Test Crashes/Failures seen on Samsung Galaxy S24

[johnzupin]

The following test issues were seen on a Samsung Galaxy S24 (build number UP1A.2310005.007.S928BXXU1AXCA) using VVL at v1.3.287-8-g28c0a5d84

        - NegativeGpuAVBufferDeviceAddress.ProxyStructLoadUint64
        - NegativeGpuAVBufferDeviceAddress.ProxyStructLoadBadAddress
        - PositiveGpuAVBufferDeviceAddress.StructLoad
        - NegativeGpuAVDescriptorIndexing.ArrayOOBGeom
        - NegativeGpuAVOOB.StorageBuffer
        - NegativeGpuAVOOB.Geometry
        - PositiveShaderSpirv.SpecializationWordBoundryOffset
        - NegativeWsi.SwapchainPresentModeInfoImplicit
        - NegativeWsi.SwapchainMaintenance1ExtensionAcquire
        - PositiveWsi.PresentFenceRetiresPresentQueueOperation
        - NegativeWsi.SwapchainMaintenance1ExtensionCaps
        # The folling tests crash
        - NegativeRenderPass.MultisampledRenderToSingleSampled
        - NegativeTransformFeedback.CmdBindTransformFeedbackBuffersEXT
        - NegativeTransformFeedback.CmdBeginTransformFeedbackEXT
        - NegativeTransformFeedback.CmdEndTransformFeedbackEXT
        - NegativeTransformFeedback.BindPipeline
        - NegativeTransformFeedback.EndRenderPass
        - NegativeTransformFeedback.CmdNextSubpass
        - NegativeWsi.SwapchainMaintenance1ExtensionRelease
        - PositiveWsi.PresentFenceWaitsForSubmission

[r-potter]

@spencer-lunarg If this hasn't already been progressed then feel free to assign to me and I'll see if I can triage VVL vs. driver issues here.

[spencer-lunarg]

I haven't looked at this in months and assigned to you

some notes I have

• NegativeTransformFeedback* these tests are all failing in one spot... but don't have a driver stack trace to figure out
• NegativeWsi*/PositiveWsi* These are all around the VK_EXT_surface_maintenance1 extension and might actually be issues with VVL (or the test)... unfortunately haven't had time to fully grasp what that extension all did
• SpecializationWordBoundryOffset - either driver bug or test is wrong
• *GpuAV* these are a little nastier to debug, would suggest to do these last

[spencer-lunarg]

closing - will track internally in blacklist.yaml instead