Programming question: memory leaks

[javasaurus]

I'm under the impression that when a program suffers from memory leaks, the memory is freed up again once the program is closed. Is it possible for memory leaks to persist after closing the offending program?

Comments

[acroyear70.livejournal.com]

Yes.

The problem is not in the program, it's in the fact that the kernel is also allocating memory to handle it's end of the resources it manages on behalf of the user-mode program. Just like Unix kernels have always had (including Linux), the Windows kernel is what actually opens and reads files, manages USB connections, does the graphics work, etc, all through low-level system calls (in fact, it was having better access to those very calls that made office so efficient in the early days and is what got the DoJ's wrath even before Netscape showed up and was attacked by the bundled IE).

The Windows kernel is, to be blunt, HORRID at cleaning up its mess. the fact that there's no easy way to actually see how much memory the kernel is using makes it all the more difficult to pin it down.

Hence, the eventual slowdown and necessary reboots over time, even if you do manage to kill off all programs.

[acroyear70.livejournal.com]

it's not as much a problem in Linux because people actually paid attention to it, having access to the memory usage and access to the code to fix any found problems.

[javasaurus.livejournal.com]

Thanks for the explanation. At least I get to blame Microsoft instead of the company who created the program I'm using.