Common.dll 0x0635C376 crash
-
So this is a new one, I’m getting a seemingly random crash at 0x0635C376 in Common.dll. I have no information about when or why it’d happen and it’s rare enough to be hard to track down. I’m also getting memory leaks (to the extent of having the server get killed by DSPM for high memory usage), so perhaps the two are related.
Has anyone run in that particular crash? I couldn’t find anything about it with a cursory search.
-
As you’ve probably noticed, that code is a copy loop [c]p = p[i+1][/c] and it’s [c]p[i+1][/c] that is failing (or maybe just [c]p[/c]). The data points to three signed ints (used as multipliers) and two table offsets (into powers of two), used at [c]635D360[/c], eventually traced back to [c]PhySys::FindRayCollisions[/c]. Don’t know if that’s any help, though.
-
It’s the stack trace I couldn’t manage to get properly. I’ll have to think about what could cause ray collision errors…
-
Actually, that might be misleading, as that’s where the data itself is used, not where the crash is happening. Reloading a system ([c]s .[/c] using Console in SP) traced that eventually back to [c]PhySys::UnMakePhysicalR[/c] from [c]CObject::unmake_physical[/c] (and a couple more), and then to [c]PhySys::PhyCollisionStateManager::enable_collisions[/c], which is just [c]PhySys::SetCollisionState[/c]. Well, maybe it is something to do with collisions, then.
-
The thing that surprises me is that this is a server crash, but I can’t think off the top of my head what collisions the server has to deal with. NPCs ignore all collisions so that shouldn’t be it. Players send their own collision events to the server. It’d have to be missiles or beams somehow using the collision infrastructure, I guess?
-
The server calls [c]PhySys::Update[/c], which eventually calls the function in question. Prior to the update, there’s calls to [c]CBeam::move[/c]; I also traced a usage of the pointer (at [c]edx+0x1C[/c]) back to [c]PhySys::BeamR[/c], coming across an [c]excessive amount of children[/c] warning and there’s also [c]Unknown anim_technique in PhySys::_BeamR_Children()[/c]. Still don’t know if it helps.
-
Sounds like it’d be related to beams somehow then. I’ll keep that in mind, thanks a lot man.
