J
jonwil
Guest
I have a program written in C++ and compiled using Visual C++ 2012 (yes I know its old but there are reasons we still use 2012 for this particular version of the program that I won't go into here). We have been getting crashdumps from some of our users where the crash point is somewhere in the Windows memory manager in ntdll.dll with the call stack featuring Windows memory manager stuff then MS CRT memory management bits then our own memory stuff then various bits of other code in our program (its not the same every time, its different) I have seen this happen on both allocate and free calls.
The best guess is that we have some sort of heap corruption issues somewhere. (already ruled out out-of-memory issues as the cause and enough people are having the same issue that its unlikely to be a hardware failure) Are there any tools or tricks or things we can use to help track down where the heap corruption is actually occurring? Some sort of heap validation setup we can have the users experiencing the crashes run that will pick up issues when they happen and help us find the cause?
Continue reading...
The best guess is that we have some sort of heap corruption issues somewhere. (already ruled out out-of-memory issues as the cause and enough people are having the same issue that its unlikely to be a hardware failure) Are there any tools or tricks or things we can use to help track down where the heap corruption is actually occurring? Some sort of heap validation setup we can have the users experiencing the crashes run that will pick up issues when they happen and help us find the cause?
Continue reading...