615.78.08 (open): external HDMI output stays black with GPU pegged at 99-100% from the first frame under Hyprland (RTX 3050 Ti Laptop); fully resolved by downgrading to 610.57.04

オープン
  • 何か問題がありますか？
#1,460 コメント 3 件 リアクション 0 件 担当者 0 名 GitHub で見る

メンテナーはふだん 1 日以内に返信

まだ誰も着手していません。

評価

難易度
4/5
見積もり時間
3〜5日
初心者へのやさしさ
36/100
issue の種類
バグ
明瞭さ
おおむね明確
活発さ
活発
技術スタック
c, linux
領域
backend, operating-systems

調査の方向性

Start by comparing the reported working 610.57.04 and failing 615.78.08 behavior in the open GPU kernel module source; the issue does not identify a file, failing test, or specific subsystem. Reproduce with the listed RTX 3050 Ti, HDMI display, and Hyprland setup, then collect nvidia-bug-report.sh output from the failing state as the report offers. Done means identifying and fixing the driver regression so HDMI displays correctly and the GPU idles.

索引モデルが issue の本文から書いたものです。

説明

Describe the bug

On nvidia-open-dkms 615.78.08, the external HDMI monitor stays solid black from the very first frame of a fresh Hyprland session, while the internal laptop panel (eDP) works normally. The GPU immediately pegs to 99-100% utilization and ~87°C, held steady, with no recovery. There are no Xid errors, no NVRM errors, and no DRM/kernel error messages at all — a completely silent failure. The kernel and the compositor both believe the HDMI output is connected, enabled, and running the correct mode; only the physical picture is missing.

To Reproduce
  1. Run nvidia-open-dkms 615.78.08 with an HDMI monitor attached (hybrid Intel+NVIDIA laptop, HDMI wired to the dGPU).
  2. Boot or log into a fresh Hyprland session.
  3. External HDMI monitor shows black immediately; nvidia-smi shows GPU utilization pegged at 99-100%, power draw ~45W, temperature climbing to ~87°C and holding.

Reproduced identically across two independent fresh sessions (one from cold boot, one from a full logout/login) — not a transient glitch.

Expected behavior

External HDMI output displays normally; GPU idles when the desktop is idle.

Bisection
  • 615.78.08: broken every time, as described above.
  • 610.57.04: same kernel, same Hyprland build, no other changes — HDMI works correctly, GPU idles at ~20% util / ~62°C, Hyprland process CPU usage normal (~10%). Downgrading via pacman -U to the cached 610.57.04 package and rebooting fully resolves the issue.
System information
  • GPU: NVIDIA GeForce RTX 3050 Ti Laptop GPU (GA106M, PCI 10de:2523)
  • Driver: nvidia-open-dkms 615.78.08 (broken) / 610.57.04 (working)
  • Kernel: 7.2.8-9-omarchy (Arch-based, x86_64)
  • Compositor: Hyprland 0.56.2 (Wayland)
  • Distro: Omarchy Linux 4.0.5
  • Laptop: ASUS ROG Zephyrus M16 (GU603ZEB), hybrid Intel iGPU + NVIDIA dGPU via supergfxd (Hybrid mode), HDMI wired to the dGPU
  • External monitor: LG TV (SSCR2), 3840x2160@60

Happy to attach nvidia-bug-report.sh output from the broken (615.78.08) state if that would help — let me know and I'll reboot back onto it to capture one.

Filed by Claude Sonnet 5 via Claude Code.

主要言語
C
スター
17.5k
フォーク
1.9k
30日以内にマージされた PR はありません

環境構築

はじめの一歩

  1. issue を最後まで読み、次にプロジェクトのコントリビューションガイドを読みます。
  2. 着手することを issue にコメントします — 二人が同じ作業をするのを防げます。
  3. リポジトリをフォークし、ブランチを切って変更します。
  4. issue 番号を参照したプルリクエストを送ります。

NVIDIA/open-gpu-kernel-modules のほかの issue

NVIDIA/open-gpu-kernel-modules の issue をすべて見る

似ている issue

C の issue をもっと見る

新しい issue をメールで受け取る

初心者向けの GitHub issue を短くまとめたダイジェスト。