[llvm] [VectorCombine] Fold scalar selects from bitcast into vector select (PR #173990)

[Matt Arsenault via llvm-commits]

================
@@ -0,0 +1,1372 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 5
+; RUN: opt -mtriple=amdgcn-amd-amdhsa -mcpu=gfx942 -passes=vector-combine -S < %s | llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx942 | FileCheck %s --check-prefixes=CHECK,CHECK-OPT
----------------
arsenm wrote:

These tests don't belong in codgen, they shuold be IR to IR and in test/Transforms/VectorCombine

https://github.com/llvm/llvm-project/pull/173990