obtained gcc-4.6.4.tar.bz2 from upstream website;upstream

verified gcc-4.6.4.tar.bz2.sig;
imported gcc-4.6.4 source tree from verified upstream tarball.

downloading a git-generated archive based on the 'upstream' tag
should provide you with a source tree that is binary identical
to the one extracted from the above tarball.

if you have obtained the source via the command 'git clone',
however, do note that line-endings of files in your working
directory might differ from line-endings of the respective
files in the upstream repository.

1 files changed, 107 insertions, 0 deletions

diff --git a/gcc/testsuite/gcc.target/i386/sse4_1-roundss-4.c b/gcc/testsuite/gcc.target/i386/sse4_1-roundss-4.c
new file mode 100644
index 000000000..71042d1b7
--- /dev/null
+++ b/gcc/testsuite/gcc.target/i386/sse4_1-roundss-4.c
@@ -0,0 +1,107 @@
+/* { dg-do run } */
+/* { dg-require-effective-target sse4 } */
+/* { dg-options "-O2 -msse4.1" } */
+/* { dg-skip-if "no M_PI" { vxworks_kernel } } */
+
+#include "sse4_1-check.h"
+
+#include <smmintrin.h>
+#include <math.h>
+#include <string.h>
+
+#define NUM 64
+
+static void
+init_round (float *src)
+{
+  int i, sign = 1;
+  float f = rand ();
+
+  for (i = 0; i < NUM; i++)
+    {
+      src[i] = (i + 1)* f * M_PI * sign;
+      if (i < (NUM / 2))
+	{
+          if ((i % 6) == 0)
+	    f = f * src[i];
+        }
+      else if (i == (NUM / 2))
+	f = rand ();
+      else if ((i % 6) == 0)
+	f = 1 / (f * (i + 1) * src[i] * M_PI *sign);
+      sign = -sign;
+    }
+}
+
+static float
+do_round (float f, int type)
+{
+  short saved_cw, new_cw, clr_mask;
+  float ret;
+
+  if ((type & 4))
+    {
+      type = 0;
+      clr_mask = 0xFFFF;
+    }
+  else
+    {
+      type = 0x003F | ((type & 3) << 10);
+      clr_mask = ~0x0C3F;
+    }
+
+  __asm__ ("flds %0" : : "m" (*&f));
+
+  __asm__ ("fstcw %0" : "=m" (*&saved_cw));
+  new_cw = saved_cw & clr_mask;
+  new_cw |= type;
+  __asm__ ("fldcw %0" : : "m" (*&new_cw));
+
+  __asm__ ("frndint\n"
+	   "fstps %0\n" : "=m" (*&ret));
+  __asm__ ("fldcw %0" : : "m" (*&saved_cw));
+  return ret;
+}
+
+static void
+sse4_1_test (void)
+{
+  int i, j;
+  float f;
+  union
+    {
+      __m128 x[NUM / 4];
+      float f[NUM];
+    } dst, src;
+
+  init_round (src.f);
+  memset (&dst, 0, NUM * sizeof(float));
+
+  for (i = 0; i < NUM / 4 ; i++)
+    dst.x[i] =  _mm_round_ss (dst.x[i], src.x[i], _MM_FROUND_RINT);
+
+  for (i = 0; i < NUM; i += 4)
+    {
+      for (j = 0; j < 3; j++)
+	if (dst.f[i + j + 1] != 0.0)
+	  abort ();
+
+      f = do_round (src.f[i], 0x04);
+      if (f != dst.f[i])
+	abort ();
+    }
+
+  for (i = 0; i < NUM / 4 ; i++)
+    dst.x[i] =  _mm_round_ss (dst.x[i], src.x[i], _MM_FROUND_NEARBYINT);
+
+  for (i = 0; i < NUM; i += 4)
+    {
+      for (j = 0; j < 3; j++)
+	if (dst.f[i + j + 1] != 0.0)
+	  abort ();
+
+      f = do_round (src.f[i], 0x0c);
+      if (f != dst.f[i])
+	abort ();
+    }
+}