On Fri, 3 Apr 1998, J. Maynard Gelinas wrote:

>        2) I've noticed that the system becomes unstable after about 24
>           hours of uptime.  Basically, it starts segfaulting upon
>           launching any and every binary; a reboot fixes things up.  I
>           suspect a memory leak here...  BTW: this is true not only
>           for 1.3E, but also 1.3 proper.  Anyone else notice this?

There is definately some sort of kernel memory leak plaguing the sun3/3x
ports.  Anyone who's attempted to build a NetBSD distribution on a sun3
can probably tell you that it takes two attempts: the first attempt will
invariably exhaust a kmem_map and things will start coring.  Eventually
the machine will panic.  I suspect something in the fork() code is
allocating something that is not being freed.

-J