fatalerror1981 commented

(Migrated from github.com)

The idea

What happens The game builds and starts fine, sound and subtitles work, but the 3D picture is completely garbled: coloured noise and smeared purple/cyan bands instead of the scene (photo attached). First seen right after the intro black screen, at the opening cutscene.

System

GPU: AMD BC-250 (RDNA, GFX1013), RADV, Mesa 26.2.4 (VK_EXT_mesh_shader reported: task/mesh shaders true)
OS: CachyOS (Arch), gamescope session, launched as a non-Steam game
RAM: 14 GB usable (shared with GPU)
Hookshot: Linux AppImage, release foliage-linux-updater, CLI build, en-US
Game: 0400486024DEC000 v131072, build e0f3ccfa (--check OK)

Log lines that look relevant:

vk: graphics pipeline for d13aac3cec1a4bb5 not made (-8) (~10 of these, VK_ERROR_FEATURE_NOT_PRESENT)
vk: storage buffer NOT bound (reads zeros): program 68642bb810ee3e7c stage 7 c5+0x60 addr 0x10001c00a size 0x3ff7bdef
vk: storage buffer NOT bound (reads zeros): program 51183fe750d5d0c8 stage 7 c5+0x1d0 addr 0x3e0000003ba478e8 size 0x449f8155
vk: texture ... NVN 0x80 -> vk 137 (ASTC decoded)

Anything else

No response

### The idea What happens The game builds and starts fine, sound and subtitles work, but the 3D picture is completely garbled: coloured noise and smeared purple/cyan bands instead of the scene (photo attached). First seen right after the intro black screen, at the opening cutscene. System GPU: AMD BC-250 (RDNA, GFX1013), RADV, Mesa 26.2.4 (VK_EXT_mesh_shader reported: task/mesh shaders true) OS: CachyOS (Arch), gamescope session, launched as a non-Steam game RAM: 14 GB usable (shared with GPU) Hookshot: Linux AppImage, release foliage-linux-updater, CLI build, en-US Game: 0400486024DEC000 v131072, build e0f3ccfa (--check OK) Log lines that look relevant: <img width="1028" height="628" alt="Image" src="https://github.com/user-attachments/assets/06dbf0f1-224a-4fd6-a569-b2ccb6dec69d" /> vk: graphics pipeline for d13aac3cec1a4bb5 not made (-8) (~10 of these, VK_ERROR_FEATURE_NOT_PRESENT) vk: storage buffer NOT bound (reads zeros): program 68642bb810ee3e7c stage 7 c5+0x60 addr 0x10001c00a size 0x3ff7bdef vk: storage buffer NOT bound (reads zeros): program 51183fe750d5d0c8 stage 7 c5+0x1d0 addr 0x3e0000003ba478e8 size 0x449f8155 vk: texture ... NVN 0x80 -> vk 137 (ASTC decoded) ### Anything else _No response_

👍 1 😕 1

August0n3 commented

2026-11-07 22:27:20 +00:00

(Migrated from github.com)

I’m having this problem too

Rx 6600
Ryzen 5500
pop os cosmic

I’m having this problem too Rx 6600 Ryzen 5500 pop os cosmic

Vinni5147 commented

2026-11-07 22:32:22 +00:00

(Migrated from github.com)

Same here

RX 580 4GB
Ryzen 7 1700x
Fedora 44

Same here RX 580 4GB Ryzen 7 1700x Fedora 44

Kaizoku85 commented

2026-11-07 22:36:27 +00:00

(Migrated from github.com)

Same issue with Linux appimage after 36 minute compilation.
Rog Ally X
Bazzite 45

Same issue with Linux appimage after 36 minute compilation. Rog Ally X Bazzite 45

👍 1

Vinni5147 commented

2026-11-07 22:43:06 +00:00

(Migrated from github.com)

Solved changing the resolution to 50% but fps still very low.

<img width="1336" height="808" alt="Image" src="https://github.com/user-attachments/assets/1501a642-7132-465b-bacc-75169ad55784" /> Solved changing the resolution to 50% but fps still very low.

👍 2

August0n3 commented

2026-11-08 00:20:48 +00:00

(Migrated from github.com)

Solved changing the resolution to 50% but fps still very low.

still 3 fps, but works

> <img alt="Image" width="1336" height="808" src="https://private-user-images.githubusercontent.com/194864481/698697212-1501a642-7132-465b-bacc-75169ad55784.png?jwt=eyJ0eXAiOiJKV1QiLCJhbGciOiJIUzI1NiJ9.eyJpc3MiOiJnaXRodWIuY29tIiwiYXVkIjoicmF3LmdpdGh1YnVzZXJjb250ZW50LmNvbSIsImtleSI6ImtleTUiLCJleHAiOjE3OTQwOTc1MjQsIm5iZiI6MTc5NDA5NzIyNCwicGF0aCI6Ii8xOTQ4NjQ0ODEvNjk4Njk3MjEyLTE1MDFhNjQyLTcxMzItNDY1Yi1iYWNjLTc1MTY5YWQ1NTc4NC5wbmc_WC1BbXotQWxnb3JpdGhtPUFXUzQtSE1BQy1TSEEyNTYmWC1BbXotQ3JlZGVudGlhbD1BS0lBVkNPRFlMU0E1M1BRSzRaQSUyRjIwMjYxMTA4JTJGdXMtZWFzdC0xJTJGczMlMkZhd3M0X3JlcXVlc3QmWC1BbXotRGF0ZT0yMDI2MTEwOFQwMDIwMjRaJlgtQW16LUV4cGlyZXM9MzAwJlgtQW16LVNpZ25hdHVyZT02ZmZlZmIwZDEzMjQzZjIzNjNlNzFiNGEwYTMxMjg1NzMzNDViOTE0MzQ0MjYxZjNkMjg2NjUwMmJkNDMwNWE2JlgtQW16LVNpZ25lZEhlYWRlcnM9aG9zdCZyZXNwb25zZS1jb250ZW50LXR5cGU9aW1hZ2UlMkZwbmcifQ.fcZQsZd-w00Phc_rqcjfSsExjQ7k9yJXMaIl-BEyo7Q"> Solved changing the resolution to 50% but fps still very low. still 3 fps, but works

NotRexed commented

2026-11-08 00:41:08 +00:00

(Migrated from github.com)

#411 adds RADV logs with the same noise after the intro (Bazzite, plus a Rog Ally Z1 Extreme and BC-250 in its comments); #395 adds an RX 580 4 GB (RADV, Fedora 44) log at 1-3 fps.

sabogalc commented

2026-11-08 05:14:28 +00:00

(Migrated from github.com)

I can reproduce this issue on my system with an AMD Radeon RX 6800 XT (Navi 21) running CachyOS.

System information:

GPU: AMD Radeon RX 6800 XT (16 GB VRAM)
iGPU: AMD Radeon 840M
Kernel: 7.3.1-1-cachyos
Mesa: 26.2.6-1
RADV / vulkan-radeon: 26.2.6-1
lib32-mesa: 26.2.6-1
lib32-vulkan-radeon: 26.2.6-1
No undervolting or overclocking. GPU is running at stock settings.

How the crash occurs:

Cutscene plays when entering Goron City.
The graphics become corrupted, with extremely bright/overexposed lighting and fog (similar to the visual corruption reported in this issue).
Shortly afterwards, Hookshot crashes.

This exact screenshot is from a different error in the Kokiri Forest, but it looked about the same in Goron City.

Relevant Vulkan errors:

vk: submit 6637 failed (-4)
vk: submit 6638 failed (-4)
vk: submit 6639 failed (-4)

The application subsequently reports:

X connection to :0 broken
corrupted double-linked list

Relevant AMDGPU kernel messages:

amdgpu: GPU reset begin!
amdgpu: MODE1 reset
amdgpu: GPU reset succeeded
amdgpu: VRAM is lost due to GPU reset!

Hookshot also reported the following memory usage shortly before the crash:

VRAM: 15643 MB used of 15827 MB allowed
textures 3101 MB (budget 2039 MB)
staging 9344 MB in 146 arenas

Other games run normally on this system, and I have not experienced similar crashes outside Hookshot.

This makes me suspect an AMD RADV/Vulkan rendering or resource-management issue triggered by cutscenes or new areas in the game.

hookshot_log.txt

I can reproduce this issue on my system with an AMD Radeon RX 6800 XT (Navi 21) running CachyOS. # System information: - GPU: AMD Radeon RX 6800 XT (16 GB VRAM) - iGPU: AMD Radeon 840M - Kernel: 7.3.1-1-cachyos - Mesa: 26.2.6-1 - RADV / vulkan-radeon: 26.2.6-1 - lib32-mesa: 26.2.6-1 - lib32-vulkan-radeon: 26.2.6-1 - No undervolting or overclocking. GPU is running at stock settings. # How the crash occurs: - Cutscene plays when entering Goron City. - The graphics become corrupted, with extremely bright/overexposed lighting and fog (similar to the visual corruption reported in this issue). - Shortly afterwards, Hookshot crashes. <img width="1920" height="1080" alt="Image" src="https://github.com/user-attachments/assets/10ef926a-1038-41b0-a818-3fbeca5b44de" /> This exact screenshot is from a different error in the Kokiri Forest, but it looked about the same in Goron City. # Relevant Vulkan errors: ``` vk: submit 6637 failed (-4) vk: submit 6638 failed (-4) vk: submit 6639 failed (-4) ``` The application subsequently reports: ``` X connection to :0 broken corrupted double-linked list ``` # Relevant AMDGPU kernel messages: ``` amdgpu: GPU reset begin! amdgpu: MODE1 reset amdgpu: GPU reset succeeded amdgpu: VRAM is lost due to GPU reset! ``` Hookshot also reported the following memory usage shortly before the crash: ``` VRAM: 15643 MB used of 15827 MB allowed textures 3101 MB (budget 2039 MB) staging 9344 MB in 146 arenas ``` Other games run normally on this system, and I have not experienced similar crashes outside Hookshot. This makes me suspect an AMD RADV/Vulkan rendering or resource-management issue triggered by cutscenes or new areas in the game. [hookshot_log.txt](https://github.com/user-attachments/files/34537886/hookshot_log.txt)

pablovera2735 commented

2026-11-08 07:06:01 +00:00

(Migrated from github.com)

It works fine for me, but the texture looks a bit off

![hookshot_log.txt](https://github.com/user-attachments/files/34539166/hookshot_log.txt) It works fine for me, but the texture looks a bit off

adnanioricce commented

2026-11-08 16:00:15 +00:00

(Migrated from github.com)

it seems that PR #425 fixes the noise corruption, without the necessity of the HOS_DYNRES=0 and HOS_RES_SCALE=0.5 workarounds.
I tried running a build with #425 , #448 and #487 in a Asus Rog Ally Z1 Extreme with Bazzite , and could pass from the initial cinematic and reach the gameplay part in link's bedroom. While the FPS low, the situation has clearly improved compared from yesterday with corruption and FPS so low that wasn't worted to go to the first gameplay and save.
I will attach the logs
PS: find out now that the screenshot of the gameplay wasn't taken, so I will leave at least the shots that were taken.

hookshot-log-20261108-123306.tar.gz

it seems that PR #425 fixes the noise corruption, without the necessity of the `HOS_DYNRES=0` and `HOS_RES_SCALE=0.5` workarounds. I tried running a build with #425 , #448 and #487 in a Asus Rog Ally Z1 Extreme with Bazzite , and could pass from the initial cinematic and reach the gameplay part in link's bedroom. While the FPS low, the situation has clearly improved compared from yesterday with corruption and FPS so low that wasn't worted to go to the first gameplay and save. I will attach the logs PS: find out now that the screenshot of the gameplay wasn't taken, so I will leave at least the shots that were taken. <img width="1280" height="720" alt="Image" src="https://github.com/user-attachments/assets/43cc1388-36a6-48e0-826a-05c1b21fa157" /> <img width="1280" height="720" alt="Image" src="https://github.com/user-attachments/assets/6ce390b1-dc2d-4a1f-ac7a-42addbe2be9c" /> <img width="1280" height="720" alt="Image" src="https://github.com/user-attachments/assets/11ff4aef-b76b-4874-8d5c-11715a34371f" /> [hookshot-log-20261108-123306.tar.gz](https://github.com/user-attachments/files/34546603/hookshot-log-20261108-123306.tar.gz)

adnanioricce commented

2026-11-08 16:07:24 +00:00

(Migrated from github.com)

an additional report, with some lines I am using to diagnose the run:
hookshot-log-20261108-123306-report.txt

an additional report, with some lines I am using to diagnose the run: [hookshot-log-20261108-123306-report.txt](https://github.com/user-attachments/files/34546701/hookshot-log-20261108-123306-report.txt)

sabogalc commented

2026-11-08 16:30:13 +00:00

(Migrated from github.com)

Update: Issue persists in Hookshot v2.0.1

I have tested the latest v2.0.1 release, but unfortunately the issue still occurs.

I noticed some differences in the new logs compared to my original report though.

1. VRAM usage

The previous crash showed VRAM usage approaching the available limit (15,643 MB / 15,827 MB), with textures exceeding their reported budget.

In the latest v2.0.1 crash, memory usage was considerably lower:

working set 5661 MB (peak 7945)
staging 3136 MB in 49 arenas

heap 0 (system): used 6604 of 7528 MB
heap 1 (device): used 12642 of 15610 MB

VRAM: 12642 MB used of 15610 MB allowed
textures 5331 MB (budget 6144 MB)

This suggests that the issue is not necessarily caused by simply exhausting the available VRAM or exceeding the texture budget.

2. Large rendered textures

I also noticed some particularly large textures being allocated:
text

288 MB: 12288x12288x1 rendered
213 MB: 1365x1365x1 x90 ASTC
213 MB: 1365x1365x1 x90 ASTC
97 MB: 2000x3200x1 x8

I'm not sure whether these are expected, but they may be relevant to the rendering or resource-management issue.

3. GPU synchronization and rendering stalls

Shortly before the crash, the log reported:
text

nvn: GPU waits in 5 s: 0 sync waits (0 ms),
14 polls timed out (443 ms), 7 queue finishes (0 ms)

Rendering performance also dropped significantly, running at a low FPS before it essentially hang still while the audio remained playing.

4. Final crash

The application eventually terminated with:
text

XIO: fatal IO error 0 (Success) on X server ":0"

double free or corruption (out)

This is somewhat different from the previous crash, which was accompanied by Vulkan submission failures and an AMDGPU GPU reset.

Conclusion

Updating to v2.0.1 has not resolved the issue. Although the reported VRAM usage is now considerably lower than in my original report, the application still appears to encounter serious rendering/synchronization problems before terminating.

Given that I can reproduce this consistently on Navi 21 with RADV, I wonder whether there is an underlying issue with Vulkan resource management, texture synchronization, or GPU command submission.

I'm attaching the updated log for further investigation.

hookshot_log.txt

# Update: Issue persists in Hookshot v2.0.1 I have tested the latest v2.0.1 release, but unfortunately the issue still occurs. I noticed some differences in the new logs compared to my original report though. # 1. VRAM usage The previous crash showed VRAM usage approaching the available limit (15,643 MB / 15,827 MB), with textures exceeding their reported budget. In the latest v2.0.1 crash, memory usage was considerably lower: ``` working set 5661 MB (peak 7945) staging 3136 MB in 49 arenas heap 0 (system): used 6604 of 7528 MB heap 1 (device): used 12642 of 15610 MB VRAM: 12642 MB used of 15610 MB allowed textures 5331 MB (budget 6144 MB) ``` This suggests that the issue is not necessarily caused by simply exhausting the available VRAM or exceeding the texture budget. # 2. Large rendered textures I also noticed some particularly large textures being allocated: text ``` 288 MB: 12288x12288x1 rendered 213 MB: 1365x1365x1 x90 ASTC 213 MB: 1365x1365x1 x90 ASTC 97 MB: 2000x3200x1 x8 ``` I'm not sure whether these are expected, but they may be relevant to the rendering or resource-management issue. # 3. GPU synchronization and rendering stalls Shortly before the crash, the log reported: text ``` nvn: GPU waits in 5 s: 0 sync waits (0 ms), 14 polls timed out (443 ms), 7 queue finishes (0 ms) ``` Rendering performance also dropped significantly, running at a low FPS before it essentially hang still while the audio remained playing. # 4. Final crash The application eventually terminated with: text ``` XIO: fatal IO error 0 (Success) on X server ":0" double free or corruption (out) ``` This is somewhat different from the previous crash, which was accompanied by Vulkan submission failures and an AMDGPU GPU reset. # Conclusion Updating to v2.0.1 has not resolved the issue. Although the reported VRAM usage is now considerably lower than in my original report, the application still appears to encounter serious rendering/synchronization problems before terminating. Given that I can reproduce this consistently on Navi 21 with RADV, I wonder whether there is an underlying issue with Vulkan resource management, texture synchronization, or GPU command submission. I'm attaching the updated log for further investigation. [hookshot_log.txt](https://github.com/user-attachments/files/34547065/hookshot_log.txt)

isorropisths commented

2026-11-08 23:51:36 +00:00

(Migrated from github.com)

Indeed, dropping the resolution fixes the glitchy image, but it's 2fps. I see GPU Utilization is 0%.

Indeed, dropping the resolution fixes the glitchy image, but it's 2fps. I see GPU Utilization is 0%. <img width="1330" height="807" alt="Image" src="https://github.com/user-attachments/assets/d3e33010-7d86-467b-9fc2-a65616280064" />

NotRexed commented

2026-11-09 05:47:15 +00:00

(Migrated from github.com)

Tracked in #1000, follow there. That issue collects every report of Linux / Steam Deck GPU hangs, crashes, low fps and build failures, and this one is linked from it with its GPU, version and log details. If it still happens on v2.0.6 (https://github.com/NotRexed/Hookshot/releases/tag/v2.0.6; Update next to Play in the launcher), please add a comment there with your hookshot_log.txt.

Tracked in #1000, follow there. That issue collects every report of Linux / Steam Deck GPU hangs, crashes, low fps and build failures, and this one is linked from it with its GPU, version and log details. If it still happens on v2.0.6 (https://github.com/NotRexed/Hookshot/releases/tag/v2.0.6; Update next to Play in the launcher), please add a comment there with your `hookshot_log.txt`.