re PR target/32000 (x86 backend uses aligned load on unaligned memory)

gcc/

2008-03-31  H.J. Lu  <hongjiu.lu@intel.com>

	PR target/32000
	* config/i386/i386.md (*movti_internal): Emit unaligned SSE
	load/store if memory is unaligned.
	(*movti_rex64): Likewise.

	* config/i386/predicates.md (misaligned_operand): New.

gcc/testsuite/

2008-03-31  H.J. Lu  <hongjiu.lu@intel.com>

	PR target/32000
	* gcc.target/i386/pr32000-1.c: New.

From-SVN: r133753

5 files changed, 77 insertions, 6 deletions

diff --git a/gcc/ChangeLog b/gcc/ChangeLog
index f7be8f1..0a426b1 100644
--- a/gcc/ChangeLog
+++ b/gcc/ChangeLog
@@ -1,3 +1,12 @@
+2008-03-31  H.J. Lu  <hongjiu.lu@intel.com>
+
+	PR target/32000
+	* config/i386/i386.md (*movti_internal): Emit unaligned SSE
+	load/store if memory is unaligned.
+	(*movti_rex64): Likewise.
+
+	* config/i386/predicates.md (misaligned_operand): New.
+
 2008-03-31  Andrew Pinski  <pinskia@gmail.com>
 
 	PR tree-opt/35431
diff --git a/gcc/config/i386/i386.md b/gcc/config/i386/i386.md
index 8b0a280..1135799 100644
--- a/gcc/config/i386/i386.md
+++ b/gcc/config/i386/i386.md
@@ -2385,10 +2385,23 @@
 	return "pxor\t%0, %0";
     case 1:
     case 2:
-      if (get_attr_mode (insn) == MODE_V4SF)
-	return "movaps\t{%1, %0|%0, %1}";
+      /* TDmode values are passed as TImode on the stack.  Moving them
+	 to stack may result in unaligned memory access.  */
+      if (misaligned_operand (operands[0], TImode)
+	  || misaligned_operand (operands[1], TImode))
+	{ 
+	  if (get_attr_mode (insn) == MODE_V4SF)
+	    return "movups\t{%1, %0|%0, %1}";
+	 else
+	   return "movdqu\t{%1, %0|%0, %1}";
+	}
       else
-	return "movdqa\t{%1, %0|%0, %1}";
+	{ 
+	  if (get_attr_mode (insn) == MODE_V4SF)
+	    return "movaps\t{%1, %0|%0, %1}";
+	 else
+	   return "movdqa\t{%1, %0|%0, %1}";
+	}
     default:
       gcc_unreachable ();
     }
@@ -2422,10 +2435,23 @@
 	return "pxor\t%0, %0";
     case 3:
     case 4:
-      if (get_attr_mode (insn) == MODE_V4SF)
-	return "movaps\t{%1, %0|%0, %1}";
+      /* TDmode values are passed as TImode on the stack.  Moving them
+	 to stack may result in unaligned memory access.  */
+      if (misaligned_operand (operands[0], TImode)
+	  || misaligned_operand (operands[1], TImode))
+	{ 
+	  if (get_attr_mode (insn) == MODE_V4SF)
+	    return "movups\t{%1, %0|%0, %1}";
+	 else
+	   return "movdqu\t{%1, %0|%0, %1}";
+	}
       else
-	return "movdqa\t{%1, %0|%0, %1}";
+	{ 
+	  if (get_attr_mode (insn) == MODE_V4SF)
+	    return "movaps\t{%1, %0|%0, %1}";
+	 else
+	   return "movdqa\t{%1, %0|%0, %1}";
+	}
     default:
       gcc_unreachable ();
     }
diff --git a/gcc/config/i386/predicates.md b/gcc/config/i386/predicates.md
index 7ba7289..5389aa9 100644
--- a/gcc/config/i386/predicates.md
+++ b/gcc/config/i386/predicates.md
@@ -1043,3 +1043,8 @@
 
 (define_predicate "absneg_operator"
   (match_code "abs,neg"))
+
+;; Return 1 if OP is misaligned memory operand
+(define_predicate "misaligned_operand"
+  (and (match_code "mem")
+       (match_test "MEM_ALIGN (op) < GET_MODE_ALIGNMENT (mode)")))
diff --git a/gcc/testsuite/ChangeLog b/gcc/testsuite/ChangeLog
index 06b43d9..4353823 100644
--- a/gcc/testsuite/ChangeLog
+++ b/gcc/testsuite/ChangeLog
@@ -1,3 +1,8 @@
+2008-03-31  H.J. Lu  <hongjiu.lu@intel.com>
+
+	PR target/32000
+	* gcc.target/i386/pr32000-1.c: New.
+
 2008-03-31  Andrew Pinski  <pinskia@gmail.com>
 
 	PR tree-opt/35431
diff --git a/gcc/testsuite/gcc.target/i386/pr32000-1.c b/gcc/testsuite/gcc.target/i386/pr32000-1.c
new file mode 100644
index 0000000..483d24a
--- /dev/null
+++ b/gcc/testsuite/gcc.target/i386/pr32000-1.c
@@ -0,0 +1,26 @@
+/* { dg-do run } */
+/* { dg-require-effective-target dfp } */
+/* { dg-options "-O -msse2 -std=gnu99" } */
+
+#include "sse2-check.h"
+
+typedef struct { _Decimal128 f __attribute__((packed)); } packed;
+
+_Decimal128 __attribute__((noinline))
+foo (_Decimal128 a1, _Decimal128 a2, _Decimal128 a3, _Decimal128 a4,
+     _Decimal128 a5, _Decimal128 a6, _Decimal128 a7, _Decimal128 a8,
+     int b1, int b2, int b3, int b4, int b5, int b6, int b7, packed y)
+{
+  return y.f;
+}
+
+void
+sse2_test (void)
+{
+  packed x;
+  _Decimal128 y = -1;
+  x.f = y;
+  y = foo (0, 0, 0, 0, 0, 0, 0, 0, 1, 2, 3, 4, 5, 6, -1, x);
+  if (__builtin_memcmp (&y, &x.f, sizeof (y)))
+    abort ();
+}