Get rid of some bogus patterns for X86vzmovl. Don't create VZEXT_MOVL

nodes for vectors with an i16 element type.  Add an optimization for 
building a vector which is all zeros/undef except for the bottom 
element, where the bottom element is an i8 or i16.

llvm-svn: 72988

1 files changed, 37 insertions, 0 deletions

diff --git a/llvm/test/CodeGen/X86/2009-06-05-VZextByteShort.ll b/llvm/test/CodeGen/X86/2009-06-05-VZextByteShort.ll
new file mode 100644
index 00000000000..220423aa986
--- /dev/null
+++ b/llvm/test/CodeGen/X86/2009-06-05-VZextByteShort.ll
@@ -0,0 +1,37 @@
+; RUN: llvm-as < %s | llc -march=x86 -mattr=+mmx,+sse2 > %t1
+; RUN: grep movzwl %t1 | count 2
+; RUN: grep movzbl %t1 | count 2
+; RUN: grep movd %t1 | count 4
+
+define <4 x i16> @a(i32* %x1) nounwind {
+  %x2 = load i32* %x1
+  %x3 = lshr i32 %x2, 1
+  %x = trunc i32 %x3 to i16
+  %r = insertelement <4 x i16> zeroinitializer, i16 %x, i32 0
+  ret <4 x i16> %r
+}
+
+define <8 x i16> @b(i32* %x1) nounwind {
+  %x2 = load i32* %x1
+  %x3 = lshr i32 %x2, 1
+  %x = trunc i32 %x3 to i16
+  %r = insertelement <8 x i16> zeroinitializer, i16 %x, i32 0
+  ret <8 x i16> %r
+}
+
+define <8 x i8> @c(i32* %x1) nounwind {
+  %x2 = load i32* %x1
+  %x3 = lshr i32 %x2, 1
+  %x = trunc i32 %x3 to i8
+  %r = insertelement <8 x i8> zeroinitializer, i8 %x, i32 0
+  ret <8 x i8> %r
+}
+
+define <16 x i8> @d(i32* %x1) nounwind {
+  %x2 = load i32* %x1
+  %x3 = lshr i32 %x2, 1
+  %x = trunc i32 %x3 to i8
+  %r = insertelement <16 x i8> zeroinitializer, i8 %x, i32 0
+  ret <16 x i8> %r
+}
+