gh-112532: Use separate mimalloc heaps for GC objects (gh-113263)

* gh-112532: Use separate mimalloc heaps for GC objects

In `--disable-gil` builds, we now use four separate heaps in
anticipation of using mimalloc to find GC objects when the GIL is
disabled. To support this, we also make a few changes to mimalloc:

* `mi_heap_t` and `mi_tld_t` initialization is split from allocation.
  This allows us to have a `mi_tld_t` per-`PyThreadState`, which is
  important to keep interpreter isolation, since the same OS thread may
  run in multiple interpreters (using different PyThreadStates.)

* Heap abandoning (mi_heap_collect_ex) can now be called from a
  different thread than the one that created the heap. This is necessary
  because we may clear and delete the containing PyThreadStates from a
  different thread during finalization and after fork().

* Use enum instead of defines and guard mimalloc includes.

* The enum typedef will be convenient for future PRs that use the type.
* Guarding the mimalloc includes allows us to unconditionally include
  pycore_mimalloc.h from other header files that rely on things like
  `struct _mimalloc_thread_state`.

* Only define _mimalloc_thread_state in Py_GIL_DISABLED builds

Include/internal/mimalloc/mimalloc/internal.h

@@ -85,6 +85,7 @@ mi_threadid_t _mi_thread_id(void) mi_attr_noexcept;
 mi_heap_t*    _mi_heap_main_get(void);     // statically allocated main backing heap
 void       _mi_thread_done(mi_heap_t* heap);
 void       _mi_thread_data_collect(void);
+void       _mi_tld_init(mi_tld_t* tld, mi_heap_t* bheap);
 
 // os.c
 void       _mi_os_init(void);                                            // called from process init
@@ -170,6 +171,7 @@ size_t     _mi_bin_size(uint8_t bin);           // for stats
 uint8_t    _mi_bin(size_t size);                // for stats
 
 // "heap.c"
+void       _mi_heap_init_ex(mi_heap_t* heap, mi_tld_t* tld, mi_arena_id_t arena_id);
 void       _mi_heap_destroy_pages(mi_heap_t* heap);
 void       _mi_heap_collect_abandon(mi_heap_t* heap);
 void       _mi_heap_set_default_direct(mi_heap_t* heap);

Include/internal/pycore_mimalloc.h

@@ -9,11 +9,37 @@
 #  error "pycore_mimalloc.h must be included before mimalloc.h"
 #endif
 
+typedef enum {
+    _Py_MIMALLOC_HEAP_MEM = 0,      // PyMem_Malloc() and friends
+    _Py_MIMALLOC_HEAP_OBJECT = 1,   // non-GC objects
+    _Py_MIMALLOC_HEAP_GC = 2,       // GC objects without pre-header
+    _Py_MIMALLOC_HEAP_GC_PRE = 3,   // GC objects with pre-header
+    _Py_MIMALLOC_HEAP_COUNT
+} _Py_mimalloc_heap_id;
+
 #include "pycore_pymem.h"
+
+#ifdef WITH_MIMALLOC
 #define MI_DEBUG_UNINIT     PYMEM_CLEANBYTE
 #define MI_DEBUG_FREED      PYMEM_DEADBYTE
 #define MI_DEBUG_PADDING    PYMEM_FORBIDDENBYTE
+#ifdef Py_DEBUG
+#  define MI_DEBUG 1
+#else
+#  define MI_DEBUG 0
+#endif
 
 #include "mimalloc.h"
+#include "mimalloc/types.h"
+#include "mimalloc/internal.h"
+#endif
+
+#ifdef Py_GIL_DISABLED
+struct _mimalloc_thread_state {
+    mi_heap_t *current_object_heap;
+    mi_heap_t heaps[_Py_MIMALLOC_HEAP_COUNT];
+    mi_tld_t tld;
+};
+#endif
 
 #endif // Py_INTERNAL_MIMALLOC_H

Include/internal/pycore_pystate.h

@@ -187,6 +187,7 @@ extern PyThreadState * _PyThreadState_New(
     int whence);
 extern void _PyThreadState_Bind(PyThreadState *tstate);
 extern void _PyThreadState_DeleteExcept(PyThreadState *tstate);
+extern void _PyThreadState_ClearMimallocHeaps(PyThreadState *tstate);
 
 // Export for '_testinternalcapi' shared extension
 PyAPI_FUNC(PyObject*) _PyThreadState_GetDict(PyThreadState *tstate);

Include/internal/pycore_tstate.h

@@ -8,6 +8,8 @@ extern "C" {
 #  error "this header requires Py_BUILD_CORE define"
 #endif
 
+#include "pycore_mimalloc.h"      // struct _mimalloc_thread_state
+
 
 // Every PyThreadState is actually allocated as a _PyThreadStateImpl. The
 // PyThreadState fields are exposed as part of the C API, although most fields
@@ -16,7 +18,10 @@ typedef struct _PyThreadStateImpl {
     // semi-public fields are in PyThreadState.
     PyThreadState base;
 
-    // TODO: add private fields here
+#ifdef Py_GIL_DISABLED
+    struct _mimalloc_thread_state mimalloc;
+#endif
+
 } _PyThreadStateImpl;