| Tue, 12 Jan 2016 15:03:56 +0100 |
Doug Simon |
register JVMCIJDKFactory at module load time
|
| Mon, 11 Jan 2016 15:10:32 +0100 |
Stefan Anzinger |
Perform a clean in jvmci.make when main clean target is called as well
|
| Fri, 08 Jan 2016 23:45:00 +0100 |
Doug Simon |
converted select_task_blocking_aware into select_task_helper (GRAAL-1387)
|
| Fri, 08 Jan 2016 22:24:51 +0100 |
Doug Simon |
improved error message and fixed macro usage (GRAAL-1387)
|
| Fri, 08 Jan 2016 17:30:33 +0100 |
Doug Simon |
improved JVMCI support for blocking compilation (GRAAL-1387)
|
| Wed, 30 Dec 2015 17:55:07 +0100 |
Doug Simon |
removed @ServiceProvider mechanism (GRAAL-1380)
|
| Sat, 26 Dec 2015 15:00:55 +0100 |
Doug Simon |
must not fold accesses to @Stable fields if -XX:-FoldStableValues (GRAAL-58)
|
| Thu, 24 Dec 2015 09:50:27 -1000 |
twisti |
8143072: [JVMCI] Port JVMCI to AArch64
|
| Mon, 21 Dec 2015 15:53:56 +0100 |
Doug Simon |
removed @Option mechanism from JVMCI (GRAAL-1371)
|
| Sun, 20 Dec 2015 23:15:24 +0100 |
Doug Simon |
tightened signature of ConstantReflectionProvider methods for reading fields
|
| Fri, 18 Dec 2015 00:11:35 -0800 |
Tom Rodriguez |
Report all recompilations in the LogCompilation tool
|
| Fri, 18 Dec 2015 00:10:30 -0800 |
Tom Rodriguez |
Make node searches look through all the graphs for a match
|
| Wed, 16 Dec 2015 20:18:20 -0800 |
Tom Rodriguez |
OptionValues should be public or package
|
| Wed, 16 Dec 2015 16:08:05 -0800 |
Christian Wimmer |
Make method public
|
| Tue, 15 Dec 2015 10:55:55 -0800 |
Tom Rodriguez |
Make JVMCIJavaClasses look a little more like JDK9 version
|
| Tue, 15 Dec 2015 09:46:00 -0800 |
Tom Rodriguez |
Add pc to deoptimized LogCompilation message
|
| Tue, 15 Dec 2015 09:44:47 -0800 |
Tom Rodriguez |
Handle missing level in LogCompilation output
|
| Tue, 15 Dec 2015 09:03:09 -0800 |
Tom Rodriguez |
Group recompilations by method name only in LogCompilation tool
|
| Tue, 15 Dec 2015 09:02:42 -0800 |
Tom Rodriguez |
Allow importing multiple files at once in IGV
|
| Mon, 14 Dec 2015 12:38:57 -0800 |
Tom Rodriguez |
For previous commit: Add missing printing logic for oop arguments in dependencies
|
| Mon, 14 Dec 2015 12:02:18 -0800 |
Tom Rodriguez |
-s
|
| Tue, 08 Dec 2015 17:58:45 +0100 |
Doug Simon |
JVMCI compiler initialization can happen on different thread than JVMCI initialization
|
| Tue, 08 Dec 2015 01:03:01 +0100 |
Doug Simon |
use bool type for TraceClassLoadingStack option
|
| Fri, 04 Dec 2015 13:47:46 +0100 |
Doug Simon |
workaround for Eclipse bug 367599
|
| Thu, 26 Nov 2015 14:49:10 +0100 |
Doug Simon |
added JVMCI specific option TraceClassLoadingStack to see why classes are loaded
|
| Thu, 26 Nov 2015 02:28:38 +0100 |
Doug Simon |
bug fix for 8143730
|
| Thu, 26 Nov 2015 00:47:01 +0100 |
Doug Simon |
8144083 [JVMCI] CompilationResult should be finalized by JVMCI compiler and made effectively final
|
| Wed, 25 Nov 2015 20:41:26 +0100 |
Doug Simon |
8143730 [JVMCI] infopoint recording is too restrictive
|
| Wed, 25 Nov 2015 16:28:10 +0100 |
Stefan Anzinger |
Finalize data section outside of HotSpotCompiledCode
|
| Wed, 25 Nov 2015 11:48:44 +0100 |
Stefan Anzinger |
Remove CompilationResult.reset()
|
| Fri, 13 Nov 2015 09:31:04 -0800 |
Tom Rodriguez |
Fix assertion failures with G1 barrier logic and guard InstalledCode updates
|
| Fri, 13 Nov 2015 14:09:57 +0100 |
Doug Simon |
eager initialization of jvmciJavaClasses must happen while thread_in_VM
|
| Thu, 12 Nov 2015 21:35:35 +0100 |
Doug Simon |
8142511: must eagerly initialize classes with static fields accessed by JVMCI native code
|
| Thu, 12 Nov 2015 20:03:43 +0100 |
Doug Simon |
must use JVMCI class loader when trying to throw a JVMCIError from native VM code
|
| Thu, 12 Nov 2015 12:12:38 +0100 |
Paul Woegerer |
In 'def run_java()' add missing forwarding of timeout argument
|
| Wed, 11 Nov 2015 18:34:15 +0100 |
Gilles Duboscq |
Fix deopt blob for x86_32
|
| Wed, 11 Nov 2015 10:33:08 +0100 |
Roland Schatz |
Backport Handle fixes from jvmci-9.
|
| Tue, 10 Nov 2015 23:16:06 +0100 |
Doug Simon |
made JVMCI providers available as a normal services to avoid need for recompiling Graal jars to deploy them on JDK9
|
| Mon, 09 Nov 2015 14:34:33 +0100 |
Roland Schatz |
Pass Handle per value.
|
| Fri, 06 Nov 2015 12:48:05 +0100 |
Gilles Duboscq |
Fix hotspot build task: needsBuild should check ProjectBuildTask.needsBuild
|
| Fri, 06 Nov 2015 12:39:06 +0100 |
Gilles Duboscq |
Add deoptimizationUnrollBlockUnpackKindOffset to HotSpotVMConfig
|
| Thu, 05 Nov 2015 15:21:30 -0800 |
Tom Rodriguez |
Use proper GC barriers when clearing _jvmci_installed_code
|
| Thu, 05 Nov 2015 15:18:36 -0800 |
Tom Rodriguez |
Add logic to verify HeapRegionRemSet when running G1
|
| Thu, 05 Nov 2015 15:17:36 -0800 |
Tom Rodriguez |
Sync G1 write_ref_nmethod_post/pre change from JDK9
|
| Thu, 05 Nov 2015 10:59:29 +0100 |
Roland Schatz |
Make kindToBasicType trapping.
|
| Tue, 03 Nov 2015 16:55:51 +0100 |
Roland Schatz |
8139589: [JVMCI] Using fthrow when throwing JVMCIError.
|
| Wed, 04 Nov 2015 16:11:58 +0100 |
Gilles Duboscq |
Add assert for Unpack_uncommon_trap for ppc and always check for rethrow_exception
|
| Mon, 02 Nov 2015 23:53:10 +0100 |
Doug Simon |
fixed logic deciding who frees the CompileTask for a JVMCI blocking compilation
|
| Mon, 02 Nov 2015 11:18:51 -0800 |
Tom Rodriguez |
Restore missing InstalledCode.version increment
|
| Mon, 02 Nov 2015 15:33:53 +0100 |
Gilles Duboscq |
Fix ppc and x86_32 after fetch_unroll_info_helper changes
|
| Fri, 30 Oct 2015 12:28:17 +0100 |
Gilles Duboscq |
Use mx.ensure_dir_exists rather than os.makedirs to avoid races
|
| Mon, 02 Nov 2015 11:24:40 +0100 |
Stefan Anzinger |
8139589: Remove left over call to pd_patch_DataSectionReference
|
| Sun, 01 Nov 2015 21:43:49 +0100 |
Doug Simon |
use timeout for JVMCI blocking compilation to avoid deadlock between compiler and app thread that submitted the blocking compilation
|
| Thu, 29 Oct 2015 10:24:08 -0700 |
Tom Rodriguez |
Remove setAddressAndEntryPoint
|
| Thu, 29 Oct 2015 16:11:45 +0100 |
Gilles Duboscq |
Let fetch_unroll_info override the exec_mode and handle rethrowing scopes
|
| Thu, 29 Oct 2015 15:21:55 +0100 |
Roland Schatz |
8139589: [JVMCI] throw exceptions in faulty code installation operations
|
| Thu, 29 Oct 2015 13:13:46 +0100 |
Stefan Anzinger |
Change access to Method::_intrinsic_id back to u1 (JDK8 and JDK9 are different on this)
|
| Wed, 28 Oct 2015 14:02:20 -0700 |
Tom Rodriguez |
Fix missing comma and add STATIC_ASSERT to check size of trap name talbe
|
| Tue, 27 Oct 2015 13:33:28 +0100 |
Doug Simon |
method intrinsic ids are 16-bit so use Unsafe.getChar(), not Unsafe.getByte() to read them
|
| Sun, 25 Oct 2015 10:49:26 +0100 |
Doug Simon |
mx: exclude JVMCI jars from unit test class path
|
