mix in another counter when the psuedo-device attachment is called,

this actually happens *after* most of the sources have attached, so microtime
is as safe now as it was when they attached.
This commit is contained in:
dan 2002-10-08 09:59:27 +00:00
parent 234b471752
commit 7c2a3f9044
1 changed files with 17 additions and 2 deletions

View File

@ -1,4 +1,4 @@
/* $NetBSD: rnd.c,v 1.32 2002/10/07 11:02:20 dan Exp $ */
/* $NetBSD: rnd.c,v 1.33 2002/10/08 09:59:27 dan Exp $ */
/*-
* Copyright (c) 1997 The NetBSD Foundation, Inc.
@ -38,7 +38,7 @@
*/
#include <sys/cdefs.h>
__KERNEL_RCSID(0, "$NetBSD: rnd.c,v 1.32 2002/10/07 11:02:20 dan Exp $");
__KERNEL_RCSID(0, "$NetBSD: rnd.c,v 1.33 2002/10/08 09:59:27 dan Exp $");
#include <sys/param.h>
#include <sys/ioctl.h>
@ -286,8 +286,19 @@ rnd_estimate_entropy(rndsource_t *rs, u_int32_t t)
void
rndattach(int num)
{
u_int32_t c;
/*
* rnd_init() is actually called very early on in the boot
* process, pseudo's don't get attached until much later
* (after many of the sources have been attached).
* XXX Maybe this is left "in case", maybe it should be an assert?
*/
rnd_init();
/* mix in another counter */
c = rnd_counter();
rndpool_add_data(&rnd_pool, &c, sizeof(u_int32_t), 0);
}
void
@ -298,6 +309,10 @@ rnd_init(void)
if (rnd_ready)
return;
/*
* take a counter early, hoping that there's some variance in
* the following operations
*/
c = rnd_counter();
LIST_INIT(&rnd_sources);