Title: [124667] trunk/Source/_javascript_Core
- Revision
- 124667
- Author
- [email protected]
- Date
- 2012-08-03 15:42:29 -0700 (Fri, 03 Aug 2012)
Log Message
DFG handling of get_by_id should always inject a ForceOSRExit node if there is no prediction
https://bugs.webkit.org/show_bug.cgi?id=93162
Reviewed by Mark Hahnenberg.
This simplifies the DFG IR by ensuring that all nodes that use value profiles will be preceded
by a ForceOSRExit if the value profile had no data.
* dfg/DFGByteCodeParser.cpp:
(JSC::DFG::ByteCodeParser::parseBlock):
Modified Paths
Diff
Modified: trunk/Source/_javascript_Core/ChangeLog (124666 => 124667)
--- trunk/Source/_javascript_Core/ChangeLog 2012-08-03 22:39:12 UTC (rev 124666)
+++ trunk/Source/_javascript_Core/ChangeLog 2012-08-03 22:42:29 UTC (rev 124667)
@@ -1,5 +1,18 @@
2012-08-03 Filip Pizlo <[email protected]>
+ DFG handling of get_by_id should always inject a ForceOSRExit node if there is no prediction
+ https://bugs.webkit.org/show_bug.cgi?id=93162
+
+ Reviewed by Mark Hahnenberg.
+
+ This simplifies the DFG IR by ensuring that all nodes that use value profiles will be preceded
+ by a ForceOSRExit if the value profile had no data.
+
+ * dfg/DFGByteCodeParser.cpp:
+ (JSC::DFG::ByteCodeParser::parseBlock):
+
+2012-08-03 Filip Pizlo <[email protected]>
+
DFG::StructureCheckHoistingPhase keeps a Node& around for too long
https://bugs.webkit.org/show_bug.cgi?id=93157
Modified: trunk/Source/_javascript_Core/dfg/DFGByteCodeParser.cpp (124666 => 124667)
--- trunk/Source/_javascript_Core/dfg/DFGByteCodeParser.cpp 2012-08-03 22:39:12 UTC (rev 124666)
+++ trunk/Source/_javascript_Core/dfg/DFGByteCodeParser.cpp 2012-08-03 22:42:29 UTC (rev 124667)
@@ -2235,7 +2235,7 @@
}
case op_get_by_id:
case op_get_by_id_out_of_line: {
- SpeculatedType prediction = getPredictionWithoutOSRExit();
+ SpeculatedType prediction = getPrediction();
NodeIndex base = get(currentInstruction[2].u.operand);
unsigned identifierNumber = m_inlineStackTop->m_identifierRemap[currentInstruction[3].u.operand];
_______________________________________________
webkit-changes mailing list
[email protected]
http://lists.webkit.org/mailman/listinfo/webkit-changes
