00007f3255a9ff99 in g_log_default_handler ()
    at /lib/x86_64-linux-gnu/libglib-2.0.so.0
#2 0x00005618851da1fd in default_log_handler
    (log_domain=0x7f3254e5b2c6 "mutter", log_level=6, message=0x5618885d6810 "Failed to create fallback offscreen framebuffer: Failed to create texture 2d due to size/format constraints", data=0x0) at ../src/main.c:297
#3 0x00007f3255aa13fa in g_logv () at /lib/x86_64-linux-gnu/libglib-2.0.so.0
#4 0x00007f3255aa16e3 in g_log () at /lib/x86_64-linux-gnu/libglib-2.0.so.0
#5 0x00007f3254dfec77 in create_fallback_offscreen
    (renderer_native=0x561885ea18f0, cogl_context=0x5618861d2750, width=1920, height=1200) at ../src/backends/native/meta-renderer-native.c:1213
#6 0x00007f3254dfef76 in meta_renderer_native_create_view
    (renderer=0x561885ea18f0, logical_monitor=0x561886d4f540, output=0x56188865d4a0, crtc=0x7f320018a610) at ../src/backends/native/meta-renderer-native.c:1288
#7 0x00007f3254c9a75d in meta_renderer_create_view
    (renderer=0x561885ea18f0, logical_monitor=0x561886d4f540, output=0x56188865d4a0, crtc=0x7f320018a610) at ../src/backends/meta-renderer.c:102
#8 0x00007f3254c9a7cd in create_crtc_view
    (logical_monitor=0x561886d4f540, monitor=0x561888572e80, output=0x56188865d4a0, crtc=0x7f320018a610, user_data=0x561885ea18f0)
    at ../src/backends/meta-renderer.c:133
#9 0x00007f3254c7ec1b in foreach_crtc
    (monitor=0x561888572e80, mode=0x56188772e740, monitor_crtc_mode=0x56188621d670, user_data=0x7fffdc9b9760, error=0x0)
    at ../src/backends/meta-logical-monitor.c:247
#10 0x00007f3254c82d04 in meta_monitor_mode_foreach_crtc
    (monitor=0x561888572e80, mode=0x56188772e740, func=0x7f3254c7ebb4 <foreach_crtc>, user_data=0x7fffdc9b9760, error=0x0)
    at ../src/backends/meta-monitor.c:1955
#11 0x00007f3254c7ecb5 in meta_logical_monitor_foreach_crtc
    (logical_monitor=0x561886d4f540, func=0x7f3254c9a78d <create_crtc_view>, user_data=0x561885ea18f0) at ../src/backends/meta-logical-monitor.c:274
#12 0x00007f3254c9a8ee in meta_renderer_real_rebuild_views
    (renderer=0x561885ea18f0) at ../src/backends/meta-renderer.c:168
#13 0x00007f3254dff4c8 in meta_renderer_native_rebuild_views
    (renderer=0x561885ea18f0)
    at ../src/backends/native/meta-renderer-native.c:1422
#14 0x00007f3254c9a78b in meta_renderer_rebuild_views (renderer=0x561885ea18f0)
    at ../src/backends/meta-renderer.c:120
#15 0x00007f3254e0ab20 in meta_stage_native_rebuild_views
    (stage_native=0x5618860ff500)
    at ../src/backends/native/meta-stage-native.c:66
#16 0x00007f3254dc7e1b in meta_backend_native_update_screen_size
    (backend=0x561885ded120, width=5043, height=3360)
    at ../src/backends/native/meta-backend-native.c:411
#17 0x00007f3254c6d7e7 in meta_backend_sync_screen_size
    (backend=0x561885ded120) at ../src/backends/meta-backend.c:285
#18 0x00007f3254c6da07 in meta_backend_monitors_changed (backend=0x561885ded120) at ../src/backends/meta-backend.c:356
#19 0x00007f3254c946a3 in meta_monitor_manager_notify_monitors_changed (manager=0x561885df0340) at ../src/backends/meta-monit...

It appears my work in progress avoids the crash nicely:

https://gitlab.gnome.org/vanvugt/mutter/-/commits/fall-back-from-secondary-gpu

Although it's slow -- it works by avoiding Nvidia GBM when the driver refuses to work in a hybrid setup, and just uses Nvidia KMS to scan out dumb buffers rendered by the primary GPU instead... but it appears that's still faster than using Xorg!?

Long term we are hoping for an Nvidia driver that doesn't fail GBM in hybrid setups.

Thanks for your attention to this bug, Daniel.

For the record, I don't think a set up like this is particularly weird; I'm seeing this issue on a laptop with embedded NVIDIA graphics. The "second video card", in my case, is the integrated Intel graphics card.

Also, I'm not sure if this is related, but yesterday I saw a regression that caused my second monitor to fail; the only workaround I found was to downgrade to nvidia-driver-470 from nvidia-driver-510. (That is, nvidia-driver-510 wasn't causing me too much trouble when I first installed Jammy from the March 21st image; it took a day or two of updating, then started causing me issues yesterday .)

Certainly using nvidia-driver-470 will avoid this crash. Although choosing 'Ubuntu on Xorg' at the login screen with nvidia-driver-510 should avoid it too.

P.S. GNOME Shell refers to the integrated GPU (Intel) as primary (because it's attached to the screen), and the more powerful GPU (usually not wired to the screen) is the secondary.

The proposed fix is https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/2341

Fixed in mutter 42.1 but we should probably patch it in sooner than that for jammy.

Distro patch proposed in https://salsa.debian.org/gnome-team/mutter/-/merge_requests/80

Just for the record, I see the same issue with a Dell Precision 7760 (CID 202103-28858) when running 22.04 beta on Wayland. I was told the HDMI connector was wired directly to the Nvidia GPU, so gnome-shell crashes when I connect an external monitor there, but even after logging back in, the monitor cannot be used.

Plugging the same monitor to the USB-C connector (which is wired to Intel GPU apparently) works like a charm...

This bug was fixed in the package mutter - 42.0-3ubuntu1

---------------
mutter (42.0-3ubuntu1) jammy; urgency=medium

  [ Marco Trevisan (Treviño) ]
  * debian/patches: Cherry-pick various upstream commits for 42.1:
    + onscreen/native: Fall back if COPY_MODE_SECONDARY_GPU fails to init
      (LP: #1964037, #1959888)
  * debian/patches: Allow any drag timestamp as drag start serial (LP: #1964541)
  * debian/patches: Move x11-fractional scaling under ubuntu namespace
  * debian/patches: Account ClutterStage grabs on Wayland key focus sync
    (LP: #1964442)
  * debian/libmutter-10-0.symbols: Update including new internal symbols
  * debian/patches: Avoid memory errors when comparing gamma values
  * debian/patches: Fix privacy-screen and connectors updates with
    triple-buffering (LP: #1966178)
  * debian/patches: Ensure privacy screen settings are applied on startup
    (LP: #1966178)

  [ Jeremy Bicha ]
  * Merge from Debian unstable. Remaining changes:
    - debian/patches/x11-Add-support-for-fractional-scaling-using-Randr.patch:
      + X11: Add support for fractional scaling using Randr
    - Add triple-buffering patch
    - debian/libmutter-10-0.symbols: Add symbols for triple buffering patch
    - ubuntu/wayland-data-device-Allow-any-drag-timestamp-as-drag-star.patch:
      + handle DnD for old (snapped) gtk apps that used wrong wayland serials

mutter (42.0-3) unstable; urgency=medium

  [ Marco Trevisan (Treviño) ]
  * debian/patches: Cherry-pick upstream fixes targeting 42.1
  * debian/patches: Ensure repick happens on actors visibility changed
    (LP: #1964545)

  [ Jeremy Bicha ]
  * Release to unstable

 -- Marco Trevisan (Treviño) <email address hidden> Wed, 13 Apr 2022 03:30:39 +0200

Confirmed the gnome will not crash but the system will back to gdm login screen in 1st hot-plugging.

Report the bug https://bugs.launchpad.net/oem-priority/+bug/1969121

Won't Fix for the Nvidia drivers because the problem is properly fixed. The remaining issue is just a performance one: bug 1970291