[llvm] [ARM] shouldFoldMaskToVariableShiftPair should be true for scalars up to the biggest legal type (PR #158070)

[David Green via llvm-commits]

================
@@ -0,0 +1,2754 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc -mtriple=thumbv7m-eabi  %s -o - | FileCheck %s --check-prefix V7M
+; RUN: llc -mtriple=armv7a-eabi  %s -o -   | FileCheck %s --check-prefix V7A
+; RUN: llc -mtriple=thumbv7a-eabi  %s -o -   | FileCheck %s --check-prefix V7A-T
+; RUN: llc -mtriple=armv6m-eabi  %s -o -   | FileCheck %s --check-prefix V6M
+
+; *Please* keep in sync with test/CodeGen/X86/extract-lowbits.ll
----------------
davemgreen wrote:

And this

https://github.com/llvm/llvm-project/pull/158070