If the machine has been running fine with no trouble and then began to
display this bad behavior without a change to its software load, then it's
most likely a hardware problem.  Your guess about memory is probably pretty
good.  
Before I tried changing stuff, I'd try reseating the memory sticks in the
mother board.  I just did this for one of our production machines, and it
did wonders.
-Brian

On Mar 17,  3:41pm, Steve Blinkhorn wrote:
} Subject: Re: kernel page fault trap
} > 
} > Steve Blinkhorn <steve@prd.co.uk> writes:
} > 
} > > Can someone interpret for me please.  One of my machines, running 2.0
} > > GENERIC, has started to misbehave.   Sometimes it freezes when running
} > > X.   Sometimes it throws an error during the boot process, of which
} > > this is an example:
} > >
} > > kernel: page fault trap, code=0
} > > Stopped at netbsd:Xdoreti+ox8a: mov0(%esp),%gs
} > >
} > > If it boots successfully, we don't seem to hit errors unless we run X.
} > >
} > > I have next to no experience of such errors in 25 years of running
} > > Unix systems; my initial guess is there may be a fault in physical
} > > memory.
} > >
} > > Any suggestions?
} > 
} > The backtrace from ddb would be useful. Try writing 'tr' at the ddb
} > prompt.
} > 
} > -- 
} >   Rui Paulo			<rpaulo@{NetBSD{,-PT}.org,fnop.net}>
} > 
} > 
} 
} Typically:
} 
} Xdoreti at netbsd: Xdoreti+0x8a
} ---interrupt---
} 0x30
} 
} But I am now getting more variety, including a:
} 
} ---interrupt---
} Bad frame pointer: 0xc0a78818
} 0x3431:
} 
} I also had a series of attempted boots where the device probe
} sequence stopped with a long backtrace with references either to
} a "hardclock" or what looked like  shorthand for the real-time
} clock.
} 
} -- 
} Steve Blinkhorn <steve@prd.co.uk>
>-- End of excerpt from Steve Blinkhorn