Reviewers: Mads Ager,

Description:
Fix incorrect deoptimization for logical not in an effect context.

The baseline compiler does not materialize a value for expressions of
the form !expr in an effect context so the graph translation should
not produce such an environment, otherwise we risk targeting it by
deoptimization.

Please review this at http://codereview.chromium.org/6537018/

SVN Base: https://v8.googlecode.com/svn/branches/bleeding_edge/build/ia32

Affected files:
  M src/hydrogen.cc
  A test/mjsunit/regress/regress-1167.js


Index: src/hydrogen.cc
diff --git a/src/hydrogen.cc b/src/hydrogen.cc
index 677d9a03c66e9de7a38ce083d2ac219e67fad344..fd5ad728ecb8fef338dda36e378251bfe68dd911 100644
--- a/src/hydrogen.cc
+++ b/src/hydrogen.cc
@@ -4654,7 +4654,7 @@ void HGraphBuilder::VisitUnaryOperation(UnaryOperation* expr) {
       VisitForControl(expr->expression(),
                       context->if_false(),
                       context->if_true());
-    } else {
+    } else if (ast_context()->IsValue()) {
       HSubgraph* true_graph = CreateEmptySubgraph();
       HSubgraph* false_graph = CreateEmptySubgraph();
       VISIT_FOR_CONTROL(expr->expression(),
@@ -4668,7 +4668,11 @@ void HGraphBuilder::VisitUnaryOperation(UnaryOperation* expr) {

       current_subgraph_->AppendJoin(true_graph, false_graph, expr);
       ast_context()->ReturnValue(Pop());
+    } else {
+      ASSERT(ast_context()->IsEffect());
+      VISIT_FOR_EFFECT(expr->expression());
     }
+
   } else if (op == Token::BIT_NOT || op == Token::SUB) {
     VISIT_FOR_VALUE(expr->expression());
     HValue* value = Pop();
Index: test/mjsunit/regress/regress-1167.js
diff --git a/test/mjsunit/regress/regress-1167.js b/test/mjsunit/regress/regress-1167.js
new file mode 100644
index 0000000000000000000000000000000000000000..ea37ba0c6337ec8ec8a4c347e62cc224881c25ec
--- /dev/null
+++ b/test/mjsunit/regress/regress-1167.js
@@ -0,0 +1,43 @@
+// Copyright 2011 the V8 project authors. All rights reserved.
+// Redistribution and use in source and binary forms, with or without
+// modification, are permitted provided that the following conditions are
+// met:
+//
+//     * Redistributions of source code must retain the above copyright
+//       notice, this list of conditions and the following disclaimer.
+//     * Redistributions in binary form must reproduce the above
+//       copyright notice, this list of conditions and the following
+//       disclaimer in the documentation and/or other materials provided
+//       with the distribution.
+//     * Neither the name of Google Inc. nor the names of its
+//       contributors may be used to endorse or promote products derived
+//       from this software without specific prior written permission.
+//
+// THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+// "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+// LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+// A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+// OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+// SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+// LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+// DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+// THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+// (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+// OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+
+// Deoptimization after a logical not in an effect context should not see a
+// value for the logical not expression.
+function test(n) {
+  var a = new Array(n);
+  for (var i = 0; i < n; ++i) {
+    // ~ of a non-numeric value is used to trigger deoptimization.
+    a[i] = void(!(delete 'object')) % ~(delete 4);
+  }
+}
+
+// OSR (after deoptimization) is used to observe the stack height mismatch.
+for (var i = 0; i < 5; ++i) {
+  for (var j = 1; j < 12; ++j) {
+    test(j * 1000);
+  }
+}


--
v8-dev mailing list
[email protected]
http://groups.google.com/group/v8-dev

Reply via email to