Title: [147290] trunk/Source/_javascript_Core
- Revision
- 147290
- Author
- [email protected]
- Date
- 2013-03-31 16:25:33 -0700 (Sun, 31 Mar 2013)
Log Message
Unreviewed, fix a comment. While thinking about TBAA for array accesses,
I realized that we have to be super careful about aliasing of typed arrays.
* dfg/DFGCSEPhase.cpp:
(JSC::DFG::CSEPhase::getByValLoadElimination):
Modified Paths
Diff
Modified: trunk/Source/_javascript_Core/ChangeLog (147289 => 147290)
--- trunk/Source/_javascript_Core/ChangeLog 2013-03-31 23:17:50 UTC (rev 147289)
+++ trunk/Source/_javascript_Core/ChangeLog 2013-03-31 23:25:33 UTC (rev 147290)
@@ -1,3 +1,11 @@
+2013-03-31 Filip Pizlo <[email protected]>
+
+ Unreviewed, fix a comment. While thinking about TBAA for array accesses,
+ I realized that we have to be super careful about aliasing of typed arrays.
+
+ * dfg/DFGCSEPhase.cpp:
+ (JSC::DFG::CSEPhase::getByValLoadElimination):
+
2013-03-30 Mark Hahnenberg <[email protected]>
Move Region into its own header
Modified: trunk/Source/_javascript_Core/dfg/DFGCSEPhase.cpp (147289 => 147290)
--- trunk/Source/_javascript_Core/dfg/DFGCSEPhase.cpp 2013-03-31 23:17:50 UTC (rev 147289)
+++ trunk/Source/_javascript_Core/dfg/DFGCSEPhase.cpp 2013-03-31 23:25:33 UTC (rev 147290)
@@ -379,6 +379,8 @@
// We must assume that the PutByVal will clobber the location we're getting from.
// FIXME: We can do better; if we know that the PutByVal is accessing an array of a
// different type than the GetByVal, then we know that they won't clobber each other.
+ // ... except of course for typed arrays, where all typed arrays clobber all other
+ // typed arrays! An Int32Array can alias a Float64Array for example, and so on.
return 0;
}
case PutStructure:
_______________________________________________
webkit-changes mailing list
[email protected]
https://lists.webkit.org/mailman/listinfo/webkit-changes
