When determining if we can fold (x >> C1) << C2, the bits that we need to verify are zero

are not the low bits of x, but the bits that WILL be the low bits after the operation completes.

llvm-svn: 122529

1 files changed, 19 insertions, 0 deletions

diff --git a/llvm/test/Transforms/InstCombine/shift.ll b/llvm/test/Transforms/InstCombine/shift.ll
index 8d1c82991fd..4f6939d6323 100644
--- a/llvm/test/Transforms/InstCombine/shift.ll
+++ b/llvm/test/Transforms/InstCombine/shift.ll
@@ -452,3 +452,22 @@ define i32 @test38(i32 %x) nounwind readnone {
 ; CHECK-NEXT: ret i32
 }
 
+; <rdar://problem/8756731>
+; CHECK: @test39
+define i8 @test39(i32 %a0) {
+entry:
+  %tmp4 = trunc i32 %a0 to i8
+; CHECK: and i8 %tmp49, 64
+  %tmp5 = shl i8 %tmp4, 5
+  %tmp48 = and i8 %tmp5, 32
+  %tmp49 = lshr i8 %tmp48, 5
+  %tmp50 = mul i8 %tmp49, 64
+  %tmp51 = xor i8 %tmp50, %tmp5
+; CHECK: and i8 %0, 16
+  %tmp52 = and i8 %tmp51, -128
+  %tmp53 = lshr i8 %tmp52, 7
+  %tmp54 = mul i8 %tmp53, 16
+  %tmp55 = xor i8 %tmp54, %tmp51
+; CHECK: ret i8 %tmp551
+  ret i8 %tmp55
+}