Reviewers: Benedikt Meurer,

Description:
[turbofan] Remove hack for dead nodes from JSGenericLowering.

Now that the graph is being trimmed after generic lowering, we can drop
this workaround. The diamond will no longer confuse the scheduler.

[email protected]

Please review this at https://codereview.chromium.org/1191913002/

Base URL: https://chromium.googlesource.com/v8/v8.git@master

Affected files (+0, -3 lines):
  M src/compiler/js-generic-lowering.cc


Index: src/compiler/js-generic-lowering.cc
diff --git a/src/compiler/js-generic-lowering.cc b/src/compiler/js-generic-lowering.cc index 851be4923ce54f49041ebbe89983f5dc20a97644..60df598d7057fccb51140969f7255fc65cfdbf28 100644
--- a/src/compiler/js-generic-lowering.cc
+++ b/src/compiler/js-generic-lowering.cc
@@ -196,9 +196,6 @@ void JSGenericLowering::ReplaceWithCompareIC(Node* node, Token::Value token) {

   // Finally patch the original node to select a boolean.
   NodeProperties::ReplaceUses(node, node, compare, compare, compare);
-  // TODO(mstarzinger): Just a work-around because SelectLowering might
-  // otherwise introduce a Phi without any uses, making Scheduler unhappy.
-  if (node->UseCount() == 0) return;
   node->TrimInputCount(3);
   node->ReplaceInput(0, booleanize);
   node->ReplaceInput(1, true_value);


--
--
v8-dev mailing list
[email protected]
http://groups.google.com/group/v8-dev
--- You received this message because you are subscribed to the Google Groups "v8-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
For more options, visit https://groups.google.com/d/optout.

Reply via email to