[llvm] [AMDGPU] Correct bitshift legality transformation for small vectors (PR #140940)

[Shilei Tian via llvm-commits]

================
@@ -0,0 +1,24 @@
+; RUN: llc -global-isel -mtriple=amdgcn -mcpu=gfx90a -O0 -print-after=legalizer %s -o /dev/null 2>&1 | FileCheck %s
----------------
shiltian wrote:

Use full triple `amdgcn-amd-amdhsa`. Auto generate check lines pls. Also use `-stop-after`.

https://github.com/llvm/llvm-project/pull/140940