harmony-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Gregory Shimansky (JIRA)" <j...@apache.org>
Subject [jira] Updated: (HARMONY-1490) [drlvm] Fixes for SingleStep implementation. Corrected phase, state structure, addded m2n transition
Date Mon, 18 Sep 2006 16:05:23 GMT
     [ http://issues.apache.org/jira/browse/HARMONY-1490?page=all ]

Gregory Shimansky updated HARMONY-1490:
---------------------------------------

    Attachment: test.zip

Simple test for single step. To build on windows use:

cl -I%JAVA_HOME%/include -I%JAVA_HOME%/include/win32 -LD -Feagent.dll SingleStep0101.cpp events.cpp
utils.cpp

on Linux

gcc -DLINUX -shared -o libagent.so -I$JAVA_HOME/include -I$JAVA_HOME/include/linux SingleStep0101.cpp
events.cpp utils.cpp

where JDK is patch to 1.5 JDK.

It looks like on Linux it crashes somewhere in bytecode predictor taken from HARMONY-1484.
I'll try to find out later what is wrong with it.

> [drlvm] Fixes for SingleStep implementation. Corrected phase, state structure, addded
m2n transition
> ----------------------------------------------------------------------------------------------------
>
>                 Key: HARMONY-1490
>                 URL: http://issues.apache.org/jira/browse/HARMONY-1490
>             Project: Harmony
>          Issue Type: Bug
>          Components: DRLVM
>            Reporter: Gregory Shimansky
>         Attachments: single-step-fixes.patch, test.zip
>
>
> I've created several bugfixes for SingleStep event implementation.
> It appears that keeping enabled flag in JVMTISingleStepState struct is not possible because
it may be deallocated when thread finishes. So I decided that enabled flag should be the pointer
to this struct itself.
> Another bugfix is that SingleStep shouldn't be sent in phases other than LIVE. This also
fixes the problem when SingleStep is sent for VM startup and agent cannot use any JNI function
yet.
> Another bugfix is for breakpoint. For JIT mode I've added an m2n transition to allow
stack iterator to work correctly inside of breakpoint callback. Otherwise it didn't see the
upper Java frames.
> On windows the test I am using still crashes when it finishes, but at least I can see
SingleStep events coming to the agent.

-- 
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: http://issues.apache.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira

        

Mime
View raw message