[Crash / GPU Reset] amdgpu ring gfx_0.1.0 timeout (MODE1 reset, VRAM lost) on AMD Radeon RX 6700 XT (RDNA2 / RADV / Linux) #1065

Closed

opened 2026-11-09 08:58:10 +00:00 by Cloudiny · 3 comments

Cloudiny commented

2026-11-09 08:58:10 +00:00

(Migrated from github.com)

Hookshot version

v2.0.7

What kind of problem

Crash, freeze or black screen

What happened

Summary
Hookshot v2.0.7 crashes and triggers an AMDGPU driver kernel reset (ring gfx_0.1.0 timeout leading to MODE1 reset and VRAM is lost) when playing on Linux with an AMD RDNA2 GPU (Radeon RX 6700 XT) under Mesa RADV. The crash consistently occurs when exiting Link's house and meeting Saria in Kokiri Forest.

System Information

OS: Arch Linux / EndeavourOS (Kernel 6.x, KDE Plasma Wayland)
GPU: AMD Radeon RX 6700 XT (NAVI22, 12GB VRAM)
Vulkan Driver: Mesa RADV 26.2.6 (Vulkan 1.4, driver 0x6802006)
Hookshot Version: v2.0.7 AppImage (source 6b1dd5d6)

Steps to Reproduce

Launch Hookshot v2.0.7 AppImage on Linux (RADV / AMD RDNA2).
Start a new game and leave Link's house.
Approach Saria outside Link's house.
Framerate suddenly drops from 60 FPS to ~11-12 FPS as texture and capture memory grows.
The screen freezes, display server loses connection, and dmesg reports a full GPU MODE1 reset.

Symptoms & Analysis

Memory usage climbs over 8.4 GB of textures (4025 textures (8488 MB)) and 9.9 GB of mirror sync captures (2104 kept (9987 MB)).
Shader compiler reports 175 failed programs (51768 programs (175 failed)).
The NVN command buffer wait queue stalls on sync events (59 polls timed out (0 ms)).
Kernel watchdog triggers on ring gfx_0.1.0 timeout. Soft reset fails, escalating to hardware MODE1 reset, dumping VRAM and invalidating the Vulkan context.

Relevant Hookshot Log (hookshot_log.txt)
[hos Presentation Thread] vk: 14160 presents, 12.7 fps; a frame: 3656 draws (0 skipped), 168 dispatches, 184 passes, 184 barriers, 6 submits, 0.3 uploads (0.1 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms; 1006 pipelines (45633 ms), 51768 programs (175 failed), 4025 textures (8488 MB)
[hos Presentation Thread] vi: 14168 frames presented (11.7 fps over the last 5.0 s, interval 1)
[hos sead::MainThread] nvn: GPU waits in 5 s: 0 sync waits (0 ms), 59 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos sead::MainThread] vk: a capture taken (ms): all 0.019; retires and stamps 0.012, bytes copied 0.000; 2104 kept (9987 MB), 0 spare buffers, 1047 texture pages copied (4 MB)

Kernel Log (journalctl -k / dmesg)
amdgpu 0000:0c:00.0: ring gfx_0.1.0 timeout, signaled seq=5780045, emitted seq=5780047
amdgpu 0000:0c:00.0: Process kwin_wayland pid 1586 thread kwin_wayla:cs0 pid 1990
amdgpu 0000:0c:00.0: Starting gfx_0.1.0 ring reset
amdgpu 0000:0c:00.0: Ring gfx_0.1.0 reset failed
amdgpu 0000:0c:00.0: GPU reset begin!. Source: 1
amdgpu 0000:0c:00.0: MODE1 reset
amdgpu 0000:0c:00.0: GPU mode1 reset
amdgpu 0000:0c:00.0: GPU reset succeeded, trying to resume
amdgpu 0000:0c:00.0: VRAM is lost due to GPU reset!

Where in the game

Kokiri Forest

How often

Every time

Graphics card and driver version

AMD Radeon RX 6700 XT (NAVI22, 12GB VRAM)

CPU, RAM and operating system

AMD Ryzen 5 5500X3D, 56GB RAM, EndeavourOS

Changes from the defaults

No response

Other software hooked into the game

No, nothing hooked in

hookshot_log.txt

Screenshots or video

No response

Before you send it

I'm on the newest release, or the problem is still there on it.
I searched the open issues for this problem.
I'm not attaching game files, keys, hookshot.exe, the .bin caches or saves.

### Hookshot version v2.0.7 ### What kind of problem Crash, freeze or black screen ### What happened Summary Hookshot v2.0.7 crashes and triggers an AMDGPU driver kernel reset (ring gfx_0.1.0 timeout leading to MODE1 reset and VRAM is lost) when playing on Linux with an AMD RDNA2 GPU (Radeon RX 6700 XT) under Mesa RADV. The crash consistently occurs when exiting Link's house and meeting Saria in Kokiri Forest. System Information - OS: Arch Linux / EndeavourOS (Kernel 6.x, KDE Plasma Wayland) - GPU: AMD Radeon RX 6700 XT (NAVI22, 12GB VRAM) - Vulkan Driver: Mesa RADV 26.2.6 (Vulkan 1.4, driver 0x6802006) - Hookshot Version: v2.0.7 AppImage (source 6b1dd5d6) Steps to Reproduce 1. Launch Hookshot v2.0.7 AppImage on Linux (RADV / AMD RDNA2). 2. Start a new game and leave Link's house. 3. Approach Saria outside Link's house. 4. Framerate suddenly drops from 60 FPS to ~11-12 FPS as texture and capture memory grows. 5. The screen freezes, display server loses connection, and dmesg reports a full GPU MODE1 reset. Symptoms & Analysis 1. Memory usage climbs over 8.4 GB of textures (4025 textures (8488 MB)) and 9.9 GB of mirror sync captures (2104 kept (9987 MB)). 2. Shader compiler reports 175 failed programs (51768 programs (175 failed)). 3. The NVN command buffer wait queue stalls on sync events (59 polls timed out (0 ms)). 4. Kernel watchdog triggers on ring gfx_0.1.0 timeout. Soft reset fails, escalating to hardware MODE1 reset, dumping VRAM and invalidating the Vulkan context. Relevant Hookshot Log (hookshot_log.txt) [hos Presentation Thread] vk: 14160 presents, 12.7 fps; a frame: 3656 draws (0 skipped), 168 dispatches, 184 passes, 184 barriers, 6 submits, 0.3 uploads (0.1 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms; 1006 pipelines (45633 ms), 51768 programs (175 failed), 4025 textures (8488 MB) [hos Presentation Thread] vi: 14168 frames presented (11.7 fps over the last 5.0 s, interval 1) [hos sead::MainThread] nvn: GPU waits in 5 s: 0 sync waits (0 ms), 59 polls timed out (0 ms), 0 queue finishes (0 ms) [hos sead::MainThread] vk: a capture taken (ms): all 0.019; retires and stamps 0.012, bytes copied 0.000; 2104 kept (9987 MB), 0 spare buffers, 1047 texture pages copied (4 MB) Kernel Log (journalctl -k / dmesg) amdgpu 0000:0c:00.0: ring gfx_0.1.0 timeout, signaled seq=5780045, emitted seq=5780047 amdgpu 0000:0c:00.0: Process kwin_wayland pid 1586 thread kwin_wayla:cs0 pid 1990 amdgpu 0000:0c:00.0: Starting gfx_0.1.0 ring reset amdgpu 0000:0c:00.0: Ring gfx_0.1.0 reset failed amdgpu 0000:0c:00.0: GPU reset begin!. Source: 1 amdgpu 0000:0c:00.0: MODE1 reset amdgpu 0000:0c:00.0: GPU mode1 reset amdgpu 0000:0c:00.0: GPU reset succeeded, trying to resume amdgpu 0000:0c:00.0: VRAM is lost due to GPU reset! ### Where in the game Kokiri Forest ### How often Every time ### Graphics card and driver version AMD Radeon RX 6700 XT (NAVI22, 12GB VRAM) ### CPU, RAM and operating system AMD Ryzen 5 5500X3D, 56GB RAM, EndeavourOS ### Changes from the defaults _No response_ ### Other software hooked into the game No, nothing hooked in ### hookshot_log.txt [hookshot_log.txt](https://github.com/user-attachments/files/34571094/hookshot_log.txt) ### Screenshots or video _No response_ ### Before you send it - [x] I'm on the newest release, or the problem is still there on it. - [x] I searched the open issues for this problem. - [x] I'm not attaching game files, keys, hookshot.exe, the .bin caches or saves.

nG64 commented

2026-11-09 09:04:59 +00:00

(Migrated from github.com)

I have the Same issue in the Steamdeck

NotRexed commented

2026-11-09 09:18:53 +00:00

(Migrated from github.com)

Thanks for the detailed report and kernel log. Linking this to the Linux and Steam Deck tracker #1000, where GPU hangs on RADV are collected (the Steam Deck report in the comments too). The log's growth to ~10 GB of kept captures and ~8.5 GB of textures before the hang is the useful part.

👍 1

NotRexed commented

2026-11-09 22:51:53 +00:00

(Migrated from github.com)

Thanks for the report. To keep each problem in one place, this is tracked in #1000 (Linux and Steam Deck: GPU hangs, crashes and low FPS), which links your details. Closing this as a duplicate; please follow #1000 for updates.