Public bug reported: Hardware - HP OMEN by HP Laptop 16-c0xxx (SKU 4S1A5EA#UUW, board 8912), BIOS F.19 (01/09/2023) - AMD Ryzen 9 5900HX, iGPU Cezanne [1002:1638] (DCN 2.1) drives the internal eDP panel (2560x1440) - dGPU NVIDIA RTX 3070 Mobile [10de:249d] (not involved, see below) - Ubuntu 26.04.1 LTS (upgraded from 24.04)
Problem With every 7.0 kernel I have tried, the internal laptop panel stays black after amdgpu takes over from the firmware framebuffer. The Ubuntu boot logo is visible (firmware/simpledrm), then the panel goes black and never comes back. External monitor works; the GDM login is placed on the (black) internal panel. Closing and reopening the lid does not recover it: every attempt to enable the panel logs the same error. amdgpu 0000:06:00.0: [drm] enabling link 0 failed: 15 (15 = link training failure). No other amdgpu errors are printed. Kernels - 6.17.0-22 / -29 / -35 / -40 (Display Core v3.2.340): panel works, 0 failures in 12+ boots since June 2026 - 7.0.0-28, 7.0.0-34, 7.0.0-38 (Display Core v3.2.369): panel black on nearly every boot (16 of 20 boots), first seen with 7.0.0-28 in July 2026 - Same result with the 24.04 HWE build and the 26.04 build of 7.0.0-38 Things ruled out - NVIDIA driver: fails with module_blacklist=nvidia,nvidia_drm,nvidia_modeset,nvidia_uvm (attached dmesg is from that boot; no NVIDIA module loaded) - External monitor / Dell WD19 dock: fails with everything unplugged - Firmware: fails with the 24.04 green_sardine firmware placed in /lib/firmware/updates (DMCUB 0x0101002B is identical in both) - Boot parameters: fails without crashkernel=, with vt.handoff=7, with nvidia-drm.fbdev=0, and with amdgpu.dcdebugmask=0x10 (PSR disabled) Possibly related Very similar reports on other Renoir/Cezanne (DCN 2.1) laptops with 7.0 kernels: - https://github.com/CachyOS/linux-cachyos/issues/832 – eDP fails to initialize on 7.0; the reporter traced it to dmub_srv_wait_for_idle() switching to a progressive backoff, which makes the 100 ms timeout expire after ~11 ms, while the eDP panel power-on DMUB command (VBIOS_LVTMA_CONTROL) takes ~180 ms on that hardware. - https://github.com/CachyOS/linux-cachyos/issues/831 – same "enabling link 0 failed: 15" on a Ryzen 7 5825U; fixed in CachyOS by reverting that change (commit 0bb166ba01aa1b35765d4af57a9345f73bdf5cec in their tree). I don't see the DMCUB diagnostic errors from those reports in my log, so I'm not certain it is the same root cause. Workaround Booting the 24.04 6.17.0-40 kernel. Attachments - dmesg-7.0.0-38-black-panel-nvidia-blocked.txt (failing boot, NVIDIA blocked) - dmesg-6.17.0-40-working.txt (same machine, working) - boot-history.txt (kernel and time of the first eDP link failure for each boot since May 2026) - lspci.txt I'm happy to test patched or -proposed kernels. ** Affects: linux (Ubuntu) Importance: Undecided Status: New ** Tags: amdgpu resolute ** Attachment added: "dmesg from failing 7.0.0-38 boot (NVIDIA modules blocked), panel black" https://bugs.launchpad.net/bugs/2172124/+attachment/6012173/+files/dmesg-7.0.0-38-black-panel-nvidia-blocked.txt -- You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. https://bugs.launchpad.net/bugs/2172124 Title: [regression] amdgpu: internal eDP panel stays black on 7.0 kernels ("enabling link 0 failed: 15") – HP OMEN 16-c0xxx, Ryzen 9 5900HX (Cezanne, DCN 2.1); 6.17 works To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/2172124/+subscriptions -- ubuntu-bugs mailing list [email protected] https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
