Feature request - Expose Android Surface swapchain frames to GPU (Vulkan) for compute processing

Hello Meta XR / OpenXR Runtime Team,

I’m developing a Quest 3 OpenXR 8K video player using Vulkan. It uses MediaCodec to decode video frames directly into a runtime-owned Android Surface swapchain (XR_KHR_android_surface_swapchain) in a zero-copy pipeline. This works well for playback, but it prevents the application from running any GPU processing on the decoded frames (e.g., sharpening, denoise, chroma upsampling, lightweight super-resolution) before the runtime compositor samples the surface for an XrCompositionLayerEquirect2KHR layer.

Feature request:
I’d like to request an official, supported way for applications to access the decoded frame content on the GPU (Vulkan) prior to composition, while retaining the benefits of the Android Surface swapchain decode path. The current copy/blit workaround is not acceptable because it adds significant memory bandwidth overhead and defeats the purpose of the zero-copy pipeline.

Sincerely,

Zurab Kargareteli - Graphics Engineer
Vrex Immersive
zurabkargareteli@vreximmersive.com