[Bug tree-optimization/54855] Unnecessary duplication when performing scalar operation on vector element

["glisse at gcc dot gnu.org" <gcc-bugzilla at gcc dot gnu dot org>]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54855

--- Comment #3 from Marc Glisse <glisse at gcc dot gnu.org> 2012-10-12 17:08:20 UTC ---
The following patch gives this loop:
.L7:
    subsd    %xmm0, %xmm1
    subl    $1, %eax
    addpd    %xmm1, %xmm1
    jne    .L7

I guess I should add the same for mul and div at the same time, but I don't
know if it is the right approach.

--- config/i386/sse.md    (revision 192405)
+++ config/i386/sse.md    (working copy)
@@ -812,20 +812,38 @@
       (const_int 1)))]
   "TARGET_SSE"
   "@
    <plusminus_mnemonic><ssescalarmodesuffix>\t{%2, %0|%0, %2}
    v<plusminus_mnemonic><ssescalarmodesuffix>\t{%2, %1, %0|%0, %1, %2}"
   [(set_attr "isa" "noavx,avx")
    (set_attr "type" "sseadd")
    (set_attr "prefix" "orig,vex")
    (set_attr "mode" "<ssescalarmode>")])

+(define_insn "*sse2_vm<plusminus_insn>v2df3"
+  [(set (match_operand:V2DF 0 "register_operand" "=x,x")
+    (vec_concat:V2DF
+      (plusminus:DF
+        (vec_select:DF 
+          (match_operand:V2DF 1 "register_operand" "0,x")
+          (parallel [(const_int 0)]))
+        (match_operand:DF 2 "nonimmediate_operand" "xm,xm"))
+      (vec_select:DF (match_dup 1) (parallel [(const_int 1)]))))]
+  "TARGET_SSE2"
+  "@
+   <plusminus_mnemonic>sd\t{%2, %0|%0, %2}
+   v<plusminus_mnemonic>sd\t{%2, %1, %0|%0, %1, %2}"
+  [(set_attr "isa" "noavx,avx")
+   (set_attr "type" "sseadd")
+   (set_attr "prefix" "orig,vex")
+   (set_attr "mode" "DF")])
+
 (define_expand "mul<mode>3"
   [(set (match_operand:VF 0 "register_operand")
     (mult:VF
       (match_operand:VF 1 "nonimmediate_operand")
       (match_operand:VF 2 "nonimmediate_operand")))]
   "TARGET_SSE"
   "ix86_fixup_binary_operands_no_copy (MULT, <MODE>mode, operands);")

 (define_insn "*mul<mode>3"
   [(set (match_operand:VF 0 "register_operand" "=x,x")