Hi all,
Looking into luni native code, I find its reflection caching is
disabled, I'd like to enable it to improve Harmony performance.
Native reflection caching is to cache reflection method/field id
of some classes which is used frequently in native, this mechanism
avoid reflect some id
Hi Mark:
I have long before thought about this. This is really a good idea.
I'd like to write more native tests, And plans to start from file related
functions (In case you have not write them).
Best regards
2007/1/18, Mark Hindess <[EMAIL PROTECTED]>:
I've checked in an initial attempt at cr
Sounds interesting, I'd like to take a look. I believe most of work
can be done after some refactor on native code :)
2007/1/23, Leo Li <[EMAIL PROTECTED]>:
Hi, all:
According to spec, "the Java virtual machine will make a best
effort to perform native I/O operations directly upon it. That
Hi, all:
According to spec, "the Java virtual machine will make a best
effort to perform native I/O operations directly upon it. That is, it
will attempt to avoid copying the buffer's content to (or from) an
intermediate buffer before (or after) each invocation of one of the
underlying operati
> On 1/22/07, Robin Garner <[EMAIL PROTECTED]> wrote:
>>
>> Couldn't you achieve exactly the same effect by using an annotation to
>> mark the called native method ?
>>
>> Such as
>>
>> @FastNative
>> public native fastNativeMethod(args...);
>>
>
> Robin,
> I think such an annotation is a good
Hi, Vladimir:
It is good that junit will kill the hanging testcase. I have
reported it as HARMONY-3036 and have given a patch. Hope it will work.
:)
Good luck!
Leo.
On 1/22/07, Vladimir Ivanov <[EMAIL PROTECTED]> wrote:
On 1/22/07, Leo Li <[EMAIL PROTECTED]> wrote:
>
> Hi, Vladimir:
>You
On Jan 22, 2007, at 6:33 AM, Tim Ellison wrote:
Yang Paulex wrote:
For windows, you can download or browse it on my personal site[1].
Or you
can create it by "ant doxygen-natives" and "ant doxygen-kernel" at
working copy>/enhanced/doc directory, supposing you have doxygen
installed.
[1]
On Jan 22, 2007, at 1:36 AM, Xiao-Feng Li wrote:
On 1/22/07, Geir Magnusson Jr. <[EMAIL PROTECTED]> wrote:
On Jan 19, 2007, at 1:17 AM, Xiao-Feng Li wrote:
> Hi, I saw many people here are from China. It's close to Chinese
> Spring Festival, do we want to have some China HUG event? We can
On 1/22/07, Robin Garner <[EMAIL PROTECTED]> wrote:
Couldn't you achieve exactly the same effect by using an annotation to
mark the called native method ?
Such as
@FastNative
public native fastNativeMethod(args...);
Robin,
I think such an annotation is a good idea for native methods wit
On 1/22/07, Pavel Afremov <[EMAIL PROTECTED]> wrote:
See my comments inline please.
On 1/22/07, Mikhail Fursov <[EMAIL PROTECTED]> wrote:
>
> Pavel,
> I see the direction of your improvement - optimization of JNI calls.
> The scenario you proposed is (correct me if I'm wrong)
It's more clos
Alexei Zakharov wrote:
> 2. However, DRLVM completely rejected to run after step 1. Even
> HelloWorld crashed. The problem was that the kernel I was using wasn't
> fresh enough - big thanks to Alexei Fedotov for pointing this out. So
> I've switched from version 2.4.27 to 2.6.8 and this brings DRLV
I'll take a look.
Thanks,
2007/1/22, Pavlenko, Andrey A <[EMAIL PROTECTED]>:
Dear committers,
Below is a list of the AWT issues which, as I think, should be closed.
Could somebody close the following issues, please?
http://issues.apache.org/jira/browse/HARMONY-2066 - non-bug diff. I
thin
HARMONY-2804 should be applied. It removes an assert from jitrino
debug.
Thanks,
Naveen
Begin forwarded message:
From: "Naveen Neelakantam (JIRA)" <[EMAIL PROTECTED]>
Date: January 19, 2007 11:04:30 AM CST
To: [EMAIL PROTECTED]
Subject: [jira] Commented: (HARMONY-2804) [drlvm][jit] StackTes
ATT. Thanks a lot!
On 1/18/07, Andrew Zhang (JIRA) <[EMAIL PROTECTED]> wrote:
[
https://issues.apache.org/jira/browse/HARMONY-3017?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12465648]
Andrew Zhang commented on HARMONY-3017:
Couldn't you achieve exactly the same effect by using an annotation to
mark the called native method ?
Such as
@FastNative
public native fastNativeMethod(args...);
then the JIT compiler emits the correct code without burdening the Java
code with details about how the fast native call linkage
Grr. It appears the code in
working_classlib\modules\portlib\src\main\native\thread\unix\x86_64\thrspinlock.s
is a complete bogus, it is copy-pasted from x86 almost verbatim and
never worked.
The tests PASS with drlvm's libhythr.so.
So we face again glorified hythr dilemma ;).
--
Alexey
2007/1/
Reproduced 100%. The problem is inside hythread_spinlock_acquire()
impl, it spins forever on free "Thread global" monitor; I'm trying to
grok asm impl for x64 now...
0 0x002a957a5cf8 in hythread_spinlock_acquire () from
./deploy/jdk/jre/bin/libhythr.so
#1 0x002a957a7cef in monitor_enter
[1] http://java.sun.com/javase/6/docs/technotes/tools/windows/java.html
On 1/22/07, Ivan Popov <[EMAIL PROTECTED]> wrote:
I've tested instrumented class against Sun JDK 1.5 and 1.6 with
-Xfuture option, which "turns on stricter class-file format checks
that enforce closer conformance to the clas
I've tested instrumented class against Sun JDK 1.5 and 1.6 with
-Xfuture option, which "turns on stricter class-file format checks
that enforce closer conformance to the class-file format
specification" [1].
The instrumented class rejected by DRLVM is still accepted by Sun JVM
launched with -Xfut
Hi, about
>Just wondering. I thought we were having a discussion about that
>stuff...
I double-checked the thread, seems like no more open navbar related
issues remain. The recent emails mostly concern restructuring of project
guidelines page. For example, the question where to place description
The native tests of portlib module are hangs up on the em64t Linux and test
run killed by timeout.
Could somebody reproduce/ fix it?
thanks, Vladimir
Execution log:
[exec] -o ../init -lc -lm -ldl \
[exec] -Xlinker -z -Xlinker origin \
[exec] -Xlinker -rpath -Xlinker \$ORIGIN/ \
On 1/22/07, Leo Li <[EMAIL PROTECTED]> wrote:
Hi, Vladimir:
You are right. The testcase seems to be unstable. The most
reliable solution, I think, is to wait in a loop till the sheduled
task does run. However, if there exists a bug in the tested class, the
testcase might trap in an endless lo
See my comments inline please.
On 1/22/07, Mikhail Fursov <[EMAIL PROTECTED]> wrote:
Pavel,
I see the direction of your improvement - optimization of JNI calls.
The scenario you proposed is (correct me if I'm wrong)
It's more close to "Replace LiL code by Magic".
Problems:
P1) GC in nat
The two attached patches provide workaround for the hanging Intel compiler
builds with cpptasks-1.0b4.
I wasn't able to find exact root cause of hangs, and I suspect that
commands 'icl -help', 'xilib -qv' and 'xilink -qv' try to paginate its output,
and thus wait indefinitely for user input. Unfor
Yang Paulex wrote:
> For windows, you can download or browse it on my personal site[1]. Or you
> can create it by "ant doxygen-natives" and "ant doxygen-kernel" at
> working copy>/enhanced/doc directory, supposing you have doxygen installed.
>
> [1] http://people.apache.org/~pyang/harmonydoc/wind
Pavel,
I see the direction of your improvement - optimization of JNI calls.
The scenario you proposed is (correct me if I'm wrong)
1) We have a native JNI call from JAVA
2) JIT replaces the call with another call -> call to stub written in Java
with a help of VMMagic package and passes original
2007/1/22, Stepan Mishura <[EMAIL PROTECTED]>:
On 1/19/07, Ruth Cao wrote:
>
> 2007/1/19, Stepan Mishura:
> >
> > On 1/19/07, Stepan Mishura wrote:
> > >
> > > On 1/19/07, Ruth Cao wrote:
> > > >
> > > > Hi all,
> > > >
> > > > When working on HARMONY-2910, we've found that current Harmony
code
Hello.
I have a proposal to improve performance of JITed code <-> VM communication
and fix some potential bugs. To reach it the Magic opportunity should be
extended. One of the new features is creating M2nFrame on the stack.
To support it, the following methods may be added to the Magic:
Dear committers,
Below is a list of the AWT issues which, as I think, should be closed.
Could somebody close the following issues, please?
http://issues.apache.org/jira/browse/HARMONY-2066 - non-bug diff. I
think it's RI bug.
http://issues.apache.org/jira/browse/HARMONY-2445 - non-bug
Hi!
Your task is very close to write barriers implementation in OPT.
I think that the best place today is IRBuilder.cpp that is called from
JavaByteCodeTranslator during HIR creation from Java bytecode.
See WB implementation as an example and add tracing code to object access
ops like genLdField..
30 matches
Mail list logo
