re PR target/37101 (wrong code: tree vectorizer omits bogus movq/movlps construct)

	PR target/37101
	* config/i386/sse.md (vec_concatv2di): Remove movlps alternative.
	(*vec_concatv2di_rex64_sse4_1): Ditto.
	(*vec_concatv2di_rex64_sse): Ditto.

testsuite/ChangeLog:

	PR target/37101
	* gcc.target/i386/pr37101.c: New test.

From-SVN: r139095

4 files changed, 101 insertions, 28 deletions

diff --git a/gcc/ChangeLog b/gcc/ChangeLog
index ed3129d..50a5355 100644
--- a/gcc/ChangeLog
+++ b/gcc/ChangeLog
@@ -1,3 +1,11 @@
+2008-08-14  Christophe Saout  <christophe@saout.de>
+	    Uros Bizjak  <ubizjak@gmail.com>
+
+	PR target/37101
+	* config/i386/sse.md (vec_concatv2di): Remove movlps alternative.
+	(*vec_concatv2di_rex64_sse4_1): Ditto.
+	(*vec_concatv2di_rex64_sse): Ditto.
+
 2008-08-14  Jakub Jelinek  <jakub@redhat.com>
 
 	PR middle-end/37103
@@ -139,7 +147,6 @@
 2008-08-12  Nathan Froyd  <froydnj@codesourcery.com>
 
 	PR libgomp/26165
-
 	* gcc.c (include_spec_function): Tweak call to find_a_file.
 
 2008-08-12  Jakub Jelinek  <jakub@redhat.com>
@@ -155,7 +162,7 @@
 
 2008-08-12  Anatoly Sokolov  <aesok@post.ru>
 
-        * final.c (final_scan_insn): Use app_enable/app_disable functions.
+	* final.c (final_scan_insn): Use app_enable/app_disable functions.
 
 2008-08-12  Ulrich Weigand  <Ulrich.Weigand@de.ibm.com>
 
@@ -685,8 +692,8 @@
 
 2008-08-07  Bob Wilson  <bob.wilson@acm.org>
 	
-	* config/xtensa/xtensa.c (xtensa_copy_incoming_a7): Copy incoming value
-	in a6 after the set_frame_ptr insn.
+	* config/xtensa/xtensa.c (xtensa_copy_incoming_a7): Copy incoming
+	value in a6 after the set_frame_ptr insn.
 	
 2008-08-07  Richard Henderson  <rth@redhat.com>
 
@@ -774,7 +781,7 @@
 	* matrix-reorg.c (compute_offset): Avoid C++ keywords.
 
 2008-08-06  Manuel Lopez-Ibanez  <manu@gcc.gnu.org>
-       
+
 	PR 26785
 	* diagnostic.c (permerror_at): New.
 	* toplev.h (permerror_at): Declare.
diff --git a/gcc/config/i386/sse.md b/gcc/config/i386/sse.md
index 1d0ae86..c8bf42d 100644
--- a/gcc/config/i386/sse.md
+++ b/gcc/config/i386/sse.md
@@ -4952,26 +4952,25 @@
    (set_attr "mode" "TI,V4SF,V2SF")])
 
 (define_insn "vec_concatv2di"
-  [(set (match_operand:V2DI 0 "register_operand"     "=Y2 ,?Y2,Y2,x,x,x")
+  [(set (match_operand:V2DI 0 "register_operand"     "=Y2 ,?Y2,Y2,x,x")
 	(vec_concat:V2DI
-	  (match_operand:DI 1 "nonimmediate_operand" " mY2,*y ,0 ,0,0,m")
-	  (match_operand:DI 2 "vector_move_operand"  " C  ,  C,Y2,x,m,0")))]
+	  (match_operand:DI 1 "nonimmediate_operand" " mY2,*y ,0 ,0,0")
+	  (match_operand:DI 2 "vector_move_operand"  " C  ,  C,Y2,x,m")))]
   "!TARGET_64BIT && TARGET_SSE"
   "@
    movq\t{%1, %0|%0, %1}
    movq2dq\t{%1, %0|%0, %1}
    punpcklqdq\t{%2, %0|%0, %2}
    movlhps\t{%2, %0|%0, %2}
-   movhps\t{%2, %0|%0, %2}
-   movlps\t{%1, %0|%0, %1}"
-  [(set_attr "type" "ssemov,ssemov,sselog,ssemov,ssemov,ssemov")
-   (set_attr "mode" "TI,TI,TI,V4SF,V2SF,V2SF")])
+   movhps\t{%2, %0|%0, %2}"
+  [(set_attr "type" "ssemov,ssemov,sselog,ssemov,ssemov")
+   (set_attr "mode" "TI,TI,TI,V4SF,V2SF")])
 
 (define_insn "*vec_concatv2di_rex64_sse4_1"
-  [(set (match_operand:V2DI 0 "register_operand"     "=x ,x ,Yi,!x,x,x,x,x")
+  [(set (match_operand:V2DI 0 "register_operand"     "=x ,x ,Yi,!x,x,x,x")
 	(vec_concat:V2DI
-	  (match_operand:DI 1 "nonimmediate_operand" " 0 ,mx,r ,*y,0,0,0,m")
-	  (match_operand:DI 2 "vector_move_operand"  " rm,C ,C ,C ,x,x,m,0")))]
+	  (match_operand:DI 1 "nonimmediate_operand" " 0 ,mx,r ,*y,0,0,0")
+	  (match_operand:DI 2 "vector_move_operand"  " rm,C ,C ,C ,x,x,m")))]
   "TARGET_64BIT && TARGET_SSE4_1"
   "@
    pinsrq\t{$0x1, %2, %0|%0, %2, 0x1}
@@ -4980,17 +4979,16 @@
    movq2dq\t{%1, %0|%0, %1}
    punpcklqdq\t{%2, %0|%0, %2}
    movlhps\t{%2, %0|%0, %2}
-   movhps\t{%2, %0|%0, %2}
-   movlps\t{%1, %0|%0, %1}"
-  [(set_attr "type" "sselog,ssemov,ssemov,ssemov,sselog,ssemov,ssemov,ssemov")
-   (set_attr "prefix_extra" "1,*,*,*,*,*,*,*")
-   (set_attr "mode" "TI,TI,TI,TI,TI,V4SF,V2SF,V2SF")])
+   movhps\t{%2, %0|%0, %2}"
+  [(set_attr "type" "sselog,ssemov,ssemov,ssemov,sselog,ssemov,ssemov")
+   (set_attr "prefix_extra" "1,*,*,*,*,*,*")
+   (set_attr "mode" "TI,TI,TI,TI,TI,V4SF,V2SF")])
 
 (define_insn "*vec_concatv2di_rex64_sse"
-  [(set (match_operand:V2DI 0 "register_operand"     "=Y2 ,Yi,!Y2,Y2,x,x,x")
+  [(set (match_operand:V2DI 0 "register_operand"     "=Y2 ,Yi,!Y2,Y2,x,x")
 	(vec_concat:V2DI
-	  (match_operand:DI 1 "nonimmediate_operand" " mY2,r ,*y ,0 ,0,0,m")
-	  (match_operand:DI 2 "vector_move_operand"  " C  ,C ,C  ,Y2,x,m,0")))]
+	  (match_operand:DI 1 "nonimmediate_operand" " mY2,r ,*y ,0 ,0,0")
+	  (match_operand:DI 2 "vector_move_operand"  " C  ,C ,C  ,Y2,x,m")))]
   "TARGET_64BIT && TARGET_SSE"
   "@
    movq\t{%1, %0|%0, %1}
@@ -4998,10 +4996,9 @@
    movq2dq\t{%1, %0|%0, %1}
    punpcklqdq\t{%2, %0|%0, %2}
    movlhps\t{%2, %0|%0, %2}
-   movhps\t{%2, %0|%0, %2}
-   movlps\t{%1, %0|%0, %1}"
-  [(set_attr "type" "ssemov,ssemov,ssemov,sselog,ssemov,ssemov,ssemov")
-   (set_attr "mode" "TI,TI,TI,TI,V4SF,V2SF,V2SF")])
+   movhps\t{%2, %0|%0, %2}"
+  [(set_attr "type" "ssemov,ssemov,ssemov,sselog,ssemov,ssemov")
+   (set_attr "mode" "TI,TI,TI,TI,V4SF,V2SF")])
 
 (define_expand "vec_unpacku_hi_v16qi"
   [(match_operand:V8HI 0 "register_operand" "")
diff --git a/gcc/testsuite/ChangeLog b/gcc/testsuite/ChangeLog
index ca3aa251..1430434 100644
--- a/gcc/testsuite/ChangeLog
+++ b/gcc/testsuite/ChangeLog
@@ -1,3 +1,8 @@
+2008-08-14  Uros Bizjak  <ubizjak@gmail.com>
+
+	PR target/37101
+	* gcc.target/i386/pr37101.c: New test.
+
 2008-08-14  Jakub Jelinek  <jakub@redhat.com>
 
 	PR middle-end/37103
@@ -126,7 +131,7 @@
 	loop.
 	Should be vectorizable on targets that support vector unpack.
 	* gcc.dg/vect/vect-widen-mult-u8.c,
-	gcc.dg/vect/wrapv-vect-reduc-dot-s8b.c,.
+	gcc.dg/vect/wrapv-vect-reduc-dot-s8b.c,
 	gcc.dg/vect/wrapv-vect-reduc-dot-s8b.c: Likewise.
 	* gcc.dg/vect/vect-35.c: Should be vectorizable on targets that
 	support vector pack. Avoid vectorization of the init loop.
diff --git a/gcc/testsuite/gcc.target/i386/pr37101.c b/gcc/testsuite/gcc.target/i386/pr37101.c
new file mode 100644
index 0000000..8fd3bfc
--- /dev/null
+++ b/gcc/testsuite/gcc.target/i386/pr37101.c
@@ -0,0 +1,64 @@
+/* { dg-do compile } */
+/* { dg-options "-O2 -msse2 -ftree-vectorize -march=nocona" } */
+
+typedef __SIZE_TYPE__ size_t;
+extern void *malloc (size_t);
+extern void free (void *);
+
+typedef struct _Resource
+{
+  struct _Resource *next;
+  unsigned int id;
+} ResourceRec, *ResourcePtr;
+
+typedef struct _ClientResource
+{
+  ResourcePtr *resources;
+  int elements;
+  int buckets;
+  int hashsize;
+} ClientResourceRec;
+
+static ClientResourceRec clientTable[256];
+
+void
+RebuildTable (int client)
+{
+  int j;
+  ResourcePtr res, next;
+  ResourcePtr **tails, *resources;
+  ResourcePtr **tptr, *rptr;
+
+  j = 2 * clientTable[client].buckets;
+
+  tails =
+    (ResourcePtr **) malloc ((unsigned long) (j * sizeof (ResourcePtr *)));
+  resources =
+    (ResourcePtr *) malloc ((unsigned long) (j * sizeof (ResourcePtr)));
+
+  for (rptr = resources, tptr = tails; --j >= 0; rptr++, tptr++)
+    {
+      *rptr = ((ResourcePtr) ((void *) 0));
+      *tptr = rptr;
+    }
+
+  clientTable[client].hashsize++;
+  for (j = clientTable[client].buckets,
+       rptr = clientTable[client].resources; --j >= 0; rptr++)
+    {
+      for (res = *rptr; res; res = next)
+	{
+	  next = res->next;
+	  res->next = ((ResourcePtr) ((void *) 0));
+	  tptr = &tails[Hash (client, res->id)];
+	  **tptr = res;
+	  *tptr = &res->next;
+	}
+    }
+  free ((void *) tails);
+  clientTable[client].buckets *= 2;
+  free ((void *) clientTable[client].resources);
+  clientTable[client].resources = resources;
+}
+
+/* { dg-final { scan-assembler-not "movlps" } } */