Public bug reported: This report was generated by an AI assistant using the user's observations and locally captured diagnostics. The user confirmed the freeze and the recovery. The root cause has not been established.
Summary One external screen freezes during ordinary desktop use while the laptop screen and the other external screen continue working. Switching to a text console and back restored the screen without reconnecting a cable, logging out, or rebooting. This may be related to bug 2162087, but the hardware differs and the cause has not been confirmed: https://bugs.launchpad.net/ubuntu/+source/mutter/+bug/2162087 Hardware and software - Lenovo ThinkPad X1 Carbon Gen 13, machine type 21NXCTO1WW - Intel Core Ultra 7 265U / Arrow Lake-U integrated graphics, PCI 8086:7d41, subsystem 17aa:2340 - Active graphics driver: i915 - Ubuntu 24.04.5 LTS, Wayland session - Kernel: 6.17.0-1033-oem (Ubuntu 6.17.0-1033.33) - GNOME Shell: 46.0-0ubuntu6~24.04.15 - libmutter-14-0: 46.2-1ubuntu0.24.04.18 - Mesa libgl1-mesa-dri: 25.2.8-0ubuntu0.24.04.4 - Intel graphics firmware package: 20240318.git3b128b60-0ubuntu3.1 - Built-in display: 1920x1200 at approximately 60 Hz - Two Sceptre F22 external screens: each 1920x1080 at 60 Hz, scale 1.0 - During this incident the external screens used DP-1 and DP-3; the built-in HDMI connector was disconnected. An Anker USB-C hub and another USB-C video adapter were present. - Extensions enabled: Ubuntu Dock, Desktop Icons NG, Ubuntu Tiling Assistant. A session without these extensions has not yet been tested. Reproduction and expected behavior The problem is intermittent during normal work, not limited to waking from sleep. There are no reliable reproduction steps yet. Expected: all three screens continue updating and accepting pointer movement. Similar display problems have occurred with other monitors and connection arrangements, but this report focuses on the captured October 28 incident. Captured incident: October 28, 2026, approximately 09:32-09:37 America/New_York (EDT, UTC-04:00) Only the far-right external monitor froze. The built-in screen and middle external monitor remained usable. The frozen screen stayed frozen throughout the diagnostic capture and subsequent live queries. 1. Both external monitors remained connected and enabled, with valid EDID checksums and the expected 1080p/60 Hz configuration. 2. During the freeze, DP-1 was assigned to pipe C / CRTC 202, with framebuffer 319 selected. DP-3 used pipe B / CRTC 145. 3. In the two protected driver snapshots, the working DP-3 output changed its selected framebuffer from 323 to 321. DP-1 stayed on framebuffer 319. 4. Further read-only libdrm queries (drmModeGetCrtc and drmCrtcGetSequence) sampled once per second for four seconds at about 09:35. DP-1 kept framebuffer 319 while its sequence advanced 13930823, 13930883, 13930943, 13931003, 13931063. The working DP-3 output continued changing framebuffer IDs. These counts show continuing refresh timing; they do not establish that the physical monitor received correct pixels. An unchanged framebuffer ID alone does not prove that no new rendering occurred. 5. /sys/class/drm/card1/error reported "No error state collected". No graphics error appeared in the recent kernel log around the capture. GNOME logged other warnings, but no confirmed error identifying the freeze. Missing errors do not rule out a driver problem. 6. The built-in panel had PSR2 enabled and reported DEEP_SLEEP. Framebuffer compression was already disabled. These observations do not establish a power-saving cause for the external freeze. Recovery The desktop session was on tty2. Ctrl+Alt+F3 followed by Ctrl+Alt+F2 restored the frozen screen without unplugging either monitor. A query after recovery showed different framebuffer IDs and that the two external connectors had swapped hardware pipe assignments. What has and has not been tested No persistent graphics settings or driver changes were made for this investigation. An Xorg session, disabling triple buffering/direct scanout, and a session without extensions have not yet been tested. The data suggests investigating per-output frame delivery, but does not establish a specific Mutter or kernel fault. Diagnostic files were retained locally but are not attached to this report. The key measurements are transcribed above. ** Affects: gnome-shell (Ubuntu) Importance: Undecided Status: New ** Tags: amd64 noble wayland-session -- You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. https://bugs.launchpad.net/bugs/2172377 Title: One external display freezes on Intel Arrow Lake-U / Wayland; VT switch restores it To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/gnome-shell/+bug/2172377/+subscriptions -- ubuntu-bugs mailing list [email protected] https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
