MDEV-7951 - sql_alloc() takes 0.25% in OLTP RO

sql_alloc() has additional costs compared to direct mem_root allocation:
- function call: it is defined in a separate translation unit and can't be
  inlined
- it needs to call pthread_getspecific() to get THD::mem_root

It is called dozens of times implicitely at least by:
- List<>::push_back()
- List<>::push_front()
- new (for Sql_alloc derived classes)
- sql_memdup()

Replaced lots of implicit sql_alloc() calls with direct mem_root allocation,
passing through THD pointer whenever it is needed.

Number of sql_alloc() calls reduced 345 -> 41 per OLTP RO transaction.
pthread_getspecific() overhead dropped 0.76 -> 0.59
sql_alloc() overhed dropped 0.25 -> 0.06

sql/sql_select.h

@@ -1805,7 +1805,7 @@ int report_error(TABLE *table, int error);
 int safe_index_read(JOIN_TAB *tab);
 COND *remove_eq_conds(THD *thd, COND *cond, Item::cond_result *cond_value);
 int get_quick_record(SQL_SELECT *select);
-SORT_FIELD * make_unireg_sortorder(ORDER *order, uint *length,
+SORT_FIELD * make_unireg_sortorder(THD *thd, ORDER *order, uint *length,
                                   SORT_FIELD *sortorder);
 int setup_order(THD *thd, Item **ref_pointer_array, TABLE_LIST *tables,
 		List<Item> &fields, List <Item> &all_fields, ORDER *order);