gh-115103: Delay reuse of mimalloc pages that store PyObjects (#115435)

This implements the delayed reuse of mimalloc pages that contain Python
objects in the free-threaded build.

Allocations of the same size class are grouped in data structures called
pages. These are different from operating system pages. For thread-safety, we
want to ensure that memory used to store PyObjects remains valid as long as
there may be concurrent lock-free readers; we want to delay using it for
other size classes, in other heaps, or returning it to the operating system.

When a mimalloc page becomes empty, instead of immediately freeing it, we tag
it with a QSBR goal and insert it into a per-thread state linked list of
pages to be freed. When mimalloc needs a fresh page, we process the queue and
free any still empty pages that are now deemed safe to be freed. Pages
waiting to be freed are still available for allocations of the same size
class and allocating from a page prevent it from being freed. There is
additional logic to handle abandoned pages when threads exit.

Include/internal/mimalloc/mimalloc/types.h

@@ -311,6 +311,7 @@ typedef struct mi_page_s {
   uint32_t              slice_offset;      // distance from the actual page data slice (0 if a page)
   uint8_t               is_committed : 1;  // `true` if the page virtual memory is committed
   uint8_t               is_zero_init : 1;  // `true` if the page was initially zero initialized
+  uint8_t               use_qsbr : 1;      // delay page freeing using qsbr
   uint8_t               tag : 4;           // tag from the owning heap
   uint8_t               debug_offset;      // number of bytes to preserve when filling freed or uninitialized memory
 
@@ -336,8 +337,13 @@ typedef struct mi_page_s {
   struct mi_page_s*     next;              // next page owned by this thread with the same `block_size`
   struct mi_page_s*     prev;              // previous page owned by this thread with the same `block_size`
 
+#ifdef Py_GIL_DISABLED
+  struct llist_node     qsbr_node;
+  uint64_t              qsbr_goal;
+#endif
+
   // 64-bit 9 words, 32-bit 12 words, (+2 for secure)
-  #if MI_INTPTR_SIZE==8
+  #if MI_INTPTR_SIZE==8 && !defined(Py_GIL_DISABLED)
   uintptr_t padding[1];
   #endif
 } mi_page_t;
@@ -555,6 +561,7 @@ struct mi_heap_s {
   bool                  no_reclaim;                          // `true` if this heap should not reclaim abandoned pages
   uint8_t               tag;                                 // custom identifier for this heap
   uint8_t               debug_offset;                        // number of bytes to preserve when filling freed or uninitialized memory
+  bool                  page_use_qsbr;                       // should freeing pages be delayed using QSBR
 };