This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

[MIPS, committed] Expect MIPS16 to use memcpy for short strings


Although I changed non-MIPS16 code to allow larger inline copies:

   http://gcc.gnu.org/ml/gcc-patches/2012-01/msg00053.html

MIPS16 code is still for now expected to use memcpy.  This patch
adjusts the testsuite accordingly.

Tested on mips64-linux-gnu and various ELF targets.  Applied.

Richard


gcc/testsuite/
	* gcc.dg/tree-prof/stringop-2.c (main): Add a nomips16 attribute
	on MIPS targets.
	* gfortran.dg/pr45636.f90: XFAIL for MIPS16 targets.

Index: gcc/testsuite/gcc.dg/tree-prof/stringop-2.c
===================================================================
--- gcc/testsuite/gcc.dg/tree-prof/stringop-2.c	2012-02-04 10:57:10.000000000 +0000
+++ gcc/testsuite/gcc.dg/tree-prof/stringop-2.c	2012-02-05 14:23:07.000000000 +0000
@@ -3,6 +3,10 @@
 int b[1000];
 int size=1;
 int max=10000;
+#ifdef __mips
+/* We allow short memcpy()s for MIPS16.  */
+int __attribute__((nomips16))
+#endif
 main()
 {
   int i;
Index: gcc/testsuite/gfortran.dg/pr45636.f90
===================================================================
--- gcc/testsuite/gfortran.dg/pr45636.f90	2012-02-05 14:23:36.000000000 +0000
+++ gcc/testsuite/gfortran.dg/pr45636.f90	2012-02-05 14:26:41.000000000 +0000
@@ -10,5 +10,5 @@ program main
   b = y
   call sub(a, b)
 end program main
-! { dg-final { scan-tree-dump-times "memset" 0 "forwprop2" } }
+! { dg-final { scan-tree-dump-times "memset" 0 "forwprop2" { xfail { mips*-*-* && { ! nomips16 } } } } }
 ! { dg-final { cleanup-tree-dump "forwprop2" } }


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]