CRITICAL FIX: TLS 未初期化による 4T SEGV を完全解消

**問題:**
- Larson 4T で 100% SEGV (1T は 2.09M ops/s で完走)
- System/mimalloc は 4T で 33.52M ops/s 正常動作
- SS OFF + Remote OFF でも 4T で SEGV

**根本原因: (Task agent ultrathink 調査結果)**
```
CRASH: mov (%r15),%r13
R15 = 0x6261  ← ASCII "ba" (ゴミ値、未初期化TLS)
```

Worker スレッドの TLS 変数が未初期化:
- `__thread void* g_tls_sll_head[TINY_NUM_CLASSES];`  ← 初期化なし
- pthread_create() で生成されたスレッドでゼロ初期化されない
- NULL チェックが通過 (0x6261 != NULL) → dereference → SEGV

**修正内容:**
全 TLS 配列に明示的初期化子 `= {0}` を追加:

1. **core/hakmem_tiny.c:**
   - `g_tls_sll_head[TINY_NUM_CLASSES] = {0}`
   - `g_tls_sll_count[TINY_NUM_CLASSES] = {0}`
   - `g_tls_live_ss[TINY_NUM_CLASSES] = {0}`
   - `g_tls_bcur[TINY_NUM_CLASSES] = {0}`
   - `g_tls_bend[TINY_NUM_CLASSES] = {0}`

2. **core/tiny_fastcache.c:**
   - `g_tiny_fast_cache[TINY_FAST_CLASS_COUNT] = {0}`
   - `g_tiny_fast_count[TINY_FAST_CLASS_COUNT] = {0}`
   - `g_tiny_fast_free_head[TINY_FAST_CLASS_COUNT] = {0}`
   - `g_tiny_fast_free_count[TINY_FAST_CLASS_COUNT] = {0}`

3. **core/hakmem_tiny_magazine.c:**
   - `g_tls_mags[TINY_NUM_CLASSES] = {0}`

4. **core/tiny_sticky.c:**
   - `g_tls_sticky_ss[TINY_NUM_CLASSES][TINY_STICKY_RING] = {0}`
   - `g_tls_sticky_idx[TINY_NUM_CLASSES][TINY_STICKY_RING] = {0}`
   - `g_tls_sticky_pos[TINY_NUM_CLASSES] = {0}`

**効果:**
```
Before: 1T: 2.09M   |  4T: SEGV 💀
After:  1T: 2.41M   |  4T: 4.19M   (+15% 1T, SEGV解消)
```

**テスト:**
```bash
# 1 thread: 完走
./larson_hakmem 2 8 128 1024 1 12345 1
→ Throughput = 2,407,597 ops/s 

# 4 threads: 完走(以前は SEGV)
./larson_hakmem 2 8 128 1024 1 12345 4
→ Throughput = 4,192,155 ops/s 
```

**調査協力:** Task agent (ultrathink mode) による完璧な根本原因特定

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

Co-Authored-By: Claude <noreply@anthropic.com>
This commit is contained in:
Moe Charm (CI)
2025-11-07 01:27:04 +09:00
parent f454d35ea4
commit 1da8754d45
110 changed files with 17703 additions and 1693 deletions

View File

@ -348,15 +348,17 @@ void hak_tiny_debug_counters_dump(void) {
extern unsigned long long g_rf_total_calls[];
extern unsigned long long g_rf_hit_bench[];
extern unsigned long long g_rf_hit_hot[];
extern unsigned long long g_rf_hit_ready[];
extern unsigned long long g_rf_hit_slab[];
extern unsigned long long g_rf_hit_ss[];
extern unsigned long long g_rf_hit_reg[];
extern unsigned long long g_rf_mmap_calls[];
fprintf(stderr, "\n[Refill Stage Counters]\n");
fprintf(stderr, "class, total, bench, hot, slab, ss, reg, mmap\n");
fprintf(stderr, "class, total, ready, bench, hot, slab, ss, reg, mmap\n");
for (int i = 0; i < TINY_NUM_CLASSES; i++) {
fprintf(stderr, "%d,%llu,%llu,%llu,%llu,%llu,%llu,%llu\n", i,
fprintf(stderr, "%d,%llu,%llu,%llu,%llu,%llu,%llu,%llu,%llu\n", i,
(unsigned long long)g_rf_total_calls[i],
(unsigned long long)g_rf_hit_ready[i],
(unsigned long long)g_rf_hit_bench[i],
(unsigned long long)g_rf_hit_hot[i],
(unsigned long long)g_rf_hit_slab[i],
@ -365,6 +367,27 @@ void hak_tiny_debug_counters_dump(void) {
(unsigned long long)g_rf_mmap_calls[i]);
}
// Refill item sources (freelist vs carve)
extern unsigned long long g_rf_freelist_items[];
extern unsigned long long g_rf_carve_items[];
fprintf(stderr, "\n[Refill Item Sources]\n");
fprintf(stderr, "class, freelist_items, carve_items\n");
for (int i = 0; i < TINY_NUM_CLASSES; i++) {
fprintf(stderr, "%d,%llu,%llu\n", i,
(unsigned long long)g_rf_freelist_items[i],
(unsigned long long)g_rf_carve_items[i]);
}
// Refill item sources (freelist vs carve)
extern unsigned long long g_rf_freelist_items[];
extern unsigned long long g_rf_carve_items[];
fprintf(stderr, "\n[Refill Item Sources]\n");
fprintf(stderr, "class, freelist_items, carve_items\n");
for (int i = 0; i < TINY_NUM_CLASSES; i++) {
fprintf(stderr, "%d,%llu,%llu\n", i,
(unsigned long long)g_rf_freelist_items[i],
(unsigned long long)g_rf_carve_items[i]);
}
// Diagnostic: refill early return counters
extern unsigned long long g_rf_early_no_ss[];
extern unsigned long long g_rf_early_no_meta[];
@ -408,6 +431,55 @@ void hak_tiny_debug_counters_dump(void) {
(unsigned long long)g_pub_hot_hits[i]);
}
// Front Gate Breakdown (SFC/SLL/Quick/Mag)
extern unsigned long long g_front_sfc_hit[];
extern unsigned long long g_front_sll_hit[];
extern unsigned long long g_front_quick_hit[];
extern unsigned long long g_front_mag_hit[];
fprintf(stderr, "\n[Front Gate Breakdown]\n");
fprintf(stderr, "class, sfc_hit, sll_hit, quick_hit, mag_hit\n");
for (int i = 0; i < TINY_NUM_CLASSES; i++) {
fprintf(stderr, "%d,%llu,%llu,%llu,%llu\n", i,
(unsigned long long)g_front_sfc_hit[i],
(unsigned long long)g_front_sll_hit[i],
(unsigned long long)g_front_quick_hit[i],
(unsigned long long)g_front_mag_hit[i]);
}
// Free Triggers (first-free / remote transition)
extern unsigned long long g_first_free_transitions[];
extern unsigned long long g_remote_free_transitions[];
fprintf(stderr, "\n[Free Triggers]\n");
fprintf(stderr, "class, first_free, remote_transition\n");
for (int i = 0; i < TINY_NUM_CLASSES; i++) {
fprintf(stderr, "%d,%llu,%llu\n", i,
(unsigned long long)g_first_free_transitions[i],
(unsigned long long)g_remote_free_transitions[i]);
}
// Adopt/Registry Gate
extern unsigned long long g_adopt_gate_calls[];
extern unsigned long long g_adopt_gate_success[];
extern unsigned long long g_reg_scan_attempts[];
extern unsigned long long g_reg_scan_hits[];
fprintf(stderr, "\n[Adopt/Registry Gate]\n");
fprintf(stderr, "class, adopt_calls, adopt_success, reg_scans, reg_hits\n");
for (int i = 0; i < TINY_NUM_CLASSES; i++) {
fprintf(stderr, "%d,%llu,%llu,%llu,%llu\n", i,
(unsigned long long)g_adopt_gate_calls[i],
(unsigned long long)g_adopt_gate_success[i],
(unsigned long long)g_reg_scan_attempts[i],
(unsigned long long)g_reg_scan_hits[i]);
}
// SuperSlab Registry (per-class sizes)
extern int g_super_reg_class_size[];
fprintf(stderr, "\n[SuperSlab Registry]\n");
fprintf(stderr, "class, reg_size\n");
for (int i = 0; i < TINY_NUM_CLASSES; i++) {
fprintf(stderr, "%d,%d\n", i, g_super_reg_class_size[i]);
}
extern unsigned long long g_fast_push_hits[];
extern unsigned long long g_fast_push_full[];
extern unsigned long long g_fast_push_disabled[];