Improve the loading of load-anyext vectors by allowing the codegen to load
multiple scalars and insert them into a vector. Next, we shuffle the elements
into the correct places, as before.
Also fix a small dagcombine bug in SimplifyBinOpWithSameOpcodeHands, when the
migration of bitcasts happened too late in the SelectionDAG process.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@159991 91177308-0d34-0410-b5e6-96231b3b80d8
diff --git a/test/CodeGen/X86/2012-07-10-extload64.ll b/test/CodeGen/X86/2012-07-10-extload64.ll
new file mode 100644
index 0000000..3284f5e
--- /dev/null
+++ b/test/CodeGen/X86/2012-07-10-extload64.ll
@@ -0,0 +1,13 @@
+; RUN: llc < %s -march=x86 -mcpu=corei7 -mtriple=i686-pc-win32 | FileCheck %s
+
+; CHECK: load_store
+define void @load_store(<4 x i16>* %in) {
+entry:
+  %A27 = load <4 x i16>* %in, align 4
+  %A28 = add <4 x i16> %A27, %A27
+  store <4 x i16> %A28, <4 x i16>* %in, align 4
+  ret void
+; CHECK: movd
+; CHECK: pinsrd
+; CHECK: ret
+}
diff --git a/test/CodeGen/X86/vec_compare-2.ll b/test/CodeGen/X86/vec_compare-2.ll
index 91777f7..46d6a23 100644
--- a/test/CodeGen/X86/vec_compare-2.ll
+++ b/test/CodeGen/X86/vec_compare-2.ll
@@ -10,8 +10,7 @@
 entry:
 ; CHECK: cfi_def_cfa_offset
 ; CHECK-NOT: set
-; CHECK: movzwl
-; CHECK: movzwl
+; CHECK: punpcklwd
 ; CHECK: pshufd
 ; CHECK: pshufb
   %shr.i = ashr <4 x i32> zeroinitializer, <i32 3, i32 3, i32 3, i32 3> ; <<4 x i32>> [#uses=1]
diff --git a/test/CodeGen/X86/widen_load-0.ll b/test/CodeGen/X86/widen_load-0.ll
index 4aeec91..d543728 100644
--- a/test/CodeGen/X86/widen_load-0.ll
+++ b/test/CodeGen/X86/widen_load-0.ll
@@ -1,18 +1,12 @@
 ; RUN: llc < %s -o - -mtriple=x86_64-linux -mcpu=corei7 | FileCheck %s
-; RUN: llc < %s -o - -mtriple=x86_64-win32 -mcpu=corei7 | FileCheck %s -check-prefix=WIN64
 ; PR4891
 
 ; Both loads should happen before either store.
 
-; CHECK: movd  ({{.*}}), {{.*}}
-; CHECK: movd  ({{.*}}), {{.*}}
-; CHECK: movd  {{.*}}, ({{.*}})
-; CHECK: movd  {{.*}}, ({{.*}})
-
-; WIN64: movd  ({{.*}}), {{.*}}
-; WIN64: movd  ({{.*}}), {{.*}}
-; WIN64: movd  {{.*}}, ({{.*}})
-; WIN64: movd  {{.*}}, ({{.*}})
+; CHECK: movl  ({{.*}}), {{.*}}
+; CHECK: movl  ({{.*}}), {{.*}}
+; CHECK: movl  {{.*}}, ({{.*}})
+; CHECK: movl  {{.*}}, ({{.*}})
 
 define void @short2_int_swap(<2 x i16>* nocapture %b, i32* nocapture %c) nounwind {
 entry: