[llvm] [AMDGPU][True16][CodeGen] fp conversion in true/fake16 format (PR #101678)

[Joe Nash via llvm-commits]

================
@@ -1447,9 +1456,7 @@ def : GCNPat <
   (i16 (trunc i64:$src)),
   (EXTRACT_SUBREG $src, lo16)
 >;
-
----------------
Sisyph wrote:

Remove whitespace change

https://github.com/llvm/llvm-project/pull/101678