Fix PR24354.

`InstCombiner::OptimizeOverflowCheck` was asserting an
invariant (operands to binary operations are ordered by decreasing
complexity) that wasn't really an invariant.  Fix this by instead having
`InstCombiner::OptimizeOverflowCheck` establish the invariant if it does
not hold.

llvm-svn: 244676

1 files changed, 33 insertions, 0 deletions

diff --git a/llvm/test/Transforms/InstCombine/pr24354.ll b/llvm/test/Transforms/InstCombine/pr24354.ll
new file mode 100644
index 00000000000..3b36fd1b74e
--- /dev/null
+++ b/llvm/test/Transforms/InstCombine/pr24354.ll
@@ -0,0 +1,33 @@
+; RUN: opt -instcombine -S < %s | FileCheck %s
+
+; This used to crash opt
+
+@c = common global i32 0, align 4
+@b = common global i32 0, align 4
+@a = common global i16 0, align 2
+@d = common global i32 0, align 4
+
+define void @fn3() {
+; CHECK: @fn3
+bb:
+  %tmp = load i32, i32* @c, align 4
+  %tmp1 = icmp eq i32 %tmp, 0
+  br i1 %tmp1, label %bb2, label %bb6
+
+bb2:                                              ; preds = %bb
+  %tmp3 = load i32, i32* @b, align 4
+  %tmp.i = add nsw i32 255, %tmp3
+  %tmp5 = icmp ugt i32 %tmp.i, 254
+  br label %bb6
+
+bb6:                                              ; preds = %bb, %bb2
+  %tmp7 = phi i1 [ true, %bb ], [ %tmp5, %bb2 ]
+  %tmp8 = zext i1 %tmp7 to i32
+  %tmp10 = icmp eq i32 %tmp8, 0
+  %tmp12 = load i16, i16* @a, align 2
+  %tmp14 = icmp ne i16 %tmp12, 0
+  %tmp16 = select i1 %tmp10, i1 false, i1 %tmp14
+  %tmp17 = zext i1 %tmp16 to i32
+  store i32 %tmp17, i32* @d, align 4
+  ret void
+}