AMDGPU Virtual Memory Protection Fault

JuanD1985 · 2024-09-05 23:39:43

When running the programs, the interface freezes. I suspect this may be related to OpenGL, Mesa, or amdgpu drivers. I am currently looking for a solution.

sep 04 11:04:39 archlinux kernel: amdgpu 0000:04:00.0: amdgpu:  for process topas pid 32347 thread
 topas:cs0 pid 32351)
sep 04 11:04:39 archlinux kernel: amdgpu 0000:04:00.0: amdgpu:   in page starting at address 0x000
0dfad6142e000 from IH client 0x1b (UTCL2)
sep 04 11:04:39 archlinux kernel: amdgpu 0000:04:00.0: amdgpu: VM_L2_PROTECTION_FAULT_STATUS:0x005
01031--
sep 04 11:04:39 archlinux kernel: amdgpu 0000:04:00.0: amdgpu:          Faulty UTCL2 client ID: TC
P (0x8)
sep 04 11:04:39 archlinux kernel: amdgpu 0000:04:00.0: amdgpu:          MORE_FAULTS: 0x1
sep 04 11:04:39 archlinux kernel: amdgpu 0000:04:00.0: amdgpu:          WALKER_ERROR: 0x0
sep 04 11:04:39 archlinux kernel: amdgpu 0000:04:00.0: amdgpu:          PERMISSION_FAULTS: 0x3
sep 04 11:04:39 archlinux kernel: amdgpu 0000:04:00.0: amdgpu:          MAPPING_ERROR: 0x0
sep 04 11:04:39 archlinux kernel: amdgpu 0000:04:00.0: amdgpu:          RW: 0x0
sep 04 11:04:51 archlinux kernel: [drm:amdgpu_job_timedout [amdgpu]] *ERROR* ring gfx timeout, but
 soft recovered

system information

» lspci -k | grep -A 3 -E "(VGA|3D)"
04:00.0 VGA compatible controller: Advanced Micro Devices, Inc. [AMD/ATI] Barcelo (rev c2)
	DeviceName: AMD Radeon Integrated Graphics
	Subsystem: Hewlett-Packard Company Device 8999
	Kernel driver in use: amdgpu

» uname -r
6.10.8-arch1-1

» sudo dmesg | grep amdgpu
[    0.000000] Command line: initrd=\amd-ucode.img initrd=\initramfs-linux.img root=PARTUUID=bcc8e8c4-3070-42c6-b105-1f011fae6097 zswap.enabled=0 rw amdgpu.cik_support=0
[    0.025342] Kernel command line: initrd=\amd-ucode.img initrd=\initramfs-linux.img root=PARTUUID=bcc8e8c4-3070-42c6-b105-1f011fae6097 zswap.enabled=0 rw amdgpu.cik_support=0
[    2.824826] [drm] amdgpu kernel modesetting enabled.
[    2.835576] amdgpu: Virtual CRAT table created for CPU
[    2.835589] amdgpu: Topology: Add CPU node
[    2.835717] amdgpu 0000:04:00.0: enabling device (0006 -> 0007)
[    3.051766] amdgpu 0000:04:00.0: amdgpu: Fetched VBIOS from VFCT
[    3.051770] amdgpu: ATOM BIOS: 113-BARCELO-004
[    3.108539] amdgpu 0000:04:00.0: vgaarb: deactivate vga console
[    3.108550] amdgpu 0000:04:00.0: amdgpu: Trusted Memory Zone (TMZ) feature enabled
[    3.108557] amdgpu 0000:04:00.0: amdgpu: MODE2 reset
[    3.108645] amdgpu 0000:04:00.0: amdgpu: VRAM: 512M 0x000000F400000000 - 0x000000F41FFFFFFF (512M used)
[    3.108648] amdgpu 0000:04:00.0: amdgpu: GART: 1024M 0x0000000000000000 - 0x000000003FFFFFFF
[    3.108878] [drm] amdgpu: 512M of VRAM memory ready
[    3.108884] [drm] amdgpu: 3644M of GTT memory ready.
[    3.110254] amdgpu 0000:04:00.0: amdgpu: Will use PSP to load VCN firmware
[    3.834775] amdgpu 0000:04:00.0: amdgpu: reserve 0x400000 from 0xf41f800000 for PSP TMR

Last edited by JuanD1985 (2024-09-05 23:46:24)

seth · 2024-09-06 08:27:48

Recent regression?
https://bbs.archlinux.org/viewtopic.php?id=299047

cryptearth · 2024-09-06 08:38:56

AMD Barcelo
Generation: Vega
Architecture: GCN 5.1

seth · 2024-09-06 08:42:27

And the linked threads has users hitting a similar regresssion on navi 21 and 31

cryptearth · 2024-09-06 10:24:53

@seth
at least this time I just wanted to give a decode of the codename Barcelo without my usual rant just because I was unaware of it and actually had a bit of trouble to find a datasheet - it's just meant as a hint that OP uses a Vega-class card

Last edited by cryptearth (2024-09-06 10:26:28)

seth · 2024-09-06 12:12:52

https://www.techpowerup.com/gpu-specs/amd-barcelo.g1045
It's actually a fairly new chip

Arch Linux

#1 2024-09-05 23:39:43

AMDGPU Virtual Memory Protection Fault

#2 2024-09-06 08:27:48

Re: AMDGPU Virtual Memory Protection Fault

#3 2024-09-06 08:38:56

Re: AMDGPU Virtual Memory Protection Fault

#4 2024-09-06 08:42:27

Re: AMDGPU Virtual Memory Protection Fault

#5 2024-09-06 10:24:53

Re: AMDGPU Virtual Memory Protection Fault

#6 2024-09-06 12:12:52

Re: AMDGPU Virtual Memory Protection Fault

Board footer