Bug 104329

Summary: Vulkan app crashes GPU
Product: Mesa Reporter: pete.marchingcubes
Component: Drivers/Vulkan/radeonAssignee: mesa-dev
Status: RESOLVED WORKSFORME QA Contact: mesa-dev
Severity: normal    
Priority: medium CC: neel84250
Version: git   
Hardware: Other   
OS: All   
Whiteboard:
i915 platform: i915 features:

Description pete.marchingcubes 2017-12-18 21:38:30 UTC
Running RX560, Fedora 26, 4.14.5 kernel, mesa git from https://copr.fedorainfracloud.org/coprs/che/mesa/

Running my Vulkan VR application (developed with NVidia, no validation errors) crashes radv resulting in completely frozen display after partially rendering a single frame with the following dmesg output:

[ 1352.348283] amdgpu 0000:02:00.0: GPU fault detected: 146 0x079a1014
[ 1352.348289] amdgpu 0000:02:00.0:   VM_CONTEXT1_PROTECTION_FAULT_ADDR   0x001000F3
[ 1352.348292] amdgpu 0000:02:00.0:   VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x0D010014
[ 1352.348296] amdgpu 0000:02:00.0: VM fault (0x14, vmid 6) at page 1048819, write from 'CB3' (0x43423300) (16)
[ 1352.348362] amdgpu 0000:02:00.0: GPU fault detected: 146 0x07ba1014
[ 1352.348363] amdgpu 0000:02:00.0:   VM_CONTEXT1_PROTECTION_FAULT_ADDR   0x00000000
[ 1352.348364] amdgpu 0000:02:00.0:   VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x0D050014
[ 1352.348366] amdgpu 0000:02:00.0: VM fault (0x14, vmid 6) at page 0, write from 'CB1' (0x43423100) (80)
[ 1352.348399] amdgpu 0000:02:00.0: GPU fault detected: 146 0x07ba2014
[ 1352.348401] amdgpu 0000:02:00.0:   VM_CONTEXT1_PROTECTION_FAULT_ADDR   0x00000000
[ 1352.348402] amdgpu 0000:02:00.0:   VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x0D050014
[ 1352.348404] amdgpu 0000:02:00.0: VM fault (0x14, vmid 6) at page 0, write from 'CB1' (0x43423100) (80)
[ 1352.348463] amdgpu 0000:02:00.0: GPU fault detected: 146 0x07fa2014
[ 1352.348464] amdgpu 0000:02:00.0:   VM_CONTEXT1_PROTECTION_FAULT_ADDR   0x00000000
[ 1352.348465] amdgpu 0000:02:00.0:   VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x0D010014
[ 1352.348467] amdgpu 0000:02:00.0: VM fault (0x14, vmid 6) at page 0, write from 'CB3' (0x43423300) (16)
[ 1352.348473] amdgpu 0000:02:00.0: GPU fault detected: 146 0x07da2014
[ 1352.348474] amdgpu 0000:02:00.0:   VM_CONTEXT1_PROTECTION_FAULT_ADDR   0x001000ED
[ 1352.348476] amdgpu 0000:02:00.0:   VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x0D050014
[ 1352.348477] amdgpu 0000:02:00.0: VM fault (0x14, vmid 6) at page 1048813, write from 'CB1' (0x43423100) (80)
[ 1352.348489] amdgpu 0000:02:00.0: GPU fault detected: 146 0x07da1014
[ 1352.348490] amdgpu 0000:02:00.0:   VM_CONTEXT1_PROTECTION_FAULT_ADDR   0x001000F3
[ 1352.348491] amdgpu 0000:02:00.0:   VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x0D050014
[ 1352.348493] amdgpu 0000:02:00.0: VM fault (0x14, vmid 6) at page 1048819, write from 'CB1' (0x43423100) (80)
[ 1352.348519] amdgpu 0000:02:00.0: GPU fault detected: 146 0x079a2014
[ 1352.348520] amdgpu 0000:02:00.0:   VM_CONTEXT1_PROTECTION_FAULT_ADDR   0x001000F3
[ 1352.348521] amdgpu 0000:02:00.0:   VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x0D010014
[ 1352.348523] amdgpu 0000:02:00.0: VM fault (0x14, vmid 6) at page 1048819, write from 'CB3' (0x43423300) (16)
[ 1352.348548] amdgpu 0000:02:00.0: GPU fault detected: 146 0x079a1014
[ 1352.348550] amdgpu 0000:02:00.0:   VM_CONTEXT1_PROTECTION_FAULT_ADDR   0x001000DF
[ 1352.348551] amdgpu 0000:02:00.0:   VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x0D010014
[ 1352.348553] amdgpu 0000:02:00.0: VM fault (0x14, vmid 6) at page 1048799, write from 'CB3' (0x43423300) (16)
[ 1352.348576] amdgpu 0000:02:00.0: GPU fault detected: 146 0x07ba2014
[ 1352.348577] amdgpu 0000:02:00.0:   VM_CONTEXT1_PROTECTION_FAULT_ADDR   0x001000D0
[ 1352.348579] amdgpu 0000:02:00.0:   VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x0D020014
[ 1352.348581] amdgpu 0000:02:00.0: VM fault (0x14, vmid 6) at page 1048784, write from 'CB2' (0x43423200) (32)
[ 1352.348607] amdgpu 0000:02:00.0: GPU fault detected: 146 0x07ba1014
[ 1352.348609] amdgpu 0000:02:00.0:   VM_CONTEXT1_PROTECTION_FAULT_ADDR   0x001000BD
[ 1352.348610] amdgpu 0000:02:00.0:   VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x0D020014
[ 1352.348612] amdgpu 0000:02:00.0: VM fault (0x14, vmid 6) at page 1048765, write from 'CB2' (0x43423200) (32)
Comment 1 Samuel Pitoiset 2017-12-18 23:12:26 UTC
Hi,

Can you reproduce the VM faults after reverting ff0f17da1446e7aa965e06c04a6ad5a55d95463d ?
Comment 2 Samuel Pitoiset 2018-05-15 20:07:34 UTC
Closing, no info provided and I doubt this can still be reproduced. If I'm wrong feel free to re-open and explain how to reproduce. Thanks!

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.