Phase ALLOC-GATE-OPT-1: tiny_alloc_gate_fast 統計計測

- AllocGateStats 構造体追加（size2class/route/env/class分布）
- malloc_tiny_fast にカウンタ埋め込み
- ENV: HAKMEM_ALLOC_GATE_STATS (default 0)
- 挙動変更なし（計測のみ）

計測結果:
- Mixed: total=542k, size2class=0, route_calls=0, env_checks=275k, C4-C7=95.2%
  - size_to_class/route_for_class は完全削減済み（LUT 効果）
  - C4-C7 が 95% → ULTRA fast path が有効
  - env_checks ≈ c7_calls → C7 ULTRA の ENV gate が毎回呼ばれる
- C6-heavy: total=11 → malloc_tiny_fast はほぼ通らない（mid/pool 主体）

結論:
- alloc gate は既に十分最適化済み（LUT + ULTRA で削減済み）
- さらなる最適化余地は小さい（env_checks は軽量化済み、数%以下の効果）
- 次フェーズでは free dispatcher (29%) や C7 ULTRA refill (7%) など、他のボトルネックを狙う

詳細: docs/analysis/ALLOC_GATE_ANALYSIS.md

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Haiku 4.5 <noreply@anthropic.com>

core/box/alloc_gate_stats_box.c

@@ -0,0 +1,27 @@
+#include "alloc_gate_stats_box.h"
+#include <stdio.h>
+
+AllocGateStats g_alloc_gate_stats = {0};
+
+__attribute__((destructor))
+static void alloc_gate_stats_dump(void) {
+    if (!alloc_gate_stats_enabled()) {
+        return;
+    }
+
+    fprintf(stderr, "[ALLOC_GATE_STATS] total=%lu size2class=%lu route_calls=%lu env_checks=%lu c0=%lu c1=%lu c2=%lu c3=%lu c4=%lu c5=%lu c6=%lu c7=%lu\n",
+            g_alloc_gate_stats.total_calls,
+            g_alloc_gate_stats.size_to_class_calls,
+            g_alloc_gate_stats.route_for_class_calls,
+            g_alloc_gate_stats.env_checks,
+            g_alloc_gate_stats.class_calls[0],
+            g_alloc_gate_stats.class_calls[1],
+            g_alloc_gate_stats.class_calls[2],
+            g_alloc_gate_stats.class_calls[3],
+            g_alloc_gate_stats.class_calls[4],
+            g_alloc_gate_stats.class_calls[5],
+            g_alloc_gate_stats.class_calls[6],
+            g_alloc_gate_stats.class_calls[7]);
+
+    fflush(stderr);
+}

core/box/alloc_gate_stats_box.h

@@ -0,0 +1,43 @@
+#ifndef HAKMEM_ALLOC_GATE_STATS_BOX_H
+#define HAKMEM_ALLOC_GATE_STATS_BOX_H
+
+#include <stdint.h>
+#include <stdbool.h>
+#include <stdlib.h>
+
+typedef struct AllocGateStats {
+    uint64_t total_calls;            // malloc_tiny_fast 入口
+
+    uint64_t size_to_class_calls;    // size→class 変換
+    uint64_t route_for_class_calls;  // class→route 判定
+    uint64_t env_checks;             // alloc-side ENV 判定数
+
+    // クラス別分布
+    uint64_t class_calls[8];         // C0〜C7 呼び出し数
+} AllocGateStats;
+
+// ENV gate
+static inline bool alloc_gate_stats_enabled(void) {
+    static int g_enabled = -1;
+    if (__builtin_expect(g_enabled == -1, 0)) {
+        const char* e = getenv("HAKMEM_ALLOC_GATE_STATS");
+        g_enabled = (e && *e && *e != '0') ? 1 : 0;
+    }
+    return g_enabled;
+}
+
+// Global stats instance
+extern AllocGateStats g_alloc_gate_stats;
+
+// Increment macros (with unlikely guard)
+#define ALLOC_GATE_STAT_INC(field) \
+    do { if (__builtin_expect(alloc_gate_stats_enabled(), 0)) { \
+        g_alloc_gate_stats.field++; \
+    } } while(0)
+
+#define ALLOC_GATE_STAT_INC_CLASS(class_idx) \
+    do { if (__builtin_expect(alloc_gate_stats_enabled(), 0)) { \
+        if ((class_idx) >= 0 && (class_idx) < 8) g_alloc_gate_stats.class_calls[class_idx]++; \
+    } } while(0)
+
+#endif // HAKMEM_ALLOC_GATE_STATS_BOX_H