[llvm] r250059 - fix typos; NFC

[Sanjay Patel via llvm-commits]

Author: spatel
Date: Mon Oct 12 11:09:59 2015
New Revision: 250059

URL: http://llvm.org/viewvc/llvm-project?rev=250059&view=rev
Log:
fix typos; NFC

Modified:
    llvm/trunk/lib/Target/X86/X86ISelDAGToDAG.cpp

Modified: llvm/trunk/lib/Target/X86/X86ISelDAGToDAG.cpp
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/X86/X86ISelDAGToDAG.cpp?rev=250059&r1=250058&r2=250059&view=diff
==============================================================================
--- llvm/trunk/lib/Target/X86/X86ISelDAGToDAG.cpp (original)
+++ llvm/trunk/lib/Target/X86/X86ISelDAGToDAG.cpp Mon Oct 12 11:09:59 2015
@@ -1648,13 +1648,12 @@ bool X86DAGToDAGISel::SelectLEAAddr(SDVa
     Complexity++;
 
   // FIXME: We are artificially lowering the criteria to turn ADD %reg, $GA
-  // to a LEA. This is determined with some expermentation but is by no means
+  // to a LEA. This is determined with some experimentation but is by no means
   // optimal (especially for code size consideration). LEA is nice because of
   // its three-address nature. Tweak the cost function again when we can run
   // convertToThreeAddress() at register allocation time.
   if (AM.hasSymbolicDisplacement()) {
-    // For X86-64, we should always use lea to materialize RIP relative
-    // addresses.
+    // For X86-64, always use LEA to materialize RIP-relative addresses.
     if (Subtarget->is64Bit())
       Complexity = 4;
     else