I've a customer using 1.4.3 on a Fujitsu Teamserver as a webserver and
proxy (Apache and Squid). He is experiencing panics quite frequently and a
trace shows (sent by the customer, so potentially slightly abridged):

debugger
panic
pmap_alloc_pvpage
pmap_enter
uvm_fault
uvm_fault_wire
uvm_fork
fork1
sysfork
syscall()
syscall (number 2)
0x401056e3

I thought this could be to do with faulty memory (not withstanding the
fact that it had been running NT successfully for a couple of years prior
to having a proper OS put on it). It has 4 256MB DIMMs in it and he has
tried virtually every combination of slots and DIMMs without any
gain. I've successfully used similar hardware for similar purposes
elsewhere, so there might be a hardware basis to this. It's difficult to
track down though.

Any clues on how to proceed?

-- 
Stephen