The last one is in current use. The others are lost.
And, yes, I'm sure it is coming from malloc because when I use
the __malloc_hook stuff to capture larger allocations, these mappings
do not happen. However, it hangs. I have to go back and add some
pthread syncronization primitives, too. Maybe that's all I need do?
Were you all having syncronization issues in glibc 2.3.2?