Retained native candidates and framebuffer targets are both linear
buffers with the same pitched layout. Publish partial damage using
CopyBufferRegion with byte offsets instead of encoding the ranges as
texture-footprint copies.
Collapse full-width rectangles into one command. Fall back to a full
copy when the partial path would exceed 256 commands, copy at least a
full frame of data, or receive invalid damage coordinates.
This removes the native-only command path introduced by 650636cc that
can fault the D3D12 copy queue after RGB24 benchmarking selects native.
Make CSRWLock own the native lock and provide scoped guards for shared,
exclusive, early-unlock, and non-blocking use.
Replace direct SRW lock management throughout the IDD, input driver,
and helper while preserving the existing lock scopes.
Move cadence and pipeline declarations under capture, and move local
named-pipe handling under ipc.
Flatten display contexts and rename files after their contained classes.
Keep transport headers limited to transport contracts and capabilities.
Consolidate CSwapChainProcessor definitions so each source file matches
the class it implements.
Introduce transport, frame, and control interfaces with an LGMP factory
backend.
Move LGMP and IVSHMEM implementation details under transport/lgmp and
expose direct frame-buffer memory through a neutral capability.
Group the IDD sources and Visual Studio filters by subsystem.
Split the device and swap-chain implementations into focused units,
rename the context classes, and reduce header coupling.
