## Summary - ChatGPT により bench_profile.h の setenv segfault を修正(RTLD_NEXT 経由に切り替え) - core/box/pool_zero_mode_box.h 新設:ENV キャッシュ経由で ZERO_MODE を統一管理 - core/hakmem_pool.c で zero mode に応じた memset 制御(FULL/header/off) - A/B テスト結果:ZERO_MODE=header で +15.34% improvement(1M iterations, C6-heavy) ## Files Modified - core/box/pool_api.inc.h: pool_zero_mode_box.h include - core/bench_profile.h: glibc setenv → malloc+putenv(segfault 回避) - core/hakmem_pool.c: zero mode 参照・制御ロジック - core/box/pool_zero_mode_box.h (新設): enum/getter - CURRENT_TASK.md: Phase ML1 結果記載 ## Test Results | Iterations | ZERO_MODE=full | ZERO_MODE=header | Improvement | |-----------|----------------|-----------------|------------| | 10K | 3.06 M ops/s | 3.17 M ops/s | +3.65% | | 1M | 23.71 M ops/s | 27.34 M ops/s | **+15.34%** | 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Haiku 4.5 <noreply@anthropic.com>
116 lines
3.6 KiB
C
116 lines
3.6 KiB
C
// capacity_box.c - Box Capacity Manager Implementation
|
||
#include "capacity_box.h"
|
||
#include "../tiny_adaptive_sizing.h" // TLSCacheStats, adaptive_sizing_init()
|
||
#include "../hakmem_tiny.h" // g_tls_sll_count
|
||
#include "../hakmem_tiny_config.h" // TINY_NUM_CLASSES, TINY_TLS_MAG_CAP
|
||
#include "../hakmem_tiny_integrity.h" // HAK_CHECK_CLASS_IDX
|
||
#include <stdatomic.h>
|
||
#include <stdio.h>
|
||
#include <stdlib.h>
|
||
|
||
// ============================================================================
|
||
// Internal State
|
||
// ============================================================================
|
||
|
||
// Initialization flag (atomic for thread-safety)
|
||
static _Atomic int g_box_cap_initialized = 0;
|
||
|
||
// External declarations (from adaptive_sizing and hakmem_tiny)
|
||
extern __thread TLSCacheStats g_tls_cache_stats[TINY_NUM_CLASSES]; // TLS variable!
|
||
extern __thread TinyTLSSLL g_tls_sll[TINY_NUM_CLASSES];
|
||
extern int g_sll_multiplier;
|
||
|
||
// ============================================================================
|
||
// Box Capacity API Implementation
|
||
// ============================================================================
|
||
|
||
void box_cap_init(void) {
|
||
// Idempotent: only initialize once
|
||
int expected = 0;
|
||
if (atomic_compare_exchange_strong(&g_box_cap_initialized, &expected, 1)) {
|
||
// First call: initialize adaptive sizing
|
||
adaptive_sizing_init();
|
||
}
|
||
// Already initialized or just initialized: safe to proceed
|
||
}
|
||
|
||
bool box_cap_is_initialized(void) {
|
||
return atomic_load(&g_box_cap_initialized) != 0;
|
||
}
|
||
|
||
uint32_t box_cap_get(int class_idx) {
|
||
// PRIORITY 1: Bounds check
|
||
HAK_CHECK_CLASS_IDX(class_idx, "box_cap_get");
|
||
|
||
// Ensure initialized
|
||
if (!box_cap_is_initialized()) {
|
||
// Auto-initialize on first use (defensive)
|
||
box_cap_init();
|
||
}
|
||
|
||
// Compute SLL capacity using same logic as sll_cap_for_class()
|
||
// This centralizes the capacity calculation(旧 g_sll_cap_override は削除済み)。
|
||
|
||
// Get base capacity from adaptive sizing
|
||
uint32_t cap = g_tls_cache_stats[class_idx].capacity;
|
||
|
||
// Apply class-specific multipliers
|
||
if (class_idx <= 3) {
|
||
// Hot classes: multiply by g_sll_multiplier
|
||
uint32_t mult = (g_sll_multiplier > 0 ? (uint32_t)g_sll_multiplier : 1u);
|
||
uint64_t want = (uint64_t)cap * (uint64_t)mult;
|
||
if (want > (uint64_t)TINY_TLS_MAG_CAP) {
|
||
cap = TINY_TLS_MAG_CAP;
|
||
} else {
|
||
cap = (uint32_t)want;
|
||
}
|
||
} else if (class_idx >= 4) {
|
||
// Mid-large classes: halve capacity
|
||
cap = (cap > 1u ? (cap / 2u) : 1u);
|
||
}
|
||
|
||
return cap;
|
||
}
|
||
|
||
bool box_cap_has_room(int class_idx, uint32_t n) {
|
||
// PRIORITY 1: Bounds check
|
||
HAK_CHECK_CLASS_IDX(class_idx, "box_cap_has_room");
|
||
|
||
uint32_t cap = box_cap_get(class_idx);
|
||
uint32_t used = g_tls_sll[class_idx].count;
|
||
|
||
// Check if adding N would exceed capacity
|
||
if (used >= cap) return false;
|
||
uint32_t avail = cap - used;
|
||
return (n <= avail);
|
||
}
|
||
|
||
uint32_t box_cap_avail(int class_idx) {
|
||
// PRIORITY 1: Bounds check
|
||
HAK_CHECK_CLASS_IDX(class_idx, "box_cap_avail");
|
||
|
||
uint32_t cap = box_cap_get(class_idx);
|
||
uint32_t used = g_tls_sll[class_idx].count;
|
||
|
||
if (used >= cap) return 0;
|
||
return (cap - used);
|
||
}
|
||
|
||
void box_cap_update(int class_idx, uint32_t new_cap) {
|
||
// PRIORITY 1: Bounds check
|
||
HAK_CHECK_CLASS_IDX(class_idx, "box_cap_update");
|
||
|
||
// Ensure initialized
|
||
if (!box_cap_is_initialized()) {
|
||
box_cap_init();
|
||
}
|
||
|
||
// Clamp to max
|
||
if (new_cap > TINY_TLS_MAG_CAP) {
|
||
new_cap = TINY_TLS_MAG_CAP;
|
||
}
|
||
|
||
// Update adaptive sizing stats
|
||
g_tls_cache_stats[class_idx].capacity = new_cap;
|
||
}
|