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, 291 insertions, 0 deletions

diff --git a/gcc/testsuite/gcc.dg/vmx/varargs-4.c b/gcc/testsuite/gcc.dg/vmx/varargs-4.c
new file mode 100644
index 000000000..a17551953
--- /dev/null
+++ b/gcc/testsuite/gcc.dg/vmx/varargs-4.c
@@ -0,0 +1,291 @@
+#include "harness.h"
+#include <stdarg.h>
+#include <stddef.h>
+#include <string.h>
+
+typedef struct n_a
+{
+  signed char m1;
+  short m2;
+  int m3;
+  double m4;
+  vector float m5;
+}
+n_a;
+
+static n_a gn_a;
+
+static int
+lay(char *p, int start, int end, int n)
+{
+  int b;
+  unsigned char ch;
+  unsigned int mask;
+
+  start *= 8;
+  end *= 8;
+  n *= 8;
+
+  for (b = 0; b + 8 <= start; b += 8)
+    {
+      ch = *p++;
+      if (ch != 0xff)
+	for (mask = 0x80; mask; b++, mask >>= 1)
+	  if ((ch & mask) != mask)
+	    return b;
+    }
+
+  if (b < start)
+    {
+      ch = *p++;
+      for (mask = 0x80; b < start; b++, mask >>= 1)
+	if ((ch & mask) != mask)
+	  return b;
+      for (; mask && b < end; b++, mask >>= 1)
+	if ((ch & mask) != 0)
+	  return b;
+    }
+
+  for (; b + 8 <= end; b += 8)
+    {
+      ch = *p++;
+      if (ch != 0)
+	for (mask = 0x80; mask; b++, mask >>= 1)
+	  if ((ch & mask) != 0)
+	    return b;
+    }
+
+  if (b < end)
+    {
+      ch = *p++;
+      for (mask = 0x80; b < end; b++, mask >>= 1)
+	if ((ch & mask) != 0)
+	  return b;
+      for (; mask && b < n; b++, mask >>= 1)
+	if ((ch & mask) != mask)
+	  return b;
+    }
+
+  for (; b + 8 <= n; b += 8)
+    {
+      ch = *p++;
+      if (ch != 0xff)
+	for (mask = 0x80; mask; b++, mask >>= 1)
+	  if ((ch & mask) != mask)
+	    return b;
+    }
+
+  return n;
+}
+
+static void
+initn_a(signed char p1, short p2, int p3, double p4, vector float p5)
+{
+  n_a i;
+
+  i.m1 = p1;
+  i.m2 = p2;
+  i.m3 = p3;
+  i.m4 = p4;
+  i.m5 = p5;
+
+  check(i.m1 == 77, "i.m1");
+  check(i.m2 == 1924, "i.m2");
+  check(i.m3 == -1471601920, "i.m3");
+  check(vec_all_eq(i.m5, ((vector float){-1.38e+09, 5.96e+08, 6.88e+08, -3.2e+08})),
+	"i.m5");
+
+  check(sizeof(n_a) == 32, "sizeof(n_a)");
+
+  check(offsetof(n_a, m1) == 0,  "offsetof(m1)");
+  check(offsetof(n_a, m2) == 2,  "offsetof(m2)");
+  check(offsetof(n_a, m3) == 4,  "offsetof(m3)");
+  check(offsetof(n_a, m4) == 8,  "offsetof(m4)");
+  check(offsetof(n_a, m5) == 16, "offsetof(m5)");
+
+  check(sizeof(i.m1) == 1,  "sizeof(m1)");
+  check(sizeof(i.m2) == 2,  "sizeof(m2)");
+  check(sizeof(i.m3) == 4,  "sizeof(m3)");
+  check(sizeof(i.m4) == 8,  "sizeof(m4)");
+  check(sizeof(i.m5) == 16, "sizeof(m5)");
+
+#define lay_check(field) do {				\
+  memset((char *)&i, 0xFF, sizeof(i));			\
+  lay_reset(field);					\
+  check(lay((char *)&i,					\
+	    offsetof(n_a, field),			\
+	    offsetof(n_a, field) + sizeof(i.field),	\
+	    sizeof(i)) == sizeof(i)*8,			\
+	"lay(" #field ")");				\
+  } while (0)
+#define lay_reset(field) i.field = 0
+
+  lay_check(m1);
+  lay_check(m2);
+  lay_check(m3);
+  lay_check(m4);
+#undef lay_reset
+#define lay_reset(field) i.field = ((vector float){0,0,0,0})
+  lay_check(m5);
+
+#undef lay_check
+#undef lay_reset
+}
+
+n_a
+valuen_a(void)
+{
+  return gn_a;
+}
+
+n_a *
+addrn_a(void)
+{
+  return &gn_a;
+}
+
+static void
+eqn_a(n_a * a)
+{
+  check(a->m1 == 77, "a->m1");
+  check(a->m2 == 1924, "a->m2");
+  check(a->m3 == -1471601920, "a->m3");
+  check(vec_all_eq(a->m5, ((vector float){-1.38e+09, 5.96e+08, 6.88e+08, -3.2e+08})),
+	"a->m5");
+}
+
+static void
+getsn_a(n_a * a)
+{
+  a->m1 = 77;
+  a->m2 = 1924;
+  a->m3 = -1471601920;
+  a->m4 = 3.65e+18;
+  a->m5 = ((vector float){-1.38e+09, 5.96e+08, 6.88e+08, -3.2e+08});
+}
+
+static void
+varlistn_a(signed char p1, va_list ap)
+{
+  n_a q;
+  q.m1 = p1;
+  q.m2 = va_arg(ap, int);
+  q.m3 = va_arg(ap, int);
+  q.m4 = va_arg(ap, double);
+  q.m5 = va_arg(ap, vector float);
+
+  check(q.m1 == 77, "q.m1");
+  check(q.m2 == 1924, "q.m2");
+  check(q.m3 == -1471601920, "q.m3");
+  check(vec_all_eq(q.m5, ((vector float){-1.38e+09, 5.96e+08, 6.88e+08, -3.2e+08})),
+	"q.m5");
+}
+
+static void
+varargsn_a(signed char p1, ...)
+{
+  n_a q, r;
+  va_list ap;
+
+  va_start(ap, p1);
+  q.m1 = p1;
+  q.m2 = va_arg(ap, int);
+  q.m3 = va_arg(ap, int);
+  q.m4 = va_arg(ap, double);
+  q.m5 = va_arg(ap, vector float);
+  va_end(ap);
+
+  check(q.m1 == 77, "q.m1");
+  check(q.m2 == 1924, "q.m2");
+  check(q.m3 == -1471601920, "q.m3");
+  check(vec_all_eq(q.m5, ((vector float){-1.38e+09, 5.96e+08, 6.88e+08, -3.2e+08})),
+	"q.m5");
+
+  va_start(ap, p1);
+  r.m1 = p1;
+  r.m2 = va_arg(ap, int);
+  r.m3 = va_arg(ap, int);
+  r.m4 = va_arg(ap, double);
+  r.m5 = va_arg(ap, vector float);
+  va_end(ap);
+
+  check(r.m1 == 77, "r.m1");
+  check(r.m2 == 1924, "r.m2");
+  check(r.m3 == -1471601920, "r.m3");
+  check(vec_all_eq(r.m5, ((vector float){-1.38e+09, 5.96e+08, 6.88e+08, -3.2e+08})),
+	"r.m5");
+
+  va_start(ap, p1);
+  varlistn_a(p1, ap);
+  va_end(ap);
+}
+
+static void
+test()
+{
+  static struct
+  {
+    char a;
+    n_a b;
+  }
+  s;
+  n_a v[3], a, *p;
+
+  static n_a i = { 77, 1924, -1471601920, 3.65e+18, {-1.38e+09, 5.96e+08, 6.88e+08, -3.2e+08} };
+
+  memset((char *)&(v), -1, sizeof(v));
+  v[1] = s.b;
+  check(lay((char *)&v, sizeof(n_a), sizeof(n_a)*2, sizeof(n_a)*3) == sizeof(n_a)*3*8,
+	"structure assignment");
+
+  check(i.m1 == 77, "i.m1");
+  check(i.m2 == 1924, "i.m2");
+  check(i.m3 == -1471601920, "i.m3");
+  check(vec_all_eq(i.m5, ((vector float){-1.38e+09, 5.96e+08, 6.88e+08, -3.2e+08})),
+	"i.m5");
+
+  initn_a(77, 1924, -1471601920, 3.65e+18, ((vector float){-1.38e+09, 5.96e+08, 6.88e+08, -3.2e+08}));
+  varargsn_a(77, 1924, -1471601920, 3.65e+18, ((vector float){-1.38e+09, 5.96e+08, 6.88e+08, -3.2e+08}));
+
+  gn_a.m1 = 77;
+  gn_a.m2 = 1924;
+  gn_a.m3 = -1471601920;
+  gn_a.m4 = 3.65e+18;
+  gn_a.m5 = ((vector float){-1.38e+09, 5.96e+08, 6.88e+08, -3.2e+08});
+  a = valuen_a();
+
+  check(a.m1 == 77, "a.m1");
+  check(a.m2 == 1924, "a.m2");
+  check(a.m3 == -1471601920, "a.m3");
+  check(vec_all_eq(a.m5, ((vector float){-1.38e+09, 5.96e+08, 6.88e+08, -3.2e+08})),
+	"a.m5");
+
+  p = addrn_a();
+
+  check(p->m1 == 77, "p->m1");
+  check(p->m2 == 1924, "p->m2");
+  check(p->m3 == -1471601920, "p->m3");
+  check(vec_all_eq(p->m5, ((vector float){-1.38e+09, 5.96e+08, 6.88e+08, -3.2e+08})),
+	"p->m5");
+
+  eqn_a(&a);
+
+  check(gn_a.m1 == 77, "gn_a.m1");
+  check(gn_a.m2 == 1924, "gn_a.m2");
+  check(gn_a.m3 == -1471601920, "gn_a.m3");
+  check(vec_all_eq(gn_a.m5, ((vector float){-1.38e+09, 5.96e+08, 6.88e+08, -3.2e+08})),
+	"gn_a.m5");
+
+  getsn_a(&v[0]);
+  v[2].m1 = v[0].m1;
+  v[2].m2 = v[0].m2;
+  v[2].m3 = v[0].m3;
+  v[2].m4 = v[0].m4;
+  v[2].m5 = v[0].m5;
+
+  check(v[2].m1 == 77, "v[2].m1");
+  check(v[2].m2 == 1924, "v[2].m2");
+  check(v[2].m3 == -1471601920, "v[2].m3");
+  check(vec_all_eq(v[2].m5, ((vector float){-1.38e+09, 5.96e+08, 6.88e+08, -3.2e+08})),
+	"v[2].m5");
+}