feat(host): vendor PyroWave + minimal Granite subset as crates/pyrowave-sys

Phase 0 of design/pyrowave-codec-plan.md — the opt-in wired-LAN ultra-low-
latency codec. Vendored at upstream 509e4f88 (API 0.4.0, Granite 44362775,
volk + vulkan-headers pins in PUNKTFUNK-VENDOR.txt), pruned to the 6.6 MB
the standalone no-renderer build needs; scripts/vendor-pyrowave.sh
reproduces the tree (a pin bump is protocol-affecting, plan §4.2).

build.rs drives the wrapper CMakeLists (static archives incl. a static
C-API lib upstream only ships shared) + bindgen over pyrowave.h; Linux and
Windows only, empty stub elsewhere (Apple gets a native Metal port, §4.7).
Offline-safe by construction: no network, no system lib, vendored Vulkan
headers — same model as the opus dep (flatpak builder has no network).

Phase-0 validation on .21 (RTX 5070 Ti, driver 610.43.03):
- upstream pyrowave-c-test + interop test (incl. dmabuf/DRM-modifier
  Vulkan<->Vulkan) pass, from the pristine AND the pruned tree
- GPU kernel times at ~1.6 bpp noise: encode/decode 0.090/0.042 ms @800p,
  0.146/0.067 @1080p, 0.226/0.103 @1440p, 0.477/0.201 @4K — order of
  magnitude under NVENC's 1-2 ms retrieve, CBR lands within ~100 B of
  target
- cargo test -p pyrowave-sys green (static link + API-version pin check)

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

crates/pyrowave-sys/vendor/pyrowave/shaders/dwt_common.h

@@ -0,0 +1,60 @@
+// Copyright (c) 2025 Hans-Kristian Arntzen
+// SPDX-License-Identifier: MIT
+#ifndef DWT_COMMON_H_
+#define DWT_COMMON_H_
+
+const int APRON = 4;
+const int APRON_HALF = APRON / 2;
+const int BLOCK_SIZE = 32;
+const int BLOCK_SIZE_HALF = BLOCK_SIZE >> 1;
+
+#if !FP16 && PRECISION == 0
+#undef PRECISION
+#define PRECISION 1
+#endif
+
+#if PRECISION == 2
+#define FLOAT float
+#define VEC2 vec2
+#define VEC4 vec4
+#define SHARED_VEC2 vec2
+#elif PRECISION == 1
+#define FLOAT float
+#define VEC2 vec2
+#define VEC4 vec4
+#if FP16
+#define SHARED_VEC2 f16vec2
+#else
+#define SHARED_VEC2 uint
+#endif
+#else
+#define FLOAT float16_t
+#define VEC2 f16vec2
+#define VEC4 f16vec4
+#define SHARED_VEC2 f16vec2
+#endif
+
+const FLOAT ALPHA = FLOAT(-1.586134342059924);
+const FLOAT BETA = FLOAT(-0.052980118572961);
+const FLOAT GAMMA = FLOAT(0.882911075530934);
+const FLOAT DELTA = FLOAT(0.443506852043971);
+const FLOAT K = FLOAT(1.230174104914001);
+const FLOAT inv_K = FLOAT(1.0 / 1.230174104914001);
+
+shared SHARED_VEC2 shared_block[(BLOCK_SIZE + 2 * APRON) / 2][(BLOCK_SIZE + 2 * APRON) + 1];
+#if !FP16 && PRECISION == 1
+VEC2 load_shared(uint y, uint x) { return unpackHalf2x16(shared_block[y][x]); }
+void store_shared(uint y, uint x, VEC2 v) { shared_block[y][x] = packHalf2x16(v); }
+#else
+VEC2 load_shared(uint y, uint x) { return VEC2(shared_block[y][x]); }
+void store_shared(uint y, uint x, VEC2 v) { shared_block[y][x] = SHARED_VEC2(v); }
+#endif
+
+bvec2 band(bvec2 a, bvec2 b)
+{
+	return bvec2(a.x && b.x, a.y && b.y);
+}
+
+#include "dwt_swizzle.h"
+
+#endif