Fri, 17 May 2013 22:02:46 +0200 |
Doug Simon |
a leaf foreign call does not need debug info
|
Fri, 17 May 2013 22:02:11 +0200 |
Doug Simon |
changed location of debug scope so that a subclass that overrides 'addMethod' still matches -G:Dump=CodeInstall option
|
Fri, 17 May 2013 16:45:16 +0200 |
Doug Simon |
added test for AES intrinsification
|
Fri, 17 May 2013 17:43:01 +0200 |
Christian Haeubl |
Bugfix.
|
Fri, 17 May 2013 17:35:10 +0200 |
Christian Haeubl |
Minor refactoring.
|
Fri, 17 May 2013 17:34:48 +0200 |
Christian Haeubl |
Removed MethodProfile testcases.
|
Fri, 17 May 2013 17:24:03 +0200 |
Christian Haeubl |
Added class unloading support to the method profiling infrastructure.
|
Fri, 17 May 2013 15:40:06 +0200 |
Christian Haeubl |
Merge.
|
Fri, 17 May 2013 15:38:22 +0200 |
Christian Haeubl |
Added infrastructure for recording invoked methods in the profiling information.
|
Fri, 17 May 2013 09:28:41 +0200 |
Christian Haeubl |
Avoid inlining of graphs with many invokes.
|
Fri, 17 May 2013 13:55:35 +0200 |
Roland Schatz |
Use correct platform kind when loading a value.
|
Fri, 17 May 2013 13:55:24 +0200 |
Roland Schatz |
Method for determining supported vector lengths.
|
Fri, 17 May 2013 15:04:09 +0200 |
Roland Schatz |
Remove magic number in macro assembler.
|
Fri, 17 May 2013 14:29:44 +0200 |
Lukas Stadler |
small refactoring of early read elimination (move logic into "virtualize" methods)
|
Thu, 16 May 2013 17:03:18 -0700 |
Christian Wimmer |
Refactoring of NodeUtil to centralize the use of reflection
|
Thu, 16 May 2013 23:30:04 +0200 |
Doug Simon |
Merge.
|
Thu, 16 May 2013 23:26:59 +0200 |
Doug Simon |
replaced VMErrorStub with use of ForeignCallStub
|
Thu, 16 May 2013 22:30:52 +0200 |
Doug Simon |
replaced OSRMigrationEndStub with use of ForeignCallStub
|
Thu, 16 May 2013 21:46:15 +0200 |
Doug Simon |
removed redundant methods
|
Thu, 16 May 2013 21:34:03 +0200 |
Doug Simon |
replaced ThreadIsInterruptedStub with use of ForeignCallStub
|
Thu, 16 May 2013 21:30:39 +0200 |
Doug Simon |
replaced Log*Stub use of ForeignCallStub
|
Thu, 16 May 2013 21:26:10 +0200 |
Doug Simon |
replaced NewMultiArrayStub use of ForeignCallStub
|
Thu, 16 May 2013 20:26:46 +0200 |
Doug Simon |
removed CRuntimeCall and replaced its usages with ForeignCallNode
|
Thu, 16 May 2013 20:23:27 +0200 |
Doug Simon |
ForeignCallNode kills all memory locations
|
Thu, 16 May 2013 18:41:22 +0200 |
Doug Simon |
rename: *RuntimeCallOp -> *ForeignCallOp
|
Thu, 16 May 2013 18:12:04 +0200 |
Doug Simon |
rename: RuntimeCallStub -> ForeignCallStub
|
Thu, 16 May 2013 18:05:30 +0200 |
Doug Simon |
moved temporaries from CallingConvention to ForeignCallLinkage
|
Thu, 16 May 2013 17:36:16 +0200 |
Doug Simon |
removed redundant parameter
|
Thu, 16 May 2013 17:23:44 +0200 |
Doug Simon |
rename: HotSpotRuntimeCallTarget -> HotSpotForeignCallLinkage
|
Thu, 16 May 2013 17:12:17 +0200 |
Doug Simon |
rename: RuntimeCallTarget -> ForeignCallLinkage
|
Thu, 16 May 2013 16:41:37 +0200 |
Doug Simon |
renamed: RuntimeCallNode -> ForeignCallNode, RuntimeCallStateSplitNode -> ForeignCallStateSplitNode
|
Thu, 16 May 2013 16:32:46 +0200 |
Doug Simon |
moved responsibility for determining if a foreign call has a side effect to the runtime
|
Thu, 16 May 2013 16:17:28 +0200 |
Doug Simon |
refactored ForeignCallDescriptor into a top level class and moved it to the api.meta project
|
Thu, 16 May 2013 16:01:40 +0200 |
Doug Simon |
rename: Descriptor -> ForeignCallDescriptor
|
Thu, 16 May 2013 15:51:16 +0200 |
Doug Simon |
made RuntimeCallNode no longer be a StateSplit; for runtime calls that are state splits, the new RuntimeCallStateSplitNode class should be used
|
Wed, 15 May 2013 22:39:26 +0200 |
Doug Simon |
Merge.
|
Wed, 15 May 2013 21:51:12 +0200 |
Doug Simon |
made identityHashCode, javaTimeMillis and javaTimeNanos be CRuntimeCalls instead of RuntimeCallNodes
|
Wed, 15 May 2013 21:37:25 +0200 |
Doug Simon |
added support for declaring a foreign function call to be a leaf
|
Wed, 15 May 2013 16:46:39 +0200 |
Doug Simon |
spelling fix
|
Thu, 16 May 2013 18:15:19 +0200 |
Lukas Stadler |
typo in StructuredGraph
|
Thu, 16 May 2013 18:14:55 +0200 |
Lukas Stadler |
set bootstrapRunning flag earlier (in VMToCompiler.startCompiler)
|
Thu, 16 May 2013 18:12:06 +0200 |
Lukas Stadler |
allow only .isAlive() nodes to be added to a FrameStateBuilder
|
Thu, 16 May 2013 18:05:42 +0200 |
Lukas Stadler |
fix remaining currentGraph.add in GraphBuilderPhase
|
Thu, 16 May 2013 17:24:02 +0200 |
Lukas Stadler |
removed killed reads from list in memory aware scheduling
|
Thu, 16 May 2013 17:23:23 +0200 |
Lukas Stadler |
simplify GraphBuilderPhase: append adds the node to the graph
|
Thu, 16 May 2013 16:53:25 +0200 |
Lukas Stadler |
use only appendConstant for appending constants in GraphBuilderPhase
|
Thu, 16 May 2013 15:16:25 +0200 |
Christian Haeubl |
Bugfixes for the inlining phase and for -XX:+PrintInlining.
|
Wed, 15 May 2013 14:30:29 -0700 |
Christian Wimmer |
Handle corner case in WordTypeVerificationPhase
|
Wed, 15 May 2013 17:29:30 +0200 |
Lukas Stadler |
fix concurrency issue in NodeClass.get
|
Wed, 15 May 2013 14:56:52 +0200 |
Doug Simon |
added links to the HotSpot Windows sources
|
Wed, 15 May 2013 14:53:34 +0200 |
Doug Simon |
Merge.
|
Tue, 14 May 2013 22:02:23 +0200 |
Doug Simon |
consolidated the two RegisterConfig instances for HotSpotRuntime into one
|
Wed, 15 May 2013 10:43:26 +0200 |
Lukas Stadler |
add check for final-ness and visibility of @Input and @Successor fields to NodeClass
|
Wed, 15 May 2013 10:43:03 +0200 |
Lukas Stadler |
fix final-ness of @Input node inputs
|
Wed, 15 May 2013 09:03:43 +0200 |
Christian Haeubl |
Merge.
|
Wed, 15 May 2013 09:03:11 +0200 |
Christian Haeubl |
Disabled a test-case temporarily.
|
Tue, 14 May 2013 17:52:51 +0200 |
Christian Haeubl |
Adjusted test-case workaround.
|
Tue, 14 May 2013 16:57:07 +0200 |
Christian Haeubl |
Merge.
|
Tue, 14 May 2013 15:44:18 +0200 |
Christian Haeubl |
Windows-specific idiv/ldiv fix.
|
Tue, 14 May 2013 11:49:10 +0200 |
Christian Haeubl |
Fix for BoxingEliminationTest.
|
Tue, 14 May 2013 10:32:02 +0200 |
Christian Haeubl |
InliningPhase cleanups.
|
Tue, 14 May 2013 10:18:31 +0200 |
Christian Haeubl |
Merge.
|
Tue, 14 May 2013 10:17:06 +0200 |
Christian Haeubl |
Fixes for the merge and inlining cleanups.
|
Mon, 13 May 2013 17:43:42 +0200 |
Christian Haeubl |
Temporary fixes for the merge.
|
Mon, 13 May 2013 17:11:31 +0200 |
Christian Haeubl |
Merge.
|
Mon, 13 May 2013 16:46:39 +0200 |
Christian Haeubl |
Added some tunable parameters to the inlining phase.
|
Mon, 13 May 2013 13:55:41 +0200 |
Christian Haeubl |
Removed memory leak and bugfixes for the InliningPhase.
|
Mon, 13 May 2013 13:14:17 +0200 |
Christian Haeubl |
some refactorings and cleanups for the InliningPhase
|
Wed, 08 May 2013 17:21:38 +0200 |
Christian Haeubl |
Enabled nmethod statistics for the HotSpot product build.
|
Wed, 08 May 2013 16:31:59 +0200 |
Christian Haeubl |
Drafted version of an inlining policy that uses the callee graph size as its metric.
|
Fri, 26 Apr 2013 12:00:50 +0200 |
Christian Haeubl |
fixed documentation for FrameMap.outgoingSize
|
Tue, 14 May 2013 22:31:28 -0700 |
Thomas Wuerthinger |
Fix unused import.
|
Tue, 14 May 2013 22:10:13 -0700 |
Thomas Wuerthinger |
Remove unused read nodes.
|
Tue, 14 May 2013 23:18:00 -0400 |
Morris Meyer |
partial SPARC assembler
|
Tue, 14 May 2013 21:43:06 +0200 |
Doug Simon |
Merge.
|
Tue, 14 May 2013 21:33:37 +0200 |
Doug Simon |
cleaned up and simplified runtime call mechanisms
|
Mon, 13 May 2013 23:15:53 +0200 |
Doug Simon |
replaced WriteBarrierPreStub and WriteBarrierPostStub with use of RuntimeCallStub
|
Tue, 14 May 2013 11:27:09 -0700 |
twisti |
implement inlining support for JSR 292
|
Tue, 14 May 2013 18:40:15 +0200 |
Doug Simon |
distinguish floating and fixed nodes when building a stub graph
|
Tue, 14 May 2013 16:21:04 +0200 |
Lukas Stadler |
Merge
|
Tue, 14 May 2013 15:56:56 +0200 |
Lukas Stadler |
simplify NewObjectSnippets
|
Tue, 14 May 2013 14:48:38 +0200 |
Lukas Stadler |
don't overwrite default entries when lowering CommitAllocationNodes
|
Tue, 14 May 2013 14:47:16 +0200 |
Lukas Stadler |
PEA: virtualizing CommitAllocationNode and AllocatedObjectNode is not useful progress
|
Tue, 14 May 2013 15:45:55 +0200 |
Doug Simon |
Merge.
|
Tue, 14 May 2013 11:19:35 +0200 |
Doug Simon |
added check that the graph inlined for an invoke with a null stateAfter contains only frame states whose bci is AFTER_BCI or INVALID_FRAMESTATE_BCI
|
Tue, 14 May 2013 11:18:14 +0200 |
Doug Simon |
fixed printf statements
|
Tue, 14 May 2013 15:02:07 +0200 |
Roland Schatz |
Allow tests to check for deoptimizations, provide Method object in before hook.
|
Tue, 14 May 2013 14:11:04 +0200 |
Lukas Stadler |
don't allocate new variables for NullCheckOps
|
Tue, 14 May 2013 09:20:48 +0200 |
Roland Schatz |
Relax kind equality assertions in LIRGenerator.
|
Tue, 14 May 2013 09:20:40 +0200 |
Roland Schatz |
Do a full canonicalization at the end of MidTier.
|
Tue, 14 May 2013 09:19:39 +0200 |
Roland Schatz |
Rename asAddress and make it protected.
|
Tue, 14 May 2013 09:15:32 +0200 |
Roland Schatz |
Javadoc comments.
|
Mon, 13 May 2013 22:57:50 +0200 |
Doug Simon |
Merge.
|
Mon, 13 May 2013 22:56:30 +0200 |
Doug Simon |
replaced MonitorEnterStub and MonitorExitStub with use of RuntimeCallStub
|
Mon, 13 May 2013 19:02:32 +0200 |
Doug Simon |
replaced CreateNullPointerExceptionStub and CreateOutOfBoundsExceptionStub with use of RuntimeCallStub
|
Mon, 13 May 2013 19:01:30 +0200 |
Doug Simon |
extended RuntimeCallStub to support calls returning an Object
|
Mon, 13 May 2013 19:00:24 +0200 |
Doug Simon |
allow inlining of invokes with a null stateAfter
|
Mon, 13 May 2013 14:54:30 +0200 |
Doug Simon |
used static import for methods in StubUtil
|
Mon, 13 May 2013 14:28:00 +0200 |
Doug Simon |
replaced RegisterFinalizerStub with use of RuntimeStubCall
|
Mon, 13 May 2013 21:59:17 +0200 |
Doug Simon |
renamed HotSpotCompilationResult to HotSpotCompiledCode and added subclasses HotSpotCompiledNmethod and HotSpotCompiledRuntimeStub
|
Mon, 13 May 2013 18:19:43 +0200 |
Doug Simon |
added subclasses for HotSpotInstalledCode: HotSpotNmethod and HotSpotRuntimeStub
|
Mon, 13 May 2013 16:09:49 +0200 |
Gilles Duboscq |
mx/commands.py: warn about arguments after '-version' when launching the vm
|
Mon, 13 May 2013 17:11:45 +0200 |
Bernhard Urban |
Value: simplify some `equals' and `hashcode' implementations even more
|
Mon, 13 May 2013 16:47:13 +0200 |
Bernhard Urban |
VerifyValueUsage: use method name and signature instead of `toString()'
|
Mon, 13 May 2013 16:37:57 +0200 |
Bernhard Urban |
VerifyValueUsage: move String creation to assert
|
Mon, 13 May 2013 15:55:41 +0200 |
Bernhard Urban |
CFGVerifier: verify post-dominator calculation
|
Mon, 13 May 2013 14:17:35 +0200 |
Doug Simon |
Merge.
|
Mon, 13 May 2013 13:59:34 +0200 |
Doug Simon |
better error message
|
Mon, 13 May 2013 13:57:49 +0200 |
Doug Simon |
javadoc fixes
|
Mon, 13 May 2013 13:57:24 +0200 |
Doug Simon |
create fake JavaMethod for non-snippet stubs so that -G:MethodFilter can be applied to them
|
Mon, 13 May 2013 13:51:59 +0200 |
Doug Simon |
changed CFGPrinterObserver to use JavaMethod instead of ResolvedJavaMethod for method scope
|
Mon, 13 May 2013 13:50:46 +0200 |
Doug Simon |
enabled creation of a HotSpotSignature from types (in addition to a string)
|
Mon, 13 May 2013 11:56:16 +0200 |
Doug Simon |
replaced IdentityHashCodeStub[Call] with use of RuntimeStubCall
|
Mon, 13 May 2013 11:44:36 +0200 |
Doug Simon |
added RuntimeCallStub for runtime calls that require no extra/special logic in the stub itself
|
Mon, 13 May 2013 11:21:24 +0200 |
Doug Simon |
RuntimeStub installation support is now more distinct from normal nmethod installation
|
Mon, 13 May 2013 11:00:56 +0200 |
Doug Simon |
replaced unnecessary call to Stub.getMethod() with Stub.toString()
|
Mon, 13 May 2013 11:00:03 +0200 |
Doug Simon |
update the start node of a StructuredGraph if it is replaced via replaceFixedWithFixed()
|
Mon, 13 May 2013 10:58:50 +0200 |
Doug Simon |
javadoc fix
|
Mon, 13 May 2013 10:58:32 +0200 |
Doug Simon |
prevent multiple method context objects for the same method from resulting in multiple IGV tree levels unless there the method actually inlines itself
|
Mon, 13 May 2013 10:50:22 +0200 |
Doug Simon |
rename parameter
|