Bug 214587 - [drm:amdgpu_job_timedout [amdgpu]] *ERROR* ring gfx timeout, signaled seq=5900910, emitted seq=5900912
Summary: [drm:amdgpu_job_timedout [amdgpu]] *ERROR* ring gfx timeout, signaled seq=590...
Status: NEW
Alias: None
Product: Drivers
Classification: Unclassified
Component: Video(DRI - non Intel) (show other bugs)
Hardware: Intel Linux
: P1 normal
Assignee: drivers_video-dri
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2021-09-30 21:36 UTC by Lahfa Samy
Modified: 2023-03-05 15:29 UTC (History)
1 user (show)

See Also:
Kernel Version: 5.14.7-arch1-1
Subsystem:
Regression: No
Bisected commit-id:


Attachments

Description Lahfa Samy 2021-09-30 21:36:58 UTC
Hi,

I've just recently hit this issue on ArchLinux kernel 5.14.7-arch1-1, linux-firmware 20210919.d526e04-1 with a Thinkpad T495 AMD Ryzen 7 3700U along a Vega Radeon RX 10 while using hashcat to brute-force hashes, hashcat was using OpenCL in order to use the GPU and then the computer just froze, and a GPU reset happened see the following logs.

Logs from dmesg:
[87507.678904] [drm] Fence fallback timer expired on ring gfx
[87512.691933] [drm] Fence fallback timer expired on ring gfx
[87517.572033] [drm] Fence fallback timer expired on ring gfx
[87523.012214] [drm] Fence fallback timer expired on ring gfx
[87533.129069] [drm:amdgpu_job_timedout [amdgpu]] *ERROR* ring gfx timeout, signaled seq=5900910, emitted seq=5900912
[87533.129518] [drm:amdgpu_job_timedout [amdgpu]] *ERROR* Process information: process Xorg pid 2937 thread Xorg:cs0 pid 3143
[87533.129957] amdgpu 0000:06:00.0: amdgpu: GPU reset begin!
[87533.138994] amdgpu 0000:06:00.0: amdgpu: Guilty job already signaled, skipping HW reset
[87533.139056] amdgpu 0000:06:00.0: amdgpu: GPU reset(2) succeeded!
Comment 1 Lahfa Samy 2021-09-30 21:38:10 UTC
The computer did unfreeze then after the reset of the GPU but it seems hashcat cannot use the GPU anymore for some reason, I'm not too sure why, but I think I need to reboot my machine.

Note You need to log in before you can comment on or make changes to this bug.