[OpenMP] Fix the build on Windows

[JohnTitor]

Summary

The code expanded from kmp_barrier.h uses some KMP_INTERNAL_*s,
so the definitions have to be placed before it.

Fixes #55815

[penzn]

LGTM

@jdoerfert - thoughts?

[jdoerfert]

This seems to be uncontroversial. LGTM

FWIW, @tlwilmar is also a good person to tag as reviewer ;)

[JohnTitor]

FWIW, @tlwilmar is also a good person to tag as reviewer ;)

Good to know, I just git-blamed the file and found you :)

Should I wait for another review from them or is it just a ping?

[JohnTitor]

I assume you just mentioned them for the next patch, pushing a commit.

openmp/runtime/src/kmp.h

 #define HWLOC_OBJ_PACKAGE HWLOC_OBJ_SOCKET
 #endif
 #endif
 
 #if KMP_ARCH_X86 || KMP_ARCH_X86_64
 #include <xmmintrin.h>
 #endif
 
+// The below has to be defined before including "kmp_barrier.h".
+#define KMP_INTERNAL_MALLOC(sz) malloc(sz)
+#define KMP_INTERNAL_FREE(p) free(p)
+#define KMP_INTERNAL_REALLOC(p, sz) realloc((p), (sz))
+#define KMP_INTERNAL_CALLOC(n, sz) calloc((n), (sz))
+
 #include "kmp_debug.h"
 #include "kmp_lock.h"
 #include "kmp_version.h"
 #include "kmp_barrier.h"
 #if USE_DEBUGGER
 #include "kmp_debugger.h"
 #endif
 #include "kmp_i18n.h"
#define __kmp_thread_malloc(th, size)                                          \
   ___kmp_thread_malloc((th), (size)KMP_SRC_LOC_CURR)
 #define __kmp_thread_calloc(th, nelem, elsize)                                 \
   ___kmp_thread_calloc((th), (nelem), (elsize)KMP_SRC_LOC_CURR)
 #define __kmp_thread_realloc(th, ptr, size)                                    \
   ___kmp_thread_realloc((th), (ptr), (size)KMP_SRC_LOC_CURR)
 #define __kmp_thread_free(th, ptr)                                             \
   ___kmp_thread_free((th), (ptr)KMP_SRC_LOC_CURR)
 
-#define KMP_INTERNAL_MALLOC(sz) malloc(sz)
-#define KMP_INTERNAL_FREE(p) free(p)
-#define KMP_INTERNAL_REALLOC(p, sz) realloc((p), (sz))
-#define KMP_INTERNAL_CALLOC(n, sz) calloc((n), (sz))
-
 extern void __kmp_push_num_threads(ident_t *loc, int gtid, int num_threads);
 
 extern void __kmp_push_proc_bind(ident_t *loc, int gtid,
                                  kmp_proc_bind_t proc_bind);
 extern void __kmp_push_num_teams(ident_t *loc, int gtid, int num_teams,
                                  int num_threads);
 extern void __kmp_push_num_teams_51(ident_t *loc, int gtid, int num_teams_lb,
                                     int num_teams_ub, int num_threads);