[llvm] [AMDGPU] Prefer lower total register usage in regions with spilling (PR #71882)

[Valery Pykhtin via llvm-commits]

================
@@ -89,7 +90,8 @@ void GCNRegPressure::inc(unsigned Reg,
 }
 
 bool GCNRegPressure::less(const GCNSubtarget &ST,
-                          const GCNRegPressure& O,
+                          const SIMachineFunctionInfo &MFI,
----------------
vpykhtin wrote:

Maybe it worth to pass MachineFunction pointer and get ST and MFI from it?

https://github.com/llvm/llvm-project/pull/71882