I am trying to debug a rather large application on a NetBSD/i386 system
(A P66 with 64M of ram and 128M of swap). The application is 2M of text
plus 10M of shared library [don't you love c++], all of them compiled
with -g.

I remove all process limits, and start gdb; five minutes later the
machine freezes. The last thing I see as I am running top is that gdb
is sleeping on "thrd_block" while it has consumed ~70Mbytes of memory.
The machine is still up (it is ping-able), but no processes seem to be
able to run. I'll have to go to the office to go into kgdb to see what
is actually going on...

[I tried with both gdb-4.11 and gdb-4.14 but that does not make any
 difference; both seem to swallow memory like there is no tomorrow]

I have been able to debug the same program on a sparc/ELC running 
SunOS4.1.3_U1B with only 48M of ram and 200 of swap.

Is that a vm inefficiency, bug or expected behaviour?

christos