> Date: Sat, 2 Oct 2010 11:54:13 -0400 (EDT)
> From: Ted Unangst <[email protected]>
>
> Speaking of the entropy pool, it should be difficult for a user to
> influence, right? So it's not the greatest idea to allow anyone who opens
> the device to feed ioctls with fairly well known parameters and stir the
> pot?
The add_timer_randomness() calls stir the pot, but we don't count the
entropy. So no harm is done if a malicious users tries to attack the
pool with well-timed ioctls.
> Index: rnd.c
> ===================================================================
> RCS file: /cvs/src/sys/dev/rnd.c,v
> retrieving revision 1.103
> diff -u -r1.103 rnd.c
> --- rnd.c 2 Oct 2010 15:49:17 -0000 1.103
> +++ rnd.c 2 Oct 2010 15:51:43 -0000
> @@ -1140,8 +1140,6 @@
> int ret = 0;
> u_int cnt;
>
> - add_timer_randomness((u_long)p ^ (u_long)data ^ cmd);
> -
> switch (cmd) {
> case FIOASYNC:
> /* rnd has no async flag in softc so this is really a no-op. */
> @@ -1201,6 +1199,5 @@
> ret = ENOTTY;
> }
>
> - add_timer_randomness((u_long)p ^ (u_long)data ^ cmd);
> return ret;
> }
