[llvm] [AMDGPU] Add flag to enable expensive trip counts for runtime unroll. (PR #171735)

[Matt Arsenault via llvm-commits]

================
@@ -124,6 +130,8 @@ void AMDGPUTTIImpl::getUnrollingPreferences(
   // We want to run unroll even for the loops which have been vectorized.
   UP.UnrollVectorizedLoop = true;
 
+  UP.AllowExpensiveTripCount = UnrollRuntimeExpensiveTripCount;
----------------
arsenm wrote:

Isn't there a generic flag that sets this in the unroll pass? The proliferation of target specific flags just to set an option in the backend implementation of some generic thing is not good 

https://github.com/llvm/llvm-project/pull/171735