Disable tail duplication of call instructions. The cost

metric is way off for these in general, and this works around
buggy code like that in PR1764.  we'll see if there is a big
performance impact of this.  If so, I'll revert it tomorrow.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43668 91177308-0d34-0410-b5e6-96231b3b80d8
This commit is contained in:
Chris Lattner 2007-11-04 06:37:55 +00:00
parent e719831b1c
commit 0647ebf1dc

View File

@ -115,6 +115,11 @@ bool TailDup::shouldEliminateUnconditionalBranch(TerminatorInst *TI) {
for (unsigned Size = 0; I != Dest->end(); ++I) {
if (Size == Threshold) return false; // The block is too large.
// Don't tail duplicate call instructions. They are very large compared to
// other instructions.
if (isa<CallInst>(I) || isa<InvokeInst>(I)) return false;
// Only count instructions that are not debugger intrinsics.
if (!isa<DbgInfoIntrinsic>(I)) ++Size;
}