> 
> 
> Actually, there is an optimization in the pmap module which causes it to
> lazily allocate the page table, such that the extra 248MB will never even
> exist until it's touched.  Do you have some indication that it's not
> working in this case?
> 

I don't see the context for this but I'll take this oportunity to explain
some other (maybe related) VM weirdness...

I'm seeing processes "freezing" when they are growing in VM size. I can't
really get detailed "ps axl" info because system events (syscalls?) seem
to thaw them. I see this especially when running X and rebuilding with
g++ or nroff. I'm watching what is happening on a terminal running top.
I see the process groff growing and then sleeping. Any other attempts to
find out what is the sleep channel result in the process thawing.

If I can get this to happen without X running, I could trap into DDB and
look around. Maybe I'll try this later. This is definitely a new behavior
since updating my sources recently.

=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
 John Brezak                    UUCP:     uunet!apollo.hp!brezak
 Hewlett Packard/Apollo         Internet: brezak@ch.hp.com
 300 Apollo Drive               Phone:    (508) 436-4915
 Chelmsford, Massachusetts      Fax:      (508) 436-5103


------------------------------------------------------------------------------