From 7b6e07ea93cfc42bfb8bdf875555d2c8c7d5e507 Mon Sep 17 00:00:00 2001
From: Peter Zhu <peter@peterzhu.ca>
Date: Tue, 18 Feb 2025 12:58:40 -0500
Subject: Add rb_gc_object_metadata API

This function replaces the internal rb_obj_gc_flags API. rb_gc_object_metadata
returns an array of name and value pairs, with the last element having
0 for the name.
---
 ext/objspace/objspace_dump.c | 29 +++++++++++++++++------------
 gc.c                         | 13 ++++++-------
 gc/default/default.c         | 41 +++++++++++++++++++++++++++--------------
 gc/gc_impl.h                 | 10 +++++++++-
 internal/gc.h                | 10 ++++++++--
 5 files changed, 67 insertions(+), 36 deletions(-)

diff --git a/ext/objspace/objspace_dump.c b/ext/objspace/objspace_dump.c
index 2b98214f69..139612526f 100644
--- a/ext/objspace/objspace_dump.c
+++ b/ext/objspace/objspace_dump.c
@@ -385,8 +385,6 @@ dump_object(VALUE obj, struct dump_config *dc)
     size_t memsize;
     struct allocation_info *ainfo = objspace_lookup_allocation_info(obj);
     rb_io_t *fptr;
-    ID flags[RB_OBJ_GC_FLAGS_MAX];
-    size_t n, i;
     ID mid;
 
     if (SPECIAL_CONST_P(obj)) {
@@ -643,14 +641,24 @@ dump_object(VALUE obj, struct dump_config *dc)
         dump_append_lu(dc, RB_NUM2ULONG(rb_obj_id(obj)));
     }
 
-    if ((n = rb_obj_gc_flags(obj, flags, sizeof(flags))) > 0) {
-        dump_append(dc, ", \"flags\":{");
-        for (i=0; i<n; i++) {
-            dump_append(dc, "\"");
-            dump_append(dc, rb_id2name(flags[i]));
-            dump_append(dc, "\":true");
-            if (i != n-1) dump_append(dc, ", ");
+    struct rb_gc_object_metadata_entry *gc_metadata = rb_gc_object_metadata(obj);
+    for (int i = 0; gc_metadata[i].name != 0; i++) {
+        if (i == 0) {
+            dump_append(dc, ", \"flags\":{");
+        }
+        else {
+            dump_append(dc, ", ");
         }
+
+        dump_append(dc, "\"");
+        dump_append(dc, rb_id2name(gc_metadata[i].name));
+        dump_append(dc, "\":");
+        dump_append_special_const(dc, gc_metadata[i].val);
+    }
+
+    /* If rb_gc_object_metadata had any entries, we need to close the opening
+     * `"flags":{`. */
+    if (gc_metadata[0].name != 0) {
         dump_append(dc, "}");
     }
 
@@ -883,7 +891,4 @@ Init_objspace_dump(VALUE rb_mObjSpace)
     rb_define_module_function(rb_mObjSpace, "_dump", objspace_dump, 2);
     rb_define_module_function(rb_mObjSpace, "_dump_all", objspace_dump_all, 4);
     rb_define_module_function(rb_mObjSpace, "_dump_shapes", objspace_dump_shapes, 2);
-
-    /* force create static IDs */
-    rb_obj_gc_flags(rb_mObjSpace, 0, 0);
 }
diff --git a/gc.c b/gc.c
index bd3e8aea4d..6510231ccf 100644
--- a/gc.c
+++ b/gc.c
@@ -678,7 +678,7 @@ typedef struct gc_function_map {
     VALUE (*stat_heap)(void *objspace_ptr, VALUE heap_name, VALUE hash_or_sym);
     const char *(*active_gc_name)(void);
     // Miscellaneous
-    size_t (*obj_flags)(void *objspace_ptr, VALUE obj, ID* flags, size_t max);
+    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);
@@ -855,7 +855,7 @@ ruby_modular_gc_init(void)
     load_modular_gc_func(stat_heap);
     load_modular_gc_func(active_gc_name);
     // Miscellaneous
-    load_modular_gc_func(obj_flags);
+    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);
@@ -938,7 +938,7 @@ ruby_modular_gc_init(void)
 # 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_obj_flags rb_gc_functions.obj_flags
+# 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
@@ -2955,11 +2955,10 @@ rb_gc_active_gc_name(void)
     return gc_name;
 }
 
-// TODO: rearchitect this function to work for a generic GC
-size_t
-rb_obj_gc_flags(VALUE obj, ID* flags, size_t max)
+struct rb_gc_object_metadata_entry *
+rb_gc_object_metadata(VALUE obj)
 {
-    return rb_gc_impl_obj_flags(rb_gc_get_objspace(), obj, flags, max);
+    return rb_gc_impl_object_metadata(rb_gc_get_objspace(), obj);
 }
 
 /* GC */
diff --git a/gc/default/default.c b/gc/default/default.c
index 4cf39a0331..03427f51a1 100644
--- a/gc/default/default.c
+++ b/gc/default/default.c
@@ -6189,33 +6189,46 @@ rb_gc_impl_writebarrier_remember(void *objspace_ptr, VALUE obj)
     }
 }
 
-// TODO: rearchitect this function to work for a generic GC
-size_t
-rb_gc_impl_obj_flags(void *objspace_ptr, VALUE obj, ID* flags, size_t max)
+#define RB_GC_OBJECT_METADATA_ENTRY_COUNT 5
+static struct rb_gc_object_metadata_entry object_metadata_entries[RB_GC_OBJECT_METADATA_ENTRY_COUNT + 1];
+
+struct rb_gc_object_metadata_entry *
+rb_gc_impl_object_metadata(void *objspace_ptr, VALUE obj)
 {
     rb_objspace_t *objspace = objspace_ptr;
     size_t n = 0;
-    static ID ID_marked;
-    static ID ID_wb_protected, ID_old, ID_marking, ID_uncollectible, ID_pinned;
+    static ID ID_wb_protected, ID_old, ID_uncollectible, ID_marking, ID_marked, ID_pinned;
 
     if (!ID_marked) {
 #define I(s) ID_##s = rb_intern(#s);
-        I(marked);
         I(wb_protected);
         I(old);
-        I(marking);
         I(uncollectible);
+        I(marking);
+        I(marked);
         I(pinned);
 #undef I
     }
 
-    if (RVALUE_WB_UNPROTECTED(objspace, obj) == 0 && n < max)                   flags[n++] = ID_wb_protected;
-    if (RVALUE_OLD_P(objspace, obj) && n < max)                                 flags[n++] = ID_old;
-    if (RVALUE_UNCOLLECTIBLE(objspace, obj) && n < max)                         flags[n++] = ID_uncollectible;
-    if (RVALUE_MARKING(objspace, obj) && n < max) flags[n++] = ID_marking;
-    if (RVALUE_MARKED(objspace, obj) && n < max)    flags[n++] = ID_marked;
-    if (RVALUE_PINNED(objspace, obj) && n < max)  flags[n++] = ID_pinned;
-    return n;
+#define SET_ENTRY(na, v) do { \
+    GC_ASSERT(n <= RB_GC_OBJECT_METADATA_ENTRY_COUNT); \
+    object_metadata_entries[n].name = ID_##na; \
+    object_metadata_entries[n].val = v; \
+    n++; \
+} while (0)
+
+    if (!RVALUE_WB_UNPROTECTED(objspace, obj)) SET_ENTRY(wb_protected, Qtrue);
+    if (RVALUE_OLD_P(objspace, obj)) SET_ENTRY(old, Qtrue);
+    if (RVALUE_UNCOLLECTIBLE(objspace, obj)) SET_ENTRY(uncollectible, Qtrue);
+    if (RVALUE_MARKING(objspace, obj)) SET_ENTRY(marking, Qtrue);
+    if (RVALUE_MARKED(objspace, obj)) SET_ENTRY(marked, Qtrue);
+    if (RVALUE_PINNED(objspace, obj)) SET_ENTRY(pinned, Qtrue);
+
+    object_metadata_entries[n].name = 0;
+    object_metadata_entries[n].val = 0;
+#undef SET_ENTRY
+
+    return object_metadata_entries;
 }
 
 void *
diff --git a/gc/gc_impl.h b/gc/gc_impl.h
index edff242bd0..5e8d209432 100644
--- a/gc/gc_impl.h
+++ b/gc/gc_impl.h
@@ -10,6 +10,14 @@
  */
 #include "ruby/ruby.h"
 
+#ifndef RB_GC_OBJECT_METADATA_ENTRY_DEFINED
+# define RB_GC_OBJECT_METADATA_ENTRY_DEFINED
+struct rb_gc_object_metadata_entry {
+    ID name;
+    VALUE val;
+};
+#endif
+
 #ifdef BUILDING_MODULAR_GC
 # define GC_IMPL_FN
 #else
@@ -108,7 +116,7 @@ GC_IMPL_FN VALUE rb_gc_impl_stat(void *objspace_ptr, VALUE hash_or_sym);
 GC_IMPL_FN VALUE rb_gc_impl_stat_heap(void *objspace_ptr, VALUE heap_name, VALUE hash_or_sym);
 GC_IMPL_FN const char *rb_gc_impl_active_gc_name(void);
 // Miscellaneous
-GC_IMPL_FN size_t rb_gc_impl_obj_flags(void *objspace_ptr, VALUE obj, ID* flags, size_t max);
+GC_IMPL_FN struct rb_gc_object_metadata_entry *rb_gc_impl_object_metadata(void *objspace_ptr, VALUE obj);
 GC_IMPL_FN bool rb_gc_impl_pointer_to_heap_p(void *objspace_ptr, const void *ptr);
 GC_IMPL_FN bool rb_gc_impl_garbage_object_p(void *objspace_ptr, VALUE obj);
 GC_IMPL_FN void rb_gc_impl_set_event_hook(void *objspace_ptr, const rb_event_flag_t event);
diff --git a/internal/gc.h b/internal/gc.h
index 1407f4b988..4e9b4554e8 100644
--- a/internal/gc.h
+++ b/internal/gc.h
@@ -127,7 +127,13 @@ struct rb_objspace; /* in vm_core.h */
             rb_wb_protected_newobj_of((ec ? ec : GET_EC()), (c), (f) & ~FL_WB_PROTECTED, s) : \
             rb_wb_unprotected_newobj_of((c), (f), s))
 
-#define RB_OBJ_GC_FLAGS_MAX 6   /* used in ext/objspace */
+#ifndef RB_GC_OBJECT_METADATA_ENTRY_DEFINED
+# define RB_GC_OBJECT_METADATA_ENTRY_DEFINED
+struct rb_gc_object_metadata_entry {
+    ID name;
+    VALUE val;
+};
+#endif
 
 #ifndef USE_UNALIGNED_MEMBER_ACCESS
 # define UNALIGNED_MEMBER_ACCESS(expr) (expr)
@@ -244,7 +250,7 @@ const char *rb_objspace_data_type_name(VALUE obj);
 VALUE rb_wb_protected_newobj_of(struct rb_execution_context_struct *, VALUE, VALUE, size_t);
 VALUE rb_wb_unprotected_newobj_of(VALUE, VALUE, size_t);
 size_t rb_obj_memsize_of(VALUE);
-size_t rb_obj_gc_flags(VALUE, ID[], size_t);
+struct rb_gc_object_metadata_entry *rb_gc_object_metadata(VALUE obj);
 void rb_gc_mark_values(long n, const VALUE *values);
 void rb_gc_mark_vm_stack_values(long n, const VALUE *values);
 void rb_gc_update_values(long n, VALUE *values);
-- 
cgit v1.2.3