| Tue, 08 Sep 2015 19:57:39 +0200 |
Roland Schatz |
Update jvmci import: Refactoring: Rename Kind to JavaKind.
|
| Thu, 03 Sep 2015 00:54:14 +0200 |
Doug Simon |
suppressed -Xlint:try warnings
|
| Mon, 10 Aug 2015 15:57:05 +0200 |
Lukas Stadler |
do not use the (possibly uninitialized) InvokeNode.stateAfter to fill BytecodePosition in PEGraphDecoder
|
| Mon, 10 Aug 2015 15:57:05 +0200 |
Lukas Stadler |
do not replace "non-replaceable" FrameStates with null in PE graph decoding
|
| Sun, 02 Aug 2015 22:55:21 -0700 |
Tom Rodriguez |
ValueNode.getKind is really getStackKind
|
| Mon, 27 Jul 2015 16:26:41 -0700 |
Tom Rodriguez |
Eliminate jdk.internal.jvmci.debug
|
| Wed, 22 Jul 2015 23:27:39 -0700 |
Tom Rodriguez |
Move most of jdk.internal.jvmci.debug back into com.oracle.graal.debug
|
| Mon, 06 Jul 2015 12:46:41 -0700 |
Tom Rodriguez |
Inlining implicitly casts the receiver to the declared type of the method
|
| Fri, 26 Jun 2015 18:02:40 +0200 |
Thomas Wuerthinger |
Fix an issue in the inlining phase in the context of deleted unwind nodes. Reported by: Christian Thalinger.
|
| Tue, 23 Jun 2015 12:00:17 +0200 |
Gilles Duboscq |
Move SpeculationLog into the structured graph
|
| Mon, 22 Jun 2015 15:06:38 +0200 |
Gilles Duboscq |
Forward hasUnsafeAccess bit during inlining
|
| Mon, 22 Jun 2015 12:12:53 -0700 |
twisti |
Rename com.oracle.jvmci to jdk.internal.jvmci
|
| Sat, 20 Jun 2015 01:29:17 +0200 |
Thomas Wuerthinger |
Remove type profile proxy node.
|
| Wed, 03 Jun 2015 02:40:53 +0200 |
Thomas Wuerthinger |
Infer stamp in newly created return value phi in the inlining utility.
|
| Tue, 02 Jun 2015 11:45:43 +0200 |
Thomas Wuerthinger |
Remove GuardingPiNode and replace with FixedGuardNode + PiNode.
|
| Thu, 28 May 2015 17:00:59 -0700 |
Christian Wimmer |
Do not rely on single/double slot information from the Kind of the value in a local variable or expression stack entry. With word type rewriting during parsing, long values can be single slot values when they are rewritten from Object values.
|
| Wed, 27 May 2015 00:36:16 +0200 |
Doug Simon |
renamed com.oracle.graal.api[meta|code] modules to com.oracle.jvmci.[meta|code] (JBS:GRAAL-53)
|
| Tue, 26 May 2015 23:21:15 +0200 |
Doug Simon |
renamed com.oracle.graal.[debug|options|hotspotvmconfig]* modules to com.oracle.jvmci.[debug|options|hotspotvmconfig]* modules (JBS:GRAAL-53)
|
| Mon, 25 May 2015 23:30:34 +0200 |
Doug Simon |
moved GraalInternalError to com.oracle.jvmci.common and renamed it to JVMCIError (JBS:GRAAL-53)
|
| Mon, 25 May 2015 01:37:31 +0200 |
Thomas Wuerthinger |
Make inlining code more robust in the context of inlining substituted graphs.
|
| Sun, 24 May 2015 00:21:20 +0200 |
Doug Simon |
compute compiled bytecodes using method inlining recording instead of a DebugMetric
|
| Mon, 11 May 2015 17:56:16 -0700 |
Tom Rodriguez |
Merge
|
| Mon, 11 May 2015 10:29:42 -0700 |
Tom Rodriguez |
Improve accuracy of info point states
|
| Mon, 11 May 2015 17:12:15 +0200 |
Doug Simon |
consolidate frame state creation and processing for intrinsics into graph parsing, removing need for CollapseFrameForSingleSideEffectPhase
|
| Tue, 28 Apr 2015 17:08:47 -0700 |
iveresov |
Make G1 and Graal be friends in JDK9
|
| Mon, 27 Apr 2015 12:20:45 -0700 |
Tom Rodriguez |
Fix construction of infopoint frame states
|
| Tue, 14 Apr 2015 14:57:41 +0200 |
Doug Simon |
expand API for retrieving method substitution graphs to indicate the BCI of the invoke being inlined or -1 if the request is not in the context of inlining
|
| Sat, 11 Apr 2015 00:16:29 -0700 |
Christian Wimmer |
Improve speed of Graph partial evaluation
|
| Wed, 08 Apr 2015 22:00:47 -0700 |
Christian Wimmer |
Refactoring to allow re-use of inlining methods during a partial evaluation
|
| Sat, 04 Apr 2015 21:57:19 +0200 |
Doug Simon |
renamed getMethodSubstitution to getSubstitution in Replacements API and removed getMethodSubstitutionMethod
|
| Sat, 04 Apr 2015 19:47:04 +0200 |
Doug Simon |
added support for inlining graphs derived from InvocationPlugins
|
| Thu, 02 Apr 2015 17:32:26 +0200 |
Roland Schatz |
Don't drop entry framestate in inline graph if it has more usages than just the start node.
|
| Fri, 27 Mar 2015 16:04:23 +0100 |
Doug Simon |
removed MacroSubstitution and its related machinery
|
| Fri, 27 Mar 2015 13:29:08 +0100 |
Doug Simon |
converted all @MacroSubstitution uses to InvocationPlugins
|
| Sun, 22 Mar 2015 17:04:03 +0100 |
Doug Simon |
made partial intrinsification work (i.e., intrinsics that call the original method to handle slow/uncommon paths)
|
| Wed, 18 Mar 2015 12:39:43 +0100 |
Doug Simon |
use cheaper test for whether or not a method has a substitution
|
| Sat, 14 Mar 2015 20:42:09 +0100 |
Thomas Wuerthinger |
Turn a check for non-null stateAfter of an invoke into an assertion.
|
| Thu, 12 Mar 2015 15:19:56 +0100 |
Doug Simon |
tidied up assertions in InliningUtil
|
| Mon, 16 Feb 2015 17:47:43 +0100 |
Thomas Wuerthinger |
Add generic parameter to NodeClass. Change Graph#getNodes(Class) to Graph#getNodes(NodeClass).
|
| Fri, 13 Feb 2015 10:01:05 +0100 |
Doug Simon |
made it explicit that a StructuredGraph only records method dependencies for inlined methods - the root method is not recorded as it is already available in the 'method' field
|
| Thu, 12 Feb 2015 11:58:46 +0100 |
Doug Simon |
copy method dependencies from inlinee to caller
|
| Thu, 12 Feb 2015 01:54:05 +0100 |
Doug Simon |
pulled method evolution dependencies out of Assumptions and directly into StructuredGraph
|
| Wed, 11 Feb 2015 10:13:08 +0100 |
Doug Simon |
made Assumptions be part of a StructuredGraph
|
| Wed, 28 Jan 2015 01:06:39 +0100 |
Thomas Wuerthinger |
Make AbstractMergeNode abstract and introduce concrete subclass MergeNode.
|
| Wed, 28 Jan 2015 01:04:20 +0100 |
Thomas Wuerthinger |
Rename MergeNode to AbstractMergeNode.
|
| Wed, 28 Jan 2015 01:01:54 +0100 |
Thomas Wuerthinger |
Make AbstractBeginNode an abstract class and create concrete subclass BeginNode.
|
| Wed, 28 Jan 2015 00:50:31 +0100 |
Thomas Wuerthinger |
Rename BeginNode => AbstractBeginNode.
|
| Tue, 20 Jan 2015 16:52:16 +0100 |
Doug Simon |
added Fingerprint points to assist debugging replay compilation
|
| Mon, 12 Jan 2015 20:39:04 +0100 |
Doug Simon |
removed Node factory methods
|
| Sun, 11 Jan 2015 17:46:47 +0100 |
Thomas Wuerthinger |
Add utilities isDirect and isIndirect to InvokeKind.
|
| Mon, 24 Nov 2014 13:53:14 +0100 |
Roland Schatz |
Generalize NULL handling to work on arbitrary pointers.
|
| Mon, 24 Nov 2014 12:13:31 +0100 |
Roland Schatz |
Use asConstant() instead of asJavaConstant() where possible.
|
| Wed, 19 Nov 2014 11:36:48 +0100 |
Gilles Duboscq |
StampTool: add more methods to create object stamps to avoid using too many boolean arguments, add some javadoc, use them.
|
| Mon, 03 Nov 2014 10:17:24 -0800 |
Tom Rodriguez |
Correct the bci used by stateDuring in ForeignCallNode
|
| Thu, 30 Oct 2014 12:21:07 +0100 |
Roland Schatz |
Rename Constant to JavaConstant.
|
| Tue, 21 Oct 2014 16:22:38 +0200 |
Bernhard Urban |
InliningUtil: tell which graph in error message.
|
| Fri, 03 Oct 2014 14:19:58 +0200 |
Doug Simon |
Backed out of changeset 17322:655f3e6b467b
|
| Fri, 03 Oct 2014 14:19:31 +0200 |
Doug Simon |
made all Node data (i.e. non-edge) fields non-final and protected to support generating a copy "constructor" using Unsafe.allocateInstance
|
| Thu, 18 Sep 2014 17:41:19 +0200 |
Gilles Duboscq |
Stamp: interface types can not be trusted except after explicit runtime checks
|
| Mon, 18 Aug 2014 14:04:21 +0200 |
Doug Simon |
added factory method to all Node classes; replaced Node classes instantiation with calls to factory methods; replaced identity tests on Node classes with ' == <node class>.getGenClass()' idiom
|
