> |     (date; dmesg ; sysctl -X; vmstat -i ) > /dev/random
> | 
> | Just playing it looks like you might get 4 so bits from the
> | rtc and clk interupt count alone.

> None.  Any data that is publically available via userland should not be
> used for cryptography.

The data from sysctl -X and vmstat -i vary quite a lot with time
and would be difficult to guess in their entrieity, even given the
their values at some later date. While any piece of data from these
commands isn't hard to guess, the idea is to take a few bits of
each of them.

I don't claim this produces hundreds of bits of entropy, but I'd
expect it to produce ten or twenty bits, even if you are given the
output of these from some stage shortly in the future.

I note from Mark's comments that writing stuff to /dev/random
doesn't change /dev/random's notion of how much entropy it has,
but does reseed the generator - so what we're talking about here
is the entropy of the seed - or how difficult to guess it is. He
does mention a very similar way of reseeding to the above:

        (ps -gauxwww; netstat -an; dmesg; vmstat -c10 1) > /dev/random

        David.


To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-current" in the body of the message

Reply via email to