Fix a bug in the scalarization of BUILD_VECTOR. BUILD_VECTOR elements may be wider than the output element type. Make sure to trunc them if needed.
Together with Michael Kuperstein <michael.m.kuperstein@intel.com>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@160235 91177308-0d34-0410-b5e6-96231b3b80d8
diff --git a/test/CodeGen/Generic/2012-07-15-BuildVectorPromote.ll b/test/CodeGen/Generic/2012-07-15-BuildVectorPromote.ll
new file mode 100644
index 0000000..6591c64
--- /dev/null
+++ b/test/CodeGen/Generic/2012-07-15-BuildVectorPromote.ll
@@ -0,0 +1,8 @@
+; RUN: llc -mcpu=corei7 < %s
+; We don't care about the output, just that it doesn't crash
+
+define <1 x i1> @buildvec_promote() {
+ %cmp = icmp ule <1 x i32> undef, undef
+ %sel = select i1 undef, <1 x i1> undef, <1 x i1> %cmp
+ ret <1 x i1> %sel
+}