Mailing-List: contact commits-help@harmony.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@harmony.apache.org
Message-ID: <4936806.1179414256652.JavaMail.jira@brutus>
Date: Thu, 17 May 2007 08:04:16 -0700 (PDT)
From: "Ilya Leviev (JIRA)" <jira@apache.org>
To: commits@harmony.apache.org
Subject: [jira] Created: (HARMONY-3902) [drlvm][thread] Race condition at
 "thread_native_thin_monitor.c":78 and "thread_java_monitors.c":122
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit

[drlvm][thread] Race condition at "thread_native_thin_monitor.c":78 and "thread_java_monitors.c":122
----------------------------------------------------------------------------------------------------

                 Key: HARMONY-3902
                 URL: https://issues.apache.org/jira/browse/HARMONY-3902
             Project: Harmony
          Issue Type: Bug
          Components: DRLVM
            Reporter: Ilya Leviev


TC report on thread unsafe access that result in race condition that occur during concurrent execution of set_fat_lock_id and jthread_monitor_enter functions.


Write -> Read data-race	
Memory read at "thread_java_monitors.c":122 conflicts with a prior memory write at "thread_native_thin_monitor.c":78


Stack Trace: 


Context
	Function void vm_execute_java_method_array(struct _jmethodID *,union jvalue *,union jvalue *) "ini.cpp":60
	Function ExecuteMethod "em_intf.cpp":43
	Function void DrlEMImpl::executeMethod(struct _jmethodID *,union jvalue *,union jvalue *) "drlemimpl.cpp":509
	Function void JIT_execute_method_default(void *,struct _jmethodID *,union jvalue *,union jvalue *) "ini_ia32.cpp":199
	Function vm_invoke_native_array_stub "ini_ia32.cpp":76
	Function class VM_thread * get_thread_ptr_stub(void) "thread_manager.cpp":138
	Function m2n_free_local_handles "m2n_ia32.cpp":268
	Function hythread_thin_monitor_try_enter "thread_native_thin_monitor.c":341
	Function class VM_thread * get_thread_ptr_stub(void) "thread_manager.cpp":138
	Function jthread_monitor_enter "thread_java_monitors.c":68


1st Access
	Function m2n_free_local_handles "m2n_ia32.cpp":268
	Function hythread_self "hythread.h":465
	Function class VM_thread * get_thread_ptr_stub(void) "thread_manager.cpp":138
	Function m2n_free_local_handles "m2n_ia32.cpp":268
	Function class VM_thread * get_thread_ptr_stub(void) "thread_manager.cpp":138
	Function hythread_suspend_enable "hythread_ext.h":370
	Function Java_java_lang_VMThreadManager_wait "java_lang_vmthreadmanager.cpp":202
	Function jthread_monitor_timed_wait "thread_java_monitors.c":308
	Function inflate_lock "thread_native_thin_monitor.c":600
	Function set_fat_lock_id "thread_native_thin_monitor.c":78

"76"	""	"     lockword&=0x7FF;"
"77"	""	"     lockword|=(monitor_id << 11) | 0x80000000;"
"78"	"*"	"     *lockword_ptr=lockword;"
"79"	""	"     apr_memory_rw_barrier();"
"80"	""	" }"


2nd Access
	Function void vm_execute_java_method_array(struct _jmethodID *,union jvalue *,union jvalue *) "ini.cpp":60
	Function ExecuteMethod "em_intf.cpp":43
	Function void DrlEMImpl::executeMethod(struct _jmethodID *,union jvalue *,union jvalue *) "drlemimpl.cpp":509
	Function void JIT_execute_method_default(void *,struct _jmethodID *,union jvalue *,union jvalue *) "ini_ia32.cpp":199
	Function vm_invoke_native_array_stub "ini_ia32.cpp":76
	Function class VM_thread * get_thread_ptr_stub(void) "thread_manager.cpp":138
	Function m2n_free_local_handles "m2n_ia32.cpp":268
	Function hythread_thin_monitor_try_enter "thread_native_thin_monitor.c":341
	Function class VM_thread * get_thread_ptr_stub(void) "thread_manager.cpp":138
	Function jthread_monitor_enter "thread_java_monitors.c":122


"120"	""	"         lockword = vm_object_get_lockword_addr(monitor);"
"121"	""	"  "
"122"	"*"	"         if (is_fat_lock(*lockword)) {"
"123"	""	"             status = hythread_thin_monitor_enter(lockword);"
"124"	""	"              if (status != TM_ERROR_NONE) {"


See also Source View screenshot.


Notes on Write->Read race condition.
------------------------------------
Write->Read data races occur when one thread writes a shared memory location (address) while another thread concurrently reads the same memory location.  
The shared memory location may be referred to by (variable) name, pointer, or even a function such as memcpy().  

The following example uses a variable name:

1st access by first thread

S1: sharedX = privateA

2nd access by second thread

S2: privateB = sharedX

If sharedX is a variable visible to all threads and privateA and privateB are local variables visible only to the thread where each was declared,
 concurrent execution of the above statements by multiple threads results in a "race" on the value to be read from sharedX.   

Since the order of execution among threads is unpredictable, it is unknown which value will be available in sharedX to be stored into privateB.  
This results in non-deterministic software, or software prone to produce different results each time it is executed. 
_______________________________________________________________________________________________________________________


If it not affect correctness of execution I will mark it by special API for prevention of further alarms on this race.


-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.