AMDGPU VM fault with Radeon Pro WX 7100 on 5.15.12 aarch64
Brief summary of the problem:
The aarch64 server uses the CentOS 7.7.1908 kernel to run the Android emulator for 5.15.12. Running for a period of time will cause GPU RESET
Hardware description:
- CPU:
HUAWEI Kunpeng 920 7260
- GPU0:
01:00.0 VGA compatible controller [0300]: Advanced Micro Devices, Inc. [AMD/ATI] Ellesmere [Radeon Pro WX 7100] [1002:67c4]
- GPU1:
81:00.0 VGA compatible controller [0300]: Advanced Micro Devices, Inc. [AMD/ATI] Ellesmere [Radeon Pro WX 7100] [1002:67c4]
System information:
- Distro name and Version:
CentOS 7.7.1908
- Kernel version:
5.15.12 #1 SMP Fri Dec 31 20:46:40 CST 2021 aarch64 aarch64 aarch64 GNU/Linux
- Custom kernel:
N/A
- AMD official driver version:
xf86-video-amdgpu 22.0.0
mesa-21.3.8
Log files (for system lockups / game freezes / crashes)
- Dmesg log (full log) dmesg.log