bug-2404a.txt -- 2024-04-07 -- Video on Diamond is flaky. Subject: Video black switching framebuffer at boot (kernel 6.8.1..4, Tiger Lake) This bug is fixed in kernel 6.8.5. Please take it as a thank-you to the developers, or a warning that a bug fixed by accident has a higher risk of regression. Symptom: Diamond, the host having problems, resumes from S4 (hibernation) or reboots and shows kernel boot messages on VT1 (640x480px) using probably the EFI framebuffer. At about the time that it switches to "colour frame buffer device 100x37" the monitor goes black, then sleeps, interpreted as the video signal being shut off. No sign of anything wrong in syslog or /var/log/boot.msg or /var/log/Xorg.0.log . My tester for display-manager passes it, the clients think X-Windows is working fine. If you unplug the monitor's HDMI cable, wait a few seconds, then replug, the monitor starts performing the video signal and will continue until Diamond hibernates or reboots (on one monitor). Description of Diamond: System model: Intel NUC11PAHi5, purchased about 2021-11-18 CPU: Intel Core i5-1135G7 4 cores 2.4GHz (turbo 4.2GHz). Tiger Lake. CPU family 6; stepping 1; microcode 0xb6 Graphics: Intel Iris Xe, output to HDMI-2.0b, and others not used. Memory: 16Gb Booter: UEFI, secure boot is disabled. The problem was first noticed shortly after 2024-04-02. Relevant package versions, with installation dates: xf86-video-intel-2.99.917.916_g31486f40-6.2.x86_64 2024-04-02 xorg-x11-server-21.1.12-1.1.x86_64 2024-04-09 kernel-default-6.7.9-1.1.x86_64 2024-03-19 kernel-default-6.8.1-1.2.x86_64 2024-04-02 kernel-default-6.8.4-1.1.x86_64 2024-04-09 Tidbits from /var/log/Xorg.0.log etc: . Integrated Graphics Chipset: Intel(R) HD Graphics. . X is using the "intel" driver at /usr/lib64/xorg/modules/drivers/intel_drv.so. GLX extension initted, also AIGLX and DRI2. . It's switching to mode 1680x1050@60.0 on DP2. HDMI is allegedly disconnected. But the physical connector is HDMI. Alibis of the usual suspects: I exchanged the following two monitors and the problem happens on both. Diamond's monitor: Dell 2007WFPb, 20.0in = 50.8cm diag, likely purchased in 2007 or 2008. IPS panel. Connection: DVI on monitor to HDMI on Diamond. If you un/replug the HDMI cable the video will come back. Holly's monitor: HP Pavilion 21.5in IPS HDMI 1920x1080px @60Hz, $90. Model number: HP 22cwa. Bought 2018-01-16, first on Amazon: 2015-10-11. Connection is HDMI on both ends. If you un/replug the HDMI cable the video stays invisible. Has Diamond's SoC (CPU + graphics) gone bad? If I boot kernel-default-6.7.9 (keeping the updated Xorg server and Intel driver), the problem isn't there, like before 2024-04-02, adding confidence in the hardware and pointing the finger of blame at the kernel, specifically at the modules i915, drm_buddy, or drm_display_helper. Evidence for involvement is, after booting any of the listed kernels, in /var/log/boot.msg we see "[drm] Initialized simpledrm 1.0.0 20200625" immediately followed by "Console: switching to colour frame buffer device 100x37" (problem appears on 6.8.1 or 6.8.4). Phoronix (Michael Larabel, 2024-01-11) reports a lot of improements in DRM, scheduled to go into the 6.8 revision. Results of a search on Google for "linux kernel 6.8 drm intel" -- just one similar bug report: https://bbs.archlinux.org/viewtopic.php?id=294593 Title: Video stops at start of boot with linux-6.8.2-arch2-1 OP replabrobin, 2024-04-05, paraphrased by jimc: "my display blanks after one or two lines of boot messages..." This is linux-6.8.2.arch2-1 reverting to linux-6.5.9-arch2-1 fixes the problem. OP believes the hardware is Intel Tigerlake Iris Xe GT2. Supplementary files may be found at http://www.jfcarter.net/~jimc/bug-intelvideo-2404/ /var/log/boot.msg /var/log/xorg.0.log /proc/modules --------- Extras, not part of bug report: https://bbs.archlinux.org/viewtopic.php?id=294593 Title: Video stops at start of boot with linux-6.8.2-arch2-1 OP replabrobin, 2024-04-05 "my display blanks after one or two lines of boot messages..." This is linux-6.8.2.arch2-1 ; reverting to linux-6.5.9-arch2-1 fixes the prob. I believe the hardware is Intel Tigerlake Iris Xe GT2. V1del (forum moderator) asks him to dump his journal and post it, or put in a pastebin and post the link, with this command line: sudo journalctl -b | curl -F 'file=@-' 0x0.st It shows the link: http://0x0.st/Xib8.txt They see several exceptions in the journal. Try installing 6.8.4. Also try kernel param: i915.enable_fbc=0 replabrobin responds, 6.8.4 didn't help, i915.enable_fbc=0 didn't help. Linux kernel mailing list: "Old" subscribe link: http://vger.kernel.org/vger-lists.html (choose subsystem) "New" subscribe link: https://subspace.kernel.org/vger.kernel.org.html Old instructions: mail to: majordomo@vger.kernel.org Empty subject. Body: subscribe linux-bluetooth (replace with desired subsystem) Server should send a welcome/confirmation e-mail. To send content: Send to linux-bluetooth@vger.kernel.org Subject: (you need one) Body: usual rules for bug reports or whatever