Sure, you might be observing a real problem, it’s just you haven’t actually shared much that would allow anyone else to debug it. Again, some of your claims don’t really make sense. For example, munmap by its own does not take any system-wide locks. It may in some circumstances trigger as a consequence system-wide effects if the system itself is under strong memory pressure, but just calling munmap is a routine thing for pretty much any long-running process.
So, instead of writing about your speculations as to where the problem is, please focus on your observations. This will be a much more productive way of providing data to developers.
As a side note, I’ve been observing a pretty high CPU use by kswapd, maybe the underlying reason is the same. Please see this post: Ubuntu 22 kswapd and general RAM issue - #12 by Toyoo