sse.md (V16): New mode iterator.

	* config/i386/sse.md (V16): New mode iterator.
	(VI1, VI8): Ditto.
	(AVXMODEQI, AVXMODEDI): Remove.
	(sse2, sse3): New mode attribute.
	(mov<mode>): Use V16 mode iterator.
	(*mov<mode>_internal): Merge with *avx_mov<mode>_internal.
	(push<mode>1): Use V16 mode iterator.
	(movmisalign<mode>): Ditto.
	(<sse>_movu<ssemodesuffix><avxmodesuffix>): Merge from
	<sse>_movu<ssemodesuffix> and avx_movu<ssemodesuffix><avxmodesuffix>.
	(*<sse>_movu<ssemodesuffix><avxmodesuffix>): Merge from
	*<sse>_movu<ssemodesuffix> and *avx_movu<ssemodesuffix><avxmodesuffix>.
	(<sse2>_movdqu<avxmodesuffix>): Merge from sse2_movdqu and
	avx_movdqu<avxmodesuffix>.
	(*<sse2>_movdqu<avxmodesuffix>): Merge from *sse2_movdqu and
	*avx_movdqu<avxmodesuffix>.
	(<sse3>_lddqu<avxmodesuffix>) Merge from sse3_lddqu and
	avx_lddqu<avxmodesuffix>.
	(<sse>_movnt<mode>): Merge with avx_movnt<AVXMODEF2P:mode>.
	(<sse2>_movnt<mode>): Merge from sse2_movntv2di and
	avx_movnt<AVXMODEDI:mode>.
	* config/i386/i386.c (ix86_expand_vector_move_misalign): Update for
	renamed sse_movups, sse2_movupd and sse2_movdqu patterns.

testsuite/ChangeLog:

	* gcc.target/i386/avx256-unaligned-load-1.c: Update scan patterns.
	* gcc.target/i386/avx256-unaligned-load-2.c: Ditto.
	* gcc.target/i386/avx256-unaligned-load-3.c: Ditto.
	* gcc.target/i386/avx256-unaligned-store-1.c: Ditto.
	* gcc.target/i386/avx256-unaligned-store-2.c: Ditto.
	* gcc.target/i386/avx256-unaligned-store-3.c: Ditto.

From-SVN: r172580

10 files changed, 197 insertions, 257 deletions

diff --git a/gcc/ChangeLog b/gcc/ChangeLog
index 99e1193..e3a583a 100644
--- a/gcc/ChangeLog
+++ b/gcc/ChangeLog
@@ -1,3 +1,29 @@
+2011-04-16  Uros Bizjak  <ubizjak@gmail.com>
+
+	* config/i386/sse.md (V16): New mode iterator.
+	(VI1, VI8): Ditto.
+	(AVXMODEQI, AVXMODEDI): Remove.
+	(sse2, sse3): New mode attribute.
+	(mov<mode>): Use V16 mode iterator.
+	(*mov<mode>_internal): Merge with *avx_mov<mode>_internal.
+	(push<mode>1): Use V16 mode iterator.
+	(movmisalign<mode>): Ditto.
+	(<sse>_movu<ssemodesuffix><avxmodesuffix>): Merge from
+	<sse>_movu<ssemodesuffix> and avx_movu<ssemodesuffix><avxmodesuffix>.
+	(*<sse>_movu<ssemodesuffix><avxmodesuffix>): Merge from
+	*<sse>_movu<ssemodesuffix> and *avx_movu<ssemodesuffix><avxmodesuffix>.
+	(<sse2>_movdqu<avxmodesuffix>): Merge from sse2_movdqu and
+	avx_movdqu<avxmodesuffix>.
+	(*<sse2>_movdqu<avxmodesuffix>): Merge from *sse2_movdqu and
+	*avx_movdqu<avxmodesuffix>.
+	(<sse3>_lddqu<avxmodesuffix>) Merge from sse3_lddqu and
+	avx_lddqu<avxmodesuffix>.
+	(<sse>_movnt<mode>): Merge with avx_movnt<AVXMODEF2P:mode>.
+	(<sse2>_movnt<mode>): Merge from sse2_movntv2di and
+	avx_movnt<AVXMODEDI:mode>.
+	* config/i386/i386.c (ix86_expand_vector_move_misalign): Update for
+	renamed sse_movups, sse2_movupd and sse2_movdqu patterns.
+
 2011-04-16  Bernd Schmidt  <bernds@codesourcery.com>
 
 	PR target/48629
@@ -24,8 +50,8 @@
 	* gimple.h (enum gf_mask): Add GF_CALL_ALLOCA_FOR_VAR.
 	(gimple_call_set_alloca_for_var): New inline function.
 	(gimple_call_alloca_for_var_p): Ditto.
-	* gimple.c (gimple_build_call_from_tree): Remember CALL_ALLOCA_FOR_VAR_P
-	state.
+	* gimple.c (gimple_build_call_from_tree): Remember
+	CALL_ALLOCA_FOR_VAR_P state.
 	* cfgexpand.c (expand_call_stmt): Restore CALL_ALLOCA_FOR_VAR_P state.
 
 	* tree-inline.c (inline_forbidden_p_stmt): Don't reject alloca
@@ -87,7 +113,7 @@
 	-fsched-pressure.
 
 2011-04-15  Georg-Johann Lay  <avr@gjlay.de>
-	
+
 	* config/avr/avr.md ("rotl<mode>3",mode=HIDI): Use match_dup
 	instead of match_operand for operand 3.
 
diff --git a/gcc/config/i386/i386.c b/gcc/config/i386/i386.c
index 6218702..36320d8 100644
--- a/gcc/config/i386/i386.c
+++ b/gcc/config/i386/i386.c
@@ -15769,12 +15769,12 @@ ix86_expand_vector_move_misalign (enum machine_mode mode, rtx operands[])
 		{
 		  op0 = gen_lowpart (V4SFmode, op0);
 		  op1 = gen_lowpart (V4SFmode, op1);
-		  emit_insn (gen_avx_movups (op0, op1));
+		  emit_insn (gen_sse_movups (op0, op1));
 		  return;
 		}
 	      op0 = gen_lowpart (V16QImode, op0);
 	      op1 = gen_lowpart (V16QImode, op1);
-	      emit_insn (gen_avx_movdqu (op0, op1));
+	      emit_insn (gen_sse2_movdqu (op0, op1));
 	      break;
 	    case 32:
 	      op0 = gen_lowpart (V32QImode, op0);
@@ -15792,7 +15792,7 @@ ix86_expand_vector_move_misalign (enum machine_mode mode, rtx operands[])
 	  switch (mode)
 	    {
 	    case V4SFmode:
-	      emit_insn (gen_avx_movups (op0, op1));
+	      emit_insn (gen_sse_movups (op0, op1));
 	      break;
 	    case V8SFmode:
 	      ix86_avx256_split_vector_move_misalign (op0, op1);
@@ -15802,10 +15802,10 @@ ix86_expand_vector_move_misalign (enum machine_mode mode, rtx operands[])
 		{
 		  op0 = gen_lowpart (V4SFmode, op0);
 		  op1 = gen_lowpart (V4SFmode, op1);
-		  emit_insn (gen_avx_movups (op0, op1));
+		  emit_insn (gen_sse_movups (op0, op1));
 		  return;
 		}
-	      emit_insn (gen_avx_movupd (op0, op1));
+	      emit_insn (gen_sse2_movupd (op0, op1));
 	      break;
 	    case V4DFmode:
 	      ix86_avx256_split_vector_move_misalign (op0, op1);
diff --git a/gcc/config/i386/sse.md b/gcc/config/i386/sse.md
index fabd35b..40e3972 100644
--- a/gcc/config/i386/sse.md
+++ b/gcc/config/i386/sse.md
@@ -18,6 +18,15 @@
 ;; along with GCC; see the file COPYING3.  If not see
 ;; <http://www.gnu.org/licenses/>.
 
+;; All vector modes including V1TImode.
+(define_mode_iterator V16
+  [(V32QI "TARGET_AVX") V16QI
+   (V16HI "TARGET_AVX") V8HI
+   (V8SI "TARGET_AVX") V4SI
+   (V4DI "TARGET_AVX") V2DI
+   V1TI
+   (V8SF "TARGET_AVX") V4SF
+   (V4DF "TARGET_AVX") V2DF])
 
 ;; All vector float modes
 (define_mode_iterator VF
@@ -43,6 +52,14 @@
    (V8SI "TARGET_AVX") V4SI
    (V4DI "TARGET_AVX") V2DI])
 
+;; All QImode vector integer modes
+(define_mode_iterator VI1
+  [(V32QI "TARGET_AVX") V16QI])
+
+;; All DImode vector integer modes
+(define_mode_iterator VI8
+  [(V4DI "TARGET_AVX") V2DI])
+
 ;; All 128bit vector integer modes
 (define_mode_iterator VI_128 [V16QI V8HI V4SI V2DI])
 
@@ -59,22 +76,13 @@
 
 ;; All 16-byte vector modes handled by SSE
 (define_mode_iterator SSEMODE [V16QI V8HI V4SI V2DI V4SF V2DF])
-(define_mode_iterator SSEMODE16 [V16QI V8HI V4SI V2DI V1TI V4SF V2DF])
 
 ;; All 32-byte vector modes handled by AVX
 (define_mode_iterator AVX256MODE [V32QI V16HI V8SI V4DI V8SF V4DF])
 
-;; All QI vector modes handled by AVX
-(define_mode_iterator AVXMODEQI [V32QI V16QI])
-
-;; All DI vector modes handled by AVX
-(define_mode_iterator AVXMODEDI [V4DI V2DI])
-
 ;; All vector modes handled by AVX
 (define_mode_iterator AVXMODE
   [V16QI V8HI V4SI V2DI V4SF V2DF V32QI V16HI V8SI V4DI V8SF V4DF])
-(define_mode_iterator AVXMODE16
-  [V16QI V8HI V4SI V2DI V1TI V4SF V2DF V32QI V16HI V8SI V4DI V8SF V4DF])
 
 ;; Mix-n-match
 (define_mode_iterator SSEMODE124 [V16QI V8HI V4SI])
@@ -107,8 +115,8 @@
 ;; Modes handled by storent patterns.
 (define_mode_iterator STORENT_MODE
   [(SF "TARGET_SSE4A") (DF "TARGET_SSE4A")
-   (SI "TARGET_SSE2") (V2DI "TARGET_SSE2") (V2DF "TARGET_SSE2")
-   (V4SF "TARGET_SSE")
+   (SI "TARGET_SSE2") (V2DI "TARGET_SSE2")
+   (V4SF "TARGET_SSE") (V2DF "TARGET_SSE2")
    (V4DF "TARGET_AVX") (V8SF "TARGET_AVX")])
 
 ;; Modes handled by vector extract patterns.
@@ -124,6 +132,13 @@
    (V4SF "sse") (V2DF "sse2")
    (V8SF "avx") (V4DF "avx")])
 
+(define_mode_attr sse2
+  [(V16QI "sse2") (V32QI "avx")
+   (V2DI "sse2") (V4DI "avx")])
+
+(define_mode_attr sse3
+  [(V16QI "sse3") (V32QI "avx")])
+
 (define_mode_attr sse4_1
   [(V4SF "sse4_1") (V2DF "sse4_1")
    (V8SF "avx") (V4DF "avx")])
@@ -192,19 +207,22 @@
 ;;
 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
 
+;; All of these patterns are enabled for SSE1 as well as SSE2.
+;; This is essential for maintaining stable calling conventions.
+
 (define_expand "mov<mode>"
-  [(set (match_operand:AVX256MODE 0 "nonimmediate_operand" "")
-	(match_operand:AVX256MODE 1 "nonimmediate_operand" ""))]
-  "TARGET_AVX"
+  [(set (match_operand:V16 0 "nonimmediate_operand" "")
+	(match_operand:V16 1 "nonimmediate_operand" ""))]
+  "TARGET_SSE"
 {
   ix86_expand_vector_move (<MODE>mode, operands);
   DONE;
 })
 
-(define_insn "*avx_mov<mode>_internal"
-  [(set (match_operand:AVXMODE16 0 "nonimmediate_operand" "=x,x ,m")
-	(match_operand:AVXMODE16 1 "nonimmediate_or_sse_const_operand"  "C ,xm,x"))]
-  "TARGET_AVX
+(define_insn "*mov<mode>_internal"
+  [(set (match_operand:V16 0 "nonimmediate_operand" "=x,x ,m")
+	(match_operand:V16 1 "nonimmediate_or_sse_const_operand"  "C ,xm,x"))]
+  "TARGET_SSE
    && (register_operand (operands[0], <MODE>mode)
        || register_operand (operands[1], <MODE>mode))"
 {
@@ -218,85 +236,51 @@
         {
 	case MODE_V8SF:
 	case MODE_V4SF:
-	  if (misaligned_operand (operands[0], <MODE>mode)
-	      || misaligned_operand (operands[1], <MODE>mode))
+	  if (TARGET_AVX
+	      && (misaligned_operand (operands[0], <MODE>mode)
+		  || misaligned_operand (operands[1], <MODE>mode)))
 	    return "vmovups\t{%1, %0|%0, %1}";
 	  else
-	    return "vmovaps\t{%1, %0|%0, %1}";
+	    return "%vmovaps\t{%1, %0|%0, %1}";
+
 	case MODE_V4DF:
 	case MODE_V2DF:
-	  if (misaligned_operand (operands[0], <MODE>mode)
-	      || misaligned_operand (operands[1], <MODE>mode))
+	  if (TARGET_AVX
+	      && (misaligned_operand (operands[0], <MODE>mode)
+		  || misaligned_operand (operands[1], <MODE>mode)))
 	    return "vmovupd\t{%1, %0|%0, %1}";
 	  else if (TARGET_SSE_PACKED_SINGLE_INSN_OPTIMAL)
-	    return "vmovaps\t{%1, %0|%0, %1}";
+	    return "%vmovaps\t{%1, %0|%0, %1}";
 	  else
-	    return "vmovapd\t{%1, %0|%0, %1}";
-	default:
-	  if (misaligned_operand (operands[0], <MODE>mode)
-	      || misaligned_operand (operands[1], <MODE>mode))
+	    return "%vmovapd\t{%1, %0|%0, %1}";
+
+	case MODE_OI:
+	case MODE_TI:
+	  if (TARGET_AVX
+	      && (misaligned_operand (operands[0], <MODE>mode)
+		  || misaligned_operand (operands[1], <MODE>mode)))
 	    return "vmovdqu\t{%1, %0|%0, %1}";
 	  else if (TARGET_SSE_PACKED_SINGLE_INSN_OPTIMAL)
-	    return "vmovaps\t{%1, %0|%0, %1}";
+	    return "%vmovaps\t{%1, %0|%0, %1}";
 	  else
-	    return "vmovdqa\t{%1, %0|%0, %1}";
-	}
-    default:
-      gcc_unreachable ();
-    }
-}
-  [(set_attr "type" "sselog1,ssemov,ssemov")
-   (set_attr "prefix" "vex")
-   (set_attr "mode" "<avxvecmode>")])
-
-;; All of these patterns are enabled for SSE1 as well as SSE2.
-;; This is essential for maintaining stable calling conventions.
-
-(define_expand "mov<mode>"
-  [(set (match_operand:SSEMODE16 0 "nonimmediate_operand" "")
-	(match_operand:SSEMODE16 1 "nonimmediate_operand" ""))]
-  "TARGET_SSE"
-{
-  ix86_expand_vector_move (<MODE>mode, operands);
-  DONE;
-})
+	    return "%vmovdqa\t{%1, %0|%0, %1}";
 
-(define_insn "*mov<mode>_internal"
-  [(set (match_operand:SSEMODE16 0 "nonimmediate_operand" "=x,x ,m")
-	(match_operand:SSEMODE16 1 "nonimmediate_or_sse_const_operand"  "C ,xm,x"))]
-  "TARGET_SSE
-   && (register_operand (operands[0], <MODE>mode)
-       || register_operand (operands[1], <MODE>mode))"
-{
-  switch (which_alternative)
-    {
-    case 0:
-      return standard_sse_constant_opcode (insn, operands[1]);
-    case 1:
-    case 2:
-      switch (get_attr_mode (insn))
-	{
-	case MODE_V4SF:
-	  return "movaps\t{%1, %0|%0, %1}";
-	case MODE_V2DF:
-	  if (TARGET_SSE_PACKED_SINGLE_INSN_OPTIMAL)
-	    return "movaps\t{%1, %0|%0, %1}";
-	  else
-	    return "movapd\t{%1, %0|%0, %1}";
 	default:
-	  if (TARGET_SSE_PACKED_SINGLE_INSN_OPTIMAL)
-	    return "movaps\t{%1, %0|%0, %1}";
-	  else
-	    return "movdqa\t{%1, %0|%0, %1}";
+	  gcc_unreachable ();
 	}
     default:
       gcc_unreachable ();
     }
 }
   [(set_attr "type" "sselog1,ssemov,ssemov")
+   (set_attr "prefix" "maybe_vex")
    (set (attr "mode")
-	(cond [(ior (ior (ne (symbol_ref "optimize_function_for_size_p (cfun)") (const_int 0))
-			 (eq (symbol_ref "TARGET_SSE2") (const_int 0)))
+	(cond [(ne (symbol_ref "TARGET_AVX") (const_int 0))
+	         (const_string "<avxvecmode>")
+	       (ior (ior
+	      	      (ne (symbol_ref "optimize_function_for_size_p (cfun)")
+		      	  (const_int 0))
+		      (eq (symbol_ref "TARGET_SSE2") (const_int 0)))
 		    (and (eq_attr "alternative" "2")
 			 (ne (symbol_ref "TARGET_SSE_TYPELESS_STORES")
 			     (const_int 0))))
@@ -308,6 +292,19 @@
 	      ]
 	  (const_string "TI")))])
 
+(define_insn "sse2_movq128"
+  [(set (match_operand:V2DI 0 "register_operand" "=x")
+	(vec_concat:V2DI
+	  (vec_select:DI
+	    (match_operand:V2DI 1 "nonimmediate_operand" "xm")
+	    (parallel [(const_int 0)]))
+	  (const_int 0)))]
+  "TARGET_SSE2"
+  "%vmovq\t{%1, %0|%0, %1}"
+  [(set_attr "type" "ssemov")
+   (set_attr "prefix" "maybe_vex")
+   (set_attr "mode" "TI")])
+
 ;; Move a DI from a 32-bit register pair (e.g. %edx:%eax) to an xmm.
 ;; We'd rather avoid this entirely; if the 32-bit reg pair was loaded
 ;; from memory, we'd prefer to load the memory directly into the %xmm
@@ -370,15 +367,7 @@
 })
 
 (define_expand "push<mode>1"
-  [(match_operand:AVX256MODE 0 "register_operand" "")]
-  "TARGET_AVX"
-{
-  ix86_expand_push (<MODE>mode, operands[0]);
-  DONE;
-})
-
-(define_expand "push<mode>1"
-  [(match_operand:SSEMODE16 0 "register_operand" "")]
+  [(match_operand:V16 0 "register_operand" "")]
   "TARGET_SSE"
 {
   ix86_expand_push (<MODE>mode, operands[0]);
@@ -386,169 +375,84 @@
 })
 
 (define_expand "movmisalign<mode>"
-  [(set (match_operand:AVX256MODE 0 "nonimmediate_operand" "")
-	(match_operand:AVX256MODE 1 "nonimmediate_operand" ""))]
-  "TARGET_AVX"
-{
-  ix86_expand_vector_move_misalign (<MODE>mode, operands);
-  DONE;
-})
-
-(define_expand "movmisalign<mode>"
-  [(set (match_operand:SSEMODE16 0 "nonimmediate_operand" "")
-	(match_operand:SSEMODE16 1 "nonimmediate_operand" ""))]
+  [(set (match_operand:V16 0 "nonimmediate_operand" "")
+	(match_operand:V16 1 "nonimmediate_operand" ""))]
   "TARGET_SSE"
 {
   ix86_expand_vector_move_misalign (<MODE>mode, operands);
   DONE;
 })
 
-(define_expand "avx_movu<ssemodesuffix><avxmodesuffix>"
-  [(set (match_operand:AVXMODEF2P 0 "nonimmediate_operand" "")
-	(unspec:AVXMODEF2P
-	  [(match_operand:AVXMODEF2P 1 "nonimmediate_operand" "")]
+(define_expand "<sse>_movu<ssemodesuffix><avxmodesuffix>"
+  [(set (match_operand:VF 0 "nonimmediate_operand" "")
+	(unspec:VF
+	  [(match_operand:VF 1 "nonimmediate_operand" "")]
 	  UNSPEC_MOVU))]
-  "AVX_VEC_FLOAT_MODE_P (<MODE>mode)"
+  ""
 {
   if (MEM_P (operands[0]) && MEM_P (operands[1]))
     operands[1] = force_reg (<MODE>mode, operands[1]);
 })
 
-(define_insn "*avx_movu<ssemodesuffix><avxmodesuffix>"
-  [(set (match_operand:AVXMODEF2P 0 "nonimmediate_operand" "=x,m")
-	(unspec:AVXMODEF2P
-	  [(match_operand:AVXMODEF2P 1 "nonimmediate_operand" "xm,x")]
+(define_insn "*<sse>_movu<ssemodesuffix><avxmodesuffix>"
+  [(set (match_operand:VF 0 "nonimmediate_operand" "=x,m")
+	(unspec:VF
+	  [(match_operand:VF 1 "nonimmediate_operand" "xm,x")]
 	  UNSPEC_MOVU))]
-  "AVX_VEC_FLOAT_MODE_P (<MODE>mode)
-   && !(MEM_P (operands[0]) && MEM_P (operands[1]))"
-  "vmovu<ssemodesuffix>\t{%1, %0|%0, %1}"
+  "!(MEM_P (operands[0]) && MEM_P (operands[1]))"
+  "%vmovu<ssemodesuffix>\t{%1, %0|%0, %1}"
   [(set_attr "type" "ssemov")
    (set_attr "movu" "1")
-   (set_attr "prefix" "vex")
-   (set_attr "mode" "<MODE>")])
-
-(define_insn "sse2_movq128"
-  [(set (match_operand:V2DI 0 "register_operand" "=x")
-	(vec_concat:V2DI
-	  (vec_select:DI
-	    (match_operand:V2DI 1 "nonimmediate_operand" "xm")
-	    (parallel [(const_int 0)]))
-	  (const_int 0)))]
-  "TARGET_SSE2"
-  "%vmovq\t{%1, %0|%0, %1}"
-  [(set_attr "type" "ssemov")
    (set_attr "prefix" "maybe_vex")
-   (set_attr "mode" "TI")])
-
-(define_expand "<sse>_movu<ssemodesuffix>"
-  [(set (match_operand:SSEMODEF2P 0 "nonimmediate_operand" "")
-	(unspec:SSEMODEF2P
-	  [(match_operand:SSEMODEF2P 1 "nonimmediate_operand" "")]
-	  UNSPEC_MOVU))]
-  "SSE_VEC_FLOAT_MODE_P (<MODE>mode)"
-{
-  if (MEM_P (operands[0]) && MEM_P (operands[1]))
-    operands[1] = force_reg (<MODE>mode, operands[1]);
-})
-
-(define_insn "*<sse>_movu<ssemodesuffix>"
-  [(set (match_operand:SSEMODEF2P 0 "nonimmediate_operand" "=x,m")
-	(unspec:SSEMODEF2P
-	  [(match_operand:SSEMODEF2P 1 "nonimmediate_operand" "xm,x")]
-	  UNSPEC_MOVU))]
-  "SSE_VEC_FLOAT_MODE_P (<MODE>mode)
-   && !(MEM_P (operands[0]) && MEM_P (operands[1]))"
-  "movu<ssemodesuffix>\t{%1, %0|%0, %1}"
-  [(set_attr "type" "ssemov")
-   (set_attr "movu" "1")
    (set_attr "mode" "<MODE>")])
 
-(define_expand "avx_movdqu<avxmodesuffix>"
-  [(set (match_operand:AVXMODEQI 0 "nonimmediate_operand" "")
-	(unspec:AVXMODEQI
-	  [(match_operand:AVXMODEQI 1 "nonimmediate_operand" "")]
-	  UNSPEC_MOVU))]
-  "TARGET_AVX"
+(define_expand "<sse2>_movdqu<avxmodesuffix>"
+  [(set (match_operand:VI1 0 "nonimmediate_operand" "")
+	(unspec:VI1 [(match_operand:VI1 1 "nonimmediate_operand" "")]
+		    UNSPEC_MOVU))]
+  "TARGET_SSE2"
 {
   if (MEM_P (operands[0]) && MEM_P (operands[1]))
     operands[1] = force_reg (<MODE>mode, operands[1]);
 })
 
-(define_insn "*avx_movdqu<avxmodesuffix>"
-  [(set (match_operand:AVXMODEQI 0 "nonimmediate_operand" "=x,m")
-	(unspec:AVXMODEQI
-	  [(match_operand:AVXMODEQI 1 "nonimmediate_operand" "xm,x")]
-	  UNSPEC_MOVU))]
-  "TARGET_AVX && !(MEM_P (operands[0]) && MEM_P (operands[1]))"
-  "vmovdqu\t{%1, %0|%0, %1}"
+(define_insn "*<sse2>_movdqu<avxmodesuffix>"
+  [(set (match_operand:VI1 0 "nonimmediate_operand" "=x,m")
+	(unspec:VI1 [(match_operand:VI1 1 "nonimmediate_operand" "xm,x")]
+		    UNSPEC_MOVU))]
+  "TARGET_SSE2 && !(MEM_P (operands[0]) && MEM_P (operands[1]))"
+  "%vmovdqu\t{%1, %0|%0, %1}"
   [(set_attr "type" "ssemov")
    (set_attr "movu" "1")
-   (set_attr "prefix" "vex")
+   (set (attr "prefix_data16")
+     (if_then_else
+       (ne (symbol_ref "TARGET_AVX") (const_int 0))
+     (const_string "*")
+     (const_string "1")))
+   (set_attr "prefix" "maybe_vex")
    (set_attr "mode" "<avxvecmode>")])
 
-(define_expand "sse2_movdqu"
-  [(set (match_operand:V16QI 0 "nonimmediate_operand" "")
-	(unspec:V16QI [(match_operand:V16QI 1 "nonimmediate_operand" "")]
-		      UNSPEC_MOVU))]
-  "TARGET_SSE2"
-{
-  if (MEM_P (operands[0]) && MEM_P (operands[1]))
-    operands[1] = force_reg (V16QImode, operands[1]);
-})
-
-(define_insn "*sse2_movdqu"
-  [(set (match_operand:V16QI 0 "nonimmediate_operand" "=x,m")
-	(unspec:V16QI [(match_operand:V16QI 1 "nonimmediate_operand" "xm,x")]
-		      UNSPEC_MOVU))]
-  "TARGET_SSE2 && !(MEM_P (operands[0]) && MEM_P (operands[1]))"
-  "movdqu\t{%1, %0|%0, %1}"
+(define_insn "<sse3>_lddqu<avxmodesuffix>"
+  [(set (match_operand:VI1 0 "register_operand" "=x")
+	(unspec:VI1 [(match_operand:VI1 1 "memory_operand" "m")]
+		    UNSPEC_LDDQU))]
+  "TARGET_SSE3"
+  "%vlddqu\t{%1, %0|%0, %1}"
   [(set_attr "type" "ssemov")
    (set_attr "movu" "1")
-   (set_attr "prefix_data16" "1")
-   (set_attr "mode" "TI")])
-
-(define_insn "avx_movnt<mode>"
-  [(set (match_operand:AVXMODEF2P 0 "memory_operand" "=m")
-	(unspec:AVXMODEF2P
-	  [(match_operand:AVXMODEF2P 1 "register_operand" "x")]
-	  UNSPEC_MOVNT))]
-  "AVX_VEC_FLOAT_MODE_P (<MODE>mode)"
-  "vmovnt<ssemodesuffix>\t{%1, %0|%0, %1}"
-  [(set_attr "type" "ssemov")
-   (set_attr "prefix" "vex")
-   (set_attr "mode" "<MODE>")])
-
-(define_insn "<sse>_movnt<mode>"
-  [(set (match_operand:SSEMODEF2P 0 "memory_operand" "=m")
-	(unspec:SSEMODEF2P
-	  [(match_operand:SSEMODEF2P 1 "register_operand" "x")]
-	  UNSPEC_MOVNT))]
-  "SSE_VEC_FLOAT_MODE_P (<MODE>mode)"
-  "movnt<ssemodesuffix>\t{%1, %0|%0, %1}"
-  [(set_attr "type" "ssemov")
-   (set_attr "mode" "<MODE>")])
-
-(define_insn "avx_movnt<mode>"
-  [(set (match_operand:AVXMODEDI 0 "memory_operand" "=m")
-	(unspec:AVXMODEDI
-	  [(match_operand:AVXMODEDI 1 "register_operand" "x")]
-	  UNSPEC_MOVNT))]
-  "TARGET_AVX"
-  "vmovntdq\t{%1, %0|%0, %1}"
-  [(set_attr "type" "ssecvt")
-   (set_attr "prefix" "vex")
+   (set (attr "prefix_data16")
+     (if_then_else
+       (ne (symbol_ref "TARGET_AVX") (const_int 0))
+     (const_string "*")
+     (const_string "0")))
+   (set (attr "prefix_rep")
+     (if_then_else
+       (ne (symbol_ref "TARGET_AVX") (const_int 0))
+     (const_string "*")
+     (const_string "1")))
+   (set_attr "prefix" "maybe_vex")
    (set_attr "mode" "<avxvecmode>")])
 
-(define_insn "sse2_movntv2di"
-  [(set (match_operand:V2DI 0 "memory_operand" "=m")
-	(unspec:V2DI [(match_operand:V2DI 1 "register_operand" "x")]
-		     UNSPEC_MOVNT))]
-  "TARGET_SSE2"
-  "movntdq\t{%1, %0|%0, %1}"
-  [(set_attr "type" "ssemov")
-   (set_attr "prefix_data16" "1")
-   (set_attr "mode" "TI")])
-
 (define_insn "sse2_movntsi"
   [(set (match_operand:SI 0 "memory_operand" "=m")
 	(unspec:SI [(match_operand:SI 1 "register_operand" "r")]
@@ -559,30 +463,31 @@
    (set_attr "prefix_data16" "0")
    (set_attr "mode" "V2DF")])
 
-(define_insn "avx_lddqu<avxmodesuffix>"
-  [(set (match_operand:AVXMODEQI 0 "register_operand" "=x")
-	(unspec:AVXMODEQI
-	  [(match_operand:AVXMODEQI 1 "memory_operand" "m")]
-	  UNSPEC_LDDQU))]
-  "TARGET_AVX"
-  "vlddqu\t{%1, %0|%0, %1}"
+(define_insn "<sse>_movnt<mode>"
+  [(set (match_operand:VF 0 "memory_operand" "=m")
+	(unspec:VF [(match_operand:VF 1 "register_operand" "x")]
+		   UNSPEC_MOVNT))]
+  "TARGET_SSE"
+  "%vmovnt<ssemodesuffix>\t{%1, %0|%0, %1}"
+  [(set_attr "type" "ssemov")
+   (set_attr "prefix" "maybe_vex")
+   (set_attr "mode" "<MODE>")])
+
+(define_insn "<sse2>_movnt<mode>"
+  [(set (match_operand:VI8 0 "memory_operand" "=m")
+	(unspec:VI8 [(match_operand:VI8 1 "register_operand" "x")]
+		    UNSPEC_MOVNT))]
+  "TARGET_SSE2"
+  "%vmovntdq\t{%1, %0|%0, %1}"
   [(set_attr "type" "ssecvt")
-   (set_attr "movu" "1")
-   (set_attr "prefix" "vex")
+   (set (attr "prefix_data16")
+     (if_then_else
+       (ne (symbol_ref "TARGET_AVX") (const_int 0))
+     (const_string "*")
+     (const_string "1")))
+   (set_attr "prefix" "maybe_vex")
    (set_attr "mode" "<avxvecmode>")])
 
-(define_insn "sse3_lddqu"
-  [(set (match_operand:V16QI 0 "register_operand" "=x")
-	(unspec:V16QI [(match_operand:V16QI 1 "memory_operand" "m")]
-		      UNSPEC_LDDQU))]
-  "TARGET_SSE3"
-  "lddqu\t{%1, %0|%0, %1}"
-  [(set_attr "type" "ssemov")
-   (set_attr "movu" "1")
-   (set_attr "prefix_data16" "0")
-   (set_attr "prefix_rep" "1")
-   (set_attr "mode" "TI")])
-
 ; Expand patterns for non-temporal stores.  At the moment, only those
 ; that directly map to insns are defined; it would be possible to
 ; define patterns for other modes that would expand to several insns.
diff --git a/gcc/testsuite/ChangeLog b/gcc/testsuite/ChangeLog
index ea66b49..97613c6 100644
--- a/gcc/testsuite/ChangeLog
+++ b/gcc/testsuite/ChangeLog
@@ -1,3 +1,12 @@
+2011-04-16  Uros Bizjak  <ubizjak@gmail.com>
+
+	* gcc.target/i386/avx256-unaligned-load-1.c: Update scan patterns.
+	* gcc.target/i386/avx256-unaligned-load-2.c: Ditto.
+	* gcc.target/i386/avx256-unaligned-load-3.c: Ditto.
+	* gcc.target/i386/avx256-unaligned-store-1.c: Ditto.
+	* gcc.target/i386/avx256-unaligned-store-2.c: Ditto.
+	* gcc.target/i386/avx256-unaligned-store-3.c: Ditto.
+
 2011-04-15  Jason Merrill  <jason@redhat.com>
 
 	* g++.dg/cpp0x/range-for17.C: New.
@@ -22,14 +31,14 @@
 2011-04-15  Nicola Pero  <nicola.pero@meta-innovation.com>
 
 	* objc.dg/naming-4.m: Updated.
-	* objc.dg/naming-5.m: Updated.	
+	* objc.dg/naming-5.m: Updated.
 	* objc.dg/naming-6.m: New.
-	* objc.dg/naming-7.m: New.	
+	* objc.dg/naming-7.m: New.
 	* obj-c++.dg/naming-1.mm: Updated.
 	* obj-c++.dg/naming-2.mm: Updated.
 	* obj-c++.dg/naming-3.mm: New.
 	* obj-c++.dg/naming-4.mm: New.
-	
+
 2011-04-15  Jerry DeLisle  <jvdelisle@gcc.gnu.org>
 
 	PR libgfortran/48589
diff --git a/gcc/testsuite/gcc.target/i386/avx256-unaligned-load-1.c b/gcc/testsuite/gcc.target/i386/avx256-unaligned-load-1.c
index 023e859..c909b94 100644
--- a/gcc/testsuite/gcc.target/i386/avx256-unaligned-load-1.c
+++ b/gcc/testsuite/gcc.target/i386/avx256-unaligned-load-1.c
@@ -15,5 +15,5 @@ avx_test (void)
 }
 
 /* { dg-final { scan-assembler-not "\\*avx_movups256/1" } } */
-/* { dg-final { scan-assembler "\\*avx_movups/1" } } */
+/* { dg-final { scan-assembler "\\*sse_movups/1" } } */
 /* { dg-final { scan-assembler "vinsertf128" } } */
diff --git a/gcc/testsuite/gcc.target/i386/avx256-unaligned-load-2.c b/gcc/testsuite/gcc.target/i386/avx256-unaligned-load-2.c
index 8394e27..7d188e2 100644
--- a/gcc/testsuite/gcc.target/i386/avx256-unaligned-load-2.c
+++ b/gcc/testsuite/gcc.target/i386/avx256-unaligned-load-2.c
@@ -25,5 +25,5 @@ avx_test (void)
 }
 
 /* { dg-final { scan-assembler-not "\\*avx_movdqu256/1" } } */
-/* { dg-final { scan-assembler "\\*avx_movdqu/1" } } */
+/* { dg-final { scan-assembler "\\*sse2_movdqu/1" } } */
 /* { dg-final { scan-assembler "vinsertf128" } } */
diff --git a/gcc/testsuite/gcc.target/i386/avx256-unaligned-load-3.c b/gcc/testsuite/gcc.target/i386/avx256-unaligned-load-3.c
index ec7d59d..6703054 100644
--- a/gcc/testsuite/gcc.target/i386/avx256-unaligned-load-3.c
+++ b/gcc/testsuite/gcc.target/i386/avx256-unaligned-load-3.c
@@ -15,5 +15,5 @@ avx_test (void)
 }
 
 /* { dg-final { scan-assembler-not "\\*avx_movupd256/1" } } */
-/* { dg-final { scan-assembler "\\*avx_movupd/1" } } */
+/* { dg-final { scan-assembler "\\*sse2_movupd/1" } } */
 /* { dg-final { scan-assembler "vinsertf128" } } */
diff --git a/gcc/testsuite/gcc.target/i386/avx256-unaligned-store-1.c b/gcc/testsuite/gcc.target/i386/avx256-unaligned-store-1.c
index 99db55c..6af02a2 100644
--- a/gcc/testsuite/gcc.target/i386/avx256-unaligned-store-1.c
+++ b/gcc/testsuite/gcc.target/i386/avx256-unaligned-store-1.c
@@ -18,5 +18,5 @@ avx_test (void)
 }
 
 /* { dg-final { scan-assembler-not "\\*avx_movups256/2" } } */
-/* { dg-final { scan-assembler "movups.*\\*avx_movv4sf_internal/3" } } */
+/* { dg-final { scan-assembler "vmovups.*\\*movv4sf_internal/3" } } */
 /* { dg-final { scan-assembler "vextractf128" } } */
diff --git a/gcc/testsuite/gcc.target/i386/avx256-unaligned-store-2.c b/gcc/testsuite/gcc.target/i386/avx256-unaligned-store-2.c
index 38ee9e2..d8ae5a9 100644
--- a/gcc/testsuite/gcc.target/i386/avx256-unaligned-store-2.c
+++ b/gcc/testsuite/gcc.target/i386/avx256-unaligned-store-2.c
@@ -25,5 +25,5 @@ avx_test (void)
 }
 
 /* { dg-final { scan-assembler-not "\\*avx_movdqu256/2" } } */
-/* { dg-final { scan-assembler "movdqu.*\\*avx_movv16qi_internal/3" } } */
+/* { dg-final { scan-assembler "vmovdqu.*\\*movv16qi_internal/3" } } */
 /* { dg-final { scan-assembler "vextractf128" } } */
diff --git a/gcc/testsuite/gcc.target/i386/avx256-unaligned-store-3.c b/gcc/testsuite/gcc.target/i386/avx256-unaligned-store-3.c
index eaab6fd..65e8204 100644
--- a/gcc/testsuite/gcc.target/i386/avx256-unaligned-store-3.c
+++ b/gcc/testsuite/gcc.target/i386/avx256-unaligned-store-3.c
@@ -18,5 +18,5 @@ avx_test (void)
 }
 
 /* { dg-final { scan-assembler-not "\\*avx_movupd256/2" } } */
-/* { dg-final { scan-assembler "movupd.*\\*avx_movv2df_internal/3" } } */
+/* { dg-final { scan-assembler "vmovupd.*\\*movv2df_internal/3" } } */
 /* { dg-final { scan-assembler "vextractf128" } } */