[DD001] Replace the temporary "soft floor" with proper boundary physics on Metal (Ihmsen 2010)

[slarson]

• Roadmap Phase: Phase 0
• Required Capabilities: physics, sph, metal
• Audit note (NEW issue surfaced by audit): The current worm_alone parity uses floor_y=2.25 as a soft proxy because the native substrate doesn't yet implement Ihmsen 2010 boundary→active SPH pressure (per docs/demos.md parity caveats). This is the real next physics step on the Metal substrate, not just an artifact-housekeeping item.
• DD Section to Read: DD001 §Backend Stabilization Roadmap
• Depends On: None ([DD001] Add a parity check for worm_alone (passive worm body) — depends on #236 #236 documents the caveat that motivates this)
• Existing Code to Reuse:
  • OpenCL reference implementation of boundary pressure in src/sphFluid.cl
  • src/metal_diff/shaders.metal for the kernel layer to add to
• Acceptance Criteria:
  • Implement boundary→active SPH pressure kernel in shaders.metal per Ihmsen 2010
  • Paired analytic backward kernel + FD validator (per DD001 Quality Criteria Configreaderwriter #7)
  • worm_alone parity test now passes with physical floor_y (not the elevated proxy value)
  • Tuned param updates committed
• Sponsor Summary Hint: Currently the worm-alone parity test passes with a floor_y elevated above the real floor — a workaround for missing physics. This issue is the actual physics: boundary particles applying SPH pressure to active particles per Ihmsen 2010.

---

This issue is part of milestone v0.0.9 — Phase 0 Modernization, derived from DD001 Body Physics Engine Architecture. Related issues in this milestone are filed in the same batch.