Investigated the DX12 present-thread overhead (~0.38 ms vs DX11 ~0.05 / GL ~0.09).
Per-stage timing of the D3D11On12 path showed the cost is NOT where you'd assume:
fence 0.005 + wrap(CreateWrappedResource+Acquire) 0.012 + copy 0.133 + flush 0.057 ms
CreateWrappedResource is cheap. The cost is the CopyResource issued on the 11On12
immediate context plus the mandatory Flush to make the shared copy visible to the
host -- both inherent to the bridge and not paid by the native-D3D11 path. The
per-frame GetDevice can't be skipped either (it's how device recreation is
detected). Documented this in the capture path.
Improving it means a native-D3D12 copy-queue path into a D3D12-shared texture, but
the host consumes the shared surface via IDXGIKeyedMutex (a D3D11 concept), so that
also requires switching the DX12 producer<->host sync to a shared ID3D12Fence -- a
cross-API rewrite. Deferred: the overhead is ~5% of a 144 Hz frame and correct.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Spamming a subsystem toggle (the "Mirror video" button) could crash the game:
remove_*_hooks freed a hook's shared D3D / Vulkan / IPC state immediately, while a
capture detour was still mid-flight on the game's render thread -> use-after-free.
Only the audio hooks had the safe-unhook drain; the video (Present/D3D9/D3D10/GL/
Vulkan) and XInput/focus/MKB hooks did not.
Test-first: mock_game_test now runs an aggressive hook/unhook storm -- a separate
thread thrashes every subsystem on/off while the game presents, across all backends.
It crashed gl + vk (0xC0000005) and failed dx9 capture-resume before the fix.
Fix (hook/src/hook_guard.hpp, DetourGate): each detour wraps its body in an RAII
active-count Guard; remove_* restores the hook first (so no new detour starts),
drains the in-flight detours to zero, and only then frees the shared state. Vulkan
is special-cased -- the game caches hk_vkQueuePresentKHR, so removal closes an
atomic capture gate (detours then pass through to the real present), drains, then
frees the read-back resources. Storm now passes on every backend.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
A pure-D3D10 game's backbuffer QIs to ID3D11Texture2D but that view reads back
empty (content lives on the game's D3D10 device), and a D3D11 backbuffer also
QIs to ID3D10Texture2D -- so GetBuffer can't discriminate. The reliable signal
is that a feature-level-10 device rejects CreateTexture2D with the NT-handle
keyed-mutex share flags (E_INVALIDARG). The present hook now tries the D3D11
fast path and, on that failure, switches (sticky) to reading the backbuffer
through the game's own D3D10 device into a staging texture and uploading it into
the shared texture on a hook-owned D3D11 device (Map blocks until the GPU copy
completes -> no cross-device race). The host reader is unchanged.
mock_game_test now decodes DX10 frames through the hook (monotonic/advancing)
alongside DX11/DX12; 15/15 ctest. Roadmap: DX10 milestone done and removed
(remaining renumbered); architecture + lessons updated.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Resolves the DX12 mirror stutter and makes dropped frames observable.
Decouple the D3D11On12 copy from the game's present queue. Submitting the
copy on the game's own present queue (the prior approach) ordered it
correctly but stalled the game's presents: GPU back-pressure, plus the
shared keyed-mutex AcquireSync is a CPU-blocking call on the render
thread. Running it on an independent queue avoids the stall but races the
game's render -> stale frames. Do both: run the copy on our own queue and
order it after the frame with an ID3D12Fence the game's present queue
signals (near-free) and our queue waits on. The present queue is still
recovered for late injection via the ExecuteCommandLists hook (now used to
signal the fence, not host the copy). Producer AcquireSync stays
non-blocking (timeout 0) so a busy mutex drops a mirror frame instead of
stalling the game.
Add drop detection (protocol v12 -> v13). The hook counts captures skipped
because the keyed mutex was busy (VideoShare.frames_dropped); the host
counts published frames it never displayed (generation gaps). The Video
panel shows "Frames lost: N/s capture N/s display", red when nonzero.
This confirmed the game-window-vs-mirror behavior is a display-path
artifact (unfocused windows lose VRR/independent flip), not a capture loss.
Add a one-shot present-pattern log: per distinct swapchain (size/format/
buffer index) and per distinct present-flags value, with DXGI_PRESENT_TEST
spelled out as an occlusion probe that draws nothing -- which is why
Miles Morales shows ~2 presents per captured frame (the test present is
counted but produces no frame).
Docs: add the DX12 capture lessons to the README (rotating back buffer,
fence/own-queue, capture-at-Present decoupling from DWM) and drop the now
-moot DX12 overhead future-work item.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Even copying the correct back buffer, the DX12 path occasionally showed a
several-frames-old frame under GPU load (visible when the game window has
true focus and the camera is whipped around with the mouse). Cause: the
On12 bridge submitted CopyResource on a command queue of our own, which
knows nothing about the game's queue. With frames in flight, our copy
could race ahead of the game's render of that buffer and capture its
previous (rotated) contents. The DX11 path never had this because it
copies on the game's immediate context, ordered after the frame.
Fix: submit the copy on the game's actual present queue so it's ordered
after the frame's rendering, matching the DX11 path. Recover the queue by
inline-hooking ID3D12CommandQueue::ExecuteCommandLists (the per-frame
method, not swapchain/queue creation) so it works for late injection --
the queue already exists when we attach. Record the last DIRECT queue
seen, preferring the one on the render thread (Present and its queue's
ExecuteCommandLists share that thread); the atomic is a cross-thread
fallback. Thread it into the On12 bridge, rebuilding if the captured
queue changes, and fall back to our own queue until it's captured.
Resolves D3D12CreateDevice dynamically from an already-loaded d3d12.dll,
so it adds no link dependency and no-ops for D3D11 / x86 games.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The D3D12 capture path grabbed GetBuffer(0) every Present. Unlike D3D11
flip-model -- where DXGI keeps GetBuffer(0) pointing at the live back
buffer -- D3D12 rotates buffers explicitly: the game renders into the
buffer at GetCurrentBackBufferIndex(), which advances each Present. So
buffer 0 only holds fresh content every Nth frame; the rest copied a
stale buffer, and the mirror silently ran at refresh/N with duplicate
frames in between.
Every metric read full rate (Present counter, published FPS, generation
bump, capture->display latency) because they count Presents, not unique
content -- which is why it looked fine but felt like missing frames,
especially on high-refresh DX12 games (DMC5/Myst at 144, Miles Morales).
Query IDXGISwapChain3::GetCurrentBackBufferIndex() before the trampoline
Present (so it's the just-rendered buffer) and copy that one; fall back
to 0 only if the interface is unavailable. The DX11 path is unaffected.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
D3D12 games (e.g. Spider-Man: Miles Morales) fire the Present hook -- the DXGI
swapchain's Present is the same vtable function for D3D11 and D3D12 -- but
GetBuffer(0) as ID3D11Texture2D fails, so the hook used to idle. Now, when the D3D11
GetBuffer fails, present_hook bridges via D3D11On12: it gets the game's ID3D12Device
from the backbuffer, creates its own DIRECT command queue on it (no need to hook the
game's ExecuteCommandLists), builds an ID3D11On12Device, CreateWrappedResource's the
D3D12 backbuffer, and CopyResource's it into the existing shared keyed-mutex texture
-- so the host side is unchanged. The bridge is created lazily and torn down with the
hook.
Verified with a new in-process dx12_present_hook_test: it drives a real D3D12
swapchain (clears a backbuffer, Presents) and asserts present fired, the backbuffer
was bridged into the shared texture, and a second device reads the exact color back
by name -- {51,102,153,255}. Full build x64 + x86 clean (the x86 hook compiles the
D3D12 path too); ctest x64 10/10, x86 3/3.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
SafetyHook's InlineHook::call() invokes the trampoline through a __cdecl
pointer (the compiler default on x86). The functions we hook are __stdcall
(IDXGISwapChain::Present/Present1, the WASAPI render interfaces, and the
WINAPI SwapBuffers/wglSwapBuffers), so on 32-bit both sides cleaned the
stack -> ESP imbalance -> Run-Time Check Failure #0 and an instant crash.
On x64 every convention collapses to one, so it only bit 32-bit games:
Slaps and Beans (Unity/Rewired, 32-bit D3D11) froze the moment the Present
hook ran. The user's "crashes as soon as a button is pressed" was the
Present, not the button.
Switch every __stdcall trampoline call to SafetyHook's stdcall() (a no-op
on x64). The XInput/focus hooks were unaffected because they never call
the trampoline -- they return synthesized data.
Reproduction + regression coverage:
- tools/input_probe (coop_input_probe): injects, reports a connected pad,
toggles a button, and takes a disable_mask to bisect which subsystem
affects a game. Isolated the freeze to the video subsystem live.
- hook_selftest_x86 + present_hook_test_x86: the x86 sub-build now builds
and runs these (the x64 present_hook_test can't see a one-convention
bug). present_hook_test_x86 drives a real swapchain through the
trampoline -- it would hit RTC #0 before this fix.
- hook_selftest strengthened to exercise every loaded xinput DLL's full
export set (GetState, ordinal-100 GetStateEx, GetCapabilities, rumble
SetState) and to dump the SharedBlock layout.
- protocol.hpp: static_asserts lock the cross-bitness front-of-block
offsets (verified byte-identical on x86 and x64).
README roadmap trimmed (this milestone done) and a lessons-learned note
added on the call()/stdcall() convention trap.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Add an injected IDXGISwapChain::Present / Present1 hook as a lower-latency,
border-free alternative to WGC. The hook copies the swapchain backbuffer into a
shared keyed-mutex texture (coop_video_<pid>); the host opens it by name and
samples it. New opt-in HookSubsys_Video (protocol v6 -> v7); the Video mirror
panel gains a WGC vs Hooked source toggle that installs/removes the subsystem.
Verified by present_hook_test (drives a real D3D11 swapchain end-to-end and reads
the rendered pixels back through the shared texture) and against Phantom Brave
(D3D9: hook installs cleanly and stays idle, WGC fallback). All 5 tests pass.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
