Hi,
>>> Something fishy with osg::Barrier
>>>
>>>Attached to this mail is is a short program that stress-test
>>>osg::Barrier. At least in Win32 it fails almost immediately. If
>>>you allow some slack by introducing some sleeps in the loop, it
>>>seems to work a bit more often. (Not that that is a pretty solution
>>>though :-)
>>
>>
>> this is a pending issue. There seem to be 2 solutions but I did
>> not have time to decide which one to use.
ok, I played around with it and my current solution would be similar
to the Linux implementation. But before I commit it and change it
again, just the short question if somebody sees a loophole that
I might have missed.
inline
void WinThreadBarrierBase::enter(void)
{
WaitForSingleObject(_pMutex1, INFINITE);
++_uiNumWaiters;
if(_uiNumWaiters == _uiNumWaitFor)
{
ReleaseSemaphore(_pBarrierSema[_uiCurrentCond],
_uiNumWaitFor - 1,
NULL);
_uiNumWaiters = 0;
_uiCurrentCond = 1 - _uiCurrentCond;
ReleaseMutex(_pMutex1);
return;
}
else
{
HANDLE waitObjects[2] =
{
_pBarrierSema[_uiCurrentCond],
_pMutex1
};
ReleaseMutex(_pMutex1);
WaitForMultipleObjects(2,
waitObjects,
true,
INFINITE);
ReleaseMutex(_pMutex1);
}
}
regards,
gerrit
-------------------------------------------------------
SF.Net email is sponsored by: Discover Easy Linux Migration Strategies
from IBM. Find simple to follow Roadmaps, straightforward articles,
informative Webcasts and more! Get everything you need to get up to
speed, fast. http://ads.osdn.com/?ad_id=7477&alloc_id=16492&op=click
_______________________________________________
Opensg-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/opensg-users
