diff options
Diffstat (limited to 'gc.c')
| -rw-r--r-- | gc.c | 10384 |
1 files changed, 4333 insertions, 6051 deletions
@@ -11,41 +11,35 @@ **********************************************************************/ -#include "ruby/ruby.h" -#include "ruby/st.h" -#include "ruby/re.h" -#include "ruby/io.h" -#include "ruby/thread.h" -#include "ruby/util.h" -#include "ruby/debug.h" -#include "eval_intern.h" -#include "vm_core.h" -#include "internal.h" -#include "gc.h" -#include "constant.h" -#include "ruby_atomic.h" -#include "probes.h" -#include <stdio.h> -#include <stdarg.h> -#include <setjmp.h> -#include <sys/types.h> -#include <assert.h> +#include "ruby/internal/config.h" +#ifdef _WIN32 +# include "ruby/ruby.h" +#endif -#ifndef __has_feature -# define __has_feature(x) 0 +#if defined(__wasm__) && !defined(__EMSCRIPTEN__) +# include "wasm/setjmp.h" +# include "wasm/machine.h" +#else +# include <setjmp.h> #endif +#include <stdarg.h> +#include <stdio.h> +/* MALLOC_HEADERS_BEGIN */ #ifndef HAVE_MALLOC_USABLE_SIZE # ifdef _WIN32 -# define HAVE_MALLOC_USABLE_SIZE -# define malloc_usable_size(a) _msize(a) +# define HAVE_MALLOC_USABLE_SIZE +# define malloc_usable_size(a) _msize(a) # elif defined HAVE_MALLOC_SIZE -# define HAVE_MALLOC_USABLE_SIZE -# define malloc_usable_size(a) malloc_size(a) +# define HAVE_MALLOC_USABLE_SIZE +# define malloc_usable_size(a) malloc_size(a) # endif #endif + #ifdef HAVE_MALLOC_USABLE_SIZE -# ifdef HAVE_MALLOC_H +# ifdef RUBY_ALTERNATIVE_MALLOC_HEADER +/* Alternative malloc header is included in ruby/missing.h */ +# elif defined(HAVE_MALLOC_H) # include <malloc.h> # elif defined(HAVE_MALLOC_NP_H) # include <malloc_np.h> @@ -54,1754 +48,1716 @@ # endif #endif -#if /* is ASAN enabled? */ \ - __has_feature(address_sanitizer) /* Clang */ || \ - defined(__SANITIZE_ADDRESS__) /* GCC 4.8.x */ - #define ATTRIBUTE_NO_ADDRESS_SAFETY_ANALYSIS \ - __attribute__((no_address_safety_analysis)) \ - __attribute__((noinline)) -#else - #define ATTRIBUTE_NO_ADDRESS_SAFETY_ANALYSIS -#endif +/* MALLOC_HEADERS_END */ #ifdef HAVE_SYS_TIME_H -#include <sys/time.h> +# include <sys/time.h> #endif #ifdef HAVE_SYS_RESOURCE_H -#include <sys/resource.h> -#endif -#if defined(__native_client__) && defined(NACL_NEWLIB) -# include "nacl/resource.h" -# undef HAVE_POSIX_MEMALIGN -# undef HAVE_MEMALIGN - +# include <sys/resource.h> #endif #if defined _WIN32 || defined __CYGWIN__ -#include <windows.h> +# include <windows.h> #elif defined(HAVE_POSIX_MEMALIGN) #elif defined(HAVE_MEMALIGN) -#include <malloc.h> +# include <malloc.h> #endif -#define rb_setjmp(env) RUBY_SETJMP(env) -#define rb_jmp_buf rb_jmpbuf_t +#include <sys/types.h> -#if defined(HAVE_RB_GC_GUARDED_PTR) && HAVE_RB_GC_GUARDED_PTR -volatile VALUE * -rb_gc_guarded_ptr(volatile VALUE *ptr) -{ - return ptr; -} +#ifdef __EMSCRIPTEN__ +#include <emscripten.h> #endif -#ifndef GC_HEAP_FREE_SLOTS -#define GC_HEAP_FREE_SLOTS 4096 -#endif -#ifndef GC_HEAP_INIT_SLOTS -#define GC_HEAP_INIT_SLOTS 10000 -#endif -#ifndef GC_HEAP_GROWTH_FACTOR -#define GC_HEAP_GROWTH_FACTOR 1.8 -#endif -#ifndef GC_HEAP_GROWTH_MAX_SLOTS -#define GC_HEAP_GROWTH_MAX_SLOTS 0 /* 0 is disable */ -#endif -#ifndef GC_HEAP_OLDOBJECT_LIMIT_FACTOR -#define GC_HEAP_OLDOBJECT_LIMIT_FACTOR 2.0 +/* For ruby_annotate_mmap */ +#ifdef HAVE_SYS_PRCTL_H +#include <sys/prctl.h> #endif -#ifndef GC_MALLOC_LIMIT_MIN -#define GC_MALLOC_LIMIT_MIN (16 * 1024 * 1024 /* 16MB */) -#endif -#ifndef GC_MALLOC_LIMIT_MAX -#define GC_MALLOC_LIMIT_MAX (32 * 1024 * 1024 /* 32MB */) -#endif -#ifndef GC_MALLOC_LIMIT_GROWTH_FACTOR -#define GC_MALLOC_LIMIT_GROWTH_FACTOR 1.4 -#endif +#undef LIST_HEAD /* ccan/list conflicts with BSD-origin sys/queue.h. */ -#ifndef GC_OLDMALLOC_LIMIT_MIN -#define GC_OLDMALLOC_LIMIT_MIN (16 * 1024 * 1024 /* 16MB */) -#endif -#ifndef GC_OLDMALLOC_LIMIT_GROWTH_FACTOR -#define GC_OLDMALLOC_LIMIT_GROWTH_FACTOR 1.2 -#endif -#ifndef GC_OLDMALLOC_LIMIT_MAX -#define GC_OLDMALLOC_LIMIT_MAX (128 * 1024 * 1024 /* 128MB */) -#endif +#include "constant.h" +#include "debug_counter.h" +#include "eval_intern.h" +#include "gc/gc.h" +#include "id_table.h" +#include "internal.h" +#include "internal/class.h" +#include "internal/compile.h" +#include "internal/complex.h" +#include "internal/concurrent_set.h" +#include "internal/cont.h" +#include "internal/error.h" +#include "internal/eval.h" +#include "internal/gc.h" +#include "internal/hash.h" +#include "internal/imemo.h" +#include "internal/io.h" +#include "internal/numeric.h" +#include "internal/object.h" +#include "internal/proc.h" +#include "internal/rational.h" +#include "internal/re.h" +#include "internal/sanitizers.h" +#include "internal/struct.h" +#include "internal/symbol.h" +#include "internal/thread.h" +#include "internal/variable.h" +#include "internal/warnings.h" +#include "probes.h" +#include "regint.h" +#include "ruby/debug.h" +#include "ruby/io.h" +#include "ruby/re.h" +#include "ruby/st.h" +#include "ruby/thread.h" +#include "ruby/util.h" +#include "ruby/vm.h" +#include "ruby_assert.h" +#include "ruby_atomic.h" +#include "symbol.h" +#include "variable.h" +#include "vm_core.h" +#include "vm_sync.h" +#include "vm_callinfo.h" +#include "ractor_core.h" +#include "yjit.h" +#include "zjit.h" -typedef struct { - unsigned int heap_init_slots; - unsigned int heap_free_slots; - double growth_factor; - unsigned int growth_max_slots; - double oldobject_limit_factor; - unsigned int malloc_limit_min; - unsigned int malloc_limit_max; - double malloc_limit_growth_factor; - unsigned int oldmalloc_limit_min; - unsigned int oldmalloc_limit_max; - double oldmalloc_limit_growth_factor; -#if defined(ENABLE_VM_OBJSPACE) && ENABLE_VM_OBJSPACE - VALUE gc_stress; -#endif -} ruby_gc_params_t; - -static ruby_gc_params_t gc_params = { - GC_HEAP_INIT_SLOTS, - GC_HEAP_FREE_SLOTS, - GC_HEAP_GROWTH_FACTOR, - GC_HEAP_GROWTH_MAX_SLOTS, - GC_HEAP_OLDOBJECT_LIMIT_FACTOR, - GC_MALLOC_LIMIT_MIN, - GC_MALLOC_LIMIT_MAX, - GC_MALLOC_LIMIT_GROWTH_FACTOR, - GC_OLDMALLOC_LIMIT_MIN, - GC_OLDMALLOC_LIMIT_MAX, - GC_OLDMALLOC_LIMIT_GROWTH_FACTOR, -#if defined(ENABLE_VM_OBJSPACE) && ENABLE_VM_OBJSPACE - FALSE, -#endif -}; +#include "builtin.h" +#include "shape.h" -/* GC_DEBUG: - * enable to embed GC debugging information. - */ -#ifndef GC_DEBUG -#define GC_DEBUG 0 -#endif +// TODO: Don't export this function in modular GC, instead MMTk should figure out +// how to combine GC thread backtrace with mutator thread backtrace. +void +rb_gc_print_backtrace(void) +{ + rb_print_backtrace(stderr); +} -#if USE_RGENGC -/* RGENGC_DEBUG: - * 1: basic information - * 2: remember set operation - * 3: mark - * 4: - * 5: sweep - */ -#ifndef RGENGC_DEBUG -#define RGENGC_DEBUG 0 -#endif +unsigned int +rb_gc_vm_lock(const char *file, int line) +{ + unsigned int lev = 0; + rb_vm_lock_enter(&lev, file, line); + return lev; +} -/* RGENGC_CHECK_MODE - * 0: disable all assertions - * 1: enable assertions (to debug RGenGC) - * 2: enable generational bits check (for debugging) - * 3: enable livness check - * 4: show all references - */ -#ifndef RGENGC_CHECK_MODE -#define RGENGC_CHECK_MODE 0 -#endif +void +rb_gc_vm_unlock(unsigned int lev, const char *file, int line) +{ + rb_vm_lock_leave(&lev, file, line); +} -/* RGENGC_PROFILE - * 0: disable RGenGC profiling - * 1: enable profiling for basic information - * 2: enable profiling for each types - */ -#ifndef RGENGC_PROFILE -#define RGENGC_PROFILE 0 -#endif +unsigned int +rb_gc_cr_lock(const char *file, int line) +{ + unsigned int lev; + rb_vm_lock_enter_cr(GET_RACTOR(), &lev, file, line); + return lev; +} -/* RGENGC_THREEGEN - * Enable/disable three gen GC. - * 0: Infant gen -> Old gen - * 1: Infant gen -> Young -> Old gen - */ -#ifndef RGENGC_THREEGEN -#define RGENGC_THREEGEN 0 -#endif +void +rb_gc_cr_unlock(unsigned int lev, const char *file, int line) +{ + rb_vm_lock_leave_cr(GET_RACTOR(), &lev, file, line); +} -/* RGENGC_ESTIMATE_OLDMALLOC - * Enable/disable to estimate increase size of malloc'ed size by old objects. - * If estimation exceeds threashold, then will invoke full GC. - * 0: disable estimation. - * 1: enable estimation. - */ -#ifndef RGENGC_ESTIMATE_OLDMALLOC -#define RGENGC_ESTIMATE_OLDMALLOC 1 -#endif +unsigned int +rb_gc_vm_lock_no_barrier(const char *file, int line) +{ + unsigned int lev = 0; + rb_vm_lock_enter_nb(&lev, file, line); + return lev; +} -#else /* USE_RGENGC */ +void +rb_gc_vm_unlock_no_barrier(unsigned int lev, const char *file, int line) +{ + rb_vm_lock_leave_nb(&lev, file, line); +} -#define RGENGC_DEBUG 0 -#define RGENGC_CHECK_MODE 0 -#define RGENGC_PROFILE 0 -#define RGENGC_THREEGEN 0 -#define RGENGC_ESTIMATE_OLDMALLOC 0 +void +rb_gc_vm_barrier(void) +{ + rb_vm_barrier(); +} -#endif /* USE_RGENGC */ +void * +rb_gc_get_ractor_newobj_cache(void) +{ + return GET_RACTOR()->newobj_cache; +} -#ifndef GC_PROFILE_MORE_DETAIL -#define GC_PROFILE_MORE_DETAIL 0 -#endif -#ifndef GC_PROFILE_DETAIL_MEMORY -#define GC_PROFILE_DETAIL_MEMORY 0 -#endif -#ifndef GC_ENABLE_LAZY_SWEEP -#define GC_ENABLE_LAZY_SWEEP 1 -#endif -#ifndef CALC_EXACT_MALLOC_SIZE -#define CALC_EXACT_MALLOC_SIZE 0 -#endif -#if defined(HAVE_MALLOC_USABLE_SIZE) || CALC_EXACT_MALLOC_SIZE > 0 -#ifndef MALLOC_ALLOCATED_SIZE -#define MALLOC_ALLOCATED_SIZE 0 -#endif -#else -#define MALLOC_ALLOCATED_SIZE 0 -#endif -#ifndef MALLOC_ALLOCATED_SIZE_CHECK -#define MALLOC_ALLOCATED_SIZE_CHECK 0 -#endif +void +rb_gc_initialize_vm_context(struct rb_gc_vm_context *context) +{ + rb_native_mutex_initialize(&context->lock); + context->ec = GET_EC(); +} -typedef enum { - GPR_FLAG_NONE = 0x000, - /* major reason */ - GPR_FLAG_MAJOR_BY_NOFREE = 0x001, - GPR_FLAG_MAJOR_BY_OLDGEN = 0x002, - GPR_FLAG_MAJOR_BY_SHADY = 0x004, - GPR_FLAG_MAJOR_BY_RESCAN = 0x008, - GPR_FLAG_MAJOR_BY_STRESS = 0x010, -#if RGENGC_ESTIMATE_OLDMALLOC - GPR_FLAG_MAJOR_BY_OLDMALLOC = 0x020, -#endif - GPR_FLAG_MAJOR_MASK = 0x0ff, - - /* gc reason */ - GPR_FLAG_NEWOBJ = 0x100, - GPR_FLAG_MALLOC = 0x200, - GPR_FLAG_METHOD = 0x400, - GPR_FLAG_CAPI = 0x800, - GPR_FLAG_STRESS = 0x1000, - - /* others */ - GPR_FLAG_IMMEDIATE_SWEEP = 0x2000, - GPR_FLAG_HAVE_FINALIZE = 0x4000 -} gc_profile_record_flag; - -typedef struct gc_profile_record { - int flags; - - double gc_time; - double gc_invoke_time; - - size_t heap_total_objects; - size_t heap_use_size; - size_t heap_total_size; - -#if GC_PROFILE_MORE_DETAIL - double gc_mark_time; - double gc_sweep_time; - - size_t heap_use_pages; - size_t heap_live_objects; - size_t heap_free_objects; - - size_t allocate_increase; - size_t allocate_limit; - - double prepare_time; - size_t removing_objects; - size_t empty_objects; -#if GC_PROFILE_DETAIL_MEMORY - long maxrss; - long minflt; - long majflt; -#endif -#endif -#if MALLOC_ALLOCATED_SIZE - size_t allocated_size; -#endif +bool +rb_gc_event_hook_required_p(rb_event_flag_t event) +{ + return ruby_vm_event_flags & event; +} -#if RGENGC_PROFILE > 0 - size_t old_objects; - size_t remembered_normal_objects; - size_t remembered_shady_objects; -#endif -} gc_profile_record; +void +rb_gc_event_hook(VALUE obj, rb_event_flag_t event) +{ + if (LIKELY(!rb_gc_event_hook_required_p(event))) return; -#if defined(_MSC_VER) || defined(__BORLANDC__) || defined(__CYGWIN__) -#pragma pack(push, 1) /* magic for reducing sizeof(RVALUE): 24 -> 20 */ -#endif + rb_execution_context_t *ec = rb_gc_get_ec(); + if (!ec->cfp) return; + +#if USE_MODULAR_GC + bool gc_thread_p = false; + if (!GET_EC()) { + gc_thread_p = true; -typedef struct RVALUE { - union { - struct { - VALUE flags; /* always 0 for freed obj */ - struct RVALUE *next; - } free; - struct RBasic basic; - struct RObject object; - struct RClass klass; - struct RFloat flonum; - struct RString string; - struct RArray array; - struct RRegexp regexp; - struct RHash hash; - struct RData data; - struct RTypedData typeddata; - struct RStruct rstruct; - struct RBignum bignum; - struct RFile file; - struct RNode node; - struct RMatch match; - struct RRational rational; - struct RComplex complex; - struct { - struct RBasic basic; - VALUE v1; - VALUE v2; - VALUE v3; - } values; - } as; -#if GC_DEBUG - const char *file; - VALUE line; +# ifdef RB_THREAD_LOCAL_SPECIFIER + rb_current_ec_set(ec); +# else + native_tls_set(ruby_current_ec_key, ec); +# endif + } #endif -} RVALUE; -#if defined(_MSC_VER) || defined(__BORLANDC__) || defined(__CYGWIN__) -#pragma pack(pop) + EXEC_EVENT_HOOK(ec, event, ec->cfp->self, 0, 0, 0, obj); + +#if USE_MODULAR_GC + if (gc_thread_p) { +# ifdef RB_THREAD_LOCAL_SPECIFIER + rb_current_ec_set(NULL); +# else + native_tls_set(ruby_current_ec_key, NULL); +# endif + } #endif +} -typedef uintptr_t bits_t; -enum { - BITS_SIZE = sizeof(bits_t), - BITS_BITLENGTH = ( BITS_SIZE * CHAR_BIT ) -}; +void * +rb_gc_get_objspace(void) +{ + return GET_VM()->gc.objspace; +} -struct heap_page_header { - struct heap_page *page; -}; +void +rb_gc_ractor_newobj_cache_foreach(void (*func)(void *cache, void *data), void *data) +{ + rb_ractor_t *r = NULL; + if (RB_LIKELY(ruby_single_main_ractor)) { + GC_ASSERT( + ccan_list_empty(&GET_VM()->ractor.set) || + (ccan_list_top(&GET_VM()->ractor.set, rb_ractor_t, vmlr_node) == ruby_single_main_ractor && + ccan_list_tail(&GET_VM()->ractor.set, rb_ractor_t, vmlr_node) == ruby_single_main_ractor) + ); -struct heap_page_body { - struct heap_page_header header; - /* char gap[]; */ - /* RVALUE values[]; */ -}; + func(ruby_single_main_ractor->newobj_cache, data); + } + else { + ccan_list_for_each(&GET_VM()->ractor.set, r, vmlr_node) { + func(r->newobj_cache, data); + } + } +} -struct gc_list { - VALUE *varptr; - struct gc_list *next; -}; +void +rb_gc_run_obj_finalizer(VALUE objid, long count, VALUE (*callback)(long i, void *data), void *data) +{ + volatile struct { + VALUE errinfo; + VALUE final; + rb_control_frame_t *cfp; + VALUE *sp; + long finished; + } saved; + + rb_execution_context_t * volatile ec = GET_EC(); +#define RESTORE_FINALIZER() (\ + ec->cfp = saved.cfp, \ + ec->cfp->sp = saved.sp, \ + ec->errinfo = saved.errinfo) + + saved.errinfo = ec->errinfo; + saved.cfp = ec->cfp; + saved.sp = ec->cfp->sp; + saved.finished = 0; + saved.final = Qundef; + + ASSERT_vm_unlocking(); + rb_ractor_ignore_belonging(true); + EC_PUSH_TAG(ec); + enum ruby_tag_type state = EC_EXEC_TAG(); + if (state != TAG_NONE) { + ++saved.finished; /* skip failed finalizer */ + + VALUE failed_final = saved.final; + saved.final = Qundef; + if (!UNDEF_P(failed_final) && !NIL_P(ruby_verbose)) { + rb_warn("Exception in finalizer %+"PRIsVALUE, failed_final); + rb_ec_error_print(ec, ec->errinfo); + } + } -#define STACK_CHUNK_SIZE 500 - -typedef struct stack_chunk { - VALUE data[STACK_CHUNK_SIZE]; - struct stack_chunk *next; -} stack_chunk_t; - -typedef struct mark_stack { - stack_chunk_t *chunk; - stack_chunk_t *cache; - size_t index; - size_t limit; - size_t cache_size; - size_t unused_cache_size; -} mark_stack_t; - -typedef struct rb_heap_struct { - struct heap_page *pages; - struct heap_page *free_pages; - struct heap_page *using_page; - struct heap_page *sweep_pages; - RVALUE *freelist; - size_t page_length; /* total page count in a heap */ - size_t total_slots; /* total slot count (page_length * HEAP_OBJ_LIMIT) */ -} rb_heap_t; - -typedef struct rb_objspace { - struct { - size_t limit; - size_t increase; -#if MALLOC_ALLOCATED_SIZE - size_t allocated_size; - size_t allocations; -#endif - } malloc_params; - - rb_heap_t eden_heap; - rb_heap_t tomb_heap; /* heap for zombies and ghosts */ - - struct { - struct heap_page **sorted; - size_t used; - size_t length; - RVALUE *range[2]; - - size_t limit; - size_t increment; - - size_t swept_slots; - size_t min_free_slots; - size_t max_free_slots; - - /* final */ - size_t final_slots; - RVALUE *deferred_final; - } heap_pages; - - struct { - int dont_gc; - int dont_lazy_sweep; - int during_gc; - rb_atomic_t finalizing; - } flags; - st_table *finalizer_table; - mark_stack_t mark_stack; - struct { - int run; - gc_profile_record *records; - gc_profile_record *current_record; - size_t next_index; - size_t size; - -#if GC_PROFILE_MORE_DETAIL - double prepare_time; -#endif - double invoke_time; - -#if USE_RGENGC - size_t minor_gc_count; - size_t major_gc_count; -#if RGENGC_PROFILE > 0 - size_t generated_normal_object_count; - size_t generated_shady_object_count; - size_t shade_operation_count; - size_t promote_infant_count; -#if RGENGC_THREEGEN - size_t promote_young_count; -#endif - size_t remembered_normal_object_count; - size_t remembered_shady_object_count; - -#if RGENGC_PROFILE >= 2 - size_t generated_normal_object_count_types[RUBY_T_MASK]; - size_t generated_shady_object_count_types[RUBY_T_MASK]; - size_t shade_operation_count_types[RUBY_T_MASK]; - size_t promote_infant_types[RUBY_T_MASK]; -#if RGENGC_THREEGEN - size_t promote_young_types[RUBY_T_MASK]; -#endif - size_t remembered_normal_object_count_types[RUBY_T_MASK]; - size_t remembered_shady_object_count_types[RUBY_T_MASK]; -#endif -#endif /* RGENGC_PROFILE */ -#endif /* USE_RGENGC */ - - /* temporary profiling space */ - double gc_sweep_start_time; - size_t total_allocated_object_num_at_gc_start; - size_t heap_used_at_gc_start; - - /* basic statistics */ - size_t count; - size_t total_allocated_object_num; - size_t total_freed_object_num; - int latest_gc_info; - } profile; - struct gc_list *global_list; - rb_event_flag_t hook_events; /* this place may be affinity with memory cache */ - VALUE gc_stress; - - struct mark_func_data_struct { - void *data; - void (*mark_func)(VALUE v, void *data); - } *mark_func_data; - -#if USE_RGENGC - struct { - int during_minor_gc; - int parent_object_is_old; - - int need_major_gc; - - size_t last_major_gc; - - size_t remembered_shady_object_count; - size_t remembered_shady_object_limit; - size_t old_object_count; - size_t old_object_limit; -#if RGENGC_THREEGEN - size_t young_object_count; -#endif + for (long i = saved.finished; RESTORE_FINALIZER(), i < count; saved.finished = ++i) { + saved.final = callback(i, data); + rb_check_funcall(saved.final, idCall, 1, &objid); + } + EC_POP_TAG(); + rb_ractor_ignore_belonging(false); +#undef RESTORE_FINALIZER +} -#if RGENGC_ESTIMATE_OLDMALLOC - size_t oldmalloc_increase; - size_t oldmalloc_increase_limit; -#endif +void +rb_gc_set_pending_interrupt(void) +{ + rb_execution_context_t *ec = GET_EC(); + ec->interrupt_mask |= PENDING_INTERRUPT_MASK; +} -#if RGENGC_CHECK_MODE >= 2 - struct st_table *allrefs_table; - size_t error_count; -#endif - } rgengc; -#endif /* USE_RGENGC */ -} rb_objspace_t; +void +rb_gc_unset_pending_interrupt(void) +{ + rb_execution_context_t *ec = GET_EC(); + ec->interrupt_mask &= ~PENDING_INTERRUPT_MASK; +} +bool +rb_gc_multi_ractor_p(void) +{ + return rb_multi_ractor_p(); +} -#ifndef HEAP_ALIGN_LOG -/* default tiny heap size: 16KB */ -#define HEAP_ALIGN_LOG 14 -#endif -#define CEILDIV(i, mod) (((i) + (mod) - 1)/(mod)) -enum { - HEAP_ALIGN = (1UL << HEAP_ALIGN_LOG), - HEAP_ALIGN_MASK = (~(~0UL << HEAP_ALIGN_LOG)), - REQUIRED_SIZE_BY_MALLOC = (sizeof(size_t) * 5), - HEAP_SIZE = (HEAP_ALIGN - REQUIRED_SIZE_BY_MALLOC), - HEAP_OBJ_LIMIT = (unsigned int)((HEAP_SIZE - sizeof(struct heap_page_header))/sizeof(struct RVALUE)), - HEAP_BITMAP_LIMIT = CEILDIV(CEILDIV(HEAP_SIZE, sizeof(struct RVALUE)), BITS_BITLENGTH), - HEAP_BITMAP_SIZE = ( BITS_SIZE * HEAP_BITMAP_LIMIT), - HEAP_BITMAP_PLANES = USE_RGENGC ? 3 : 1 /* RGENGC: mark bits, rememberset bits and oldgen bits */ -}; +bool +rb_gc_shutdown_call_finalizer_p(VALUE obj) +{ + switch (BUILTIN_TYPE(obj)) { + case T_DATA: + if (!ruby_free_at_exit_p()) { + if (!RDATA(obj)->type) return false; + if (!rbimpl_typeddata_embedded_p(obj) && !RTYPEDDATA(obj)->data) return false; + } + if (rb_obj_is_thread(obj)) return false; + if (rb_obj_is_mutex(obj)) return false; + if (rb_obj_is_fiber(obj)) return false; + if (rb_ractor_p(obj)) return false; + if (rb_obj_is_fstring_table(obj)) return false; + if (rb_obj_is_symbol_table(obj)) return false; -struct heap_page { - struct heap_page_body *body; - RVALUE *freelist; - RVALUE *start; - size_t final_slots; - size_t limit; - struct heap_page *next; - struct heap_page *prev; - struct heap_page *free_next; - rb_heap_t *heap; - int before_sweep; - - bits_t mark_bits[HEAP_BITMAP_LIMIT]; -#if USE_RGENGC - bits_t rememberset_bits[HEAP_BITMAP_LIMIT]; - bits_t oldgen_bits[HEAP_BITMAP_LIMIT]; -#endif -}; + return true; -#define GET_PAGE_BODY(x) ((struct heap_page_body *)((bits_t)(x) & ~(HEAP_ALIGN_MASK))) -#define GET_PAGE_HEADER(x) (&GET_PAGE_BODY(x)->header) -#define GET_HEAP_PAGE(x) (GET_PAGE_HEADER(x)->page) -#define GET_HEAP_MARK_BITS(x) (&GET_HEAP_PAGE(x)->mark_bits[0]) -#define GET_HEAP_REMEMBERSET_BITS(x) (&GET_HEAP_PAGE(x)->rememberset_bits[0]) -#define GET_HEAP_OLDGEN_BITS(x) (&GET_HEAP_PAGE(x)->oldgen_bits[0]) -#define NUM_IN_PAGE(p) (((bits_t)(p) & HEAP_ALIGN_MASK)/sizeof(RVALUE)) -#define BITMAP_INDEX(p) (NUM_IN_PAGE(p) / BITS_BITLENGTH ) -#define BITMAP_OFFSET(p) (NUM_IN_PAGE(p) & (BITS_BITLENGTH-1)) -#define BITMAP_BIT(p) ((bits_t)1 << BITMAP_OFFSET(p)) -/* Bitmap Operations */ -#define MARKED_IN_BITMAP(bits, p) ((bits)[BITMAP_INDEX(p)] & BITMAP_BIT(p)) -#define MARK_IN_BITMAP(bits, p) ((bits)[BITMAP_INDEX(p)] = (bits)[BITMAP_INDEX(p)] | BITMAP_BIT(p)) -#define CLEAR_IN_BITMAP(bits, p) ((bits)[BITMAP_INDEX(p)] = (bits)[BITMAP_INDEX(p)] & ~BITMAP_BIT(p)) - -/* Aliases */ -#if defined(ENABLE_VM_OBJSPACE) && ENABLE_VM_OBJSPACE -#define rb_objspace (*GET_VM()->objspace) -#define ruby_initial_gc_stress gc_params.gc_stress -VALUE *ruby_initial_gc_stress_ptr = &ruby_initial_gc_stress; -#else -static rb_objspace_t rb_objspace = {{GC_MALLOC_LIMIT_MIN}}; -VALUE *ruby_initial_gc_stress_ptr = &rb_objspace.gc_stress; -#endif + case T_FILE: + return true; -#define malloc_limit objspace->malloc_params.limit -#define malloc_increase objspace->malloc_params.increase -#define malloc_allocated_size objspace->malloc_params.allocated_size -#define heap_pages_sorted objspace->heap_pages.sorted -#define heap_pages_used objspace->heap_pages.used -#define heap_pages_length objspace->heap_pages.length -#define heap_pages_lomem objspace->heap_pages.range[0] -#define heap_pages_himem objspace->heap_pages.range[1] -#define heap_pages_swept_slots objspace->heap_pages.swept_slots -#define heap_pages_increment objspace->heap_pages.increment -#define heap_pages_min_free_slots objspace->heap_pages.min_free_slots -#define heap_pages_max_free_slots objspace->heap_pages.max_free_slots -#define heap_pages_final_slots objspace->heap_pages.final_slots -#define heap_pages_deferred_final objspace->heap_pages.deferred_final -#define heap_eden (&objspace->eden_heap) -#define heap_tomb (&objspace->tomb_heap) -#define dont_gc objspace->flags.dont_gc -#define during_gc objspace->flags.during_gc -#define finalizing objspace->flags.finalizing -#define finalizer_table objspace->finalizer_table -#define global_List objspace->global_list -#define ruby_gc_stress objspace->gc_stress -#define monitor_level objspace->rgengc.monitor_level -#define monitored_object_table objspace->rgengc.monitored_object_table - -#define is_lazy_sweeping(heap) ((heap)->sweep_pages != 0) -#if SIZEOF_LONG == SIZEOF_VOIDP -# define nonspecial_obj_id(obj) (VALUE)((SIGNED_VALUE)(obj)|FIXNUM_FLAG) -# define obj_id_to_ref(objid) ((objid) ^ FIXNUM_FLAG) /* unset FIXNUM_FLAG */ -#elif SIZEOF_LONG_LONG == SIZEOF_VOIDP -# define nonspecial_obj_id(obj) LL2NUM((SIGNED_VALUE)(obj) / 2) -# define obj_id_to_ref(objid) (FIXNUM_P(objid) ? \ - ((objid) ^ FIXNUM_FLAG) : (NUM2PTR(objid) << 1)) -#else -# error not supported -#endif + case T_SYMBOL: + return true; -#define RANY(o) ((RVALUE*)(o)) + case T_NONE: + return false; -#define nomem_error GET_VM()->special_exceptions[ruby_error_nomemory] + default: + return ruby_free_at_exit_p(); + } +} -int ruby_gc_debug_indent = 0; -VALUE rb_mGC; -int ruby_disable_gc_stress = 0; - -void rb_gcdebug_print_obj_condition(VALUE obj); - -static void rb_objspace_call_finalizer(rb_objspace_t *objspace); -static VALUE define_final0(VALUE obj, VALUE block); - -static void negative_size_allocation_error(const char *); -static void *aligned_malloc(size_t, size_t); -static void aligned_free(void *); - -static void init_mark_stack(mark_stack_t *stack); - -static VALUE lazy_sweep_enable(void); -static int ready_to_gc(rb_objspace_t *objspace); -static int heap_ready_to_gc(rb_objspace_t *objspace, rb_heap_t *heap); -static int garbage_collect(rb_objspace_t *, int full_mark, int immediate_sweep, int reason); -static int garbage_collect_body(rb_objspace_t *, int full_mark, int immediate_sweep, int reason); -static int gc_heap_lazy_sweep(rb_objspace_t *objspace, rb_heap_t *heap); -static void gc_rest_sweep(rb_objspace_t *objspace); -static void gc_heap_rest_sweep(rb_objspace_t *objspace, rb_heap_t *heap); - -static void gc_mark_stacked_objects(rb_objspace_t *); -static void gc_mark(rb_objspace_t *objspace, VALUE ptr); -static void gc_mark_maybe(rb_objspace_t *objspace, VALUE ptr); -static void gc_mark_children(rb_objspace_t *objspace, VALUE ptr); - -static size_t obj_memsize_of(VALUE obj, int use_tdata); - -static double getrusage_time(void); -static inline void gc_prof_setup_new_record(rb_objspace_t *objspace, int reason); -static inline void gc_prof_timer_start(rb_objspace_t *); -static inline void gc_prof_timer_stop(rb_objspace_t *); -static inline void gc_prof_mark_timer_start(rb_objspace_t *); -static inline void gc_prof_mark_timer_stop(rb_objspace_t *); -static inline void gc_prof_sweep_timer_start(rb_objspace_t *); -static inline void gc_prof_sweep_timer_stop(rb_objspace_t *); -static inline void gc_prof_set_malloc_info(rb_objspace_t *); -static inline void gc_prof_set_heap_info(rb_objspace_t *); - -#define gc_prof_record(objspace) (objspace)->profile.current_record -#define gc_prof_enabled(objspace) ((objspace)->profile.run && (objspace)->profile.current_record) - -#define rgengc_report if (RGENGC_DEBUG) rgengc_report_body -static void rgengc_report_body(int level, rb_objspace_t *objspace, const char *fmt, ...); -static const char * type_name(int type, VALUE obj); -static const char *obj_type_name(VALUE obj); +void +rb_gc_obj_changed_pool(VALUE obj, size_t heap_id) +{ + RUBY_ASSERT(RB_TYPE_P(obj, T_OBJECT)); + + RBASIC_SET_SHAPE_ID(obj, rb_obj_shape_transition_heap(obj, heap_id)); +} -#if USE_RGENGC -static int rgengc_remembered(rb_objspace_t *objspace, VALUE obj); -static int rgengc_remember(rb_objspace_t *objspace, VALUE obj); -static void rgengc_mark_and_rememberset_clear(rb_objspace_t *objspace, rb_heap_t *heap); -static void rgengc_rememberset_mark(rb_objspace_t *objspace, rb_heap_t *heap); +void rb_vm_update_references(void *ptr); -#define FL_TEST2(x,f) ((RGENGC_CHECK_MODE && SPECIAL_CONST_P(x)) ? (rb_bug("FL_TEST2: SPECIAL_CONST"), 0) : FL_TEST_RAW((x),(f)) != 0) -#define FL_SET2(x,f) do {if (RGENGC_CHECK_MODE && SPECIAL_CONST_P(x)) rb_bug("FL_SET2: SPECIAL_CONST"); RBASIC(x)->flags |= (f);} while (0) -#define FL_UNSET2(x,f) do {if (RGENGC_CHECK_MODE && SPECIAL_CONST_P(x)) rb_bug("FL_UNSET2: SPECIAL_CONST"); RBASIC(x)->flags &= ~(f);} while (0) +#define rb_setjmp(env) RUBY_SETJMP(env) +#define rb_jmp_buf rb_jmpbuf_t -#define RVALUE_WB_PROTECTED_RAW(obj) FL_TEST2((obj), FL_WB_PROTECTED) -#define RVALUE_WB_PROTECTED(obj) RVALUE_WB_PROTECTED_RAW(check_gen_consistency((VALUE)obj)) +#if !defined(MAP_ANONYMOUS) && defined(MAP_ANON) +#define MAP_ANONYMOUS MAP_ANON +#endif -#define RVALUE_OLDGEN_BITMAP(obj) MARKED_IN_BITMAP(GET_HEAP_OLDGEN_BITS(obj), (obj)) +#define unless_objspace(objspace) \ + void *objspace; \ + rb_vm_t *unless_objspace_vm = GET_VM(); \ + if (unless_objspace_vm) objspace = unless_objspace_vm->gc.objspace; \ + else /* return; or objspace will be warned uninitialized */ -static inline int is_pointer_to_heap(rb_objspace_t *objspace, void *ptr); -static inline int gc_marked(rb_objspace_t *objspace, VALUE ptr); +#define RMOVED(obj) ((struct RMoved *)(obj)) -static inline VALUE -check_gen_consistency(VALUE obj) -{ - if (RGENGC_CHECK_MODE > 0) { - int old_flag = RVALUE_OLDGEN_BITMAP(obj) != 0; - int promoted_flag = FL_TEST2(obj, FL_PROMOTED); - rb_objspace_t *objspace = &rb_objspace; +#define TYPED_UPDATE_IF_MOVED(_objspace, _type, _thing) do { \ + if (gc_object_moved_p_internal((_objspace), (VALUE)(_thing))) { \ + *(_type *)&(_thing) = (_type)gc_location_internal(_objspace, (VALUE)_thing); \ + } \ +} while (0) - obj_memsize_of((VALUE)obj, FALSE); +#define UPDATE_IF_MOVED(_objspace, _thing) TYPED_UPDATE_IF_MOVED(_objspace, VALUE, _thing) - if (!is_pointer_to_heap(objspace, (void *)obj)) { - rb_bug("check_gen_consistency: %p (%s) is not Ruby object.", (void *)obj, obj_type_name(obj)); - } +#if RUBY_MARK_FREE_DEBUG +int ruby_gc_debug_indent = 0; +#endif - if (promoted_flag) { - if (!RVALUE_WB_PROTECTED_RAW(obj)) { - const char *type = old_flag ? "old" : "young"; - rb_bug("check_gen_consistency: %p (%s) is not WB protected, but %s object.", (void *)obj, obj_type_name(obj), type); - } +#ifndef RGENGC_OBJ_INFO +# define RGENGC_OBJ_INFO RGENGC_CHECK_MODE +#endif -#if !RGENGC_THREEGEN - if (!old_flag) { - rb_bug("check_gen_consistency: %p (%s) is not infant, but is not old (on 2gen).", (void *)obj, obj_type_name(obj)); - } +#ifndef CALC_EXACT_MALLOC_SIZE +# define CALC_EXACT_MALLOC_SIZE 0 #endif - if (old_flag && objspace->rgengc.during_minor_gc && !gc_marked(objspace, obj)) { - rb_bug("check_gen_consistency: %p (%s) is old, but is not marked while minor marking.", (void *)obj, obj_type_name(obj)); - } - } - else { - if (old_flag) { - rb_bug("check_gen_consistency: %p (%s) is not infant, but is old.", (void *)obj, obj_type_name(obj)); - } - } +VALUE rb_mGC; + +static size_t malloc_offset = 0; +#if defined(HAVE_MALLOC_USABLE_SIZE) +static size_t +gc_compute_malloc_offset(void) +{ + // Different allocators use different metadata storage strategies which result in different + // ideal sizes. + // For instance malloc(64) will waste 8B with glibc, but waste 0B with jemalloc. + // But malloc(56) will waste 0B with glibc, but waste 8B with jemalloc. + // So we try allocating 64, 56 and 48 bytes and select the first offset that doesn't + // waste memory. + // This was tested on Linux with glibc 2.35 and jemalloc 5, and for both it result in + // no wasted memory. + size_t offset = 0; + for (offset = 0; offset <= 16; offset += 8) { + size_t allocated = (64 - offset); + void *test_ptr = malloc(allocated); + size_t wasted = malloc_usable_size(test_ptr) - allocated; + free(test_ptr); + + if (wasted == 0) { + return offset; + } } - return obj; + return 0; } - -static inline VALUE -RVALUE_INFANT_P(VALUE obj) +#else +static size_t +gc_compute_malloc_offset(void) { - check_gen_consistency(obj); - return !FL_TEST2(obj, FL_PROMOTED); + // If we don't have malloc_usable_size, we use powers of 2. + return 0; } +#endif -static inline VALUE -RVALUE_OLD_BITMAP_P(VALUE obj) +size_t +rb_malloc_grow_capa(size_t current, size_t type_size) { - check_gen_consistency(obj); - return (RVALUE_OLDGEN_BITMAP(obj) != 0); -} + size_t current_capacity = current; + if (current_capacity < 4) { + current_capacity = 4; + } + current_capacity *= type_size; -static inline VALUE -RVALUE_OLD_P(VALUE obj) -{ - check_gen_consistency(obj); -#if RGENGC_THREEGEN - return FL_TEST2(obj, FL_PROMOTED) && RVALUE_OLD_BITMAP_P(obj); -#else - return FL_TEST2(obj, FL_PROMOTED); -#endif + // We double the current capacity. + size_t new_capacity = (current_capacity * 2); + + // And round up to the next power of 2 if it's not already one. + if (rb_popcount64(new_capacity) != 1) { + new_capacity = (size_t)(1 << (64 - nlz_int64(new_capacity))); + } + + new_capacity -= malloc_offset; + new_capacity /= type_size; + if (current > new_capacity) { + rb_bug("rb_malloc_grow_capa: current_capacity=%zu, new_capacity=%zu, malloc_offset=%zu", current, new_capacity, malloc_offset); + } + RUBY_ASSERT(new_capacity > current); + return new_capacity; } -static inline VALUE -RVALUE_PROMOTED_P(VALUE obj) +static inline struct rbimpl_size_overflow_tag +size_mul_add_overflow(size_t x, size_t y, size_t z) /* x * y + z */ { - check_gen_consistency(obj); - return FL_TEST2(obj, FL_PROMOTED); + struct rbimpl_size_overflow_tag t = rbimpl_size_mul_overflow(x, y); + struct rbimpl_size_overflow_tag u = rbimpl_size_add_overflow(t.result, z); + return (struct rbimpl_size_overflow_tag) { t.overflowed || u.overflowed, u.result }; } -static inline void -RVALUE_PROMOTE_INFANT(VALUE obj) +static inline struct rbimpl_size_overflow_tag +size_mul_add_mul_overflow(size_t x, size_t y, size_t z, size_t w) /* x * y + z * w */ { - check_gen_consistency(obj); - if (RGENGC_CHECK_MODE && !RVALUE_INFANT_P(obj)) rb_bug("RVALUE_PROMOTE_INFANT: %p (%s) is not infant object.", (void *)obj, obj_type_name(obj)); - FL_SET2(obj, FL_PROMOTED); -#if !RGENGC_THREEGEN - MARK_IN_BITMAP(GET_HEAP_OLDGEN_BITS(obj), obj); -#endif - check_gen_consistency(obj); + struct rbimpl_size_overflow_tag t = rbimpl_size_mul_overflow(x, y); + struct rbimpl_size_overflow_tag u = rbimpl_size_mul_overflow(z, w); + struct rbimpl_size_overflow_tag v = rbimpl_size_add_overflow(t.result, u.result); + return (struct rbimpl_size_overflow_tag) { t.overflowed || u.overflowed || v.overflowed, v.result }; +} -#if RGENGC_PROFILE >= 1 - { - rb_objspace_t *objspace = &rb_objspace; - objspace->profile.promote_infant_count++; +PRINTF_ARGS(NORETURN(static void gc_raise(VALUE, const char*, ...)), 2, 3); -#if RGENGC_PROFILE >= 2 - objspace->profile.promote_infant_types[BUILTIN_TYPE(obj)]++; -#endif +static inline size_t +size_mul_or_raise(size_t x, size_t y, VALUE exc) +{ + struct rbimpl_size_overflow_tag t = rbimpl_size_mul_overflow(x, y); + if (LIKELY(!t.overflowed)) { + return t.result; + } + else if (rb_during_gc()) { + rb_memerror(); /* or...? */ + } + else { + gc_raise( + exc, + "integer overflow: %"PRIuSIZE + " * %"PRIuSIZE + " > %"PRIuSIZE, + x, y, (size_t)SIZE_MAX); } -#endif } -#if RGENGC_THREEGEN -/* - * Two gen: Infant -> Old. - * Three gen: Infant -> Young -> Old. - */ -static inline VALUE -RVALUE_YOUNG_P(VALUE obj) +size_t +rb_size_mul_or_raise(size_t x, size_t y, VALUE exc) { - check_gen_consistency(obj); - return FL_TEST2(obj, FL_PROMOTED) && (RVALUE_OLDGEN_BITMAP(obj) == 0); + return size_mul_or_raise(x, y, exc); } -static inline void -RVALUE_PROMOTE_YOUNG(VALUE obj) +static inline size_t +size_mul_add_or_raise(size_t x, size_t y, size_t z, VALUE exc) { - check_gen_consistency(obj); - if (RGENGC_CHECK_MODE && !RVALUE_YOUNG_P(obj)) rb_bug("RVALUE_PROMOTE_YOUNG: %p (%s) is not young object.", (void *)obj, obj_type_name(obj)); - MARK_IN_BITMAP(GET_HEAP_OLDGEN_BITS(obj), obj); - check_gen_consistency(obj); - -#if RGENGC_PROFILE >= 1 - { - rb_objspace_t *objspace = &rb_objspace; - objspace->profile.promote_young_count++; -#if RGENGC_PROFILE >= 2 - objspace->profile.promote_young_types[BUILTIN_TYPE(obj)]++; -#endif + struct rbimpl_size_overflow_tag t = size_mul_add_overflow(x, y, z); + if (LIKELY(!t.overflowed)) { + return t.result; + } + else if (rb_during_gc()) { + rb_memerror(); /* or...? */ + } + else { + gc_raise( + exc, + "integer overflow: %"PRIuSIZE + " * %"PRIuSIZE + " + %"PRIuSIZE + " > %"PRIuSIZE, + x, y, z, (size_t)SIZE_MAX); } -#endif } -static inline void -RVALUE_DEMOTE_FROM_YOUNG(VALUE obj) +size_t +rb_size_mul_add_or_raise(size_t x, size_t y, size_t z, VALUE exc) { - if (RGENGC_CHECK_MODE && !RVALUE_YOUNG_P(obj)) - rb_bug("RVALUE_DEMOTE_FROM_YOUNG: %p (%s) is not young object.", (void *)obj, obj_type_name(obj)); - - check_gen_consistency(obj); - FL_UNSET2(obj, FL_PROMOTED); - check_gen_consistency(obj); + return size_mul_add_or_raise(x, y, z, exc); } -#endif -static inline void -RVALUE_DEMOTE_FROM_OLD(VALUE obj) +static inline size_t +size_mul_add_mul_or_raise(size_t x, size_t y, size_t z, size_t w, VALUE exc) { - if (RGENGC_CHECK_MODE && !RVALUE_OLD_P(obj)) - rb_bug("RVALUE_DEMOTE_FROM_OLD: %p (%s) is not old object.", (void *)obj, obj_type_name(obj)); - - check_gen_consistency(obj); - FL_UNSET2(obj, FL_PROMOTED); - CLEAR_IN_BITMAP(GET_HEAP_OLDGEN_BITS(obj), obj); - check_gen_consistency(obj); + struct rbimpl_size_overflow_tag t = size_mul_add_mul_overflow(x, y, z, w); + if (LIKELY(!t.overflowed)) { + return t.result; + } + else if (rb_during_gc()) { + rb_memerror(); /* or...? */ + } + else { + gc_raise( + exc, + "integer overflow: %"PRIdSIZE + " * %"PRIdSIZE + " + %"PRIdSIZE + " * %"PRIdSIZE + " > %"PRIdSIZE, + x, y, z, w, (size_t)SIZE_MAX); + } } -#endif /* USE_RGENGC */ - -/* - --------------------------- ObjectSpace ----------------------------- -*/ - -#if defined(ENABLE_VM_OBJSPACE) && ENABLE_VM_OBJSPACE -rb_objspace_t * -rb_objspace_alloc(void) +#if defined(HAVE_RB_GC_GUARDED_PTR_VAL) && HAVE_RB_GC_GUARDED_PTR_VAL +/* trick the compiler into thinking a external signal handler uses this */ +volatile VALUE rb_gc_guarded_val; +volatile VALUE * +rb_gc_guarded_ptr_val(volatile VALUE *ptr, VALUE val) { - rb_objspace_t *objspace = malloc(sizeof(rb_objspace_t)); - memset(objspace, 0, sizeof(*objspace)); - ruby_gc_stress = ruby_initial_gc_stress; - - malloc_limit = gc_params.malloc_limit_min; + rb_gc_guarded_val = val; - return objspace; + return ptr; } #endif -#if defined(ENABLE_VM_OBJSPACE) && ENABLE_VM_OBJSPACE -static void free_stack_chunks(mark_stack_t *); -static void heap_page_free(rb_objspace_t *objspace, struct heap_page *page); +static const char *obj_type_name(VALUE obj); +static st_table *id2ref_tbl; +#include "gc/default/default.c" + +#if USE_MODULAR_GC && !defined(HAVE_DLOPEN) +# error "Modular GC requires dlopen" +#elif USE_MODULAR_GC +#include <dlfcn.h> + +typedef struct gc_function_map { + // Bootup + void *(*objspace_alloc)(void); + void (*objspace_init)(void *objspace_ptr); + void *(*ractor_cache_alloc)(void *objspace_ptr, void *ractor); + void (*set_params)(void *objspace_ptr); + void (*init)(void); + size_t *(*heap_sizes)(void *objspace_ptr); + // Shutdown + void (*shutdown_free_objects)(void *objspace_ptr); + void (*objspace_free)(void *objspace_ptr); + void (*ractor_cache_free)(void *objspace_ptr, void *cache); + // GC + void (*start)(void *objspace_ptr, bool full_mark, bool immediate_mark, bool immediate_sweep, bool compact); + bool (*during_gc_p)(void *objspace_ptr); + void (*prepare_heap)(void *objspace_ptr); + void (*gc_enable)(void *objspace_ptr); + void (*gc_disable)(void *objspace_ptr, bool finish_current_gc); + bool (*gc_enabled_p)(void *objspace_ptr); + VALUE (*config_get)(void *objpace_ptr); + void (*config_set)(void *objspace_ptr, VALUE hash); + void (*stress_set)(void *objspace_ptr, VALUE flag); + VALUE (*stress_get)(void *objspace_ptr); + struct rb_gc_vm_context *(*get_vm_context)(void *objspace_ptr); + // Object allocation + VALUE (*new_obj)(void *objspace_ptr, void *cache_ptr, VALUE klass, VALUE flags, bool wb_protected, size_t alloc_size); + size_t (*obj_slot_size)(VALUE obj); + size_t (*heap_id_for_size)(void *objspace_ptr, size_t size); + bool (*size_allocatable_p)(size_t size); + // Malloc + void *(*malloc)(void *objspace_ptr, size_t size, bool gc_allowed); + void *(*calloc)(void *objspace_ptr, size_t size, bool gc_allowed); + void *(*realloc)(void *objspace_ptr, void *ptr, size_t new_size, size_t old_size, bool gc_allowed); + void (*free)(void *objspace_ptr, void *ptr, size_t old_size); + void (*adjust_memory_usage)(void *objspace_ptr, ssize_t diff); + // Marking + void (*mark)(void *objspace_ptr, VALUE obj); + void (*mark_and_move)(void *objspace_ptr, VALUE *ptr); + void (*mark_and_pin)(void *objspace_ptr, VALUE obj); + void (*mark_maybe)(void *objspace_ptr, VALUE obj); + // Weak references + void (*declare_weak_references)(void *objspace_ptr, VALUE obj); + bool (*handle_weak_references_alive_p)(void *objspace_ptr, VALUE obj); + // Compaction + void (*register_pinning_obj)(void *objspace_ptr, VALUE obj); + bool (*object_moved_p)(void *objspace_ptr, VALUE obj); + VALUE (*location)(void *objspace_ptr, VALUE value); + // Write barriers + void (*writebarrier)(void *objspace_ptr, VALUE a, VALUE b); + void (*writebarrier_unprotect)(void *objspace_ptr, VALUE obj); + void (*writebarrier_remember)(void *objspace_ptr, VALUE obj); + // Heap walking + void (*each_objects)(void *objspace_ptr, int (*callback)(void *, void *, size_t, void *), void *data); + void (*each_object)(void *objspace_ptr, void (*func)(VALUE obj, void *data), void *data); + // Finalizers + void (*make_zombie)(void *objspace_ptr, VALUE obj, void (*dfree)(void *), void *data); + VALUE (*define_finalizer)(void *objspace_ptr, VALUE obj, VALUE block); + void (*undefine_finalizer)(void *objspace_ptr, VALUE obj); + void (*copy_finalizer)(void *objspace_ptr, VALUE dest, VALUE obj); + void (*shutdown_call_finalizer)(void *objspace_ptr); + // Forking + void (*before_fork)(void *objspace_ptr); + void (*after_fork)(void *objspace_ptr, rb_pid_t pid); + // Statistics + void (*set_measure_total_time)(void *objspace_ptr, VALUE flag); + bool (*get_measure_total_time)(void *objspace_ptr); + unsigned long long (*get_total_time)(void *objspace_ptr); + size_t (*gc_count)(void *objspace_ptr); + VALUE (*latest_gc_info)(void *objspace_ptr, VALUE key); + VALUE (*stat)(void *objspace_ptr, VALUE hash_or_sym); + VALUE (*stat_heap)(void *objspace_ptr, VALUE heap_name, VALUE hash_or_sym); + const char *(*active_gc_name)(void); + // Miscellaneous + struct rb_gc_object_metadata_entry *(*object_metadata)(void *objspace_ptr, VALUE obj); + bool (*pointer_to_heap_p)(void *objspace_ptr, const void *ptr); + bool (*garbage_object_p)(void *objspace_ptr, VALUE obj); + void (*set_event_hook)(void *objspace_ptr, const rb_event_flag_t event); + void (*copy_attributes)(void *objspace_ptr, VALUE dest, VALUE obj); + + bool modular_gc_loaded_p; +} rb_gc_function_map_t; + +static rb_gc_function_map_t rb_gc_functions; + +# define RUBY_GC_LIBRARY "RUBY_GC_LIBRARY" +# define MODULAR_GC_DIR STRINGIZE(modular_gc_dir) -void -rb_objspace_free(rb_objspace_t *objspace) -{ - gc_rest_sweep(objspace); +static void +ruby_modular_gc_init(void) +{ + // Assert that the directory path ends with a / + RUBY_ASSERT_ALWAYS(MODULAR_GC_DIR[sizeof(MODULAR_GC_DIR) - 2] == '/'); + + const char *gc_so_file = getenv(RUBY_GC_LIBRARY); + + rb_gc_function_map_t gc_functions = { 0 }; + + char *gc_so_path = NULL; + void *handle = NULL; + if (gc_so_file) { + /* Check to make sure that gc_so_file matches /[\w-_]+/ so that it does + * not load a shared object outside of the directory. */ + for (size_t i = 0; i < strlen(gc_so_file); i++) { + char c = gc_so_file[i]; + if (isalnum(c)) continue; + switch (c) { + case '-': + case '_': + break; + default: + fprintf(stderr, "Only alphanumeric, dash, and underscore is allowed in "RUBY_GC_LIBRARY"\n"); + exit(EXIT_FAILURE); + } + } - if (objspace->profile.records) { - free(objspace->profile.records); - objspace->profile.records = 0; - } + size_t gc_so_path_size = strlen(MODULAR_GC_DIR "librubygc." DLEXT) + strlen(gc_so_file) + 1; +#ifdef LOAD_RELATIVE + Dl_info dli; + size_t prefix_len = 0; + if (dladdr((void *)(uintptr_t)ruby_modular_gc_init, &dli)) { + const char *base = strrchr(dli.dli_fname, '/'); + if (base) { + size_t tail = 0; +# define end_with_p(lit) \ + (prefix_len >= (tail = rb_strlen_lit(lit)) && \ + memcmp(base - tail, lit, tail) == 0) + + prefix_len = base - dli.dli_fname; + if (end_with_p("/bin") || end_with_p("/lib")) { + prefix_len -= tail; + } + prefix_len += MODULAR_GC_DIR[0] != '/'; + gc_so_path_size += prefix_len; + } + } +#endif + gc_so_path = alloca(gc_so_path_size); + { + size_t gc_so_path_idx = 0; +#define GC_SO_PATH_APPEND(str) do { \ + gc_so_path_idx += strlcpy(gc_so_path + gc_so_path_idx, str, gc_so_path_size - gc_so_path_idx); \ +} while (0) +#ifdef LOAD_RELATIVE + if (prefix_len > 0) { + memcpy(gc_so_path, dli.dli_fname, prefix_len); + gc_so_path_idx = prefix_len; + } +#endif + GC_SO_PATH_APPEND(MODULAR_GC_DIR "librubygc."); + GC_SO_PATH_APPEND(gc_so_file); + GC_SO_PATH_APPEND(DLEXT); + GC_ASSERT(gc_so_path_idx == gc_so_path_size - 1); +#undef GC_SO_PATH_APPEND + } - if (global_List) { - struct gc_list *list, *next; - for (list = global_List; list; list = next) { - next = list->next; - xfree(list); - } - } - if (heap_pages_sorted) { - size_t i; - for (i = 0; i < heap_pages_used; ++i) { - heap_page_free(objspace, heap_pages_sorted[i]); - } - free(heap_pages_sorted); - heap_pages_used = 0; - heap_pages_length = 0; - heap_pages_lomem = 0; - heap_pages_himem = 0; + handle = dlopen(gc_so_path, RTLD_LAZY | RTLD_GLOBAL); + if (!handle) { + fprintf(stderr, "ruby_modular_gc_init: Shared library %s cannot be opened: %s\n", gc_so_path, dlerror()); + exit(EXIT_FAILURE); + } - objspace->eden_heap.page_length = 0; - objspace->eden_heap.total_slots = 0; - objspace->eden_heap.pages = NULL; + gc_functions.modular_gc_loaded_p = true; } - free_stack_chunks(&objspace->mark_stack); - free(objspace); -} -#endif + unsigned int err_count = 0; + +# define load_modular_gc_func(name) do { \ + if (handle) { \ + const char *func_name = "rb_gc_impl_" #name; \ + gc_functions.name = dlsym(handle, func_name); \ + if (!gc_functions.name) { \ + fprintf(stderr, "ruby_modular_gc_init: %s function not exported by library %s\n", func_name, gc_so_path); \ + err_count++; \ + } \ + } \ + else { \ + gc_functions.name = rb_gc_impl_##name; \ + } \ +} while (0) + + // Bootup + load_modular_gc_func(objspace_alloc); + load_modular_gc_func(objspace_init); + load_modular_gc_func(ractor_cache_alloc); + load_modular_gc_func(set_params); + load_modular_gc_func(init); + load_modular_gc_func(heap_sizes); + // Shutdown + load_modular_gc_func(shutdown_free_objects); + load_modular_gc_func(objspace_free); + load_modular_gc_func(ractor_cache_free); + // GC + load_modular_gc_func(start); + load_modular_gc_func(during_gc_p); + load_modular_gc_func(prepare_heap); + load_modular_gc_func(gc_enable); + load_modular_gc_func(gc_disable); + load_modular_gc_func(gc_enabled_p); + load_modular_gc_func(config_set); + load_modular_gc_func(config_get); + load_modular_gc_func(stress_set); + load_modular_gc_func(stress_get); + load_modular_gc_func(get_vm_context); + // Object allocation + load_modular_gc_func(new_obj); + load_modular_gc_func(obj_slot_size); + load_modular_gc_func(heap_id_for_size); + load_modular_gc_func(size_allocatable_p); + // Malloc + load_modular_gc_func(malloc); + load_modular_gc_func(calloc); + load_modular_gc_func(realloc); + load_modular_gc_func(free); + load_modular_gc_func(adjust_memory_usage); + // Marking + load_modular_gc_func(mark); + load_modular_gc_func(mark_and_move); + load_modular_gc_func(mark_and_pin); + load_modular_gc_func(mark_maybe); + // Weak references + load_modular_gc_func(declare_weak_references); + load_modular_gc_func(handle_weak_references_alive_p); + // Compaction + load_modular_gc_func(register_pinning_obj); + load_modular_gc_func(object_moved_p); + load_modular_gc_func(location); + // Write barriers + load_modular_gc_func(writebarrier); + load_modular_gc_func(writebarrier_unprotect); + load_modular_gc_func(writebarrier_remember); + // Heap walking + load_modular_gc_func(each_objects); + load_modular_gc_func(each_object); + // Finalizers + load_modular_gc_func(make_zombie); + load_modular_gc_func(define_finalizer); + load_modular_gc_func(undefine_finalizer); + load_modular_gc_func(copy_finalizer); + load_modular_gc_func(shutdown_call_finalizer); + // Forking + load_modular_gc_func(before_fork); + load_modular_gc_func(after_fork); + // Statistics + load_modular_gc_func(set_measure_total_time); + load_modular_gc_func(get_measure_total_time); + load_modular_gc_func(get_total_time); + load_modular_gc_func(gc_count); + load_modular_gc_func(latest_gc_info); + load_modular_gc_func(stat); + load_modular_gc_func(stat_heap); + load_modular_gc_func(active_gc_name); + // Miscellaneous + load_modular_gc_func(object_metadata); + load_modular_gc_func(pointer_to_heap_p); + load_modular_gc_func(garbage_object_p); + load_modular_gc_func(set_event_hook); + load_modular_gc_func(copy_attributes); + + if (err_count > 0) { + fprintf(stderr, "ruby_modular_gc_init: found %u missing exports in library %s\n", err_count, gc_so_path); + exit(EXIT_FAILURE); + } + +# undef load_modular_gc_func + + rb_gc_functions = gc_functions; +} + +// Bootup +# define rb_gc_impl_objspace_alloc rb_gc_functions.objspace_alloc +# define rb_gc_impl_objspace_init rb_gc_functions.objspace_init +# define rb_gc_impl_ractor_cache_alloc rb_gc_functions.ractor_cache_alloc +# define rb_gc_impl_set_params rb_gc_functions.set_params +# define rb_gc_impl_init rb_gc_functions.init +# define rb_gc_impl_heap_sizes rb_gc_functions.heap_sizes +// Shutdown +# define rb_gc_impl_shutdown_free_objects rb_gc_functions.shutdown_free_objects +# define rb_gc_impl_objspace_free rb_gc_functions.objspace_free +# define rb_gc_impl_ractor_cache_free rb_gc_functions.ractor_cache_free +// GC +# define rb_gc_impl_start rb_gc_functions.start +# define rb_gc_impl_during_gc_p rb_gc_functions.during_gc_p +# define rb_gc_impl_prepare_heap rb_gc_functions.prepare_heap +# define rb_gc_impl_gc_enable rb_gc_functions.gc_enable +# define rb_gc_impl_gc_disable rb_gc_functions.gc_disable +# define rb_gc_impl_gc_enabled_p rb_gc_functions.gc_enabled_p +# define rb_gc_impl_config_get rb_gc_functions.config_get +# define rb_gc_impl_config_set rb_gc_functions.config_set +# define rb_gc_impl_stress_set rb_gc_functions.stress_set +# define rb_gc_impl_stress_get rb_gc_functions.stress_get +# define rb_gc_impl_get_vm_context rb_gc_functions.get_vm_context +// Object allocation +# define rb_gc_impl_new_obj rb_gc_functions.new_obj +# define rb_gc_impl_obj_slot_size rb_gc_functions.obj_slot_size +# define rb_gc_impl_heap_id_for_size rb_gc_functions.heap_id_for_size +# define rb_gc_impl_size_allocatable_p rb_gc_functions.size_allocatable_p +// Malloc +# define rb_gc_impl_malloc rb_gc_functions.malloc +# define rb_gc_impl_calloc rb_gc_functions.calloc +# define rb_gc_impl_realloc rb_gc_functions.realloc +# define rb_gc_impl_free rb_gc_functions.free +# define rb_gc_impl_adjust_memory_usage rb_gc_functions.adjust_memory_usage +// Marking +# define rb_gc_impl_mark rb_gc_functions.mark +# define rb_gc_impl_mark_and_move rb_gc_functions.mark_and_move +# define rb_gc_impl_mark_and_pin rb_gc_functions.mark_and_pin +# define rb_gc_impl_mark_maybe rb_gc_functions.mark_maybe +// Weak references +# define rb_gc_impl_declare_weak_references rb_gc_functions.declare_weak_references +# define rb_gc_impl_handle_weak_references_alive_p rb_gc_functions.handle_weak_references_alive_p +// Compaction +# define rb_gc_impl_register_pinning_obj rb_gc_functions.register_pinning_obj +# define rb_gc_impl_object_moved_p rb_gc_functions.object_moved_p +# define rb_gc_impl_location rb_gc_functions.location +// Write barriers +# define rb_gc_impl_writebarrier rb_gc_functions.writebarrier +# define rb_gc_impl_writebarrier_unprotect rb_gc_functions.writebarrier_unprotect +# define rb_gc_impl_writebarrier_remember rb_gc_functions.writebarrier_remember +// Heap walking +# define rb_gc_impl_each_objects rb_gc_functions.each_objects +# define rb_gc_impl_each_object rb_gc_functions.each_object +// Finalizers +# define rb_gc_impl_make_zombie rb_gc_functions.make_zombie +# define rb_gc_impl_define_finalizer rb_gc_functions.define_finalizer +# define rb_gc_impl_undefine_finalizer rb_gc_functions.undefine_finalizer +# define rb_gc_impl_copy_finalizer rb_gc_functions.copy_finalizer +# define rb_gc_impl_shutdown_call_finalizer rb_gc_functions.shutdown_call_finalizer +// Forking +# define rb_gc_impl_before_fork rb_gc_functions.before_fork +# define rb_gc_impl_after_fork rb_gc_functions.after_fork +// Statistics +# define rb_gc_impl_set_measure_total_time rb_gc_functions.set_measure_total_time +# define rb_gc_impl_get_measure_total_time rb_gc_functions.get_measure_total_time +# define rb_gc_impl_get_total_time rb_gc_functions.get_total_time +# define rb_gc_impl_gc_count rb_gc_functions.gc_count +# define rb_gc_impl_latest_gc_info rb_gc_functions.latest_gc_info +# define rb_gc_impl_stat rb_gc_functions.stat +# define rb_gc_impl_stat_heap rb_gc_functions.stat_heap +# define rb_gc_impl_active_gc_name rb_gc_functions.active_gc_name +// Miscellaneous +# define rb_gc_impl_object_metadata rb_gc_functions.object_metadata +# define rb_gc_impl_pointer_to_heap_p rb_gc_functions.pointer_to_heap_p +# define rb_gc_impl_garbage_object_p rb_gc_functions.garbage_object_p +# define rb_gc_impl_set_event_hook rb_gc_functions.set_event_hook +# define rb_gc_impl_copy_attributes rb_gc_functions.copy_attributes +#endif + +#ifdef RUBY_ASAN_ENABLED static void -heap_pages_expand_sorted(rb_objspace_t *objspace) +asan_death_callback(void) { - size_t next_length = heap_pages_increment; - next_length += heap_eden->page_length; - next_length += heap_tomb->page_length; + if (GET_VM()) { + rb_bug_without_die("ASAN error"); + } +} +#endif - if (next_length > heap_pages_length) { - struct heap_page **sorted; - size_t size = next_length * sizeof(struct heap_page *); +static VALUE initial_stress = Qfalse; - rgengc_report(3, objspace, "heap_pages_expand_sorted: next_length: %d, size: %d\n", (int)next_length, (int)size); +void * +rb_objspace_alloc(void) +{ +#if USE_MODULAR_GC + ruby_modular_gc_init(); +#endif - if (heap_pages_length > 0) { - sorted = (struct heap_page **)realloc(heap_pages_sorted, size); - if (sorted) heap_pages_sorted = sorted; - } - else { - sorted = heap_pages_sorted = (struct heap_page **)malloc(size); - } + void *objspace = rb_gc_impl_objspace_alloc(); + ruby_current_vm_ptr->gc.objspace = objspace; + rb_gc_impl_objspace_init(objspace); + rb_gc_impl_stress_set(objspace, initial_stress); - if (sorted == 0) { - during_gc = 0; - rb_memerror(); - } +#ifdef RUBY_ASAN_ENABLED + __sanitizer_set_death_callback(asan_death_callback); +#endif - heap_pages_length = next_length; - } + return objspace; } -static inline void -heap_page_add_freeobj(rb_objspace_t *objspace, struct heap_page *page, VALUE obj) +void +rb_objspace_free(void *objspace) { - RVALUE *p = (RVALUE *)obj; - p->as.free.flags = 0; - p->as.free.next = page->freelist; - page->freelist = p; - rgengc_report(3, objspace, "heap_page_add_freeobj: %p (%s) is added to freelist\n", p, obj_type_name(obj)); + rb_gc_impl_objspace_free(objspace); } -static inline void -heap_add_freepage(rb_objspace_t *objspace, rb_heap_t *heap, struct heap_page *page) +size_t +rb_gc_obj_slot_size(VALUE obj) { - if (page->freelist) { - page->free_next = heap->free_pages; - heap->free_pages = page; - } + return rb_gc_impl_obj_slot_size(obj); } -static void -heap_unlink_page(rb_objspace_t *objspace, rb_heap_t *heap, struct heap_page *page) +static inline void +gc_validate_pc(VALUE obj) { - if (page->prev) page->prev->next = page->next; - if (page->next) page->next->prev = page->prev; - if (heap->pages == page) heap->pages = page->next; - page->prev = NULL; - page->next = NULL; - page->heap = NULL; - heap->page_length--; - heap->total_slots -= page->limit; -} +#if RUBY_DEBUG + // IMEMOs and objects without a class (e.g managed id table) are not traceable + if (RB_TYPE_P(obj, T_IMEMO) || !CLASS_OF(obj)) return; -static void -heap_page_free(rb_objspace_t *objspace, struct heap_page *page) -{ - heap_pages_used--; - aligned_free(page->body); - free(page); + rb_execution_context_t *ec = GET_EC(); + const rb_control_frame_t *cfp = ec->cfp; + if (cfp && VM_FRAME_RUBYFRAME_P(cfp) && CFP_PC(cfp)) { + const VALUE *iseq_encoded = ISEQ_BODY(CFP_ISEQ(cfp))->iseq_encoded; + const VALUE *iseq_encoded_end = iseq_encoded + ISEQ_BODY(CFP_ISEQ(cfp))->iseq_size; + RUBY_ASSERT(CFP_PC(cfp) >= iseq_encoded, "PC not set when allocating, breaking tracing"); + RUBY_ASSERT(CFP_PC(cfp) <= iseq_encoded_end, "PC not set when allocating, breaking tracing"); + } +#endif } +NOINLINE(static void gc_newobj_hook(VALUE obj)); static void -heap_pages_free_unused_pages(rb_objspace_t *objspace) -{ - size_t i, j; - - for (i = j = 1; j < heap_pages_used; i++) { - struct heap_page *page = heap_pages_sorted[i]; - - if (page->heap == heap_tomb && page->final_slots == 0) { - if (heap_pages_swept_slots - page->limit > heap_pages_max_free_slots) { - if (0) fprintf(stderr, "heap_pages_free_unused_pages: %d free page %p, heap_pages_swept_slots: %d, heap_pages_max_free_slots: %d\n", - (int)i, page, (int)heap_pages_swept_slots, (int)heap_pages_max_free_slots); - heap_pages_swept_slots -= page->limit; - heap_unlink_page(objspace, heap_tomb, page); - heap_page_free(objspace, page); - continue; - } - else { - /* fprintf(stderr, "heap_pages_free_unused_pages: remain!!\n"); */ - } - } - if (i != j) { - heap_pages_sorted[j] = page; - } - j++; +gc_newobj_hook(VALUE obj) +{ + int lev = RB_GC_VM_LOCK_NO_BARRIER(); + { + size_t slot_size = rb_gc_obj_slot_size(obj); + memset((char *)obj + sizeof(struct RBasic), 0, slot_size - sizeof(struct RBasic)); + + /* We must disable GC here because the callback could call xmalloc + * which could potentially trigger a GC, and a lot of code is unsafe + * to trigger a GC right after an object has been allocated because + * they perform initialization for the object and assume that the + * GC does not trigger before then. */ + bool gc_disabled = RTEST(rb_gc_disable_no_rest()); + { + rb_gc_event_hook(obj, RUBY_INTERNAL_EVENT_NEWOBJ); + } + if (!gc_disabled) rb_gc_enable(); } - assert(j == heap_pages_used); + RB_GC_VM_UNLOCK_NO_BARRIER(lev); } -static struct heap_page * -heap_page_allocate(rb_objspace_t *objspace) +VALUE +rb_newobj(rb_execution_context_t *ec, VALUE klass, VALUE flags, shape_id_t shape_id, bool wb_protected, size_t size) { - RVALUE *start, *end, *p; - struct heap_page *page; - struct heap_page_body *page_body = 0; - size_t hi, lo, mid; - size_t limit = HEAP_OBJ_LIMIT; + GC_ASSERT((flags & FL_WB_PROTECTED) == 0); + rb_ractor_t *cr = rb_ec_ractor_ptr(ec); + VALUE obj = rb_gc_impl_new_obj(rb_gc_get_objspace(), cr->newobj_cache, klass, flags, wb_protected, size); - /* assign heap_page body (contains heap_page_header and RVALUEs) */ - page_body = (struct heap_page_body *)aligned_malloc(HEAP_ALIGN, HEAP_SIZE); - if (page_body == 0) { - during_gc = 0; - rb_memerror(); - } - - /* assign heap_page entry */ - page = (struct heap_page *)malloc(sizeof(struct heap_page)); - if (page == 0) { - aligned_free(page_body); - during_gc = 0; - rb_memerror(); - } - MEMZERO((void*)page, struct heap_page, 1); - - page->body = page_body; - - /* setup heap_pages_sorted */ - lo = 0; - hi = heap_pages_used; - while (lo < hi) { - struct heap_page *mid_page; - - mid = (lo + hi) / 2; - mid_page = heap_pages_sorted[mid]; - if (mid_page->body < page_body) { - lo = mid + 1; - } - else if (mid_page->body > page_body) { - hi = mid; - } - else { - rb_bug("same heap page is allocated: %p at %"PRIuVALUE, (void *)page_body, (VALUE)mid); - } - } - if (hi < heap_pages_used) { - MEMMOVE(&heap_pages_sorted[hi+1], &heap_pages_sorted[hi], struct heap_page_header*, heap_pages_used - hi); - } +#if RACTOR_CHECK_MODE + void rb_ractor_setup_belonging(VALUE obj); + rb_ractor_setup_belonging(obj); +#endif - heap_pages_sorted[hi] = page; + RBASIC_SET_SHAPE_ID_NO_CHECKS(obj, shape_id); - heap_pages_used++; - assert(heap_pages_used <= heap_pages_length); + gc_validate_pc(obj); - /* adjust obj_limit (object number available in this page) */ - start = (RVALUE*)((VALUE)page_body + sizeof(struct heap_page_header)); - if ((VALUE)start % sizeof(RVALUE) != 0) { - int delta = (int)(sizeof(RVALUE) - ((VALUE)start % sizeof(RVALUE))); - start = (RVALUE*)((VALUE)start + delta); - limit = (HEAP_SIZE - (size_t)((VALUE)start - (VALUE)page_body))/sizeof(RVALUE); + if (UNLIKELY(rb_gc_event_hook_required_p(RUBY_INTERNAL_EVENT_NEWOBJ))) { + gc_newobj_hook(obj); } - end = start + limit; - - if (heap_pages_lomem == 0 || heap_pages_lomem > start) heap_pages_lomem = start; - if (heap_pages_himem < end) heap_pages_himem = end; - page->start = start; - page->limit = limit; - page_body->header.page = page; +#if RGENGC_CHECK_MODE +# ifndef GC_DEBUG_SLOT_FILL_SPECIAL_VALUE +# define GC_DEBUG_SLOT_FILL_SPECIAL_VALUE 255 +# endif - for (p = start; p != end; p++) { - rgengc_report(3, objspace, "assign_heap_page: %p is added to freelist\n"); - heap_page_add_freeobj(objspace, page, (VALUE)p); - } + memset( + (void *)(obj + sizeof(struct RBasic)), + GC_DEBUG_SLOT_FILL_SPECIAL_VALUE, + rb_gc_obj_slot_size(obj) - sizeof(struct RBasic) + ); +#endif - return page; + return obj; } -static struct heap_page * -heap_page_resurrect(rb_objspace_t *objspace) +VALUE +rb_ec_newobj_of(rb_execution_context_t *ec, VALUE klass, VALUE flags, size_t size) { - struct heap_page *page; + VALUE type = flags & T_MASK; + RUBY_ASSERT(type != T_OBJECT); + RUBY_ASSERT(type != T_DATA); + RUBY_ASSERT(type != T_CLASS); + RUBY_ASSERT(type != T_MODULE); + RUBY_ASSERT(type != T_ICLASS); + (void)type; - if ((page = heap_tomb->pages) != NULL) { - heap_unlink_page(objspace, heap_tomb, page); - return page; - } - return NULL; + return rb_newobj(ec, klass, flags, ROOT_SHAPE_ID | SHAPE_ID_LAYOUT_OTHER, true, size); } -static struct heap_page * -heap_page_create(rb_objspace_t *objspace) +VALUE +rb_newobj_of_with_shape(VALUE klass, VALUE flags, shape_id_t shape_id, size_t size) { - struct heap_page *page = heap_page_resurrect(objspace); - const char *method = "recycle"; - if (page == NULL) { - page = heap_page_allocate(objspace); - method = "allocate"; - } - if (0) fprintf(stderr, "heap_page_create: %s - %p, heap_pages_used: %d, heap_pages_used: %d, tomb->page_length: %d\n", - method, page, (int)heap_pages_length, (int)heap_pages_used, (int)heap_tomb->page_length); - return page; + return rb_newobj(GET_EC(), klass, flags, shape_id, true, size); } -static void -heap_add_page(rb_objspace_t *objspace, rb_heap_t *heap, struct heap_page *page) +VALUE +rb_newobj_of(VALUE klass, VALUE flags, size_t size) { - page->heap = heap; - page->next = heap->pages; - if (heap->pages) heap->pages->prev = page; - heap->pages = page; - heap->page_length++; - heap->total_slots += page->limit; + return rb_newobj(GET_EC(), klass, flags, ROOT_SHAPE_ID | SHAPE_ID_LAYOUT_OTHER, true, size); } -static void -heap_assign_page(rb_objspace_t *objspace, rb_heap_t *heap) +static +VALUE class_allocate_complex_instance(VALUE klass, uint32_t capacity) { - struct heap_page *page = heap_page_create(objspace); - heap_add_page(objspace, heap, page); - heap_add_freepage(objspace, heap, page); + shape_id_t initial_shape_id = rb_shape_id_with_robject_layout(rb_shape_root(rb_gc_heap_id_for_size(sizeof(struct RObject)))); + VALUE obj = rb_newobj_of_with_shape(klass, T_OBJECT, initial_shape_id, sizeof(struct RObject)); + rb_obj_init_complex(obj, rb_st_init_numtable_with_size(capacity)); + return obj; } -static void -heap_add_pages(rb_objspace_t *objspace, rb_heap_t *heap, size_t add) +VALUE +rb_class_allocate_instance(VALUE klass) { - size_t i; + uint32_t index_tbl_num_entries = RCLASS_MAX_IV_COUNT(klass); + VALUE obj; - heap_pages_increment = add; - heap_pages_expand_sorted(objspace); - for (i = 0; i < add; i++) { - heap_assign_page(objspace, heap); + // Directly start as COMPLEX if we know we're over the limit. + RUBY_ASSERT(rb_shape_tree.max_capacity > 0); + if (RB_UNLIKELY(index_tbl_num_entries > rb_shape_tree.max_capacity)) { + obj = class_allocate_complex_instance(klass, index_tbl_num_entries); } - heap_pages_increment = 0; -} + else { + size_t size = rb_obj_embedded_size(index_tbl_num_entries); + if (!rb_gc_size_allocatable_p(size)) { + size = sizeof(struct RObject); + } -static void -heap_set_increment(rb_objspace_t *objspace, size_t minimum_limit) -{ - size_t used = heap_pages_used - heap_tomb->page_length; - size_t next_used_limit = (size_t)(used * gc_params.growth_factor); - if (gc_params.growth_max_slots > 0) { - size_t max_used_limit = (size_t)(used + gc_params.growth_max_slots/HEAP_OBJ_LIMIT); - if (next_used_limit > max_used_limit) next_used_limit = max_used_limit; - } - if (next_used_limit == heap_pages_used) next_used_limit++; + // There might be a NEWOBJ tracepoint callback, and it may set fields. + // So the shape must be passed to `NEWOBJ_OF`. + obj = rb_newobj_of_with_shape(klass, T_OBJECT, rb_shape_id_with_robject_layout(rb_shape_root(rb_gc_heap_id_for_size(size))), size); - if (next_used_limit < minimum_limit) { - next_used_limit = minimum_limit; + #if RUBY_DEBUG + VALUE *ptr = ROBJECT_FIELDS(obj); + size_t fields_count = RSHAPE_LEN(RBASIC_SHAPE_ID(obj)); + for (size_t i = fields_count; i < ROBJECT_FIELDS_CAPACITY(obj); i++) { + ptr[i] = Qundef; + } + #endif } - heap_pages_increment = next_used_limit - used; - heap_pages_expand_sorted(objspace); +#if RUBY_DEBUG + if (rb_obj_class(obj) != rb_class_real(klass)) { + rb_bug("Expected rb_class_allocate_instance to set the class correctly"); + } +#endif - if (0) fprintf(stderr, "heap_set_increment: heap_pages_length: %d, heap_pages_used: %d, heap_pages_increment: %d, next_used_limit: %d\n", - (int)heap_pages_length, (int)heap_pages_used, (int)heap_pages_increment, (int)next_used_limit); + return obj; } -static int -heap_increment(rb_objspace_t *objspace, rb_heap_t *heap) +void +rb_gc_register_pinning_obj(VALUE obj) { - rgengc_report(5, objspace, "heap_increment: heap_pages_length: %d, heap_pages_inc: %d, heap->page_length: %d\n", - (int)heap_pages_length, (int)heap_pages_increment, (int)heap->page_length); + rb_gc_impl_register_pinning_obj(rb_gc_get_objspace(), obj); +} + +#define UNEXPECTED_NODE(func) \ + rb_bug(#func"(): GC does not handle T_NODE 0x%x(%p) 0x%"PRIxVALUE, \ + BUILTIN_TYPE(obj), (void*)(obj), RBASIC(obj)->flags) - if (heap_pages_increment > 0) { - heap_pages_increment--; - heap_assign_page(objspace, heap); - return TRUE; +static inline void +rb_data_object_check(VALUE klass) +{ + RUBY_ASSERT(!RCLASS_SINGLETON_P(klass)); + if (klass != rb_cObject && (rb_get_alloc_func(klass) == rb_class_allocate_instance)) { + rb_undef_alloc_func(klass); + rb_warn("undefining the allocator of T_DATA class %"PRIsVALUE, klass); } - return FALSE; } -static struct heap_page * -heap_prepare_freepage(rb_objspace_t *objspace, rb_heap_t *heap) +#define RTYPEDDATA_EMBEDDED_P rbimpl_typeddata_embedded_p +#define RB_DATA_TYPE_EMBEDDABLE_P(type) ((type)->flags & RUBY_TYPED_EMBEDDABLE) +#define RTYPEDDATA_EMBEDDABLE_P(obj) RB_DATA_TYPE_EMBEDDABLE_P(RTYPEDDATA_TYPE(obj)) + +static VALUE +typed_data_alloc(VALUE klass, VALUE typed_flag, void *datap, const rb_data_type_t *type, size_t size) { - if (!GC_ENABLE_LAZY_SWEEP && objspace->flags.dont_lazy_sweep) { - if (heap_increment(objspace, heap) == 0 && - garbage_collect(objspace, FALSE, TRUE, GPR_FLAG_NEWOBJ) == 0) { - goto err; - } - goto ok; - } + RBIMPL_NONNULL_ARG(type); + if (klass) rb_data_object_check(klass); + bool wb_protected = (type->flags & RUBY_FL_WB_PROTECTED) || !type->function.dmark; + VALUE obj = rb_newobj(GET_EC(), klass, T_DATA, ROOT_SHAPE_ID | SHAPE_ID_LAYOUT_RDATA, wb_protected, size); - if (!heap_ready_to_gc(objspace, heap)) return heap->free_pages; + rb_gc_register_pinning_obj(obj); - during_gc++; + struct RTypedData *data = (struct RTypedData *)obj; + data->fields_obj = 0; + *(VALUE *)&data->type = ((VALUE)type) | typed_flag; + data->data = datap; - if ((is_lazy_sweeping(heap) && gc_heap_lazy_sweep(objspace, heap)) || heap_increment(objspace, heap)) { - goto ok; - } + return obj; +} -#if GC_PROFILE_MORE_DETAIL - objspace->profile.prepare_time = 0; -#endif - if (garbage_collect_body(objspace, 0, 0, GPR_FLAG_NEWOBJ) == 0) { - err: - during_gc = 0; - rb_memerror(); +VALUE +rb_data_typed_object_wrap(VALUE klass, void *datap, const rb_data_type_t *type) +{ + if (UNLIKELY(RB_DATA_TYPE_EMBEDDABLE_P(type))) { + rb_raise(rb_eTypeError, "Cannot wrap an embeddable TypedData"); } - ok: - during_gc = 0; - return heap->free_pages; + + return typed_data_alloc(klass, 0, datap, type, sizeof(struct RTypedData)); } -static RVALUE * -heap_get_freeobj_from_next_freepage(rb_objspace_t *objspace, rb_heap_t *heap) +VALUE +rb_data_typed_object_zalloc(VALUE klass, size_t size, const rb_data_type_t *type) { - struct heap_page *page; - RVALUE *p; + if (RB_DATA_TYPE_EMBEDDABLE_P(type)) { + if (!(type->flags & RUBY_TYPED_FREE_IMMEDIATELY)) { + rb_raise(rb_eTypeError, "Embeddable TypedData must be freed immediately"); + } - page = heap->free_pages; - while (page == NULL) { - page = heap_prepare_freepage(objspace, heap); + size_t embed_size = offsetof(struct RTypedData, data) + size; + if (rb_gc_size_allocatable_p(embed_size)) { + VALUE obj = typed_data_alloc(klass, TYPED_DATA_EMBEDDED, 0, type, embed_size); + memset((char *)obj + offsetof(struct RTypedData, data), 0, size); + return obj; + } } - heap->free_pages = page->free_next; - heap->using_page = page; - p = page->freelist; - page->freelist = NULL; + VALUE obj = typed_data_alloc(klass, 0, NULL, type, sizeof(struct RTypedData)); + DATA_PTR(obj) = xcalloc(1, size); + return obj; +} - return p; +static size_t +ruby_xmalloc_usable_size(void *ptr) +{ +#ifdef HAVE_MALLOC_USABLE_SIZE +#if CALC_EXACT_MALLOC_SIZE + struct malloc_obj_info *info = (struct malloc_obj_info *)ptr - 1; + return malloc_usable_size(info) - sizeof(struct malloc_obj_info); +#else + return malloc_usable_size(ptr); +#endif +#else + return 0; +#endif } -static inline VALUE -heap_get_freeobj(rb_objspace_t *objspace, rb_heap_t *heap) +static size_t +rb_objspace_data_type_memsize(VALUE obj) { - RVALUE *p = heap->freelist; + size_t size = 0; + const void *ptr = RTYPEDDATA_GET_DATA(obj); + + if (ptr) { + if (RTYPEDDATA_EMBEDDABLE_P(obj) && !RTYPEDDATA_EMBEDDED_P(obj)) { + size += ruby_xmalloc_usable_size((void *)ptr); + } - while (1) { - if (p) { - heap->freelist = p->as.free.next; - return (VALUE)p; - } - else { - p = heap_get_freeobj_from_next_freepage(objspace, heap); - } + const rb_data_type_t *type = RTYPEDDATA_TYPE(obj); + if (type->function.dsize) { + size += type->function.dsize(ptr); + } } + + return size; } -void -rb_objspace_set_event_hook(const rb_event_flag_t event) +const char * +rb_objspace_data_type_name(VALUE obj) { - rb_objspace_t *objspace = &rb_objspace; - objspace->hook_events = event & RUBY_INTERNAL_EVENT_OBJSPACE_MASK; + return RTYPEDDATA_TYPE(obj)->wrap_struct_name; } -static void -gc_event_hook_body(rb_objspace_t *objspace, const rb_event_flag_t event, VALUE data) +void +rb_gc_declare_weak_references(VALUE obj) { - rb_thread_t *th = GET_THREAD(); - EXEC_EVENT_HOOK(th, event, th->cfp->self, 0, 0, data); + rb_gc_impl_declare_weak_references(rb_gc_get_objspace(), obj); } -#define gc_event_hook(objspace, event, data) do { \ - if (UNLIKELY((objspace)->hook_events & (event))) { \ - gc_event_hook_body((objspace), (event), (data)); \ - } \ -} while (0) - -static VALUE -newobj_of(VALUE klass, VALUE flags, VALUE v1, VALUE v2, VALUE v3) +bool +rb_gc_handle_weak_references_alive_p(VALUE obj) { - rb_objspace_t *objspace = &rb_objspace; - VALUE obj; + if (SPECIAL_CONST_P(obj)) return true; - if (UNLIKELY(during_gc)) { - dont_gc = 1; - during_gc = 0; - rb_bug("object allocation during garbage collection phase"); - } - - if (UNLIKELY(ruby_gc_stress && !ruby_disable_gc_stress)) { - if (!garbage_collect(objspace, FALSE, FALSE, GPR_FLAG_NEWOBJ)) { - during_gc = 0; - rb_memerror(); - } - } + return rb_gc_impl_handle_weak_references_alive_p(rb_gc_get_objspace(), obj); +} - obj = heap_get_freeobj(objspace, heap_eden); +void +rb_gc_handle_weak_references(VALUE obj) +{ + switch (BUILTIN_TYPE(obj)) { + case T_DATA: + { + const rb_data_type_t *type = RTYPEDDATA_TYPE(obj); - /* OBJSETUP */ - RBASIC(obj)->flags = flags; - RBASIC_SET_CLASS_RAW(obj, klass); - if (rb_safe_level() >= 3) FL_SET((obj), FL_TAINT); - RANY(obj)->as.values.v1 = v1; - RANY(obj)->as.values.v2 = v2; - RANY(obj)->as.values.v3 = v3; + if (type->function.handle_weak_references) { + (type->function.handle_weak_references)(RTYPEDDATA_GET_DATA(obj)); + } + else { + rb_bug( + "rb_gc_handle_weak_references: TypedData %s does not implement handle_weak_references", + RTYPEDDATA_TYPE(obj)->wrap_struct_name + ); + } + } + break; -#if GC_DEBUG - RANY(obj)->file = rb_sourcefile(); - RANY(obj)->line = rb_sourceline(); - assert(!SPECIAL_CONST_P(obj)); /* check alignment */ -#endif + case T_IMEMO: { + switch (imemo_type(obj)) { + case imemo_callcache: { + struct rb_callcache *cc = (struct rb_callcache *)obj; + if (cc->klass != Qundef && + (!rb_gc_handle_weak_references_alive_p(cc->klass) || + !rb_gc_handle_weak_references_alive_p((VALUE)cc->cme_))) { + vm_cc_invalidate(cc); + } + break; + } + case imemo_subclasses: { + struct rb_subclasses *subs = (struct rb_subclasses *)obj; + VALUE *entries = rb_imemo_subclasses_entries(obj); + for (uint32_t i = 0; i < subs->count; i++) { + if (entries[i] && !rb_gc_handle_weak_references_alive_p(entries[i])) { + entries[i] = 0; + } + } + break; + } + default: + rb_bug("rb_gc_handle_weak_references: unexpected imemo type"); + } -#if RGENGC_PROFILE - if (flags & FL_WB_PROTECTED) { - objspace->profile.generated_normal_object_count++; -#if RGENGC_PROFILE >= 2 - objspace->profile.generated_normal_object_count_types[BUILTIN_TYPE(obj)]++; -#endif - } - else { - objspace->profile.generated_shady_object_count++; -#if RGENGC_PROFILE >= 2 - objspace->profile.generated_shady_object_count_types[BUILTIN_TYPE(obj)]++; -#endif + break; + } + default: + rb_bug("rb_gc_handle_weak_references: type not supported\n"); } -#endif +} - rgengc_report(5, objspace, "newobj: %p (%s)\n", (void *)obj, obj_type_name(obj)); +static inline bool +rb_gc_imemo_needs_cleanup_p(VALUE obj) +{ + switch (imemo_type(obj)) { + case imemo_constcache: + case imemo_cref: + case imemo_ifunc: + case imemo_memo: + case imemo_svar: + case imemo_callcache: + case imemo_throw_data: + case imemo_cvar_entry: + return false; -#if USE_RGENGC && RGENGC_CHECK_MODE - if (RVALUE_PROMOTED_P(obj)) rb_bug("newobj: %p (%s) is promoted.\n", (void *)obj, obj_type_name(obj)); - if (rgengc_remembered(objspace, (VALUE)obj)) rb_bug("newobj: %p (%s) is remembered.\n", (void *)obj, obj_type_name(obj)); -#endif + case imemo_env: + case imemo_ment: + case imemo_iseq: + case imemo_callinfo: + case imemo_cdhash: + return true; - objspace->profile.total_allocated_object_num++; - gc_event_hook(objspace, RUBY_INTERNAL_EVENT_NEWOBJ, obj); + case imemo_subclasses: + return FL_TEST_RAW(obj, IMEMO_SUBCLASSES_HEAP); - return obj; -} + case imemo_tmpbuf: + return ((rb_imemo_tmpbuf_t *)obj)->ptr != NULL; -VALUE -rb_newobj(void) -{ - return newobj_of(0, T_NONE, 0, 0, 0); + case imemo_fields: + return FL_TEST_RAW(obj, OBJ_FIELD_HEAP) || (id2ref_tbl && rb_obj_shape_has_id(obj)); + } + UNREACHABLE_RETURN(true); } -VALUE -rb_newobj_of(VALUE klass, VALUE flags) +/* + * Returns true if the object requires a full rb_gc_obj_free() call during sweep, + * false if it can be freed quickly without calling destructors or cleanup. + * + * Objects that return false are: + * - Simple embedded objects without external allocations + * - Objects without finalizers + * - Objects without object IDs registered in id2ref + * - Objects without generic instance variables + * + * This is used by the GC sweep fast path to avoid function call overhead + * for the majority of simple objects. + */ +bool +rb_gc_obj_needs_cleanup_p(VALUE obj) { - return newobj_of(klass, flags, 0, 0, 0); -} + VALUE flags = RBASIC(obj)->flags; -NODE* -rb_node_newnode(enum node_type type, VALUE a0, VALUE a1, VALUE a2) -{ - VALUE flags = (RGENGC_WB_PROTECTED_NODE_CREF && type == NODE_CREF ? FL_WB_PROTECTED : 0); - NODE *n = (NODE *)newobj_of(0, T_NODE | flags, a0, a1, a2); - nd_set_type(n, type); - return n; -} + if (flags & FL_FINALIZE) return true; -VALUE -rb_data_object_alloc(VALUE klass, void *datap, RUBY_DATA_FUNC dmark, RUBY_DATA_FUNC dfree) -{ - if (klass) Check_Type(klass, T_CLASS); - return newobj_of(klass, T_DATA, (VALUE)dmark, (VALUE)dfree, (VALUE)datap); -} + switch (flags & RUBY_T_MASK) { + case T_IMEMO: + return rb_gc_imemo_needs_cleanup_p(obj); -VALUE -rb_data_typed_object_alloc(VALUE klass, void *datap, const rb_data_type_t *type) -{ - if (klass) Check_Type(klass, T_CLASS); - return newobj_of(klass, T_DATA | (type->flags & ~T_MASK), (VALUE)type, (VALUE)1, (VALUE)datap); -} + case T_DATA: + case T_OBJECT: + case T_STRING: + case T_ARRAY: + case T_HASH: + case T_BIGNUM: + case T_STRUCT: + case T_FLOAT: + case T_RATIONAL: + case T_COMPLEX: + case T_MATCH: + break; -size_t -rb_objspace_data_type_memsize(VALUE obj) -{ - if (RTYPEDDATA_P(obj) && RTYPEDDATA_TYPE(obj)->function.dsize) { - return RTYPEDDATA_TYPE(obj)->function.dsize(RTYPEDDATA_DATA(obj)); - } - else { - return 0; + case T_FILE: + case T_SYMBOL: + case T_CLASS: + case T_ICLASS: + case T_MODULE: + case T_REGEXP: + return true; } -} -const char * -rb_objspace_data_type_name(VALUE obj) -{ - if (RTYPEDDATA_P(obj)) { - return RTYPEDDATA_TYPE(obj)->wrap_struct_name; - } - else { - return 0; - } -} + shape_id_t shape_id = RBASIC_SHAPE_ID(obj); + if (id2ref_tbl && rb_shape_has_object_id(shape_id)) return true; -static inline int -is_pointer_to_heap(rb_objspace_t *objspace, void *ptr) -{ - register RVALUE *p = RANY(ptr); - register struct heap_page *page; - register size_t hi, lo, mid; + switch (flags & RUBY_T_MASK) { + case T_OBJECT: + if (flags & ROBJECT_HEAP) return true; + return false; + + case T_DATA: + { + uintptr_t type = (uintptr_t)RTYPEDDATA(obj)->type; + if (type & TYPED_DATA_EMBEDDED) { + RUBY_DATA_FUNC dfree = ((const rb_data_type_t *)(type & TYPED_DATA_PTR_MASK))->function.dfree; + if (dfree == RUBY_NEVER_FREE || dfree == RUBY_TYPED_DEFAULT_FREE) { + return false; + } + } + } + return true; - if (p < heap_pages_lomem || p > heap_pages_himem) return FALSE; - if ((VALUE)p % sizeof(RVALUE) != 0) return FALSE; + case T_STRING: + if (flags & (RSTRING_NOEMBED | RSTRING_FSTR)) return true; + return rb_shape_has_fields(shape_id); + + case T_ARRAY: + if (!(flags & RARRAY_EMBED_FLAG)) return true; + return rb_shape_has_fields(shape_id); + + case T_HASH: + if (flags & RHASH_ST_TABLE_FLAG) return true; + return rb_shape_has_fields(shape_id); + + case T_MATCH: + if ((flags & (RMATCH_ONIG | RMATCH_OFFSETS_EXTERNAL)) || USE_DEBUG_COUNTER) return true; + return rb_shape_has_fields(shape_id); + + case T_BIGNUM: + if (!(flags & BIGNUM_EMBED_FLAG)) return true; + return rb_shape_has_fields(shape_id); + + case T_STRUCT: + if (!(flags & RSTRUCT_EMBED_LEN_MASK)) return true; + if (flags & RSTRUCT_GEN_FIELDS) return rb_shape_has_fields(shape_id); + return false; - /* check if p looks like a pointer using bsearch*/ - lo = 0; - hi = heap_pages_used; - while (lo < hi) { - mid = (lo + hi) / 2; - page = heap_pages_sorted[mid]; - if (page->start <= p) { - if (p < page->start + page->limit) { - return TRUE; - } - lo = mid + 1; - } - else { - hi = mid; - } + case T_FLOAT: + case T_RATIONAL: + case T_COMPLEX: + return rb_shape_has_fields(shape_id); + + default: + UNREACHABLE_RETURN(true); } - return FALSE; } -static int -free_method_entry_i(ID key, rb_method_entry_t *me, st_data_t data) +static void +io_fptr_finalize(void *fptr) { - if (!me->mark) { - rb_free_method_entry(me); - } - return ST_CONTINUE; + rb_io_fptr_finalize((struct rb_io *)fptr); } -void -rb_free_m_tbl(st_table *tbl) +static inline void +make_io_zombie(void *objspace, VALUE obj) { - st_foreach(tbl, free_method_entry_i, 0); - st_free_table(tbl); + rb_io_t *fptr = RFILE(obj)->fptr; + rb_gc_impl_make_zombie(objspace, obj, io_fptr_finalize, fptr); } -void -rb_free_m_tbl_wrapper(struct method_table_wrapper *wrapper) +static bool +rb_data_free(void *objspace, VALUE obj) { - if (wrapper->tbl) { - rb_free_m_tbl(wrapper->tbl); + void *data = RTYPEDDATA_GET_DATA(obj); + if (data) { + int free_immediately = false; + void (*dfree)(void *); + + free_immediately = (RTYPEDDATA_TYPE(obj)->flags & RUBY_TYPED_FREE_IMMEDIATELY) != 0; + dfree = RTYPEDDATA_TYPE(obj)->function.dfree; + + if (dfree) { + if (dfree == RUBY_DEFAULT_FREE) { + if (!RTYPEDDATA_EMBEDDED_P(obj)) { + xfree(data); + RB_DEBUG_COUNTER_INC(obj_data_xfree); + } + } + else if (free_immediately) { + (*dfree)(data); + if (RTYPEDDATA_EMBEDDABLE_P(obj) && !RTYPEDDATA_EMBEDDED_P(obj)) { + xfree(data); + } + + RB_DEBUG_COUNTER_INC(obj_data_imm_free); + } + else { + rb_gc_impl_make_zombie(objspace, obj, dfree, data); + RB_DEBUG_COUNTER_INC(obj_data_zombie); + return FALSE; + } + } + else { + RB_DEBUG_COUNTER_INC(obj_data_empty); + } } - xfree(wrapper); -} -static int -free_const_entry_i(ID key, rb_const_entry_t *ce, st_data_t data) -{ - xfree(ce); - return ST_CONTINUE; + return true; } -void -rb_free_const_table(st_table *tbl) -{ - st_foreach(tbl, free_const_entry_i, 0); - st_free_table(tbl); -} +struct classext_foreach_args { + VALUE klass; + rb_objspace_t *objspace; // used for update_* +}; -static inline void -make_deferred(rb_objspace_t *objspace,RVALUE *p) +static void +classext_free(rb_classext_t *ext, bool is_prime, VALUE box_value, void *arg) { - p->as.basic.flags = T_ZOMBIE; - p->as.free.next = heap_pages_deferred_final; - heap_pages_deferred_final = p; + struct classext_foreach_args *args = (struct classext_foreach_args *)arg; + + rb_class_classext_free(args->klass, ext, is_prime); } -static inline void -make_io_deferred(rb_objspace_t *objspace,RVALUE *p) +static void +classext_iclass_free(rb_classext_t *ext, bool is_prime, VALUE box_value, void *arg) { - rb_io_t *fptr = p->as.file.fptr; - make_deferred(objspace, p); - p->as.data.dfree = (void (*)(void*))rb_io_fptr_finalize; - p->as.data.data = fptr; + struct classext_foreach_args *args = (struct classext_foreach_args *)arg; + + rb_iclass_classext_free(args->klass, ext, is_prime); } -static int -obj_free(rb_objspace_t *objspace, VALUE obj) +bool +rb_gc_obj_free(void *objspace, VALUE obj) { - gc_event_hook(objspace, RUBY_INTERNAL_EVENT_FREEOBJ, obj); + struct classext_foreach_args args; + + RB_DEBUG_COUNTER_INC(obj_free); switch (BUILTIN_TYPE(obj)) { case T_NIL: case T_FIXNUM: case T_TRUE: case T_FALSE: - rb_bug("obj_free() called for broken object"); - break; - } - - if (FL_TEST(obj, FL_EXIVAR)) { - rb_free_generic_ivar((VALUE)obj); - FL_UNSET(obj, FL_EXIVAR); + rb_bug("obj_free() called for broken object"); + break; + default: + break; } -#if USE_RGENGC - if (MARKED_IN_BITMAP(GET_HEAP_OLDGEN_BITS(obj),obj)) - CLEAR_IN_BITMAP(GET_HEAP_OLDGEN_BITS(obj),obj); -#endif - switch (BUILTIN_TYPE(obj)) { case T_OBJECT: - if (!(RANY(obj)->as.basic.flags & ROBJECT_EMBED) && - RANY(obj)->as.object.as.heap.ivptr) { - xfree(RANY(obj)->as.object.as.heap.ivptr); - } - break; + if (FL_TEST_RAW(obj, ROBJECT_HEAP)) { + if (rb_obj_shape_complex_p(obj)) { + RB_DEBUG_COUNTER_INC(obj_obj_complex); + st_free_table(ROBJECT_FIELDS_HASH(obj)); + } + else { + SIZED_FREE_N(ROBJECT(obj)->as.heap.fields, ROBJECT_FIELDS_CAPACITY(obj)); + RB_DEBUG_COUNTER_INC(obj_obj_ptr); + } + } + else { + RB_DEBUG_COUNTER_INC(obj_obj_embed); + } + break; case T_MODULE: case T_CLASS: - if (RCLASS_M_TBL_WRAPPER(obj)) { - rb_free_m_tbl_wrapper(RCLASS_M_TBL_WRAPPER(obj)); +#if USE_ZJIT + rb_zjit_klass_free(obj); +#endif + args.klass = obj; + rb_class_classext_foreach(obj, classext_free, (void *)&args); + if (RCLASS_CLASSEXT_TBL(obj)) { + st_free_table(RCLASS_CLASSEXT_TBL(obj)); } - if (RCLASS_IV_TBL(obj)) { - st_free_table(RCLASS_IV_TBL(obj)); - } - if (RCLASS_CONST_TBL(obj)) { - rb_free_const_table(RCLASS_CONST_TBL(obj)); - } - if (RCLASS_IV_INDEX_TBL(obj)) { - st_free_table(RCLASS_IV_INDEX_TBL(obj)); - } - if (RCLASS_EXT(obj)->subclasses) { - if (BUILTIN_TYPE(obj) == T_MODULE) { - rb_class_detach_module_subclasses(obj); - } - else { - rb_class_detach_subclasses(obj); - } - RCLASS_EXT(obj)->subclasses = NULL; - } - rb_class_remove_from_module_subclasses(obj); - rb_class_remove_from_super_subclasses(obj); - if (RANY(obj)->as.klass.ptr) - xfree(RANY(obj)->as.klass.ptr); - RANY(obj)->as.klass.ptr = NULL; - break; + (void)RB_DEBUG_COUNTER_INC_IF(obj_module_ptr, BUILTIN_TYPE(obj) == T_MODULE); + (void)RB_DEBUG_COUNTER_INC_IF(obj_class_ptr, BUILTIN_TYPE(obj) == T_CLASS); + break; case T_STRING: - rb_str_free(obj); - break; + rb_str_free(obj); + break; case T_ARRAY: - rb_ary_free(obj); - break; + rb_ary_free(obj); + break; case T_HASH: - if (RANY(obj)->as.hash.ntbl) { - st_free_table(RANY(obj)->as.hash.ntbl); - } - break; +#if USE_DEBUG_COUNTER + switch (RHASH_SIZE(obj)) { + case 0: + RB_DEBUG_COUNTER_INC(obj_hash_empty); + break; + case 1: + RB_DEBUG_COUNTER_INC(obj_hash_1); + break; + case 2: + RB_DEBUG_COUNTER_INC(obj_hash_2); + break; + case 3: + RB_DEBUG_COUNTER_INC(obj_hash_3); + break; + case 4: + RB_DEBUG_COUNTER_INC(obj_hash_4); + break; + case 5: + case 6: + case 7: + case 8: + RB_DEBUG_COUNTER_INC(obj_hash_5_8); + break; + default: + GC_ASSERT(RHASH_SIZE(obj) > 8); + RB_DEBUG_COUNTER_INC(obj_hash_g8); + } + + if (RHASH_AR_TABLE_P(obj)) { + if (RHASH_AR_TABLE(obj) == NULL) { + RB_DEBUG_COUNTER_INC(obj_hash_null); + } + else { + RB_DEBUG_COUNTER_INC(obj_hash_ar); + } + } + else { + RB_DEBUG_COUNTER_INC(obj_hash_st); + } +#endif + + rb_hash_free(obj); + break; case T_REGEXP: - if (RANY(obj)->as.regexp.ptr) { - onig_free(RANY(obj)->as.regexp.ptr); - } - break; + if (RREGEXP(obj)->ptr) { + onig_free(RREGEXP(obj)->ptr); + RB_DEBUG_COUNTER_INC(obj_regexp_ptr); + } + break; case T_DATA: - if (DATA_PTR(obj)) { - int free_immediately = FALSE; - - if (RTYPEDDATA_P(obj)) { - free_immediately = (RANY(obj)->as.typeddata.type->flags & RUBY_TYPED_FREE_IMMEDIATELY) != 0; - RDATA(obj)->dfree = RANY(obj)->as.typeddata.type->function.dfree; - if (0 && free_immediately == 0) /* to expose non-free-immediate T_DATA */ - fprintf(stderr, "not immediate -> %s\n", RANY(obj)->as.typeddata.type->wrap_struct_name); - } - if (RANY(obj)->as.data.dfree == RUBY_DEFAULT_FREE) { - xfree(DATA_PTR(obj)); - } - else if (RANY(obj)->as.data.dfree) { - if (free_immediately) { - (RDATA(obj)->dfree)(DATA_PTR(obj)); - } - else { - make_deferred(objspace, RANY(obj)); - return 1; - } - } - } - break; + if (!rb_data_free(objspace, obj)) return false; + break; case T_MATCH: - if (RANY(obj)->as.match.rmatch) { - struct rmatch *rm = RANY(obj)->as.match.rmatch; - onig_region_free(&rm->regs, 0); - if (rm->char_offset) - xfree(rm->char_offset); - xfree(rm); - } - break; + { + struct RMatch *rm = RMATCH(obj); +#if USE_DEBUG_COUNTER + if (rm->num_regs >= 8) { + RB_DEBUG_COUNTER_INC(obj_match_ge8); + } + else if (rm->num_regs >= 4) { + RB_DEBUG_COUNTER_INC(obj_match_ge4); + } + else if (rm->num_regs >= 1) { + RB_DEBUG_COUNTER_INC(obj_match_under4); + } +#endif + if (FL_TEST_RAW(obj, RMATCH_ONIG)) { + onig_region_free(&rm->as.onig, 0); + } + SIZED_FREE_N(rm->char_offset, rm->char_offset_num_allocated); + + RB_DEBUG_COUNTER_INC(obj_match_ptr); + } + break; case T_FILE: - if (RANY(obj)->as.file.fptr) { - make_io_deferred(objspace, RANY(obj)); - return 1; - } - break; + if (RFILE(obj)->fptr) { + bool closed = rb_io_fptr_finalize_closed(RFILE(obj)->fptr); + if (!closed) make_io_zombie(objspace, obj); + RB_DEBUG_COUNTER_INC(obj_file_ptr); + return closed; + } + break; case T_RATIONAL: + RB_DEBUG_COUNTER_INC(obj_rational); + break; case T_COMPLEX: - break; + RB_DEBUG_COUNTER_INC(obj_complex); + break; + case T_MOVED: + break; case T_ICLASS: - /* iClass shares table with the module */ - if (RCLASS_EXT(obj)->subclasses) { - rb_class_detach_subclasses(obj); - RCLASS_EXT(obj)->subclasses = NULL; - } - rb_class_remove_from_module_subclasses(obj); - rb_class_remove_from_super_subclasses(obj); - xfree(RANY(obj)->as.klass.ptr); - RANY(obj)->as.klass.ptr = NULL; - break; + args.klass = obj; + + rb_class_classext_foreach(obj, classext_iclass_free, (void *)&args); + if (RCLASS_CLASSEXT_TBL(obj)) { + st_free_table(RCLASS_CLASSEXT_TBL(obj)); + } + + RB_DEBUG_COUNTER_INC(obj_iclass_ptr); + break; case T_FLOAT: - break; + RB_DEBUG_COUNTER_INC(obj_float); + break; case T_BIGNUM: - if (!(RBASIC(obj)->flags & RBIGNUM_EMBED_FLAG) && RBIGNUM_DIGITS(obj)) { - xfree(RBIGNUM_DIGITS(obj)); - } - break; + if (!BIGNUM_EMBED_P(obj) && BIGNUM_DIGITS(obj)) { + SIZED_FREE_N(BIGNUM_DIGITS(obj), BIGNUM_LEN(obj)); + RB_DEBUG_COUNTER_INC(obj_bignum_ptr); + } + else { + RB_DEBUG_COUNTER_INC(obj_bignum_embed); + } + break; + case T_NODE: - switch (nd_type(obj)) { - case NODE_SCOPE: - if (RANY(obj)->as.node.u1.tbl) { - xfree(RANY(obj)->as.node.u1.tbl); - } - break; - case NODE_ARGS: - if (RANY(obj)->as.node.u3.args) { - xfree(RANY(obj)->as.node.u3.args); - } - break; - case NODE_ALLOCA: - xfree(RANY(obj)->as.node.u1.node); - break; - } - break; /* no need to free iv_tbl */ + UNEXPECTED_NODE(obj_free); + break; case T_STRUCT: - if ((RBASIC(obj)->flags & RSTRUCT_EMBED_LEN_MASK) == 0 && - RANY(obj)->as.rstruct.as.heap.ptr) { - xfree((void *)RANY(obj)->as.rstruct.as.heap.ptr); - } - break; - - default: - rb_bug("gc_sweep(): unknown data type 0x%x(%p) 0x%"PRIxVALUE, - BUILTIN_TYPE(obj), (void*)obj, RBASIC(obj)->flags); - } - - return 0; -} - -void -Init_heap(void) -{ - rb_objspace_t *objspace = &rb_objspace; - -#if RGENGC_ESTIMATE_OLDMALLOC - objspace->rgengc.oldmalloc_increase_limit = gc_params.oldmalloc_limit_min; -#endif + if ((RBASIC(obj)->flags & RSTRUCT_EMBED_LEN_MASK) || + RSTRUCT(obj)->as.heap.ptr == NULL) { + RB_DEBUG_COUNTER_INC(obj_struct_embed); + } + else { + SIZED_FREE_N(RSTRUCT(obj)->as.heap.ptr, RSTRUCT(obj)->as.heap.len); + RB_DEBUG_COUNTER_INC(obj_struct_ptr); + } + break; - heap_add_pages(objspace, heap_eden, gc_params.heap_init_slots / HEAP_OBJ_LIMIT); + case T_SYMBOL: + RB_DEBUG_COUNTER_INC(obj_symbol); + break; - init_mark_stack(&objspace->mark_stack); + case T_IMEMO: + rb_imemo_free((VALUE)obj); + break; -#ifdef USE_SIGALTSTACK - { - /* altstack of another threads are allocated in another place */ - rb_thread_t *th = GET_THREAD(); - void *tmp = th->altstack; - th->altstack = malloc(rb_sigaltstack_size()); - free(tmp); /* free previously allocated area */ + default: + rb_bug("gc_sweep(): unknown data type 0x%x(%p) 0x%"PRIxVALUE, + BUILTIN_TYPE(obj), (void*)obj, RBASIC(obj)->flags); } -#endif - - objspace->profile.invoke_time = getrusage_time(); - finalizer_table = st_init_numtable(); -} - -typedef int each_obj_callback(void *, void *, size_t, void *); - -struct each_obj_args { - each_obj_callback *callback; - void *data; -}; -static VALUE -objspace_each_objects(VALUE arg) -{ - size_t i; - struct heap_page_body *last_body = 0; - struct heap_page *page; - RVALUE *pstart, *pend; - rb_objspace_t *objspace = &rb_objspace; - struct each_obj_args *args = (struct each_obj_args *)arg; - - i = 0; - while (i < heap_pages_used) { - while (0 < i && last_body < heap_pages_sorted[i-1]->body) i--; - while (i < heap_pages_used && heap_pages_sorted[i]->body <= last_body) i++; - if (heap_pages_used <= i) break; - - page = heap_pages_sorted[i]; - last_body = page->body; - - pstart = page->start; - pend = pstart + page->limit; - - if ((*args->callback)(pstart, pend, sizeof(RVALUE), args->data)) { - break; - } + if (FL_TEST_RAW(obj, FL_FINALIZE)) { + rb_gc_impl_make_zombie(objspace, obj, 0, 0); + return FALSE; + } + else { + return TRUE; } - - return Qnil; } -/* - * rb_objspace_each_objects() is special C API to walk through - * Ruby object space. This C API is too difficult to use it. - * To be frank, you should not use it. Or you need to read the - * source code of this function and understand what this function does. - * - * 'callback' will be called several times (the number of heap page, - * at current implementation) with: - * vstart: a pointer to the first living object of the heap_page. - * vend: a pointer to next to the valid heap_page area. - * stride: a distance to next VALUE. - * - * If callback() returns non-zero, the iteration will be stopped. - * - * This is a sample callback code to iterate liveness objects: - * - * int - * sample_callback(void *vstart, void *vend, int stride, void *data) { - * VALUE v = (VALUE)vstart; - * for (; v != (VALUE)vend; v += stride) { - * if (RBASIC(v)->flags) { // liveness check - * // do something with live object 'v' - * } - * return 0; // continue to iteration - * } - * - * Note: 'vstart' is not a top of heap_page. This point the first - * living object to grasp at least one object to avoid GC issue. - * This means that you can not walk through all Ruby object page - * including freed object page. - * - * Note: On this implementation, 'stride' is same as sizeof(RVALUE). - * However, there are possibilities to pass variable values with - * 'stride' with some reasons. You must use stride instead of - * use some constant value in the iteration. - */ void -rb_objspace_each_objects(each_obj_callback *callback, void *data) +rb_objspace_set_event_hook(const rb_event_flag_t event) { - struct each_obj_args args; - rb_objspace_t *objspace = &rb_objspace; - int prev_dont_lazy_sweep = objspace->flags.dont_lazy_sweep; - - gc_rest_sweep(objspace); - objspace->flags.dont_lazy_sweep = TRUE; - - args.callback = callback; - args.data = data; - - if (prev_dont_lazy_sweep) { - objspace_each_objects((VALUE)&args); - } - else { - rb_ensure(objspace_each_objects, (VALUE)&args, lazy_sweep_enable, Qnil); - } + rb_gc_impl_set_event_hook(rb_gc_get_objspace(), event); } -struct os_each_struct { - size_t num; - VALUE of; -}; - static int internal_object_p(VALUE obj) { - RVALUE *p = (RVALUE *)obj; - - if (p->as.basic.flags) { - switch (BUILTIN_TYPE(p)) { - case T_NONE: - case T_ICLASS: - case T_NODE: - case T_ZOMBIE: - break; - case T_CLASS: - if (FL_TEST(p, FL_SINGLETON)) - break; - default: - if (!p->as.basic.klass) break; - return 0; - } + void *ptr = asan_unpoison_object_temporary(obj); + + if (RBASIC(obj)->flags) { + switch (BUILTIN_TYPE(obj)) { + case T_NODE: + UNEXPECTED_NODE(internal_object_p); + break; + case T_NONE: + case T_MOVED: + case T_IMEMO: + case T_ICLASS: + case T_ZOMBIE: + break; + case T_CLASS: + if (obj == rb_mRubyVMFrozenCore) + return 1; + + if (!RBASIC_CLASS(obj)) break; + if (RCLASS_SINGLETON_P(obj)) { + return rb_singleton_class_internal_p(obj); + } + return 0; + default: + if (!RBASIC(obj)->klass) break; + return 0; + } + } + if (ptr || !RBASIC(obj)->flags) { + rb_asan_poison_object(obj); } return 1; } @@ -1812,20 +1768,26 @@ rb_objspace_internal_object_p(VALUE obj) return internal_object_p(obj); } +struct os_each_struct { + size_t num; + VALUE of; +}; + static int os_obj_of_i(void *vstart, void *vend, size_t stride, void *data) { struct os_each_struct *oes = (struct os_each_struct *)data; - RVALUE *p = (RVALUE *)vstart, *pend = (RVALUE *)vend; - for (; p != pend; p++) { - volatile VALUE v = (VALUE)p; - if (!internal_object_p(v)) { - if (!oes->of || rb_obj_is_kind_of(v, oes->of)) { - rb_yield(v); - oes->num++; - } - } + VALUE v = (VALUE)vstart; + for (; v != (VALUE)vend; v += stride) { + if (!internal_object_p(v)) { + if (!oes->of || rb_obj_is_kind_of(v, oes->of)) { + if (!rb_multi_ractor_p() || rb_ractor_shareable_p(v)) { + rb_yield(v); + oes->num++; + } + } + } } return 0; @@ -1844,37 +1806,41 @@ os_obj_of(VALUE of) /* * call-seq: - * ObjectSpace.each_object([module]) {|obj| ... } -> fixnum + * ObjectSpace.each_object([module]) {|obj| ... } -> integer * ObjectSpace.each_object([module]) -> an_enumerator * * Calls the block once for each living, nonimmediate object in this * Ruby process. If <i>module</i> is specified, calls the block * for only those classes or modules that match (or are a subclass of) * <i>module</i>. Returns the number of objects found. Immediate - * objects (<code>Fixnum</code>s, <code>Symbol</code>s - * <code>true</code>, <code>false</code>, and <code>nil</code>) are - * never returned. In the example below, <code>each_object</code> - * returns both the numbers we defined and several constants defined in - * the <code>Math</code> module. + * objects (such as <code>Fixnum</code>s, static <code>Symbol</code>s + * <code>true</code>, <code>false</code> and <code>nil</code>) are + * never returned. * * If no block is given, an enumerator is returned instead. * - * a = 102.7 - * b = 95 # Won't be returned - * c = 12345678987654321 - * count = ObjectSpace.each_object(Numeric) {|x| p x } + * Job = Class.new + * jobs = [Job.new, Job.new] + * count = ObjectSpace.each_object(Job) {|x| p x } * puts "Total count: #{count}" * * <em>produces:</em> * - * 12345678987654321 - * 102.7 - * 2.71828182845905 - * 3.14159265358979 - * 2.22044604925031e-16 - * 1.7976931348623157e+308 - * 2.2250738585072e-308 - * Total count: 7 + * #<Job:0x000000011d6cbbf0> + * #<Job:0x000000011d6cbc68> + * Total count: 2 + * + * Due to a current Ractor implementation issue, this method does not yield + * Ractor-unshareable objects when the process is in multi-Ractor mode. Multi-ractor + * mode is enabled when <code>Ractor.new</code> has been called for the first time. + * See https://bugs.ruby-lang.org/issues/19387 for more information. + * + * a = 12345678987654321 # shareable + * b = [].freeze # shareable + * c = {} # not shareable + * ObjectSpace.each_object {|x| x } # yields a, b, and c + * Ractor.new {} # enter multi-Ractor mode + * ObjectSpace.each_object {|x| x } # does not yield c * */ @@ -1883,22 +1849,19 @@ os_each_obj(int argc, VALUE *argv, VALUE os) { VALUE of; - if (argc == 0) { - of = 0; - } - else { - rb_scan_args(argc, argv, "01", &of); - } + of = (!rb_check_arity(argc, 0, 1) ? 0 : argv[0]); RETURN_ENUMERATOR(os, 1, &of); return os_obj_of(of); } /* * call-seq: - * ObjectSpace.undefine_finalizer(obj) + * ObjectSpace.undefine_finalizer(obj) -> obj * - * Removes all finalizers for <i>obj</i>. + * Removes all finalizers registered for +obj+ with + * ObjectSpace.define_finalizer, and returns +obj+. * + * Does nothing if +obj+ has no finalizers. */ static VALUE @@ -1910,39 +1873,103 @@ undefine_final(VALUE os, VALUE obj) VALUE rb_undefine_finalizer(VALUE obj) { - rb_objspace_t *objspace = &rb_objspace; - st_data_t data = obj; rb_check_frozen(obj); - st_delete(finalizer_table, &data, 0); - FL_UNSET(obj, FL_FINALIZE); + + rb_gc_impl_undefine_finalizer(rb_gc_get_objspace(), obj); + return obj; } static void should_be_callable(VALUE block) { - if (!rb_obj_respond_to(block, rb_intern("call"), TRUE)) { - rb_raise(rb_eArgError, "wrong type argument %s (should be callable)", - rb_obj_classname(block)); + if (!rb_obj_respond_to(block, idCall, TRUE)) { + rb_raise(rb_eArgError, "wrong type argument %"PRIsVALUE" (should be callable)", + rb_obj_class(block)); } } + static void should_be_finalizable(VALUE obj) { - rb_check_frozen(obj); if (!FL_ABLE(obj)) { - rb_raise(rb_eArgError, "cannot define finalizer for %s", - rb_obj_classname(obj)); + rb_raise(rb_eArgError, "cannot define finalizer for %s", + rb_obj_classname(obj)); } + rb_check_frozen(obj); +} + +void +rb_gc_copy_finalizer(VALUE dest, VALUE obj) +{ + rb_gc_impl_copy_finalizer(rb_gc_get_objspace(), dest, obj); } /* * call-seq: - * ObjectSpace.define_finalizer(obj, aProc=proc()) + * ObjectSpace.define_finalizer(obj) {|id| ... } -> array + * ObjectSpace.define_finalizer(obj, finalizer) -> array * - * Adds <i>aProc</i> as a finalizer, to be called after <i>obj</i> - * was destroyed. + * Adds a new finalizer for +obj+ that is called when +obj+ is destroyed + * by the garbage collector or when Ruby shuts down (which ever comes first). * + * With a block given, uses the block as the callback. Without a block given, + * uses a callable object +finalizer+ as the callback. The callback is called + * when +obj+ is destroyed with a single argument +id+ which is the object + * ID of +obj+ (see Object#object_id). + * + * The return value is an array <code>[0, callback]</code>, where +callback+ + * is a Proc created from the block if one was given or +finalizer+ otherwise. + * + * Note that defining a finalizer in an instance method of the object may prevent + * the object from being garbage collected since if the block or +finalizer+ refers + * to +obj+ then +obj+ will never be reclaimed by the garbage collector. For example, + * the following script demonstrates the issue: + * + * class Foo + * def define_final + * ObjectSpace.define_finalizer(self) do |id| + * puts "Running finalizer for #{id}!" + * end + * end + * end + * + * obj = Foo.new + * obj.define_final + * + * There are two patterns to solve this issue: + * + * - Create the finalizer in a non-instance method so it can safely capture + * the needed state: + * + * class Foo + * def define_final + * ObjectSpace.define_finalizer(self, self.class.create_finalizer) + * end + * + * def self.create_finalizer + * proc do |id| + * puts "Running finalizer for #{id}!" + * end + * end + * end + * + * - Use a callable object: + * + * class Foo + * class Finalizer + * def call(id) + * puts "Running finalizer for #{id}!" + * end + * end + * + * def define_final + * ObjectSpace.define_finalizer(self, Finalizer.new) + * end + * end + * + * Note that finalization can be unpredictable and is never guaranteed + * to be run except on exit. */ static VALUE @@ -1951,337 +1978,394 @@ define_final(int argc, VALUE *argv, VALUE os) VALUE obj, block; rb_scan_args(argc, argv, "11", &obj, &block); - should_be_finalizable(obj); if (argc == 1) { - block = rb_block_proc(); + block = rb_block_proc(); } - else { - should_be_callable(block); + + if (rb_callable_receiver(block) == obj) { + rb_warn("finalizer references object to be finalized"); } - return define_final0(obj, block); + return rb_define_finalizer(obj, block); } -static VALUE -define_final0(VALUE obj, VALUE block) +VALUE +rb_define_finalizer(VALUE obj, VALUE block) { - rb_objspace_t *objspace = &rb_objspace; - VALUE table; - st_data_t data; + should_be_finalizable(obj); + should_be_callable(block); - RBASIC(obj)->flags |= FL_FINALIZE; + block = rb_gc_impl_define_finalizer(rb_gc_get_objspace(), obj, block); - block = rb_ary_new3(2, INT2FIX(rb_safe_level()), block); + block = rb_ary_new3(2, INT2FIX(0), block); OBJ_FREEZE(block); - - if (st_lookup(finalizer_table, obj, &data)) { - table = (VALUE)data; - rb_ary_push(table, block); - } - else { - table = rb_ary_new3(1, block); - RBASIC_CLEAR_CLASS(table); - st_add_direct(finalizer_table, obj, table); - } return block; } -VALUE -rb_define_finalizer(VALUE obj, VALUE block) +void +rb_objspace_call_finalizer(void) { - should_be_finalizable(obj); - should_be_callable(block); - return define_final0(obj, block); + rb_gc_impl_shutdown_call_finalizer(rb_gc_get_objspace()); } void -rb_gc_copy_finalizer(VALUE dest, VALUE obj) +rb_objspace_free_objects(void *objspace) { - rb_objspace_t *objspace = &rb_objspace; - VALUE table; - st_data_t data; + rb_gc_impl_shutdown_free_objects(objspace); +} - if (!FL_TEST(obj, FL_FINALIZE)) return; - if (st_lookup(finalizer_table, obj, &data)) { - table = (VALUE)data; - st_insert(finalizer_table, dest, table); - } - FL_SET(dest, FL_FINALIZE); +int +rb_objspace_garbage_object_p(VALUE obj) +{ + return !SPECIAL_CONST_P(obj) && rb_gc_impl_garbage_object_p(rb_gc_get_objspace(), obj); } -static VALUE -run_single_final(VALUE arg) +bool +rb_gc_pointer_to_heap_p(VALUE obj) { - VALUE *args = (VALUE *)arg; - rb_eval_cmd(args[0], args[1], (int)args[2]); - return Qnil; + return rb_gc_impl_pointer_to_heap_p(rb_gc_get_objspace(), (void *)obj); } -static void -run_finalizer(rb_objspace_t *objspace, VALUE obj, VALUE table) +#define OBJ_ID_INCREMENT (RUBY_IMMEDIATE_MASK + 1) +#define LAST_OBJECT_ID() (object_id_counter * OBJ_ID_INCREMENT) +static VALUE id2ref_value = 0; + +#if SIZEOF_SIZE_T == SIZEOF_LONG_LONG +static size_t object_id_counter = 1; +#else +static unsigned long long object_id_counter = 1; +#endif + +static inline VALUE +generate_next_object_id(void) { - long i; - int status; - VALUE args[3]; - VALUE objid = nonspecial_obj_id(obj); +#if SIZEOF_SIZE_T == SIZEOF_LONG_LONG + // 64bit atomics are available + return SIZET2NUM(RUBY_ATOMIC_SIZE_FETCH_ADD(object_id_counter, 1) * OBJ_ID_INCREMENT); +#else + unsigned int lock_lev = RB_GC_VM_LOCK(); + VALUE id = ULL2NUM(++object_id_counter * OBJ_ID_INCREMENT); + RB_GC_VM_UNLOCK(lock_lev); + return id; +#endif +} - if (RARRAY_LEN(table) > 0) { - args[1] = rb_obj_freeze(rb_ary_new3(1, objid)); +void +rb_gc_obj_id_moved(VALUE obj) +{ + if (UNLIKELY(id2ref_tbl)) { + st_insert(id2ref_tbl, (st_data_t)rb_obj_id(obj), (st_data_t)obj); + } +} + +static int +object_id_cmp(st_data_t x, st_data_t y) +{ + if (RB_TYPE_P(x, T_BIGNUM)) { + return !rb_big_eql(x, y); } else { - args[1] = 0; + return x != y; } +} + +static st_index_t +object_id_hash(st_data_t n) +{ + return FIX2LONG(rb_hash((VALUE)n)); +} + +static const struct st_hash_type object_id_hash_type = { + object_id_cmp, + object_id_hash, +}; - args[2] = (VALUE)rb_safe_level(); - for (i=0; i<RARRAY_LEN(table); i++) { - VALUE final = RARRAY_AREF(table, i); - args[0] = RARRAY_AREF(final, 1); - args[2] = FIX2INT(RARRAY_AREF(final, 0)); - status = 0; - rb_protect(run_single_final, (VALUE)args, &status); - if (status) - rb_set_errinfo(Qnil); +static void gc_mark_tbl_no_pin(st_table *table); + +static void +id2ref_tbl_mark(void *data) +{ + st_table *table = (st_table *)data; + if (UNLIKELY(!RB_POSFIXABLE(LAST_OBJECT_ID()))) { + // It's very unlikely, but if enough object ids were generated, keys may be T_BIGNUM + rb_mark_set(table); } + // We purposely don't mark values, as they are weak references. + // rb_gc_obj_free_vm_weak_references takes care of cleaning them up. +} + +static size_t +id2ref_tbl_memsize(const void *data) +{ + return rb_st_memsize(data); } static void -run_final(rb_objspace_t *objspace, VALUE obj) +id2ref_tbl_free(void *data) { - RUBY_DATA_FUNC free_func = 0; - st_data_t key, table; + id2ref_tbl = NULL; // clear global ref + st_table *table = (st_table *)data; + st_free_table(table); +} - heap_pages_final_slots--; +static const rb_data_type_t id2ref_tbl_type = { + .wrap_struct_name = "VM/_id2ref_table", + .function = { + .dmark = id2ref_tbl_mark, + .dfree = id2ref_tbl_free, + .dsize = id2ref_tbl_memsize, + // dcompact function not required because the table is reference updated + // in rb_gc_vm_weak_table_foreach + }, + .flags = RUBY_TYPED_WB_PROTECTED | RUBY_TYPED_FREE_IMMEDIATELY +}; - RBASIC_CLEAR_CLASS(obj); +static VALUE +class_object_id(VALUE klass) +{ + VALUE id = RUBY_ATOMIC_VALUE_LOAD(RCLASS(klass)->object_id); + if (!id) { + unsigned int lock_lev = RB_GC_VM_LOCK(); + id = generate_next_object_id(); + VALUE existing_id = RUBY_ATOMIC_VALUE_CAS(RCLASS(klass)->object_id, 0, id); + if (existing_id) { + id = existing_id; + } + else if (RB_UNLIKELY(id2ref_tbl)) { + st_insert(id2ref_tbl, id, klass); + } + RB_GC_VM_UNLOCK(lock_lev); + } + return id; +} - if (RTYPEDDATA_P(obj)) { - free_func = RTYPEDDATA_TYPE(obj)->function.dfree; +static inline VALUE +object_id_get(VALUE obj, shape_id_t shape_id) +{ + VALUE id; + if (rb_shape_complex_p(shape_id)) { + id = rb_obj_field_get(obj, ROOT_COMPLEX_WITH_OBJ_ID); } else { - free_func = RDATA(obj)->dfree; - } - if (free_func) { - (*free_func)(DATA_PTR(obj)); + id = rb_obj_field_get(obj, rb_shape_object_id(shape_id)); } - key = (st_data_t)obj; - if (st_delete(finalizer_table, &key, &table)) { - run_finalizer(objspace, obj, (VALUE)table); +#if RUBY_DEBUG + if (!(FIXNUM_P(id) || RB_TYPE_P(id, T_BIGNUM))) { + rb_p(obj); + rb_bug("Object's shape includes object_id, but it's missing %s", rb_obj_info(obj)); } +#endif + + return id; } -static void -finalize_list(rb_objspace_t *objspace, RVALUE *p) +static VALUE +object_id0(VALUE obj) { - while (p) { - RVALUE *tmp = p->as.free.next; - struct heap_page *page = GET_HEAP_PAGE(p); + VALUE id = Qfalse; + shape_id_t shape_id = RBASIC_SHAPE_ID(obj); + + if (rb_shape_has_object_id(shape_id)) { + return object_id_get(obj, shape_id); + } - run_final(objspace, (VALUE)p); - objspace->profile.total_freed_object_num++; + shape_id_t object_id_shape_id = rb_obj_shape_transition_object_id(obj); - page->final_slots--; - heap_page_add_freeobj(objspace, GET_HEAP_PAGE(p), (VALUE)p); - heap_pages_swept_slots++; + id = generate_next_object_id(); + rb_obj_field_set(obj, object_id_shape_id, 0, id); - p = tmp; + RUBY_ASSERT(RBASIC_SHAPE_ID(obj) == object_id_shape_id); + RUBY_ASSERT(rb_obj_shape_has_id(obj)); + + if (RB_UNLIKELY(id2ref_tbl)) { + RB_VM_LOCKING() { + st_insert(id2ref_tbl, (st_data_t)id, (st_data_t)obj); + } } + return id; } -static void -finalize_deferred(rb_objspace_t *objspace) +static VALUE +object_id(VALUE obj) { - RVALUE *p; + switch (BUILTIN_TYPE(obj)) { + case T_CLASS: + case T_MODULE: + // With Ruby Box, classes and modules have different fields + // in different boxes, so we cannot store the object id + // in fields. + return class_object_id(obj); + case T_IMEMO: + RUBY_ASSERT(IMEMO_TYPE_P(obj, imemo_fields)); + break; + default: + break; + } - while ((p = ATOMIC_PTR_EXCHANGE(heap_pages_deferred_final, 0)) != 0) { - finalize_list(objspace, p); + if (UNLIKELY(rb_gc_multi_ractor_p() && rb_ractor_shareable_p(obj))) { + unsigned int lock_lev = RB_GC_VM_LOCK(); + VALUE id = object_id0(obj); + RB_GC_VM_UNLOCK(lock_lev); + return id; } + + return object_id0(obj); } static void -gc_finalize_deferred(void *dmy) +build_id2ref_i(VALUE obj, void *data) { - rb_objspace_t *objspace = &rb_objspace; - if (ATOMIC_EXCHANGE(finalizing, 1)) return; - finalize_deferred(objspace); - ATOMIC_SET(finalizing, 0); -} + st_table *id2ref_tbl = (st_table *)data; -/* TODO: to keep compatibility, maybe unused. */ -void -rb_gc_finalize_deferred(void) -{ - gc_finalize_deferred(0); + switch (BUILTIN_TYPE(obj)) { + case T_CLASS: + case T_MODULE: + RUBY_ASSERT(!rb_objspace_garbage_object_p(obj)); + if (RCLASS(obj)->object_id) { + st_insert(id2ref_tbl, RCLASS(obj)->object_id, obj); + } + break; + case T_IMEMO: + RUBY_ASSERT(!rb_objspace_garbage_object_p(obj)); + if (IMEMO_TYPE_P(obj, imemo_fields) && rb_obj_shape_has_id(obj)) { + st_insert(id2ref_tbl, rb_obj_id(obj), rb_imemo_fields_owner(obj)); + } + break; + case T_OBJECT: + RUBY_ASSERT(!rb_objspace_garbage_object_p(obj)); + if (rb_obj_shape_has_id(obj)) { + st_insert(id2ref_tbl, rb_obj_id(obj), obj); + } + break; + default: + // For generic_fields, the T_IMEMO/fields is responsible for populating the entry. + break; + } } -static void -gc_finalize_deferred_register(void) +static VALUE +object_id_to_ref(void *objspace_ptr, VALUE object_id) { - if (rb_postponed_job_register_one(0, gc_finalize_deferred, 0) == 0) { - rb_bug("gc_finalize_deferred_register: can't register finalizer."); + rb_objspace_t *objspace = objspace_ptr; + + unsigned int lev = RB_GC_VM_LOCK(); + + if (!id2ref_tbl) { + rb_gc_vm_barrier(); // stop other ractors + + // GC Must not trigger while we build the table, otherwise if we end + // up freeing an object that had an ID, we might try to delete it from + // the table even though it wasn't inserted yet. + st_table *tmp_id2ref_tbl = st_init_table(&object_id_hash_type); + VALUE tmp_id2ref_value = TypedData_Wrap_Struct(0, &id2ref_tbl_type, tmp_id2ref_tbl); + + // build_id2ref_i will most certainly malloc, which could trigger GC and sweep + // objects we just added to the table. + // By calling rb_gc_disable() we also save having to handle potentially garbage objects. + bool gc_disabled = RTEST(rb_gc_disable()); + { + id2ref_tbl = tmp_id2ref_tbl; + id2ref_value = tmp_id2ref_value; + + rb_gc_impl_each_object(objspace, build_id2ref_i, (void *)id2ref_tbl); + } + if (!gc_disabled) rb_gc_enable(); } -} -struct force_finalize_list { VALUE obj; - VALUE table; - struct force_finalize_list *next; -}; + bool found = st_lookup(id2ref_tbl, object_id, &obj) && !rb_gc_impl_garbage_object_p(objspace, obj); -static int -force_chain_object(st_data_t key, st_data_t val, st_data_t arg) -{ - struct force_finalize_list **prev = (struct force_finalize_list **)arg; - struct force_finalize_list *curr = ALLOC(struct force_finalize_list); - curr->obj = key; - curr->table = val; - curr->next = *prev; - *prev = curr; - return ST_CONTINUE; -} + RB_GC_VM_UNLOCK(lev); -void -rb_gc_call_finalizer_at_exit(void) -{ - rb_objspace_call_finalizer(&rb_objspace); -} + if (found) { + return obj; + } -static void -rb_objspace_call_finalizer(rb_objspace_t *objspace) -{ - RVALUE *p, *pend; - size_t i; - - gc_rest_sweep(objspace); - - if (ATOMIC_EXCHANGE(finalizing, 1)) return; - - /* run finalizers */ - finalize_deferred(objspace); - assert(heap_pages_deferred_final == 0); - - /* force to run finalizer */ - while (finalizer_table->num_entries) { - struct force_finalize_list *list = 0; - st_foreach(finalizer_table, force_chain_object, (st_data_t)&list); - while (list) { - struct force_finalize_list *curr = list; - st_data_t obj = (st_data_t)curr->obj; - run_finalizer(objspace, curr->obj, curr->table); - st_delete(finalizer_table, &obj, 0); - list = curr->next; - xfree(curr); - } - } - - /* finalizers are part of garbage collection */ - during_gc++; - - /* run data object's finalizers */ - for (i = 0; i < heap_pages_used; i++) { - p = heap_pages_sorted[i]->start; pend = p + heap_pages_sorted[i]->limit; - while (p < pend) { - switch (BUILTIN_TYPE(p)) { - case T_DATA: - if (!DATA_PTR(p) || !RANY(p)->as.data.dfree) break; - if (rb_obj_is_thread((VALUE)p)) break; - if (rb_obj_is_mutex((VALUE)p)) break; - if (rb_obj_is_fiber((VALUE)p)) break; - p->as.free.flags = 0; - if (RTYPEDDATA_P(p)) { - RDATA(p)->dfree = RANY(p)->as.typeddata.type->function.dfree; - } - if (RANY(p)->as.data.dfree == (RUBY_DATA_FUNC)-1) { - xfree(DATA_PTR(p)); - } - else if (RANY(p)->as.data.dfree) { - make_deferred(objspace, RANY(p)); - } - break; - case T_FILE: - if (RANY(p)->as.file.fptr) { - make_io_deferred(objspace, RANY(p)); - } - break; - } - p++; - } - } - during_gc = 0; - if (heap_pages_deferred_final) { - finalize_list(objspace, heap_pages_deferred_final); - } - - st_free_table(finalizer_table); - finalizer_table = 0; - ATOMIC_SET(finalizing, 0); -} - -static inline int -is_id_value(rb_objspace_t *objspace, VALUE ptr) -{ - if (!is_pointer_to_heap(objspace, (void *)ptr)) return FALSE; - if (BUILTIN_TYPE(ptr) > T_FIXNUM) return FALSE; - if (BUILTIN_TYPE(ptr) == T_ICLASS) return FALSE; - return TRUE; -} - -static inline int -heap_is_swept_object(rb_objspace_t *objspace, rb_heap_t *heap, VALUE ptr) -{ - struct heap_page *page = GET_HEAP_PAGE(ptr); - return page->before_sweep ? FALSE : TRUE; -} - -static inline int -is_swept_object(rb_objspace_t *objspace, VALUE ptr) -{ - if (heap_is_swept_object(objspace, heap_eden, ptr)) { - return TRUE; + if (rb_funcall(object_id, rb_intern(">="), 1, ULL2NUM(LAST_OBJECT_ID()))) { + rb_raise(rb_eRangeError, "%+"PRIsVALUE" is not an id value", rb_funcall(object_id, rb_intern("to_s"), 1, INT2FIX(10))); } else { - return FALSE; + rb_raise(rb_eRangeError, "%+"PRIsVALUE" is a recycled object", rb_funcall(object_id, rb_intern("to_s"), 1, INT2FIX(10))); } } -static inline int -is_dead_object(rb_objspace_t *objspace, VALUE ptr) -{ - if (!is_lazy_sweeping(heap_eden) || MARKED_IN_BITMAP(GET_HEAP_MARK_BITS(ptr), ptr)) return FALSE; - if (!is_swept_object(objspace, ptr)) return TRUE; - return FALSE; -} +static inline void +obj_free_object_id(VALUE obj) +{ + VALUE obj_id = 0; + if (RB_UNLIKELY(id2ref_tbl)) { + switch (BUILTIN_TYPE(obj)) { + case T_CLASS: + case T_MODULE: + obj_id = RCLASS(obj)->object_id; + break; + case T_IMEMO: + if (!IMEMO_TYPE_P(obj, imemo_fields)) { + return; + } + // fallthrough + case T_OBJECT: + { + shape_id_t shape_id = RBASIC_SHAPE_ID(obj); + if (rb_shape_has_object_id(shape_id)) { + obj_id = object_id_get(obj, shape_id); + } + break; + } + default: + // For generic_fields, the T_IMEMO/fields is responsible for freeing the id. + return; + } -static inline int -is_live_object(rb_objspace_t *objspace, VALUE ptr) -{ - switch (BUILTIN_TYPE(ptr)) { - case 0: case T_ZOMBIE: - return FALSE; + if (RB_UNLIKELY(obj_id)) { + RUBY_ASSERT(FIXNUM_P(obj_id) || RB_TYPE_P(obj_id, T_BIGNUM)); + + if (!st_delete(id2ref_tbl, (st_data_t *)&obj_id, NULL)) { + // The the object is a T_IMEMO/fields, then it's possible the actual object + // has been garbage collected already. + if (!RB_TYPE_P(obj, T_IMEMO)) { + rb_bug("Object ID seen, but not in _id2ref table: object_id=%llu object=%s", NUM2ULL(obj_id), rb_obj_info(obj)); + } + } + } } - if (is_dead_object(objspace, ptr)) return FALSE; - return TRUE; } -static inline int -is_markable_object(rb_objspace_t *objspace, VALUE obj) +void +rb_gc_obj_free_vm_weak_references(VALUE obj) { - if (rb_special_const_p(obj)) return 0; /* special const is not markable */ + ASSUME(!RB_SPECIAL_CONST_P(obj)); + obj_free_object_id(obj); - if (RGENGC_CHECK_MODE) { - if (!is_pointer_to_heap(objspace, (void *)obj)) rb_bug("is_markable_object: %p is not pointer to heap", (void *)obj); - if (BUILTIN_TYPE(obj) == T_NONE) rb_bug("is_markable_object: %p is T_NONE", (void *)obj); - if (BUILTIN_TYPE(obj) == T_ZOMBIE) rb_bug("is_markable_object: %p is T_ZOMBIE", (void *)obj); + if (rb_obj_gen_fields_p(obj)) { + rb_free_generic_ivar(obj); } - return 1; -} - -int -rb_objspace_markable_object_p(VALUE obj) -{ - rb_objspace_t *objspace = &rb_objspace; - return is_markable_object(objspace, obj) && is_live_object(objspace, obj); + switch (BUILTIN_TYPE(obj)) { + case T_STRING: + if (FL_TEST_RAW(obj, RSTRING_FSTR)) { + rb_gc_free_fstring(obj); + } + break; + case T_SYMBOL: + rb_gc_free_dsymbol(obj); + break; + case T_IMEMO: + switch (imemo_type(obj)) { + case imemo_callinfo: + rb_vm_ci_free((const struct rb_callinfo *)obj); + break; + case imemo_ment: + rb_free_method_entry_vm_weak_references((const rb_method_entry_t *)obj); + break; + default: + break; + } + break; + default: + break; + } } /* @@ -2295,47 +2379,86 @@ rb_objspace_markable_object_p(VALUE obj) * r = ObjectSpace._id2ref(s.object_id) #=> "I am a string" * r == s #=> true * + * On multi-ractor mode, if the object is not shareable, it raises + * RangeError. + * + * This method is deprecated and should no longer be used. */ static VALUE -id2ref(VALUE obj, VALUE objid) -{ -#if SIZEOF_LONG == SIZEOF_VOIDP -#define NUM2PTR(x) NUM2ULONG(x) -#elif SIZEOF_LONG_LONG == SIZEOF_VOIDP -#define NUM2PTR(x) NUM2ULL(x) -#endif - rb_objspace_t *objspace = &rb_objspace; - VALUE ptr; - void *p0; - - ptr = NUM2PTR(objid); - p0 = (void *)ptr; - - if (ptr == Qtrue) return Qtrue; - if (ptr == Qfalse) return Qfalse; - if (ptr == Qnil) return Qnil; - if (FIXNUM_P(ptr)) return (VALUE)ptr; - if (FLONUM_P(ptr)) return (VALUE)ptr; - ptr = obj_id_to_ref(objid); +id2ref(VALUE objid) +{ + objid = rb_to_int(objid); + if (FIXNUM_P(objid) || rb_big_size(objid) <= SIZEOF_VOIDP) { + VALUE ptr = (VALUE)NUM2PTR(objid); + if (SPECIAL_CONST_P(ptr)) { + if (ptr == Qtrue) return Qtrue; + if (ptr == Qfalse) return Qfalse; + if (NIL_P(ptr)) return Qnil; + if (FIXNUM_P(ptr)) return ptr; + if (FLONUM_P(ptr)) return ptr; + + if (SYMBOL_P(ptr)) { + // Check that the symbol is valid + if (rb_static_id_valid_p(SYM2ID(ptr))) { + return ptr; + } + else { + rb_raise(rb_eRangeError, "%p is not a symbol id value", (void *)ptr); + } + } - if ((ptr % sizeof(RVALUE)) == (4 << 2)) { - ID symid = ptr / sizeof(RVALUE); - if (rb_id2name(symid) == 0) - rb_raise(rb_eRangeError, "%p is not symbol id value", p0); - return ID2SYM(symid); + rb_raise(rb_eRangeError, "%+"PRIsVALUE" is not an id value", rb_int2str(objid, 10)); + } } - if (!is_id_value(objspace, ptr)) { - rb_raise(rb_eRangeError, "%p is not id value", p0); + VALUE obj = object_id_to_ref(rb_gc_get_objspace(), objid); + if (!rb_multi_ractor_p() || rb_ractor_shareable_p(obj)) { + return obj; } - if (!is_live_object(objspace, ptr)) { - rb_raise(rb_eRangeError, "%p is recycled object", p0); + else { + rb_raise(rb_eRangeError, "%+"PRIsVALUE" is the id of an unshareable object on multi-ractor", rb_int2str(objid, 10)); } - if (RBASIC(ptr)->klass == 0) { - rb_raise(rb_eRangeError, "%p is internal object", p0); +} + +/* :nodoc: */ +static VALUE +os_id2ref(VALUE os, VALUE objid) +{ + rb_category_warn(RB_WARN_CATEGORY_DEPRECATED, "ObjectSpace._id2ref is deprecated"); + return id2ref(objid); +} + +static VALUE +rb_find_object_id(void *objspace, VALUE obj, VALUE (*get_heap_object_id)(VALUE)) +{ + if (SPECIAL_CONST_P(obj)) { +#if SIZEOF_LONG == SIZEOF_VOIDP + return LONG2NUM((SIGNED_VALUE)obj); +#else + return LL2NUM((SIGNED_VALUE)obj); +#endif } - return (VALUE)ptr; + + return get_heap_object_id(obj); +} + +static VALUE +nonspecial_obj_id(VALUE obj) +{ +#if SIZEOF_LONG == SIZEOF_VOIDP + return (VALUE)((SIGNED_VALUE)(obj)|FIXNUM_FLAG); +#elif SIZEOF_LONG_LONG == SIZEOF_VOIDP + return LL2NUM((SIGNED_VALUE)(obj) / 2); +#else +# error not supported +#endif +} + +VALUE +rb_memory_id(VALUE obj) +{ + return rb_find_object_id(NULL, obj, nonspecial_obj_id); } /* @@ -2348,209 +2471,209 @@ id2ref(VALUE obj, VALUE objid) * * Returns an integer identifier for +obj+. * - * The same number will be returned on all calls to +id+ for a given object, - * and no two active objects will share an id. + * The same number will be returned on all calls to +object_id+ for a given + * object, and no two active objects will share an id. * - * Object#object_id is a different concept from the +:name+ notation, which - * returns the symbol id of +name+. + * Note: that some objects of builtin classes are reused for optimization. + * This is the case for immediate values and frozen string literals. * - * Replaces the deprecated Object#id. - */ - -/* - * call-seq: - * obj.hash -> fixnum + * BasicObject implements +__id__+, Kernel implements +object_id+. * - * Generates a Fixnum hash value for this object. + * Immediate values are not passed by reference but are passed by value: + * +nil+, +true+, +false+, Fixnums, Symbols, and some Floats. * - * This function must have the property that <code>a.eql?(b)</code> implies - * <code>a.hash == b.hash</code>. - * - * The hash value is used by Hash class. - * - * Any hash value that exceeds the capacity of a Fixnum will be truncated - * before being used. + * Object.new.object_id == Object.new.object_id # => false + * (21 * 2).object_id == (21 * 2).object_id # => true + * "hello".object_id == "hello".object_id # => false + * "hi".freeze.object_id == "hi".freeze.object_id # => true */ VALUE rb_obj_id(VALUE obj) { - /* - * 32-bit VALUE space - * MSB ------------------------ LSB - * false 00000000000000000000000000000000 - * true 00000000000000000000000000000010 - * nil 00000000000000000000000000000100 - * undef 00000000000000000000000000000110 - * symbol ssssssssssssssssssssssss00001110 - * object oooooooooooooooooooooooooooooo00 = 0 (mod sizeof(RVALUE)) - * fixnum fffffffffffffffffffffffffffffff1 - * - * object_id space - * LSB - * false 00000000000000000000000000000000 - * true 00000000000000000000000000000010 - * nil 00000000000000000000000000000100 - * undef 00000000000000000000000000000110 - * symbol 000SSSSSSSSSSSSSSSSSSSSSSSSSSS0 S...S % A = 4 (S...S = s...s * A + 4) - * object oooooooooooooooooooooooooooooo0 o...o % A = 0 - * fixnum fffffffffffffffffffffffffffffff1 bignum if required - * - * where A = sizeof(RVALUE)/4 - * - * sizeof(RVALUE) is - * 20 if 32-bit, double is 4-byte aligned - * 24 if 32-bit, double is 8-byte aligned - * 40 if 64-bit - */ - if (SYMBOL_P(obj)) { - return (SYM2ID(obj) * sizeof(RVALUE) + (4 << 2)) | FIXNUM_FLAG; - } - else if (FLONUM_P(obj)) { -#if SIZEOF_LONG == SIZEOF_VOIDP - return LONG2NUM((SIGNED_VALUE)obj); -#else - return LL2NUM((SIGNED_VALUE)obj); + /* If obj is an immediate, the object ID is obj directly converted to a Numeric. + * Otherwise, the object ID is a Numeric that is a non-zero multiple of + * (RUBY_IMMEDIATE_MASK + 1) which guarantees that it does not collide with + * any immediates. */ + return rb_find_object_id(rb_gc_get_objspace(), obj, object_id); +} + +bool +rb_obj_id_p(VALUE obj) +{ + return !RB_TYPE_P(obj, T_IMEMO) && rb_obj_shape_has_id(obj); +} + +/* + * GC implementations should call this function before the GC phase that updates references + * embedded in the machine code generated by JIT compilers. JIT compilers usually enforce the + * "W^X" policy and protect the code memory from being modified during execution. This function + * makes the code memory writeable. + */ +void +rb_gc_before_updating_jit_code(void) +{ +#if USE_YJIT + rb_yjit_mark_all_writeable(); +#endif +#if USE_ZJIT + rb_zjit_mark_all_writable(); +#endif +} + +/* + * GC implementations should call this function before the GC phase that updates references + * embedded in the machine code generated by JIT compilers. This function makes the code memory + * executable again. + */ +void +rb_gc_after_updating_jit_code(void) +{ +#if USE_YJIT + rb_yjit_mark_all_executable(); #endif +#if USE_ZJIT + rb_zjit_mark_all_executable(); +#endif +} + +static void +classext_memsize(rb_classext_t *ext, bool prime, VALUE box_value, void *arg) +{ + size_t *size = (size_t *)arg; + size_t s = 0; + + if (RCLASSEXT_M_TBL(ext)) { + s += rb_id_table_memsize(RCLASSEXT_M_TBL(ext)); + } + if (RCLASSEXT_CONST_TBL(ext)) { + s += rb_id_table_memsize(RCLASSEXT_CONST_TBL(ext)); + } + if (RCLASSEXT_SUPERCLASSES_WITH_SELF(ext)) { + s += (RCLASSEXT_SUPERCLASS_DEPTH(ext) + 1) * sizeof(VALUE); } - else if (SPECIAL_CONST_P(obj)) { - return LONG2NUM((SIGNED_VALUE)obj); + if (!prime) { + s += sizeof(rb_classext_t); } - return nonspecial_obj_id(obj); + *size += s; } -size_t rb_str_memsize(VALUE); -size_t rb_ary_memsize(VALUE); -size_t rb_io_memsize(const rb_io_t *); -size_t rb_generic_ivar_memsize(VALUE); -#include "regint.h" +static void +classext_superclasses_memsize(rb_classext_t *ext, bool prime, VALUE box_value, void *arg) +{ + size_t *size = (size_t *)arg; + size_t array_size; + if (RCLASSEXT_SUPERCLASSES_WITH_SELF(ext)) { + RUBY_ASSERT(prime); + array_size = RCLASSEXT_SUPERCLASS_DEPTH(ext) + 1; + *size += array_size * sizeof(VALUE); + } +} -static size_t -obj_memsize_of(VALUE obj, int use_tdata) +size_t +rb_obj_memsize_of(VALUE obj) { size_t size = 0; if (SPECIAL_CONST_P(obj)) { - return 0; - } - - if (FL_TEST(obj, FL_EXIVAR)) { - size += rb_generic_ivar_memsize(obj); + return 0; } switch (BUILTIN_TYPE(obj)) { case T_OBJECT: - if (!(RBASIC(obj)->flags & ROBJECT_EMBED) && - ROBJECT(obj)->as.heap.ivptr) { - size += ROBJECT(obj)->as.heap.numiv * sizeof(VALUE); - } - break; + if (FL_TEST_RAW(obj, ROBJECT_HEAP)) { + if (rb_obj_shape_complex_p(obj)) { + size += rb_st_memsize(ROBJECT_FIELDS_HASH(obj)); + } + else { + size += ROBJECT_FIELDS_CAPACITY(obj) * sizeof(VALUE); + } + } + break; case T_MODULE: case T_CLASS: - if (RCLASS_M_TBL_WRAPPER(obj)) { - size += sizeof(struct method_table_wrapper); - } - if (RCLASS_M_TBL(obj)) { - size += st_memsize(RCLASS_M_TBL(obj)); - } - if (RCLASS_EXT(obj)) { - if (RCLASS_IV_TBL(obj)) { - size += st_memsize(RCLASS_IV_TBL(obj)); - } - if (RCLASS_IV_INDEX_TBL(obj)) { - size += st_memsize(RCLASS_IV_INDEX_TBL(obj)); - } - if (RCLASS(obj)->ptr->iv_tbl) { - size += st_memsize(RCLASS(obj)->ptr->iv_tbl); - } - if (RCLASS(obj)->ptr->const_tbl) { - size += st_memsize(RCLASS(obj)->ptr->const_tbl); - } - size += sizeof(rb_classext_t); - } - break; + rb_class_classext_foreach(obj, classext_memsize, (void *)&size); + rb_class_classext_foreach(obj, classext_superclasses_memsize, (void *)&size); + break; + case T_ICLASS: + if (RICLASS_OWNS_M_TBL_P(obj)) { + if (RCLASS_M_TBL(obj)) { + size += rb_id_table_memsize(RCLASS_M_TBL(obj)); + } + } + break; case T_STRING: - size += rb_str_memsize(obj); - break; + size += rb_str_memsize(obj); + break; case T_ARRAY: - size += rb_ary_memsize(obj); - break; + size += rb_ary_memsize(obj); + break; case T_HASH: - if (RHASH(obj)->ntbl) { - size += st_memsize(RHASH(obj)->ntbl); - } - break; + if (RHASH_ST_TABLE_P(obj)) { + VM_ASSERT(RHASH_ST_TABLE(obj) != NULL); + /* st_table is in the slot */ + size += st_memsize(RHASH_ST_TABLE(obj)) - sizeof(st_table); + } + break; case T_REGEXP: - if (RREGEXP(obj)->ptr) { - size += onig_memsize(RREGEXP(obj)->ptr); - } - break; + if (RREGEXP_PTR(obj)) { + size += onig_memsize(RREGEXP_PTR(obj)); + } + break; case T_DATA: - if (use_tdata) size += rb_objspace_data_type_memsize(obj); - break; + size += rb_objspace_data_type_memsize(obj); + break; case T_MATCH: - if (RMATCH(obj)->rmatch) { - struct rmatch *rm = RMATCH(obj)->rmatch; - size += onig_region_memsize(&rm->regs); - size += sizeof(struct rmatch_offset) * rm->char_offset_num_allocated; - size += sizeof(struct rmatch); - } - break; + { + struct RMatch *rm = RMATCH(obj); + if (FL_TEST_RAW(obj, RMATCH_ONIG)) { + size += onig_region_memsize(&rm->as.onig); + } + size += sizeof(struct rmatch_offset) * rm->char_offset_num_allocated; + } + break; case T_FILE: - if (RFILE(obj)->fptr) { - size += rb_io_memsize(RFILE(obj)->fptr); - } - break; + if (RFILE(obj)->fptr) { + size += rb_io_memsize(RFILE(obj)->fptr); + } + break; case T_RATIONAL: case T_COMPLEX: - break; - case T_ICLASS: - /* iClass shares table with the module */ - break; + break; + case T_IMEMO: + size += rb_imemo_memsize(obj); + break; case T_FLOAT: - break; + case T_SYMBOL: + break; case T_BIGNUM: - if (!(RBASIC(obj)->flags & RBIGNUM_EMBED_FLAG) && RBIGNUM_DIGITS(obj)) { - size += RBIGNUM_LEN(obj) * sizeof(BDIGIT); - } - break; + if (!(RBASIC(obj)->flags & BIGNUM_EMBED_FLAG) && BIGNUM_DIGITS(obj)) { + size += BIGNUM_LEN(obj) * sizeof(BDIGIT); + } + break; + case T_NODE: - switch (nd_type(obj)) { - case NODE_SCOPE: - if (RNODE(obj)->u1.tbl) { - /* TODO: xfree(RANY(obj)->as.node.u1.tbl); */ - } - break; - case NODE_ALLOCA: - /* TODO: xfree(RANY(obj)->as.node.u1.node); */ - ; - } - break; /* no need to free iv_tbl */ + UNEXPECTED_NODE(obj_memsize_of); + break; case T_STRUCT: - if ((RBASIC(obj)->flags & RSTRUCT_EMBED_LEN_MASK) == 0 && - RSTRUCT(obj)->as.heap.ptr) { - size += sizeof(VALUE) * RSTRUCT_LEN(obj); - } - break; + if (RSTRUCT_EMBED_LEN(obj) == 0) { + size += sizeof(VALUE) * RSTRUCT_LEN_RAW(obj); + } + break; case T_ZOMBIE: - break; + case T_MOVED: + break; default: - rb_bug("objspace/memsize_of(): unknown data type 0x%x(%p)", - BUILTIN_TYPE(obj), (void*)obj); + rb_bug("objspace/memsize_of(): unknown data type 0x%x(%p)", + BUILTIN_TYPE(obj), (void*)obj); } - return size; -} - -size_t -rb_obj_memsize_of(VALUE obj) -{ - return obj_memsize_of(obj, TRUE); + return size + rb_gc_obj_slot_size(obj); } static int @@ -2562,26 +2685,62 @@ set_zero(st_data_t key, st_data_t val, st_data_t arg) return ST_CONTINUE; } +struct count_objects_data { + size_t counts[T_MASK+1]; + size_t freed; + size_t total; +}; + +static void +count_objects_i(VALUE obj, void *d) +{ + struct count_objects_data *data = (struct count_objects_data *)d; + + if (RBASIC(obj)->flags) { + data->counts[BUILTIN_TYPE(obj)]++; + } + else { + data->freed++; + } + + data->total++; +} + /* * call-seq: - * ObjectSpace.count_objects([result_hash]) -> hash + * ObjectSpace.count_objects(result_hash = {}) -> hash + * + * Counts the number of objects, grouped by type. + * + * It returns a hash that looks like: + * + * { + * TOTAL: 10000, + * FREE: 3011, + * T_OBJECT: 6, + * T_CLASS: 404, + * # ... + * } * - * Counts objects for each type. + * The contents of the returned hash are implementation specific and + * may be changed in future versions without notice. * - * It returns a hash, such as: - * { - * :TOTAL=>10000, - * :FREE=>3011, - * :T_OBJECT=>6, - * :T_CLASS=>404, - * # ... - * } + * The keys starting with +:T_+ are live objects of a particular type. + * For example, +:T_ARRAY+ is the number of arrays. * - * The contents of the returned hash are implementation specific. - * It may be changed in future. + * The key +:FREE+ is the number of object slots which are empty. + * + * The key +:TOTAL+ is the total number of slots (which is the sum of + * all of the other values). * * If the optional argument +result_hash+ is given, - * it is overwritten and returned. This is intended to avoid probe effect. + * it is overwritten and returned. + * This is intended to avoid the probe effect. + * + * h = {} + * ObjectSpace.count_objects(h) + * puts h + * # => { TOTAL: 10000, T_CLASS: 158280, T_MODULE: 20672, T_STRING: 527249 } * * This method is only expected to work on C Ruby. * @@ -2590,2596 +2749,1779 @@ set_zero(st_data_t key, st_data_t val, st_data_t arg) static VALUE count_objects(int argc, VALUE *argv, VALUE os) { - rb_objspace_t *objspace = &rb_objspace; - size_t counts[T_MASK+1]; - size_t freed = 0; - size_t total = 0; - size_t i; - VALUE hash; + struct count_objects_data data = { 0 }; + VALUE hash = Qnil; + VALUE types[T_MASK + 1]; - if (rb_scan_args(argc, argv, "01", &hash) == 1) { + if (rb_check_arity(argc, 0, 1) == 1) { + hash = argv[0]; if (!RB_TYPE_P(hash, T_HASH)) rb_raise(rb_eTypeError, "non-hash given"); } - for (i = 0; i <= T_MASK; i++) { - counts[i] = 0; + for (size_t i = 0; i <= T_MASK; i++) { + // type_sym can allocate an object, + // so we need to create all key symbols in advance + // not to disturb the result + types[i] = type_sym(i); } - for (i = 0; i < heap_pages_used; i++) { - struct heap_page *page = heap_pages_sorted[i]; - RVALUE *p, *pend; + // Same as type_sym, we need to create all key symbols in advance + VALUE total = ID2SYM(rb_intern("TOTAL")); + VALUE free = ID2SYM(rb_intern("FREE")); - p = page->start; pend = p + page->limit; - for (;p < pend; p++) { - if (p->as.basic.flags) { - counts[BUILTIN_TYPE(p)]++; - } - else { - freed++; - } - } - total += page->limit; - } + rb_gc_impl_each_object(rb_gc_get_objspace(), count_objects_i, &data); - if (hash == Qnil) { + if (NIL_P(hash)) { hash = rb_hash_new(); } else if (!RHASH_EMPTY_P(hash)) { - st_foreach(RHASH_TBL_RAW(hash), set_zero, hash); - } - rb_hash_aset(hash, ID2SYM(rb_intern("TOTAL")), SIZET2NUM(total)); - rb_hash_aset(hash, ID2SYM(rb_intern("FREE")), SIZET2NUM(freed)); - - for (i = 0; i <= T_MASK; i++) { - VALUE type; - switch (i) { -#define COUNT_TYPE(t) case (t): type = ID2SYM(rb_intern(#t)); break; - COUNT_TYPE(T_NONE); - COUNT_TYPE(T_OBJECT); - COUNT_TYPE(T_CLASS); - COUNT_TYPE(T_MODULE); - COUNT_TYPE(T_FLOAT); - COUNT_TYPE(T_STRING); - COUNT_TYPE(T_REGEXP); - COUNT_TYPE(T_ARRAY); - COUNT_TYPE(T_HASH); - COUNT_TYPE(T_STRUCT); - COUNT_TYPE(T_BIGNUM); - COUNT_TYPE(T_FILE); - COUNT_TYPE(T_DATA); - COUNT_TYPE(T_MATCH); - COUNT_TYPE(T_COMPLEX); - COUNT_TYPE(T_RATIONAL); - COUNT_TYPE(T_NIL); - COUNT_TYPE(T_TRUE); - COUNT_TYPE(T_FALSE); - COUNT_TYPE(T_SYMBOL); - COUNT_TYPE(T_FIXNUM); - COUNT_TYPE(T_UNDEF); - COUNT_TYPE(T_NODE); - COUNT_TYPE(T_ICLASS); - COUNT_TYPE(T_ZOMBIE); -#undef COUNT_TYPE - default: type = INT2NUM(i); break; + rb_hash_stlike_foreach(hash, set_zero, hash); + } + rb_hash_aset(hash, total, SIZET2NUM(data.total)); + rb_hash_aset(hash, free, SIZET2NUM(data.freed)); + + for (size_t i = 0; i <= T_MASK; i++) { + if (data.counts[i]) { + rb_hash_aset(hash, types[i], SIZET2NUM(data.counts[i])); } - if (counts[i]) - rb_hash_aset(hash, type, SIZET2NUM(counts[i])); } return hash; } -/* - ------------------------ Garbage Collection ------------------------ -*/ +#define SET_STACK_END SET_MACHINE_STACK_END(&ec->machine.stack_end) -/* Sweeping */ +#define STACK_START (ec->machine.stack_start) +#define STACK_END (ec->machine.stack_end) +#define STACK_LEVEL_MAX (ec->machine.stack_maxsize/sizeof(VALUE)) -static VALUE -lazy_sweep_enable(void) +#if STACK_GROW_DIRECTION < 0 +# define STACK_LENGTH (size_t)(STACK_START - STACK_END) +#elif STACK_GROW_DIRECTION > 0 +# define STACK_LENGTH (size_t)(STACK_END - STACK_START + 1) +#else +# define STACK_LENGTH ((STACK_END < STACK_START) ? (size_t)(STACK_START - STACK_END) \ + : (size_t)(STACK_END - STACK_START + 1)) +#endif +#if !STACK_GROW_DIRECTION +int ruby_stack_grow_direction; +int +ruby_get_stack_grow_direction(volatile VALUE *addr) { - rb_objspace_t *objspace = &rb_objspace; + VALUE *end; + SET_MACHINE_STACK_END(&end); - objspace->flags.dont_lazy_sweep = FALSE; - return Qnil; + if (end > addr) return ruby_stack_grow_direction = 1; + return ruby_stack_grow_direction = -1; } +#endif -static size_t -objspace_live_slot(rb_objspace_t *objspace) +size_t +ruby_stack_length(VALUE **p) { - return objspace->profile.total_allocated_object_num - objspace->profile.total_freed_object_num; + rb_execution_context_t *ec = GET_EC(); + SET_STACK_END; + if (p) *p = STACK_UPPER(STACK_END, STACK_START, STACK_END); + return STACK_LENGTH; } -static size_t -objspace_total_slot(rb_objspace_t *objspace) +#define PREVENT_STACK_OVERFLOW 1 +#ifndef PREVENT_STACK_OVERFLOW +#if !(defined(POSIX_SIGNAL) && defined(SIGSEGV) && defined(HAVE_SIGALTSTACK)) +# define PREVENT_STACK_OVERFLOW 1 +#else +# define PREVENT_STACK_OVERFLOW 0 +#endif +#endif +#if PREVENT_STACK_OVERFLOW && !defined(__EMSCRIPTEN__) +static int +stack_check(rb_execution_context_t *ec, int water_mark) { - return heap_eden->total_slots + heap_tomb->total_slots; + SET_STACK_END; + + size_t length = STACK_LENGTH; + size_t maximum_length = STACK_LEVEL_MAX - water_mark; + + return length > maximum_length; } +#else +#define stack_check(ec, water_mark) FALSE +#endif -static size_t -objspace_free_slot(rb_objspace_t *objspace) +#define STACKFRAME_FOR_CALL_CFUNC 2048 + +int +rb_ec_stack_check(rb_execution_context_t *ec) { - return objspace_total_slot(objspace) - (objspace_live_slot(objspace) - heap_pages_final_slots); + return stack_check(ec, STACKFRAME_FOR_CALL_CFUNC); } -static void -gc_setup_mark_bits(struct heap_page *page) +int +ruby_stack_check(void) { -#if USE_RGENGC - /* copy oldgen bitmap to mark bitmap */ - memcpy(&page->mark_bits[0], &page->oldgen_bits[0], HEAP_BITMAP_SIZE); -#else - /* clear mark bitmap */ - memset(&page->mark_bits[0], 0, HEAP_BITMAP_SIZE); -#endif -} + return stack_check(GET_EC(), STACKFRAME_FOR_CALL_CFUNC); +} + +/* ==================== Marking ==================== */ + +#define RB_GC_MARK_OR_TRAVERSE(func, obj_or_ptr, obj, check_obj) do { \ + if (!RB_SPECIAL_CONST_P(obj)) { \ + rb_vm_t *vm = GET_VM(); \ + void *objspace = vm->gc.objspace; \ + if (LIKELY(vm->gc.mark_func_data == NULL)) { \ + GC_ASSERT(rb_gc_impl_during_gc_p(objspace)); \ + (func)(objspace, (obj_or_ptr)); \ + } \ + else if (check_obj ? \ + rb_gc_impl_pointer_to_heap_p(objspace, (const void *)obj) && \ + !rb_gc_impl_garbage_object_p(objspace, obj) : \ + true) { \ + GC_ASSERT(!rb_gc_impl_during_gc_p(objspace)); \ + struct gc_mark_func_data_struct *mark_func_data = vm->gc.mark_func_data; \ + vm->gc.mark_func_data = NULL; \ + mark_func_data->mark_func((obj), mark_func_data->data); \ + vm->gc.mark_func_data = mark_func_data; \ + } \ + } \ +} while (0) static inline void -gc_page_sweep(rb_objspace_t *objspace, rb_heap_t *heap, struct heap_page *sweep_page) -{ - int i; - size_t empty_slots = 0, freed_slots = 0, final_slots = 0; - RVALUE *p, *pend,*offset; - bits_t *bits, bitset; - - rgengc_report(1, objspace, "page_sweep: start.\n"); - - sweep_page->before_sweep = 0; - - p = sweep_page->start; pend = p + sweep_page->limit; - offset = p - NUM_IN_PAGE(p); - bits = sweep_page->mark_bits; - - /* create guard : fill 1 out-of-range */ - bits[BITMAP_INDEX(p)] |= BITMAP_BIT(p)-1; - bits[BITMAP_INDEX(pend)] |= ~(BITMAP_BIT(pend) - 1); - - for (i=0; i < HEAP_BITMAP_LIMIT; i++) { - bitset = ~bits[i]; - if (bitset) { - p = offset + i * BITS_BITLENGTH; - do { - if ((bitset & 1) && BUILTIN_TYPE(p) != T_ZOMBIE) { - if (p->as.basic.flags) { - rgengc_report(3, objspace, "page_sweep: free %p (%s)\n", p, obj_type_name((VALUE)p)); -#if USE_RGENGC && RGENGC_CHECK_MODE - if (objspace->rgengc.during_minor_gc && RVALUE_OLD_P((VALUE)p)) rb_bug("page_sweep: %p (%s) is old while minor GC.\n", p, obj_type_name((VALUE)p)); - if (rgengc_remembered(objspace, (VALUE)p)) rb_bug("page_sweep: %p (%s) is remembered.\n", p, obj_type_name((VALUE)p)); -#endif - if (obj_free(objspace, (VALUE)p)) { - final_slots++; - } - else if (FL_TEST(p, FL_FINALIZE)) { - RDATA(p)->dfree = 0; - make_deferred(objspace,p); - final_slots++; - } - else { - (void)VALGRIND_MAKE_MEM_UNDEFINED((void*)p, sizeof(RVALUE)); - heap_page_add_freeobj(objspace, sweep_page, (VALUE)p); - rgengc_report(3, objspace, "page_sweep: %p (%s) is added to freelist\n", p, obj_type_name((VALUE)p)); - freed_slots++; - } - } - else { - empty_slots++; - } - } - p++; - bitset >>= 1; - } while (bitset); - } - } - - gc_setup_mark_bits(sweep_page); - -#if GC_PROFILE_MORE_DETAIL - if (gc_prof_enabled(objspace)) { - gc_profile_record *record = gc_prof_record(objspace); - record->removing_objects += final_slots + freed_slots; - record->empty_objects += empty_slots; - } -#endif - - if (final_slots + freed_slots + empty_slots == sweep_page->limit) { - /* there are no living objects -> move this page to tomb heap */ - heap_unlink_page(objspace, heap, sweep_page); - heap_add_page(objspace, heap_tomb, sweep_page); - } - else { - if (freed_slots + empty_slots > 0) { - heap_add_freepage(objspace, heap, sweep_page); - } - else { - sweep_page->free_next = NULL; - } - } - heap_pages_swept_slots += freed_slots + empty_slots; - objspace->profile.total_freed_object_num += freed_slots; - heap_pages_final_slots += final_slots; - sweep_page->final_slots = final_slots; - - if (0) fprintf(stderr, "gc_page_sweep(%d): freed?: %d, limt: %d, freed_slots: %d, empty_slots: %d, final_slots: %d\n", - (int)rb_gc_count(), - final_slots + freed_slots + empty_slots == sweep_page->limit, - (int)sweep_page->limit, (int)freed_slots, (int)empty_slots, (int)final_slots); - - if (heap_pages_deferred_final && !finalizing) { - rb_thread_t *th = GET_THREAD(); - if (th) { - gc_finalize_deferred_register(); - } - } - - rgengc_report(1, objspace, "page_sweep: end.\n"); +gc_mark_internal(VALUE obj) +{ + RB_GC_MARK_OR_TRAVERSE(rb_gc_impl_mark, obj, obj, false); } -/* allocate additional minimum page to work */ -static void -gc_heap_prepare_minimum_pages(rb_objspace_t *objspace, rb_heap_t *heap) +void +rb_gc_mark_movable(VALUE obj) { - if (!heap->free_pages) { - /* there is no free after page_sweep() */ - heap_set_increment(objspace, 0); - if (!heap_increment(objspace, heap)) { /* can't allocate additional free objects */ - during_gc = 0; - rb_memerror(); - } - } + gc_mark_internal(obj); } -static void -gc_before_heap_sweep(rb_objspace_t *objspace, rb_heap_t *heap) +void +rb_gc_mark_and_move(VALUE *ptr) { - heap->sweep_pages = heap->pages; - heap->free_pages = NULL; - - if (heap->using_page) { - RVALUE **p = &heap->using_page->freelist; - while (*p) { - p = &(*p)->as.free.next; - } - *p = heap->freelist; - heap->using_page = NULL; - } - heap->freelist = NULL; + RB_GC_MARK_OR_TRAVERSE(rb_gc_impl_mark_and_move, ptr, *ptr, false); } -#if defined(__GNUC__) && __GNUC__ == 4 && __GNUC_MINOR__ == 4 -__attribute__((noinline)) -#endif -static void -gc_before_sweep(rb_objspace_t *objspace) +static inline void +gc_mark_and_pin_internal(VALUE obj) { - rb_heap_t *heap; - size_t total_limit_slot; - - rgengc_report(1, objspace, "gc_before_sweep\n"); - - /* sweep unlinked method entries */ - if (GET_VM()->unlinked_method_entry_list) { - rb_sweep_method_entry(GET_VM()); - } - - heap_pages_swept_slots = 0; - total_limit_slot = objspace_total_slot(objspace); - - heap_pages_min_free_slots = (size_t)(total_limit_slot * 0.30); - if (heap_pages_min_free_slots < gc_params.heap_free_slots) { - heap_pages_min_free_slots = gc_params.heap_free_slots; - } - heap_pages_max_free_slots = (size_t)(total_limit_slot * 0.80); - if (heap_pages_max_free_slots < gc_params.heap_init_slots) { - heap_pages_max_free_slots = gc_params.heap_init_slots; - } - if (0) fprintf(stderr, "heap_pages_min_free_slots: %d, heap_pages_max_free_slots: %d\n", - (int)heap_pages_min_free_slots, (int)heap_pages_max_free_slots); - - heap = heap_eden; - gc_before_heap_sweep(objspace, heap); - - gc_prof_set_malloc_info(objspace); - - /* reset malloc info */ - if (0) fprintf(stderr, "%d\t%d\t%d\n", (int)rb_gc_count(), (int)malloc_increase, (int)malloc_limit); - - { - size_t inc = ATOMIC_SIZE_EXCHANGE(malloc_increase, 0); - size_t old_limit = malloc_limit; - - if (inc > malloc_limit) { - malloc_limit = (size_t)(inc * gc_params.malloc_limit_growth_factor); - if (gc_params.malloc_limit_max > 0 && /* ignore max-check if 0 */ - malloc_limit > gc_params.malloc_limit_max) { - malloc_limit = gc_params.malloc_limit_max; - } - } - else { - malloc_limit = (size_t)(malloc_limit * 0.98); /* magic number */ - if (malloc_limit < gc_params.malloc_limit_min) { - malloc_limit = gc_params.malloc_limit_min; - } - } - - if (0) { - if (old_limit != malloc_limit) { - fprintf(stderr, "[%"PRIuSIZE"] malloc_limit: %"PRIuSIZE" -> %"PRIuSIZE"\n", - rb_gc_count(), old_limit, malloc_limit); - } - else { - fprintf(stderr, "[%"PRIuSIZE"] malloc_limit: not changed (%"PRIuSIZE")\n", - rb_gc_count(), malloc_limit); - } - } - } - - /* reset oldmalloc info */ -#if RGENGC_ESTIMATE_OLDMALLOC - if (objspace->rgengc.during_minor_gc) { - if (objspace->rgengc.oldmalloc_increase > objspace->rgengc.oldmalloc_increase_limit) { - objspace->rgengc.need_major_gc = GPR_FLAG_MAJOR_BY_OLDMALLOC;; - objspace->rgengc.oldmalloc_increase_limit = - (size_t)(objspace->rgengc.oldmalloc_increase_limit * gc_params.oldmalloc_limit_growth_factor); - - if (objspace->rgengc.oldmalloc_increase_limit > gc_params.oldmalloc_limit_max) { - objspace->rgengc.oldmalloc_increase_limit = gc_params.oldmalloc_limit_max; - } - } - - if (0) fprintf(stderr, "%d\t%d\t%u\t%u\t%d\n", - (int)rb_gc_count(), - (int)objspace->rgengc.need_major_gc, - (unsigned int)objspace->rgengc.oldmalloc_increase, - (unsigned int)objspace->rgengc.oldmalloc_increase_limit, - (unsigned int)gc_params.oldmalloc_limit_max); - } - else { - /* major GC */ - objspace->rgengc.oldmalloc_increase = 0; + RB_GC_MARK_OR_TRAVERSE(rb_gc_impl_mark_and_pin, obj, obj, false); +} - if ((objspace->profile.latest_gc_info & GPR_FLAG_MAJOR_BY_OLDMALLOC) == 0) { - objspace->rgengc.oldmalloc_increase_limit = - (size_t)(objspace->rgengc.oldmalloc_increase_limit / ((gc_params.oldmalloc_limit_growth_factor - 1)/10 + 1)); - if (objspace->rgengc.oldmalloc_increase_limit < gc_params.oldmalloc_limit_min) { - objspace->rgengc.oldmalloc_increase_limit = gc_params.oldmalloc_limit_min; - } - } - } +void +rb_gc_mark(VALUE obj) +{ + gc_mark_and_pin_internal(obj); +} -#endif +static inline void +gc_mark_maybe_internal(VALUE obj) +{ + RB_GC_MARK_OR_TRAVERSE(rb_gc_impl_mark_maybe, obj, obj, true); +} +void +rb_gc_mark_maybe(VALUE obj) +{ + gc_mark_maybe_internal(obj); } +ATTRIBUTE_NO_ADDRESS_SAFETY_ANALYSIS(static void each_location(register const VALUE *x, register long n, void (*cb)(VALUE, void *), void *data)); static void -gc_after_sweep(rb_objspace_t *objspace) -{ - rb_heap_t *heap = heap_eden; - - rgengc_report(1, objspace, "after_gc_sweep: heap->total_slots: %d, heap->swept_slots: %d, min_free_slots: %d\n", - (int)heap->total_slots, (int)heap_pages_swept_slots, (int)heap_pages_min_free_slots); - - if (heap_pages_swept_slots < heap_pages_min_free_slots) { -#if USE_RGENGC - if (objspace->rgengc.during_minor_gc && objspace->profile.count - objspace->rgengc.last_major_gc > 2 /* magic number */) { - objspace->rgengc.need_major_gc = GPR_FLAG_MAJOR_BY_NOFREE; - } - else { - heap_set_increment(objspace, (heap_pages_min_free_slots - heap_pages_swept_slots) / HEAP_OBJ_LIMIT); - heap_increment(objspace, heap); - } -#else - heap_set_increment(objspace, (heap_pages_min_free_slots - heap_pages_swept_slots) / HEAP_OBJ_LIMIT); - heap_increment(objspace, heap); -#endif - } - - gc_prof_set_heap_info(objspace); - - heap_pages_free_unused_pages(objspace); - - /* if heap_pages has unused pages, then assign them to increment */ - if (heap_pages_increment < heap_tomb->page_length) { - heap_pages_increment = heap_tomb->page_length; - } - -#if RGENGC_PROFILE > 0 - if (0) { - fprintf(stderr, "%d\t%d\t%d\t%d\t%d\t%d\t%d\n", - (int)rb_gc_count(), - (int)objspace->profile.major_gc_count, - (int)objspace->profile.minor_gc_count, - (int)objspace->profile.promote_infant_count, -#if RGENGC_THREEGEN - (int)objspace->profile.promote_young_count, -#else - 0, -#endif - (int)objspace->profile.remembered_normal_object_count, - (int)objspace->rgengc.remembered_shady_object_count); +each_location(register const VALUE *x, register long n, void (*cb)(VALUE, void *), void *data) +{ + VALUE v; + while (n--) { + v = *x; + cb(v, data); + x++; } -#endif - - gc_event_hook(objspace, RUBY_INTERNAL_EVENT_GC_END_SWEEP, 0); } -static int -gc_heap_lazy_sweep(rb_objspace_t *objspace, rb_heap_t *heap) +static void +each_location_ptr(const VALUE *start, const VALUE *end, void (*cb)(VALUE, void *), void *data) { - struct heap_page *page = heap->sweep_pages, *next; - int result = FALSE; - - if (page == NULL) return FALSE; - -#if GC_ENABLE_LAZY_SWEEP - gc_prof_sweep_timer_start(objspace); -#endif - - while (page) { - heap->sweep_pages = next = page->next; - - gc_page_sweep(objspace, heap, page); - - if (!next) gc_after_sweep(objspace); - - if (heap->free_pages) { - result = TRUE; - break; - } - - page = next; - } - -#if GC_ENABLE_LAZY_SWEEP - gc_prof_sweep_timer_stop(objspace); -#endif - - return result; + if (end <= start) return; + each_location(start, end - start, cb, data); } static void -gc_heap_rest_sweep(rb_objspace_t *objspace, rb_heap_t *heap) +gc_mark_maybe_each_location(VALUE obj, void *data) { - if (is_lazy_sweeping(heap)) { - during_gc++; - while (is_lazy_sweeping(heap)) { - gc_heap_lazy_sweep(objspace, heap); - } - during_gc = 0; - } + gc_mark_maybe_internal(obj); } -static void -gc_rest_sweep(rb_objspace_t *objspace) +void +rb_gc_mark_locations(const VALUE *start, const VALUE *end) { - rb_heap_t *heap = heap_eden; /* lazy sweep only for eden */ - gc_heap_rest_sweep(objspace, heap); + each_location_ptr(start, end, gc_mark_maybe_each_location, NULL); } -static void -gc_sweep(rb_objspace_t *objspace, int immediate_sweep) +void +rb_gc_mark_values(long n, const VALUE *values) { - if (immediate_sweep) { -#if !GC_ENABLE_LAZY_SWEEP - gc_prof_sweep_timer_start(objspace); -#endif - gc_before_sweep(objspace); - gc_heap_rest_sweep(objspace, heap_eden); -#if !GC_ENABLE_LAZY_SWEEP - gc_prof_sweep_timer_stop(objspace); -#endif - } - else { - struct heap_page *page; - gc_before_sweep(objspace); - page = heap_eden->sweep_pages; - while (page) { - page->before_sweep = 1; - page = page->next; - } - gc_heap_lazy_sweep(objspace, heap_eden); + for (long i = 0; i < n; i++) { + gc_mark_internal(values[i]); } +} - gc_heap_prepare_minimum_pages(objspace, heap_eden); +void +rb_gc_mark_vm_stack_values(long n, const VALUE *values) +{ + for (long i = 0; i < n; i++) { + gc_mark_and_pin_internal(values[i]); + } } -/* Marking - Marking stack */ +static int +mark_key(st_data_t key, st_data_t value, st_data_t data) +{ + gc_mark_and_pin_internal((VALUE)key); -static void push_mark_stack(mark_stack_t *, VALUE); -static int pop_mark_stack(mark_stack_t *, VALUE *); -static void shrink_stack_chunk_cache(mark_stack_t *stack); + return ST_CONTINUE; +} -static stack_chunk_t * -stack_chunk_alloc(void) +void +rb_mark_set(st_table *tbl) { - stack_chunk_t *res; - - res = malloc(sizeof(stack_chunk_t)); - if (!res) - rb_memerror(); + if (!tbl) return; - return res; + st_foreach(tbl, mark_key, (st_data_t)rb_gc_get_objspace()); } -static inline int -is_mark_stack_empty(mark_stack_t *stack) +static int +mark_keyvalue(st_data_t key, st_data_t value, st_data_t data) { - return stack->chunk == NULL; + gc_mark_internal((VALUE)key); + gc_mark_internal((VALUE)value); + + return ST_CONTINUE; } -static void -add_stack_chunk_cache(mark_stack_t *stack, stack_chunk_t *chunk) +static int +pin_key_pin_value(st_data_t key, st_data_t value, st_data_t data) { - chunk->next = stack->cache; - stack->cache = chunk; - stack->cache_size++; + gc_mark_and_pin_internal((VALUE)key); + gc_mark_and_pin_internal((VALUE)value); + + return ST_CONTINUE; } -static void -shrink_stack_chunk_cache(mark_stack_t *stack) +static int +pin_key_mark_value(st_data_t key, st_data_t value, st_data_t data) { - stack_chunk_t *chunk; + gc_mark_and_pin_internal((VALUE)key); + gc_mark_internal((VALUE)value); - if (stack->unused_cache_size > (stack->cache_size/2)) { - chunk = stack->cache; - stack->cache = stack->cache->next; - stack->cache_size--; - free(chunk); - } - stack->unused_cache_size = stack->cache_size; + return ST_CONTINUE; } static void -push_mark_stack_chunk(mark_stack_t *stack) +mark_hash(VALUE hash) { - stack_chunk_t *next; - - assert(stack->index == stack->limit); - if (stack->cache_size > 0) { - next = stack->cache; - stack->cache = stack->cache->next; - stack->cache_size--; - if (stack->unused_cache_size > stack->cache_size) - stack->unused_cache_size = stack->cache_size; + if (rb_hash_compare_by_id_p(hash)) { + rb_hash_stlike_foreach(hash, pin_key_mark_value, 0); } else { - next = stack_chunk_alloc(); + rb_hash_stlike_foreach(hash, mark_keyvalue, 0); } - next->next = stack->chunk; - stack->chunk = next; - stack->index = 0; + + gc_mark_internal(RHASH(hash)->ifnone); } -static void -pop_mark_stack_chunk(mark_stack_t *stack) +void +rb_mark_hash(st_table *tbl) { - stack_chunk_t *prev; + if (!tbl) return; - prev = stack->chunk->next; - assert(stack->index == 0); - add_stack_chunk_cache(stack, stack->chunk); - stack->chunk = prev; - stack->index = stack->limit; + st_foreach(tbl, pin_key_pin_value, 0); } -#if defined(ENABLE_VM_OBJSPACE) && ENABLE_VM_OBJSPACE -static void -free_stack_chunks(mark_stack_t *stack) +static enum rb_id_table_iterator_result +mark_method_entry_i(VALUE me, void *objspace) { - stack_chunk_t *chunk = stack->chunk; - stack_chunk_t *next = NULL; + gc_mark_internal(me); - while (chunk != NULL) { - next = chunk->next; - free(chunk); - chunk = next; - } + return ID_TABLE_CONTINUE; } -#endif static void -push_mark_stack(mark_stack_t *stack, VALUE data) +mark_m_tbl(void *objspace, struct rb_id_table *tbl) { - if (stack->index == stack->limit) { - push_mark_stack_chunk(stack); + if (tbl) { + rb_id_table_foreach_values(tbl, mark_method_entry_i, objspace); } - stack->chunk->data[stack->index++] = data; } -static int -pop_mark_stack(mark_stack_t *stack, VALUE *data) +static enum rb_id_table_iterator_result +mark_const_entry_i(VALUE value, void *objspace) { - if (is_mark_stack_empty(stack)) { - return FALSE; - } - if (stack->index == 1) { - *data = stack->chunk->data[--stack->index]; - pop_mark_stack_chunk(stack); - } - else { - *data = stack->chunk->data[--stack->index]; + const rb_const_entry_t *ce = (const rb_const_entry_t *)value; + + if (!rb_gc_checking_shareable()) { + gc_mark_internal(ce->value); + gc_mark_internal(ce->file); // TODO: ce->file should be shareable? } - return TRUE; + return ID_TABLE_CONTINUE; } static void -init_mark_stack(mark_stack_t *stack) +mark_const_tbl(rb_objspace_t *objspace, struct rb_id_table *tbl) { - int i; - - if (0) push_mark_stack_chunk(stack); - stack->index = stack->limit = STACK_CHUNK_SIZE; - - for (i=0; i < 4; i++) { - add_stack_chunk_cache(stack, stack_chunk_alloc()); - } - stack->unused_cache_size = stack->cache_size; + if (!tbl) return; + rb_id_table_foreach_values(tbl, mark_const_entry_i, objspace); } -/* Marking */ - -#ifdef __ia64 -#define SET_STACK_END (SET_MACHINE_STACK_END(&th->machine.stack_end), th->machine.register_stack_end = rb_ia64_bsp()) -#else -#define SET_STACK_END SET_MACHINE_STACK_END(&th->machine.stack_end) -#endif - -#define STACK_START (th->machine.stack_start) -#define STACK_END (th->machine.stack_end) -#define STACK_LEVEL_MAX (th->machine.stack_maxsize/sizeof(VALUE)) - #if STACK_GROW_DIRECTION < 0 -# define STACK_LENGTH (size_t)(STACK_START - STACK_END) +#define GET_STACK_BOUNDS(start, end, appendix) ((start) = STACK_END, (end) = STACK_START) #elif STACK_GROW_DIRECTION > 0 -# define STACK_LENGTH (size_t)(STACK_END - STACK_START + 1) +#define GET_STACK_BOUNDS(start, end, appendix) ((start) = STACK_START, (end) = STACK_END+(appendix)) #else -# define STACK_LENGTH ((STACK_END < STACK_START) ? (size_t)(STACK_START - STACK_END) \ - : (size_t)(STACK_END - STACK_START + 1)) +#define GET_STACK_BOUNDS(start, end, appendix) \ + ((STACK_END < STACK_START) ? \ + ((start) = STACK_END, (end) = STACK_START) : ((start) = STACK_START, (end) = STACK_END+(appendix))) #endif -#if !STACK_GROW_DIRECTION -int ruby_stack_grow_direction; -int -ruby_get_stack_grow_direction(volatile VALUE *addr) -{ - VALUE *end; - SET_MACHINE_STACK_END(&end); - if (end > addr) return ruby_stack_grow_direction = 1; - return ruby_stack_grow_direction = -1; -} +static void +gc_mark_machine_stack_location_maybe(VALUE obj, void *data) +{ + gc_mark_maybe_internal(obj); + +#ifdef RUBY_ASAN_ENABLED + const rb_execution_context_t *ec = (const rb_execution_context_t *)data; + void *fake_frame_start; + void *fake_frame_end; + bool is_fake_frame = asan_get_fake_stack_extents( + ec->machine.asan_fake_stack_handle, obj, + ec->machine.stack_start, ec->machine.stack_end, + &fake_frame_start, &fake_frame_end + ); + if (is_fake_frame) { + each_location_ptr(fake_frame_start, fake_frame_end, gc_mark_maybe_each_location, NULL); + } #endif - -size_t -ruby_stack_length(VALUE **p) -{ - rb_thread_t *th = GET_THREAD(); - SET_STACK_END; - if (p) *p = STACK_UPPER(STACK_END, STACK_START, STACK_END); - return STACK_LENGTH; } -#if !(defined(POSIX_SIGNAL) && defined(SIGSEGV) && defined(HAVE_SIGALTSTACK)) -static int -stack_check(int water_mark) +static bool +gc_object_moved_p_internal(void *objspace, VALUE obj) { - int ret; - rb_thread_t *th = GET_THREAD(); - SET_STACK_END; - ret = STACK_LENGTH > STACK_LEVEL_MAX - water_mark; -#ifdef __ia64 - if (!ret) { - ret = (VALUE*)rb_ia64_bsp() - th->machine.register_stack_start > - th->machine.register_stack_maxsize/sizeof(VALUE) - water_mark; + if (SPECIAL_CONST_P(obj)) { + return false; } -#endif - return ret; -} -#endif - -#define STACKFRAME_FOR_CALL_CFUNC 512 -int -ruby_stack_check(void) -{ -#if defined(POSIX_SIGNAL) && defined(SIGSEGV) && defined(HAVE_SIGALTSTACK) - return 0; -#else - return stack_check(STACKFRAME_FOR_CALL_CFUNC); -#endif + return rb_gc_impl_object_moved_p(objspace, obj); } -ATTRIBUTE_NO_ADDRESS_SAFETY_ANALYSIS -static void -mark_locations_array(rb_objspace_t *objspace, register VALUE *x, register long n) +static VALUE +gc_location_internal(void *objspace, VALUE value) { - VALUE v; - while (n--) { - v = *x; - gc_mark_maybe(objspace, v); - x++; + if (SPECIAL_CONST_P(value)) { + return value; } -} -static void -gc_mark_locations(rb_objspace_t *objspace, VALUE *start, VALUE *end) -{ - long n; + GC_ASSERT(rb_gc_impl_pointer_to_heap_p(objspace, (void *)value)); - if (end <= start) return; - n = end - start; - mark_locations_array(objspace, start, n); + return rb_gc_impl_location(objspace, value); } -void -rb_gc_mark_locations(VALUE *start, VALUE *end) +VALUE +rb_gc_location(VALUE value) { - gc_mark_locations(&rb_objspace, start, end); + return gc_location_internal(rb_gc_get_objspace(), value); } -#define rb_gc_mark_locations(start, end) gc_mark_locations(objspace, (start), (end)) +#if defined(__wasm__) -struct mark_tbl_arg { - rb_objspace_t *objspace; -}; -static int -mark_entry(st_data_t key, st_data_t value, st_data_t data) -{ - struct mark_tbl_arg *arg = (void*)data; - gc_mark(arg->objspace, (VALUE)value); - return ST_CONTINUE; -} +static VALUE *rb_stack_range_tmp[2]; static void -mark_tbl(rb_objspace_t *objspace, st_table *tbl) +rb_mark_locations(void *begin, void *end) { - struct mark_tbl_arg arg; - if (!tbl || tbl->num_entries == 0) return; - arg.objspace = objspace; - st_foreach(tbl, mark_entry, (st_data_t)&arg); + rb_stack_range_tmp[0] = begin; + rb_stack_range_tmp[1] = end; } -static int -mark_key(st_data_t key, st_data_t value, st_data_t data) +void +rb_gc_save_machine_context(void) { - struct mark_tbl_arg *arg = (void*)data; - gc_mark(arg->objspace, (VALUE)key); - return ST_CONTINUE; + // no-op } -static void -mark_set(rb_objspace_t *objspace, st_table *tbl) -{ - struct mark_tbl_arg arg; - if (!tbl) return; - arg.objspace = objspace; - st_foreach(tbl, mark_key, (st_data_t)&arg); -} +# if defined(__EMSCRIPTEN__) -void -rb_mark_set(st_table *tbl) +static void +mark_current_machine_context(const rb_execution_context_t *ec) { - mark_set(&rb_objspace, tbl); -} + emscripten_scan_stack(rb_mark_locations); + each_location_ptr(rb_stack_range_tmp[0], rb_stack_range_tmp[1], gc_mark_maybe_each_location, NULL); -static int -mark_keyvalue(st_data_t key, st_data_t value, st_data_t data) -{ - struct mark_tbl_arg *arg = (void*)data; - gc_mark(arg->objspace, (VALUE)key); - gc_mark(arg->objspace, (VALUE)value); - return ST_CONTINUE; + emscripten_scan_registers(rb_mark_locations); + each_location_ptr(rb_stack_range_tmp[0], rb_stack_range_tmp[1], gc_mark_maybe_each_location, NULL); } +# else // use Asyncify version static void -mark_hash(rb_objspace_t *objspace, st_table *tbl) +mark_current_machine_context(rb_execution_context_t *ec) { - struct mark_tbl_arg arg; - if (!tbl) return; - arg.objspace = objspace; - st_foreach(tbl, mark_keyvalue, (st_data_t)&arg); -} + VALUE *stack_start, *stack_end; + SET_STACK_END; + GET_STACK_BOUNDS(stack_start, stack_end, 1); + each_location_ptr(stack_start, stack_end, gc_mark_maybe_each_location, NULL); -void -rb_mark_hash(st_table *tbl) -{ - mark_hash(&rb_objspace, tbl); + rb_wasm_scan_locals(rb_mark_locations); + each_location_ptr(rb_stack_range_tmp[0], rb_stack_range_tmp[1], gc_mark_maybe_each_location, NULL); } -static void -mark_method_entry(rb_objspace_t *objspace, const rb_method_entry_t *me) -{ - const rb_method_definition_t *def = me->def; - - gc_mark(objspace, me->klass); - again: - if (!def) return; - switch (def->type) { - case VM_METHOD_TYPE_ISEQ: - gc_mark(objspace, def->body.iseq->self); - break; - case VM_METHOD_TYPE_BMETHOD: - gc_mark(objspace, def->body.proc); - break; - case VM_METHOD_TYPE_ATTRSET: - case VM_METHOD_TYPE_IVAR: - gc_mark(objspace, def->body.attr.location); - break; - case VM_METHOD_TYPE_REFINED: - if (def->body.orig_me) { - def = def->body.orig_me->def; - goto again; - } - break; - default: - break; /* ignore */ - } -} +# endif + +#else // !defined(__wasm__) void -rb_mark_method_entry(const rb_method_entry_t *me) +rb_gc_save_machine_context(void) { - mark_method_entry(&rb_objspace, me); -} + rb_thread_t *thread = GET_THREAD(); -static int -mark_method_entry_i(ID key, const rb_method_entry_t *me, st_data_t data) -{ - struct mark_tbl_arg *arg = (void*)data; - mark_method_entry(arg->objspace, me); - return ST_CONTINUE; + RB_VM_SAVE_MACHINE_CONTEXT(thread); } -static void -mark_m_tbl_wrapper(rb_objspace_t *objspace, struct method_table_wrapper *wrapper) -{ - struct mark_tbl_arg arg; - if (!wrapper || !wrapper->tbl) return; - if (LIKELY(objspace->mark_func_data == 0)) { - /* prevent multiple marking during same GC cycle, - * since m_tbl is shared between several T_ICLASS */ - size_t serial = rb_gc_count(); - if (wrapper->serial == serial) return; - wrapper->serial = serial; - } - arg.objspace = objspace; - st_foreach(wrapper->tbl, mark_method_entry_i, (st_data_t)&arg); -} - -static int -mark_const_entry_i(ID key, const rb_const_entry_t *ce, st_data_t data) -{ - struct mark_tbl_arg *arg = (void*)data; - gc_mark(arg->objspace, ce->value); - gc_mark(arg->objspace, ce->file); - return ST_CONTINUE; -} static void -mark_const_tbl(rb_objspace_t *objspace, st_table *tbl) +mark_current_machine_context(const rb_execution_context_t *ec) { - struct mark_tbl_arg arg; - if (!tbl) return; - arg.objspace = objspace; - st_foreach(tbl, mark_const_entry_i, (st_data_t)&arg); + rb_gc_mark_machine_context(ec); } - -#if STACK_GROW_DIRECTION < 0 -#define GET_STACK_BOUNDS(start, end, appendix) ((start) = STACK_END, (end) = STACK_START) -#elif STACK_GROW_DIRECTION > 0 -#define GET_STACK_BOUNDS(start, end, appendix) ((start) = STACK_START, (end) = STACK_END+(appendix)) -#else -#define GET_STACK_BOUNDS(start, end, appendix) \ - ((STACK_END < STACK_START) ? \ - ((start) = STACK_END, (end) = STACK_START) : ((start) = STACK_START, (end) = STACK_END+(appendix))) #endif -static void -mark_current_machine_context(rb_objspace_t *objspace, rb_thread_t *th) +void +rb_gc_mark_machine_context(const rb_execution_context_t *ec) { - union { - rb_jmp_buf j; - VALUE v[sizeof(rb_jmp_buf) / sizeof(VALUE)]; - } save_regs_gc_mark; VALUE *stack_start, *stack_end; - FLUSH_REGISTER_WINDOWS; - /* This assumes that all registers are saved into the jmp_buf (and stack) */ - rb_setjmp(save_regs_gc_mark.j); - - /* SET_STACK_END must be called in this function because - * the stack frame of this function may contain - * callee save registers and they should be marked. */ - SET_STACK_END; - GET_STACK_BOUNDS(stack_start, stack_end, 1); - - mark_locations_array(objspace, save_regs_gc_mark.v, numberof(save_regs_gc_mark.v)); + GET_STACK_BOUNDS(stack_start, stack_end, 0); + RUBY_DEBUG_LOG("ec->th:%u stack_start:%p stack_end:%p", rb_ec_thread_ptr(ec)->serial, stack_start, stack_end); - rb_gc_mark_locations(stack_start, stack_end); -#ifdef __ia64 - rb_gc_mark_locations(th->machine.register_stack_start, th->machine.register_stack_end); -#endif -#if defined(__mc68000__) - mark_locations_array(objspace, (VALUE*)((char*)STACK_END + 2), - (STACK_START - STACK_END)); + void *data = +#ifdef RUBY_ASAN_ENABLED + /* gc_mark_machine_stack_location_maybe() uses data as const */ + (rb_execution_context_t *)ec; +#else + NULL; #endif + + each_location_ptr(stack_start, stack_end, gc_mark_machine_stack_location_maybe, data); + int num_regs = sizeof(ec->machine.regs)/(sizeof(VALUE)); + each_location((VALUE*)&ec->machine.regs, num_regs, gc_mark_machine_stack_location_maybe, data); } -void -rb_gc_mark_machine_stack(rb_thread_t *th) +static int +rb_mark_tbl_i(st_data_t key, st_data_t value, st_data_t data) { - rb_objspace_t *objspace = &rb_objspace; - VALUE *stack_start, *stack_end; + gc_mark_and_pin_internal((VALUE)value); - GET_STACK_BOUNDS(stack_start, stack_end, 0); - rb_gc_mark_locations(stack_start, stack_end); -#ifdef __ia64 - rb_gc_mark_locations(th->machine.register_stack_start, th->machine.register_stack_end); -#endif + return ST_CONTINUE; } void rb_mark_tbl(st_table *tbl) { - mark_tbl(&rb_objspace, tbl); + if (!tbl || tbl->num_entries == 0) return; + + st_foreach(tbl, rb_mark_tbl_i, 0); } static void -gc_mark_maybe(rb_objspace_t *objspace, VALUE obj) +gc_mark_tbl_no_pin(st_table *tbl) { - (void)VALGRIND_MAKE_MEM_DEFINED(&obj, sizeof(obj)); - if (is_pointer_to_heap(objspace, (void *)obj)) { - int type = BUILTIN_TYPE(obj); - if (type != T_ZOMBIE && type != T_NONE) { - gc_mark(objspace, obj); - } - } + if (!tbl || tbl->num_entries == 0) return; + + st_foreach(tbl, gc_mark_tbl_no_pin_i, 0); } void -rb_gc_mark_maybe(VALUE obj) +rb_mark_tbl_no_pin(st_table *tbl) { - gc_mark_maybe(&rb_objspace, obj); + gc_mark_tbl_no_pin(tbl); } -static inline int -gc_marked(rb_objspace_t *objspace, VALUE ptr) +void +rb_gc_mark_set_no_pin(st_table *tbl) { - register bits_t *bits = GET_HEAP_MARK_BITS(ptr); - if (MARKED_IN_BITMAP(bits, ptr)) return 1; - return 0; -} + if (!tbl || tbl->num_entries == 0) return; -static inline int -gc_mark_ptr(rb_objspace_t *objspace, VALUE ptr) -{ - register bits_t *bits = GET_HEAP_MARK_BITS(ptr); - if (gc_marked(objspace, ptr)) return 0; - MARK_IN_BITMAP(bits, ptr); - return 1; + st_foreach(tbl, gc_mark_set_no_pin_i, 0); } -static void -rgengc_check_relation(rb_objspace_t *objspace, VALUE obj) -{ -#if USE_RGENGC - if (objspace->rgengc.parent_object_is_old) { - if (!RVALUE_WB_PROTECTED(obj)) { - if (rgengc_remember(objspace, obj)) { - objspace->rgengc.remembered_shady_object_count++; - } - } -#if RGENGC_THREEGEN - else { - if (gc_marked(objspace, obj)) { - if (!RVALUE_OLD_P(obj)) { - /* An object pointed from an OLD object should be OLD. */ - rgengc_remember(objspace, obj); - } - } - else { - if (RVALUE_INFANT_P(obj)) { - RVALUE_PROMOTE_INFANT(obj); - } - } - } -#endif - } -#endif +static bool +gc_declarative_marking_p(const rb_data_type_t *type) +{ + return (type->flags & RUBY_TYPED_DECL_MARKING) != 0; } -static void -gc_mark(rb_objspace_t *objspace, VALUE ptr) +rb_execution_context_t * +rb_gc_get_ec(void) { - if (!is_markable_object(objspace, ptr)) return; + void *objspace = rb_gc_get_objspace(); - if (LIKELY(objspace->mark_func_data == 0)) { - rgengc_check_relation(objspace, ptr); - if (!gc_mark_ptr(objspace, ptr)) return; /* already marked */ - push_mark_stack(&objspace->mark_stack, ptr); + if (RB_LIKELY(rb_gc_impl_during_gc_p(objspace))) { + return rb_gc_impl_get_vm_context(objspace)->ec; } else { - objspace->mark_func_data->mark_func(ptr, objspace->mark_func_data->data); + return GET_EC(); } } void -rb_gc_mark(VALUE ptr) +rb_gc_mark_roots(void *objspace, const char **categoryp) { - gc_mark(&rb_objspace, ptr); -} + rb_execution_context_t *ec = rb_gc_get_ec(); + rb_vm_t *vm = rb_ec_vm_ptr(ec); -/* resurrect non-marked `obj' if obj is before swept */ +#define MARK_CHECKPOINT(category) do { \ + if (categoryp) *categoryp = category; \ +} while (0) -void -rb_gc_resurrect(VALUE obj) -{ - rb_objspace_t *objspace = &rb_objspace; + MARK_CHECKPOINT("vm"); + rb_vm_mark(vm); + + MARK_CHECKPOINT("end_proc"); + rb_mark_end_proc(); + + MARK_CHECKPOINT("global_tbl"); + rb_gc_mark_global_tbl(); + +#if USE_YJIT + void rb_yjit_root_mark(void); // in Rust + + if (rb_yjit_enabled_p) { + MARK_CHECKPOINT("YJIT"); + rb_yjit_root_mark(); + } +#endif - if (is_lazy_sweeping(heap_eden) && - !gc_marked(objspace, obj) && - !is_swept_object(objspace, obj)) { - gc_mark_ptr(objspace, obj); +#if USE_ZJIT + void rb_zjit_root_mark(void); + if (rb_zjit_enabled_p) { + MARK_CHECKPOINT("ZJIT"); + rb_zjit_root_mark(); } +#endif + + MARK_CHECKPOINT("machine_context"); + mark_current_machine_context(ec); + + MARK_CHECKPOINT("global_symbols"); + rb_sym_global_symbols_mark_and_move(); + + MARK_CHECKPOINT("finish"); + +#undef MARK_CHECKPOINT } +struct gc_mark_classext_foreach_arg { + rb_objspace_t *objspace; + VALUE obj; +}; + static void -gc_mark_children(rb_objspace_t *objspace, VALUE ptr) +gc_mark_classext_module(rb_classext_t *ext, bool prime, VALUE box_value, void *arg) { - register RVALUE *obj = RANY(ptr); + struct gc_mark_classext_foreach_arg *foreach_arg = (struct gc_mark_classext_foreach_arg *)arg; + rb_objspace_t *objspace = foreach_arg->objspace; + + if (RCLASSEXT_SUPER(ext)) { + gc_mark_internal(RCLASSEXT_SUPER(ext)); + } + mark_m_tbl(objspace, RCLASSEXT_M_TBL(ext)); - goto marking; /* skip */ + if (!rb_gc_checking_shareable()) { + // unshareable + gc_mark_internal(RCLASSEXT_FIELDS_OBJ(ext)); + gc_mark_internal(RCLASSEXT_CVC_TBL(ext)); + } - again: - if (LIKELY(objspace->mark_func_data == 0)) { - obj = RANY(ptr); - if (!is_markable_object(objspace, ptr)) return; - rgengc_check_relation(objspace, ptr); - if (!gc_mark_ptr(objspace, ptr)) return; /* already marked */ + if (!RCLASSEXT_SHARED_CONST_TBL(ext) && RCLASSEXT_CONST_TBL(ext)) { + mark_const_tbl(objspace, RCLASSEXT_CONST_TBL(ext)); } - else { - gc_mark(objspace, ptr); - return; + mark_m_tbl(objspace, RCLASSEXT_CALLABLE_M_TBL(ext)); + gc_mark_internal(RCLASSEXT_CC_TBL(ext)); + if (RCLASSEXT_SUBCLASSES(ext)) { + gc_mark_internal(RCLASSEXT_SUBCLASSES(ext)); } + gc_mark_internal(RCLASSEXT_CLASSPATH(ext)); +} - marking: +static void +gc_mark_classext_iclass(rb_classext_t *ext, bool prime, VALUE box_value, void *arg) +{ + struct gc_mark_classext_foreach_arg *foreach_arg = (struct gc_mark_classext_foreach_arg *)arg; + rb_objspace_t *objspace = foreach_arg->objspace; -#if USE_RGENGC - check_gen_consistency((VALUE)obj); + if (RCLASSEXT_SUPER(ext)) { + gc_mark_internal(RCLASSEXT_SUPER(ext)); + } + if (RCLASSEXT_ICLASS_IS_ORIGIN(ext) && !RCLASSEXT_ICLASS_ORIGIN_SHARED_MTBL(ext)) { + mark_m_tbl(objspace, RCLASSEXT_M_TBL(ext)); + } + if (RCLASSEXT_INCLUDER(ext)) { + gc_mark_internal(RCLASSEXT_INCLUDER(ext)); + } + mark_m_tbl(objspace, RCLASSEXT_CALLABLE_M_TBL(ext)); + gc_mark_internal(RCLASSEXT_CC_TBL(ext)); + if (RCLASSEXT_SUBCLASSES(ext)) { + gc_mark_internal(RCLASSEXT_SUBCLASSES(ext)); + } +} - if (LIKELY(objspace->mark_func_data == 0)) { - /* minor/major common */ - if (RVALUE_WB_PROTECTED(obj)) { - if (RVALUE_INFANT_P((VALUE)obj)) { - /* infant -> young */ - RVALUE_PROMOTE_INFANT((VALUE)obj); -#if RGENGC_THREEGEN - /* infant -> young */ - objspace->rgengc.young_object_count++; - objspace->rgengc.parent_object_is_old = FALSE; -#else - /* infant -> old */ - objspace->rgengc.old_object_count++; - objspace->rgengc.parent_object_is_old = TRUE; -#endif - rgengc_report(3, objspace, "gc_mark_children: promote infant -> young %p (%s).\n", (void *)obj, obj_type_name((VALUE)obj)); - } - else { - objspace->rgengc.parent_object_is_old = TRUE; - -#if RGENGC_THREEGEN - if (RVALUE_YOUNG_P((VALUE)obj)) { - /* young -> old */ - RVALUE_PROMOTE_YOUNG((VALUE)obj); - objspace->rgengc.old_object_count++; - rgengc_report(3, objspace, "gc_mark_children: promote young -> old %p (%s).\n", (void *)obj, obj_type_name((VALUE)obj)); - } - else { -#endif - if (!objspace->rgengc.during_minor_gc) { - /* major/full GC */ - objspace->rgengc.old_object_count++; - } -#if RGENGC_THREEGEN - } -#endif - } - } - else { - rgengc_report(3, objspace, "gc_mark_children: do not promote non-WB-protected %p (%s).\n", (void *)obj, obj_type_name((VALUE)obj)); - objspace->rgengc.parent_object_is_old = FALSE; - } +#define TYPED_DATA_REFS_OFFSET_LIST(d) (size_t *)(uintptr_t)RTYPEDDATA_TYPE(d)->function.dmark + +void +rb_gc_move_obj_during_marking(VALUE from, VALUE to) +{ + if (rb_obj_using_gen_fields_table_p(to)) { + rb_mark_generic_ivar(from); } +} - check_gen_consistency((VALUE)obj); -#endif /* USE_RGENGC */ +void +rb_gc_mark_children(void *objspace, VALUE obj) +{ + struct gc_mark_classext_foreach_arg foreach_args; - if (FL_TEST(obj, FL_EXIVAR)) { - rb_mark_generic_ivar(ptr); + if (rb_obj_using_gen_fields_table_p(obj)) { + rb_mark_generic_ivar(obj); } switch (BUILTIN_TYPE(obj)) { + case T_FLOAT: + case T_BIGNUM: + return; + case T_NIL: case T_FIXNUM: - rb_bug("rb_gc_mark() called for broken object"); - break; + rb_bug("rb_gc_mark() called for broken object"); + break; case T_NODE: - switch (nd_type(obj)) { - case NODE_IF: /* 1,2,3 */ - case NODE_FOR: - case NODE_ITER: - case NODE_WHEN: - case NODE_MASGN: - case NODE_RESCUE: - case NODE_RESBODY: - case NODE_CLASS: - case NODE_BLOCK_PASS: - gc_mark(objspace, (VALUE)obj->as.node.u2.node); - /* fall through */ - case NODE_BLOCK: /* 1,3 */ - case NODE_ARRAY: - case NODE_DSTR: - case NODE_DXSTR: - case NODE_DREGX: - case NODE_DREGX_ONCE: - case NODE_ENSURE: - case NODE_CALL: - case NODE_DEFS: - case NODE_OP_ASGN1: - gc_mark(objspace, (VALUE)obj->as.node.u1.node); - /* fall through */ - case NODE_SUPER: /* 3 */ - case NODE_FCALL: - case NODE_DEFN: - case NODE_ARGS_AUX: - ptr = (VALUE)obj->as.node.u3.node; - goto again; - - case NODE_WHILE: /* 1,2 */ - case NODE_UNTIL: - case NODE_AND: - case NODE_OR: - case NODE_CASE: - case NODE_SCLASS: - case NODE_DOT2: - case NODE_DOT3: - case NODE_FLIP2: - case NODE_FLIP3: - case NODE_MATCH2: - case NODE_MATCH3: - case NODE_OP_ASGN_OR: - case NODE_OP_ASGN_AND: - case NODE_MODULE: - case NODE_ALIAS: - case NODE_VALIAS: - case NODE_ARGSCAT: - gc_mark(objspace, (VALUE)obj->as.node.u1.node); - /* fall through */ - case NODE_GASGN: /* 2 */ - case NODE_LASGN: - case NODE_DASGN: - case NODE_DASGN_CURR: - case NODE_IASGN: - case NODE_IASGN2: - case NODE_CVASGN: - case NODE_COLON3: - case NODE_OPT_N: - case NODE_EVSTR: - case NODE_UNDEF: - case NODE_POSTEXE: - ptr = (VALUE)obj->as.node.u2.node; - goto again; - - case NODE_HASH: /* 1 */ - case NODE_LIT: - case NODE_STR: - case NODE_XSTR: - case NODE_DEFINED: - case NODE_MATCH: - case NODE_RETURN: - case NODE_BREAK: - case NODE_NEXT: - case NODE_YIELD: - case NODE_COLON2: - case NODE_SPLAT: - case NODE_TO_ARY: - ptr = (VALUE)obj->as.node.u1.node; - goto again; - - case NODE_SCOPE: /* 2,3 */ - case NODE_CDECL: - case NODE_OPT_ARG: - gc_mark(objspace, (VALUE)obj->as.node.u3.node); - ptr = (VALUE)obj->as.node.u2.node; - goto again; - - case NODE_ARGS: /* custom */ - { - struct rb_args_info *args = obj->as.node.u3.args; - if (args) { - if (args->pre_init) gc_mark(objspace, (VALUE)args->pre_init); - if (args->post_init) gc_mark(objspace, (VALUE)args->post_init); - if (args->opt_args) gc_mark(objspace, (VALUE)args->opt_args); - if (args->kw_args) gc_mark(objspace, (VALUE)args->kw_args); - if (args->kw_rest_arg) gc_mark(objspace, (VALUE)args->kw_rest_arg); - } - } - ptr = (VALUE)obj->as.node.u2.node; - goto again; - - case NODE_ZARRAY: /* - */ - case NODE_ZSUPER: - case NODE_VCALL: - case NODE_GVAR: - case NODE_LVAR: - case NODE_DVAR: - case NODE_IVAR: - case NODE_CVAR: - case NODE_NTH_REF: - case NODE_BACK_REF: - case NODE_REDO: - case NODE_RETRY: - case NODE_SELF: - case NODE_NIL: - case NODE_TRUE: - case NODE_FALSE: - case NODE_ERRINFO: - case NODE_BLOCK_ARG: - break; - case NODE_ALLOCA: - mark_locations_array(objspace, - (VALUE*)obj->as.node.u1.value, - obj->as.node.u3.cnt); - gc_mark(objspace, (VALUE)obj->as.node.u2.node); - break; - - case NODE_CREF: - gc_mark(objspace, obj->as.node.nd_refinements); - gc_mark(objspace, (VALUE)obj->as.node.nd_clss); - ptr = (VALUE)obj->as.node.nd_next; - goto again; - - default: /* unlisted NODE */ - gc_mark_maybe(objspace, (VALUE)obj->as.node.u1.node); - gc_mark_maybe(objspace, (VALUE)obj->as.node.u2.node); - gc_mark_maybe(objspace, (VALUE)obj->as.node.u3.node); - } - return; /* no need to mark class. */ - } - - gc_mark(objspace, obj->as.basic.klass); + UNEXPECTED_NODE(rb_gc_mark); + break; + + case T_IMEMO: + rb_imemo_mark_and_move(obj, false); + return; + + default: + break; + } + + gc_mark_internal(RBASIC(obj)->klass); + switch (BUILTIN_TYPE(obj)) { - case T_ICLASS: case T_CLASS: + if (FL_TEST_RAW(obj, FL_SINGLETON) && + !rb_gc_checking_shareable()) { + gc_mark_internal(RCLASS_ATTACHED_OBJECT(obj)); + } + // Continue to the shared T_CLASS/T_MODULE case T_MODULE: - mark_m_tbl_wrapper(objspace, RCLASS_M_TBL_WRAPPER(obj)); - if (!RCLASS_EXT(obj)) break; - mark_tbl(objspace, RCLASS_IV_TBL(obj)); - mark_const_tbl(objspace, RCLASS_CONST_TBL(obj)); - ptr = RCLASS_SUPER((VALUE)obj); - goto again; + foreach_args.objspace = objspace; + foreach_args.obj = obj; + rb_class_classext_foreach(obj, gc_mark_classext_module, (void *)&foreach_args); + if (BOX_USER_P(RCLASS_PRIME_BOX(obj))) { + gc_mark_internal(RCLASS_PRIME_BOX(obj)->box_object); + } + break; + + case T_ICLASS: + foreach_args.objspace = objspace; + foreach_args.obj = obj; + rb_class_classext_foreach(obj, gc_mark_classext_iclass, (void *)&foreach_args); + if (BOX_USER_P(RCLASS_PRIME_BOX(obj))) { + gc_mark_internal(RCLASS_PRIME_BOX(obj)->box_object); + } + break; case T_ARRAY: - if (FL_TEST(obj, ELTS_SHARED)) { - ptr = obj->as.array.as.heap.aux.shared; - goto again; - } - else { - long i, len = RARRAY_LEN(obj); - const VALUE *ptr = RARRAY_CONST_PTR(obj); - for (i=0; i < len; i++) { - gc_mark(objspace, *ptr++); - } - } - break; + if (ARY_SHARED_P(obj)) { + VALUE root = ARY_SHARED_ROOT(obj); + gc_mark_internal(root); + } + else { + long len = RARRAY_LEN(obj); + const VALUE *ptr = RARRAY_CONST_PTR(obj); + for (long i = 0; i < len; i++) { + gc_mark_internal(ptr[i]); + } + } + break; case T_HASH: - mark_hash(objspace, obj->as.hash.ntbl); - ptr = obj->as.hash.ifnone; - goto again; + mark_hash(obj); + break; + + case T_SYMBOL: + gc_mark_internal(RSYMBOL(obj)->fstr); + break; case T_STRING: -#define STR_ASSOC FL_USER3 /* copied from string.c */ - if (FL_TEST(obj, RSTRING_NOEMBED) && FL_ANY(obj, ELTS_SHARED|STR_ASSOC)) { - ptr = obj->as.string.as.heap.aux.shared; - goto again; - } - break; + if (STR_SHARED_P(obj)) { + if (STR_EMBED_P(RSTRING(obj)->as.heap.aux.shared)) { + /* Embedded shared strings cannot be moved because this string + * points into the slot of the shared string. There may be code + * using the RSTRING_PTR on the stack, which would pin this + * string but not pin the shared string, causing it to move. */ + gc_mark_and_pin_internal(RSTRING(obj)->as.heap.aux.shared); + } + else { + gc_mark_internal(RSTRING(obj)->as.heap.aux.shared); + } + } + break; - case T_DATA: - if (RTYPEDDATA_P(obj)) { - RUBY_DATA_FUNC mark_func = obj->as.typeddata.type->function.dmark; - if (mark_func) (*mark_func)(DATA_PTR(obj)); - } - else { - if (obj->as.data.dmark) (*obj->as.data.dmark)(DATA_PTR(obj)); - } - break; + case T_DATA: { + void *const ptr = RTYPEDDATA_GET_DATA(obj); - case T_OBJECT: - { - long i, len = ROBJECT_NUMIV(obj); - VALUE *ptr = ROBJECT_IVPTR(obj); - for (i = 0; i < len; i++) { - gc_mark(objspace, *ptr++); + gc_mark_internal(RTYPEDDATA(obj)->fields_obj); + + if (ptr) { + if (gc_declarative_marking_p(RTYPEDDATA_TYPE(obj))) { + size_t *offset_list = TYPED_DATA_REFS_OFFSET_LIST(obj); + + for (size_t offset = *offset_list; offset != RUBY_REF_END; offset = *offset_list++) { + gc_mark_internal(*(VALUE *)((char *)ptr + offset)); + } + } + else { + RUBY_DATA_FUNC mark_func = RTYPEDDATA_TYPE(obj)->function.dmark; + if (mark_func) (*mark_func)(ptr); + } + } + + break; + } + + case T_OBJECT: { + uint32_t len; + if (rb_obj_shape_complex_p(obj)) { + gc_mark_tbl_no_pin(ROBJECT_FIELDS_HASH(obj)); + len = ROBJECT_FIELDS_COUNT_COMPLEX(obj); + } + else { + const VALUE * const ptr = ROBJECT_FIELDS(obj); + + len = ROBJECT_FIELDS_COUNT_NOT_COMPLEX(obj); + for (uint32_t i = 0; i < len; i++) { + gc_mark_internal(ptr[i]); } } - break; + break; + } case T_FILE: - if (obj->as.file.fptr) { - gc_mark(objspace, obj->as.file.fptr->pathv); - gc_mark(objspace, obj->as.file.fptr->tied_io_for_writing); - gc_mark(objspace, obj->as.file.fptr->writeconv_asciicompat); - gc_mark(objspace, obj->as.file.fptr->writeconv_pre_ecopts); - gc_mark(objspace, obj->as.file.fptr->encs.ecopts); - gc_mark(objspace, obj->as.file.fptr->write_lock); + if (RFILE(obj)->fptr) { + gc_mark_internal(RFILE(obj)->fptr->self); + gc_mark_internal(RFILE(obj)->fptr->pathv); + gc_mark_internal(RFILE(obj)->fptr->tied_io_for_writing); + gc_mark_internal(RFILE(obj)->fptr->writeconv_asciicompat); + gc_mark_internal(RFILE(obj)->fptr->writeconv_pre_ecopts); + gc_mark_internal(RFILE(obj)->fptr->encs.ecopts); + gc_mark_internal(RFILE(obj)->fptr->write_lock); + gc_mark_internal(RFILE(obj)->fptr->timeout); + gc_mark_internal(RFILE(obj)->fptr->wakeup_mutex); } break; case T_REGEXP: - ptr = obj->as.regexp.src; - goto again; - - case T_FLOAT: - case T_BIGNUM: - break; + gc_mark_internal(RREGEXP(obj)->src); + break; case T_MATCH: - gc_mark(objspace, obj->as.match.regexp); - if (obj->as.match.str) { - ptr = obj->as.match.str; - goto again; - } - break; + gc_mark_internal(RMATCH(obj)->regexp); + if (RMATCH(obj)->str) { + gc_mark_internal(RMATCH(obj)->str); + } + break; case T_RATIONAL: - gc_mark(objspace, obj->as.rational.num); - ptr = obj->as.rational.den; - goto again; + gc_mark_internal(RRATIONAL(obj)->num); + gc_mark_internal(RRATIONAL(obj)->den); + break; case T_COMPLEX: - gc_mark(objspace, obj->as.complex.real); - ptr = obj->as.complex.imag; - goto again; + gc_mark_internal(RCOMPLEX(obj)->real); + gc_mark_internal(RCOMPLEX(obj)->imag); + break; - case T_STRUCT: - { - long len = RSTRUCT_LEN(obj); - const VALUE *ptr = RSTRUCT_CONST_PTR(obj); + case T_STRUCT: { + const long len = RSTRUCT_LEN(obj); + const VALUE * const ptr = RSTRUCT_CONST_PTR(obj); - while (len--) { - gc_mark(objspace, *ptr++); - } - } - break; + for (long i = 0; i < len; i++) { + gc_mark_internal(ptr[i]); + } + + if (rb_obj_shape_has_fields(obj) && !FL_TEST_RAW(obj, RSTRUCT_GEN_FIELDS)) { + gc_mark_internal(RSTRUCT_FIELDS_OBJ(obj)); + } + + break; + } default: -#if GC_DEBUG - rb_gcdebug_print_obj_condition((VALUE)obj); -#endif - if (BUILTIN_TYPE(obj) == T_NONE) rb_bug("rb_gc_mark(): %p is T_NONE", (void *)obj); - if (BUILTIN_TYPE(obj) == T_ZOMBIE) rb_bug("rb_gc_mark(): %p is T_ZOMBIE", (void *)obj); - rb_bug("rb_gc_mark(): unknown data type 0x%x(%p) %s", - BUILTIN_TYPE(obj), (void *)obj, - is_pointer_to_heap(objspace, obj) ? "corrupted object" : "non object"); + if (BUILTIN_TYPE(obj) == T_MOVED) rb_bug("rb_gc_mark(): %p is T_MOVED", (void *)obj); + if (BUILTIN_TYPE(obj) == T_NONE) rb_bug("rb_gc_mark(): %p is T_NONE", (void *)obj); + if (BUILTIN_TYPE(obj) == T_ZOMBIE) rb_bug("rb_gc_mark(): %p is T_ZOMBIE", (void *)obj); + rb_bug("rb_gc_mark(): unknown data type 0x%x(%p) %s", + BUILTIN_TYPE(obj), (void *)obj, + rb_gc_impl_pointer_to_heap_p(objspace, (void *)obj) ? "corrupted object" : "non object"); } } -static void -gc_mark_stacked_objects(rb_objspace_t *objspace) +size_t +rb_gc_obj_optimal_size(VALUE obj) { - mark_stack_t *mstack = &objspace->mark_stack; - VALUE obj = 0; + switch (BUILTIN_TYPE(obj)) { + case T_ARRAY: + { + size_t size = rb_ary_size_as_embedded(obj); + if (rb_gc_size_allocatable_p(size)) { + return size; + } + else { + return sizeof(struct RArray); + } + } - if (!mstack->index) return; - while (pop_mark_stack(mstack, &obj)) { - if (RGENGC_CHECK_MODE > 0 && !gc_marked(objspace, obj)) { - rb_bug("gc_mark_stacked_objects: %p (%s) is infant, but not marked.", (void *)obj, obj_type_name(obj)); - } - gc_mark_children(objspace, obj); - } - shrink_stack_chunk_cache(mstack); -} + case T_OBJECT: + if (rb_obj_shape_complex_p(obj)) { + return sizeof(struct RObject); + } + else { + size_t size = rb_obj_embedded_size(ROBJECT_FIELDS_CAPACITY(obj)); + if (rb_gc_size_allocatable_p(size)) { + return size; + } + else { + return sizeof(struct RObject); + } + } -#ifndef RGENGC_PRINT_TICK -#define RGENGC_PRINT_TICK 0 -#endif -/* the following code is only for internal tuning. */ + case T_STRING: + { + size_t size = rb_str_size_as_embedded(obj); + if (rb_gc_size_allocatable_p(size)) { + return size; + } + else { + return sizeof(struct RString); + } + } -/* Source code to use RDTSC is quoted and modified from - * http://www.mcs.anl.gov/~kazutomo/rdtsc.html - * written by Kazutomo Yoshii <kazutomo@mcs.anl.gov> - */ + case T_HASH: + { + if (RB_OBJ_FROZEN(obj) && RHASH_AR_TABLE_P(obj)) { + return sizeof(struct RHash) + offsetof(ar_table, pairs) + RHASH_AR_TABLE_BOUND(obj) * sizeof(ar_table_pair); + } + return sizeof(struct RHash) + (RHASH_ST_TABLE_P(obj) ? sizeof(st_table) : sizeof(ar_table)); + } -#if RGENGC_PRINT_TICK -#if defined(__GNUC__) && defined(__i386__) -typedef unsigned long long tick_t; + default: + return 0; + } +} -static inline tick_t -tick(void) +void +rb_gc_writebarrier(VALUE a, VALUE b) { - unsigned long long int x; - __asm__ __volatile__ ("rdtsc" : "=A" (x)); - return x; + rb_gc_impl_writebarrier(rb_gc_get_objspace(), a, b); } -#elif defined(__GNUC__) && defined(__x86_64__) -typedef unsigned long long tick_t; - -static __inline__ tick_t -tick(void) +void +rb_gc_writebarrier_unprotect(VALUE obj) { - unsigned long hi, lo; - __asm__ __volatile__ ("rdtsc" : "=a"(lo), "=d"(hi)); - return ((unsigned long long)lo)|( ((unsigned long long)hi)<<32); + rb_gc_impl_writebarrier_unprotect(rb_gc_get_objspace(), obj); } -#elif defined(_WIN32) && defined(_MSC_VER) -#include <intrin.h> -typedef unsigned __int64 tick_t; - -static inline tick_t -tick(void) +/* + * remember `obj' if needed. + */ +void +rb_gc_writebarrier_remember(VALUE obj) { - return __rdtsc(); + rb_gc_impl_writebarrier_remember(rb_gc_get_objspace(), obj); } -#else /* use clock */ -typedef clock_t tick_t; -static inline tick_t -tick(void) +void +rb_gc_copy_attributes(VALUE dest, VALUE obj) { - return clock(); + rb_gc_impl_copy_attributes(rb_gc_get_objspace(), dest, obj); } -#endif -#define MAX_TICKS 0x100 -static tick_t mark_ticks[MAX_TICKS]; -static const char *mark_ticks_categories[MAX_TICKS]; - -static void -show_mark_ticks(void) +#if USE_MODULAR_GC +int +rb_gc_modular_gc_loaded_p(void) { - int i; - fprintf(stderr, "mark ticks result:\n"); - for (i=0; i<MAX_TICKS; i++) { - const char *category = mark_ticks_categories[i]; - if (category) { - fprintf(stderr, "%s\t%8lu\n", category, (unsigned long)mark_ticks[i]); - } - else { - break; - } - } + return rb_gc_functions.modular_gc_loaded_p; } -#endif /* RGENGC_PRINT_TICK */ - -static void -gc_mark_roots(rb_objspace_t *objspace, int full_mark, const char **categoryp) +const char * +rb_gc_active_gc_name(void) { - struct gc_list *list; - rb_thread_t *th = GET_THREAD(); - if (categoryp) *categoryp = "xxx"; - -#if RGENGC_PRINT_TICK - tick_t start_tick = tick(); - int tick_count = 0; - const char *prev_category = 0; + const char *gc_name = rb_gc_impl_active_gc_name(); - if (mark_ticks_categories[0] == 0) { - atexit(show_mark_ticks); + const size_t len = strlen(gc_name); + if (len > RB_GC_MAX_NAME_LEN) { + rb_bug("GC should have a name no more than %d chars long. Currently: %zu (%s)", + RB_GC_MAX_NAME_LEN, len, gc_name); } -#endif - -#if RGENGC_PRINT_TICK -#define MARK_CHECKPOINT_PRINT_TICK(category) do { \ - if (prev_category) { \ - tick_t t = tick(); \ - mark_ticks[tick_count] = t - start_tick; \ - mark_ticks_categories[tick_count] = prev_category; \ - tick_count++; \ - } \ - prev_category = category; \ - start_tick = tick(); \ -} while (0) -#else /* RGENGC_PRINT_TICK */ -#define MARK_CHECKPOINT_PRINT_TICK(category) -#endif -#define MARK_CHECKPOINT(category) do { \ - if (categoryp) *categoryp = category; \ - MARK_CHECKPOINT_PRINT_TICK(category); \ -} while (0) - - MARK_CHECKPOINT("vm"); - SET_STACK_END; - th->vm->self ? rb_gc_mark(th->vm->self) : rb_vm_mark(th->vm); - - MARK_CHECKPOINT("finalizers"); - mark_tbl(objspace, finalizer_table); - - MARK_CHECKPOINT("machine_context"); - mark_current_machine_context(objspace, th); - - MARK_CHECKPOINT("symbols"); -#if USE_RGENGC - objspace->rgengc.parent_object_is_old = TRUE; - rb_gc_mark_symbols(full_mark); - objspace->rgengc.parent_object_is_old = FALSE; -#else - rb_gc_mark_symbols(full_mark); + return gc_name; +} #endif - MARK_CHECKPOINT("encodings"); - rb_gc_mark_encodings(); - - /* mark protected global variables */ - MARK_CHECKPOINT("global_list"); - for (list = global_List; list; list = list->next) { - rb_gc_mark_maybe(*list->varptr); - } - - MARK_CHECKPOINT("end_proc"); - rb_mark_end_proc(); +struct rb_gc_object_metadata_entry * +rb_gc_object_metadata(VALUE obj) +{ + return rb_gc_impl_object_metadata(rb_gc_get_objspace(), obj); +} - MARK_CHECKPOINT("global_tbl"); - rb_gc_mark_global_tbl(); +/* GC */ - /* mark generic instance variables for special constants */ - MARK_CHECKPOINT("generic_ivars"); - rb_mark_generic_ivar_tbl(); +void * +rb_gc_ractor_cache_alloc(rb_ractor_t *ractor) +{ + return rb_gc_impl_ractor_cache_alloc(rb_gc_get_objspace(), ractor); +} - MARK_CHECKPOINT("parser"); - rb_gc_mark_parser(); +void +rb_gc_ractor_cache_free(void *cache) +{ + rb_gc_impl_ractor_cache_free(rb_gc_get_objspace(), cache); +} - MARK_CHECKPOINT("live_method_entries"); - rb_gc_mark_unlinked_live_method_entries(th->vm); +void +rb_gc_register_mark_object(VALUE obj) +{ + if (!rb_gc_impl_pointer_to_heap_p(rb_gc_get_objspace(), (void *)obj)) + return; - MARK_CHECKPOINT("finish"); -#undef MARK_CHECKPOINT + rb_vm_register_global_object(obj); } -static void -gc_marks_body(rb_objspace_t *objspace, int full_mark) +void +rb_gc_register_address(VALUE *addr) { - /* start marking */ - rgengc_report(1, objspace, "gc_marks_body: start (%s)\n", full_mark ? "full" : "minor"); + rb_vm_t *vm = GET_VM(); -#if USE_RGENGC - objspace->rgengc.parent_object_is_old = FALSE; - objspace->rgengc.during_minor_gc = full_mark ? FALSE : TRUE; + VALUE obj = *addr; - if (objspace->rgengc.during_minor_gc) { - objspace->profile.minor_gc_count++; - rgengc_rememberset_mark(objspace, heap_eden); - } - else { - objspace->profile.major_gc_count++; - rgengc_mark_and_rememberset_clear(objspace, heap_eden); + RB_VM_LOCKING() { + if (vm->global_object_list_size == vm->global_object_list_capa) { + size_t new_capa = vm->global_object_list_capa ? vm->global_object_list_capa * 2 : 64; + SIZED_REALLOC_N(vm->global_object_list, VALUE *, new_capa, vm->global_object_list_capa); + vm->global_object_list_capa = new_capa; + } + + vm->global_object_list[vm->global_object_list_size++] = addr; } -#endif - gc_mark_roots(objspace, full_mark, 0); - gc_mark_stacked_objects(objspace); - gc_event_hook(objspace, RUBY_INTERNAL_EVENT_GC_END_MARK, 0); - rgengc_report(1, objspace, "gc_marks_body: end (%s)\n", full_mark ? "full" : "minor"); + /* + * Because some C extensions have assignment-then-register bugs, + * we guard `obj` here so that it would not get swept defensively. + */ + RB_GC_GUARD(obj); + if (0 && !SPECIAL_CONST_P(obj)) { + rb_warn("Object is assigned to registering address already: %"PRIsVALUE, + rb_obj_class(obj)); + rb_print_backtrace(stderr); + } } -struct verify_internal_consistency_struct { - rb_objspace_t *objspace; - int err_count; - VALUE parent; -}; - -#if USE_RGENGC -static void -verify_internal_consistency_reachable_i(VALUE child, void *ptr) +void +rb_gc_unregister_address(VALUE *addr) { - struct verify_internal_consistency_struct *data = (struct verify_internal_consistency_struct *)ptr; - - assert(RVALUE_OLD_P(data->parent)); - - if (!RVALUE_OLD_P(child)) { - if (!MARKED_IN_BITMAP(GET_HEAP_PAGE(data->parent)->rememberset_bits, data->parent) && - !MARKED_IN_BITMAP(GET_HEAP_PAGE(child)->rememberset_bits, child)) { - fprintf(stderr, "verify_internal_consistency_reachable_i: WB miss %p (%s) -> %p (%s)\n", - (void *)data->parent, obj_type_name(data->parent), - (void *)child, obj_type_name(child)); - data->err_count++; - } + rb_vm_t *vm = GET_VM(); + RB_VM_LOCKING() { + size_t index; + for (index = 0; index < vm->global_object_list_size; index++) { + if (addr == vm->global_object_list[index]) { + MEMMOVE( + vm->global_object_list[index], + &vm->global_object_list[index + 1], + VALUE *, + vm->global_object_list_size - index - 1 + ); + vm->global_object_list_size--; + break; + } + } } } -static int -verify_internal_consistency_i(void *page_start, void *page_end, size_t stride, void *ptr) +void +rb_global_variable(VALUE *var) { - struct verify_internal_consistency_struct *data = (struct verify_internal_consistency_struct *)ptr; - VALUE v; + rb_gc_register_address(var); +} - for (v = (VALUE)page_start; v != (VALUE)page_end; v += stride) { - if (is_live_object(data->objspace, v)) { - if (RVALUE_OLD_P(v)) { - data->parent = v; - /* reachable objects from an oldgen object should be old or (young with remember) */ - rb_objspace_reachable_objects_from(v, verify_internal_consistency_reachable_i, (void *)data); - } - } - } +static VALUE +gc_start_internal(rb_execution_context_t *ec, VALUE self, VALUE full_mark, VALUE immediate_mark, VALUE immediate_sweep, VALUE compact) +{ + rb_gc_impl_start(rb_gc_get_objspace(), RTEST(full_mark), RTEST(immediate_mark), RTEST(immediate_sweep), RTEST(compact)); - return 0; + return Qnil; } -#endif /* USE_RGENGC */ /* - * call-seq: - * GC.verify_internal_consistency -> nil + * rb_objspace_each_objects() is special C API to walk through + * Ruby object space. This C API is too difficult to use it. + * To be frank, you should not use it. Or you need to read the + * source code of this function and understand what this function does. + * + * 'callback' will be called several times (the number of heap page, + * at current implementation) with: + * vstart: a pointer to the first living object of the heap_page. + * vend: a pointer to next to the valid heap_page area. + * stride: a distance to next VALUE. * - * Verify internal consistency. + * If callback() returns non-zero, the iteration will be stopped. + * + * This is a sample callback code to iterate liveness objects: * - * This method is implementation specific. - * Now this method checks generatioanl consistency - * if RGenGC is supported. + * static int + * sample_callback(void *vstart, void *vend, int stride, void *data) + * { + * VALUE v = (VALUE)vstart; + * for (; v != (VALUE)vend; v += stride) { + * if (!rb_objspace_internal_object_p(v)) { // liveness check + * // do something with live object 'v' + * } + * } + * return 0; // continue to iteration + * } + * + * Note: 'vstart' is not a top of heap_page. This point the first + * living object to grasp at least one object to avoid GC issue. + * This means that you can not walk through all Ruby object page + * including freed object page. + * + * Note: On this implementation, 'stride' is the same as sizeof(RVALUE). + * However, there are possibilities to pass variable values with + * 'stride' with some reasons. You must use stride instead of + * use some constant value in the iteration. */ -static VALUE -gc_verify_internal_consistency(VALUE self) +void +rb_objspace_each_objects(int (*callback)(void *, void *, size_t, void *), void *data) { - struct verify_internal_consistency_struct data; - data.objspace = &rb_objspace; - data.err_count = 0; + rb_gc_impl_each_objects(rb_gc_get_objspace(), callback, data); +} -#if USE_RGENGC - { - struct each_obj_args eo_args; - eo_args.callback = verify_internal_consistency_i; - eo_args.data = (void *)&data; - objspace_each_objects((VALUE)&eo_args); +static void +gc_ref_update_array(void *objspace, VALUE v) +{ + if (ARY_SHARED_P(v)) { + VALUE old_root = RARRAY(v)->as.heap.aux.shared_root; + + UPDATE_IF_MOVED(objspace, RARRAY(v)->as.heap.aux.shared_root); + + VALUE new_root = RARRAY(v)->as.heap.aux.shared_root; + // If the root is embedded and its location has changed + if (ARY_EMBED_P(new_root) && new_root != old_root) { + size_t offset = (size_t)(RARRAY(v)->as.heap.ptr - RARRAY(old_root)->as.ary); + GC_ASSERT(RARRAY(v)->as.heap.ptr >= RARRAY(old_root)->as.ary); + RARRAY(v)->as.heap.ptr = RARRAY(new_root)->as.ary + offset; + } } -#endif - if (data.err_count != 0) { - rb_bug("gc_verify_internal_consistency: found internal consistency.\n"); + else { + long len = RARRAY_LEN(v); + + if (len > 0) { + VALUE *ptr = (VALUE *)RARRAY_CONST_PTR(v); + for (long i = 0; i < len; i++) { + UPDATE_IF_MOVED(objspace, ptr[i]); + } + } + + if (rb_gc_obj_slot_size(v) >= rb_ary_size_as_embedded(v)) { + if (rb_ary_embeddable_p(v)) { + rb_ary_make_embedded(v); + } + } } - return Qnil; } -#if RGENGC_CHECK_MODE >= 3 +static void +gc_ref_update_object(void *objspace, VALUE v) +{ + VALUE *ptr = ROBJECT_FIELDS(v); -#define MAKE_ROOTSIG(obj) (((VALUE)(obj) << 1) | 0x01) -#define IS_ROOTSIG(obj) ((VALUE)(obj) & 0x01) -#define GET_ROOTSIG(obj) ((const char *)((VALUE)(obj) >> 1)) + if (FL_TEST_RAW(v, ROBJECT_HEAP)) { + if (rb_obj_shape_complex_p(v)) { + gc_ref_update_table_values_only(ROBJECT_FIELDS_HASH(v)); + return; + } -struct reflist { - VALUE *list; - int pos; - int size; -}; + size_t slot_size = rb_gc_obj_slot_size(v); + size_t embed_size = rb_obj_embedded_size(ROBJECT_FIELDS_CAPACITY(v)); + if (slot_size >= embed_size) { + // Object can be re-embedded + memcpy(ROBJECT(v)->as.ary, ptr, sizeof(VALUE) * ROBJECT_FIELDS_COUNT(v)); + SIZED_FREE_N(ptr, ROBJECT_FIELDS_CAPACITY(v)); + FL_UNSET_RAW(v, ROBJECT_HEAP); + ptr = ROBJECT(v)->as.ary; + } + } -static struct reflist * -reflist_create(VALUE obj) -{ - struct reflist *refs = xmalloc(sizeof(struct reflist)); - refs->size = 1; - refs->list = ALLOC_N(VALUE, refs->size); - refs->list[0] = obj; - refs->pos = 1; - return refs; + for (uint32_t i = 0; i < ROBJECT_FIELDS_COUNT(v); i++) { + UPDATE_IF_MOVED(objspace, ptr[i]); + } } -static void -reflist_destruct(struct reflist *refs) +void +rb_gc_ref_update_table_values_only(st_table *tbl) { - xfree(refs->list); - xfree(refs); + gc_ref_update_table_values_only(tbl); } -static void -reflist_add(struct reflist *refs, VALUE obj) +/* Update MOVED references in a VALUE=>VALUE st_table */ +void +rb_gc_update_tbl_refs(st_table *ptr) { - if (refs->pos == refs->size) { - refs->size *= 2; - SIZED_REALLOC_N(refs->list, VALUE, refs->size, refs->size/2); - } - - refs->list[refs->pos++] = obj; + gc_update_table_refs(ptr); } -static void -reflist_dump(struct reflist *refs) +static int +rb_gc_update_set_refs_i(st_data_t key, st_data_t value, st_data_t argp, int error) { - int i; - for (i=0; i<refs->pos; i++) { - VALUE obj = refs->list[i]; - if (IS_ROOTSIG(obj)) { /* root */ - fprintf(stderr, "<root@%s>", GET_ROOTSIG(obj)); - } - else { - fprintf(stderr, "<%p@%s>", (void *)obj, obj_type_name(obj)); - } - if (i+1 < refs->pos) fprintf(stderr, ", "); + if (rb_gc_location((VALUE)key) != (VALUE)key) { + return ST_REPLACE; } + + return ST_CONTINUE; } -#if RGENGC_CHECK_MODE >= 3 static int -reflist_refered_from_machine_context(struct reflist *refs) +rb_gc_update_set_refs_replace_i(st_data_t *key, st_data_t *value, st_data_t argp, int existing) { - int i; - for (i=0; i<refs->pos; i++) { - VALUE obj = refs->list[i]; - if (IS_ROOTSIG(obj) && strcmp(GET_ROOTSIG(obj), "machine_context") == 0) return 1; + if (rb_gc_location((VALUE)*key) != (VALUE)*key) { + *key = rb_gc_location((VALUE)*key); } - return 0; -} -#endif -struct allrefs { - rb_objspace_t *objspace; - /* a -> obj1 - * b -> obj1 - * c -> obj1 - * c -> obj2 - * d -> obj3 - * #=> {obj1 => [a, b, c], obj2 => [c, d]} - */ - struct st_table *references; - const char *category; - VALUE root_obj; -}; + return ST_CONTINUE; +} -static int -allrefs_add(struct allrefs *data, VALUE obj) +void +rb_gc_update_set_refs(st_table *tbl) { - struct reflist *refs; + if (!tbl || tbl->num_entries == 0) return; - if (st_lookup(data->references, obj, (st_data_t *)&refs)) { - reflist_add(refs, data->root_obj); - return 0; - } - else { - refs = reflist_create(data->root_obj); - st_insert(data->references, obj, (st_data_t)refs); - return 1; + if (st_foreach_with_replace(tbl, rb_gc_update_set_refs_i, rb_gc_update_set_refs_replace_i, 0)) { + rb_raise(rb_eRuntimeError, "hash modified during iteration"); } } static void -allrefs_i(VALUE obj, void *ptr) +gc_ref_update_hash(void *objspace, VALUE v) { - struct allrefs *data = (struct allrefs *)ptr; - - if (allrefs_add(data, obj)) { - push_mark_stack(&data->objspace->mark_stack, obj); - } + rb_hash_stlike_foreach_with_replace(v, hash_foreach_replace, hash_replace_ref, (st_data_t)objspace); } static void -allrefs_roots_i(VALUE obj, void *ptr) +gc_update_values(void *objspace, long n, VALUE *values) { - struct allrefs *data = (struct allrefs *)ptr; - if (strlen(data->category) == 0) rb_bug("!!!"); - data->root_obj = MAKE_ROOTSIG(data->category); - - if (allrefs_add(data, obj)) { - push_mark_stack(&data->objspace->mark_stack, obj); + for (long i = 0; i < n; i++) { + UPDATE_IF_MOVED(objspace, values[i]); } } -static st_table * -objspace_allrefs(rb_objspace_t *objspace) +void +rb_gc_update_values(long n, VALUE *values) { - struct allrefs data; - struct mark_func_data_struct mfd; - VALUE obj; - - data.objspace = objspace; - data.references = st_init_numtable(); - - mfd.mark_func = allrefs_roots_i; - mfd.data = &data; + gc_update_values(rb_gc_get_objspace(), n, values); +} - /* traverse root objects */ - objspace->mark_func_data = &mfd; - gc_mark_roots(objspace, TRUE, &data.category); - objspace->mark_func_data = 0; +static enum rb_id_table_iterator_result +check_id_table_move(VALUE value, void *data) +{ + void *objspace = (void *)data; - /* traverse rest objects reachable from root objects */ - while (pop_mark_stack(&objspace->mark_stack, &obj)) { - rb_objspace_reachable_objects_from(data.root_obj = obj, allrefs_i, &data); + if (gc_object_moved_p_internal(objspace, (VALUE)value)) { + return ID_TABLE_REPLACE; } - shrink_stack_chunk_cache(&objspace->mark_stack); - return data.references; + return ID_TABLE_CONTINUE; } -static int -objspaec_allrefs_destruct_i(st_data_t key, st_data_t value, void *ptr) +void +rb_gc_prepare_heap_process_object(VALUE obj) { - struct reflist *refs = (struct reflist *)value; - reflist_destruct(refs); - return ST_CONTINUE; + switch (BUILTIN_TYPE(obj)) { + case T_STRING: + // Precompute the string coderange. This both save time for when it will be + // eventually needed, and avoid mutating heap pages after a potential fork. + rb_enc_str_coderange(obj); + break; + default: + break; + } } -static void -objspace_allrefs_destruct(struct st_table *refs) +void +rb_gc_prepare_heap(void) { - st_foreach(refs, objspaec_allrefs_destruct_i, 0); - st_free_table(refs); + rb_gc_impl_prepare_heap(rb_gc_get_objspace()); } -#if RGENGC_CHECK_MODE >= 4 -static int -allrefs_dump_i(st_data_t k, st_data_t v, st_data_t ptr) -{ - VALUE obj = (VALUE)k; - struct reflist *refs = (struct reflist *)v; - fprintf(stderr, "[allrefs_dump_i] %p (%s%s%s%s) <- ", - (void *)obj, obj_type_name(obj), - RVALUE_OLD_P(obj) ? "[O]" : "[Y]", - RVALUE_WB_PROTECTED(obj) ? "[W]" : "", - MARKED_IN_BITMAP(GET_HEAP_REMEMBERSET_BITS(obj), obj) ? "[R]" : ""); - reflist_dump(refs); - fprintf(stderr, "\n"); - return ST_CONTINUE; +size_t +rb_gc_heap_id_for_size(size_t size) +{ + return rb_gc_impl_heap_id_for_size(rb_gc_get_objspace(), size); } -static void -allrefs_dump(rb_objspace_t *objspace) +bool +rb_gc_size_allocatable_p(size_t size) { - fprintf(stderr, "[all refs] (size: %d)\n", (int)objspace->rgengc.allrefs_table->num_entries); - st_foreach(objspace->rgengc.allrefs_table, allrefs_dump_i, 0); + return rb_gc_impl_size_allocatable_p(size); } -#endif -#if RGENGC_CHECK_MODE >= 3 -static int -gc_check_after_marks_i(st_data_t k, st_data_t v, void *ptr) -{ - VALUE obj = k; - struct reflist *refs = (struct reflist *)v; - rb_objspace_t *objspace = (rb_objspace_t *)ptr; - - /* object should be marked or oldgen */ - if (!MARKED_IN_BITMAP(GET_HEAP_MARK_BITS(obj), obj)) { - fprintf(stderr, "gc_check_after_marks_i: %p (%s) is not marked and not oldgen.\n", (void *)obj, obj_type_name(obj)); - fprintf(stderr, "gc_check_after_marks_i: %p is referred from ", (void *)obj); - reflist_dump(refs); - - if (reflist_refered_from_machine_context(refs)) { - fprintf(stderr, " (marked from machine stack).\n"); - /* marked from machine context can be false positive */ - } - else { - objspace->rgengc.error_count++; - fprintf(stderr, "\n"); - } +static enum rb_id_table_iterator_result +update_id_table(VALUE *value, void *data, int existing) +{ + void *objspace = (void *)data; + + if (gc_object_moved_p_internal(objspace, (VALUE)*value)) { + *value = gc_location_internal(objspace, (VALUE)*value); } - return ST_CONTINUE; + + return ID_TABLE_CONTINUE; } -#endif static void -gc_marks_check(rb_objspace_t *objspace, int (*checker_func)(ANYARGS), const char *checker_name) +update_m_tbl(void *objspace, struct rb_id_table *tbl) { + if (tbl) { + rb_id_table_foreach_values_with_replace(tbl, check_id_table_move, update_id_table, objspace); + } +} - size_t saved_malloc_increase = objspace->malloc_params.increase; -#if RGENGC_ESTIMATE_OLDMALLOC - size_t saved_oldmalloc_increase = objspace->rgengc.oldmalloc_increase; -#endif - VALUE already_disabled = rb_gc_disable(); - - objspace->rgengc.allrefs_table = objspace_allrefs(objspace); - st_foreach(objspace->rgengc.allrefs_table, checker_func, (st_data_t)objspace); +static enum rb_id_table_iterator_result +update_const_tbl_i(VALUE value, void *objspace) +{ + rb_const_entry_t *ce = (rb_const_entry_t *)value; - if (objspace->rgengc.error_count > 0) { -#if RGENGC_CHECK_MODE >= 4 - allrefs_dump(objspace); -#endif - rb_bug("%s: GC has problem.", checker_name); + if (gc_object_moved_p_internal(objspace, ce->value)) { + ce->value = gc_location_internal(objspace, ce->value); } - objspace_allrefs_destruct(objspace->rgengc.allrefs_table); - objspace->rgengc.allrefs_table = 0; + if (gc_object_moved_p_internal(objspace, ce->file)) { + ce->file = gc_location_internal(objspace, ce->file); + } - if (already_disabled == Qfalse) rb_gc_enable(); - objspace->malloc_params.increase = saved_malloc_increase; -#if RGENGC_ESTIMATE_OLDMALLOC - objspace->rgengc.oldmalloc_increase = saved_oldmalloc_increase; -#endif + return ID_TABLE_CONTINUE; } -#endif /* RGENGC_CHECK_MODE >= 2 */ - static void -gc_marks(rb_objspace_t *objspace, int full_mark) +update_const_tbl(void *objspace, struct rb_id_table *tbl) { - struct mark_func_data_struct *prev_mark_func_data; - - gc_prof_mark_timer_start(objspace); - { - /* setup marking */ - prev_mark_func_data = objspace->mark_func_data; - objspace->mark_func_data = 0; - -#if USE_RGENGC - -#if RGENGC_CHECK_MODE >= 2 - gc_verify_internal_consistency(Qnil); -#endif - if (full_mark == TRUE) { /* major/full GC */ - objspace->rgengc.remembered_shady_object_count = 0; - objspace->rgengc.old_object_count = 0; -#if RGENGC_THREEGEN - objspace->rgengc.young_object_count = 0; -#endif - - gc_marks_body(objspace, TRUE); - { - /* See the comment about RUBY_GC_HEAP_OLDOBJECT_LIMIT_FACTOR */ - const double r = gc_params.oldobject_limit_factor; - objspace->rgengc.remembered_shady_object_limit = (size_t)(objspace->rgengc.remembered_shady_object_count * r); - objspace->rgengc.old_object_limit = (size_t)(objspace->rgengc.old_object_count * r); - } - } - else { /* minor GC */ - gc_marks_body(objspace, FALSE); - } - -#if RGENGC_PROFILE > 0 - if (gc_prof_record(objspace)) { - gc_profile_record *record = gc_prof_record(objspace); - record->old_objects = objspace->rgengc.old_object_count; - } -#endif - -#if RGENGC_CHECK_MODE >= 3 - gc_marks_check(objspace, gc_check_after_marks_i, "after_marks"); -#endif - -#else /* USE_RGENGC */ - gc_marks_body(objspace, TRUE); -#endif + if (!tbl) return; + rb_id_table_foreach_values(tbl, update_const_tbl_i, objspace); +} - objspace->mark_func_data = prev_mark_func_data; +static void +update_superclasses(rb_objspace_t *objspace, rb_classext_t *ext) +{ + if (RCLASSEXT_SUPERCLASSES_WITH_SELF(ext)) { + size_t array_size = RCLASSEXT_SUPERCLASS_DEPTH(ext) + 1; + for (size_t i = 0; i < array_size; i++) { + UPDATE_IF_MOVED(objspace, RCLASSEXT_SUPERCLASSES(ext)[i]); + } } - gc_prof_mark_timer_stop(objspace); } -/* RGENGC */ - static void -rgengc_report_body(int level, rb_objspace_t *objspace, const char *fmt, ...) -{ - if (level <= RGENGC_DEBUG) { - char buf[1024]; - FILE *out = stderr; - va_list args; - const char *status = " "; - -#if USE_RGENGC - if (during_gc) { - status = objspace->rgengc.during_minor_gc ? "-" : "+"; - } -#endif +update_classext_values(rb_objspace_t *objspace, rb_classext_t *ext, bool is_iclass) +{ + UPDATE_IF_MOVED(objspace, RCLASSEXT_ORIGIN(ext)); + UPDATE_IF_MOVED(objspace, RCLASSEXT_REFINED_CLASS(ext)); + UPDATE_IF_MOVED(objspace, RCLASSEXT_CLASSPATH(ext)); + if (is_iclass) { + UPDATE_IF_MOVED(objspace, RCLASSEXT_INCLUDER(ext)); + } +} - va_start(args, fmt); - vsnprintf(buf, 1024, fmt, args); - va_end(args); +static void +update_classext(rb_classext_t *ext, bool is_prime, VALUE box_value, void *arg) +{ + struct classext_foreach_args *args = (struct classext_foreach_args *)arg; + rb_objspace_t *objspace = args->objspace; - fprintf(out, "%s|", status); - fputs(buf, out); + if (RCLASSEXT_SUPER(ext)) { + UPDATE_IF_MOVED(objspace, RCLASSEXT_SUPER(ext)); } -} -#if USE_RGENGC + update_m_tbl(objspace, RCLASSEXT_M_TBL(ext)); -/* bit operations */ + UPDATE_IF_MOVED(objspace, ext->fields_obj); + if (!RCLASSEXT_SHARED_CONST_TBL(ext)) { + update_const_tbl(objspace, RCLASSEXT_CONST_TBL(ext)); + } + UPDATE_IF_MOVED(objspace, RCLASSEXT_CC_TBL(ext)); + UPDATE_IF_MOVED(objspace, RCLASSEXT_CVC_TBL(ext)); + update_superclasses(objspace, ext); + if (RCLASSEXT_SUBCLASSES(ext)) { + UPDATE_IF_MOVED(objspace, RCLASSEXT_SUBCLASSES(ext)); + } -static int -rgengc_remembersetbits_get(rb_objspace_t *objspace, VALUE obj) -{ - bits_t *bits = GET_HEAP_REMEMBERSET_BITS(obj); - return MARKED_IN_BITMAP(bits, obj) ? 1 : 0; + update_classext_values(objspace, ext, false); } -static int -rgengc_remembersetbits_set(rb_objspace_t *objspace, VALUE obj) +static void +update_iclass_classext(rb_classext_t *ext, bool is_prime, VALUE box_value, void *arg) { - bits_t *bits = GET_HEAP_REMEMBERSET_BITS(obj); - if (MARKED_IN_BITMAP(bits, obj)) { - return FALSE; + struct classext_foreach_args *args = (struct classext_foreach_args *)arg; + rb_objspace_t *objspace = args->objspace; + + if (RCLASSEXT_SUPER(ext)) { + UPDATE_IF_MOVED(objspace, RCLASSEXT_SUPER(ext)); } - else { - MARK_IN_BITMAP(bits, obj); - return TRUE; + update_m_tbl(objspace, RCLASSEXT_M_TBL(ext)); + update_m_tbl(objspace, RCLASSEXT_CALLABLE_M_TBL(ext)); + UPDATE_IF_MOVED(objspace, RCLASSEXT_CC_TBL(ext)); + UPDATE_IF_MOVED(objspace, RCLASSEXT_CVC_TBL(ext)); + if (RCLASSEXT_SUBCLASSES(ext)) { + UPDATE_IF_MOVED(objspace, RCLASSEXT_SUBCLASSES(ext)); } + + update_classext_values(objspace, ext, true); } -/* wb, etc */ +struct global_vm_table_foreach_data { + vm_table_foreach_callback_func callback; + vm_table_update_callback_func update_callback; + void *data; + bool weak_only; +}; -/* return FALSE if already remembered */ static int -rgengc_remember(rb_objspace_t *objspace, VALUE obj) +vm_weak_table_foreach_weak_key(st_data_t key, st_data_t value, st_data_t data, int error) { - rgengc_report(2, objspace, "rgengc_remember: %p (%s, %s) %s\n", (void *)obj, obj_type_name(obj), - RVALUE_WB_PROTECTED(obj) ? "WB-protected" : "non-WB-protected", - rgengc_remembersetbits_get(objspace, obj) ? "was already remembered" : "is remembered now"); + struct global_vm_table_foreach_data *iter_data = (struct global_vm_table_foreach_data *)data; -#if RGENGC_CHECK_MODE > 0 - { - switch (BUILTIN_TYPE(obj)) { - case T_NONE: - case T_ZOMBIE: - rb_bug("rgengc_remember: should not remember %p (%s)\n", - (void *)obj, obj_type_name(obj)); - default: - ; - } - } -#endif + int ret = iter_data->callback((VALUE)key, iter_data->data); - if (RGENGC_PROFILE) { - if (!rgengc_remembered(objspace, obj)) { -#if RGENGC_PROFILE > 0 - if (RVALUE_WB_PROTECTED(obj)) { - objspace->profile.remembered_normal_object_count++; -#if RGENGC_PROFILE >= 2 - objspace->profile.remembered_normal_object_count_types[BUILTIN_TYPE(obj)]++; -#endif - } - else { - objspace->profile.remembered_shady_object_count++; -#if RGENGC_PROFILE >= 2 - objspace->profile.remembered_shady_object_count_types[BUILTIN_TYPE(obj)]++; -#endif - } -#endif /* RGENGC_PROFILE > 0 */ - } + if (!iter_data->weak_only) { + if (ret != ST_CONTINUE) return ret; + + ret = iter_data->callback((VALUE)value, iter_data->data); } - return rgengc_remembersetbits_set(objspace, obj); + return ret; } static int -rgengc_remembered(rb_objspace_t *objspace, VALUE obj) +vm_weak_table_foreach_update_weak_key(st_data_t *key, st_data_t *value, st_data_t data, int existing) { - int result = rgengc_remembersetbits_get(objspace, obj); - check_gen_consistency(obj); - rgengc_report(6, objspace, "gc_remembered: %p (%s) => %d\n", (void *)obj, obj_type_name(obj), result); - return result; -} + struct global_vm_table_foreach_data *iter_data = (struct global_vm_table_foreach_data *)data; -static void -rgengc_rememberset_mark(rb_objspace_t *objspace, rb_heap_t *heap) -{ - size_t j; - RVALUE *p, *offset; - bits_t *bits, bitset; - struct heap_page *page = heap->pages; + int ret = iter_data->update_callback((VALUE *)key, iter_data->data); -#if RGENGC_PROFILE > 0 - size_t shady_object_count = 0, clear_count = 0; -#endif - - while (page) { - p = page->start; - bits = page->rememberset_bits; - offset = p - NUM_IN_PAGE(p); - - for (j=0; j < HEAP_BITMAP_LIMIT; j++) { - if (bits[j]) { - p = offset + j * BITS_BITLENGTH; - bitset = bits[j]; - do { - if (bitset & 1) { - /* mark before RVALUE_PROMOTE_... */ - gc_mark_ptr(objspace, (VALUE)p); - - if (RVALUE_WB_PROTECTED(p)) { - rgengc_report(2, objspace, "rgengc_rememberset_mark: clear %p (%s)\n", p, obj_type_name((VALUE)p)); -#if RGENGC_THREEGEN - if (RVALUE_INFANT_P((VALUE)p)) RVALUE_PROMOTE_INFANT((VALUE)p); - if (RVALUE_YOUNG_P((VALUE)p)) RVALUE_PROMOTE_YOUNG((VALUE)p); -#endif - CLEAR_IN_BITMAP(bits, p); -#if RGENGC_PROFILE > 0 - clear_count++; -#endif - } - else { -#if RGENGC_PROFILE > 0 - shady_object_count++; -#endif - } + if (!iter_data->weak_only) { + if (ret != ST_CONTINUE) return ret; - rgengc_report(2, objspace, "rgengc_rememberset_mark: mark %p (%s)\n", p, obj_type_name((VALUE)p)); - gc_mark_children(objspace, (VALUE) p); - } - p++; - bitset >>= 1; - } while (bitset); - } - } - page = page->next; + ret = iter_data->update_callback((VALUE *)value, iter_data->data); } - rgengc_report(2, objspace, "rgengc_rememberset_mark: finished\n"); - -#if RGENGC_PROFILE > 0 - rgengc_report(2, objspace, "rgengc_rememberset_mark: clear_count: %"PRIdSIZE", shady_object_count: %"PRIdSIZE"\n", clear_count, shady_object_count); - if (gc_prof_record(objspace)) { - gc_profile_record *record = gc_prof_record(objspace); - record->remembered_normal_objects = clear_count; - record->remembered_shady_objects = shady_object_count; - } -#endif + return ret; } -static void -rgengc_mark_and_rememberset_clear(rb_objspace_t *objspace, rb_heap_t *heap) +static int +vm_weak_table_sym_set_foreach(VALUE *sym_ptr, void *data) { - struct heap_page *page = heap->pages; + VALUE sym = *sym_ptr; + struct global_vm_table_foreach_data *iter_data = (struct global_vm_table_foreach_data *)data; - while (page) { - memset(&page->mark_bits[0], 0, HEAP_BITMAP_SIZE); - memset(&page->rememberset_bits[0], 0, HEAP_BITMAP_SIZE); - page = page->next; - } -} + if (RB_SPECIAL_CONST_P(sym)) return ST_CONTINUE; -/* RGENGC: APIs */ + int ret = iter_data->callback(sym, iter_data->data); -void -rb_gc_writebarrier(VALUE a, VALUE b) -{ - if (RGENGC_CHECK_MODE) { - if (!RVALUE_PROMOTED_P(a)) rb_bug("rb_gc_writebarrier: referer object %p (%s) is not promoted.\n", (void *)a, obj_type_name(a)); + if (ret == ST_REPLACE) { + ret = iter_data->update_callback(sym_ptr, iter_data->data); } - if (!RVALUE_OLD_P(b) && RVALUE_OLD_BITMAP_P(a)) { - rb_objspace_t *objspace = &rb_objspace; - - if (!rgengc_remembered(objspace, a)) { - rgengc_report(2, objspace, "rb_gc_wb: %p (%s) -> %p (%s)\n", - (void *)a, obj_type_name(a), (void *)b, obj_type_name(b)); - rgengc_remember(objspace, a); - } - } + return ret; } -void -rb_gc_writebarrier_unprotect_promoted(VALUE obj) +struct st_table *rb_generic_fields_tbl_get(void); + +static int +vm_weak_table_id2ref_foreach(st_data_t key, st_data_t value, st_data_t data, int error) { - rb_objspace_t *objspace = &rb_objspace; + struct global_vm_table_foreach_data *iter_data = (struct global_vm_table_foreach_data *)data; - if (RGENGC_CHECK_MODE) { - if (!RVALUE_PROMOTED_P(obj)) rb_bug("rb_gc_writebarrier_unprotect_promoted: called on non-promoted object"); - if (!RVALUE_WB_PROTECTED(obj)) rb_bug("rb_gc_writebarrier_unprotect_promoted: called on shady object"); + if (!iter_data->weak_only && !FIXNUM_P((VALUE)key)) { + int ret = iter_data->callback((VALUE)key, iter_data->data); + if (ret != ST_CONTINUE) return ret; } - rgengc_report(0, objspace, "rb_gc_writebarrier_unprotect_promoted: %p (%s)%s\n", (void *)obj, obj_type_name(obj), - rgengc_remembered(objspace, obj) ? " (already remembered)" : ""); - - if (RVALUE_OLD_P(obj)) { - RVALUE_DEMOTE_FROM_OLD(obj); - - rgengc_remember(objspace, obj); - objspace->rgengc.remembered_shady_object_count++; - -#if RGENGC_PROFILE - objspace->profile.shade_operation_count++; -#if RGENGC_PROFILE >= 2 - objspace->profile.shade_operation_count_types[BUILTIN_TYPE(obj)]++; -#endif /* RGENGC_PROFILE >= 2 */ -#endif /* RGENGC_PROFILE */ - } -#if RGENGC_THREEGEN - else { - RVALUE_DEMOTE_FROM_YOUNG(obj); - } -#endif + return iter_data->callback((VALUE)value, iter_data->data); } -void -rb_gc_writebarrier_remember_promoted(VALUE obj) +static int +vm_weak_table_id2ref_foreach_update(st_data_t *key, st_data_t *value, st_data_t data, int existing) { - rb_objspace_t *objspace = &rb_objspace; - rgengc_remember(objspace, obj); -} + struct global_vm_table_foreach_data *iter_data = (struct global_vm_table_foreach_data *)data; -static st_table *rgengc_unprotect_logging_table; + iter_data->update_callback((VALUE *)value, iter_data->data); + + if (!iter_data->weak_only && !FIXNUM_P((VALUE)*key)) { + iter_data->update_callback((VALUE *)key, iter_data->data); + } -static int -rgengc_unprotect_logging_exit_func_i(st_data_t key, st_data_t val) -{ - fprintf(stderr, "%s\t%d\n", (char *)key, (int)val); return ST_CONTINUE; } -static void -rgengc_unprotect_logging_exit_func(void) +static int +vm_weak_table_gen_fields_foreach(st_data_t key, st_data_t value, st_data_t data) { - st_foreach(rgengc_unprotect_logging_table, rgengc_unprotect_logging_exit_func_i, 0); -} + struct global_vm_table_foreach_data *iter_data = (struct global_vm_table_foreach_data *)data; -void -rb_gc_unprotect_logging(void *objptr, const char *filename, int line) -{ - VALUE obj = (VALUE)objptr; + int ret = iter_data->callback((VALUE)key, iter_data->data); - if (rgengc_unprotect_logging_table == 0) { - rgengc_unprotect_logging_table = st_init_strtable(); - atexit(rgengc_unprotect_logging_exit_func); - } + VALUE new_value = (VALUE)value; + VALUE new_key = (VALUE)key; - if (OBJ_WB_PROTECTED(obj)) { - char buff[0x100]; - st_data_t cnt = 1; - char *ptr = buff; + switch (ret) { + case ST_CONTINUE: + break; - snprintf(ptr, 0x100 - 1, "%s|%s:%d", obj_type_name(obj), filename, line); + case ST_DELETE: + // When we're removing an object from the weak ref table, we need to + // set the shape on it so that the GC finalizer won't try to remove + // it again. A "root shape" indicates to the GC that this object + // has no fields on it, hence it won't be in the gen fields table. + RBASIC_SET_SHAPE_ID((VALUE)key, ROOT_SHAPE_ID | SHAPE_ID_LAYOUT_OTHER); + return ST_DELETE; + + case ST_REPLACE: { + ret = iter_data->update_callback(&new_key, iter_data->data); + if (key != new_key) { + ret = ST_DELETE; + } + break; + } - if (st_lookup(rgengc_unprotect_logging_table, (st_data_t)ptr, &cnt)) { - cnt++; - } - else { - ptr = (char *)malloc(strlen(buff) + 1); - strcpy(ptr, buff); - } - st_insert(rgengc_unprotect_logging_table, (st_data_t)ptr, cnt); + default: + rb_bug("vm_weak_table_gen_fields_foreach: return value %d not supported", ret); } -} -#endif /* USE_RGENGC */ + if (!iter_data->weak_only) { + int ivar_ret = iter_data->callback(new_value, iter_data->data); + switch (ivar_ret) { + case ST_CONTINUE: + break; -/* RGENGC analysis information */ + case ST_REPLACE: + iter_data->update_callback(&new_value, iter_data->data); + break; -VALUE -rb_obj_rgengc_writebarrier_protected_p(VALUE obj) -{ - return OBJ_WB_PROTECTED(obj) ? Qtrue : Qfalse; -} - -VALUE -rb_obj_rgengc_promoted_p(VALUE obj) -{ - return OBJ_PROMOTED(obj) ? Qtrue : Qfalse; -} - -size_t -rb_obj_gc_flags(VALUE obj, ID* flags, size_t max) -{ - size_t n = 0; - static ID ID_marked; -#if USE_RGENGC - static ID ID_wb_protected, ID_old, ID_remembered; -#if RGENGC_THREEGEN - static ID ID_young, ID_infant; -#endif -#endif + default: + rb_bug("vm_weak_table_gen_fields_foreach: return value %d not supported", ivar_ret); + } + } - if (!ID_marked) { -#define I(s) ID_##s = rb_intern(#s); - I(marked); -#if USE_RGENGC - I(wb_protected); - I(old); - I(remembered); -#if RGENGC_THREEGEN - I(young); - I(infant); -#endif -#endif -#undef I - } - -#if USE_RGENGC - if (OBJ_WB_PROTECTED(obj) && n<max) - flags[n++] = ID_wb_protected; - if (RVALUE_OLD_P(obj) && n<max) - flags[n++] = ID_old; -#if RGENGC_THREEGEN - if (RVALUE_YOUNG_P(obj) && n<max) - flags[n++] = ID_young; - if (RVALUE_INFANT_P(obj) && n<max) - flags[n++] = ID_infant; -#endif - if (MARKED_IN_BITMAP(GET_HEAP_REMEMBERSET_BITS(obj), obj) && n<max) - flags[n++] = ID_remembered; -#endif - if (MARKED_IN_BITMAP(GET_HEAP_MARK_BITS(obj), obj) && n<max) - flags[n++] = ID_marked; + if (key != new_key || value != new_value) { + DURING_GC_COULD_MALLOC_REGION_START(); + { + st_insert(rb_generic_fields_tbl_get(), (st_data_t)new_key, new_value); + } + DURING_GC_COULD_MALLOC_REGION_END(); + } - return n; + return ret; } -/* GC */ - -void -rb_gc_force_recycle(VALUE p) +static int +vm_weak_table_frozen_strings_foreach(VALUE *str, void *data) { - rb_objspace_t *objspace = &rb_objspace; + // int retval = vm_weak_table_foreach_weak_key(key, value, data, error); + struct global_vm_table_foreach_data *iter_data = (struct global_vm_table_foreach_data *)data; + int retval = iter_data->callback(*str, iter_data->data); -#if USE_RGENGC - CLEAR_IN_BITMAP(GET_HEAP_REMEMBERSET_BITS(p), p); - CLEAR_IN_BITMAP(GET_HEAP_OLDGEN_BITS(p), p); - if (!GET_HEAP_PAGE(p)->before_sweep) { - CLEAR_IN_BITMAP(GET_HEAP_MARK_BITS(p), p); + if (retval == ST_REPLACE) { + retval = iter_data->update_callback(str, iter_data->data); } -#endif - objspace->profile.total_freed_object_num++; - heap_page_add_freeobj(objspace, GET_HEAP_PAGE(p), p); + if (retval == ST_DELETE) { + FL_UNSET(*str, RSTRING_FSTR); + } - /* Disable counting swept_slots because there are no meaning. - * if (!MARKED_IN_BITMAP(GET_HEAP_MARK_BITS(p), p)) { - * objspace->heap.swept_slots++; - * } - */ + return retval; } +void rb_fstring_foreach_with_replace(int (*callback)(VALUE *str, void *data), void *data); void -rb_gc_register_mark_object(VALUE obj) -{ - VALUE ary = GET_THREAD()->vm->mark_object_ary; - rb_ary_push(ary, obj); +rb_gc_vm_weak_table_foreach(vm_table_foreach_callback_func callback, + vm_table_update_callback_func update_callback, + void *data, + bool weak_only, + enum rb_gc_vm_weak_tables table) +{ + rb_vm_t *vm = GET_VM(); + + struct global_vm_table_foreach_data foreach_data = { + .callback = callback, + .update_callback = update_callback, + .data = data, + .weak_only = weak_only, + }; + + switch (table) { + case RB_GC_VM_CI_TABLE: { + st_foreach_with_replace( + &vm->ci_table, + vm_weak_table_foreach_weak_key, + vm_weak_table_foreach_update_weak_key, + (st_data_t)&foreach_data + ); + break; + } + case RB_GC_VM_OVERLOADED_CME_TABLE: { + st_foreach_with_replace( + &vm->overloaded_cme_table, + vm_weak_table_foreach_weak_key, + vm_weak_table_foreach_update_weak_key, + (st_data_t)&foreach_data + ); + break; + } + case RB_GC_VM_GLOBAL_SYMBOLS_TABLE: { + rb_sym_global_symbol_table_foreach_weak_reference( + vm_weak_table_sym_set_foreach, + &foreach_data + ); + break; + } + case RB_GC_VM_ID2REF_TABLE: { + if (id2ref_tbl) { + st_foreach_with_replace( + id2ref_tbl, + vm_weak_table_id2ref_foreach, + vm_weak_table_id2ref_foreach_update, + (st_data_t)&foreach_data + ); + } + break; + } + case RB_GC_VM_GENERIC_FIELDS_TABLE: { + st_table *generic_fields_tbl = rb_generic_fields_tbl_get(); + if (generic_fields_tbl) { + st_foreach( + generic_fields_tbl, + vm_weak_table_gen_fields_foreach, + (st_data_t)&foreach_data + ); + } + break; + } + case RB_GC_VM_FROZEN_STRINGS_TABLE: { + rb_fstring_foreach_with_replace( + vm_weak_table_frozen_strings_foreach, + &foreach_data + ); + break; + } + case RB_GC_VM_WEAK_TABLE_COUNT: + rb_bug("Unreachable"); + default: + rb_bug("rb_gc_vm_weak_table_foreach: unknown table %d", table); + } } void -rb_gc_register_address(VALUE *addr) +rb_gc_update_vm_references(void *objspace) { - rb_objspace_t *objspace = &rb_objspace; - struct gc_list *tmp; + rb_execution_context_t *ec = GET_EC(); + rb_vm_t *vm = rb_ec_vm_ptr(ec); - tmp = ALLOC(struct gc_list); - tmp->next = global_List; - tmp->varptr = addr; - global_List = tmp; -} + rb_vm_update_references(vm); + rb_gc_update_global_tbl(); + rb_sym_global_symbols_mark_and_move(); -void -rb_gc_unregister_address(VALUE *addr) -{ - rb_objspace_t *objspace = &rb_objspace; - struct gc_list *tmp = global_List; +#if USE_YJIT + void rb_yjit_root_update_references(void); // in Rust - if (tmp->varptr == addr) { - global_List = tmp->next; - xfree(tmp); - return; + if (rb_yjit_enabled_p) { + rb_yjit_root_update_references(); } - while (tmp->next) { - if (tmp->next->varptr == addr) { - struct gc_list *t = tmp->next; +#endif - tmp->next = tmp->next->next; - xfree(t); - break; - } - tmp = tmp->next; +#if USE_ZJIT + void rb_zjit_root_update_references(void); // in Rust + + if (rb_zjit_enabled_p) { + rb_zjit_root_update_references(); } +#endif } void -rb_global_variable(VALUE *var) +rb_gc_update_object_references(void *objspace, VALUE obj) { - rb_gc_register_address(var); -} - -#define GC_NOTIFY 0 - -static int -garbage_collect_body(rb_objspace_t *objspace, int full_mark, int immediate_sweep, int reason) -{ - if (ruby_gc_stress && !ruby_disable_gc_stress) { - int flag = FIXNUM_P(ruby_gc_stress) ? FIX2INT(ruby_gc_stress) : 0; - - if (flag & 0x01) - reason &= ~GPR_FLAG_MAJOR_MASK; - else - reason |= GPR_FLAG_MAJOR_BY_STRESS; - immediate_sweep = !(flag & 0x02); - } - else { - if (!GC_ENABLE_LAZY_SWEEP || objspace->flags.dont_lazy_sweep) { - immediate_sweep = TRUE; - } -#if USE_RGENGC - if (full_mark) { - reason |= GPR_FLAG_MAJOR_BY_NOFREE; - } - if (objspace->rgengc.need_major_gc) { - reason |= objspace->rgengc.need_major_gc; - objspace->rgengc.need_major_gc = GPR_FLAG_NONE; - } - if (objspace->rgengc.remembered_shady_object_count > objspace->rgengc.remembered_shady_object_limit) { - reason |= GPR_FLAG_MAJOR_BY_SHADY; - } - if (objspace->rgengc.old_object_count > objspace->rgengc.old_object_limit) { - reason |= GPR_FLAG_MAJOR_BY_OLDGEN; - } -#endif - } - - if (immediate_sweep) reason |= GPR_FLAG_IMMEDIATE_SWEEP; - full_mark = (reason & GPR_FLAG_MAJOR_MASK) ? TRUE : FALSE; + struct classext_foreach_args args; - if (GC_NOTIFY) fprintf(stderr, "start garbage_collect(%d, %d, %d)\n", full_mark, immediate_sweep, reason); + switch (BUILTIN_TYPE(obj)) { + case T_CLASS: + if (FL_TEST_RAW(obj, FL_SINGLETON)) { + UPDATE_IF_MOVED(objspace, RCLASS_ATTACHED_OBJECT(obj)); + } + // Continue to the shared T_CLASS/T_MODULE + case T_MODULE: + args.klass = obj; + args.objspace = objspace; + rb_class_classext_foreach(obj, update_classext, (void *)&args); + break; - objspace->profile.count++; - objspace->profile.latest_gc_info = reason; + case T_ICLASS: + args.objspace = objspace; + rb_class_classext_foreach(obj, update_iclass_classext, (void *)&args); + break; - gc_event_hook(objspace, RUBY_INTERNAL_EVENT_GC_START, 0 /* TODO: pass minor/immediate flag? */); + case T_IMEMO: + rb_imemo_mark_and_move(obj, true); + return; - objspace->profile.total_allocated_object_num_at_gc_start = objspace->profile.total_allocated_object_num; - objspace->profile.heap_used_at_gc_start = heap_pages_used; + case T_NIL: + case T_FIXNUM: + case T_NODE: + case T_MOVED: + case T_NONE: + /* These can't move */ + return; - gc_prof_setup_new_record(objspace, reason); - gc_prof_timer_start(objspace); - { - if (during_gc == 0) { - rb_bug("during_gc should not be 0. RUBY_INTERNAL_EVENT_GC_START user should not cause GC in events."); - } - gc_marks(objspace, full_mark); - gc_sweep(objspace, immediate_sweep); - during_gc = 0; - } - gc_prof_timer_stop(objspace); + case T_ARRAY: + gc_ref_update_array(objspace, obj); + break; - if (GC_NOTIFY) fprintf(stderr, "end garbage_collect()\n"); - return TRUE; -} + case T_HASH: + gc_ref_update_hash(objspace, obj); + UPDATE_IF_MOVED(objspace, RHASH(obj)->ifnone); + break; -static int -heap_ready_to_gc(rb_objspace_t *objspace, rb_heap_t *heap) -{ - if (dont_gc || during_gc) { - if (!heap->freelist && !heap->free_pages) { - if (!heap_increment(objspace, heap)) { - heap_set_increment(objspace, 0); - heap_increment(objspace, heap); + case T_STRING: + { + if (STR_SHARED_P(obj)) { + UPDATE_IF_MOVED(objspace, RSTRING(obj)->as.heap.aux.shared); } - } - return FALSE; - } - return TRUE; -} - -static int -ready_to_gc(rb_objspace_t *objspace) -{ - return heap_ready_to_gc(objspace, heap_eden); -} - -static int -garbage_collect(rb_objspace_t *objspace, int full_mark, int immediate_sweep, int reason) -{ - if (!heap_pages_used) { - during_gc = 0; - return FALSE; - } - if (!ready_to_gc(objspace)) { - during_gc = 0; - return TRUE; - } -#if GC_PROFILE_MORE_DETAIL - objspace->profile.prepare_time = getrusage_time(); -#endif - gc_rest_sweep(objspace); -#if GC_PROFILE_MORE_DETAIL - objspace->profile.prepare_time = getrusage_time() - objspace->profile.prepare_time; -#endif - - during_gc++; - - return garbage_collect_body(objspace, full_mark, immediate_sweep, reason); -} + /* If, after move the string is not embedded, and can fit in the + * slot it's been placed in, then re-embed it. */ + if (rb_gc_obj_slot_size(obj) >= rb_str_size_as_embedded(obj)) { + if (!STR_EMBED_P(obj) && rb_str_reembeddable_p(obj)) { + rb_str_make_embedded(obj); + } + } -struct objspace_and_reason { - rb_objspace_t *objspace; - int reason; - int full_mark; - int immediate_sweep; -}; + break; + } + case T_DATA: + /* Call the compaction callback, if it exists */ + { + void *const ptr = RTYPEDDATA_GET_DATA(obj); + + UPDATE_IF_MOVED(objspace, RTYPEDDATA(obj)->fields_obj); + + if (ptr) { + if (gc_declarative_marking_p(RTYPEDDATA_TYPE(obj))) { + size_t *offset_list = TYPED_DATA_REFS_OFFSET_LIST(obj); + + for (size_t offset = *offset_list; offset != RUBY_REF_END; offset = *offset_list++) { + VALUE *ref = (VALUE *)((char *)ptr + offset); + *ref = gc_location_internal(objspace, *ref); + } + } + else { + RUBY_DATA_FUNC compact_func = RTYPEDDATA_TYPE(obj)->function.dcompact; + if (compact_func) (*compact_func)(ptr); + } + } + } + break; -static void * -gc_with_gvl(void *ptr) -{ - struct objspace_and_reason *oar = (struct objspace_and_reason *)ptr; - return (void *)(VALUE)garbage_collect(oar->objspace, oar->full_mark, oar->immediate_sweep, oar->reason); -} + case T_OBJECT: + gc_ref_update_object(objspace, obj); + break; -static int -garbage_collect_with_gvl(rb_objspace_t *objspace, int full_mark, int immediate_sweep, int reason) -{ - if (dont_gc) return TRUE; - if (ruby_thread_has_gvl_p()) { - return garbage_collect(objspace, full_mark, immediate_sweep, reason); - } - else { - if (ruby_native_thread_p()) { - struct objspace_and_reason oar; - oar.objspace = objspace; - oar.reason = reason; - oar.full_mark = full_mark; - oar.immediate_sweep = immediate_sweep; - return (int)(VALUE)rb_thread_call_with_gvl(gc_with_gvl, (void *)&oar); - } - else { - /* no ruby thread */ - fprintf(stderr, "[FATAL] failed to allocate memory\n"); - exit(EXIT_FAILURE); - } - } -} + case T_FILE: + if (RFILE(obj)->fptr) { + UPDATE_IF_MOVED(objspace, RFILE(obj)->fptr->self); + UPDATE_IF_MOVED(objspace, RFILE(obj)->fptr->pathv); + UPDATE_IF_MOVED(objspace, RFILE(obj)->fptr->tied_io_for_writing); + UPDATE_IF_MOVED(objspace, RFILE(obj)->fptr->writeconv_asciicompat); + UPDATE_IF_MOVED(objspace, RFILE(obj)->fptr->writeconv_pre_ecopts); + UPDATE_IF_MOVED(objspace, RFILE(obj)->fptr->encs.ecopts); + UPDATE_IF_MOVED(objspace, RFILE(obj)->fptr->write_lock); + UPDATE_IF_MOVED(objspace, RFILE(obj)->fptr->timeout); + UPDATE_IF_MOVED(objspace, RFILE(obj)->fptr->wakeup_mutex); + } + break; + case T_REGEXP: + UPDATE_IF_MOVED(objspace, RREGEXP(obj)->src); + break; -int -rb_garbage_collect(void) -{ - return garbage_collect(&rb_objspace, TRUE, TRUE, GPR_FLAG_CAPI); -} + case T_SYMBOL: + UPDATE_IF_MOVED(objspace, RSYMBOL(obj)->fstr); + break; -#undef Init_stack + case T_FLOAT: + case T_BIGNUM: + break; -void -Init_stack(volatile VALUE *addr) -{ - ruby_init_stack(addr); -} + case T_MATCH: + UPDATE_IF_MOVED(objspace, RMATCH(obj)->regexp); -/* - * call-seq: - * GC.start -> nil - * GC.garbage_collect -> nil - * ObjectSpace.garbage_collect -> nil - * GC.start(full_mark: false) -> nil - * - * Initiates garbage collection, unless manually disabled. - * - * This method is defined with keyword arguments that default to true: - * - * def GC.start(full_mark: true, immediate_sweep: true) end - * - * Use full_mark: false to perform a minor GC. - * Use immediate_sweep: false to defer sweeping (use lazy sweep). - * - * Note: These keyword arguments are implementation and version dependent. They - * are not guaranteed to be future-compatible, and may be ignored if the - * underlying implementation does not support them. - */ + if (RMATCH(obj)->str) { + UPDATE_IF_MOVED(objspace, RMATCH(obj)->str); + } + break; -static VALUE -gc_start_internal(int argc, VALUE *argv, VALUE self) -{ - rb_objspace_t *objspace = &rb_objspace; - int full_mark = TRUE, immediate_sweep = TRUE; - VALUE opt = Qnil; - static ID keyword_ids[2]; + case T_RATIONAL: + UPDATE_IF_MOVED(objspace, RRATIONAL(obj)->num); + UPDATE_IF_MOVED(objspace, RRATIONAL(obj)->den); + break; - rb_scan_args(argc, argv, "0:", &opt); + case T_COMPLEX: + UPDATE_IF_MOVED(objspace, RCOMPLEX(obj)->real); + UPDATE_IF_MOVED(objspace, RCOMPLEX(obj)->imag); - if (!NIL_P(opt)) { - VALUE kwvals[2]; + break; - if (!keyword_ids[0]) { - keyword_ids[0] = rb_intern("full_mark"); - keyword_ids[1] = rb_intern("immediate_sweep"); - } + case T_STRUCT: + { + long i, len = RSTRUCT_LEN(obj); + VALUE *ptr = (VALUE *)RSTRUCT_CONST_PTR(obj); - rb_get_kwargs(opt, keyword_ids, 0, 2, kwvals); + for (i = 0; i < len; i++) { + UPDATE_IF_MOVED(objspace, ptr[i]); + } - if (kwvals[0] != Qundef) - full_mark = RTEST(kwvals[0]); - if (kwvals[1] != Qundef) - immediate_sweep = RTEST(kwvals[1]); + if (RSTRUCT_EMBED_LEN(obj)) { + if (!FL_TEST_RAW(obj, RSTRUCT_GEN_FIELDS)) { + UPDATE_IF_MOVED(objspace, ptr[len]); + } + } + else { + UPDATE_IF_MOVED(objspace, RSTRUCT(obj)->as.heap.fields_obj); + } + } + break; + default: + rb_bug("unreachable"); + break; } - garbage_collect(objspace, full_mark, immediate_sweep, GPR_FLAG_METHOD); - if (!finalizing) finalize_deferred(objspace); - - return Qnil; + UPDATE_IF_MOVED(objspace, RBASIC(obj)->klass); } VALUE @@ -5192,604 +4534,247 @@ rb_gc_start(void) void rb_gc(void) { - rb_objspace_t *objspace = &rb_objspace; - garbage_collect(objspace, TRUE, TRUE, GPR_FLAG_CAPI); - if (!finalizing) finalize_deferred(objspace); + unless_objspace(objspace) { return; } + + rb_gc_impl_start(objspace, true, true, true, false); } int rb_during_gc(void) { - rb_objspace_t *objspace = &rb_objspace; - return during_gc; -} + unless_objspace(objspace) { return FALSE; } -#if RGENGC_PROFILE >= 2 -static void -gc_count_add_each_types(VALUE hash, const char *name, const size_t *types) -{ - VALUE result = rb_hash_new(); - int i; - for (i=0; i<T_MASK; i++) { - const char *type = type_name(i, 0); - rb_hash_aset(result, ID2SYM(rb_intern(type)), SIZET2NUM(types[i])); - } - rb_hash_aset(hash, ID2SYM(rb_intern(name)), result); + return rb_gc_impl_during_gc_p(objspace); } -#endif size_t rb_gc_count(void) { - return rb_objspace.profile.count; + return rb_gc_impl_gc_count(rb_gc_get_objspace()); } -/* - * call-seq: - * GC.count -> Integer - * - * The number of times GC occurred. - * - * It returns the number of times GC occurred since the process started. - * - */ - static VALUE -gc_count(VALUE self) +gc_count(rb_execution_context_t *ec, VALUE self) { return SIZET2NUM(rb_gc_count()); } -static VALUE -gc_info_decode(int flags, VALUE hash_or_key) +VALUE +rb_gc_latest_gc_info(VALUE key) { - static VALUE sym_major_by = Qnil, sym_gc_by, sym_immediate_sweep, sym_have_finalizer; - static VALUE sym_nofree, sym_oldgen, sym_shady, sym_rescan, sym_stress; -#if RGENGC_ESTIMATE_OLDMALLOC - static VALUE sym_oldmalloc; -#endif - static VALUE sym_newobj, sym_malloc, sym_method, sym_capi; - VALUE hash = Qnil, key = Qnil; - VALUE major_by; - - if (SYMBOL_P(hash_or_key)) - key = hash_or_key; - else if (RB_TYPE_P(hash_or_key, T_HASH)) - hash = hash_or_key; - else - rb_raise(rb_eTypeError, "non-hash or symbol given"); - - if (sym_major_by == Qnil) { -#define S(s) sym_##s = ID2SYM(rb_intern_const(#s)) - S(major_by); - S(gc_by); - S(immediate_sweep); - S(have_finalizer); - S(nofree); - S(oldgen); - S(shady); - S(rescan); - S(stress); -#if RGENGC_ESTIMATE_OLDMALLOC - S(oldmalloc); -#endif - S(newobj); - S(malloc); - S(method); - S(capi); -#undef S - } - -#define SET(name, attr) \ - if (key == sym_##name) \ - return (attr); \ - else if (hash != Qnil) \ - rb_hash_aset(hash, sym_##name, (attr)); - - major_by = - (flags & GPR_FLAG_MAJOR_BY_OLDGEN) ? sym_oldgen : - (flags & GPR_FLAG_MAJOR_BY_SHADY) ? sym_shady : - (flags & GPR_FLAG_MAJOR_BY_RESCAN) ? sym_rescan : - (flags & GPR_FLAG_MAJOR_BY_STRESS) ? sym_stress : -#if RGENGC_ESTIMATE_OLDMALLOC - (flags & GPR_FLAG_MAJOR_BY_OLDMALLOC) ? sym_oldmalloc : -#endif - (flags & GPR_FLAG_MAJOR_BY_NOFREE) ? sym_nofree : - Qnil; - SET(major_by, major_by); - - SET(gc_by, - (flags & GPR_FLAG_NEWOBJ) ? sym_newobj : - (flags & GPR_FLAG_MALLOC) ? sym_malloc : - (flags & GPR_FLAG_METHOD) ? sym_method : - (flags & GPR_FLAG_CAPI) ? sym_capi : - (flags & GPR_FLAG_STRESS) ? sym_stress : - Qnil - ); + if (!SYMBOL_P(key) && !RB_TYPE_P(key, T_HASH)) { + rb_raise(rb_eTypeError, "non-hash or symbol given"); + } - SET(have_finalizer, (flags & GPR_FLAG_HAVE_FINALIZE) ? Qtrue : Qfalse); - SET(immediate_sweep, (flags & GPR_FLAG_IMMEDIATE_SWEEP) ? Qtrue : Qfalse); -#undef SET + VALUE val = rb_gc_impl_latest_gc_info(rb_gc_get_objspace(), key); - if (key != Qnil) /* matched key should return above */ - rb_raise(rb_eArgError, "unknown key: %s", RSTRING_PTR(rb_id2str(SYM2ID(key)))); + if (val == Qundef) { + rb_raise(rb_eArgError, "unknown key: %"PRIsVALUE, rb_sym2str(key)); + } - return hash; + return val; } -VALUE -rb_gc_latest_gc_info(VALUE key) +static VALUE +gc_stat(rb_execution_context_t *ec, VALUE self, VALUE arg) // arg is (nil || hash || symbol) { - rb_objspace_t *objspace = &rb_objspace; - return gc_info_decode(objspace->profile.latest_gc_info, key); -} + if (NIL_P(arg)) { + arg = rb_hash_new(); + } + else if (!RB_TYPE_P(arg, T_HASH) && !SYMBOL_P(arg)) { + rb_raise(rb_eTypeError, "non-hash or symbol given"); + } -/* - * call-seq: - * GC.latest_gc_info -> {:gc_by=>:newobj} - * GC.latest_gc_info(hash) -> hash - * GC.latest_gc_info(:major_by) -> :malloc - * - * Returns information about the most recent garbage collection. - */ + VALUE ret = rb_gc_impl_stat(rb_gc_get_objspace(), arg); -static VALUE -gc_latest_gc_info(int argc, VALUE *argv, VALUE self) -{ - rb_objspace_t *objspace = &rb_objspace; - VALUE arg = Qnil; + if (ret == Qundef) { + GC_ASSERT(SYMBOL_P(arg)); - if (rb_scan_args(argc, argv, "01", &arg) == 1) { - if (!SYMBOL_P(arg) && !RB_TYPE_P(arg, T_HASH)) { - rb_raise(rb_eTypeError, "non-hash or symbol given"); - } + rb_raise(rb_eArgError, "unknown key: %"PRIsVALUE, rb_sym2str(arg)); } - if (arg == Qnil) - arg = rb_hash_new(); - - return gc_info_decode(objspace->profile.latest_gc_info, arg); + return ret; } -static VALUE -gc_stat_internal(VALUE hash_or_sym, size_t *out) -{ - static VALUE sym_count; - static VALUE sym_heap_used, sym_heap_length, sym_heap_increment; - static VALUE sym_heap_live_slot, sym_heap_free_slot, sym_heap_final_slot, sym_heap_swept_slot; - static VALUE sym_heap_eden_page_length, sym_heap_tomb_page_length; - static VALUE sym_total_allocated_object, sym_total_freed_object; - static VALUE sym_malloc_increase, sym_malloc_limit; -#if USE_RGENGC - static VALUE sym_minor_gc_count, sym_major_gc_count; - static VALUE sym_remembered_shady_object, sym_remembered_shady_object_limit; - static VALUE sym_old_object, sym_old_object_limit; -#if RGENGC_ESTIMATE_OLDMALLOC - static VALUE sym_oldmalloc_increase, sym_oldmalloc_limit; -#endif -#if RGENGC_PROFILE - static VALUE sym_generated_normal_object_count, sym_generated_shady_object_count; - static VALUE sym_shade_operation_count, sym_promote_infant_count, sym_promote_young_count; - static VALUE sym_remembered_normal_object_count, sym_remembered_shady_object_count; -#endif /* RGENGC_PROFILE */ -#endif /* USE_RGENGC */ - - rb_objspace_t *objspace = &rb_objspace; - VALUE hash = Qnil, key = Qnil; - - if (RB_TYPE_P(hash_or_sym, T_HASH)) - hash = hash_or_sym; - else if (SYMBOL_P(hash_or_sym) && out) - key = hash_or_sym; - else - rb_raise(rb_eTypeError, "non-hash or symbol argument"); - - if (sym_count == 0) { -#define S(s) sym_##s = ID2SYM(rb_intern_const(#s)) - S(count); - S(heap_used); - S(heap_length); - S(heap_increment); - S(heap_live_slot); - S(heap_free_slot); - S(heap_final_slot); - S(heap_swept_slot); - S(heap_eden_page_length); - S(heap_tomb_page_length); - S(total_allocated_object); - S(total_freed_object); - S(malloc_increase); - S(malloc_limit); -#if USE_RGENGC - S(minor_gc_count); - S(major_gc_count); - S(remembered_shady_object); - S(remembered_shady_object_limit); - S(old_object); - S(old_object_limit); -#if RGENGC_ESTIMATE_OLDMALLOC - S(oldmalloc_increase); - S(oldmalloc_limit); -#endif -#if RGENGC_PROFILE - S(generated_normal_object_count); - S(generated_shady_object_count); - S(shade_operation_count); - S(promote_infant_count); - S(promote_young_count); - S(remembered_normal_object_count); - S(remembered_shady_object_count); -#endif /* USE_RGENGC */ -#endif /* RGENGC_PROFILE */ -#undef S - } - -#define SET(name, attr) \ - if (key == sym_##name) \ - return (*out = attr, Qnil); \ - else if (hash != Qnil) \ - rb_hash_aset(hash, sym_##name, SIZET2NUM(attr)); - - SET(count, objspace->profile.count); - - /* implementation dependent counters */ - SET(heap_used, heap_pages_used); - SET(heap_length, heap_pages_length); - SET(heap_increment, heap_pages_increment); - SET(heap_live_slot, objspace_live_slot(objspace)); - SET(heap_free_slot, objspace_free_slot(objspace)); - SET(heap_final_slot, heap_pages_final_slots); - SET(heap_swept_slot, heap_pages_swept_slots); - SET(heap_eden_page_length, heap_eden->page_length); - SET(heap_tomb_page_length, heap_tomb->page_length); - SET(total_allocated_object, objspace->profile.total_allocated_object_num); - SET(total_freed_object, objspace->profile.total_freed_object_num); - SET(malloc_increase, malloc_increase); - SET(malloc_limit, malloc_limit); -#if USE_RGENGC - SET(minor_gc_count, objspace->profile.minor_gc_count); - SET(major_gc_count, objspace->profile.major_gc_count); - SET(remembered_shady_object, objspace->rgengc.remembered_shady_object_count); - SET(remembered_shady_object_limit, objspace->rgengc.remembered_shady_object_limit); - SET(old_object, objspace->rgengc.old_object_count); - SET(old_object_limit, objspace->rgengc.old_object_limit); -#if RGENGC_ESTIMATE_OLDMALLOC - SET(oldmalloc_increase, objspace->rgengc.oldmalloc_increase); - SET(oldmalloc_limit, objspace->rgengc.oldmalloc_increase_limit); -#endif +size_t +rb_gc_stat(VALUE arg) +{ + if (!RB_TYPE_P(arg, T_HASH) && !SYMBOL_P(arg)) { + rb_raise(rb_eTypeError, "non-hash or symbol given"); + } -#if RGENGC_PROFILE - SET(generated_normal_object_count, objspace->profile.generated_normal_object_count); - SET(generated_shady_object_count, objspace->profile.generated_shady_object_count); - SET(shade_operation_count, objspace->profile.shade_operation_count); - SET(promote_infant_count, objspace->profile.promote_infant_count); -#if RGENGC_THREEGEN - SET(promote_young_count, objspace->profile.promote_young_count); -#endif - SET(remembered_normal_object_count, objspace->profile.remembered_normal_object_count); - SET(remembered_shady_object_count, objspace->profile.remembered_shady_object_count); -#endif /* RGENGC_PROFILE */ -#endif /* USE_RGENGC */ -#undef SET - - if (key != Qnil) /* matched key should return above */ - rb_raise(rb_eArgError, "unknown key: %s", RSTRING_PTR(rb_id2str(SYM2ID(key)))); - -#if defined(RGENGC_PROFILE) && RGENGC_PROFILE >= 2 - if (hash != Qnil) { - gc_count_add_each_types(hash, "generated_normal_object_count_types", objspace->profile.generated_normal_object_count_types); - gc_count_add_each_types(hash, "generated_shady_object_count_types", objspace->profile.generated_shady_object_count_types); - gc_count_add_each_types(hash, "shade_operation_count_types", objspace->profile.shade_operation_count_types); - gc_count_add_each_types(hash, "promote_infant_types", objspace->profile.promote_infant_types); -#if RGENGC_THREEGEN - gc_count_add_each_types(hash, "promote_young_types", objspace->profile.promote_young_types); -#endif - gc_count_add_each_types(hash, "remembered_normal_object_count_types", objspace->profile.remembered_normal_object_count_types); - gc_count_add_each_types(hash, "remembered_shady_object_count_types", objspace->profile.remembered_shady_object_count_types); + VALUE ret = rb_gc_impl_stat(rb_gc_get_objspace(), arg); + + if (ret == Qundef) { + GC_ASSERT(SYMBOL_P(arg)); + + rb_raise(rb_eArgError, "unknown key: %"PRIsVALUE, rb_sym2str(arg)); } -#endif - return hash; + if (SYMBOL_P(arg)) { + return NUM2SIZET(ret); + } + else { + return 0; + } } -/* - * call-seq: - * GC.stat -> Hash - * GC.stat(hash) -> hash - * GC.stat(:key) -> Numeric - * - * Returns a Hash containing information about the GC. - * - * The hash includes information about internal statistics about GC such as: - * - * { - * :count=>2, - * :heap_used=>9, - * :heap_length=>11, - * :heap_increment=>2, - * :heap_live_slot=>6836, - * :heap_free_slot=>519, - * :heap_final_slot=>0, - * :heap_swept_slot=>818, - * :total_allocated_object=>7674, - * :total_freed_object=>838, - * :malloc_increase=>181034, - * :malloc_limit=>16777216, - * :minor_gc_count=>2, - * :major_gc_count=>0, - * :remembered_shady_object=>55, - * :remembered_shady_object_limit=>0, - * :old_object=>2422, - * :old_object_limit=>0, - * :oldmalloc_increase=>277386, - * :oldmalloc_limit=>16777216 - * } - * - * The contents of the hash are implementation specific and may be changed in - * the future. - * - * This method is only expected to work on C Ruby. - * - */ - static VALUE -gc_stat(int argc, VALUE *argv, VALUE self) +gc_stat_heap(rb_execution_context_t *ec, VALUE self, VALUE heap_name, VALUE arg) { - VALUE arg = Qnil; + if (NIL_P(arg)) { + arg = rb_hash_new(); + } - if (rb_scan_args(argc, argv, "01", &arg) == 1) { - if (SYMBOL_P(arg)) { - size_t value = 0; - gc_stat_internal(arg, &value); - return SIZET2NUM(value); - } else if (!RB_TYPE_P(arg, T_HASH)) { - rb_raise(rb_eTypeError, "non-hash or symbol given"); - } + if (NIL_P(heap_name)) { + if (!RB_TYPE_P(arg, T_HASH)) { + rb_raise(rb_eTypeError, "non-hash given"); + } + } + else if (FIXNUM_P(heap_name)) { + if (!SYMBOL_P(arg) && !RB_TYPE_P(arg, T_HASH)) { + rb_raise(rb_eTypeError, "non-hash or symbol given"); + } + } + else { + rb_raise(rb_eTypeError, "heap_name must be nil or an Integer"); } - if (arg == Qnil) { - arg = rb_hash_new(); + VALUE ret = rb_gc_impl_stat_heap(rb_gc_get_objspace(), heap_name, arg); + + if (ret == Qundef) { + GC_ASSERT(SYMBOL_P(arg)); + + rb_raise(rb_eArgError, "unknown key: %"PRIsVALUE, rb_sym2str(arg)); } - gc_stat_internal(arg, 0); - return arg; + + return ret; } -size_t -rb_gc_stat(VALUE key) +static VALUE +gc_config_get(rb_execution_context_t *ec, VALUE self) { - if (SYMBOL_P(key)) { - size_t value = 0; - gc_stat_internal(key, &value); - return value; - } else { - gc_stat_internal(key, 0); - return 0; - } -} + VALUE cfg_hash = rb_gc_impl_config_get(rb_gc_get_objspace()); + rb_hash_aset(cfg_hash, sym("implementation"), rb_fstring_cstr(rb_gc_impl_active_gc_name())); -/* - * call-seq: - * GC.stress -> fixnum, true or false - * - * Returns current status of GC stress mode. - */ + return cfg_hash; +} static VALUE -gc_stress_get(VALUE self) +gc_config_set(rb_execution_context_t *ec, VALUE self, VALUE hash) { - rb_objspace_t *objspace = &rb_objspace; - return ruby_gc_stress; + void *objspace = rb_gc_get_objspace(); + + rb_gc_impl_config_set(objspace, hash); + + return Qnil; } -/* - * call-seq: - * GC.stress = bool -> bool - * - * Updates the GC stress mode. - * - * When stress mode is enabled, the GC is invoked at every GC opportunity: - * all memory and object allocations. - * - * Enabling stress mode will degrade performance, it is only for debugging. - */ +static VALUE +gc_stress_get(rb_execution_context_t *ec, VALUE self) +{ + return rb_gc_impl_stress_get(rb_gc_get_objspace()); +} static VALUE -gc_stress_set(VALUE self, VALUE flag) +gc_stress_set_m(rb_execution_context_t *ec, VALUE self, VALUE flag) { - rb_objspace_t *objspace = &rb_objspace; - rb_secure(2); - ruby_gc_stress = FIXNUM_P(flag) ? flag : (RTEST(flag) ? Qtrue : Qfalse); + rb_gc_impl_stress_set(rb_gc_get_objspace(), flag); + return flag; } -/* - * call-seq: - * GC.enable -> true or false - * - * Enables garbage collection, returning +true+ if garbage - * collection was previously disabled. - * - * GC.disable #=> false - * GC.enable #=> true - * GC.enable #=> false - * - */ +void +rb_gc_initial_stress_set(VALUE flag) +{ + initial_stress = flag; +} + +size_t * +rb_gc_heap_sizes(void) +{ + return rb_gc_impl_heap_sizes(rb_gc_get_objspace()); +} VALUE rb_gc_enable(void) { - rb_objspace_t *objspace = &rb_objspace; - int old = dont_gc; - - dont_gc = FALSE; - return old ? Qtrue : Qfalse; + return rb_objspace_gc_enable(rb_gc_get_objspace()); } -/* - * call-seq: - * GC.disable -> true or false - * - * Disables garbage collection, returning +true+ if garbage - * collection was already disabled. - * - * GC.disable #=> false - * GC.disable #=> true - * - */ - VALUE -rb_gc_disable(void) +rb_objspace_gc_enable(void *objspace) { - rb_objspace_t *objspace = &rb_objspace; - int old = dont_gc; - - gc_rest_sweep(objspace); - - dont_gc = TRUE; - return old ? Qtrue : Qfalse; + bool disabled = !rb_gc_impl_gc_enabled_p(objspace); + rb_gc_impl_gc_enable(objspace); + return RBOOL(disabled); } -static int -get_envparam_int(const char *name, unsigned int *default_value, int lower_bound) -{ - char *ptr = getenv(name); - int val; - - if (ptr != NULL) { - val = atoi(ptr); - if (val > lower_bound) { - if (RTEST(ruby_verbose)) fprintf(stderr, "%s=%d (default value: %d)\n", name, val, *default_value); - *default_value = val; - return 1; - } - else { - if (RTEST(ruby_verbose)) fprintf(stderr, "%s=%d (default value: %d) is ignored because it must be greater than %d.\n", name, val, *default_value, lower_bound); - } - } - return 0; +static VALUE +gc_enable(rb_execution_context_t *ec, VALUE _) +{ + return rb_gc_enable(); } -static int -get_envparam_double(const char *name, double *default_value, double lower_bound) -{ - char *ptr = getenv(name); - double val; - - if (ptr != NULL) { - val = strtod(ptr, NULL); - if (val > lower_bound) { - if (RTEST(ruby_verbose)) fprintf(stderr, "%s=%f (%f)\n", name, val, *default_value); - *default_value = val; - return 1; - } - else { - if (RTEST(ruby_verbose)) fprintf(stderr, "%s=%f (default value: %f) is ignored because it must be greater than %f.\n", name, val, *default_value, lower_bound); - } - } - return 0; +static VALUE +gc_disable_no_rest(void *objspace) +{ + bool disabled = !rb_gc_impl_gc_enabled_p(objspace); + rb_gc_impl_gc_disable(objspace, false); + return RBOOL(disabled); } -static void -gc_set_initial_pages(void) +VALUE +rb_gc_disable_no_rest(void) { - size_t min_pages; - rb_objspace_t *objspace = &rb_objspace; - - min_pages = gc_params.heap_init_slots / HEAP_OBJ_LIMIT; - if (min_pages > heap_eden->page_length) { - heap_add_pages(objspace, heap_eden, min_pages - heap_eden->page_length); - } + return gc_disable_no_rest(rb_gc_get_objspace()); } -/* - * GC tuning environment variables - * - * * RUBY_GC_HEAP_INIT_SLOTS - * - Initial allocation slots. - * * RUBY_GC_HEAP_FREE_SLOTS - * - Prepare at least this ammount of slots after GC. - * - Allocate slots if there are not enough slots. - * * RUBY_GC_HEAP_GROWTH_FACTOR (new from 2.1) - * - Allocate slots by this factor. - * - (next slots number) = (current slots number) * (this factor) - * * RUBY_GC_HEAP_GROWTH_MAX_SLOTS (new from 2.1) - * - Allocation rate is limited to this factor. - * * RUBY_GC_HEAP_OLDOBJECT_LIMIT_FACTOR (new from 2.1.1) - * - Do full GC when the number of old objects is more than R * N - * where R is this factor and - * N is the number of old objects just after last full GC. - * - * * obsolete - * * RUBY_FREE_MIN -> RUBY_GC_HEAP_FREE_SLOTS (from 2.1) - * * RUBY_HEAP_MIN_SLOTS -> RUBY_GC_HEAP_INIT_SLOTS (from 2.1) - * - * * RUBY_GC_MALLOC_LIMIT - * * RUBY_GC_MALLOC_LIMIT_MAX (new from 2.1) - * * RUBY_GC_MALLOC_LIMIT_GROWTH_FACTOR (new from 2.1) - * - * * RUBY_GC_OLDMALLOC_LIMIT (new from 2.1) - * * RUBY_GC_OLDMALLOC_LIMIT_MAX (new from 2.1) - * * RUBY_GC_OLDMALLOC_LIMIT_GROWTH_FACTOR (new from 2.1) - */ - -void -ruby_gc_set_params(int safe_level) +VALUE +rb_gc_disable(void) { - if (safe_level > 0) return; - - /* RUBY_GC_HEAP_FREE_SLOTS */ - if (get_envparam_int("RUBY_GC_HEAP_FREE_SLOTS", &gc_params.heap_free_slots, 0)) { - /* ok */ - } - else if (get_envparam_int("RUBY_FREE_MIN", &gc_params.heap_free_slots, 0)) { - rb_warn("RUBY_FREE_MIN is obsolete. Use RUBY_GC_HEAP_FREE_SLOTS instead."); - } - - /* RUBY_GC_HEAP_INIT_SLOTS */ - if (get_envparam_int("RUBY_GC_HEAP_INIT_SLOTS", &gc_params.heap_init_slots, 0)) { - gc_set_initial_pages(); - } - else if (get_envparam_int("RUBY_HEAP_MIN_SLOTS", &gc_params.heap_init_slots, 0)) { - rb_warn("RUBY_HEAP_MIN_SLOTS is obsolete. Use RUBY_GC_HEAP_INIT_SLOTS instead."); - gc_set_initial_pages(); - } - - get_envparam_double("RUBY_GC_HEAP_GROWTH_FACTOR", &gc_params.growth_factor, 1.0); - get_envparam_int ("RUBY_GC_HEAP_GROWTH_MAX_SLOTS", &gc_params.growth_max_slots, 0); - get_envparam_double("RUBY_GC_HEAP_OLDOBJECT_LIMIT_FACTOR", &gc_params.oldobject_limit_factor, 0.0); + return rb_objspace_gc_disable(rb_gc_get_objspace()); +} - get_envparam_int("RUBY_GC_MALLOC_LIMIT", &gc_params.malloc_limit_min, 0); - get_envparam_int("RUBY_GC_MALLOC_LIMIT_MAX", &gc_params.malloc_limit_max, 0); - get_envparam_double("RUBY_GC_MALLOC_LIMIT_GROWTH_FACTOR", &gc_params.malloc_limit_growth_factor, 1.0); +VALUE +rb_objspace_gc_disable(void *objspace) +{ + bool disabled = !rb_gc_impl_gc_enabled_p(objspace); + rb_gc_impl_gc_disable(objspace, true); + return RBOOL(disabled); +} -#if RGENGC_ESTIMATE_OLDMALLOC - if (get_envparam_int("RUBY_GC_OLDMALLOC_LIMIT", &gc_params.oldmalloc_limit_min, 0)) { - rb_objspace_t *objspace = &rb_objspace; - objspace->rgengc.oldmalloc_increase_limit = gc_params.oldmalloc_limit_min; - } - get_envparam_int("RUBY_GC_OLDMALLOC_LIMIT_MAX", &gc_params.oldmalloc_limit_max, 0); - get_envparam_double("RUBY_GC_OLDMALLOC_LIMIT_GROWTH_FACTOR", &gc_params.oldmalloc_limit_growth_factor, 1.0); -#endif +static VALUE +gc_disable(rb_execution_context_t *ec, VALUE _) +{ + return rb_gc_disable(); } +// TODO: think about moving ruby_gc_set_params into Init_heap or Init_gc void -rb_gc_set_params(void) +ruby_gc_set_params(void) { - ruby_gc_set_params(rb_safe_level()); + rb_gc_impl_set_params(rb_gc_get_objspace()); } void rb_objspace_reachable_objects_from(VALUE obj, void (func)(VALUE, void *), void *data) { - rb_objspace_t *objspace = &rb_objspace; + RB_VM_LOCKING() { + if (rb_gc_impl_during_gc_p(rb_gc_get_objspace())) rb_bug("rb_objspace_reachable_objects_from() is not supported while during GC"); + + if (!RB_SPECIAL_CONST_P(obj)) { + rb_vm_t *vm = GET_VM(); + struct gc_mark_func_data_struct *prev_mfd = vm->gc.mark_func_data; + struct gc_mark_func_data_struct mfd = { + .mark_func = func, + .data = data, + }; - if (is_markable_object(objspace, obj)) { - struct mark_func_data_struct mfd; - mfd.mark_func = func; - mfd.data = data; - objspace->mark_func_data = &mfd; - gc_mark_children(objspace, obj); - objspace->mark_func_data = 0; + vm->gc.mark_func_data = &mfd; + rb_gc_mark_children(rb_gc_get_objspace(), obj); + vm->gc.mark_func_data = prev_mfd; + } } } @@ -5809,1605 +4794,960 @@ root_objects_from(VALUE obj, void *ptr) void rb_objspace_reachable_objects_from_root(void (func)(const char *category, VALUE, void *), void *passing_data) { - rb_objspace_t *objspace = &rb_objspace; - struct root_objects_data data; - struct mark_func_data_struct mfd; + if (rb_gc_impl_during_gc_p(rb_gc_get_objspace())) rb_bug("rb_gc_impl_objspace_reachable_objects_from_root() is not supported while during GC"); - data.func = func; - data.data = passing_data; + rb_vm_t *vm = GET_VM(); - mfd.mark_func = root_objects_from; - mfd.data = &data; + struct root_objects_data data = { + .func = func, + .data = passing_data, + }; - objspace->mark_func_data = &mfd; - { - gc_mark_roots(objspace, TRUE, &data.category); - } - objspace->mark_func_data = 0; + struct gc_mark_func_data_struct *prev_mfd = vm->gc.mark_func_data; + struct gc_mark_func_data_struct mfd = { + .mark_func = root_objects_from, + .data = &data, + }; + + vm->gc.mark_func_data = &mfd; + rb_gc_save_machine_context(); + rb_gc_mark_roots(vm->gc.objspace, &data.category); + vm->gc.mark_func_data = prev_mfd; } /* - ------------------------ Extended allocator ------------------------ + ------------------------------ DEBUG ------------------------------ */ -static void objspace_xfree(rb_objspace_t *objspace, void *ptr, size_t size); - -static void * -negative_size_allocation_error_with_gvl(void *ptr) -{ - rb_raise(rb_eNoMemError, "%s", (const char *)ptr); - return 0; /* should not be reached */ -} - -static void -negative_size_allocation_error(const char *msg) -{ - if (ruby_thread_has_gvl_p()) { - rb_raise(rb_eNoMemError, "%s", msg); - } - else { - if (ruby_native_thread_p()) { - rb_thread_call_with_gvl(negative_size_allocation_error_with_gvl, (void *)msg); - } - else { - fprintf(stderr, "[FATAL] %s\n", msg); - exit(EXIT_FAILURE); - } - } -} - -static void * -ruby_memerror_body(void *dummy) -{ - rb_memerror(); - return 0; -} - -static void -ruby_memerror(void) -{ - if (ruby_thread_has_gvl_p()) { - rb_memerror(); - } - else { - if (ruby_native_thread_p()) { - rb_thread_call_with_gvl(ruby_memerror_body, 0); - } - else { - /* no ruby thread */ - fprintf(stderr, "[FATAL] failed to allocate memory\n"); - exit(EXIT_FAILURE); - } - } -} - -void -rb_memerror(void) -{ - rb_thread_t *th = GET_THREAD(); - if (!nomem_error || - rb_thread_raised_p(th, RAISED_NOMEMORY)) { - fprintf(stderr, "[FATAL] failed to allocate memory\n"); - exit(EXIT_FAILURE); - } - if (rb_thread_raised_p(th, RAISED_NOMEMORY)) { - rb_thread_raised_clear(th); - GET_THREAD()->errinfo = nomem_error; - JUMP_TAG(TAG_RAISE); - } - rb_thread_raised_set(th, RAISED_NOMEMORY); - rb_exc_raise(nomem_error); -} - -static void * -aligned_malloc(size_t alignment, size_t size) +static const char * +type_name(int type, VALUE obj) { - void *res; - -#if defined __MINGW32__ - res = __mingw_aligned_malloc(size, alignment); -#elif defined _WIN32 && !defined __CYGWIN__ - void *_aligned_malloc(size_t, size_t); - res = _aligned_malloc(size, alignment); -#elif defined(HAVE_POSIX_MEMALIGN) - if (posix_memalign(&res, alignment, size) == 0) { - return res; - } - else { - return NULL; + switch (type) { +#define TYPE_NAME(t) case (t): return #t; + TYPE_NAME(T_NONE); + TYPE_NAME(T_OBJECT); + TYPE_NAME(T_CLASS); + TYPE_NAME(T_MODULE); + TYPE_NAME(T_FLOAT); + TYPE_NAME(T_STRING); + TYPE_NAME(T_REGEXP); + TYPE_NAME(T_ARRAY); + TYPE_NAME(T_HASH); + TYPE_NAME(T_STRUCT); + TYPE_NAME(T_BIGNUM); + TYPE_NAME(T_FILE); + TYPE_NAME(T_MATCH); + TYPE_NAME(T_COMPLEX); + TYPE_NAME(T_RATIONAL); + TYPE_NAME(T_NIL); + TYPE_NAME(T_TRUE); + TYPE_NAME(T_FALSE); + TYPE_NAME(T_SYMBOL); + TYPE_NAME(T_FIXNUM); + TYPE_NAME(T_UNDEF); + TYPE_NAME(T_IMEMO); + TYPE_NAME(T_ICLASS); + TYPE_NAME(T_MOVED); + TYPE_NAME(T_ZOMBIE); + case T_DATA: + if (obj && rb_objspace_data_type_name(obj)) { + return rb_objspace_data_type_name(obj); + } + return "T_DATA"; +#undef TYPE_NAME } -#elif defined(HAVE_MEMALIGN) - res = memalign(alignment, size); -#else - char* aligned; - res = malloc(alignment + size + sizeof(void*)); - aligned = (char*)res + alignment + sizeof(void*); - aligned -= ((VALUE)aligned & (alignment - 1)); - ((void**)aligned)[-1] = res; - res = (void*)aligned; -#endif - -#if defined(_DEBUG) || GC_DEBUG - /* alignment must be a power of 2 */ - assert(((alignment - 1) & alignment) == 0); - assert(alignment % sizeof(void*) == 0); -#endif - return res; -} - -static void -aligned_free(void *ptr) -{ -#if defined __MINGW32__ - __mingw_aligned_free(ptr); -#elif defined _WIN32 && !defined __CYGWIN__ - _aligned_free(ptr); -#elif defined(HAVE_MEMALIGN) || defined(HAVE_POSIX_MEMALIGN) - free(ptr); -#else - free(((void**)ptr)[-1]); -#endif -} - -static inline size_t -objspace_malloc_size(rb_objspace_t *objspace, void *ptr, size_t hint) -{ -#ifdef HAVE_MALLOC_USABLE_SIZE - return malloc_usable_size(ptr); -#else - return hint; -#endif + return "unknown"; } -enum memop_type { - MEMOP_TYPE_MALLOC = 1, - MEMOP_TYPE_FREE = 2, - MEMOP_TYPE_REALLOC = 3 -}; - -static inline void -atomic_sub_nounderflow(size_t *var, size_t sub) +static const char * +obj_type_name(VALUE obj) { - if (sub == 0) return; - - while (1) { - size_t val = *var; - if (val < sub) sub = val; - if (ATOMIC_SIZE_CAS(*var, val, val-sub) == val) break; - } + return type_name(TYPE(obj), obj); } -static void -objspace_malloc_increase(rb_objspace_t *objspace, void *mem, size_t new_size, size_t old_size, enum memop_type type) +const char * +rb_method_type_name(rb_method_type_t type) { - if (new_size > old_size) { - ATOMIC_SIZE_ADD(malloc_increase, new_size - old_size); -#if RGENGC_ESTIMATE_OLDMALLOC - ATOMIC_SIZE_ADD(objspace->rgengc.oldmalloc_increase, new_size - old_size); -#endif - } - else { - atomic_sub_nounderflow(&malloc_increase, old_size - new_size); -#if RGENGC_ESTIMATE_OLDMALLOC - atomic_sub_nounderflow(&objspace->rgengc.oldmalloc_increase, old_size - new_size); -#endif - } - - if (type == MEMOP_TYPE_MALLOC) { - if (ruby_gc_stress && !ruby_disable_gc_stress) { - garbage_collect_with_gvl(objspace, FALSE, TRUE, GPR_FLAG_MALLOC); - } - else { - retry: - if (malloc_increase > malloc_limit) { - if (ruby_thread_has_gvl_p() && is_lazy_sweeping(heap_eden)) { - gc_rest_sweep(objspace); /* rest_sweep can reduce malloc_increase */ - goto retry; - } - garbage_collect_with_gvl(objspace, FALSE, TRUE, GPR_FLAG_MALLOC); - } - } - } - -#if MALLOC_ALLOCATED_SIZE - if (new_size >= old_size) { - ATOMIC_SIZE_ADD(objspace->malloc_params.allocated_size, new_size - old_size); - } - else { - size_t dec_size = old_size - new_size; - size_t allocated_size = objspace->malloc_params.allocated_size; - -#if MALLOC_ALLOCATED_SIZE_CHECK - if (allocated_size < dec_size) { - rb_bug("objspace_malloc_increase: underflow malloc_params.allocated_size."); - } -#endif - atomic_sub_nounderflow(objspace->malloc_params.allocated_size, dec_size); - } - - if (0) fprintf(stderr, "incraese - ptr: %p, type: %s, new_size: %d, old_size: %d\n", - mem, - type == MEMOP_TYPE_MALLOC ? "malloc" : - type == MEMOP_TYPE_FREE ? "free " : - type == MEMOP_TYPE_REALLOC ? "realloc": "error", - (int)new_size, (int)old_size); - switch (type) { - case MEMOP_TYPE_MALLOC: - ATOMIC_SIZE_INC(objspace->malloc_params.allocations); - break; - case MEMOP_TYPE_FREE: - { - size_t allocations = objspace->malloc_params.allocations; - if (allocations > 0) { - atomic_sub_nounderflow(objspace->malloc_params.allocations, 1); - } -#if MALLOC_ALLOCATED_SIZE_CHECK - else { - assert(objspace->malloc_params.allocations > 0); - } -#endif - } - break; - case MEMOP_TYPE_REALLOC: /* ignore */ break; - } -#endif + case VM_METHOD_TYPE_ISEQ: return "iseq"; + case VM_METHOD_TYPE_ATTRSET: return "attrset"; + case VM_METHOD_TYPE_IVAR: return "ivar"; + case VM_METHOD_TYPE_BMETHOD: return "bmethod"; + case VM_METHOD_TYPE_ALIAS: return "alias"; + case VM_METHOD_TYPE_REFINED: return "refined"; + case VM_METHOD_TYPE_CFUNC: return "cfunc"; + case VM_METHOD_TYPE_ZSUPER: return "zsuper"; + case VM_METHOD_TYPE_MISSING: return "missing"; + case VM_METHOD_TYPE_OPTIMIZED: return "optimized"; + case VM_METHOD_TYPE_UNDEF: return "undef"; + case VM_METHOD_TYPE_NOTIMPLEMENTED: return "notimplemented"; + } + rb_bug("rb_method_type_name: unreachable (type: %d)", type); } -static inline size_t -objspace_malloc_prepare(rb_objspace_t *objspace, size_t size) +static void +rb_raw_iseq_info(char *const buff, const size_t buff_size, const rb_iseq_t *iseq) { - if ((ssize_t)size < 0) { - negative_size_allocation_error("negative allocation size (or too big)"); + if (buff_size > 0 && ISEQ_BODY(iseq) && ISEQ_BODY(iseq)->location.label && !RB_TYPE_P(ISEQ_BODY(iseq)->location.pathobj, T_MOVED)) { + VALUE path = rb_iseq_path(iseq); + int n = ISEQ_BODY(iseq)->location.first_lineno; + snprintf(buff, buff_size, " %s@%s:%d", + RSTRING_PTR(ISEQ_BODY(iseq)->location.label), + RSTRING_PTR(path), n); } - if (size == 0) size = 1; - -#if CALC_EXACT_MALLOC_SIZE - size += sizeof(size_t); -#endif - - return size; -} - -static inline void * -objspace_malloc_fixup(rb_objspace_t *objspace, void *mem, size_t size) -{ -#if CALC_EXACT_MALLOC_SIZE - ((size_t *)mem)[0] = size; - mem = (size_t *)mem + 1; -#endif - - return mem; } -#define TRY_WITH_GC(alloc) do { \ - if (!(alloc) && \ - (!garbage_collect_with_gvl(objspace, 1, 1, GPR_FLAG_MALLOC) || /* full mark && immediate sweep */ \ - !(alloc))) { \ - ruby_memerror(); \ - } \ +static int +str_len_no_raise(VALUE str) +{ + long len = RSTRING_LEN(str); + if (len < 0) return 0; + if (len > INT_MAX) return INT_MAX; + return (int)len; +} + +#define BUFF_ARGS buff + pos, buff_size - pos +#define APPEND_F(...) if ((pos += snprintf(BUFF_ARGS, "" __VA_ARGS__)) >= buff_size) goto end +#define APPEND_S(s) do { \ + if ((pos + (int)rb_strlen_lit(s)) >= buff_size) { \ + goto end; \ + } \ + else { \ + memcpy(buff + pos, (s), rb_strlen_lit(s) + 1); \ + } \ } while (0) +#define C(c, s) ((c) != 0 ? (s) : " ") -static void * -objspace_xmalloc(rb_objspace_t *objspace, size_t size) +static size_t +rb_raw_obj_info_common(char *const buff, const size_t buff_size, const VALUE obj) { - void *mem; - - size = objspace_malloc_prepare(objspace, size); - TRY_WITH_GC(mem = malloc(size)); - size = objspace_malloc_size(objspace, mem, size); - objspace_malloc_increase(objspace, mem, size, 0, MEMOP_TYPE_MALLOC); - return objspace_malloc_fixup(objspace, mem, size); -} + size_t pos = 0; -static void * -objspace_xrealloc(rb_objspace_t *objspace, void *ptr, size_t new_size, size_t old_size) -{ - void *mem; + if (SPECIAL_CONST_P(obj)) { + APPEND_F("%s", obj_type_name(obj)); - if ((ssize_t)new_size < 0) { - negative_size_allocation_error("negative re-allocation size"); + if (FIXNUM_P(obj)) { + APPEND_F(" %ld", FIX2LONG(obj)); + } + else if (SYMBOL_P(obj)) { + APPEND_F(" %s", rb_id2name(SYM2ID(obj))); + } } + else { + // const int age = RVALUE_AGE_GET(obj); + + if (rb_gc_impl_pointer_to_heap_p(rb_gc_get_objspace(), (void *)obj)) { + APPEND_F("%p %s/", (void *)obj, obj_type_name(obj)); + // TODO: fixme + // APPEND_F("%p [%d%s%s%s%s%s%s] %s ", + // (void *)obj, age, + // C(RVALUE_UNCOLLECTIBLE_BITMAP(obj), "L"), + // C(RVALUE_MARK_BITMAP(obj), "M"), + // C(RVALUE_PIN_BITMAP(obj), "P"), + // C(RVALUE_MARKING_BITMAP(obj), "R"), + // C(RVALUE_WB_UNPROTECTED_BITMAP(obj), "U"), + // C(rb_objspace_garbage_object_p(obj), "G"), + // obj_type_name(obj)); + } + else { + /* fake */ + // APPEND_F("%p [%dXXXX] %s", + // (void *)obj, age, + // obj_type_name(obj)); + } - if (!ptr) return objspace_xmalloc(objspace, new_size); - - /* - * The behavior of realloc(ptr, 0) is implementation defined. - * Therefore we don't use realloc(ptr, 0) for portability reason. - * see http://www.open-std.org/jtc1/sc22/wg14/www/docs/dr_400.htm - */ - if (new_size == 0) { - objspace_xfree(objspace, ptr, old_size); - return 0; + if (internal_object_p(obj)) { + /* ignore */ + } + else if (RBASIC(obj)->klass == 0) { + APPEND_S("(temporary internal)"); + } + else if (RTEST(RBASIC(obj)->klass)) { + VALUE class_path = rb_mod_name(RBASIC(obj)->klass); + if (!NIL_P(class_path)) { + APPEND_F("%s ", RSTRING_PTR(class_path)); + } + } } + end: -#if CALC_EXACT_MALLOC_SIZE - new_size += sizeof(size_t); - ptr = (size_t *)ptr - 1; - oldsize = ((size_t *)ptr)[0]; -#endif - - old_size = objspace_malloc_size(objspace, ptr, old_size); - TRY_WITH_GC(mem = realloc(ptr, new_size)); - new_size = objspace_malloc_size(objspace, mem, new_size); - -#if CALC_EXACT_MALLOC_SIZE - ((size_t *)mem)[0] = new_size; - mem = (size_t *)mem + 1; -#endif - - objspace_malloc_increase(objspace, mem, new_size, old_size, MEMOP_TYPE_REALLOC); - - return mem; + return pos; } -static void -objspace_xfree(rb_objspace_t *objspace, void *ptr, size_t old_size) -{ -#if CALC_EXACT_MALLOC_SIZE - ptr = ((size_t *)ptr) - 1; - oldsize = ((size_t*)ptr)[0]; -#endif - old_size = objspace_malloc_size(objspace, ptr, old_size); - - free(ptr); - - objspace_malloc_increase(objspace, ptr, 0, old_size, MEMOP_TYPE_FREE); -} +const char *rb_raw_obj_info(char *const buff, const size_t buff_size, VALUE obj); -void * -ruby_xmalloc(size_t size) -{ - return objspace_xmalloc(&rb_objspace, size); -} +static size_t +rb_raw_obj_info_buitin_type(char *const buff, const size_t buff_size, const VALUE obj, size_t pos) +{ + if (LIKELY(pos < buff_size) && !SPECIAL_CONST_P(obj)) { + const enum ruby_value_type type = BUILTIN_TYPE(obj); + + switch (type) { + case T_NODE: + UNEXPECTED_NODE(rb_raw_obj_info); + break; + case T_ARRAY: + if (ARY_SHARED_P(obj)) { + APPEND_S("shared -> "); + rb_raw_obj_info(BUFF_ARGS, ARY_SHARED_ROOT(obj)); + } + else { + APPEND_F("[%s%s%s] ", + C(ARY_EMBED_P(obj), "E"), + C(ARY_SHARED_P(obj), "S"), + C(ARY_SHARED_ROOT_P(obj), "R")); + + if (ARY_EMBED_P(obj)) { + APPEND_F("len: %ld (embed)", + RARRAY_LEN(obj)); + } + else { + APPEND_F("len: %ld, capa:%ld ptr:%p", + RARRAY_LEN(obj), + RARRAY(obj)->as.heap.aux.capa, + (void *)RARRAY_CONST_PTR(obj)); + } + } + break; + case T_STRING: { + APPEND_F("[%s%s] ", + C(FL_TEST(obj, RSTRING_FSTR), "F"), + C(RB_OBJ_FROZEN(obj), "R")); + + if (STR_SHARED_P(obj)) { + APPEND_F(" [shared] len: %ld", RSTRING_LEN(obj)); + } + else { + if (STR_EMBED_P(obj)) APPEND_S(" [embed]"); -static inline size_t -xmalloc2_size(size_t n, size_t size) -{ - size_t len = size * n; - if (n != 0 && size != len / n) { - rb_raise(rb_eArgError, "malloc: possible integer overflow"); + APPEND_F(" len: %ld, capa: %" PRIdSIZE, RSTRING_LEN(obj), rb_str_capacity(obj)); + } + APPEND_F(" \"%.*s\"", str_len_no_raise(obj), RSTRING_PTR(obj)); + break; + } + case T_SYMBOL: { + VALUE fstr = RSYMBOL(obj)->fstr; + ID id = RSYMBOL(obj)->id; + if (RB_TYPE_P(fstr, T_STRING)) { + APPEND_F(":%s id:%d", RSTRING_PTR(fstr), (unsigned int)id); + } + else { + APPEND_F("(%p) id:%d", (void *)fstr, (unsigned int)id); + } + break; + } + case T_MOVED: { + APPEND_F("-> %p", (void*)gc_location_internal(rb_gc_get_objspace(), obj)); + break; + } + case T_HASH: { + APPEND_F("[%c] %"PRIdSIZE, + RHASH_AR_TABLE_P(obj) ? 'A' : 'S', + RHASH_SIZE(obj)); + break; + } + case T_CLASS: + case T_MODULE: + { + VALUE class_path = rb_mod_name(obj); + if (!NIL_P(class_path)) { + APPEND_F("%s", RSTRING_PTR(class_path)); + } + else { + APPEND_S("(anon)"); + } + break; + } + case T_ICLASS: + { + VALUE class_path = rb_mod_name(RBASIC_CLASS(obj)); + if (!NIL_P(class_path)) { + APPEND_F("src:%s", RSTRING_PTR(class_path)); + } + break; + } + case T_OBJECT: + { + if (FL_TEST_RAW(obj, ROBJECT_HEAP)) { + if (rb_obj_shape_complex_p(obj)) { + size_t hash_len = rb_st_table_size(ROBJECT_FIELDS_HASH(obj)); + APPEND_F("(complex) len:%zu", hash_len); + } + else { + APPEND_F("(embed) len:%d capa:%d", RSHAPE_LEN(RBASIC_SHAPE_ID(obj)), ROBJECT_FIELDS_CAPACITY(obj)); + } + } + else { + APPEND_F("len:%d capa:%d ptr:%p", RSHAPE_LEN(RBASIC_SHAPE_ID(obj)), ROBJECT_FIELDS_CAPACITY(obj), (void *)ROBJECT_FIELDS(obj)); + } + } + break; + case T_DATA: { + const struct rb_block *block; + const rb_iseq_t *iseq; + if (rb_obj_is_proc(obj) && + (block = vm_proc_block(obj)) != NULL && + (vm_block_type(block) == block_type_iseq) && + (iseq = vm_block_iseq(block)) != NULL) { + rb_raw_iseq_info(BUFF_ARGS, iseq); + } + else if (rb_ractor_p(obj)) { + rb_ractor_t *r = (void *)DATA_PTR(obj); + if (r) { + APPEND_F("r:%d", r->pub.id); + } + } + break; + } + case T_IMEMO: { + APPEND_F("<%s> ", rb_imemo_name(imemo_type(obj))); + + switch (imemo_type(obj)) { + case imemo_ment: + { + const rb_method_entry_t *me = (const rb_method_entry_t *)obj; + + APPEND_F(":%s (%s%s%s%s) type:%s aliased:%d owner:%p defined_class:%p", + rb_id2name(me->called_id), + METHOD_ENTRY_VISI(me) == METHOD_VISI_PUBLIC ? "pub" : + METHOD_ENTRY_VISI(me) == METHOD_VISI_PRIVATE ? "pri" : "pro", + METHOD_ENTRY_COMPLEMENTED(me) ? ",cmp" : "", + METHOD_ENTRY_CACHED(me) ? ",cc" : "", + METHOD_ENTRY_INVALIDATED(me) ? ",inv" : "", + me->def ? rb_method_type_name(me->def->type) : "NULL", + me->def ? me->def->aliased : -1, + (void *)me->owner, // obj_info(me->owner), + (void *)me->defined_class); //obj_info(me->defined_class))); + + if (me->def) { + switch (me->def->type) { + case VM_METHOD_TYPE_ISEQ: + APPEND_S(" (iseq:"); + rb_raw_obj_info(BUFF_ARGS, (VALUE)me->def->body.iseq.iseqptr); + APPEND_S(")"); + break; + default: + break; + } + } + + break; + } + case imemo_iseq: { + const rb_iseq_t *iseq = (const rb_iseq_t *)obj; + rb_raw_iseq_info(BUFF_ARGS, iseq); + break; + } + case imemo_callinfo: + { + const struct rb_callinfo *ci = (const struct rb_callinfo *)obj; + APPEND_F("(mid:%s, flag:%x argc:%d, kwarg:%s)", + rb_id2name(vm_ci_mid(ci)), + vm_ci_flag(ci), + vm_ci_argc(ci), + vm_ci_kwarg(ci) ? "available" : "NULL"); + break; + } + case imemo_callcache: + { + const struct rb_callcache *cc = (const struct rb_callcache *)obj; + VALUE class_path = vm_cc_valid(cc) ? rb_mod_name(cc->klass) : Qnil; + const rb_callable_method_entry_t *cme = vm_cc_cme(cc); + + APPEND_F("(klass:%s cme:%s%s (%p) call:%p", + NIL_P(class_path) ? (vm_cc_valid(cc) ? "??" : "<NULL>") : RSTRING_PTR(class_path), + cme ? rb_id2name(cme->called_id) : "<NULL>", + cme ? (METHOD_ENTRY_INVALIDATED(cme) ? " [inv]" : "") : "", + (void *)cme, + (void *)(uintptr_t)vm_cc_call(cc)); + break; + } + default: + break; + } + } + default: + break; + } } - return len; -} + end: -void * -ruby_xmalloc2(size_t n, size_t size) -{ - return objspace_xmalloc(&rb_objspace, xmalloc2_size(n, size)); + return pos; } -static void * -objspace_xcalloc(rb_objspace_t *objspace, size_t count, size_t elsize) -{ - void *mem; - size_t size; - - size = xmalloc2_size(count, elsize); - size = objspace_malloc_prepare(objspace, size); - - TRY_WITH_GC(mem = calloc(1, size)); - return objspace_malloc_fixup(objspace, mem, size); -} +#undef C -void * -ruby_xcalloc(size_t n, size_t size) +#ifdef RUBY_ASAN_ENABLED +void +rb_asan_poison_object(VALUE obj) { - return objspace_xcalloc(&rb_objspace, n, size); + MAYBE_UNUSED(struct RVALUE *) ptr = (void *)obj; + asan_poison_memory_region(ptr, rb_gc_obj_slot_size(obj)); } -#ifdef ruby_sized_xrealloc -#undef ruby_sized_xrealloc -#endif -void * -ruby_sized_xrealloc(void *ptr, size_t new_size, size_t old_size) +void +rb_asan_unpoison_object(VALUE obj, bool newobj_p) { - return objspace_xrealloc(&rb_objspace, ptr, new_size, old_size); + MAYBE_UNUSED(struct RVALUE *) ptr = (void *)obj; + asan_unpoison_memory_region(ptr, rb_gc_obj_slot_size(obj), newobj_p); } void * -ruby_xrealloc(void *ptr, size_t new_size) +rb_asan_poisoned_object_p(VALUE obj) { - return ruby_sized_xrealloc(ptr, new_size, 0); + MAYBE_UNUSED(struct RVALUE *) ptr = (void *)obj; + return __asan_region_is_poisoned(ptr, rb_gc_obj_slot_size(obj)); } - -#ifdef ruby_sized_xrealloc2 -#undef ruby_sized_xrealloc2 #endif -void * -ruby_sized_xrealloc2(void *ptr, size_t n, size_t size, size_t old_n) + +static void +raw_obj_info(char *const buff, const size_t buff_size, VALUE obj) { - size_t len = size * n; - if (n != 0 && size != len / n) { - rb_raise(rb_eArgError, "realloc: possible integer overflow"); - } - return objspace_xrealloc(&rb_objspace, ptr, len, old_n * size); + size_t pos = rb_raw_obj_info_common(buff, buff_size, obj); + pos = rb_raw_obj_info_buitin_type(buff, buff_size, obj, pos); + if (pos >= buff_size) {} // truncated } -void * -ruby_xrealloc2(void *ptr, size_t n, size_t size) +const char * +rb_raw_obj_info(char *const buff, const size_t buff_size, VALUE obj) { - return ruby_sized_xrealloc2(ptr, n, size, 0); -} + void *objspace = rb_gc_get_objspace(); -#ifdef ruby_sized_xfree -#undef ruby_sized_xfree + if (SPECIAL_CONST_P(obj)) { + raw_obj_info(buff, buff_size, obj); + } + else if (!rb_gc_impl_pointer_to_heap_p(objspace, (const void *)obj)) { + snprintf(buff, buff_size, "out-of-heap:%p", (void *)obj); + } +#if 0 // maybe no need to check it? + else if (0 && rb_gc_impl_garbage_object_p(objspace, obj)) { + snprintf(buff, buff_size, "garbage:%p", (void *)obj); + } #endif -void -ruby_sized_xfree(void *x, size_t size) -{ - if (x) { - objspace_xfree(&rb_objspace, x, size); + else { + asan_unpoisoning_object(obj) { + raw_obj_info(buff, buff_size, obj); + } } + return buff; } -void -ruby_xfree(void *x) -{ - ruby_sized_xfree(x, 0); -} - -/* Mimic ruby_xmalloc, but need not rb_objspace. - * should return pointer suitable for ruby_xfree - */ -void * -ruby_mimmalloc(size_t size) -{ - void *mem; -#if CALC_EXACT_MALLOC_SIZE - size += sizeof(size_t); -#endif - mem = malloc(size); -#if CALC_EXACT_MALLOC_SIZE - /* set 0 for consistency of allocated_size/allocations */ - ((size_t *)mem)[0] = 0; - mem = (size_t *)mem + 1; -#endif - return mem; -} +#undef APPEND_S +#undef APPEND_F +#undef BUFF_ARGS -void -ruby_mimfree(void *ptr) +/* Increments *var atomically and resets *var to 0 when maxval is + * reached. Returns the wraparound old *var value (0...maxval). */ +static rb_atomic_t +atomic_inc_wraparound(rb_atomic_t *var, const rb_atomic_t maxval) { - size_t *mem = (size_t *)ptr; -#if CALC_EXACT_MALLOC_SIZE - mem = mem - 1; -#endif - free(mem); + rb_atomic_t oldval = RUBY_ATOMIC_FETCH_ADD(*var, 1); + if (RB_UNLIKELY(oldval >= maxval - 1)) { // wraparound *var + const rb_atomic_t newval = oldval + 1; + RUBY_ATOMIC_CAS(*var, newval, newval % maxval); + oldval %= maxval; + } + return oldval; } -#if MALLOC_ALLOCATED_SIZE -/* - * call-seq: - * GC.malloc_allocated_size -> Integer - * - * Returns the size of memory allocated by malloc(). - * - * Only available if ruby was built with +CALC_EXACT_MALLOC_SIZE+. - */ - -static VALUE -gc_malloc_allocated_size(VALUE self) +static const char * +obj_info(VALUE obj) { - return UINT2NUM(rb_objspace.malloc_params.allocated_size); -} - -/* - * call-seq: - * GC.malloc_allocations -> Integer - * - * Returns the number of malloc() allocations. - * - * Only available if ruby was built with +CALC_EXACT_MALLOC_SIZE+. - */ + if (RGENGC_OBJ_INFO) { + static struct { + rb_atomic_t index; + char buffers[10][0x100]; + } info = {0}; -static VALUE -gc_malloc_allocations(VALUE self) -{ - return UINT2NUM(rb_objspace.malloc_params.allocations); + rb_atomic_t index = atomic_inc_wraparound(&info.index, numberof(info.buffers)); + char *const buff = info.buffers[index]; + return rb_raw_obj_info(buff, sizeof(info.buffers[0]), obj); + } + return obj_type_name(obj); } -#endif /* - ------------------------------ WeakMap ------------------------------ + ------------------------ Extended allocator ------------------------ */ -struct weakmap { - st_table *obj2wmap; /* obj -> [ref,...] */ - st_table *wmap2obj; /* ref -> obj */ - VALUE final; +struct gc_raise_tag { + VALUE exc; + const char *fmt; + va_list *ap; }; -#define WMAP_DELETE_DEAD_OBJECT_IN_MARK 0 - -#if WMAP_DELETE_DEAD_OBJECT_IN_MARK -static int -wmap_mark_map(st_data_t key, st_data_t val, st_data_t arg) +static void * +gc_vraise(void *ptr) { - rb_objspace_t *objspace = (rb_objspace_t *)arg; - VALUE obj = (VALUE)val; - if (!is_live_object(objspace, obj)) return ST_DELETE; - return ST_CONTINUE; + struct gc_raise_tag *argv = ptr; + rb_vraise(argv->exc, argv->fmt, *argv->ap); + UNREACHABLE_RETURN(NULL); } -#endif static void -wmap_mark(void *ptr) +gc_raise(VALUE exc, const char *fmt, ...) { - struct weakmap *w = ptr; -#if WMAP_DELETE_DEAD_OBJECT_IN_MARK - if (w->obj2wmap) st_foreach(w->obj2wmap, wmap_mark_map, (st_data_t)&rb_objspace); -#endif - rb_gc_mark(w->final); -} + va_list ap; + va_start(ap, fmt); + struct gc_raise_tag argv = { + exc, fmt, &ap, + }; -static int -wmap_free_map(st_data_t key, st_data_t val, st_data_t arg) -{ - VALUE *ptr = (VALUE *)val; - ruby_sized_xfree(ptr, (ptr[0] + 1) * sizeof(VALUE)); - return ST_CONTINUE; -} + if (ruby_native_thread_p()) { + rb_thread_call_with_gvl(gc_vraise, &argv); + UNREACHABLE; + } + else { + /* Not in a ruby thread */ + fprintf(stderr, "%s", "[FATAL] "); + vfprintf(stderr, fmt, ap); + } -static void -wmap_free(void *ptr) -{ - struct weakmap *w = ptr; - st_foreach(w->obj2wmap, wmap_free_map, 0); - st_free_table(w->obj2wmap); - st_free_table(w->wmap2obj); + va_end(ap); + abort(); } -static int -wmap_memsize_map(st_data_t key, st_data_t val, st_data_t arg) +NORETURN(static void negative_size_allocation_error(const char *)); +static void +negative_size_allocation_error(const char *msg) { - VALUE *ptr = (VALUE *)val; - *(size_t *)arg += (ptr[0] + 1) * sizeof(VALUE); - return ST_CONTINUE; + gc_raise(rb_eNoMemError, "%s", msg); } -static size_t -wmap_memsize(const void *ptr) -{ - size_t size; - const struct weakmap *w = ptr; - if (!w) return 0; - size = sizeof(*w); - size += st_memsize(w->obj2wmap); - size += st_memsize(w->wmap2obj); - st_foreach(w->obj2wmap, wmap_memsize_map, (st_data_t)&size); - return size; -} - -static const rb_data_type_t weakmap_type = { - "weakmap", - { - wmap_mark, - wmap_free, - wmap_memsize, - }, - NULL, NULL, RUBY_TYPED_FREE_IMMEDIATELY -}; - -static VALUE -wmap_allocate(VALUE klass) +static void * +ruby_memerror_body(void *dummy) { - struct weakmap *w; - VALUE obj = TypedData_Make_Struct(klass, struct weakmap, &weakmap_type, w); - w->obj2wmap = st_init_numtable(); - w->wmap2obj = st_init_numtable(); - w->final = rb_obj_method(obj, ID2SYM(rb_intern("finalize"))); - return obj; + rb_memerror(); + return 0; } -static int -wmap_final_func(st_data_t *key, st_data_t *value, st_data_t arg, int existing) +NORETURN(static void ruby_memerror(void)); +RBIMPL_ATTR_MAYBE_UNUSED() +static void +ruby_memerror(void) { - VALUE wmap, *ptr, size, i, j; - if (!existing) return ST_STOP; - wmap = (VALUE)arg, ptr = (VALUE *)*value; - for (i = j = 1, size = ptr[0]; i <= size; ++i) { - if (ptr[i] != wmap) { - ptr[j++] = ptr[i]; - } - } - if (j == 1) { - ruby_sized_xfree(ptr, i * sizeof(VALUE)); - return ST_DELETE; + if (ruby_thread_has_gvl_p()) { + rb_memerror(); } - if (j < i) { - ptr = ruby_sized_xrealloc2(ptr, j + 1, sizeof(VALUE), i); - ptr[0] = j; - *value = (st_data_t)ptr; + else { + if (ruby_native_thread_p()) { + rb_thread_call_with_gvl(ruby_memerror_body, 0); + } + else { + /* no ruby thread */ + fprintf(stderr, "[FATAL] failed to allocate memory\n"); + } } - return ST_CONTINUE; + + /* We have discussions whether we should die here; */ + /* We might rethink about it later. */ + exit(EXIT_FAILURE); } -static VALUE -wmap_finalize(VALUE self, VALUE objid) +void +rb_memerror(void) { - st_data_t orig, wmap, data; - VALUE obj, *rids, i, size; - struct weakmap *w; + /* the `GET_VM()->special_exceptions` below assumes that + * the VM is reachable from the current thread. We should + * definitely make sure of that. */ + RUBY_ASSERT_ALWAYS(ruby_thread_has_gvl_p()); - TypedData_Get_Struct(self, struct weakmap, &weakmap_type, w); - /* Get reference from object id. */ - obj = obj_id_to_ref(objid); + rb_execution_context_t *ec = GET_EC(); + VALUE exc = GET_VM()->special_exceptions[ruby_error_nomemory]; - /* obj is original referenced object and/or weak reference. */ - orig = (st_data_t)obj; - if (st_delete(w->obj2wmap, &orig, &data)) { - rids = (VALUE *)data; - size = *rids++; - for (i = 0; i < size; ++i) { - wmap = (st_data_t)rids[i]; - st_delete(w->wmap2obj, &wmap, NULL); - } - ruby_sized_xfree((VALUE *)data, (size + 1) * sizeof(VALUE)); + if (!exc || + rb_ec_raised_p(ec, RAISED_NOMEMORY) || + rb_ec_vm_lock_rec(ec) != ec->tag->lock_rec) { + fprintf(stderr, "[FATAL] failed to allocate memory\n"); + exit(EXIT_FAILURE); } - - wmap = (st_data_t)obj; - if (st_delete(w->wmap2obj, &wmap, &orig)) { - wmap = (st_data_t)obj; - st_update(w->obj2wmap, orig, wmap_final_func, wmap); + if (rb_ec_raised_p(ec, RAISED_NOMEMORY)) { + rb_ec_raised_clear(ec); + } + else { + rb_ec_raised_set(ec, RAISED_NOMEMORY); + exc = ruby_vm_special_exception_copy(exc); } - return self; + ec->errinfo = exc; + EC_JUMP_TAG(ec, TAG_RAISE); } -struct wmap_iter_arg { - rb_objspace_t *objspace; - VALUE value; -}; - -static int -wmap_inspect_i(st_data_t key, st_data_t val, st_data_t arg) +bool +rb_memerror_reentered(void) { - VALUE str = (VALUE)arg; - VALUE k = (VALUE)key, v = (VALUE)val; + rb_execution_context_t *ec = GET_EC(); + return (ec && rb_ec_raised_p(ec, RAISED_NOMEMORY)); +} - if (RSTRING_PTR(str)[0] == '#') { - rb_str_cat2(str, ", "); +static void * +handle_malloc_failure(void *ptr) +{ + if (LIKELY(ptr)) { + return ptr; } else { - rb_str_cat2(str, ": "); - RSTRING_PTR(str)[0] = '#'; + ruby_memerror(); + UNREACHABLE_RETURN(ptr); } - k = SPECIAL_CONST_P(k) ? rb_inspect(k) : rb_any_to_s(k); - rb_str_append(str, k); - rb_str_cat2(str, " => "); - v = SPECIAL_CONST_P(v) ? rb_inspect(v) : rb_any_to_s(v); - rb_str_append(str, v); - OBJ_INFECT(str, k); - OBJ_INFECT(str, v); - - return ST_CONTINUE; } -static VALUE -wmap_inspect(VALUE self) -{ - VALUE str; - VALUE c = rb_class_name(CLASS_OF(self)); - struct weakmap *w; +static void *ruby_xmalloc_body(size_t size); - TypedData_Get_Struct(self, struct weakmap, &weakmap_type, w); - str = rb_sprintf("-<%"PRIsVALUE":%p", c, (void *)self); - if (w->wmap2obj) { - st_foreach(w->wmap2obj, wmap_inspect_i, str); - } - RSTRING_PTR(str)[0] = '#'; - rb_str_cat2(str, ">"); - return str; -} - -static int -wmap_each_i(st_data_t key, st_data_t val, st_data_t arg) +void * +ruby_xmalloc(size_t size) { - rb_objspace_t *objspace = (rb_objspace_t *)arg; - VALUE obj = (VALUE)val; - if (is_id_value(objspace, obj) && is_live_object(objspace, obj)) { - rb_yield_values(2, (VALUE)key, obj); - } - return ST_CONTINUE; + return handle_malloc_failure(ruby_xmalloc_body(size)); } -/* Iterates over keys and objects in a weakly referenced object */ -static VALUE -wmap_each(VALUE self) +static bool +malloc_gc_allowed(void) { - struct weakmap *w; - rb_objspace_t *objspace = &rb_objspace; + rb_ractor_t *r = rb_current_ractor_raw(false); - TypedData_Get_Struct(self, struct weakmap, &weakmap_type, w); - st_foreach(w->wmap2obj, wmap_each_i, (st_data_t)objspace); - return self; + return r == NULL || !r->malloc_gc_disabled; } -static int -wmap_each_key_i(st_data_t key, st_data_t val, st_data_t arg) +static void * +ruby_xmalloc_body(size_t size) { - rb_objspace_t *objspace = (rb_objspace_t *)arg; - VALUE obj = (VALUE)val; - if (is_id_value(objspace, obj) && is_live_object(objspace, obj)) { - rb_yield((VALUE)key); + if ((ssize_t)size < 0) { + negative_size_allocation_error("too large allocation size"); } - return ST_CONTINUE; + + return rb_gc_impl_malloc(rb_gc_get_objspace(), size, malloc_gc_allowed()); } -/* Iterates over keys and objects in a weakly referenced object */ -static VALUE -wmap_each_key(VALUE self) +void +ruby_malloc_size_overflow(size_t count, size_t elsize) { - struct weakmap *w; - rb_objspace_t *objspace = &rb_objspace; - - TypedData_Get_Struct(self, struct weakmap, &weakmap_type, w); - st_foreach(w->wmap2obj, wmap_each_key_i, (st_data_t)objspace); - return self; + rb_raise(rb_eArgError, + "malloc: possible integer overflow (%"PRIuSIZE"*%"PRIuSIZE")", + count, elsize); } -static int -wmap_each_value_i(st_data_t key, st_data_t val, st_data_t arg) +void +ruby_malloc_add_size_overflow(size_t x, size_t y) { - rb_objspace_t *objspace = (rb_objspace_t *)arg; - VALUE obj = (VALUE)val; - if (is_id_value(objspace, obj) && is_live_object(objspace, obj)) { - rb_yield(obj); - } - return ST_CONTINUE; + rb_raise(rb_eArgError, + "malloc: possible integer overflow (%"PRIuSIZE"+%"PRIuSIZE")", + x, y); } -/* Iterates over keys and objects in a weakly referenced object */ -static VALUE -wmap_each_value(VALUE self) -{ - struct weakmap *w; - rb_objspace_t *objspace = &rb_objspace; +static void *ruby_xmalloc2_body(size_t n, size_t size); - TypedData_Get_Struct(self, struct weakmap, &weakmap_type, w); - st_foreach(w->wmap2obj, wmap_each_value_i, (st_data_t)objspace); - return self; +void * +ruby_xmalloc2(size_t n, size_t size) +{ + return handle_malloc_failure(ruby_xmalloc2_body(n, size)); } -static int -wmap_keys_i(st_data_t key, st_data_t val, st_data_t arg) +static void * +ruby_xmalloc2_body(size_t n, size_t size) { - struct wmap_iter_arg *argp = (struct wmap_iter_arg *)arg; - rb_objspace_t *objspace = argp->objspace; - VALUE ary = argp->value; - VALUE obj = (VALUE)val; - if (is_id_value(objspace, obj) && is_live_object(objspace, obj)) { - rb_ary_push(ary, (VALUE)key); - } - return ST_CONTINUE; + return rb_gc_impl_malloc(rb_gc_get_objspace(), xmalloc2_size(n, size), malloc_gc_allowed()); } -/* Iterates over keys and objects in a weakly referenced object */ -static VALUE -wmap_keys(VALUE self) -{ - struct weakmap *w; - struct wmap_iter_arg args; +static void *ruby_xcalloc_body(size_t n, size_t size); - TypedData_Get_Struct(self, struct weakmap, &weakmap_type, w); - args.objspace = &rb_objspace; - args.value = rb_ary_new(); - st_foreach(w->wmap2obj, wmap_keys_i, (st_data_t)&args); - return args.value; +void * +ruby_xcalloc(size_t n, size_t size) +{ + return handle_malloc_failure(ruby_xcalloc_body(n, size)); } -static int -wmap_values_i(st_data_t key, st_data_t val, st_data_t arg) +static void * +ruby_xcalloc_body(size_t n, size_t size) { - struct wmap_iter_arg *argp = (struct wmap_iter_arg *)arg; - rb_objspace_t *objspace = argp->objspace; - VALUE ary = argp->value; - VALUE obj = (VALUE)val; - if (is_id_value(objspace, obj) && is_live_object(objspace, obj)) { - rb_ary_push(ary, obj); - } - return ST_CONTINUE; + return rb_gc_impl_calloc(rb_gc_get_objspace(), xmalloc2_size(n, size), malloc_gc_allowed()); } -/* Iterates over values and objects in a weakly referenced object */ -static VALUE -wmap_values(VALUE self) -{ - struct weakmap *w; - struct wmap_iter_arg args; +static void *ruby_xrealloc_sized_body(void *ptr, size_t new_size, size_t old_size); - TypedData_Get_Struct(self, struct weakmap, &weakmap_type, w); - args.objspace = &rb_objspace; - args.value = rb_ary_new(); - st_foreach(w->wmap2obj, wmap_values_i, (st_data_t)&args); - return args.value; +#ifdef ruby_xrealloc_sized +#undef ruby_xrealloc_sized +#endif +void * +ruby_xrealloc_sized(void *ptr, size_t new_size, size_t old_size) +{ + return handle_malloc_failure(ruby_xrealloc_sized_body(ptr, new_size, old_size)); } -static int -wmap_aset_update(st_data_t *key, st_data_t *val, st_data_t arg, int existing) +static void * +ruby_xrealloc_sized_body(void *ptr, size_t new_size, size_t old_size) { - VALUE size, *ptr, *optr; - if (existing) { - size = (ptr = optr = (VALUE *)*val)[0]; - ++size; - ptr = ruby_sized_xrealloc2(ptr, size + 1, sizeof(VALUE), size); + if ((ssize_t)new_size < 0) { + negative_size_allocation_error("too large allocation size"); } - else { - optr = 0; - size = 1; - ptr = ruby_xmalloc2(2, sizeof(VALUE)); - } - ptr[0] = size; - ptr[size] = (VALUE)arg; - if (ptr == optr) return ST_STOP; - *val = (st_data_t)ptr; - return ST_CONTINUE; + + return rb_gc_impl_realloc(rb_gc_get_objspace(), ptr, new_size, old_size, malloc_gc_allowed()); } -/* Creates a weak reference from the given key to the given value */ -static VALUE -wmap_aset(VALUE self, VALUE wmap, VALUE orig) +void * +ruby_xrealloc(void *ptr, size_t new_size) { - struct weakmap *w; - - TypedData_Get_Struct(self, struct weakmap, &weakmap_type, w); - should_be_finalizable(orig); - should_be_finalizable(wmap); - define_final0(orig, w->final); - define_final0(wmap, w->final); - st_update(w->obj2wmap, (st_data_t)orig, wmap_aset_update, wmap); - st_insert(w->wmap2obj, (st_data_t)wmap, (st_data_t)orig); - return nonspecial_obj_id(orig); + return ruby_xrealloc_sized(ptr, new_size, 0); } -/* Retrieves a weakly referenced object with the given key */ -static VALUE -wmap_aref(VALUE self, VALUE wmap) +static void *ruby_xrealloc2_sized_body(void *ptr, size_t n, size_t size, size_t old_n); + +#ifdef ruby_xrealloc2_sized +#undef ruby_xrealloc2_sized +#endif +void * +ruby_xrealloc2_sized(void *ptr, size_t n, size_t size, size_t old_n) { - st_data_t data; - VALUE obj; - struct weakmap *w; - rb_objspace_t *objspace = &rb_objspace; - - TypedData_Get_Struct(self, struct weakmap, &weakmap_type, w); - if (!st_lookup(w->wmap2obj, (st_data_t)wmap, &data)) return Qnil; - obj = (VALUE)data; - if (!is_id_value(objspace, obj)) return Qnil; - if (!is_live_object(objspace, obj)) return Qnil; - return obj; + return handle_malloc_failure(ruby_xrealloc2_sized_body(ptr, n, size, old_n)); } -/* Returns +true+ if +key+ is registered */ -static VALUE -wmap_has_key(VALUE self, VALUE key) +static void * +ruby_xrealloc2_sized_body(void *ptr, size_t n, size_t size, size_t old_n) { - return NIL_P(wmap_aref(self, key)) ? Qfalse : Qtrue; + size_t len = xmalloc2_size(n, size); + return rb_gc_impl_realloc(rb_gc_get_objspace(), ptr, len, old_n * size, malloc_gc_allowed()); } -static VALUE -wmap_size(VALUE self) +void * +ruby_xrealloc2(void *ptr, size_t n, size_t size) { - struct weakmap *w; - st_index_t n; - - TypedData_Get_Struct(self, struct weakmap, &weakmap_type, w); - n = w->wmap2obj->num_entries; -#if SIZEOF_ST_INDEX_T <= SIZEOF_LONG - return ULONG2NUM(n); -#else - return ULL2NUM(n); -#endif + return ruby_xrealloc2_sized(ptr, n, size, 0); } -/* - ------------------------------ GC profiler ------------------------------ -*/ - -#define GC_PROFILE_RECORD_DEFAULT_SIZE 100 - -static double -getrusage_time(void) -{ -#if defined(HAVE_CLOCK_GETTIME) && defined(CLOCK_PROCESS_CPUTIME_ID) - { - static int try_clock_gettime = 1; - struct timespec ts; - if (try_clock_gettime && clock_gettime(CLOCK_PROCESS_CPUTIME_ID, &ts) == 0) { - return ts.tv_sec + ts.tv_nsec * 1e-9; - } - else { - try_clock_gettime = 0; - } - } +#ifdef ruby_xfree_sized +#undef ruby_xfree_sized #endif - -#ifdef RUSAGE_SELF - { - struct rusage usage; - struct timeval time; - if (getrusage(RUSAGE_SELF, &usage) == 0) { - time = usage.ru_utime; - return time.tv_sec + time.tv_usec * 1e-6; +void +ruby_xfree_sized(void *x, size_t size) +{ + if (LIKELY(x)) { + /* It's possible for a C extension's pthread destructor function set by pthread_key_create + * to be called after ruby_vm_destruct and attempt to free memory. Fall back to mimfree in + * that case. */ + if (LIKELY(GET_VM())) { + rb_gc_impl_free(rb_gc_get_objspace(), x, size); } - } -#endif - -#ifdef _WIN32 - { - FILETIME creation_time, exit_time, kernel_time, user_time; - ULARGE_INTEGER ui; - LONG_LONG q; - double t; - - if (GetProcessTimes(GetCurrentProcess(), - &creation_time, &exit_time, &kernel_time, &user_time) != 0) { - memcpy(&ui, &user_time, sizeof(FILETIME)); - q = ui.QuadPart / 10L; - t = (DWORD)(q % 1000000L) * 1e-6; - q /= 1000000L; -#ifdef __GNUC__ - t += q; -#else - t += (double)(DWORD)(q >> 16) * (1 << 16); - t += (DWORD)q & ~(~0 << 16); -#endif - return t; + else { + ruby_mimfree(x); } } -#endif - - return 0.0; -} - -static inline void -gc_prof_setup_new_record(rb_objspace_t *objspace, int reason) -{ - if (objspace->profile.run) { - size_t index = objspace->profile.next_index; - gc_profile_record *record; - - /* create new record */ - objspace->profile.next_index++; - - if (!objspace->profile.records) { - objspace->profile.size = GC_PROFILE_RECORD_DEFAULT_SIZE; - objspace->profile.records = malloc(sizeof(gc_profile_record) * objspace->profile.size); - } - if (index >= objspace->profile.size) { - objspace->profile.size += 1000; - objspace->profile.records = realloc(objspace->profile.records, sizeof(gc_profile_record) * objspace->profile.size); - } - if (!objspace->profile.records) { - rb_bug("gc_profile malloc or realloc miss"); - } - record = objspace->profile.current_record = &objspace->profile.records[objspace->profile.next_index - 1]; - MEMZERO(record, gc_profile_record, 1); - - /* setup before-GC parameter */ - record->flags = reason | ((ruby_gc_stress && !ruby_disable_gc_stress) ? GPR_FLAG_STRESS : 0); -#if MALLOC_ALLOCATED_SIZE - record->allocated_size = malloc_allocated_size; -#endif -#if GC_PROFILE_DETAIL_MEMORY -#ifdef RUSAGE_SELF - { - struct rusage usage; - if (getrusage(RUSAGE_SELF, &usage) == 0) { - record->maxrss = usage.ru_maxrss; - record->minflt = usage.ru_minflt; - record->majflt = usage.ru_majflt; - } - } -#endif -#endif - } } -static inline void -gc_prof_timer_start(rb_objspace_t *objspace) +void +ruby_xfree(void *x) { - if (gc_prof_enabled(objspace)) { - gc_profile_record *record = gc_prof_record(objspace); -#if GC_PROFILE_MORE_DETAIL - record->prepare_time = objspace->profile.prepare_time; -#endif - record->gc_time = 0; - record->gc_invoke_time = getrusage_time(); - } + ruby_xfree_sized(x, 0); } -static double -elapsed_time_from(double time) +void * +rb_xmalloc_mul_add(size_t x, size_t y, size_t z) /* x * y + z */ { - double now = getrusage_time(); - if (now > time) { - return now - time; - } - else { - return 0; - } + size_t w = size_mul_add_or_raise(x, y, z, rb_eArgError); + return ruby_xmalloc(w); } -static inline void -gc_prof_timer_stop(rb_objspace_t *objspace) +void * +rb_xcalloc_mul_add(size_t x, size_t y, size_t z) /* x * y + z */ { - if (gc_prof_enabled(objspace)) { - gc_profile_record *record = gc_prof_record(objspace); - record->gc_time = elapsed_time_from(record->gc_invoke_time); - record->gc_invoke_time -= objspace->profile.invoke_time; - } + size_t w = size_mul_add_or_raise(x, y, z, rb_eArgError); + return ruby_xcalloc(w, 1); } -static inline void -gc_prof_mark_timer_start(rb_objspace_t *objspace) +void * +rb_xrealloc_mul_add(const void *p, size_t x, size_t y, size_t z) /* x * y + z */ { - if (RUBY_DTRACE_GC_MARK_BEGIN_ENABLED()) { - RUBY_DTRACE_GC_MARK_BEGIN(); - } -#if GC_PROFILE_MORE_DETAIL - if (gc_prof_enabled(objspace)) { - gc_prof_record(objspace)->gc_mark_time = getrusage_time(); - } -#endif + size_t w = size_mul_add_or_raise(x, y, z, rb_eArgError); + return ruby_xrealloc((void *)p, w); } -static inline void -gc_prof_mark_timer_stop(rb_objspace_t *objspace) +void * +rb_xmalloc_mul_add_mul(size_t x, size_t y, size_t z, size_t w) /* x * y + z * w */ { - if (RUBY_DTRACE_GC_MARK_END_ENABLED()) { - RUBY_DTRACE_GC_MARK_END(); - } -#if GC_PROFILE_MORE_DETAIL - if (gc_prof_enabled(objspace)) { - gc_profile_record *record = gc_prof_record(objspace); - record->gc_mark_time = elapsed_time_from(record->gc_mark_time); - } -#endif + size_t u = size_mul_add_mul_or_raise(x, y, z, w, rb_eArgError); + return ruby_xmalloc(u); } -static inline void -gc_prof_sweep_timer_start(rb_objspace_t *objspace) +void * +rb_xcalloc_mul_add_mul(size_t x, size_t y, size_t z, size_t w) /* x * y + z * w */ { - if (RUBY_DTRACE_GC_SWEEP_BEGIN_ENABLED()) { - RUBY_DTRACE_GC_SWEEP_BEGIN(); - } - if (gc_prof_enabled(objspace)) { - gc_profile_record *record = gc_prof_record(objspace); - - if (record->gc_time > 0 || GC_PROFILE_MORE_DETAIL) { - objspace->profile.gc_sweep_start_time = getrusage_time(); - } - } + size_t u = size_mul_add_mul_or_raise(x, y, z, w, rb_eArgError); + return ruby_xcalloc(u, 1); } -static inline void -gc_prof_sweep_timer_stop(rb_objspace_t *objspace) +/* Mimic ruby_xmalloc, but need not rb_objspace. + * should return pointer suitable for ruby_xfree + */ +void * +ruby_mimmalloc(size_t size) { - if (RUBY_DTRACE_GC_SWEEP_END_ENABLED()) { - RUBY_DTRACE_GC_SWEEP_END(); - } - - if (gc_prof_enabled(objspace)) { - double sweep_time; - gc_profile_record *record = gc_prof_record(objspace); - - if (record->gc_time > 0) { - sweep_time = elapsed_time_from(objspace->profile.gc_sweep_start_time); - /* need to accumulate GC time for lazy sweep after gc() */ - record->gc_time += sweep_time; - } - else if (GC_PROFILE_MORE_DETAIL) { - sweep_time = elapsed_time_from(objspace->profile.gc_sweep_start_time); - } - -#if GC_PROFILE_MORE_DETAIL - record->gc_sweep_time += sweep_time; - if (heap_pages_deferred_final) record->flags |= GPR_FLAG_HAVE_FINALIZE; + void *mem; +#if CALC_EXACT_MALLOC_SIZE + size += sizeof(struct malloc_obj_info); #endif - if (heap_pages_deferred_final) objspace->profile.latest_gc_info |= GPR_FLAG_HAVE_FINALIZE; + mem = malloc(size); +#if CALC_EXACT_MALLOC_SIZE + if (!mem) { + return NULL; } -} - -static inline void -gc_prof_set_malloc_info(rb_objspace_t *objspace) -{ -#if GC_PROFILE_MORE_DETAIL - if (gc_prof_enabled(objspace)) { - gc_profile_record *record = gc_prof_record(objspace); - record->allocate_increase = malloc_increase; - record->allocate_limit = malloc_limit; + else + /* set 0 for consistency of allocated_size/allocations */ + { + struct malloc_obj_info *info = mem; + info->size = 0; + mem = info + 1; } #endif + return mem; } -static inline void -gc_prof_set_heap_info(rb_objspace_t *objspace) +void * +ruby_mimcalloc(size_t num, size_t size) { - if (gc_prof_enabled(objspace)) { - gc_profile_record *record = gc_prof_record(objspace); - size_t live = objspace->profile.total_allocated_object_num_at_gc_start - objspace->profile.total_freed_object_num; - size_t total = objspace->profile.heap_used_at_gc_start * HEAP_OBJ_LIMIT; - -#if GC_PROFILE_MORE_DETAIL - record->heap_use_pages = objspace->profile.heap_used_at_gc_start; - record->heap_live_objects = live; - record->heap_free_objects = total - live; -#endif - - record->heap_total_objects = total; - record->heap_use_size = live * sizeof(RVALUE); - record->heap_total_size = total * sizeof(RVALUE); - } -} - -/* - * call-seq: - * GC::Profiler.clear -> nil - * - * Clears the GC profiler data. - * - */ - -static VALUE -gc_profile_clear(void) -{ - rb_objspace_t *objspace = &rb_objspace; - if (GC_PROFILE_RECORD_DEFAULT_SIZE * 2 < objspace->profile.size) { - objspace->profile.size = GC_PROFILE_RECORD_DEFAULT_SIZE * 2; - objspace->profile.records = realloc(objspace->profile.records, sizeof(gc_profile_record) * objspace->profile.size); - if (!objspace->profile.records) { - rb_memerror(); - } + void *mem; +#if CALC_EXACT_MALLOC_SIZE + struct rbimpl_size_overflow_tag t = rbimpl_size_mul_overflow(num, size); + if (UNLIKELY(t.overflowed)) { + return NULL; } - MEMZERO(objspace->profile.records, gc_profile_record, objspace->profile.size); - objspace->profile.next_index = 0; - objspace->profile.current_record = 0; - return Qnil; -} - -/* - * call-seq: - * GC::Profiler.raw_data -> [Hash, ...] - * - * Returns an Array of individual raw profile data Hashes ordered - * from earliest to latest by +:GC_INVOKE_TIME+. - * - * For example: - * - * [ - * { - * :GC_TIME=>1.3000000000000858e-05, - * :GC_INVOKE_TIME=>0.010634999999999999, - * :HEAP_USE_SIZE=>289640, - * :HEAP_TOTAL_SIZE=>588960, - * :HEAP_TOTAL_OBJECTS=>14724, - * :GC_IS_MARKED=>false - * }, - * # ... - * ] - * - * The keys mean: - * - * +:GC_TIME+:: - * Time elapsed in seconds for this GC run - * +:GC_INVOKE_TIME+:: - * Time elapsed in seconds from startup to when the GC was invoked - * +:HEAP_USE_SIZE+:: - * Total bytes of heap used - * +:HEAP_TOTAL_SIZE+:: - * Total size of heap in bytes - * +:HEAP_TOTAL_OBJECTS+:: - * Total number of objects - * +:GC_IS_MARKED+:: - * Returns +true+ if the GC is in mark phase - * - * If ruby was built with +GC_PROFILE_MORE_DETAIL+, you will also have access - * to the following hash keys: - * - * +:GC_MARK_TIME+:: - * +:GC_SWEEP_TIME+:: - * +:ALLOCATE_INCREASE+:: - * +:ALLOCATE_LIMIT+:: - * +:HEAP_USE_PAGES+:: - * +:HEAP_LIVE_OBJECTS+:: - * +:HEAP_FREE_OBJECTS+:: - * +:HAVE_FINALIZE+:: - * - */ - -static VALUE -gc_profile_record_get(void) -{ - VALUE prof; - VALUE gc_profile = rb_ary_new(); - size_t i; - rb_objspace_t *objspace = (&rb_objspace); - - if (!objspace->profile.run) { - return Qnil; - } - - for (i =0; i < objspace->profile.next_index; i++) { - gc_profile_record *record = &objspace->profile.records[i]; - - prof = rb_hash_new(); - rb_hash_aset(prof, ID2SYM(rb_intern("GC_FLAGS")), gc_info_decode(record->flags, rb_hash_new())); - rb_hash_aset(prof, ID2SYM(rb_intern("GC_TIME")), DBL2NUM(record->gc_time)); - rb_hash_aset(prof, ID2SYM(rb_intern("GC_INVOKE_TIME")), DBL2NUM(record->gc_invoke_time)); - rb_hash_aset(prof, ID2SYM(rb_intern("HEAP_USE_SIZE")), SIZET2NUM(record->heap_use_size)); - rb_hash_aset(prof, ID2SYM(rb_intern("HEAP_TOTAL_SIZE")), SIZET2NUM(record->heap_total_size)); - rb_hash_aset(prof, ID2SYM(rb_intern("HEAP_TOTAL_OBJECTS")), SIZET2NUM(record->heap_total_objects)); - rb_hash_aset(prof, ID2SYM(rb_intern("GC_IS_MARKED")), Qtrue); -#if GC_PROFILE_MORE_DETAIL - rb_hash_aset(prof, ID2SYM(rb_intern("GC_MARK_TIME")), DBL2NUM(record->gc_mark_time)); - rb_hash_aset(prof, ID2SYM(rb_intern("GC_SWEEP_TIME")), DBL2NUM(record->gc_sweep_time)); - rb_hash_aset(prof, ID2SYM(rb_intern("ALLOCATE_INCREASE")), SIZET2NUM(record->allocate_increase)); - rb_hash_aset(prof, ID2SYM(rb_intern("ALLOCATE_LIMIT")), SIZET2NUM(record->allocate_limit)); - rb_hash_aset(prof, ID2SYM(rb_intern("HEAP_USE_PAGES")), SIZET2NUM(record->heap_use_pages)); - rb_hash_aset(prof, ID2SYM(rb_intern("HEAP_LIVE_OBJECTS")), SIZET2NUM(record->heap_live_objects)); - rb_hash_aset(prof, ID2SYM(rb_intern("HEAP_FREE_OBJECTS")), SIZET2NUM(record->heap_free_objects)); - - rb_hash_aset(prof, ID2SYM(rb_intern("REMOVING_OBJECTS")), SIZET2NUM(record->removing_objects)); - rb_hash_aset(prof, ID2SYM(rb_intern("EMPTY_OBJECTS")), SIZET2NUM(record->empty_objects)); - - rb_hash_aset(prof, ID2SYM(rb_intern("HAVE_FINALIZE")), (record->flags & GPR_FLAG_HAVE_FINALIZE) ? Qtrue : Qfalse); -#endif - -#if RGENGC_PROFILE > 0 - rb_hash_aset(prof, ID2SYM(rb_intern("OLD_OBJECTS")), SIZET2NUM(record->old_objects)); - rb_hash_aset(prof, ID2SYM(rb_intern("REMEMBED_NORMAL_OBJECTS")), SIZET2NUM(record->remembered_normal_objects)); - rb_hash_aset(prof, ID2SYM(rb_intern("REMEMBED_SHADY_OBJECTS")), SIZET2NUM(record->remembered_shady_objects)); -#endif - rb_ary_push(gc_profile, prof); + size = t.result + sizeof(struct malloc_obj_info); + mem = calloc1(size); + if (!mem) { + return NULL; } - - return gc_profile; -} - -#if GC_PROFILE_MORE_DETAIL -#define MAJOR_REASON_MAX 0x10 - -static char * -gc_profile_dump_major_reason(int flags, char *buff) -{ - int reason = flags & GPR_FLAG_MAJOR_MASK; - int i = 0; - - if (reason == GPR_FLAG_NONE) { - buff[0] = '-'; - buff[1] = 0; + else + /* set 0 for consistency of allocated_size/allocations */ + { + struct malloc_obj_info *info = mem; + info->size = 0; + mem = info + 1; } - else { -#define C(x, s) \ - if (reason & GPR_FLAG_MAJOR_BY_##x) { \ - buff[i++] = #x[0]; \ - if (i >= MAJOR_REASON_MAX) rb_bug("gc_profile_dump_major_reason: overflow"); \ - buff[i] = 0; \ - } - C(NOFREE, N); - C(OLDGEN, O); - C(SHADY, S); - C(RESCAN, R); - C(STRESS, T); -#if RGENGC_ESTIMATE_OLDMALLOC - C(OLDMALLOC, M); +#else + mem = calloc(num, size); #endif -#undef C - } - return buff; + return mem; } -#endif -static void -gc_profile_dump_on(VALUE out, VALUE (*append)(VALUE, VALUE)) +void +ruby_mimfree(void *ptr) { - rb_objspace_t *objspace = &rb_objspace; - size_t count = objspace->profile.next_index; -#ifdef MAJOR_REASON_MAX - char reason_str[MAJOR_REASON_MAX]; -#endif - - if (objspace->profile.run && count /* > 1 */) { - size_t i; - const gc_profile_record *record; - - append(out, rb_sprintf("GC %"PRIuSIZE" invokes.\n", objspace->profile.count)); - append(out, rb_str_new_cstr("Index Invoke Time(sec) Use Size(byte) Total Size(byte) Total Object GC Time(ms)\n")); - - for (i = 0; i < count; i++) { - record = &objspace->profile.records[i]; - append(out, rb_sprintf("%5"PRIdSIZE" %19.3f %20"PRIuSIZE" %20"PRIuSIZE" %20"PRIuSIZE" %30.20f\n", - i+1, record->gc_invoke_time, record->heap_use_size, - record->heap_total_size, record->heap_total_objects, record->gc_time*1000)); - } - -#if GC_PROFILE_MORE_DETAIL - append(out, rb_str_new_cstr("\n\n" \ - "More detail.\n" \ - "Prepare Time = Previously GC's rest sweep time\n" - "Index Flags Allocate Inc. Allocate Limit" -#if CALC_EXACT_MALLOC_SIZE - " Allocated Size" -#endif - " Use Page Mark Time(ms) Sweep Time(ms) Prepare Time(ms) LivingObj FreeObj RemovedObj EmptyObj" -#if RGENGC_PROFILE - " OldgenObj RemNormObj RemShadObj" -#endif -#if GC_PROFILE_DETAIL_MEMORY - " MaxRSS(KB) MinorFLT MajorFLT" -#endif - "\n")); - - for (i = 0; i < count; i++) { - record = &objspace->profile.records[i]; - append(out, rb_sprintf("%5"PRIdSIZE" %4s/%c/%6s%c %13"PRIuSIZE" %15"PRIuSIZE #if CALC_EXACT_MALLOC_SIZE - " %15"PRIuSIZE -#endif - " %9"PRIuSIZE" %17.12f %17.12f %17.12f %10"PRIuSIZE" %10"PRIuSIZE" %10"PRIuSIZE" %10"PRIuSIZE -#if RGENGC_PROFILE - "%10"PRIuSIZE" %10"PRIuSIZE" %10"PRIuSIZE -#endif -#if GC_PROFILE_DETAIL_MEMORY - "%11ld %8ld %8ld" -#endif - - "\n", - i+1, - gc_profile_dump_major_reason(record->flags, reason_str), - (record->flags & GPR_FLAG_HAVE_FINALIZE) ? 'F' : '.', - (record->flags & GPR_FLAG_NEWOBJ) ? "NEWOBJ" : - (record->flags & GPR_FLAG_MALLOC) ? "MALLOC" : - (record->flags & GPR_FLAG_METHOD) ? "METHOD" : - (record->flags & GPR_FLAG_CAPI) ? "CAPI__" : "??????", - (record->flags & GPR_FLAG_STRESS) ? '!' : ' ', - record->allocate_increase, record->allocate_limit, -#if CALC_EXACT_MALLOC_SIZE - record->allocated_size, -#endif - record->heap_use_pages, - record->gc_mark_time*1000, - record->gc_sweep_time*1000, - record->prepare_time*1000, - - record->heap_live_objects, - record->heap_free_objects, - record->removing_objects, - record->empty_objects -#if RGENGC_PROFILE - , - record->old_objects, - record->remembered_normal_objects, - record->remembered_shady_objects -#endif -#if GC_PROFILE_DETAIL_MEMORY - , - record->maxrss / 1024, - record->minflt, - record->majflt -#endif - - )); - } + struct malloc_obj_info *info = (struct malloc_obj_info *)ptr - 1; + ptr = info; #endif - } + free(ptr); } -/* - * call-seq: - * GC::Profiler.result -> String - * - * Returns a profile data report such as: - * - * GC 1 invokes. - * Index Invoke Time(sec) Use Size(byte) Total Size(byte) Total Object GC time(ms) - * 1 0.012 159240 212940 10647 0.00000000000001530000 - */ - -static VALUE -gc_profile_result(void) +void +rb_gc_adjust_memory_usage(ssize_t diff) { - VALUE str = rb_str_buf_new(0); - gc_profile_dump_on(str, rb_str_buf_append); - return str; -} + unless_objspace(objspace) { return; } -/* - * call-seq: - * GC::Profiler.report - * GC::Profiler.report(io) - * - * Writes the GC::Profiler.result to <tt>$stdout</tt> or the given IO object. - * - */ + rb_gc_impl_adjust_memory_usage(objspace, diff); +} -static VALUE -gc_profile_report(int argc, VALUE *argv, VALUE self) +const char * +rb_obj_info(VALUE obj) { - VALUE out; - - if (argc == 0) { - out = rb_stdout; - } - else { - rb_scan_args(argc, argv, "01", &out); - } - gc_profile_dump_on(out, rb_io_write); - - return Qnil; + return obj_info(obj); } -/* - * call-seq: - * GC::Profiler.total_time -> float - * - * The total time used for garbage collection in seconds - */ - -static VALUE -gc_profile_total_time(VALUE self) +void +rb_obj_info_dump(VALUE obj) { - double time = 0; - rb_objspace_t *objspace = &rb_objspace; - - if (objspace->profile.run && objspace->profile.next_index > 0) { - size_t i; - size_t count = objspace->profile.next_index; - - for (i = 0; i < count; i++) { - time += objspace->profile.records[i].gc_time; - } - } - return DBL2NUM(time); + char buff[0x100]; + fprintf(stderr, "rb_obj_info_dump: %s\n", rb_raw_obj_info(buff, 0x100, obj)); } -/* - * call-seq: - * GC::Profiler.enabled? -> true or false - * - * The current status of GC profile mode. - */ - -static VALUE -gc_profile_enable_get(VALUE self) +void +rb_obj_info_dump_loc(VALUE obj, const char *file, int line, const char *func) { - rb_objspace_t *objspace = &rb_objspace; - return objspace->profile.run ? Qtrue : Qfalse; + char buff[0x100]; + fprintf(stderr, "<OBJ_INFO:%s@%s:%d> %s\n", func, file, line, rb_raw_obj_info(buff, 0x100, obj)); } -/* - * call-seq: - * GC::Profiler.enable -> nil - * - * Starts the GC profiler. - * - */ - -static VALUE -gc_profile_enable(void) +void +rb_gc_before_fork(void) { - rb_objspace_t *objspace = &rb_objspace; - objspace->profile.run = TRUE; - objspace->profile.current_record = 0; - return Qnil; + rb_gc_impl_before_fork(rb_gc_get_objspace()); } -/* - * call-seq: - * GC::Profiler.disable -> nil - * - * Stops the GC profiler. - * - */ - -static VALUE -gc_profile_disable(void) +void +rb_gc_after_fork(rb_pid_t pid) { - rb_objspace_t *objspace = &rb_objspace; - - objspace->profile.run = FALSE; - objspace->profile.current_record = 0; - return Qnil; + rb_gc_impl_after_fork(rb_gc_get_objspace(), pid); } -/* - ------------------------------ DEBUG ------------------------------ -*/ - -static const char * -type_name(int type, VALUE obj) +bool +rb_gc_obj_shareable_p(VALUE obj) { - switch (type) { -#define TYPE_NAME(t) case (t): return #t; - TYPE_NAME(T_NONE); - TYPE_NAME(T_OBJECT); - TYPE_NAME(T_CLASS); - TYPE_NAME(T_MODULE); - TYPE_NAME(T_FLOAT); - TYPE_NAME(T_STRING); - TYPE_NAME(T_REGEXP); - TYPE_NAME(T_ARRAY); - TYPE_NAME(T_HASH); - TYPE_NAME(T_STRUCT); - TYPE_NAME(T_BIGNUM); - TYPE_NAME(T_FILE); - TYPE_NAME(T_MATCH); - TYPE_NAME(T_COMPLEX); - TYPE_NAME(T_RATIONAL); - TYPE_NAME(T_NIL); - TYPE_NAME(T_TRUE); - TYPE_NAME(T_FALSE); - TYPE_NAME(T_SYMBOL); - TYPE_NAME(T_FIXNUM); - TYPE_NAME(T_UNDEF); - TYPE_NAME(T_NODE); - TYPE_NAME(T_ICLASS); - TYPE_NAME(T_ZOMBIE); - case T_DATA: - if (obj && rb_objspace_data_type_name(obj)) { - return rb_objspace_data_type_name(obj); - } - return "T_DATA"; -#undef TYPE_NAME - } - return "unknown"; + return RB_OBJ_SHAREABLE_P(obj); } -static const char * -obj_type_name(VALUE obj) +void +rb_gc_rp(VALUE obj) { - return type_name(TYPE(obj), obj); + rp(obj); } -#if GC_DEBUG +struct check_shareable_data { + VALUE parent; + long err_count; +}; -void -rb_gcdebug_print_obj_condition(VALUE obj) +static void +check_shareable_i(const VALUE child, void *ptr) { - rb_objspace_t *objspace = &rb_objspace; + struct check_shareable_data *data = (struct check_shareable_data *)ptr; - fprintf(stderr, "created at: %s:%d\n", RSTRING_PTR(RANY(obj)->file), FIX2INT(RANY(obj)->line)); + if (!rb_gc_obj_shareable_p(child)) { + fprintf(stderr, "(a) "); + rb_gc_rp(data->parent); + fprintf(stderr, "(b) "); + rb_gc_rp(child); + fprintf(stderr, "check_shareable_i: shareable (a) -> unshareable (b)\n"); - if (is_pointer_to_heap(objspace, (void *)obj)) { - fprintf(stderr, "pointer to heap?: true\n"); - } - else { - fprintf(stderr, "pointer to heap?: false\n"); - return; + data->err_count++; + rb_bug("!! violate shareable constraint !!"); } +} - fprintf(stderr, "marked? : %s\n", MARKED_IN_BITMAP(GET_HEAP_MARK_BITS(obj), obj) ? "true" : "false"); -#if USE_RGENGC -#if RGENGC_THREEGEN - fprintf(stderr, "young? : %s\n", RVALUE_YOUNG_P(obj) ? "true" : "false"); -#endif - fprintf(stderr, "old? : %s\n", RVALUE_OLD_P(obj) ? "true" : "false"); - fprintf(stderr, "WB-protected?: %s\n", RVALUE_WB_PROTECTED(obj) ? "true" : "false"); - fprintf(stderr, "remembered? : %s\n", MARKED_IN_BITMAP(GET_HEAP_REMEMBERSET_BITS(obj), obj) ? "true" : "false"); -#endif +static bool gc_checking_shareable = false; - if (is_lazy_sweeping(heap_eden)) { - fprintf(stderr, "lazy sweeping?: true\n"); - fprintf(stderr, "swept?: %s\n", is_swept_object(objspace, obj) ? "done" : "not yet"); - } - else { - fprintf(stderr, "lazy sweeping?: false\n"); +static void +gc_verify_shareable(void *objspace, VALUE obj, void *data) +{ + // while gc_checking_shareable is true, + // other Ractors should not run the GC, until the flag is not local. + // TODO: remove VM locking if the flag is Ractor local + + unsigned int lev = RB_GC_VM_LOCK(); + { + gc_checking_shareable = true; + rb_objspace_reachable_objects_from(obj, check_shareable_i, (void *)data); + gc_checking_shareable = false; } + RB_GC_VM_UNLOCK(lev); } -static VALUE -gcdebug_sential(VALUE obj, VALUE name) +// TODO: only one level (non-recursive) +void +rb_gc_verify_shareable(VALUE obj) { - fprintf(stderr, "WARNING: object %s(%p) is inadvertently collected\n", (char *)name, (void *)obj); - return Qnil; + rb_objspace_t *objspace = rb_gc_get_objspace(); + struct check_shareable_data data = { + .parent = obj, + .err_count = 0, + }; + gc_verify_shareable(objspace, obj, &data); + + if (data.err_count > 0) { + rb_bug("rb_gc_verify_shareable"); + } } -void -rb_gcdebug_sentinel(VALUE obj, const char *name) +bool +rb_gc_checking_shareable(void) { - rb_define_finalizer(obj, rb_proc_new(gcdebug_sential, (VALUE)name)); + return gc_checking_shareable; } -#endif /* GC_DEBUG */ /* * Document-module: ObjectSpace @@ -7417,8 +5757,9 @@ rb_gcdebug_sentinel(VALUE obj, const char *name) * traverse all living objects with an iterator. * * ObjectSpace also provides support for object finalizers, procs that will be - * called when a specific object is about to be destroyed by garbage - * collection. + * called after a specific object was destroyed by garbage collection. See + * the documentation for +ObjectSpace.define_finalizer+ for important + * information on how to use this method correctly. * * a = "A" * b = "B" @@ -7426,22 +5767,15 @@ rb_gcdebug_sentinel(VALUE obj, const char *name) * ObjectSpace.define_finalizer(a, proc {|id| puts "Finalizer one on #{id}" }) * ObjectSpace.define_finalizer(b, proc {|id| puts "Finalizer two on #{id}" }) * + * a = nil + * b = nil + * * _produces:_ * * Finalizer two on 537763470 * Finalizer one on 537763480 */ -/* - * Document-class: ObjectSpace::WeakMap - * - * An ObjectSpace::WeakMap object holds references to - * any objects, but those objects can get garbage collected. - * - * This class is mostly used internally by WeakRef, please use - * +lib/weakref.rb+ for the public interface. - */ - /* Document-class: GC::Profiler * * The GC profiler provides access to information on GC runs including time, @@ -7460,118 +5794,66 @@ rb_gcdebug_sentinel(VALUE obj, const char *name) * See also GC.count, GC.malloc_allocated_size and GC.malloc_allocations */ -/* - * The GC module provides an interface to Ruby's mark and - * sweep garbage collection mechanism. - * - * Some of the underlying methods are also available via the ObjectSpace - * module. - * - * You may obtain information about the operation of the GC through - * GC::Profiler. - */ +#include "gc.rbinc" void Init_GC(void) { - VALUE rb_mObjSpace; - VALUE rb_mProfiler; - VALUE gc_constants; +#undef rb_intern + rb_gc_register_address(&id2ref_value); + + malloc_offset = gc_compute_malloc_offset(); rb_mGC = rb_define_module("GC"); - rb_define_singleton_method(rb_mGC, "start", gc_start_internal, -1); - rb_define_singleton_method(rb_mGC, "enable", rb_gc_enable, 0); - rb_define_singleton_method(rb_mGC, "disable", rb_gc_disable, 0); - rb_define_singleton_method(rb_mGC, "stress", gc_stress_get, 0); - rb_define_singleton_method(rb_mGC, "stress=", gc_stress_set, 1); - rb_define_singleton_method(rb_mGC, "count", gc_count, 0); - rb_define_singleton_method(rb_mGC, "stat", gc_stat, -1); - rb_define_singleton_method(rb_mGC, "latest_gc_info", gc_latest_gc_info, -1); - rb_define_method(rb_mGC, "garbage_collect", gc_start_internal, -1); - - gc_constants = rb_hash_new(); - rb_hash_aset(gc_constants, ID2SYM(rb_intern("RVALUE_SIZE")), SIZET2NUM(sizeof(RVALUE))); - rb_hash_aset(gc_constants, ID2SYM(rb_intern("HEAP_OBJ_LIMIT")), SIZET2NUM(HEAP_OBJ_LIMIT)); - rb_hash_aset(gc_constants, ID2SYM(rb_intern("HEAP_BITMAP_SIZE")), SIZET2NUM(HEAP_BITMAP_SIZE)); - rb_hash_aset(gc_constants, ID2SYM(rb_intern("HEAP_BITMAP_PLANES")), SIZET2NUM(HEAP_BITMAP_PLANES)); - OBJ_FREEZE(gc_constants); - rb_define_const(rb_mGC, "INTERNAL_CONSTANTS", gc_constants); - - rb_mProfiler = rb_define_module_under(rb_mGC, "Profiler"); - rb_define_singleton_method(rb_mProfiler, "enabled?", gc_profile_enable_get, 0); - rb_define_singleton_method(rb_mProfiler, "enable", gc_profile_enable, 0); - rb_define_singleton_method(rb_mProfiler, "raw_data", gc_profile_record_get, 0); - rb_define_singleton_method(rb_mProfiler, "disable", gc_profile_disable, 0); - rb_define_singleton_method(rb_mProfiler, "clear", gc_profile_clear, 0); - rb_define_singleton_method(rb_mProfiler, "result", gc_profile_result, 0); - rb_define_singleton_method(rb_mProfiler, "report", gc_profile_report, -1); - rb_define_singleton_method(rb_mProfiler, "total_time", gc_profile_total_time, 0); - - rb_mObjSpace = rb_define_module("ObjectSpace"); + + VALUE rb_mObjSpace = rb_define_module("ObjectSpace"); + rb_define_module_function(rb_mObjSpace, "each_object", os_each_obj, -1); - rb_define_module_function(rb_mObjSpace, "garbage_collect", gc_start_internal, -1); rb_define_module_function(rb_mObjSpace, "define_finalizer", define_final, -1); rb_define_module_function(rb_mObjSpace, "undefine_finalizer", undefine_final, 1); - rb_define_module_function(rb_mObjSpace, "_id2ref", id2ref, 1); + rb_define_module_function(rb_mObjSpace, "_id2ref", os_id2ref, 1); - nomem_error = rb_exc_new3(rb_eNoMemError, - rb_obj_freeze(rb_str_new2("failed to allocate memory"))); - OBJ_TAINT(nomem_error); - OBJ_FREEZE(nomem_error); + rb_vm_register_special_exception(ruby_error_nomemory, rb_eNoMemError, "failed to allocate memory"); rb_define_method(rb_cBasicObject, "__id__", rb_obj_id, 0); rb_define_method(rb_mKernel, "object_id", rb_obj_id, 0); rb_define_module_function(rb_mObjSpace, "count_objects", count_objects, -1); - { - VALUE rb_cWeakMap = rb_define_class_under(rb_mObjSpace, "WeakMap", rb_cObject); - rb_define_alloc_func(rb_cWeakMap, wmap_allocate); - rb_define_method(rb_cWeakMap, "[]=", wmap_aset, 2); - rb_define_method(rb_cWeakMap, "[]", wmap_aref, 1); - rb_define_method(rb_cWeakMap, "include?", wmap_has_key, 1); - rb_define_method(rb_cWeakMap, "member?", wmap_has_key, 1); - rb_define_method(rb_cWeakMap, "key?", wmap_has_key, 1); - rb_define_method(rb_cWeakMap, "inspect", wmap_inspect, 0); - rb_define_method(rb_cWeakMap, "each", wmap_each, 0); - rb_define_method(rb_cWeakMap, "each_pair", wmap_each, 0); - rb_define_method(rb_cWeakMap, "each_key", wmap_each_key, 0); - rb_define_method(rb_cWeakMap, "each_value", wmap_each_value, 0); - rb_define_method(rb_cWeakMap, "keys", wmap_keys, 0); - rb_define_method(rb_cWeakMap, "values", wmap_values, 0); - rb_define_method(rb_cWeakMap, "size", wmap_size, 0); - rb_define_method(rb_cWeakMap, "length", wmap_size, 0); - rb_define_private_method(rb_cWeakMap, "finalize", wmap_finalize, 1); - rb_include_module(rb_cWeakMap, rb_mEnumerable); - } - - /* internal methods */ - rb_define_singleton_method(rb_mGC, "verify_internal_consistency", gc_verify_internal_consistency, 0); -#if MALLOC_ALLOCATED_SIZE - rb_define_singleton_method(rb_mGC, "malloc_allocated_size", gc_malloc_allocated_size, 0); - rb_define_singleton_method(rb_mGC, "malloc_allocations", gc_malloc_allocations, 0); + rb_gc_impl_init(); +} + +// Set a name for the anonymous virtual memory area. `addr` is the starting +// address of the area and `size` is its length in bytes. `name` is a +// NUL-terminated human-readable string. +// +// This function is usually called after calling `mmap()`. The human-readable +// annotation helps developers identify the call site of `mmap()` that created +// the memory mapping. +// +// This function currently only works on Linux 5.17 or higher. After calling +// this function, we can see annotations in the form of "[anon:...]" in +// `/proc/self/maps`, where `...` is the content of `name`. This function has +// no effect when called on other platforms. +void +ruby_annotate_mmap(const void *addr, unsigned long size, const char *name) +{ +#if defined(HAVE_SYS_PRCTL_H) && defined(PR_SET_VMA) && defined(PR_SET_VMA_ANON_NAME) + // The name length cannot exceed 80 (including the '\0'). + RUBY_ASSERT(strlen(name) < 80); + prctl(PR_SET_VMA, PR_SET_VMA_ANON_NAME, (unsigned long)addr, size, name); + // We ignore errors in prctl. prctl may set errno to EINVAL for several + // reasons. + // 1. The attr (PR_SET_VMA_ANON_NAME) is not a valid attribute. + // 2. addr is an invalid address. + // 3. The string pointed by name is too long. + // The first error indicates PR_SET_VMA_ANON_NAME is not available, and may + // happen if we run the compiled binary on an old kernel. In theory, all + // other errors should result in a failure. But since EINVAL cannot tell + // the first error from others, and this function is mainly used for + // debugging, we silently ignore the error. + errno = 0; #endif - - /* ::GC::OPTS, which shows GC build options */ - { - VALUE opts; - rb_define_const(rb_mGC, "OPTS", opts = rb_ary_new()); -#define OPT(o) if (o) rb_ary_push(opts, rb_str_new2(#o)) - OPT(GC_DEBUG); - OPT(USE_RGENGC); - OPT(RGENGC_DEBUG); - OPT(RGENGC_CHECK_MODE); - OPT(RGENGC_PROFILE); - OPT(RGENGC_THREEGEN); - OPT(RGENGC_ESTIMATE_OLDMALLOC); - OPT(GC_PROFILE_MORE_DETAIL); - OPT(GC_ENABLE_LAZY_SWEEP); - OPT(CALC_EXACT_MALLOC_SIZE); - OPT(MALLOC_ALLOCATED_SIZE); - OPT(MALLOC_ALLOCATED_SIZE_CHECK); - OPT(GC_PROFILE_DETAIL_MEMORY); -#undef OPT - } } |