1994-11-02 04:42:14 +00:00
|
|
|
/*
|
1995-02-14 23:04:52 +00:00
|
|
|
* Copyright 1994, 1995 Massachusetts Institute of Technology
|
1994-11-02 04:42:14 +00:00
|
|
|
*
|
1995-02-14 23:04:52 +00:00
|
|
|
* Permission to use, copy, modify, and distribute this software and
|
|
|
|
* its documentation for any purpose and without fee is hereby
|
|
|
|
* granted, provided that both the above copyright notice and this
|
|
|
|
* permission notice appear in all copies, that both the above
|
|
|
|
* copyright notice and this permission notice appear in all
|
|
|
|
* supporting documentation, and that the name of M.I.T. not be used
|
|
|
|
* in advertising or publicity pertaining to distribution of the
|
|
|
|
* software without specific, written prior permission. M.I.T. makes
|
|
|
|
* no representations about the suitability of this software for any
|
|
|
|
* purpose. It is provided "as is" without express or implied
|
|
|
|
* warranty.
|
1995-05-30 08:16:23 +00:00
|
|
|
*
|
1995-02-14 23:04:52 +00:00
|
|
|
* THIS SOFTWARE IS PROVIDED BY M.I.T. ``AS IS''. M.I.T. DISCLAIMS
|
|
|
|
* ALL EXPRESS OR IMPLIED WARRANTIES WITH REGARD TO THIS SOFTWARE,
|
|
|
|
* INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
|
|
|
|
* MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. IN NO EVENT
|
|
|
|
* SHALL M.I.T. BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
|
|
|
|
* SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
|
|
|
|
* LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF
|
|
|
|
* USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
|
|
|
|
* ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
|
|
|
|
* OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT
|
|
|
|
* OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
|
|
|
|
* SUCH DAMAGE.
|
1994-11-02 04:42:14 +00:00
|
|
|
*
|
1995-10-29 15:33:36 +00:00
|
|
|
* $Id: in_rmx.c,v 1.16 1995/09/18 15:51:32 wollman Exp $
|
1994-11-02 04:42:14 +00:00
|
|
|
*/
|
|
|
|
|
|
|
|
/*
|
|
|
|
* This code does two things necessary for the enhanced TCP metrics to
|
|
|
|
* function in a useful manner:
|
|
|
|
* 1) It marks all non-host routes as `cloning', thus ensuring that
|
|
|
|
* every actual reference to such a route actually gets turned
|
|
|
|
* into a reference to a host route to the specific destination
|
|
|
|
* requested.
|
|
|
|
* 2) When such routes lose all their references, it arranges for them
|
|
|
|
* to be deleted in some random collection of circumstances, so that
|
|
|
|
* a large quantity of stale routing data is not kept in kernel memory
|
|
|
|
* indefinitely. See in_rtqtimo() below for the exact mechanism.
|
|
|
|
*/
|
|
|
|
|
|
|
|
#include <sys/param.h>
|
|
|
|
#include <sys/systm.h>
|
|
|
|
#include <sys/kernel.h>
|
|
|
|
#include <sys/queue.h>
|
|
|
|
#include <sys/socket.h>
|
|
|
|
#include <sys/socketvar.h>
|
|
|
|
#include <sys/mbuf.h>
|
1994-12-01 23:19:48 +00:00
|
|
|
#include <sys/syslog.h>
|
1994-11-02 04:42:14 +00:00
|
|
|
|
|
|
|
#include <net/if.h>
|
|
|
|
#include <net/route.h>
|
|
|
|
#include <netinet/in.h>
|
|
|
|
#include <netinet/in_systm.h>
|
|
|
|
#include <netinet/in_var.h>
|
|
|
|
|
1995-07-10 15:39:16 +00:00
|
|
|
#include <netinet/ip.h>
|
|
|
|
#include <netinet/ip_var.h>
|
|
|
|
|
|
|
|
#include <netinet/tcp.h>
|
|
|
|
#include <netinet/tcp_seq.h>
|
|
|
|
#include <netinet/tcp_timer.h>
|
|
|
|
#include <netinet/tcp_var.h>
|
1995-09-18 15:51:40 +00:00
|
|
|
#ifndef MTUDISC
|
1995-07-10 15:39:16 +00:00
|
|
|
#include <netinet/tcpip.h>
|
1995-09-18 15:51:40 +00:00
|
|
|
#endif /* not MTUDISC */
|
1995-07-10 15:39:16 +00:00
|
|
|
|
1994-12-13 22:32:45 +00:00
|
|
|
#define RTPRF_OURS RTF_PROTO3 /* set on routes we manage */
|
1994-11-02 04:42:14 +00:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Do what we need to do when inserting a route.
|
|
|
|
*/
|
|
|
|
static struct radix_node *
|
|
|
|
in_addroute(void *v_arg, void *n_arg, struct radix_node_head *head,
|
|
|
|
struct radix_node *treenodes)
|
|
|
|
{
|
|
|
|
struct rtentry *rt = (struct rtentry *)treenodes;
|
1995-07-10 15:39:16 +00:00
|
|
|
struct sockaddr_in *sin = (struct sockaddr_in *)rt_key(rt);
|
1994-11-02 04:42:14 +00:00
|
|
|
|
|
|
|
/*
|
1995-01-23 02:02:50 +00:00
|
|
|
* For IP, all unicast non-host routes are automatically cloning.
|
1994-11-02 04:42:14 +00:00
|
|
|
*/
|
1995-01-23 02:02:50 +00:00
|
|
|
if(!(rt->rt_flags & (RTF_HOST | RTF_CLONING))) {
|
|
|
|
if(!IN_MULTICAST(ntohl(sin->sin_addr.s_addr))) {
|
|
|
|
rt->rt_flags |= RTF_PRCLONING;
|
|
|
|
}
|
|
|
|
}
|
1994-11-02 04:42:14 +00:00
|
|
|
|
1995-07-10 15:39:16 +00:00
|
|
|
/*
|
|
|
|
* We also specify a send and receive pipe size for every
|
|
|
|
* route added, to help TCP a bit. TCP doesn't actually
|
|
|
|
* want a true pipe size, which would be prohibitive in memory
|
|
|
|
* costs and is hard to compute anyway; it simply uses these
|
|
|
|
* values to size its buffers. So, we fill them in with the
|
|
|
|
* same values that TCP would have used anyway, and allow the
|
|
|
|
* installing program or the link layer to override these values
|
|
|
|
* as it sees fit. This will hopefully allow TCP more
|
|
|
|
* opportunities to save its ssthresh value.
|
|
|
|
*/
|
|
|
|
if (!rt->rt_rmx.rmx_sendpipe && !(rt->rt_rmx.rmx_locks & RTV_SPIPE))
|
|
|
|
rt->rt_rmx.rmx_sendpipe = tcp_sendspace;
|
|
|
|
|
|
|
|
if (!rt->rt_rmx.rmx_recvpipe && !(rt->rt_rmx.rmx_locks & RTV_RPIPE))
|
|
|
|
rt->rt_rmx.rmx_recvpipe = tcp_recvspace;
|
|
|
|
|
1995-09-18 15:51:40 +00:00
|
|
|
#ifndef MTUDISC
|
1995-07-10 15:39:16 +00:00
|
|
|
/*
|
|
|
|
* Finally, set an MTU, again duplicating logic in TCP.
|
|
|
|
* The in_localaddr() business will go away when we have
|
|
|
|
* proper PMTU discovery.
|
|
|
|
*/
|
1995-09-18 15:51:40 +00:00
|
|
|
#endif /* not MTUDISC */
|
1995-07-10 15:39:16 +00:00
|
|
|
if (!rt->rt_rmx.rmx_mtu && !(rt->rt_rmx.rmx_locks & RTV_MTU)
|
|
|
|
&& rt->rt_ifp)
|
1995-09-18 15:51:40 +00:00
|
|
|
#ifndef MTUDISC
|
1995-07-10 15:39:16 +00:00
|
|
|
rt->rt_rmx.rmx_mtu = (in_localaddr(sin->sin_addr)
|
|
|
|
? rt->rt_ifp->if_mtu
|
|
|
|
: tcp_mssdflt + sizeof(struct tcpiphdr));
|
1995-09-18 15:51:40 +00:00
|
|
|
#else /* MTUDISC */
|
|
|
|
rt->rt_rmx.rmx_mtu = rt->rt_ifp->if_mtu;
|
|
|
|
#endif /* MTUDISC */
|
1995-07-10 15:39:16 +00:00
|
|
|
|
1994-11-03 01:05:34 +00:00
|
|
|
return rn_addroute(v_arg, n_arg, head, treenodes);
|
1994-11-02 04:42:14 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
1994-11-03 01:05:34 +00:00
|
|
|
* This code is the inverse of in_clsroute: on first reference, if we
|
|
|
|
* were managing the route, stop doing so and set the expiration timer
|
|
|
|
* back off again.
|
1994-11-02 04:42:14 +00:00
|
|
|
*/
|
|
|
|
static struct radix_node *
|
|
|
|
in_matroute(void *v_arg, struct radix_node_head *head)
|
|
|
|
{
|
|
|
|
struct radix_node *rn = rn_match(v_arg, head);
|
|
|
|
struct rtentry *rt = (struct rtentry *)rn;
|
|
|
|
|
|
|
|
if(rt && rt->rt_refcnt == 0) { /* this is first reference */
|
1994-12-13 22:32:45 +00:00
|
|
|
if(rt->rt_flags & RTPRF_OURS) {
|
|
|
|
rt->rt_flags &= ~RTPRF_OURS;
|
1994-11-03 01:05:34 +00:00
|
|
|
rt->rt_rmx.rmx_expire = 0;
|
1994-11-02 04:42:14 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
return rn;
|
|
|
|
}
|
|
|
|
|
1995-02-14 23:11:26 +00:00
|
|
|
/* MIB variables: net.inet.ip.{rtexpire,rtmaxcache,rtminexpire}. */
|
1995-02-14 23:04:52 +00:00
|
|
|
int rtq_reallyold = 60*60; /* one hour is ``really old'' */
|
|
|
|
int rtq_toomany = 128; /* 128 cached routes is ``too many'' */
|
|
|
|
int rtq_minreallyold = 10; /* never automatically crank down to less */
|
1994-11-03 01:05:34 +00:00
|
|
|
|
1994-11-02 04:42:14 +00:00
|
|
|
/*
|
1995-01-23 02:02:50 +00:00
|
|
|
* On last reference drop, mark the route as belong to us so that it can be
|
1994-11-02 04:42:14 +00:00
|
|
|
* timed out.
|
|
|
|
*/
|
|
|
|
static void
|
|
|
|
in_clsroute(struct radix_node *rn, struct radix_node_head *head)
|
|
|
|
{
|
|
|
|
struct rtentry *rt = (struct rtentry *)rn;
|
1995-05-30 08:16:23 +00:00
|
|
|
|
1995-01-23 02:02:50 +00:00
|
|
|
if(!(rt->rt_flags & RTF_UP))
|
|
|
|
return; /* prophylactic measures */
|
|
|
|
|
1994-11-03 01:05:34 +00:00
|
|
|
if((rt->rt_flags & (RTF_LLINFO | RTF_HOST)) != RTF_HOST)
|
1994-11-02 04:42:14 +00:00
|
|
|
return;
|
|
|
|
|
1995-05-30 08:16:23 +00:00
|
|
|
if((rt->rt_flags & (RTF_WASCLONED | RTPRF_OURS))
|
1994-12-13 22:32:45 +00:00
|
|
|
!= RTF_WASCLONED)
|
1994-11-03 01:05:34 +00:00
|
|
|
return;
|
1994-11-02 04:42:14 +00:00
|
|
|
|
1995-02-14 23:04:52 +00:00
|
|
|
/*
|
|
|
|
* As requested by David Greenman:
|
|
|
|
* If rtq_reallyold is 0, just delete the route without
|
|
|
|
* waiting for a timeout cycle to kill it.
|
|
|
|
*/
|
|
|
|
if(rtq_reallyold != 0) {
|
|
|
|
rt->rt_flags |= RTPRF_OURS;
|
|
|
|
rt->rt_rmx.rmx_expire = time.tv_sec + rtq_reallyold;
|
|
|
|
} else {
|
|
|
|
rtrequest(RTM_DELETE,
|
|
|
|
(struct sockaddr *)rt_key(rt),
|
|
|
|
rt->rt_gateway, rt_mask(rt),
|
|
|
|
rt->rt_flags, 0);
|
|
|
|
}
|
1994-11-02 04:42:14 +00:00
|
|
|
}
|
|
|
|
|
1994-12-01 23:19:48 +00:00
|
|
|
struct rtqk_arg {
|
|
|
|
struct radix_node_head *rnh;
|
1994-12-13 22:32:45 +00:00
|
|
|
int draining;
|
1994-12-01 23:19:48 +00:00
|
|
|
int killed;
|
|
|
|
int found;
|
1995-02-14 23:11:26 +00:00
|
|
|
int updating;
|
1994-12-01 23:19:48 +00:00
|
|
|
time_t nextstop;
|
|
|
|
};
|
1994-11-02 04:42:14 +00:00
|
|
|
|
|
|
|
/*
|
1994-12-13 22:32:45 +00:00
|
|
|
* Get rid of old routes. When draining, this deletes everything, even when
|
1995-02-14 23:11:26 +00:00
|
|
|
* the timeout is not expired yet. When updating, this makes sure that
|
|
|
|
* nothing has a timeout longer than the current value of rtq_reallyold.
|
1994-11-02 04:42:14 +00:00
|
|
|
*/
|
1994-12-01 23:19:48 +00:00
|
|
|
static int
|
|
|
|
in_rtqkill(struct radix_node *rn, void *rock)
|
|
|
|
{
|
|
|
|
struct rtqk_arg *ap = rock;
|
|
|
|
struct rtentry *rt = (struct rtentry *)rn;
|
|
|
|
int err;
|
|
|
|
|
1994-12-13 22:32:45 +00:00
|
|
|
if(rt->rt_flags & RTPRF_OURS) {
|
1994-12-01 23:19:48 +00:00
|
|
|
ap->found++;
|
|
|
|
|
1995-02-14 23:11:26 +00:00
|
|
|
if(ap->draining || rt->rt_rmx.rmx_expire <= time.tv_sec) {
|
1994-12-01 23:19:48 +00:00
|
|
|
if(rt->rt_refcnt > 0)
|
1995-03-19 14:29:26 +00:00
|
|
|
panic("rtqkill route really not free");
|
1994-12-01 23:19:48 +00:00
|
|
|
|
|
|
|
err = rtrequest(RTM_DELETE,
|
|
|
|
(struct sockaddr *)rt_key(rt),
|
|
|
|
rt->rt_gateway, rt_mask(rt),
|
|
|
|
rt->rt_flags, 0);
|
|
|
|
if(err) {
|
1995-02-20 15:48:46 +00:00
|
|
|
log(LOG_WARNING, "in_rtqkill: error %d\n", err);
|
1994-12-01 23:19:48 +00:00
|
|
|
} else {
|
|
|
|
ap->killed++;
|
|
|
|
}
|
|
|
|
} else {
|
1995-05-30 08:16:23 +00:00
|
|
|
if(ap->updating
|
1995-06-21 19:48:53 +00:00
|
|
|
&& (rt->rt_rmx.rmx_expire - time.tv_sec
|
1995-02-14 23:11:26 +00:00
|
|
|
> rtq_reallyold)) {
|
1995-05-30 08:16:23 +00:00
|
|
|
rt->rt_rmx.rmx_expire = time.tv_sec
|
1995-02-14 23:11:26 +00:00
|
|
|
+ rtq_reallyold;
|
|
|
|
}
|
1994-12-01 23:19:48 +00:00
|
|
|
ap->nextstop = lmin(ap->nextstop,
|
|
|
|
rt->rt_rmx.rmx_expire);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
1995-02-14 23:04:52 +00:00
|
|
|
#define RTQ_TIMEOUT 60*10 /* run no less than once every ten minutes */
|
1995-01-23 02:02:50 +00:00
|
|
|
int rtq_timeout = RTQ_TIMEOUT;
|
|
|
|
|
1994-11-02 04:42:14 +00:00
|
|
|
static void
|
|
|
|
in_rtqtimo(void *rock)
|
|
|
|
{
|
1994-12-01 23:19:48 +00:00
|
|
|
struct radix_node_head *rnh = rock;
|
|
|
|
struct rtqk_arg arg;
|
|
|
|
struct timeval atv;
|
1995-02-14 23:04:52 +00:00
|
|
|
static time_t last_adjusted_timeout = 0;
|
1994-12-21 17:25:52 +00:00
|
|
|
int s;
|
1994-12-01 23:19:48 +00:00
|
|
|
|
|
|
|
arg.found = arg.killed = 0;
|
|
|
|
arg.rnh = rnh;
|
1995-01-23 02:02:50 +00:00
|
|
|
arg.nextstop = time.tv_sec + rtq_timeout;
|
1995-02-14 23:11:26 +00:00
|
|
|
arg.draining = arg.updating = 0;
|
1994-12-21 17:25:52 +00:00
|
|
|
s = splnet();
|
1994-12-01 23:19:48 +00:00
|
|
|
rnh->rnh_walktree(rnh, in_rtqkill, &arg);
|
1994-12-21 17:25:52 +00:00
|
|
|
splx(s);
|
1995-02-14 23:04:52 +00:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Attempt to be somewhat dynamic about this:
|
|
|
|
* If there are ``too many'' routes sitting around taking up space,
|
|
|
|
* then crank down the timeout, and see if we can't make some more
|
|
|
|
* go away. However, we make sure that we will never adjust more
|
|
|
|
* than once in rtq_timeout seconds, to keep from cranking down too
|
|
|
|
* hard.
|
|
|
|
*/
|
|
|
|
if((arg.found - arg.killed > rtq_toomany)
|
|
|
|
&& (time.tv_sec - last_adjusted_timeout >= rtq_timeout)
|
|
|
|
&& rtq_reallyold > rtq_minreallyold) {
|
|
|
|
rtq_reallyold = 2*rtq_reallyold / 3;
|
|
|
|
if(rtq_reallyold < rtq_minreallyold) {
|
|
|
|
rtq_reallyold = rtq_minreallyold;
|
|
|
|
}
|
|
|
|
|
|
|
|
last_adjusted_timeout = time.tv_sec;
|
1995-02-20 15:48:46 +00:00
|
|
|
log(LOG_DEBUG, "in_rtqtimo: adjusted rtq_reallyold to %d\n",
|
1995-02-14 23:04:52 +00:00
|
|
|
rtq_reallyold);
|
|
|
|
arg.found = arg.killed = 0;
|
1995-02-14 23:11:26 +00:00
|
|
|
arg.updating = 1;
|
1995-02-14 23:04:52 +00:00
|
|
|
s = splnet();
|
|
|
|
rnh->rnh_walktree(rnh, in_rtqkill, &arg);
|
|
|
|
splx(s);
|
|
|
|
}
|
|
|
|
|
1994-12-01 23:19:48 +00:00
|
|
|
atv.tv_usec = 0;
|
|
|
|
atv.tv_sec = arg.nextstop;
|
|
|
|
timeout(in_rtqtimo, rock, hzto(&atv));
|
|
|
|
}
|
1994-11-02 04:42:14 +00:00
|
|
|
|
1994-12-01 23:19:48 +00:00
|
|
|
void
|
|
|
|
in_rtqdrain(void)
|
|
|
|
{
|
1994-12-13 22:32:45 +00:00
|
|
|
struct radix_node_head *rnh = rt_tables[AF_INET];
|
|
|
|
struct rtqk_arg arg;
|
1994-12-21 17:25:52 +00:00
|
|
|
int s;
|
1994-12-13 22:32:45 +00:00
|
|
|
arg.found = arg.killed = 0;
|
|
|
|
arg.rnh = rnh;
|
|
|
|
arg.nextstop = 0;
|
|
|
|
arg.draining = 1;
|
1995-02-14 23:11:26 +00:00
|
|
|
arg.updating = 0;
|
1994-12-21 17:25:52 +00:00
|
|
|
s = splnet();
|
1994-12-13 22:32:45 +00:00
|
|
|
rnh->rnh_walktree(rnh, in_rtqkill, &arg);
|
|
|
|
splx(s);
|
1994-11-02 04:42:14 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Initialize our routing tree.
|
|
|
|
*/
|
|
|
|
int
|
|
|
|
in_inithead(void **head, int off)
|
|
|
|
{
|
|
|
|
struct radix_node_head *rnh;
|
|
|
|
|
|
|
|
if(!rn_inithead(head, off))
|
|
|
|
return 0;
|
|
|
|
|
1994-12-02 03:32:24 +00:00
|
|
|
if(head != (void **)&rt_tables[AF_INET]) /* BOGUS! */
|
|
|
|
return 1; /* only do this for the real routing table */
|
|
|
|
|
1994-11-02 04:42:14 +00:00
|
|
|
rnh = *head;
|
|
|
|
rnh->rnh_addaddr = in_addroute;
|
|
|
|
rnh->rnh_matchaddr = in_matroute;
|
|
|
|
rnh->rnh_close = in_clsroute;
|
1994-11-03 01:05:34 +00:00
|
|
|
in_rtqtimo(rnh); /* kick off timeout first time */
|
1994-11-02 04:42:14 +00:00
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
|