wcsmbs: Add wcscpy loop unroll option

This allows an architecture to use the old generic implementation
and also set explicit loop unrolling.

Checked on aarch64-linux-gnu.

	* include/loop_unroll.h: New file.
	* wcsmbs/wcscpy (__wcscpy): Add option to use loop unrolling
	besides generic implementation.

1 files changed, 19 insertions, 0 deletions

diff --git a/wcsmbs/wcscpy.c b/wcsmbs/wcscpy.c
index 6fb2969513..36c001f69a 100644
--- a/wcsmbs/wcscpy.c
+++ b/wcsmbs/wcscpy.c
@@ -17,6 +17,7 @@
    <http://www.gnu.org/licenses/>.  */
 
 #include <wchar.h>
+#include <loop_unroll.h>
 
 
 #ifdef WCSCPY
@@ -27,7 +28,25 @@
 wchar_t *
 __wcscpy (wchar_t *dest, const wchar_t *src)
 {
+#ifndef UNROLL_NTIMES
   return __wmemcpy (dest, src, __wcslen (src) + 1);
+#else
+  /* Some architectures might have costly tail function call (powerpc
+     for instance) where wmemcpy call overhead for smalls sizes might
+     be more costly than just unrolling the main loop.  */
+  wchar_t *wcp = dest;
+
+#define ITERATION(index)		\
+  ({					\
+     wchar_t c = *src++;		\
+     *wcp++ = c;			\
+     c != L'\0';			\
+  })
+
+  while (1)
+    UNROLL_REPEAT(UNROLL_NTIMES, ITERATION);
+  return dest;
+#endif
 }
 #ifndef WCSCPY
 weak_alias (__wcscpy, wcscpy)