// hak_free_api.inc.h — Box: hak_free_at() implementation
#ifndef HAK_FREE_API_INC_H
#define HAK_FREE_API_INC_H

#include "hakmem_tiny_superslab.h"  // For SUPERSLAB_MAGIC, SuperSlab
#include "../tiny_free_fast_v2.inc.h"  // Phase 7: Header-based ultra-fast free

#ifdef HAKMEM_POOL_TLS_PHASE1
#include "../pool_tls.h"
#endif

// Optional route trace: print first N classification lines when enabled by env
static inline int hak_free_route_trace_on(void) {
    static int g_trace = -1;
    if (__builtin_expect(g_trace == -1, 0)) {
        const char* e = getenv("HAKMEM_FREE_ROUTE_TRACE");
        g_trace = (e && *e && *e != '0') ? 1 : 0;
    }
    return g_trace;
}
static inline int* hak_free_route_budget_ptr(void) {
    static int g_budget = 32; // first 32 frees only
    return &g_budget;
}
static inline void hak_free_route_log(const char* tag, void* p) {
    if (!hak_free_route_trace_on()) return;
    int* budget = hak_free_route_budget_ptr();
    if (*budget <= 0) return;
    (*budget)--;
    fprintf(stderr, "[FREE_ROUTE] %s ptr=%p\n", tag, p);
}

// Optional: request-trace for invalid-magic cases (first N hits)
static inline int hak_super_reg_reqtrace_on(void) {
    static int g_on = -1;
    if (__builtin_expect(g_on == -1, 0)) {
        const char* e = getenv("HAKMEM_SUPER_REG_REQTRACE");
        g_on = (e && *e && *e != '0') ? 1 : 0;
    }
    return g_on;
}
static inline int* hak_super_reg_reqtrace_budget_ptr(void) {
    static int g_budget = 16; // trace first 16 occurrences
    return &g_budget;
}
static inline void hak_super_reg_reqtrace_dump(void* ptr) {
    if (!hak_super_reg_reqtrace_on()) return;
    int* b = hak_super_reg_reqtrace_budget_ptr();
    if (*b <= 0) return;
    (*b)--;
    uintptr_t p = (uintptr_t)ptr;
    uintptr_t m20 = ((uintptr_t)1 << 20) - 1;
    uintptr_t m21 = ((uintptr_t)1 << 21) - 1;
    SuperSlab* s20 = (SuperSlab*)(p & ~m20);
    SuperSlab* s21 = (SuperSlab*)(p & ~m21);
    unsigned long long mg20 = 0, mg21 = 0;
    // Best-effort reads (may be unmapped; wrap in volatile access)
    mg20 = (unsigned long long)(s20 ? s20->magic : 0);
    mg21 = (unsigned long long)(s21 ? s21->magic : 0);
    fprintf(stderr,
            "[SUPER_REG_REQTRACE] ptr=%p base1M=%p magic1M=0x%llx base2M=%p magic2M=0x%llx\n",
            ptr, (void*)s20, mg20, (void*)s21, mg21);
}

#ifndef HAKMEM_TINY_PHASE6_BOX_REFACTOR
__attribute__((always_inline))
inline
#endif
void hak_free_at(void* ptr, size_t size, hak_callsite_t site) {
#if HAKMEM_DEBUG_TIMING
    HKM_TIME_START(t0);
#endif
    (void)site; (void)size;
    if (!ptr) {
#if HAKMEM_DEBUG_TIMING
        HKM_TIME_END(HKM_CAT_HAK_FREE, t0);
#endif
        return;
    }

#ifdef HAKMEM_POOL_TLS_PHASE1
    // Phase 1: Try Pool TLS free FIRST for 8KB-52KB range
    // CRITICAL: Must come before Phase 7 Tiny to avoid magic mismatch SEGV
    // Pool TLS uses magic 0xb0, Tiny uses magic 0xa0 - must distinguish!
    {
        void* header_addr = (char*)ptr - 1;

        // Safety vs performance trade-off:
        // - If HAKMEM_TINY_SAFE_FREE=1 (strict), validate with mincore() always
        // - Else (default), only validate on page-boundary risk to avoid syscall cost
        #if HAKMEM_TINY_SAFE_FREE
        if (!hak_is_memory_readable(header_addr)) { goto skip_pool_tls; }
        #else
        uintptr_t off = (uintptr_t)header_addr & 0xFFF;
        if (__builtin_expect(off == 0, 0)) {
            if (!hak_is_memory_readable(header_addr)) { goto skip_pool_tls; }
        }
        #endif

        uint8_t header = *(uint8_t*)header_addr;

        if ((header & 0xF0) == POOL_MAGIC) {
            pool_free(ptr);
            hak_free_route_log("pool_tls", ptr);
            goto done;
        }
        // Not Pool TLS - fall through to other paths
    }
skip_pool_tls:
#endif

#if HAKMEM_TINY_HEADER_CLASSIDX
    // Phase 7: Dual-header dispatch (1-byte Tiny header OR 16-byte malloc/mmap header)
    //
    // Step 1: Try 1-byte Tiny header (fast path: 5-10 cycles)
    if (__builtin_expect(hak_tiny_free_fast_v2(ptr), 1)) {
        hak_free_route_log("header_fast", ptr);
#if !HAKMEM_BUILD_RELEASE
        hak_free_v2_track_fast();  // Track hit rate in debug
#endif
        goto done;  // Success - done in 5-10 cycles! NO SuperSlab lookup!
    }

    // Step 2: Try 16-byte AllocHeader (malloc/mmap allocations)
    // CRITICAL: Must check this BEFORE calling hak_tiny_free() to avoid silent failures!
    {
        void* raw = (char*)ptr - HEADER_SIZE;

        // SAFETY: Check if raw header is accessible before dereferencing
        // This prevents SEGV when malloc metadata is unmapped
        //
        // OPTIMIZATION: raw = ptr - HEADER_SIZE (16 bytes)
        // Page boundary case: if ptr is in first 16 bytes of page, raw crosses page boundary
        // Check: (ptr & 0xFFF) < HEADER_SIZE → raw might be on previous (unmapped) page
        uintptr_t offset_in_page = (uintptr_t)ptr & 0xFFF;
        if (__builtin_expect(offset_in_page < HEADER_SIZE, 0)) {
            // Potential page boundary crossing - do safety check
            if (!hak_is_memory_readable(raw)) {
                goto slow_path_after_step2;
            }
        }
        // Normal case (99.6%): raw is on same page as ptr (no mincore call!)

        // Safe to dereference now
        AllocHeader* hdr = (AllocHeader*)raw;

        if (hdr->magic == HAKMEM_MAGIC) {
            // Valid 16-byte header found (malloc/mmap allocation)
            hak_free_route_log("header_16byte", ptr);

            if (hdr->method == ALLOC_METHOD_MALLOC) {
                // CRITICAL: raw was allocated with __libc_malloc, so free with __libc_free
                extern void __libc_free(void*);
                __libc_free(raw);
                goto done;
            }

            // Handle other methods (mmap, etc) - continue to slow path below
        }
slow_path_after_step2:;
    }

    // Fallback: Invalid header (non-tiny) or TLS cache full
#if !HAKMEM_BUILD_RELEASE
    hak_free_v2_track_slow();
#endif
#endif

    // SS-first free（既定ON）
#if !HAKMEM_TINY_HEADER_CLASSIDX
    // Only run SS-first if Phase 7 header-based free is not enabled
    // (Phase 7 already does the SS lookup and handles SS allocations)
    {
        static int s_free_to_ss = -2;
        if (s_free_to_ss == -2) {
            const char* e = getenv("HAKMEM_TINY_FREE_TO_SS");
            s_free_to_ss = (e && *e) ? ((*e!='0')?1:0) : 1;
        }
        if (s_free_to_ss) {
            extern int g_use_superslab;
            if (__builtin_expect(g_use_superslab != 0, 1)) {
                SuperSlab* ss = hak_super_lookup(ptr);
                if (ss && ss->magic == SUPERSLAB_MAGIC) {
                    int sidx = slab_index_for(ss, ptr);
                    int cap = ss_slabs_capacity(ss);
                    if (__builtin_expect(sidx >= 0 && sidx < cap, 1)) { hak_free_route_log("ss_hit", ptr); hak_tiny_free(ptr); goto done; }
                }
                // FIX: Removed dangerous "guess loop" (lines 92-95)
                // The loop dereferenced unmapped memory causing SEGV
                // If registry lookup fails, allocation is not from SuperSlab
            }
        }
    }
#endif

    // Mid/L25 headerless経路
    {
        extern int hak_pool_mid_lookup(void* ptr, size_t* out_size);
        extern void hak_pool_free_fast(void* ptr, uintptr_t site_id);
        size_t mid_sz = 0; if (hak_pool_mid_lookup(ptr, &mid_sz)) { hak_free_route_log("mid_hit", ptr); hak_pool_free_fast(ptr, (uintptr_t)site); goto done; }
    }
    {
        extern int hak_l25_lookup(void* ptr, size_t* out_size);
        extern void hak_l25_pool_free_fast(void* ptr, uintptr_t site_id);
        size_t l25_sz = 0; if (hak_l25_lookup(ptr, &l25_sz)) { hak_free_route_log("l25_hit", ptr); hkm_ace_stat_large_free(); hak_l25_pool_free_fast(ptr, (uintptr_t)site); goto done; }
    }

    // Raw header dispatch（mmap/malloc/BigCacheなど）
    {
        void* raw = (char*)ptr - HEADER_SIZE;

        // CRITICAL FIX (2025-11-07): Check if memory is accessible before dereferencing
        // This prevents SEGV when ptr has no header (Tiny alloc where SS lookup failed, or libc alloc)
        if (!hak_is_memory_readable(raw)) {
            // Memory not accessible, ptr likely has no header
            hak_free_route_log("unmapped_header_fallback", ptr);

            // In direct-link mode, try tiny_free (handles headerless Tiny allocs)
            if (!g_ldpreload_mode && g_invalid_free_mode) {
                hak_tiny_free(ptr);
                goto done;
            }

            // LD_PRELOAD mode: route to libc (might be libc allocation)
            extern void __libc_free(void*);
            __libc_free(ptr);
            goto done;
        }

        // Safe to dereference header now
        AllocHeader* hdr = (AllocHeader*)raw;
        if (hdr->magic != HAKMEM_MAGIC) {
            // CRITICAL FIX (2025-11-07): Invalid magic could mean:
            // 1. Tiny allocation where SuperSlab lookup failed (NO header exists)
            // 2. Libc allocation from mixed environment
            // 3. Double-free or corrupted pointer

            if (g_invalid_free_log) fprintf(stderr, "[hakmem] ERROR: Invalid magic 0x%X (expected 0x%X)\n", hdr->magic, HAKMEM_MAGIC);

            // One-shot request-trace to help diagnose SS registry lookups
            hak_super_reg_reqtrace_dump(ptr);

            // In direct-link mode, try routing to tiny free as best-effort recovery
            // This handles case #1 where SuperSlab lookup failed but allocation is valid
            if (!g_ldpreload_mode && g_invalid_free_mode) {
                // Attempt tiny free (will validate internally and handle gracefully if invalid)
                hak_free_route_log("invalid_magic_tiny_recovery", ptr);
                hak_tiny_free(ptr);
                goto done;
            }

            // LD_PRELOAD mode or fallback mode: route to libc
            // IMPORTANT: Use ptr (not raw), as NO header exists
            if (g_invalid_free_mode) {
                // Skip mode: leak memory (original behavior, but logged)
                static int leak_warn = 0;
                if (!leak_warn) {
                    fprintf(stderr, "[hakmem] WARNING: Skipping free of invalid pointer %p (may leak memory)\n", ptr);
                    leak_warn = 1;
                }
                goto done;
            } else {
                // Fallback mode: route to libc
                extern void __libc_free(void*);
                __libc_free(ptr);  // Use ptr, not raw!
                goto done;
            }
        }
        if (HAK_ENABLED_CACHE(HAKMEM_FEATURE_BIGCACHE) && hdr->class_bytes >= 2097152) {
            if (hak_bigcache_put(ptr, hdr->size, hdr->alloc_site)) goto done;
        }
        {
            static int g_bc_l25_en_free = -1; if (g_bc_l25_en_free == -1) { const char* e = getenv("HAKMEM_BIGCACHE_L25"); g_bc_l25_en_free = (e && atoi(e) != 0) ? 1 : 0; }
            if (g_bc_l25_en_free && HAK_ENABLED_CACHE(HAKMEM_FEATURE_BIGCACHE) && hdr->size >= 524288 && hdr->size < 2097152) {
                if (hak_bigcache_put(ptr, hdr->size, hdr->alloc_site)) goto done;
            }
        }
        switch (hdr->method) {
            case ALLOC_METHOD_POOL: if (HAK_ENABLED_ALLOC(HAKMEM_FEATURE_POOL)) { hkm_ace_stat_mid_free(); hak_pool_free(ptr, hdr->size, hdr->alloc_site); goto done; } break;
            case ALLOC_METHOD_L25_POOL: hkm_ace_stat_large_free(); hak_l25_pool_free(ptr, hdr->size, hdr->alloc_site); goto done;
            case ALLOC_METHOD_MALLOC:
                // CRITICAL FIX: raw was allocated with __libc_malloc, so free with __libc_free
                // Using free(raw) would go through wrapper → infinite recursion
                hak_free_route_log("malloc_hdr", ptr);
                extern void __libc_free(void*);
                __libc_free(raw);
                break;
            case ALLOC_METHOD_MMAP:
#ifdef __linux__
                if (HAK_ENABLED_MEMORY(HAKMEM_FEATURE_BATCH_MADVISE) && hdr->size >= BATCH_MIN_SIZE) { hak_batch_add(raw, hdr->size); goto done; }
                if (hkm_whale_put(raw, hdr->size) != 0) { hkm_sys_munmap(raw, hdr->size); }
#else
                // CRITICAL FIX: Same as ALLOC_METHOD_MALLOC
                extern void __libc_free(void*);
                __libc_free(raw);
#endif
                break;
            default: fprintf(stderr, "[hakmem] ERROR: Unknown allocation method: %d\n", hdr->method); break;
        }
    }

done:
#if HAKMEM_DEBUG_TIMING
    HKM_TIME_END(HKM_CAT_HAK_FREE, t0);
#endif
    return;
}

#endif // HAK_FREE_API_INC_H