[llvm] [AMDGPU] Use S_CMP to lower a copy of a lane mask to SCC (PR #221445)

[Jay Foad via llvm-commits]

================
@@ -5134,6 +5134,67 @@ bool SIInstrInfo::hasVALU32BitEncoding(unsigned Opcode) const {
   return pseudoToMCOpcode(Op32) != -1;
 }
 
+/// Return true if \p MI is a VALU comparison, i.e. an instruction that writes
+/// a lane mask with one bit per lane, and zeroes the bits of lanes that were
+/// inactive when it executed.
+///
+/// TODO: Also handle the sdst result of V_ADD_CO_U32 and V_SUB_CO_U32 and
----------------
jayfoad wrote:

Yes that'll come with test updates that show the benefit.

https://github.com/llvm/llvm-project/pull/221445