Reword this comment to reference a more fundamental issue.
authorDan Gohman <gohman@apple.com>
Mon, 11 Jan 2010 17:14:46 +0000 (17:14 +0000)
committerDan Gohman <gohman@apple.com>
Mon, 11 Jan 2010 17:14:46 +0000 (17:14 +0000)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@93154 91177308-0d34-0410-b5e6-96231b3b80d8

lib/Target/X86/X86FastISel.cpp

index 10fba2461edf122b10423cb36572bec948ba86ed..7e02d59c1bca72a359d07b214bdc7860a42022e6 100644 (file)
@@ -1230,8 +1230,8 @@ bool X86FastISel::X86SelectCall(Instruction *I) {
       CC != CallingConv::X86_FastCall)
     return false;
 
-  // On X86, -tailcallopt changes the fastcc ABI. FastISel doesn't
-  // handle this for now.
+  // fastcc with -tailcallopt is intended to provide a guaranteed
+  // tail call optimization. Fastisel doesn't know how to do that.
   if (CC == CallingConv::Fast && PerformTailCallOpt)
     return false;