[llvm] [VectorCombine] fix division by zero in foldSelectShuffle (PR #156779)

[Mikael Holmen via llvm-commits]

================
@@ -0,0 +1,21 @@
+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --version 5
+; RUN: opt -passes=vector-combine -S < %s | FileCheck %s
+
+define ptx_kernel void @shuffle_ptx_i64() {
----------------
mikaelholmen wrote:

For info: I posted an x86 reproducer in https://github.com/llvm/llvm-project/issues/157401#issuecomment-3265423695

https://github.com/llvm/llvm-project/pull/156779