HI Andreas, Could you explain how you can reproduce these deadlocks/bugs so that I might be able to reproduce them at my end and investigate.
Thanks, Robert. On Fri, Apr 23, 2010 at 3:51 PM, Andreas Malzahn <[email protected]> wrote: > Hi, > > the current osgTerrain code has some bugs and will lead to deadlocks. > Two bugs are fixed in the appended file, while for the deadlocks the solution > is not clear to me. > > One deadlock can be avoided by setting updateNeighboursImmediately = false - > but I don't think this is the right solution. If updateNeighboursImmediately > is true, the deadlock appears when the GeometryTechnique locks itself for > update, and then locks the neighbour - and the neighbour does the same in > another database thread. > > There is still one deadlock in the Terrain and TerrainTile/GeometryTechnique > - Terrain::traverse() locks first the terrain and then the > TerrainTile/GeometryTechnique while in GeometryTechnique::init() first the > GeometryTechnique is locked, and later in the function > Terrain::updateTerrainTileOnNextFrame() is called which gets the lock on > Terrain. > > I have no idea yet how to resolve this. > > Thank you! > > Cheers, > Andreas > > ------------------ > Read this topic online here: > http://forum.openscenegraph.org/viewtopic.php?p=27225#27225 > > > > > Attachments: > http://forum.openscenegraph.org//files/geometrytechnique_424.cpp > > > _______________________________________________ > osg-users mailing list > [email protected] > http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org > _______________________________________________ osg-users mailing list [email protected] http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
