Free kernel memory

[Michael ODonnell]

Detecting leaks by observing total system memory usage seems
backwards; wouldn't it be more direct to track the various VM
statistics for the suspect processes?  Maybe something like:

  watch "ps -eaflw | fgrep -e SZ -e PIDofInterest -e otherPIDofInterest -e yetAnotherPID"

...and then just watch to see who bloats first?

This topic has been covered here in the past so you may find joy by
scanning the archives, and you're guaranteed to get a diaper-load of
info by yahooing or dogpiling.  There are also a number of payware
and freeware libraries with instrumented versions of malloc()/free()
to help you find badness in your code.  valgrind is quite useful, too.

But, of course, the most important question we need to ask you is this:

   WTF has your stupid memory leak investigation got to do with
   the vitally important topic of textbooks and why have you
   interrupted the most relevant and fascinating discussion
   to grace this channel in months with such annoying drivel?

Sheesh!