SparseBitVector is SLOW.

Use a Bitvector instead, we didn't need the smaller memory footprint anyway.
This makes the greedy register allocator 10% faster.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@129390 91177308-0d34-0410-b5e6-96231b3b80d8
diff --git a/lib/CodeGen/SplitKit.h b/lib/CodeGen/SplitKit.h
index a387896..688e4a5 100644
--- a/lib/CodeGen/SplitKit.h
+++ b/lib/CodeGen/SplitKit.h
@@ -144,6 +144,9 @@
   /// isThroughBlock - Return true if CurLI is live through MBB without uses.
   bool isThroughBlock(unsigned MBB) const { return ThroughBlocks.test(MBB); }
 
+  /// getThroughBlocks - Return the set of through blocks.
+  const BitVector &getThroughBlocks() const { return ThroughBlocks; }
+
   typedef SmallPtrSet<const MachineBasicBlock*, 16> BlockPtrSet;
 
   /// getMultiUseBlocks - Add basic blocks to Blocks that may benefit from