Tiny: fix header/stride mismatch and harden refill paths

- Root cause: header-based class indexing (HEADER_CLASSIDX=1) wrote a 1-byte
  header during allocation, but linear carve/refill and initial slab capacity
  still used bare class block sizes. This mismatch could overrun slab usable
  space and corrupt freelists, causing reproducible SEGV at ~100k iters.

Changes
- Superslab: compute capacity with effective stride (block_size + header for
  classes 0..6; class7 remains headerless) in superslab_init_slab(). Add a
  debug-only bound check in superslab_alloc_from_slab() to fail fast if carve
  would exceed usable bytes.
- Refill (non-P0 and P0): use header-aware stride for all linear carving and
  TLS window bump operations. Ensure alignment/validation in tiny_refill_opt.h
  also uses stride, not raw class size.
- Drain: keep existing defense-in-depth for remote sentinel and sanitize nodes
  before splicing into freelist (already present).

Notes
- This unifies the memory layout across alloc/linear-carve/refill with a single
  stride definition and keeps class7 (1024B) headerless as designed.
- Debug builds add fail-fast checks; release builds remain lean.

Next
- Re-run Tiny benches (256/1024B) in debug to confirm stability, then in
  release. If any remaining crash persists, bisect with HAKMEM_TINY_P0_BATCH_REFILL=0
  to isolate P0 batch carve, and continue reducing branch-miss as planned.

run_pool_bench.sh

@@ -0,0 +1,92 @@
+#!/usr/bin/env bash
+# run_pool_bench.sh - Pool TLS Phase 1.5a パフォーマンステスト
+#
+# Pool TLS の性能を System malloc / mimalloc と比較
+
+set -euo pipefail
+
+echo "========================================="
+echo " 🧪 Pool TLS Phase 1.5a Benchmark"
+echo "========================================="
+echo ""
+
+# 1. Build Pool TLS version
+echo "📦 Building Pool TLS Phase 1.5a..."
+./build_pool_tls.sh bench_mid_large_mt_hakmem >/dev/null 2>&1
+echo "✓ HAKMEM built"
+
+# 2. Build System malloc version
+echo "📦 Building System malloc version..."
+./build.sh bench_mid_large_mt_system >/dev/null 2>&1
+echo "✓ System malloc built"
+
+echo ""
+echo "========================================="
+echo " 📊 Running Benchmarks"
+echo "========================================="
+echo ""
+
+# Test parameters
+THREADS=1
+CYCLES=100000
+WORKSET=256
+SEED=42
+
+# Run HAKMEM
+echo "🔵 HAKMEM (Pool TLS Phase 1.5a)..."
+HAKMEM_OUTPUT=$(./bench_mid_large_mt_hakmem $THREADS $CYCLES $WORKSET $SEED 2>&1)
+HAKMEM_RESULT=$(echo "$HAKMEM_OUTPUT" | grep "Throughput" | awk '{print $3}')
+HAKMEM_TIME=$(echo "$HAKMEM_OUTPUT" | grep "Throughput" | awk '{print $NF}' | tr -d 's.')
+
+# Run System
+echo "🟢 System malloc..."
+SYSTEM_OUTPUT=$(./bench_mid_large_mt_system $THREADS $CYCLES $WORKSET $SEED 2>&1)
+SYSTEM_RESULT=$(echo "$SYSTEM_OUTPUT" | grep "Throughput" | awk '{print $3}')
+SYSTEM_TIME=$(echo "$SYSTEM_OUTPUT" | grep "Throughput" | awk '{print $NF}' | tr -d 's.')
+
+# Calculate performance ratio
+RATIO=$(echo "scale=2; $HAKMEM_RESULT / $SYSTEM_RESULT * 100" | bc)
+SPEEDUP=$(echo "scale=2; $HAKMEM_RESULT / $SYSTEM_RESULT" | bc)
+
+# Determine status emoji
+if (( $(echo "$RATIO >= 100" | bc -l) )); then
+    STATUS="🏆"
+    VERDICT="HAKMEM WINS!"
+elif (( $(echo "$RATIO >= 80" | bc -l) )); then
+    STATUS="✅"
+    VERDICT="Competitive"
+elif (( $(echo "$RATIO >= 50" | bc -l) )); then
+    STATUS="⚠️ "
+    VERDICT="Below target"
+else
+    STATUS="❌"
+    VERDICT="Needs optimization"
+fi
+
+# Pretty print results
+echo ""
+echo "========================================="
+echo " 📊 Performance Results"
+echo "========================================="
+printf "%-25s %12s ops/s  (%.3fs)\n" "HAKMEM (Pool TLS):" "$HAKMEM_RESULT" "$HAKMEM_TIME"
+printf "%-25s %12s ops/s  (%.3fs)\n" "System malloc:" "$SYSTEM_RESULT" "$SYSTEM_TIME"
+echo "----------------------------------------"
+printf "%-25s %11s%%  (${SPEEDUP}x)\n" "Performance ratio:" "$RATIO"
+echo "========================================="
+echo " $STATUS $VERDICT"
+echo "========================================="
+echo ""
+
+# Expected performance targets
+echo "📈 Pool TLS Phase 1.5a Targets:"
+echo "  - Current:  1.0-2.0M ops/s (baseline)"
+echo "  - Phase 1.5b: 5-15M ops/s (with optimizations)"
+echo "  - Phase 2:  15-30M ops/s (full learning)"
+echo ""
+
+# Exit code based on performance
+if (( $(echo "$RATIO >= 50" | bc -l) )); then
+    exit 0
+else
+    exit 1
+fi