[Isrrael101]

What

The GPU selector accepts a discrete adapter with graphics, compute and presentation queues before checking its required BDA features. WSL's Dozen adapter for an RTX 3060 is selected despite lacking VK_KHR_8bit_storage, so device initialization fails instead of using the available compatible lavapipe device.

Probe the existing byte-storage and buffer-device-address requirements before accepting each candidate. Log the rejection reason and continue selecting among compatible adapters. Explicit ANYPS5_GPU filtering still applies. Document the selection requirements.

Tested

Ubuntu 26.04 under WSL, Intel i7-11370H, Mesa 26.2.4. Full relinker and patched system-library build completed.

• agc_driver_device_selection passed with Dozen and lavapipe both installed; the unsupported NVIDIA and Intel Dozen adapters were skipped.
• agc_driver_bda_device passed with VK_DRIVER_FILES=/usr/share/vulkan/icd.d/lvp_icd.json (the default adapter skips this test because it lacks byte storage).
• Eight existing tests covering guest exception handlers, raised exceptions, address synchronization, _umtx_op, guest module directories and Intel conversion passed.
• Conventions and git diff --check passed.
• 7 Days to Die - Console Edition (PPSA16645) rendered its loading screen and reached the license-agreement screen with automatic lavapipe selection. A five-minute run and a subsequent six-minute run completed without the startup abort, ending at the configured time limits. The agreement was not accepted. Gameplay and online services are not verified; this is not a full compatibility claim.

Checklist

• Based on current main; no other open PR implements the same functions
• One topic per PR; follow-ups go in a new PR
• No comments in code except technical debt
• Unimplemented paths throw; no silent stubs added
• No notes, investigation files or images added to the repository
• No new third-party code
• Generic behaviour, not specific to one title
• Depends on: none
• Unblocks: PPSA16645 startup GPU selection; no existing Game bring-up discussion found; gameplay remains unverified
• AI-assisted: yes, with Codex

[github-actions]

Some comparisons could not be completed because a pull request conflicts with the base branch. Review these shared files manually before concluding that there is no overlap.

Pull request	Shared files not compared
PR 1395	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp
PR 2434	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp
PR 4040	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp
PR 4472	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp
PR 5021	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp docs/user/USAGE.md
PR 5925	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp docs/user/USAGE.md
PR 6188	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp
PR 6804	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp
PR 6924	docs/user/USAGE.md
PR 7275	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp
PR 7449	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp
PR 7464	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp
PR 7642	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp docs/user/USAGE.md
PR 8004	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp
PR 9345	docs/user/USAGE.md
PR 9356	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp
PR 9387	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp
PR 9898	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp
PR 10193	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp docs/user/USAGE.md
PR 10442	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp
PR 10479	docs/user/USAGE.md
PR 10861	docs/user/USAGE.md
PR 10867	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp
PR 11351	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp
PR 11435	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp
PR 11436	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp
PR 11483	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp
PR 12120	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp
PR 12567	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp
PR 12678	docs/user/USAGE.md
PR 12746	docs/user/USAGE.md
PR 12828	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp
PR 12863	docs/user/USAGE.md
PR 12930	docs/user/USAGE.md
PR 12944	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp
PR 12970	docs/user/USAGE.md
PR 13148	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp
PR 13646	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp
PR 13707	docs/user/USAGE.md
PR 13863	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp
PR 13920	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp
PR 13928	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp docs/user/USAGE.md
PR 13929	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp docs/user/USAGE.md
PR 14217	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp
PR 14218	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp
PR 14264	docs/user/USAGE.md
PR 14344	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp
PR 14943	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp

[BunnyFl]

Reads right. QueryBdaByteFeatures and QueryBdaFeatures refuse a missing extension or feature through Graphics::Require, which throws std::runtime_error (Graphics/include/Context.hpp:31-41), so the new catch in the selection loop sees exactly those refusals and moves to the next device. The device chosen is then queried again with the same two calls (Execution/src/VulkanDevice.cpp:1061-1062 at this head). Those calls throw on the same conditions there (VulkanDevice.cpp:1061-1062), so a device skipped by the new check would have failed that setup anyway and no previously working GPU is lost. With ANYPS5_GPU naming only an unsupported device, selection still ends in the "no usable device whose name contains ANYPS5_GPU=..." error (:1027), after the skip reason is logged.

Tested on Windows 11, WinLibs GCC 15.2.0, with this head merged onto main: agc_driver_device_selection, agc_driver_bda_device and agc_driver_bda_device_selection_cpu pass, nothing skipped. The ctest output of this run has no "Skipping physical device" line. The Dozen case itself wasn't reproduced here.