Launcher/build-dawn-linux.sh builds the pinned Dawn with Aurora's Vulkan hook
and density map patches into a package local-build.sh takes through the new
--dawn-package option, together with --openxr, --headset steam_frame and
--cpu. docs/steam-frame.md now leads with the native SteamOS build (building
in a podman container on the Frame, running it, the log lines to expect);
OPENXR.md's backend table records the Linux Vulkan backend and why the Lepton
flavour cannot present.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_019HBRGKTE1GnN2ah8gcZKr3
MKW_VR_STANDALONE (runtime_config.h) names the standalone headsets: the
Quest, and the Steam Frame in its Android flavour and its native SteamOS
build. They share render scale 0.8, foveation medium, the GX thread,
the game's object culling, and the headset panel's foveation and
tracked-hands rows; passthrough stays Quest-only.
The Frame's native build also sets AuroraConfig::xrHeadsetOnly, so
Aurora neither presents the desktop window nobody watches nor renders
it past the last pass the eyes sample, which Android does always (4 to
6 ms of a 12 ms GPU frame on a Quest 3).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_019HBRGKTE1GnN2ah8gcZKr3
Lepton's Vulkan driver has no external memory or sync fd extensions,
so the Steam Frame gets a native SteamOS build instead. Its backend is
the PC's: OpenXR creates Dawn's own Vulkan instance and device, and the
eyes are copied into the swapchain on Dawn's queue, with nothing shared
between devices.
- openxr_vulkan_win32 and Aurora's vulkan_win32_interop now compile on
desktop Linux. Dawn links statically there, so the bridge calls the
patched package's hooks directly, and exists only when the package's
aurora-dawn.json declares the Vulkan hook ABI (AuroraDawnProvider
turns AuroraVulkanAbi into AURORA_DAWN_VULKAN_HOOKS); a stock Dawn
keeps the stubs and VR falls back to the desktop.
- openxr_integration.cpp gains a Linux branch: XR_KHR_vulkan_enable2,
the timespec clock, refresh rate, performance level, the Frame
controllers and eye gaze; fragment density maps are requested on
Linux as on the Quest (fdm.cpp and gpu.cpp, AURORA_FDM=0/1 overrides).
- Controller motion uses XR_KHR_convert_timespec_time on any Linux.
- CMake: Vulkan headers are fetched for Linux OpenXR builds, the
headset option is MKW_HEADSET on Android and Linux, and Linux aarch64
builds take MKW_LINUX_CPU (cortex-x4 for the Steam Frame, else native).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_019HBRGKTE1GnN2ah8gcZKr3
The Capability Viewer inside Lepton shows no AHardwareBuffer, external
memory fd, semaphore fd or fence fd extension, so the Quest backend's
two-device eye handoff cannot present there; the eyes need to be copied
on Dawn's own device bound to the session, as on Windows Vulkan.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_019HBRGKTE1GnN2ah8gcZKr3
Read from Walkabout Mini Golf's Unity log on the Frame: the Frame
controller profile, display refresh rate, eye gaze and performance
settings are all there; XR_KHR_convert_timespec_time is not, so VR
frame interpolation is unavailable on the Frame.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_019HBRGKTE1GnN2ah8gcZKr3
Lepton ships /vendor/etc/openxr/1/active_runtime.json and Valve's
implicit fdm_injection layer instead of a runtime broker; the statically
linked Khronos loader falls back to that file, so the app needs no
change.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_019HBRGKTE1GnN2ah8gcZKr3
The Frame's kernel reports sve, sve2 and the SVE2 extensions in its
HWCAP, on SteamOS and inside Lepton alike, unlike phones with the same
Snapdragon 8 Gen 3. The steamFrame flavour now builds with
-mcpu=cortex-x4 instead of cortex-x4+nosve.
docs/steam-frame.md records what the headset reported: 4 KB pages,
Android 11 (API 30) in a Waydroid-based Lepton, Mesa's Turnip as its
Vulkan driver, and SteamVR's native linuxarm64 OpenXR runtime on the
host, with the questions still open.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_019HBRGKTE1GnN2ah8gcZKr3
The adb commands named the device through a PowerShell variable, which
fish and bash do not read; each line now names it, and two grep lines
pull the driver and the relevant extensions out of the vkjson dump.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_019HBRGKTE1GnN2ah8gcZKr3
docs/steam-frame.md covers the steamFrame flavour: the CPU target and
why it excludes SVE, the Lepton entry activity and manifest, the Frame
controller map, the refresh rate and eye-tracked foveation, building and
installing, the device checklist (what to collect from the headset, the
log lines a first session should show, the open questions), and what a
native SteamOS build would need. OPENXR.md documents refresh_rate and
eye_tracked_foveation and the Frame controller profile; the Quest doc,
the Android README and the README point to it.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_019HBRGKTE1GnN2ah8gcZKr3
With [vr] eye_tracked_foveation (on by default on the Steam Frame, off
elsewhere) the runtime asks for XR_EXT_eye_gaze_interaction. When the
system reports an eye tracker, OpenXRInput binds the gaze pose and
locates it for each packet's display time, in the space the eye views
are located in; vr/eye_gaze.h turns it into tangents of each eye's own
view, which AuroraStereoFrame now carries (appended, after the existing
prefix).
Aurora centres the eye's fragment density map on the gaze snapped to a
cell of two map texels (about 3 degrees). Each eye keeps up to 32 maps,
one per cell looked at, so a glance back reuses its map; a new map is
bound once its upload completes, and until then the eye keeps the map
it had. Without a tracked gaze (a blink, no tracker, the setting off)
foveation centres on the forward direction exactly as before: the
forward maps are byte-identical.
Also logs every extension the OpenXR runtime offers at startup, so the
first Steam Frame session shows what SteamVR's Android runtime has.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_019HBRGKTE1GnN2ah8gcZKr3
SteamVR has no XR_FB_passthrough, so the Steam Frame build neither asks
for it nor offers the setting: [vr] passthrough defaults to false there,
and the F10/headset panel and the launcher hide the switch. The Frame
keeps the Android defaults that suit it (render scale 0.8, foveation
medium, the GX thread, object culling, performance level boost) and
starts at 120 Hz (previous commit).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_019HBRGKTE1GnN2ah8gcZKr3
[vr] refresh_rate (Hz, 0 = the headset's own) is requested through
XR_FB_display_refresh_rate each time the session starts running and
whenever the setting changes, matched against the runtime's own list
within half a hertz. Setting it back to 0 restores the rate the session
started at. A rate that is not offered, or one the runtime declines, is
logged and changes nothing.
The game renders 60 frames a second, so at 120 Hz every frame shows for
exactly two refreshes, where 72 or 90 Hz hold some frames longer than
others. The Steam Frame build defaults to 120; every other build keeps
the headset's own rate unless the player picks one.
The F10/headset panel and the Android launcher's Headset section offer
the choice; the interpolation tooltip points at it.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_019HBRGKTE1GnN2ah8gcZKr3
The Steam Frame runs Android apps through Lepton with SteamVR's OpenXR
runtime. A third headset flavour, steamFrame, targets it:
- -mcpu=cortex-x4+nosve for the Snapdragon 8 Gen 3 (its firmware does
not expose SVE, which clang otherwise auto-vectorises with), from one
flavour-to-CPU map that the kit export now reads instead of guessing
from the variant name.
- MKW_ANDROID_HEADSET=steam_frame defines MKW_HEADSET_STEAM_FRAME for
the runtime's own targets, for Frame-specific defaults.
- A manifest without the Horizon OS entries, and FrameEntryActivity as
the single real MAIN/LAUNCHER activity with the Khronos and Oculus VR
categories, which Lepton needs to start an app in VR. It opens the
setup panel and, when the selected game can start, the game on top.
- XR_VALVE_frame_controller_interaction: the Frame controller profile
with its left D-pad (new dpad_* actions, the Wii Remote's D-pad or the
gamepad's), View as menu and the left shoulder as the panel button.
Also requested on Windows for SteamVR streaming to a Frame.
- Build-Quest.ps1, Build-QuestGame.ps1 and Run-Quest.ps1 take
-Headset frame.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_019HBRGKTE1GnN2ah8gcZKr3
- Updated stereo_frame_worker_smoke.cpp to allow dynamic headset rates and prediction lead time.
- Improved logging to include motion diagnostics and adjusted frame submission logic based on headset frequency.
- Enhanced stereo_interpolation_test.cpp with additional tests for camera motion separation and playback cadence.
- Introduced MkwVRReadSceneView function to read the camera view matrix for improved scene rendering.
- Modified VR first-person logic to support scene view reading and validation.
- Added scene_camera.hpp to encapsulate camera motion handling and inverse view calculations.
- Ensured that the VR integration layer correctly logs motion diagnostics and handles scene playback accurately.
- Introduced `vrCockpitItemThrow` configuration option to enable throwing items in VR.
- Implemented `ThrowDetector` and `ThrowSequence` classes to manage item throwing mechanics based on hand movements.
- Updated `OpenXRInput` to handle item throwing based on hand swings, integrating with existing VR input systems.
- Enhanced settings overlay to allow users to toggle item throwing functionality.
- Added tests for item throwing mechanics, ensuring correct behavior for various hand movements and conditions.
- Implemented DVDReadVrAsset function to read mapped disc paths for VR assets.
- Created HeldItem structure and ReadHeldItem function for managing held items in the game.
- Developed unit tests for ReadHeldItem to ensure correct functionality and edge case handling.
- Added cockpit item data tests to validate model indexing and parsing of archives.
- Introduced a new configuration option for object culling in VR, allowing the game to hide objects outside the camera's view.
- Implemented native replacements for the Mario Kart functions responsible for scene culling, ensuring accurate behavior in VR.
- Added a new header file `mkw_vr_culling.h` to define the culling logic and structures.
- Created `mkw_vr_culling.cpp` to implement the culling logic, including frustum intersection checks and screen info updates.
- Updated runtime configuration to include the new object culling option, with appropriate getters and setters.
- Enhanced the settings overlay to allow users to toggle object culling in VR.
- Added tests in `vr_culling_tests.cpp` to validate the frustum intersection logic and ensure compliance with original behavior.
- Updated CMake files to include new source files and tests.
- Added support for hiding model arrays in the rendering pipeline to optimize performance during VR gameplay.
- Introduced new functions in `gx_model_visibility` to manage hidden model arrays and their visibility based on game state.
- Enhanced cockpit recentering logic to ensure accurate seat measurements and eye positioning during gameplay.
- Updated `FirstPersonState` to track cockpit height and forward direction, improving VR experience.
- Added tests for Bullet Bill model visibility and cockpit height adjustments to ensure functionality and stability.
- Refactored existing code to accommodate new features and improve overall code organization.
Bump the backend to 0.2.44 and the Quest app to versionCode 5 / 0.5.0-quest,
the first release to ship Darthcircuit's Quest 1 build beside the modern one.
The headset flavours split the kit export into one task per variant, which
picks the CMake tree by the flavour's MKW_ANDROID_CPU.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Adds the modernQuest and quest1 headset flavours (Kryo CPU target, direct-VR
library entry on Quest 1), records the CPU target in the game kit and checks
it wherever a kit or game package is used, and gates the Quest 1 EFB-copy and
pipeline-scheduling workarounds on the monterey device.
Resolved docs/quest-port.md by keeping both sides: the foveated rendering
section and the Quest 1 renderer compatibility section, and both Build-Quest.ps1
command lines.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- Added SetRenderScale method to both D3D12 and Vulkan backends to adjust the render scale dynamically.
- Updated eye size calculations to reflect the new render scale in both backends.
- Implemented ResizeWritablePair method to handle resizing of swapchains based on the requested eye sizes.
- Modified BeginFrame methods to ensure swapchains are resized appropriately before rendering.
- Enhanced error handling to maintain current eye sizes when the requested sizes cannot be allocated.
- Added tests to validate render scale functionality, ensuring correct behavior when scaling up and down, including handling of refused sizes.
- Started with the controllers connected, a race stayed on them: Horizon OS switches all input
between controllers and hands, and back to the controllers as soon as one lying on a table moves
(a Quest 3 log went touch_controller, simple_controller, touch_controller within seconds). An
app cannot disconnect the controllers.
- With tracked hands on, the input now resumes XR_META_simultaneous_hands_and_controllers (Meta's
multimodal), which overrides that switching: a controller not in a hand no longer owns it, so the
cameras track that hand at once, and a held controller keeps working. Paused again when the
option goes off; a refused resume is logged and not retried until the option is toggled.
- A hand is bare when its squeeze action is inactive and it either drives khr/simple_controller or
has camera-tracked joints, since a free hand under simultaneous tracking may get a profile our
actions are not bound in. Docs: OPENXR.md.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- Now that holding the wheel engages the race controls, closing the other hand on the rim could
bring thumb and index together on the way and fire an item. A pinch uses an item only while the
hand's grasp is under 0.5; a relaxed hand reads about 0, a hand closed on a rim about 0.9.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- On a Quest 3 the race controls never engaged: grabbing the wheel gave no gas and a right pinch
stayed A. They waited for the game to switch the remote's pointer off, and MKW keeps it on in a
race (the log only ever said "the game's pointer is on").
- The race controls now apply while a bare hand holds the wheel: that holds the gas, and a free
hand's pinch (either hand) uses an item. With no bare hand on the wheel a right pinch stays A,
which is what the pause menu and the results need. The flick still works in the whole cockpit.
- The game-pointer publication (KPAD hook and bridge) is gone. The "tracked hands" source line is
logged at most once a second: camera-tracked hands drop in and out of view often.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- A bare hand already feeds no Wii Remote motion (camera-tracked poses are too noisy to
differentiate twice). In a cockpit race, both hands on the wheel rising together or a free bare
hand rising fast now plays one shake on the remote's accelerometer: 150 ms, so the guest sees it
on at least three frames, one cycle up to +2 g and down to the -3.6 g limit, as Dolphin's
emulated shake does. A turn (one hand up, the other down) never flicks, nor does a lone hand on
the wheel; a rise must last three samples and cover 5 cm within 150 ms, tracking jumps reset it,
and flicks are 0.5 s apart. Wii Remote presentation only.
- debug.wiicompiled.inject <n>:flick plays the same shake, with the controllers or unattended, to
tune it apart from the gesture. Docs: OPENXR.md, docs/quest-port.md.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- With tracked hands on, a hand driving khr/simple_controller with camera-tracked joints is bare:
latched through wheel_tracking_grace (Meta drops the select action while a hand is lost),
cleared as soon as a controller's squeeze is back. Its palm joint stands in for the grip and a
grasp from the middle, ring and little fingers' flexion (0 below 1.2 rad, 1 from 3.0) for the
squeeze, so closing a hand on the rim takes hold under the wheel's own press and release. The
grasp never reaches the game's buttons.
- In a cockpit race a bare hand holding the wheel holds the gas (A / South) and a pinch from a free
bare hand uses an item (Z / L) once the hand has been off the wheel for 0.15 s. Only while the
game has the remote's pointer off: KPAD publishes the game's pointer switch for the VR remote,
so the pause menu and the results keep a right pinch as A. The pacing thread logs each change,
to confirm on the headset that MKW turns the pointer off while driving.
- A held bare hand keeps its last joints drawn through a short loss; bare hands get no haptics.
The headset panel's readout shows grasp, hold and pinch. Docs: OPENXR.md.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- New [vr] hand_tracking (default off, Quest only for now): the Quest launcher's Settings > VR and
the headset panel's VR tab, under hand steering. Two XR_EXT_hand_tracking trackers are located
every XR frame: with the controllers held the Quest builds the joints from their touch sensors
(XR_EXT_hand_tracking_data_source's controller source), once they are put down from its
cameras. The trackers exist only while the option and hand steering are on and also serve the
runtime hand mesh; the extensions (plus XR_FB_hand_tracking_aim) are asked for when either is
on at launch.
- Aurora skins the runtime mesh with the joints themselves (tracked pose times inverse bind pose,
no curl, no grip); runtimes with joints but no mesh get a skeleton; non-finite joints put only
that hand back on its grip curl. AuroraCockpitHand carries the 26 seated-frame joints and radii.
- The manifest declares horizonos.permission.HAND_TRACKING (and the deprecated
com.oculus.permission.HAND_TRACKING), both normal permissions with no prompt on a Quest 3, and
oculus.software.handtracking as optional; without it Horizon OS keeps the app controllers-only.
Bare hands then drive khr/simple_controller, so on Android a hand whose squeeze action is
inactive and select active is treated as bare: with the option off it presses nothing but the
menu gesture, is not drawn and feeds no Wii Remote motion.
- Interaction-profile changes and tracker sources are logged. The pure rules (grasp from finger
flexion, bare latch, pinch gate, bare-hand buttons, flick) live in the OpenXR-free
vr/openxr_hand_tracking.h with mkw_vr_hand_tracking_tests; the driving and flick parts are
wired in the next commits. Docs: OPENXR.md, docs/quest-port.md, README.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- The pacing thread aims each immersive window eye through the window itself (AimEyesThroughWindow):
it keeps its position but looks square-on at the window's plane through an off-axis frustum just
around it, so the eye image is the window, at the display's pixel density (about 680x380 per eye
at render_scale 0.8 instead of 1344x1408), with a two-pixel border the mask leaves transparent.
The frame's views carry that pose and field of view to the projection layer.
- vulkan_interop.cpp copies an eye smaller than its AHardwareBuffer into the buffer's corner, and
the Quest layer's imageRect is the rendered part of the swapchain image.
- These eyes are not foveated: their field of view follows the head, which would rebuild the
density map every frame.
- Quest only (kWindowShapedEyesSupported); the PC backends copy whole eyes and keep masking them.
debug.wiicompiled.window_eyes 0 renders them whole and masked again for A/B timing.
- Quest 3, paused Retro Rewind race, render_scale 1.0: Immersive window runs the GPU at level 1
(456 MHz, app GPU 11.0 ms, eyes 8.4 ms) where Immersive needs level 3 (599-640 MHz, 13.4 ms,
10.3 ms), about 42% fewer GPU cycles. No edge artifacts, image as sharp. Docs: OPENXR.md.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- Introduced a new configuration option for immersive window mode in runtime_config.h.
- Updated the parsing and setting functions to handle the immersive window state.
- Modified the OpenXR backend to support rendering with the immersive window, blending the race view with the surrounding environment.
- Enhanced the settings overlay to allow users to select between immersive, immersive window, and flat screen race views.
- Implemented GPU rendering logic for the immersive window mask, ensuring correct visual output in various rendering paths.
- Added tests to validate the immersive window functionality and its interaction with existing race view settings.
- The Quest's GPU applies the fragment density per screen tile and samples textures at a coarser
level inside reduced-rate tiles, so fine animated detail such as Retro Rewind's swamp goo shows
square tile edges at the periphery.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- Retro Rewind's SNES Ghost Valley 2 at render_scale 1.0, GPU-bound at about 40 FPS: merging the
eye passes lifts it from 38.9 to 41.5 FPS (app GPU 23.3 to 21.7 ms), and no foveation level
raises the frame rate further.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- Luigi Circuit Grand Prix start on a Quest 3, same-session A/B through the debug properties:
one render pass per eye saves 12% of eye GPU time at render_scale 0.8 and 8% at 1.3.
- Foveation is neutral at 0.8 (eyes are geometry and full-resolution tile-store bound) and saves
8/14/22% at Low/Medium/High at 1.3, so it stays off by default.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- Eyes render under a VK_EXT_fragment_density_map: full rate around each eye's forward direction,
2x2 then 4x4 pixel blocks towards the edges ([vr] foveation = off|low|medium|high, default off).
XR_FB_foveation cannot help here: the runtime's maps only shape passes drawing into its
swapchain, and the eyes reach it through a copy.
- aurora-main/patches/dawn/aurora_fdm.inc: Dawn enables the extension only on request and for
dynamic rendering, flags every render pipeline, and chains an immutable RG8 map into any pass
whose first color attachment is a view bound to one (ABI: include/aurora/dawn_fdm_abi.h).
- android/Build-QuestDawn.ps1 builds the pinned Dawn revision with those patches for arm64
(dawn-build CI flags, protobuf off) into a cached package; Build-Quest.ps1 links it
(-StockDawn opts out) and AuroraDawnProvider.cmake enables the ABI from its manifest.
- lib/gfx/foveation.hpp generates the maps (32 px per texel, densities 255/127/63); an eye is
foveated only when single_pass_eyes draws it in one render pass. Menus never are.
- Live level from the headset panel's VR tab and the launcher; the launch decides whether the
device has maps. debug.wiicompiled.foveation and debug.wiicompiled.fdm for A/B.
- Tests: Foveation cases in gx_fifo_tests, mkw_vr_config_tests. Docs: OPENXR.md, quest-port.md.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- Added eye_pass_plan.hpp: an eye keeps drawing in the render pass it has open across the frame's
GX copies (only the mono render performs them), skips passes a later clear of the whole color and
depth erases, and splits only for a clear of color alone or depth alone.
- render_stereo_eye follows that plan; the cockpit fallback is drawn inside the last pass instead of
a render pass of its own that loaded the eye back.
- The stencil is now cleared where the eye actually starts, so a cockpit mask drawn in an erased pass
can no longer punch holes in the HUD.
- [vr] single_pass_eyes (default on, live from F10) and debug.wiicompiled.eye_passes 0/1 on the Quest
switch back to one render pass per recorded pass for A/B timing.
- Each new plan shape is logged once ("Eye replay plan: ...").
- Tests: EyePassPlan cases in gx_fifo_tests; stereo_multiplayer_smoke and stereo_frame_worker_smoke
pass on D3D12.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
mitch030504