Bug 98785

Summary:	Talos Principle causes GPU faults when rendering main menu animation
Product:	Mesa	Reporter:	Vedran Miletić <vedran>
Component:	Drivers/Gallium/radeonsi	Assignee:	Default DRI bug account <dri-devel>
Status:	RESOLVED FIXED	QA Contact:	Default DRI bug account <dri-devel>
Severity:	major
Priority:	medium	CC:	arsenm2
Version:	git	Keywords:	regression
Hardware:	Other
OS:	All
Whiteboard:
i915 platform:		i915 features:

Description Vedran Miletić 2016-11-19 17:16:23 UTC

I'm using Mesa and LLVM git snapshots on kernel 4.8.8 on Tonga (R9 380X). Running Talos public beta gives:

[  220.000810] amdgpu 0000:01:00.0: GPU fault detected: 147 0x0620c001
[  220.000816] amdgpu 0000:01:00.0:   VM_CONTEXT1_PROTECTION_FAULT_ADDR   0x0F33DCC4
[  220.000819] amdgpu 0000:01:00.0:   VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x060C0001
[  220.000823] VM fault (0x01, vmid 3) at page 255057092, read from 'TC5' (0x54433500) (192)
[  220.050610] amdgpu 0000:01:00.0: GPU fault detected: 147 0x02604402
[  220.050619] amdgpu 0000:01:00.0:   VM_CONTEXT1_PROTECTION_FAULT_ADDR   0x0002524C
[  220.050623] amdgpu 0000:01:00.0:   VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x06044002
[  220.050628] VM fault (0x02, vmid 3) at page 152140, read from 'TC7' (0x54433700) (68)
[  220.069652] amdgpu 0000:01:00.0: GPU fault detected: 147 0x07e08002
[  220.069659] amdgpu 0000:01:00.0:   VM_CONTEXT1_PROTECTION_FAULT_ADDR   0x0013E8FC
[  220.069663] amdgpu 0000:01:00.0:   VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x06080002
[  220.069666] VM fault (0x02, vmid 3) at page 1304828, read from 'TC11' (0x54433131) (128)
[  220.134453] amdgpu 0000:01:00.0: GPU fault detected: 147 0x00004802
[  220.134458] amdgpu 0000:01:00.0:   VM_CONTEXT1_PROTECTION_FAULT_ADDR   0x00000000
[  220.134460] amdgpu 0000:01:00.0:   VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x0A048002
[  220.134462] VM fault (0x02, vmid 5) at page 0, read from 'TC6' (0x54433600) (72)
[  220.174059] amdgpu 0000:01:00.0: GPU fault detected: 147 0x0620c001
[  220.174064] amdgpu 0000:01:00.0:   VM_CONTEXT1_PROTECTION_FAULT_ADDR   0x0F03DCC4
[  220.174066] amdgpu 0000:01:00.0:   VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x080C0001
[  220.174069] VM fault (0x01, vmid 4) at page 251911364, read from 'TC5' (0x54433500) (192)
[  220.174290] amdgpu 0000:01:00.0: GPU fault detected: 147 0x07a8c002
[  220.174297] amdgpu 0000:01:00.0:   VM_CONTEXT1_PROTECTION_FAULT_ADDR   0x0013E8F5
[  220.174300] amdgpu 0000:01:00.0:   VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x080C0002
[  220.174303] VM fault (0x02, vmid 4) at page 1304821, read from 'TC5' (0x54433500) (192)

The rendering looks correct, unlike with Vulkan in bug 98783.

Comment 1 Vedran Miletić 2016-11-20 14:00:28 UTC

Related LLVM bug: https://llvm.org/bugs/show_bug.cgi?id=31019

Comment 2 Vedran Miletić 2016-11-25 00:50:41 UTC

Still broken with llvm r287905287905.

Comment 3 Vedran Miletić 2016-11-25 00:51:11 UTC

(In reply to Vedran Miletić from comment #2)
> Still broken with llvm r287905287905.

*r287905

Comment 4 Vedran Miletić 2017-01-05 11:53:42 UTC

> Related LLVM bug: https://llvm.org/bugs/show_bug.cgi?id=31019

Fixed.

Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.