I'm seeing crashes in both GMod (server) and TF2 (server & client) that point to vphysics. I don't know if anything has changed recently, but I know that these crashes are happening extremely frequently. I'm hoping someone has some insight or Valve is looking into what might be causing these issues.
The first crash I have is one I've been helping a user diagnose. His server consistently crashes with the same stack trace on constraint activation: http://dl.dropbox.com/u/759758/crash_srcds.exe_20111231043430_1.dmp http://dl.dropbox.com/u/759758/crash_srcds.exe_20111219090657_1.dmp The value on top of the stack is -1, which is the counter register from the previous call. It seems to have iterated an array and rather than exit it called the function that's sitting on top. I can't figure out where exactly this is, but I know it's always as a constraint is activated. The second crash is from when all the Windows users on our TF2 server crash. http://dl.dropbox.com/u/759758/crash_hl2.exe_20111231221200_1.mdmp I'm not as certain about this crash as it seems that memalloc returned an invalid pointer. This crash probably takes place in CreatePhysicsObject. The one interesting trait is that the pointer is "sold" perhaps a string could have overwritten the pointer? This third crash happens on our TF2 server, which also consistently gets the same stack trace. http://dl.dropbox.com/u/759758/crash_srcds.exe_20111231174805_1.dmp http://dl.dropbox.com/u/759758/crash_srcds.exe_20111228021207_1.dmp This one appears to be some IVP_Mindist_Solver. I'm not sure why this one is happening. It is however the same function on the stack every time. The crashes don't seem to be the same between servers, but they are happening consistently in the same function. It also happens to be that it's during the initialization of an object. _______________________________________________ To unsubscribe, edit your list preferences, or view the list archives, please visit: https://list.valvesoftware.com/cgi-bin/mailman/listinfo/hlds
