>Etch. The kernel stack is something ~6K in size, I would feel very bad
>about having to increase it. But looking at your stack trace, it may be
>something else also that causes the problem. Can you do a ps and ps/a in
>ddb also please?

	with IPsec, kernel call chain will be longer than normal cases
	(due to chained headers like [IP AH ESP payload]), and crypto
	logic asks for a big buffer.  so it eats up more stack.
	i'm still not sure if the kernel stack overflow is the problem,
	i'm no expert in decoding vax debugging outputs.

itojun