Propogate the Depth argument when calling
TLI.computeMaskedBitsForTargetNode from ComputeMaskedBits, since
the former may call back into the latter. This fixes a major
compile time problem on a testcase that happnened to hit this
in a particularly bad way, PR4643.

llvm-svn: 78023
diff --git a/llvm/lib/CodeGen/SelectionDAG/SelectionDAG.cpp b/llvm/lib/CodeGen/SelectionDAG/SelectionDAG.cpp
index 095d23a..a0e6054 100644
--- a/llvm/lib/CodeGen/SelectionDAG/SelectionDAG.cpp
+++ b/llvm/lib/CodeGen/SelectionDAG/SelectionDAG.cpp
@@ -1987,7 +1987,8 @@
   case ISD::INTRINSIC_WO_CHAIN:
   case ISD::INTRINSIC_W_CHAIN:
   case ISD::INTRINSIC_VOID:
-      TLI.computeMaskedBitsForTargetNode(Op, Mask, KnownZero, KnownOne, *this);
+      TLI.computeMaskedBitsForTargetNode(Op, Mask, KnownZero, KnownOne, *this,
+                                         Depth);
     }
     return;
   }