You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@harmony.apache.org by "Ivan Popov (JIRA)" <ji...@apache.org> on 2007/03/02 09:48:51 UTC
[jira] Commented: (HARMONY-3047) [drlvm] EUT jdtdebug suit passes
with JVMTI_ERROR_unknown error
[ https://issues.apache.org/jira/browse/HARMONY-3047?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12477227 ]
Ivan Popov commented on HARMONY-3047:
-------------------------------------
JDWP agent code is still sensitive to JVMTI errors. If agent cannot release raw monitor due to JVMTI error, this may cause deadlock in agent threads and lead to VM hang up. I've submitted HARMONY-3269 enhancement to change JDWP agent cleanup procedure in order to minimize such risk.
> [drlvm] EUT jdtdebug suit passes with JVMTI_ERROR_unknown error
> ---------------------------------------------------------------
>
> Key: HARMONY-3047
> URL: https://issues.apache.org/jira/browse/HARMONY-3047
> Project: Harmony
> Issue Type: Bug
> Components: DRLVM
> Environment: reproduced on Windows
> Reporter: Aleksey Ignatenko
> Attachments: java_hangs.dmp, java_jdtdebug_assert_gc_slide_mov_all_assert.dmp, java_jdtdebug_class_h_1700_assert_failed.dmp, jdwp_agent_threads.patch, jdwp_agent_threads.patch, readPacket.txt
>
>
> When runningn jdtdebug suilt from EUT Eclipse writes to console window the following error (I consider it equal to crash):
> FATAL ERROR occurred in native method: JDWP on raw monitor wait, jvmtiError=JVMTI_ERROR_unknown(70007)
> The stack trace of the 02B73788 java thread:
> JDWP exit error JVMTI_ERROR_unknown(70007): on raw monitor wait
> I succeded to see the crash in debugger, stack trace:
> kernel32.dll!77e55dea()
> kernel32.dll!77e55dea()
> rpcrt4.dll!77c71abe()
> rpcrt4.dll!77c71122()
> rpcrt4.dll!77c84664()
> rpcrt4.dll!77c848aa()
> rpcrt4.dll!77ce2125()
> harmonyvm.dll!std::_Tree<std::_Tmap_traits<String const *,Class *,std::less<String const *>,std::allocator<std::pair<String const * const,Class *> >,0> >::_Key(std::_Tree_nod<std::_Tmap_traits<String const *,Class *,std::less<String const *>,std::allocator<std::pair<String const * const,Class *> >,0> >::_Node * _Pnode=0x01252840) Line 142 + 0x12 C++
> harmonyvm.dll!std::_Tree<std::_Tmap_traits<String const *,Class *,std::less<String const *>,std::allocator<std::pair<String const * const,Class *> >,0> >::iterator::iterator(std::_Tree_nod<std::_Tmap_traits<String const *,Class *,std::less<String const *>,std::allocator<std::pair<String const * const,Class *> >,0> >::_Node * _Pnode=0x00000000) Line 313 C++
> 0cab2048()
> hythr.dll!apr_thread_mutex_trylock(apr_thread_mutex_t * mutex=0x5f29c8f4) Line 106 + 0x36 C
> harmonyvm.dll!interpreter_enabled() Line 48 + 0xf C++
> 00000009()
> > harmonyvm.dll!Class::prepare(Global_Env * env=0x71b2423a) Line 1535 + 0xf C++
> mswsock.dll!71b24206()
> harmonyvm.dll!get_vm_thread(HyThread * thr=0x4bbbdbb4) Line 119 + 0xf C++
> hythr.dll!tmn_self_macro() Line 96 + 0xa C
> b87ae8ec()
> Looks like the problem is somethere in classloader, but Eclipse reports JVMTI error.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.