惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

L
LangChain Blog
博客园_首页
钛媒体:引领未来商业与生活新知
钛媒体:引领未来商业与生活新知
月光博客
月光博客
S
SegmentFault 最新的问题
量子位
Apple Machine Learning Research
Apple Machine Learning Research
博客园 - 司徒正美
博客园 - Franky
Google DeepMind News
Google DeepMind News
Recent Announcements
Recent Announcements
B
Blog RSS Feed
C
Check Point Blog
The Cloudflare Blog
M
MIT News - Artificial intelligence
OSCHINA 社区最新新闻
OSCHINA 社区最新新闻
F
Fortinet All Blogs
Hugging Face - Blog
Hugging Face - Blog
博客园 - 叶小钗
V
Visual Studio Blog
V
V2EX
Microsoft Azure Blog
Microsoft Azure Blog
博客园 - 聂微东
freeCodeCamp Programming Tutorials: Python, JavaScript, Git & More

CachyOS Forum - Latest posts

Ran into a frozen black screen issue with latest update Niri Settings GUI Not Appearing Update to Gnome 50 and I'm getting black/blank screen on NVIDIA with hybrid graphics My JBL Quantum 350 headphones dont play sound 90% of the time on linux while on windows it plays audio 99% of the time Niri/Noctalia Shell notifications Getting a black screen after alt-tabbing too many times Rog-control-center Duplicate keyboard and mouse in system settings? Games running very poorly after first session (semi-solved, still need help though!) Amdgpu random hangs (does cachy perhaps plan to carry patches for this)? A little something. The adj script that replaces GNU fmt for non-ASCII text Mouse scroll wheel issues - ratchet Can't start steam after updates of today (17 April) Balena etcher or ventoy QEMU Bridge setup - Maybe already known but I found it pretty difficult to do and needed much reading Multi-boot install: what do I do about my /boot and /boot/efi partition? Do you collect all the solutions and post them on Wiki or Forum? System update - lts headers invalid or corrupted Sick to death of updates killing my games! I messed up my stuff again and cant download any thing now. error below Got Logitech MX4 with Haptic working in CachyOS Need help with the performance button on my Acer Nitro V16 AI Bailing out, you are at your own Troubles in the Starfield What is so special about the new Linux Kernel 7? "Issue: Display brightness is locked at a low level after waking from sleep on CachyOS (latest update) Balena etcher or ventoy Help with kernel update, emergency mode Issue happening only on wayland Display not recognised - low resolution
RX 9060 XT (RDNA 4) — Hard reboot only in Diablo 4 via Pr...
@goofballexp · 2026-05-10 · via CachyOS Forum - Latest posts

1

System Specs:

  • OS: CachyOS x86_64
  • Kernel: 7.0.4-1-cachyos
  • CPU: AMD Ryzen 5 7500F
  • GPU: PowerColor Radeon RX 9060 XT 16GB (Navi 44 / RDNA 4)
  • RAM: 32GB Corsair Vengeance DDR5 (CMH32GX5M2E6000C36) @ 4800MHz JEDEC
  • Motherboard: MSI PRO B650-S WIFI (BIOS 1.N2, March 2026)
  • PSU: CoolerMaster MWE Gold 850W
  • Display: 2x monitors via DisplayPort
  • DE: KDE Plasma 6.6.4 on Wayland

The Problem:
Diablo 4 launched via Lutris/Proton causes random hard system reboots with zero log trace. No MCE errors, no amdgpu driver timeout, no GPU reset message — the system just instantly dies and reboots. Other games (e.g. Borderlands 4) run completely stable at stock GPU clocks for hours with no issues.

What the logs show:
After extensive investigation, the previous boot logs showed:

amdgpu 0000:03:00.0: MODE1 reset
amdgpu 0000:03:00.0: GPU mode1 reset
amdgpu 0000:03:00.0: GPU smu mode1 reset

And on boot:

amdgpu 0000:03:00.0: [drm] REG_WAIT timeout 1us * 150000 tries - optc401_disable_crtc line:237

The Diablo 4 Fenris engine log simply cuts off mid-session with no error — the OS dies underneath the game before it can write anything.

What I’ve already tried:

  • Updated BIOS to latest (1.N2, March 2026 — fixed initial data fabric sync flood)
  • Added kernel parameters: amdgpu.dcdebugmask=0x10 amdgpu.runpm=0
  • GPU power cap via LACT (130W) — reduces frequency but doesn’t eliminate crashes
  • Lutris env vars: RADV_DEBUG=nomeshshaders, VKD3D_CONFIG=no_upload_hvv, VKD3D_LIMIT_TGSM_ALLOCATIONS=1, RADV_PERFTEST=gpl
  • Proton version: GE-Proton10-34 (pinned)
  • DXVK enabled, VKD3D handling DX12
  • RAM running at stock JEDEC 4800MHz (XMP/EXPO disabled)
  • rasdaemon running — captures zero hardware errors during crashes

Current state:
The MODE1 resets no longer appear in logs after adding the kernel parameters, but hard reboots continue during D4 gameplay. No other game is affected. Crash leaves absolutely no trace in journalctl, dmesg, rasdaemon, or the game’s own Fenris debug log.

I believe this is an RDNA 4 + VKD3D/Proton interaction bug specific to Diablo 4’s DX12 rendering path. Crash report bundle attached.

Any help or pointers to relevant upstream bugs appreciated. Happy to test patches or provide additional logs.