Bypass cadence retention when the software render adapter is active.
Publish each available source frame immediately with one damage-aware
D3D copy into its final IVSHMEM resource.
Keep only one software copy in flight so newer frames are dropped
instead of queued behind stale work. Retain accumulated damage for the
next frame and skip static re-encodes when no image update is pending.
Use a row-major IVSHMEM texture when the shared heap supports it and
fall back to a direct IVSHMEM buffer copy otherwise. In indirect mode,
copy only damaged rows from readback memory into IVSHMEM.
Check a busy slot fence before waiting for its thread-pool callback.
This lets the caller complete finished GPU work inline and avoids a
scheduler delay before the allocator and command list can be reused.
Signal a per-slot event after callbacks finish and the command slot has
returned to the free state. This removes the one millisecond polling delay
when candidate state becomes visible just before its slot is reusable.
Replace the rotating D3D12 copy queues with two framebuffer-bound
recording slots on one physical COPY queue. Keep allocator, command
list, query range, callback state, and fence target independent per
slot.
Submit source waits, execution, and signaling on the shared timeline.
Track framebuffer ownership through completion and serialize LGMP
publication around the last successfully published frame.
Use calibrated copy-queue timestamps to separate source and effect
waits from the actual framebuffer copy.
Exclude producer readiness waits from client import timing so the
same interval is not counted in both Copy and Import.
The callback runs in a random thread, we can't call directx methods
safely from it, so move reset so it's called automatically when a free
copy list is obtained.
