d/freebsd-skq
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
26ae2b86b6
freebsd-skq/sys/net/rtsock.c

1310 lines
33 KiB
C
Raw Normal View History

/* -> /*- for license, minor formatting changes
2005-01-07 01:45:51 +00:00
/*-
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
* Copyright (c) 1988, 1991, 1993
* The Regents of the University of California. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* 4. Neither the name of the University nor the names of its contributors
* may be used to endorse or promote products derived from this software
* without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS 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.
*
Bring in latest CSRG revisions to this file: - Report destination address of a P2P link when servicing routing socket messages. - Report interface name, address, and destination address of a P2P link when servicing NET_RT_{DUMP,FLAGS} sysctls. Part of CSRG revision 8.6 coresponds to revision 1.12. CSRG revision 8.7 corresponds to revision 1.15.
2001-10-17 11:23:59 +00:00
* @(#)rtsock.c 8.7 (Berkeley) 10/12/95
$Id$ -> $FreeBSD$
1999-08-28 01:08:13 +00:00
* $FreeBSD$
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
*/
Ok, here it is, we finally add SCTP to current. Note that this work is not just mine, but it is also the works of Peter Lei and Michael Tuexen. They both are my two key other developers working on the project.. and they need ata-boy's too: **** peterlei@cisco.com tuexen@fh-muenster.de **** I did do a make sysent which updated the syscall's and sysproto.. I hope that is correct... without it you don't build since we have new syscalls for SCTP :-0 So go out and look at the NOTES, add option SCTP (make sure inet and inet6 are present too) and play with SCTP. I will see about comitting some test tools I have after I figure out where I should place them. I also have a lib (libsctp.a) that adds some of the missing socketapi functions that I need to put into lib's.. I will talk to George about this :-) There may still be some 64 bit issues in here, none of us have a 64 bit processor to test with yet.. Michael may have a MAC but thats another beast too.. If you have a mac and want to use SCTP contact Michael he maintains a web site with a loadable module with this code :-) Reviewed by: gnn Approved by: gnn
2006-11-03 15:23:16 +00:00
#include "opt_sctp.h"
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
#include <sys/param.h>
Revert the change of #includes in sys/filedesc.h and sys/socketvar.h. Requested by: bde Since locking sigio_lock is usually followed by calling pgsigio(), move the declaration of sigio_lock and the definitions of SIGIO_*() to sys/signalvar.h. While I am here, sort include files alphabetically, where possible.
2002-04-30 01:54:54 +00:00
#include <sys/domain.h>
Make networking domains drop-ins, through the magic of GNU ld. (Some day, there may even be LKMs.) Also, change the internal name of `unixdomain' to `localdomain' since AF_LOCAL is now the preferred name of this family. Declare netisr correctly and in the right place.
1995-05-11 00:13:26 +00:00
#include <sys/kernel.h>
Revert the change of #includes in sys/filedesc.h and sys/socketvar.h. Requested by: bde Since locking sigio_lock is usually followed by calling pgsigio(), move the declaration of sigio_lock and the definitions of SIGIO_*() to sys/signalvar.h. While I am here, sort include files alphabetically, where possible.
2002-04-30 01:54:54 +00:00
#include <sys/jail.h>
Added used #include - don't depend on <sys/mbuf.h> including <sys/malloc.h> (unless we only use the bogusly shared M*WAIT flags).
1997-09-02 01:19:47 +00:00
#include <sys/malloc.h>
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
#include <sys/mbuf.h>
Sweep kernel replacing suser(9) calls with priv(9) calls, assigning specific privilege names to a broad range of privileges. These may require some future tweaking. Sponsored by: nCircle Network Security, Inc. Obtained from: TrustedBSD Project Discussed on: arch@ Reviewed (at least in part) by: mlaier, jmg, pjd, bde, ceri, Alex Lyashkov <umka at sevcity dot net>, Skip Ford <skip dot ford at verizon dot net>, Antoine Brodin <antoine dot brodin at laposte dot net>
2006-11-06 13:42:10 +00:00
#include <sys/priv.h>
Revert the change of #includes in sys/filedesc.h and sys/socketvar.h. Requested by: bde Since locking sigio_lock is usually followed by calling pgsigio(), move the declaration of sigio_lock and the definitions of SIGIO_*() to sys/signalvar.h. While I am here, sort include files alphabetically, where possible.
2002-04-30 01:54:54 +00:00
#include <sys/proc.h>
#include <sys/protosw.h>
#include <sys/signalvar.h>
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
#include <sys/socket.h>
#include <sys/socketvar.h>
Revert the change of #includes in sys/filedesc.h and sys/socketvar.h. Requested by: bde Since locking sigio_lock is usually followed by calling pgsigio(), move the declaration of sigio_lock and the definitions of SIGIO_*() to sys/signalvar.h. While I am here, sort include files alphabetically, where possible.
2002-04-30 01:54:54 +00:00
#include <sys/sysctl.h>
#include <sys/systm.h>
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
#include <net/if.h>
Introduce a netisr to deliver kernel-generated routing, avoiding recursive entering of the socket code from the routing code: - Modify rt_dispatch() to bundle up the sockaddr family, if any, associated with a pending mbuf to dispatch to routing sockets, in an m_tag on the mbuf. - Allocate NETISR_ROUTE for use by routing sockets. - Introduce rtsintrq, an ifqueue to be used by the netisr, and introduce rts_input(), a function to unbundle the tagged sockaddr and inject the mbuf and address into raw_input(), which previously occurred in rt_dispatch(). - Introduce rts_init() to initialize rtsintrq, its mutex, and register the netisr. Perform this at the same point in system initialization as setup of the domains. This change introduces asynchrony between the generation of a pending routing socket message and delivery to sockets for use by userspace. It avoids socket->routing->rtsock->socket use and helps to avoid lock order reversals between the routing code and socket code (in particular, raw socket control blocks), as route locks are held over calls to rt_dispatch(). Reviewed by: "George V.Neville-Neil" <gnn@neville-neil.com> Conceptual head nod by: sam
2004-06-09 02:48:23 +00:00
#include <net/netisr.h>
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
#include <net/raw_cb.h>
Revert the change of #includes in sys/filedesc.h and sys/socketvar.h. Requested by: bde Since locking sigio_lock is usually followed by calling pgsigio(), move the declaration of sigio_lock and the definitions of SIGIO_*() to sys/signalvar.h. While I am here, sort include files alphabetically, where possible.
2002-04-30 01:54:54 +00:00
#include <net/route.h>
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
Give jail(8) the feature to allow raw sockets from within a jail, which is less restrictive but allows for more flexible jail usage (for those who are willing to make the sacrifice). The default is off, but allowing raw sockets within jails can now be accomplished by tuning security.jail.allow_raw_sockets to 1. Turning this on will allow you to use things like ping(8) or traceroute(8) from within a jail. The patch being committed is not identical to the patch in the PR. The committed version is more friendly to APIs which pjd is working on, so it should integrate into his work quite nicely. This change has also been presented and addressed on the freebsd-hackers mailing list. Submitted by: Christian S.J. Peron <maneo@bsdpro.com> PR: kern/65800
2004-04-26 19:46:52 +00:00
#include <netinet/in.h>
Ok, here it is, we finally add SCTP to current. Note that this work is not just mine, but it is also the works of Peter Lei and Michael Tuexen. They both are my two key other developers working on the project.. and they need ata-boy's too: **** peterlei@cisco.com tuexen@fh-muenster.de **** I did do a make sysent which updated the syscall's and sysproto.. I hope that is correct... without it you don't build since we have new syscalls for SCTP :-0 So go out and look at the NOTES, add option SCTP (make sure inet and inet6 are present too) and play with SCTP. I will see about comitting some test tools I have after I figure out where I should place them. I also have a lib (libsctp.a) that adds some of the missing socketapi functions that I need to put into lib's.. I will talk to George about this :-) There may still be some 64 bit issues in here, none of us have a 64 bit processor to test with yet.. Michael may have a MAC but thats another beast too.. If you have a mac and want to use SCTP contact Michael he maintains a web site with a loadable module with this code :-) Reviewed by: gnn Approved by: gnn
2006-11-03 15:23:16 +00:00
#ifdef SCTP
extern void sctp_addr_change(struct ifaddr *ifa, int cmd);
#endif /* SCTP */
Last major round (Unless Bruce thinks of somthing :-) of malloc changes. Distribute all but the most fundamental malloc types. This time I also remembered the trick to making things static: Put "static" in front of them. A couple of finer points by: bde
1997-10-12 20:26:33 +00:00
MALLOC_DEFINE(M_RTABLE, "routetbl", "routing tables");
cleanups prior to adding locking (and in some cases to eliminate locking): o move route_cb to be private to rtsock.c o replace global static route_proto by locals o eliminate global #define shorthands for info references o remove some register decls o ansi-fy function decls o move items to be close in scope to their usage o add rt_dispatch function for dispatching the actual message o cleanup tangled logic for doing all-but-me msg send Support by: FreeBSD Foundation
2003-10-03 18:15:54 +00:00
/* NB: these are not modified */
All net.* sysctl converted now.
1995-11-16 19:00:27 +00:00
static struct sockaddr route_dst = { 2, PF_ROUTE, };
static struct sockaddr route_src = { 2, PF_ROUTE, };
An actual fix for the routing default crashes that 1/ is compatible with the old route(1) in case needed. 2/ actually fixes the problem while vetting bad user input. note: I have already fixed route(1) so the problem shouldn't occur. if it does. use 0.0.0.0/0 instead of the word 'default' :)
1997-07-18 11:44:24 +00:00
static struct sockaddr sa_zero = { sizeof(sa_zero), AF_INET, };
cleanups prior to adding locking (and in some cases to eliminate locking): o move route_cb to be private to rtsock.c o replace global static route_proto by locals o eliminate global #define shorthands for info references o remove some register decls o ansi-fy function decls o move items to be close in scope to their usage o add rt_dispatch function for dispatching the actual message o cleanup tangled logic for doing all-but-me msg send Support by: FreeBSD Foundation
2003-10-03 18:15:54 +00:00
static struct {
Fix a typo (attacked -> attached). Approved by: sam
2004-08-24 08:47:15 +00:00
int ip_count; /* attached w/ AF_INET */
cleanups prior to adding locking (and in some cases to eliminate locking): o move route_cb to be private to rtsock.c o replace global static route_proto by locals o eliminate global #define shorthands for info references o remove some register decls o ansi-fy function decls o move items to be close in scope to their usage o add rt_dispatch function for dispatching the actual message o cleanup tangled logic for doing all-but-me msg send Support by: FreeBSD Foundation
2003-10-03 18:15:54 +00:00
int ip6_count; /* attached w/ AF_INET6 */
int ipx_count; /* attached w/ AF_IPX */
int any_count; /* total attached */
} route_cb;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
trivial locking rtsock_cb Sponsored by: FreeBSD Foundation
2003-10-03 18:27:02 +00:00
struct mtx rtsock_mtx;
MTX_SYSINIT(rtsock, &rtsock_mtx, "rtsock route_cb lock", MTX_DEF);
#define RTSOCK_LOCK() mtx_lock(&rtsock_mtx)
#define RTSOCK_UNLOCK() mtx_unlock(&rtsock_mtx)
#define RTSOCK_LOCK_ASSERT() mtx_assert(&rtsock_mtx, MA_OWNED)
Introduce a netisr to deliver kernel-generated routing, avoiding recursive entering of the socket code from the routing code: - Modify rt_dispatch() to bundle up the sockaddr family, if any, associated with a pending mbuf to dispatch to routing sockets, in an m_tag on the mbuf. - Allocate NETISR_ROUTE for use by routing sockets. - Introduce rtsintrq, an ifqueue to be used by the netisr, and introduce rts_input(), a function to unbundle the tagged sockaddr and inject the mbuf and address into raw_input(), which previously occurred in rt_dispatch(). - Introduce rts_init() to initialize rtsintrq, its mutex, and register the netisr. Perform this at the same point in system initialization as setup of the domains. This change introduces asynchrony between the generation of a pending routing socket message and delivery to sockets for use by userspace. It avoids socket->routing->rtsock->socket use and helps to avoid lock order reversals between the routing code and socket code (in particular, raw socket control blocks), as route locks are held over calls to rt_dispatch(). Reviewed by: "George V.Neville-Neil" <gnn@neville-neil.com> Conceptual head nod by: sam
2004-06-09 02:48:23 +00:00
static struct ifqueue rtsintrq;
Allow the size of the routing socket netisr queue to be configured using the tunable or sysctl 'net.route.netisr_maxqlen'. Default the maximum depth to 256 rather than IFQ_MAXLEN due to the downsides of dropping routing messages. MT5 candidate. Discussed with: mdodd, mlaier, Vincent Jardin <jardin at 6wind.com>
2004-08-21 21:20:06 +00:00
SYSCTL_NODE(_net, OID_AUTO, route, CTLFLAG_RD, 0, "");
SYSCTL_INT(_net_route, OID_AUTO, netisr_maxqlen, CTLFLAG_RW,
&rtsintrq.ifq_maxlen, 0, "maximum routing socket dispatch queue length");
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
struct walkarg {
All net.* sysctl converted now.
1995-11-16 19:00:27 +00:00
int w_tmemsize;
int w_op, w_arg;
caddr_t w_tmem;
struct sysctl_req *w_req;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
};
Introduce a netisr to deliver kernel-generated routing, avoiding recursive entering of the socket code from the routing code: - Modify rt_dispatch() to bundle up the sockaddr family, if any, associated with a pending mbuf to dispatch to routing sockets, in an m_tag on the mbuf. - Allocate NETISR_ROUTE for use by routing sockets. - Introduce rtsintrq, an ifqueue to be used by the netisr, and introduce rts_input(), a function to unbundle the tagged sockaddr and inject the mbuf and address into raw_input(), which previously occurred in rt_dispatch(). - Introduce rts_init() to initialize rtsintrq, its mutex, and register the netisr. Perform this at the same point in system initialization as setup of the domains. This change introduces asynchrony between the generation of a pending routing socket message and delivery to sockets for use by userspace. It avoids socket->routing->rtsock->socket use and helps to avoid lock order reversals between the routing code and socket code (in particular, raw socket control blocks), as route locks are held over calls to rt_dispatch(). Reviewed by: "George V.Neville-Neil" <gnn@neville-neil.com> Conceptual head nod by: sam
2004-06-09 02:48:23 +00:00
static void rts_input(struct mbuf *m);
Minor changes to improve code readability (no actual code changes): + replace 0 with NULL where appropriate (not complete) + remove register declaration while there + add argument names to function prototypes to have a better idea of what they are used for + add 'const' qualifiers in 3 places
2004-04-18 00:56:44 +00:00
static struct mbuf *rt_msg1(int type, struct rt_addrinfo *rtinfo);
static int rt_msg2(int type, struct rt_addrinfo *rtinfo,
caddr_t cp, struct walkarg *w);
static int rt_xaddrs(caddr_t cp, caddr_t cplim,
struct rt_addrinfo *rtinfo);
Remove __P.
2002-03-19 21:54:18 +00:00
static int sysctl_dumpentry(struct radix_node *rn, void *vw);
static int sysctl_iflist(int af, struct walkarg *w);
Add a sysctl MIB, NET_RT_IFMALIST, to retrieve multicast group memberships in a protocol-independent way. Submitted by: harti
2003-11-14 18:48:15 +00:00
static int sysctl_ifmalist(int af, struct walkarg *w);
Minor changes to improve code readability (no actual code changes): + replace 0 with NULL where appropriate (not complete) + remove register declaration while there + add argument names to function prototypes to have a better idea of what they are used for + add 'const' qualifiers in 3 places
2004-04-18 00:56:44 +00:00
static int route_output(struct mbuf *m, struct socket *so);
static void rt_setmetrics(u_long which, const struct rt_metrics *in,
struct rt_metrics_lite *out);
static void rt_getmetrics(const struct rt_metrics_lite *in,
struct rt_metrics *out);
static void rt_dispatch(struct mbuf *, const struct sockaddr *);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
Introduce a netisr to deliver kernel-generated routing, avoiding recursive entering of the socket code from the routing code: - Modify rt_dispatch() to bundle up the sockaddr family, if any, associated with a pending mbuf to dispatch to routing sockets, in an m_tag on the mbuf. - Allocate NETISR_ROUTE for use by routing sockets. - Introduce rtsintrq, an ifqueue to be used by the netisr, and introduce rts_input(), a function to unbundle the tagged sockaddr and inject the mbuf and address into raw_input(), which previously occurred in rt_dispatch(). - Introduce rts_init() to initialize rtsintrq, its mutex, and register the netisr. Perform this at the same point in system initialization as setup of the domains. This change introduces asynchrony between the generation of a pending routing socket message and delivery to sockets for use by userspace. It avoids socket->routing->rtsock->socket use and helps to avoid lock order reversals between the routing code and socket code (in particular, raw socket control blocks), as route locks are held over calls to rt_dispatch(). Reviewed by: "George V.Neville-Neil" <gnn@neville-neil.com> Conceptual head nod by: sam
2004-06-09 02:48:23 +00:00
static void
rts_init(void)
{
If a tunable for the routing socket netisr queue max is defined, allow it to override the default value, rather than the default value overriding the tunable.
2004-08-21 21:45:40 +00:00
int tmp;
Introduce a netisr to deliver kernel-generated routing, avoiding recursive entering of the socket code from the routing code: - Modify rt_dispatch() to bundle up the sockaddr family, if any, associated with a pending mbuf to dispatch to routing sockets, in an m_tag on the mbuf. - Allocate NETISR_ROUTE for use by routing sockets. - Introduce rtsintrq, an ifqueue to be used by the netisr, and introduce rts_input(), a function to unbundle the tagged sockaddr and inject the mbuf and address into raw_input(), which previously occurred in rt_dispatch(). - Introduce rts_init() to initialize rtsintrq, its mutex, and register the netisr. Perform this at the same point in system initialization as setup of the domains. This change introduces asynchrony between the generation of a pending routing socket message and delivery to sockets for use by userspace. It avoids socket->routing->rtsock->socket use and helps to avoid lock order reversals between the routing code and socket code (in particular, raw socket control blocks), as route locks are held over calls to rt_dispatch(). Reviewed by: "George V.Neville-Neil" <gnn@neville-neil.com> Conceptual head nod by: sam
2004-06-09 02:48:23 +00:00
Allow the size of the routing socket netisr queue to be configured using the tunable or sysctl 'net.route.netisr_maxqlen'. Default the maximum depth to 256 rather than IFQ_MAXLEN due to the downsides of dropping routing messages. MT5 candidate. Discussed with: mdodd, mlaier, Vincent Jardin <jardin at 6wind.com>
2004-08-21 21:20:06 +00:00
rtsintrq.ifq_maxlen = 256;
If a tunable for the routing socket netisr queue max is defined, allow it to override the default value, rather than the default value overriding the tunable.
2004-08-21 21:45:40 +00:00
if (TUNABLE_INT_FETCH("net.route.netisr_maxqlen", &tmp))
rtsintrq.ifq_maxlen = tmp;
Introduce a netisr to deliver kernel-generated routing, avoiding recursive entering of the socket code from the routing code: - Modify rt_dispatch() to bundle up the sockaddr family, if any, associated with a pending mbuf to dispatch to routing sockets, in an m_tag on the mbuf. - Allocate NETISR_ROUTE for use by routing sockets. - Introduce rtsintrq, an ifqueue to be used by the netisr, and introduce rts_input(), a function to unbundle the tagged sockaddr and inject the mbuf and address into raw_input(), which previously occurred in rt_dispatch(). - Introduce rts_init() to initialize rtsintrq, its mutex, and register the netisr. Perform this at the same point in system initialization as setup of the domains. This change introduces asynchrony between the generation of a pending routing socket message and delivery to sockets for use by userspace. It avoids socket->routing->rtsock->socket use and helps to avoid lock order reversals between the routing code and socket code (in particular, raw socket control blocks), as route locks are held over calls to rt_dispatch(). Reviewed by: "George V.Neville-Neil" <gnn@neville-neil.com> Conceptual head nod by: sam
2004-06-09 02:48:23 +00:00
mtx_init(&rtsintrq.ifq_mtx, "rts_inq", NULL, MTX_DEF);
netisr_register(NETISR_ROUTE, rts_input, &rtsintrq, NETISR_MPSAFE);
}
SYSINIT(rtsock, SI_SUB_PROTO_DOMAIN, SI_ORDER_THIRD, rts_init, 0)
static void
rts_input(struct mbuf *m)
{
struct sockproto route_proto;
unsigned short *family;
struct m_tag *tag;
route_proto.sp_family = PF_ROUTE;
tag = m_tag_find(m, PACKET_TAG_RTSOCKFAM, NULL);
if (tag != NULL) {
family = (unsigned short *)(tag + 1);
route_proto.sp_protocol = *family;
m_tag_delete(m, tag);
} else
route_proto.sp_protocol = 0;
raw_input(m, &route_proto, &route_src, &route_dst);
}
The long-awaited mega-massive-network-code- cleanup. Part I. This commit includes the following changes: 1) Old-style (pr_usrreq()) protocols are no longer supported, the compatibility glue for them is deleted, and the kernel will panic on boot if any are compiled in. 2) Certain protocol entry points are modified to take a process structure, so they they can easily tell whether or not it is possible to sleep, and also to access credentials. 3) SS_PRIV is no more, and with it goes the SO_PRIVSTATE setsockopt() call. Protocols should use the process pointer they are now passed. 4) The PF_LOCAL and PF_ROUTE families have been updated to use the new style, as has the `raw' skeleton family. 5) PF_LOCAL sockets now obey the process's umask when creating a socket in the filesystem. As a result, LINT is now broken. I'm hoping that some enterprising hacker with a bit more time will either make the broken bits work (should be easy for netipx) or dike them out.
1997-04-27 20:01:29 +00:00
/*
* It really doesn't make any sense at all for this code to share much
* with raw_usrreq.c, since its functionality is so restricted. XXX
*/
Change protocol switch pru_abort() API so that it returns void rather than an int, as an error here is not meaningful. Modify soabort() to unconditionally free the socket on the return of pru_abort(), and modify most protocols to no longer conditionally free the socket, since the caller will do this. This commit likely leaves parts of netinet and netinet6 in a situation where they may panic or leak memory, as they have not are not fully updated by this commit. This will be corrected shortly in followup commits to these components. MFC after: 3 months
2006-04-01 15:15:05 +00:00
static void
The long-awaited mega-massive-network-code- cleanup. Part I. This commit includes the following changes: 1) Old-style (pr_usrreq()) protocols are no longer supported, the compatibility glue for them is deleted, and the kernel will panic on boot if any are compiled in. 2) Certain protocol entry points are modified to take a process structure, so they they can easily tell whether or not it is possible to sleep, and also to access credentials. 3) SS_PRIV is no more, and with it goes the SO_PRIVSTATE setsockopt() call. Protocols should use the process pointer they are now passed. 4) The PF_LOCAL and PF_ROUTE families have been updated to use the new style, as has the `raw' skeleton family. 5) PF_LOCAL sockets now obey the process's umask when creating a socket in the filesystem. As a result, LINT is now broken. I'm hoping that some enterprising hacker with a bit more time will either make the broken bits work (should be easy for netipx) or dike them out.
1997-04-27 20:01:29 +00:00
rts_abort(struct socket *so)
{
De-spl parts of the routing socket code now generally protected through locking; leave some spl references around code where there are open questions about global variable references. Also, add an XXX regarding locking in sysctl. MFC after: 3 days
2005-08-25 13:30:04 +00:00
Change protocol switch pru_abort() API so that it returns void rather than an int, as an error here is not meaningful. Modify soabort() to unconditionally free the socket on the return of pru_abort(), and modify most protocols to no longer conditionally free the socket, since the caller will do this. This commit likely leaves parts of netinet and netinet6 in a situation where they may panic or leak memory, as they have not are not fully updated by this commit. This will be corrected shortly in followup commits to these components. MFC after: 3 months
2006-04-01 15:15:05 +00:00
raw_usrreqs.pru_abort(so);
The long-awaited mega-massive-network-code- cleanup. Part I. This commit includes the following changes: 1) Old-style (pr_usrreq()) protocols are no longer supported, the compatibility glue for them is deleted, and the kernel will panic on boot if any are compiled in. 2) Certain protocol entry points are modified to take a process structure, so they they can easily tell whether or not it is possible to sleep, and also to access credentials. 3) SS_PRIV is no more, and with it goes the SO_PRIVSTATE setsockopt() call. Protocols should use the process pointer they are now passed. 4) The PF_LOCAL and PF_ROUTE families have been updated to use the new style, as has the `raw' skeleton family. 5) PF_LOCAL sockets now obey the process's umask when creating a socket in the filesystem. As a result, LINT is now broken. I'm hoping that some enterprising hacker with a bit more time will either make the broken bits work (should be easy for netipx) or dike them out.
1997-04-27 20:01:29 +00:00
}
Change semantics of socket close and detach. Add a new protocol switch function, pru_close, to notify protocols that the file descriptor or other consumer of a socket is closing the socket. pru_abort is now a notification of close also, and no longer detaches. pru_detach is no longer used to notify of close, and will be called during socket tear-down by sofree() when all references to a socket evaporate after an earlier call to abort or close the socket. This means detach is now an unconditional teardown of a socket, whereas previously sockets could persist after detach of the protocol retained a reference. This faciliates sharing mutexes between layers of the network stack as the mutex is required during the checking and removal of references at the head of sofree(). With this change, pru_detach can now assume that the mutex will no longer be required by the socket layer after completion, whereas before this was not necessarily true. Reviewed by: gnn
2006-07-21 17:11:15 +00:00
static void
rts_close(struct socket *so)
{
raw_usrreqs.pru_close(so);
}
The long-awaited mega-massive-network-code- cleanup. Part I. This commit includes the following changes: 1) Old-style (pr_usrreq()) protocols are no longer supported, the compatibility glue for them is deleted, and the kernel will panic on boot if any are compiled in. 2) Certain protocol entry points are modified to take a process structure, so they they can easily tell whether or not it is possible to sleep, and also to access credentials. 3) SS_PRIV is no more, and with it goes the SO_PRIVSTATE setsockopt() call. Protocols should use the process pointer they are now passed. 4) The PF_LOCAL and PF_ROUTE families have been updated to use the new style, as has the `raw' skeleton family. 5) PF_LOCAL sockets now obey the process's umask when creating a socket in the filesystem. As a result, LINT is now broken. I'm hoping that some enterprising hacker with a bit more time will either make the broken bits work (should be easy for netipx) or dike them out.
1997-04-27 20:01:29 +00:00
/* pru_accept is EOPNOTSUPP */
All net.* sysctl converted now.
1995-11-16 19:00:27 +00:00
static int
KSE Milestone 2 Note ALL MODULES MUST BE RECOMPILED make the kernel aware that there are smaller units of scheduling than the process. (but only allow one thread per process at this time). This is functionally equivalent to teh previousl -current except that there is a thread associated with each process. Sorry john! (your next MFC will be a doosie!) Reviewed by: peter@freebsd.org, dillon@freebsd.org X-MFC after: ha ha ha ha
2001-09-12 08:38:13 +00:00
rts_attach(struct socket *so, int proto, struct thread *td)
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
{
The long-awaited mega-massive-network-code- cleanup. Part I. This commit includes the following changes: 1) Old-style (pr_usrreq()) protocols are no longer supported, the compatibility glue for them is deleted, and the kernel will panic on boot if any are compiled in. 2) Certain protocol entry points are modified to take a process structure, so they they can easily tell whether or not it is possible to sleep, and also to access credentials. 3) SS_PRIV is no more, and with it goes the SO_PRIVSTATE setsockopt() call. Protocols should use the process pointer they are now passed. 4) The PF_LOCAL and PF_ROUTE families have been updated to use the new style, as has the `raw' skeleton family. 5) PF_LOCAL sockets now obey the process's umask when creating a socket in the filesystem. As a result, LINT is now broken. I'm hoping that some enterprising hacker with a bit more time will either make the broken bits work (should be easy for netipx) or dike them out.
1997-04-27 20:01:29 +00:00
struct rawcb *rp;
int s, error;
Chance protocol switch method pru_detach() so that it returns void rather than an error. Detaches do not "fail", they other occur or the protocol flags SS_PROTOREF to take ownership of the socket. soclose() no longer looks at so_pcb to see if it's NULL, relying entirely on the protocol to decide whether it's time to free the socket or not using SS_PROTOREF. so_pcb is now entirely owned and managed by the protocol code. Likewise, no longer test so_pcb in other socket functions, such as soreceive(), which have no business digging into protocol internals. Protocol detach routines no longer try to free the socket on detach, this is performed in the socket code if the protocol permits it. In rts_detach(), no longer test for rp != NULL in detach, and likewise in other protocols that don't permit a NULL so_pcb, reduce the incidence of testing for it during detach. netinet and netinet6 are not fully updated to this change, which will be in an upcoming commit. In their current state they may leak memory or panic. MFC after: 3 months
2006-04-01 15:42:02 +00:00
KASSERT(so->so_pcb == NULL, ("rts_attach: so_pcb != NULL"));
Convert more malloc+bzero to malloc+M_ZERO. Submitted by: josh@zipperup.org Submitted by: Robert Drehmel <robd@gmx.net>
2000-12-08 21:51:06 +00:00
/* XXX */
Back out M_* changes, per decision of the TRB. Approved by: trb
2003-02-19 05:47:46 +00:00
MALLOC(rp, struct rawcb *, sizeof *rp, M_PCB, M_WAITOK | M_ZERO);
Minor changes to improve code readability (no actual code changes): + replace 0 with NULL where appropriate (not complete) + remove register declaration while there + add argument names to function prototypes to have a better idea of what they are used for + add 'const' qualifiers in 3 places
2004-04-18 00:56:44 +00:00
if (rp == NULL)
The long-awaited mega-massive-network-code- cleanup. Part I. This commit includes the following changes: 1) Old-style (pr_usrreq()) protocols are no longer supported, the compatibility glue for them is deleted, and the kernel will panic on boot if any are compiled in. 2) Certain protocol entry points are modified to take a process structure, so they they can easily tell whether or not it is possible to sleep, and also to access credentials. 3) SS_PRIV is no more, and with it goes the SO_PRIVSTATE setsockopt() call. Protocols should use the process pointer they are now passed. 4) The PF_LOCAL and PF_ROUTE families have been updated to use the new style, as has the `raw' skeleton family. 5) PF_LOCAL sockets now obey the process's umask when creating a socket in the filesystem. As a result, LINT is now broken. I'm hoping that some enterprising hacker with a bit more time will either make the broken bits work (should be easy for netipx) or dike them out.
1997-04-27 20:01:29 +00:00
return ENOBUFS;
/*
* The splnet() is necessary to block protocols from sending
* error notifications (like RTM_REDIRECT or RTM_LOSING) while
* this PCB is extant but incompletely initialized.
* Probably we should try to do more of this work beforehand and
* eliminate the spl.
*/
s = splnet();
so->so_pcb = (caddr_t)rp;
Synch with NetBSD and OpenBSD. Allow non-superuser to open, listen to, and send safe commands on the routing socket. Superuser priviledge is required for all commands but RTM_GET. Lose `setuid root' bit of route(8). Reviewed by: wollman, dd
2001-08-31 12:31:09 +00:00
error = raw_attach(so, proto);
The long-awaited mega-massive-network-code- cleanup. Part I. This commit includes the following changes: 1) Old-style (pr_usrreq()) protocols are no longer supported, the compatibility glue for them is deleted, and the kernel will panic on boot if any are compiled in. 2) Certain protocol entry points are modified to take a process structure, so they they can easily tell whether or not it is possible to sleep, and also to access credentials. 3) SS_PRIV is no more, and with it goes the SO_PRIVSTATE setsockopt() call. Protocols should use the process pointer they are now passed. 4) The PF_LOCAL and PF_ROUTE families have been updated to use the new style, as has the `raw' skeleton family. 5) PF_LOCAL sockets now obey the process's umask when creating a socket in the filesystem. As a result, LINT is now broken. I'm hoping that some enterprising hacker with a bit more time will either make the broken bits work (should be easy for netipx) or dike them out.
1997-04-27 20:01:29 +00:00
rp = sotorawcb(so);
if (error) {
splx(s);
fix memory leak when error during opening of routing socket PR: kern/29336 Submitted by: Richard Andrades <richard@xebeo.com> MFC after: 1 month
2001-08-02 19:56:29 +00:00
so->so_pcb = NULL;
The long-awaited mega-massive-network-code- cleanup. Part I. This commit includes the following changes: 1) Old-style (pr_usrreq()) protocols are no longer supported, the compatibility glue for them is deleted, and the kernel will panic on boot if any are compiled in. 2) Certain protocol entry points are modified to take a process structure, so they they can easily tell whether or not it is possible to sleep, and also to access credentials. 3) SS_PRIV is no more, and with it goes the SO_PRIVSTATE setsockopt() call. Protocols should use the process pointer they are now passed. 4) The PF_LOCAL and PF_ROUTE families have been updated to use the new style, as has the `raw' skeleton family. 5) PF_LOCAL sockets now obey the process's umask when creating a socket in the filesystem. As a result, LINT is now broken. I'm hoping that some enterprising hacker with a bit more time will either make the broken bits work (should be easy for netipx) or dike them out.
1997-04-27 20:01:29 +00:00
free(rp, M_PCB);
return error;
}
trivial locking rtsock_cb Sponsored by: FreeBSD Foundation
2003-10-03 18:27:02 +00:00
RTSOCK_LOCK();
The long-awaited mega-massive-network-code- cleanup. Part I. This commit includes the following changes: 1) Old-style (pr_usrreq()) protocols are no longer supported, the compatibility glue for them is deleted, and the kernel will panic on boot if any are compiled in. 2) Certain protocol entry points are modified to take a process structure, so they they can easily tell whether or not it is possible to sleep, and also to access credentials. 3) SS_PRIV is no more, and with it goes the SO_PRIVSTATE setsockopt() call. Protocols should use the process pointer they are now passed. 4) The PF_LOCAL and PF_ROUTE families have been updated to use the new style, as has the `raw' skeleton family. 5) PF_LOCAL sockets now obey the process's umask when creating a socket in the filesystem. As a result, LINT is now broken. I'm hoping that some enterprising hacker with a bit more time will either make the broken bits work (should be easy for netipx) or dike them out.
1997-04-27 20:01:29 +00:00
switch(rp->rcb_proto.sp_protocol) {
case AF_INET:
route_cb.ip_count++;
break;
Count AF_INET6 attachement to routing socket. Obtained from: KAME project
2000-01-28 20:10:51 +00:00
case AF_INET6:
route_cb.ip6_count++;
break;
The long-awaited mega-massive-network-code- cleanup. Part I. This commit includes the following changes: 1) Old-style (pr_usrreq()) protocols are no longer supported, the compatibility glue for them is deleted, and the kernel will panic on boot if any are compiled in. 2) Certain protocol entry points are modified to take a process structure, so they they can easily tell whether or not it is possible to sleep, and also to access credentials. 3) SS_PRIV is no more, and with it goes the SO_PRIVSTATE setsockopt() call. Protocols should use the process pointer they are now passed. 4) The PF_LOCAL and PF_ROUTE families have been updated to use the new style, as has the `raw' skeleton family. 5) PF_LOCAL sockets now obey the process's umask when creating a socket in the filesystem. As a result, LINT is now broken. I'm hoping that some enterprising hacker with a bit more time will either make the broken bits work (should be easy for netipx) or dike them out.
1997-04-27 20:01:29 +00:00
case AF_IPX:
route_cb.ipx_count++;
break;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
The long-awaited mega-massive-network-code- cleanup. Part I. This commit includes the following changes: 1) Old-style (pr_usrreq()) protocols are no longer supported, the compatibility glue for them is deleted, and the kernel will panic on boot if any are compiled in. 2) Certain protocol entry points are modified to take a process structure, so they they can easily tell whether or not it is possible to sleep, and also to access credentials. 3) SS_PRIV is no more, and with it goes the SO_PRIVSTATE setsockopt() call. Protocols should use the process pointer they are now passed. 4) The PF_LOCAL and PF_ROUTE families have been updated to use the new style, as has the `raw' skeleton family. 5) PF_LOCAL sockets now obey the process's umask when creating a socket in the filesystem. As a result, LINT is now broken. I'm hoping that some enterprising hacker with a bit more time will either make the broken bits work (should be easy for netipx) or dike them out.
1997-04-27 20:01:29 +00:00
rp->rcb_faddr = &route_src;
route_cb.any_count++;
trivial locking rtsock_cb Sponsored by: FreeBSD Foundation
2003-10-03 18:27:02 +00:00
RTSOCK_UNLOCK();
Remove so*_locked(), which were backed out by mistake.
2002-06-18 07:42:02 +00:00
soisconnected(so);
The long-awaited mega-massive-network-code- cleanup. Part I. This commit includes the following changes: 1) Old-style (pr_usrreq()) protocols are no longer supported, the compatibility glue for them is deleted, and the kernel will panic on boot if any are compiled in. 2) Certain protocol entry points are modified to take a process structure, so they they can easily tell whether or not it is possible to sleep, and also to access credentials. 3) SS_PRIV is no more, and with it goes the SO_PRIVSTATE setsockopt() call. Protocols should use the process pointer they are now passed. 4) The PF_LOCAL and PF_ROUTE families have been updated to use the new style, as has the `raw' skeleton family. 5) PF_LOCAL sockets now obey the process's umask when creating a socket in the filesystem. As a result, LINT is now broken. I'm hoping that some enterprising hacker with a bit more time will either make the broken bits work (should be easy for netipx) or dike them out.
1997-04-27 20:01:29 +00:00
so->so_options |= SO_USELOOPBACK;
splx(s);
return 0;
}
static int
KSE Milestone 2 Note ALL MODULES MUST BE RECOMPILED make the kernel aware that there are smaller units of scheduling than the process. (but only allow one thread per process at this time). This is functionally equivalent to teh previousl -current except that there is a thread associated with each process. Sorry john! (your next MFC will be a doosie!) Reviewed by: peter@freebsd.org, dillon@freebsd.org X-MFC after: ha ha ha ha
2001-09-12 08:38:13 +00:00
rts_bind(struct socket *so, struct sockaddr *nam, struct thread *td)
The long-awaited mega-massive-network-code- cleanup. Part I. This commit includes the following changes: 1) Old-style (pr_usrreq()) protocols are no longer supported, the compatibility glue for them is deleted, and the kernel will panic on boot if any are compiled in. 2) Certain protocol entry points are modified to take a process structure, so they they can easily tell whether or not it is possible to sleep, and also to access credentials. 3) SS_PRIV is no more, and with it goes the SO_PRIVSTATE setsockopt() call. Protocols should use the process pointer they are now passed. 4) The PF_LOCAL and PF_ROUTE families have been updated to use the new style, as has the `raw' skeleton family. 5) PF_LOCAL sockets now obey the process's umask when creating a socket in the filesystem. As a result, LINT is now broken. I'm hoping that some enterprising hacker with a bit more time will either make the broken bits work (should be easy for netipx) or dike them out.
1997-04-27 20:01:29 +00:00
{
De-spl parts of the routing socket code now generally protected through locking; leave some spl references around code where there are open questions about global variable references. Also, add an XXX regarding locking in sysctl. MFC after: 3 days
2005-08-25 13:30:04 +00:00
return (raw_usrreqs.pru_bind(so, nam, td)); /* xxx just EINVAL */
The long-awaited mega-massive-network-code- cleanup. Part I. This commit includes the following changes: 1) Old-style (pr_usrreq()) protocols are no longer supported, the compatibility glue for them is deleted, and the kernel will panic on boot if any are compiled in. 2) Certain protocol entry points are modified to take a process structure, so they they can easily tell whether or not it is possible to sleep, and also to access credentials. 3) SS_PRIV is no more, and with it goes the SO_PRIVSTATE setsockopt() call. Protocols should use the process pointer they are now passed. 4) The PF_LOCAL and PF_ROUTE families have been updated to use the new style, as has the `raw' skeleton family. 5) PF_LOCAL sockets now obey the process's umask when creating a socket in the filesystem. As a result, LINT is now broken. I'm hoping that some enterprising hacker with a bit more time will either make the broken bits work (should be easy for netipx) or dike them out.
1997-04-27 20:01:29 +00:00
}
static int
KSE Milestone 2 Note ALL MODULES MUST BE RECOMPILED make the kernel aware that there are smaller units of scheduling than the process. (but only allow one thread per process at this time). This is functionally equivalent to teh previousl -current except that there is a thread associated with each process. Sorry john! (your next MFC will be a doosie!) Reviewed by: peter@freebsd.org, dillon@freebsd.org X-MFC after: ha ha ha ha
2001-09-12 08:38:13 +00:00
rts_connect(struct socket *so, struct sockaddr *nam, struct thread *td)
The long-awaited mega-massive-network-code- cleanup. Part I. This commit includes the following changes: 1) Old-style (pr_usrreq()) protocols are no longer supported, the compatibility glue for them is deleted, and the kernel will panic on boot if any are compiled in. 2) Certain protocol entry points are modified to take a process structure, so they they can easily tell whether or not it is possible to sleep, and also to access credentials. 3) SS_PRIV is no more, and with it goes the SO_PRIVSTATE setsockopt() call. Protocols should use the process pointer they are now passed. 4) The PF_LOCAL and PF_ROUTE families have been updated to use the new style, as has the `raw' skeleton family. 5) PF_LOCAL sockets now obey the process's umask when creating a socket in the filesystem. As a result, LINT is now broken. I'm hoping that some enterprising hacker with a bit more time will either make the broken bits work (should be easy for netipx) or dike them out.
1997-04-27 20:01:29 +00:00
{
De-spl parts of the routing socket code now generally protected through locking; leave some spl references around code where there are open questions about global variable references. Also, add an XXX regarding locking in sysctl. MFC after: 3 days
2005-08-25 13:30:04 +00:00
return (raw_usrreqs.pru_connect(so, nam, td)); /* XXX just EINVAL */
The long-awaited mega-massive-network-code- cleanup. Part I. This commit includes the following changes: 1) Old-style (pr_usrreq()) protocols are no longer supported, the compatibility glue for them is deleted, and the kernel will panic on boot if any are compiled in. 2) Certain protocol entry points are modified to take a process structure, so they they can easily tell whether or not it is possible to sleep, and also to access credentials. 3) SS_PRIV is no more, and with it goes the SO_PRIVSTATE setsockopt() call. Protocols should use the process pointer they are now passed. 4) The PF_LOCAL and PF_ROUTE families have been updated to use the new style, as has the `raw' skeleton family. 5) PF_LOCAL sockets now obey the process's umask when creating a socket in the filesystem. As a result, LINT is now broken. I'm hoping that some enterprising hacker with a bit more time will either make the broken bits work (should be easy for netipx) or dike them out.
1997-04-27 20:01:29 +00:00
}
/* pru_connect2 is EOPNOTSUPP */
/* pru_control is EOPNOTSUPP */
Chance protocol switch method pru_detach() so that it returns void rather than an error. Detaches do not "fail", they other occur or the protocol flags SS_PROTOREF to take ownership of the socket. soclose() no longer looks at so_pcb to see if it's NULL, relying entirely on the protocol to decide whether it's time to free the socket or not using SS_PROTOREF. so_pcb is now entirely owned and managed by the protocol code. Likewise, no longer test so_pcb in other socket functions, such as soreceive(), which have no business digging into protocol internals. Protocol detach routines no longer try to free the socket on detach, this is performed in the socket code if the protocol permits it. In rts_detach(), no longer test for rp != NULL in detach, and likewise in other protocols that don't permit a NULL so_pcb, reduce the incidence of testing for it during detach. netinet and netinet6 are not fully updated to this change, which will be in an upcoming commit. In their current state they may leak memory or panic. MFC after: 3 months
2006-04-01 15:42:02 +00:00
static void
The long-awaited mega-massive-network-code- cleanup. Part I. This commit includes the following changes: 1) Old-style (pr_usrreq()) protocols are no longer supported, the compatibility glue for them is deleted, and the kernel will panic on boot if any are compiled in. 2) Certain protocol entry points are modified to take a process structure, so they they can easily tell whether or not it is possible to sleep, and also to access credentials. 3) SS_PRIV is no more, and with it goes the SO_PRIVSTATE setsockopt() call. Protocols should use the process pointer they are now passed. 4) The PF_LOCAL and PF_ROUTE families have been updated to use the new style, as has the `raw' skeleton family. 5) PF_LOCAL sockets now obey the process's umask when creating a socket in the filesystem. As a result, LINT is now broken. I'm hoping that some enterprising hacker with a bit more time will either make the broken bits work (should be easy for netipx) or dike them out.
1997-04-27 20:01:29 +00:00
rts_detach(struct socket *so)
{
struct rawcb *rp = sotorawcb(so);
Chance protocol switch method pru_detach() so that it returns void rather than an error. Detaches do not "fail", they other occur or the protocol flags SS_PROTOREF to take ownership of the socket. soclose() no longer looks at so_pcb to see if it's NULL, relying entirely on the protocol to decide whether it's time to free the socket or not using SS_PROTOREF. so_pcb is now entirely owned and managed by the protocol code. Likewise, no longer test so_pcb in other socket functions, such as soreceive(), which have no business digging into protocol internals. Protocol detach routines no longer try to free the socket on detach, this is performed in the socket code if the protocol permits it. In rts_detach(), no longer test for rp != NULL in detach, and likewise in other protocols that don't permit a NULL so_pcb, reduce the incidence of testing for it during detach. netinet and netinet6 are not fully updated to this change, which will be in an upcoming commit. In their current state they may leak memory or panic. MFC after: 3 months
2006-04-01 15:42:02 +00:00
KASSERT(rp != NULL, ("rts_detach: rp == NULL"));
RTSOCK_LOCK();
switch(rp->rcb_proto.sp_protocol) {
case AF_INET:
route_cb.ip_count--;
break;
case AF_INET6:
route_cb.ip6_count--;
break;
case AF_IPX:
route_cb.ipx_count--;
break;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
Chance protocol switch method pru_detach() so that it returns void rather than an error. Detaches do not "fail", they other occur or the protocol flags SS_PROTOREF to take ownership of the socket. soclose() no longer looks at so_pcb to see if it's NULL, relying entirely on the protocol to decide whether it's time to free the socket or not using SS_PROTOREF. so_pcb is now entirely owned and managed by the protocol code. Likewise, no longer test so_pcb in other socket functions, such as soreceive(), which have no business digging into protocol internals. Protocol detach routines no longer try to free the socket on detach, this is performed in the socket code if the protocol permits it. In rts_detach(), no longer test for rp != NULL in detach, and likewise in other protocols that don't permit a NULL so_pcb, reduce the incidence of testing for it during detach. netinet and netinet6 are not fully updated to this change, which will be in an upcoming commit. In their current state they may leak memory or panic. MFC after: 3 months
2006-04-01 15:42:02 +00:00
route_cb.any_count--;
RTSOCK_UNLOCK();
raw_usrreqs.pru_detach(so);
The long-awaited mega-massive-network-code- cleanup. Part I. This commit includes the following changes: 1) Old-style (pr_usrreq()) protocols are no longer supported, the compatibility glue for them is deleted, and the kernel will panic on boot if any are compiled in. 2) Certain protocol entry points are modified to take a process structure, so they they can easily tell whether or not it is possible to sleep, and also to access credentials. 3) SS_PRIV is no more, and with it goes the SO_PRIVSTATE setsockopt() call. Protocols should use the process pointer they are now passed. 4) The PF_LOCAL and PF_ROUTE families have been updated to use the new style, as has the `raw' skeleton family. 5) PF_LOCAL sockets now obey the process's umask when creating a socket in the filesystem. As a result, LINT is now broken. I'm hoping that some enterprising hacker with a bit more time will either make the broken bits work (should be easy for netipx) or dike them out.
1997-04-27 20:01:29 +00:00
}
static int
rts_disconnect(struct socket *so)
{
De-spl parts of the routing socket code now generally protected through locking; leave some spl references around code where there are open questions about global variable references. Also, add an XXX regarding locking in sysctl. MFC after: 3 days
2005-08-25 13:30:04 +00:00
return (raw_usrreqs.pru_disconnect(so));
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
The long-awaited mega-massive-network-code- cleanup. Part I. This commit includes the following changes: 1) Old-style (pr_usrreq()) protocols are no longer supported, the compatibility glue for them is deleted, and the kernel will panic on boot if any are compiled in. 2) Certain protocol entry points are modified to take a process structure, so they they can easily tell whether or not it is possible to sleep, and also to access credentials. 3) SS_PRIV is no more, and with it goes the SO_PRIVSTATE setsockopt() call. Protocols should use the process pointer they are now passed. 4) The PF_LOCAL and PF_ROUTE families have been updated to use the new style, as has the `raw' skeleton family. 5) PF_LOCAL sockets now obey the process's umask when creating a socket in the filesystem. As a result, LINT is now broken. I'm hoping that some enterprising hacker with a bit more time will either make the broken bits work (should be easy for netipx) or dike them out.
1997-04-27 20:01:29 +00:00
/* pru_listen is EOPNOTSUPP */
static int
Fix all areas of the system (or at least all those in LINT) to avoid storing socket addresses in mbufs. (Socket buffers are the one exception.) A number of kernel APIs needed to get fixed in order to make this happen. Also, fix three protocol families which kept PCBs in mbufs to not malloc them instead. Delete some old compatibility cruft while we're at it, and add some new routines in the in_cksum family.
1997-08-16 19:16:27 +00:00
rts_peeraddr(struct socket *so, struct sockaddr **nam)
The long-awaited mega-massive-network-code- cleanup. Part I. This commit includes the following changes: 1) Old-style (pr_usrreq()) protocols are no longer supported, the compatibility glue for them is deleted, and the kernel will panic on boot if any are compiled in. 2) Certain protocol entry points are modified to take a process structure, so they they can easily tell whether or not it is possible to sleep, and also to access credentials. 3) SS_PRIV is no more, and with it goes the SO_PRIVSTATE setsockopt() call. Protocols should use the process pointer they are now passed. 4) The PF_LOCAL and PF_ROUTE families have been updated to use the new style, as has the `raw' skeleton family. 5) PF_LOCAL sockets now obey the process's umask when creating a socket in the filesystem. As a result, LINT is now broken. I'm hoping that some enterprising hacker with a bit more time will either make the broken bits work (should be easy for netipx) or dike them out.
1997-04-27 20:01:29 +00:00
{
De-spl parts of the routing socket code now generally protected through locking; leave some spl references around code where there are open questions about global variable references. Also, add an XXX regarding locking in sysctl. MFC after: 3 days
2005-08-25 13:30:04 +00:00
return (raw_usrreqs.pru_peeraddr(so, nam));
The long-awaited mega-massive-network-code- cleanup. Part I. This commit includes the following changes: 1) Old-style (pr_usrreq()) protocols are no longer supported, the compatibility glue for them is deleted, and the kernel will panic on boot if any are compiled in. 2) Certain protocol entry points are modified to take a process structure, so they they can easily tell whether or not it is possible to sleep, and also to access credentials. 3) SS_PRIV is no more, and with it goes the SO_PRIVSTATE setsockopt() call. Protocols should use the process pointer they are now passed. 4) The PF_LOCAL and PF_ROUTE families have been updated to use the new style, as has the `raw' skeleton family. 5) PF_LOCAL sockets now obey the process's umask when creating a socket in the filesystem. As a result, LINT is now broken. I'm hoping that some enterprising hacker with a bit more time will either make the broken bits work (should be easy for netipx) or dike them out.
1997-04-27 20:01:29 +00:00
}
/* pru_rcvd is EOPNOTSUPP */
/* pru_rcvoob is EOPNOTSUPP */
static int
Fix all areas of the system (or at least all those in LINT) to avoid storing socket addresses in mbufs. (Socket buffers are the one exception.) A number of kernel APIs needed to get fixed in order to make this happen. Also, fix three protocol families which kept PCBs in mbufs to not malloc them instead. Delete some old compatibility cruft while we're at it, and add some new routines in the in_cksum family.
1997-08-16 19:16:27 +00:00
rts_send(struct socket *so, int flags, struct mbuf *m, struct sockaddr *nam,
KSE Milestone 2 Note ALL MODULES MUST BE RECOMPILED make the kernel aware that there are smaller units of scheduling than the process. (but only allow one thread per process at this time). This is functionally equivalent to teh previousl -current except that there is a thread associated with each process. Sorry john! (your next MFC will be a doosie!) Reviewed by: peter@freebsd.org, dillon@freebsd.org X-MFC after: ha ha ha ha
2001-09-12 08:38:13 +00:00
struct mbuf *control, struct thread *td)
The long-awaited mega-massive-network-code- cleanup. Part I. This commit includes the following changes: 1) Old-style (pr_usrreq()) protocols are no longer supported, the compatibility glue for them is deleted, and the kernel will panic on boot if any are compiled in. 2) Certain protocol entry points are modified to take a process structure, so they they can easily tell whether or not it is possible to sleep, and also to access credentials. 3) SS_PRIV is no more, and with it goes the SO_PRIVSTATE setsockopt() call. Protocols should use the process pointer they are now passed. 4) The PF_LOCAL and PF_ROUTE families have been updated to use the new style, as has the `raw' skeleton family. 5) PF_LOCAL sockets now obey the process's umask when creating a socket in the filesystem. As a result, LINT is now broken. I'm hoping that some enterprising hacker with a bit more time will either make the broken bits work (should be easy for netipx) or dike them out.
1997-04-27 20:01:29 +00:00
{
De-spl parts of the routing socket code now generally protected through locking; leave some spl references around code where there are open questions about global variable references. Also, add an XXX regarding locking in sysctl. MFC after: 3 days
2005-08-25 13:30:04 +00:00
return (raw_usrreqs.pru_send(so, flags, m, nam, control, td));
The long-awaited mega-massive-network-code- cleanup. Part I. This commit includes the following changes: 1) Old-style (pr_usrreq()) protocols are no longer supported, the compatibility glue for them is deleted, and the kernel will panic on boot if any are compiled in. 2) Certain protocol entry points are modified to take a process structure, so they they can easily tell whether or not it is possible to sleep, and also to access credentials. 3) SS_PRIV is no more, and with it goes the SO_PRIVSTATE setsockopt() call. Protocols should use the process pointer they are now passed. 4) The PF_LOCAL and PF_ROUTE families have been updated to use the new style, as has the `raw' skeleton family. 5) PF_LOCAL sockets now obey the process's umask when creating a socket in the filesystem. As a result, LINT is now broken. I'm hoping that some enterprising hacker with a bit more time will either make the broken bits work (should be easy for netipx) or dike them out.
1997-04-27 20:01:29 +00:00
}
/* pru_sense is null */
static int
rts_shutdown(struct socket *so)
{
De-spl parts of the routing socket code now generally protected through locking; leave some spl references around code where there are open questions about global variable references. Also, add an XXX regarding locking in sysctl. MFC after: 3 days
2005-08-25 13:30:04 +00:00
return (raw_usrreqs.pru_shutdown(so));
The long-awaited mega-massive-network-code- cleanup. Part I. This commit includes the following changes: 1) Old-style (pr_usrreq()) protocols are no longer supported, the compatibility glue for them is deleted, and the kernel will panic on boot if any are compiled in. 2) Certain protocol entry points are modified to take a process structure, so they they can easily tell whether or not it is possible to sleep, and also to access credentials. 3) SS_PRIV is no more, and with it goes the SO_PRIVSTATE setsockopt() call. Protocols should use the process pointer they are now passed. 4) The PF_LOCAL and PF_ROUTE families have been updated to use the new style, as has the `raw' skeleton family. 5) PF_LOCAL sockets now obey the process's umask when creating a socket in the filesystem. As a result, LINT is now broken. I'm hoping that some enterprising hacker with a bit more time will either make the broken bits work (should be easy for netipx) or dike them out.
1997-04-27 20:01:29 +00:00
}
static int
Fix all areas of the system (or at least all those in LINT) to avoid storing socket addresses in mbufs. (Socket buffers are the one exception.) A number of kernel APIs needed to get fixed in order to make this happen. Also, fix three protocol families which kept PCBs in mbufs to not malloc them instead. Delete some old compatibility cruft while we're at it, and add some new routines in the in_cksum family.
1997-08-16 19:16:27 +00:00
rts_sockaddr(struct socket *so, struct sockaddr **nam)
The long-awaited mega-massive-network-code- cleanup. Part I. This commit includes the following changes: 1) Old-style (pr_usrreq()) protocols are no longer supported, the compatibility glue for them is deleted, and the kernel will panic on boot if any are compiled in. 2) Certain protocol entry points are modified to take a process structure, so they they can easily tell whether or not it is possible to sleep, and also to access credentials. 3) SS_PRIV is no more, and with it goes the SO_PRIVSTATE setsockopt() call. Protocols should use the process pointer they are now passed. 4) The PF_LOCAL and PF_ROUTE families have been updated to use the new style, as has the `raw' skeleton family. 5) PF_LOCAL sockets now obey the process's umask when creating a socket in the filesystem. As a result, LINT is now broken. I'm hoping that some enterprising hacker with a bit more time will either make the broken bits work (should be easy for netipx) or dike them out.
1997-04-27 20:01:29 +00:00
{
De-spl parts of the routing socket code now generally protected through locking; leave some spl references around code where there are open questions about global variable references. Also, add an XXX regarding locking in sysctl. MFC after: 3 days
2005-08-25 13:30:04 +00:00
return (raw_usrreqs.pru_sockaddr(so, nam));
The long-awaited mega-massive-network-code- cleanup. Part I. This commit includes the following changes: 1) Old-style (pr_usrreq()) protocols are no longer supported, the compatibility glue for them is deleted, and the kernel will panic on boot if any are compiled in. 2) Certain protocol entry points are modified to take a process structure, so they they can easily tell whether or not it is possible to sleep, and also to access credentials. 3) SS_PRIV is no more, and with it goes the SO_PRIVSTATE setsockopt() call. Protocols should use the process pointer they are now passed. 4) The PF_LOCAL and PF_ROUTE families have been updated to use the new style, as has the `raw' skeleton family. 5) PF_LOCAL sockets now obey the process's umask when creating a socket in the filesystem. As a result, LINT is now broken. I'm hoping that some enterprising hacker with a bit more time will either make the broken bits work (should be easy for netipx) or dike them out.
1997-04-27 20:01:29 +00:00
}
static struct pr_usrreqs route_usrreqs = {
Initialize struct pr_userreqs in new/sparse style and fill in common default elements in net_init_domain(). This makes it possible to grep these structures and see any bogosities.
2004-11-08 14:44:54 +00:00
.pru_abort = rts_abort,
.pru_attach = rts_attach,
.pru_bind = rts_bind,
.pru_connect = rts_connect,
.pru_detach = rts_detach,
.pru_disconnect = rts_disconnect,
.pru_peeraddr = rts_peeraddr,
.pru_send = rts_send,
.pru_shutdown = rts_shutdown,
.pru_sockaddr = rts_sockaddr,
Change semantics of socket close and detach. Add a new protocol switch function, pru_close, to notify protocols that the file descriptor or other consumer of a socket is closing the socket. pru_abort is now a notification of close also, and no longer detaches. pru_detach is no longer used to notify of close, and will be called during socket tear-down by sofree() when all references to a socket evaporate after an earlier call to abort or close the socket. This means detach is now an unconditional teardown of a socket, whereas previously sockets could persist after detach of the protocol retained a reference. This faciliates sharing mutexes between layers of the network stack as the mutex is required during the checking and removal of references at the head of sofree(). With this change, pru_detach can now assume that the mutex will no longer be required by the socket layer after completion, whereas before this was not necessarily true. Reviewed by: gnn
2006-07-21 17:11:15 +00:00
.pru_close = rts_close,
The long-awaited mega-massive-network-code- cleanup. Part I. This commit includes the following changes: 1) Old-style (pr_usrreq()) protocols are no longer supported, the compatibility glue for them is deleted, and the kernel will panic on boot if any are compiled in. 2) Certain protocol entry points are modified to take a process structure, so they they can easily tell whether or not it is possible to sleep, and also to access credentials. 3) SS_PRIV is no more, and with it goes the SO_PRIVSTATE setsockopt() call. Protocols should use the process pointer they are now passed. 4) The PF_LOCAL and PF_ROUTE families have been updated to use the new style, as has the `raw' skeleton family. 5) PF_LOCAL sockets now obey the process's umask when creating a socket in the filesystem. As a result, LINT is now broken. I'm hoping that some enterprising hacker with a bit more time will either make the broken bits work (should be easy for netipx) or dike them out.
1997-04-27 20:01:29 +00:00
};
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
/*ARGSUSED*/
All net.* sysctl converted now.
1995-11-16 19:00:27 +00:00
static int
Minor changes to improve code readability (no actual code changes): + replace 0 with NULL where appropriate (not complete) + remove register declaration while there + add argument names to function prototypes to have a better idea of what they are used for + add 'const' qualifiers in 3 places
2004-04-18 00:56:44 +00:00
route_output(struct mbuf *m, struct socket *so)
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
{
cleanups prior to adding locking (and in some cases to eliminate locking): o move route_cb to be private to rtsock.c o replace global static route_proto by locals o eliminate global #define shorthands for info references o remove some register decls o ansi-fy function decls o move items to be close in scope to their usage o add rt_dispatch function for dispatching the actual message o cleanup tangled logic for doing all-but-me msg send Support by: FreeBSD Foundation
2003-10-03 18:15:54 +00:00
#define sa_equal(a1, a2) (bcmp((a1), (a2), (a1)->sa_len) == 0)
Minor changes to improve code readability (no actual code changes): + replace 0 with NULL where appropriate (not complete) + remove register declaration while there + add argument names to function prototypes to have a better idea of what they are used for + add 'const' qualifiers in 3 places
2004-04-18 00:56:44 +00:00
struct rt_msghdr *rtm = NULL;
struct rtentry *rt = NULL;
Updated routing-socket code from Berkeley Obtained from: Keith Bostic by way of Paul Traina
1995-05-10 16:48:52 +00:00
struct radix_node_head *rnh;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
struct rt_addrinfo info;
int len, error = 0;
Minor changes to improve code readability (no actual code changes): + replace 0 with NULL where appropriate (not complete) + remove register declaration while there + add argument names to function prototypes to have a better idea of what they are used for + add 'const' qualifiers in 3 places
2004-04-18 00:56:44 +00:00
struct ifnet *ifp = NULL;
Give jail(8) the feature to allow raw sockets from within a jail, which is less restrictive but allows for more flexible jail usage (for those who are willing to make the sacrifice). The default is off, but allowing raw sockets within jails can now be accomplished by tuning security.jail.allow_raw_sockets to 1. Turning this on will allow you to use things like ping(8) or traceroute(8) from within a jail. The patch being committed is not identical to the patch in the PR. The committed version is more friendly to APIs which pjd is working on, so it should integrate into his work quite nicely. This change has also been presented and addressed on the freebsd-hackers mailing list. Submitted by: Christian S.J. Peron <maneo@bsdpro.com> PR: kern/65800
2004-04-26 19:46:52 +00:00
struct sockaddr_in jail;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
#define senderr(e) { error = e; goto flush;}
Minor changes to improve code readability (no actual code changes): + replace 0 with NULL where appropriate (not complete) + remove register declaration while there + add argument names to function prototypes to have a better idea of what they are used for + add 'const' qualifiers in 3 places
2004-04-18 00:56:44 +00:00
if (m == NULL || ((m->m_len < sizeof(long)) &&
(m = m_pullup(m, sizeof(long))) == NULL))
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
return (ENOBUFS);
if ((m->m_flags & M_PKTHDR) == 0)
panic("route_output");
len = m->m_pkthdr.len;
if (len < sizeof(*rtm) ||
len != mtod(m, struct rt_msghdr *)->rtm_msglen) {
Minor changes to improve code readability (no actual code changes): + replace 0 with NULL where appropriate (not complete) + remove register declaration while there + add argument names to function prototypes to have a better idea of what they are used for + add 'const' qualifiers in 3 places
2004-04-18 00:56:44 +00:00
info.rti_info[RTAX_DST] = NULL;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
senderr(EINVAL);
}
R_Malloc(rtm, struct rt_msghdr *, len);
Minor changes to improve code readability (no actual code changes): + replace 0 with NULL where appropriate (not complete) + remove register declaration while there + add argument names to function prototypes to have a better idea of what they are used for + add 'const' qualifiers in 3 places
2004-04-18 00:56:44 +00:00
if (rtm == NULL) {
info.rti_info[RTAX_DST] = NULL;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
senderr(ENOBUFS);
}
m_copydata(m, 0, len, (caddr_t)rtm);
if (rtm->rtm_version != RTM_VERSION) {
Minor changes to improve code readability (no actual code changes): + replace 0 with NULL where appropriate (not complete) + remove register declaration while there + add argument names to function prototypes to have a better idea of what they are used for + add 'const' qualifiers in 3 places
2004-04-18 00:56:44 +00:00
info.rti_info[RTAX_DST] = NULL;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
senderr(EPROTONOSUPPORT);
}
rtm->rtm_pid = curproc->p_pid;
Pull post-4.4BSD change to sys/net/route.c from BSD/OS 4.2. Have sys/net/route.c:rtrequest1(), which takes ``rt_addrinfo *'' as the argument. Pass rt_addrinfo all the way down to rtrequest1 and ifa->ifa_rtrequest. 3rd argument of ifa->ifa_rtrequest is now ``rt_addrinfo *'' instead of ``sockaddr *'' (almost noone is using it anyways). Benefit: the following command now works. Previously we needed two route(8) invocations, "add" then "change". # route add -inet6 default ::1 -ifp gif0 Remove unsafe typecast in rtrequest(), from ``rtentry *'' to ``sockaddr *''. It was introduced by 4.3BSD-Reno and never corrected. Obtained from: BSD/OS, NetBSD MFC after: 1 month PR: kern/28360
2001-10-17 18:07:05 +00:00
bzero(&info, sizeof(info));
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
info.rti_addrs = rtm->rtm_addrs;
An actual fix for the routing default crashes that 1/ is compatible with the old route(1) in case needed. 2/ actually fixes the problem while vetting bad user input. note: I have already fixed route(1) so the problem shouldn't occur. if it does. use 0.0.0.0/0 instead of the word 'default' :)
1997-07-18 11:44:24 +00:00
if (rt_xaddrs((caddr_t)(rtm + 1), len + (caddr_t)rtm, &info)) {
Minor changes to improve code readability (no actual code changes): + replace 0 with NULL where appropriate (not complete) + remove register declaration while there + add argument names to function prototypes to have a better idea of what they are used for + add 'const' qualifiers in 3 places
2004-04-18 00:56:44 +00:00
info.rti_info[RTAX_DST] = NULL;
An actual fix for the routing default crashes that 1/ is compatible with the old route(1) in case needed. 2/ actually fixes the problem while vetting bad user input. note: I have already fixed route(1) so the problem shouldn't occur. if it does. use 0.0.0.0/0 instead of the word 'default' :)
1997-07-18 11:44:24 +00:00
senderr(EINVAL);
}
Pull post-4.4BSD change to sys/net/route.c from BSD/OS 4.2. Have sys/net/route.c:rtrequest1(), which takes ``rt_addrinfo *'' as the argument. Pass rt_addrinfo all the way down to rtrequest1 and ifa->ifa_rtrequest. 3rd argument of ifa->ifa_rtrequest is now ``rt_addrinfo *'' instead of ``sockaddr *'' (almost noone is using it anyways). Benefit: the following command now works. Previously we needed two route(8) invocations, "add" then "change". # route add -inet6 default ::1 -ifp gif0 Remove unsafe typecast in rtrequest(), from ``rtentry *'' to ``sockaddr *''. It was introduced by 4.3BSD-Reno and never corrected. Obtained from: BSD/OS, NetBSD MFC after: 1 month PR: kern/28360
2001-10-17 18:07:05 +00:00
info.rti_flags = rtm->rtm_flags;
Minor changes to improve code readability (no actual code changes): + replace 0 with NULL where appropriate (not complete) + remove register declaration while there + add argument names to function prototypes to have a better idea of what they are used for + add 'const' qualifiers in 3 places
2004-04-18 00:56:44 +00:00
if (info.rti_info[RTAX_DST] == NULL ||
cleanups prior to adding locking (and in some cases to eliminate locking): o move route_cb to be private to rtsock.c o replace global static route_proto by locals o eliminate global #define shorthands for info references o remove some register decls o ansi-fy function decls o move items to be close in scope to their usage o add rt_dispatch function for dispatching the actual message o cleanup tangled logic for doing all-but-me msg send Support by: FreeBSD Foundation
2003-10-03 18:15:54 +00:00
info.rti_info[RTAX_DST]->sa_family >= AF_MAX ||
Minor changes to improve code readability (no actual code changes): + replace 0 with NULL where appropriate (not complete) + remove register declaration while there + add argument names to function prototypes to have a better idea of what they are used for + add 'const' qualifiers in 3 places
2004-04-18 00:56:44 +00:00
(info.rti_info[RTAX_GATEWAY] != NULL &&
cleanups prior to adding locking (and in some cases to eliminate locking): o move route_cb to be private to rtsock.c o replace global static route_proto by locals o eliminate global #define shorthands for info references o remove some register decls o ansi-fy function decls o move items to be close in scope to their usage o add rt_dispatch function for dispatching the actual message o cleanup tangled logic for doing all-but-me msg send Support by: FreeBSD Foundation
2003-10-03 18:15:54 +00:00
info.rti_info[RTAX_GATEWAY]->sa_family >= AF_MAX))
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
senderr(EINVAL);
cleanups prior to adding locking (and in some cases to eliminate locking): o move route_cb to be private to rtsock.c o replace global static route_proto by locals o eliminate global #define shorthands for info references o remove some register decls o ansi-fy function decls o move items to be close in scope to their usage o add rt_dispatch function for dispatching the actual message o cleanup tangled logic for doing all-but-me msg send Support by: FreeBSD Foundation
2003-10-03 18:15:54 +00:00
if (info.rti_info[RTAX_GENMASK]) {
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
struct radix_node *t;
cleanups prior to adding locking (and in some cases to eliminate locking): o move route_cb to be private to rtsock.c o replace global static route_proto by locals o eliminate global #define shorthands for info references o remove some register decls o ansi-fy function decls o move items to be close in scope to their usage o add rt_dispatch function for dispatching the actual message o cleanup tangled logic for doing all-but-me msg send Support by: FreeBSD Foundation
2003-10-03 18:15:54 +00:00
t = rn_addmask((caddr_t) info.rti_info[RTAX_GENMASK], 0, 1);
Style and code unobfuscation.
2004-04-18 19:38:20 +00:00
if (t != NULL &&
More style and deobfuscation fixes. Submitted by: bde
2004-04-19 07:20:32 +00:00
bcmp((char *)(void *)info.rti_info[RTAX_GENMASK] + 1,
(char *)(void *)t->rn_key + 1,
((struct sockaddr *)t->rn_key)->sa_len - 1) == 0)
cleanups prior to adding locking (and in some cases to eliminate locking): o move route_cb to be private to rtsock.c o replace global static route_proto by locals o eliminate global #define shorthands for info references o remove some register decls o ansi-fy function decls o move items to be close in scope to their usage o add rt_dispatch function for dispatching the actual message o cleanup tangled logic for doing all-but-me msg send Support by: FreeBSD Foundation
2003-10-03 18:15:54 +00:00
info.rti_info[RTAX_GENMASK] =
More style and deobfuscation fixes. Submitted by: bde
2004-04-19 07:20:32 +00:00
(struct sockaddr *)t->rn_key;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
else
senderr(ENOBUFS);
}
Synch with NetBSD and OpenBSD. Allow non-superuser to open, listen to, and send safe commands on the routing socket. Superuser priviledge is required for all commands but RTM_GET. Lose `setuid root' bit of route(8). Reviewed by: wollman, dd
2001-08-31 12:31:09 +00:00
/*
* Verify that the caller has the appropriate privilege; RTM_GET
* is the only operation the non-superuser is allowed.
*/
Sweep kernel replacing suser(9) calls with priv(9) calls, assigning specific privilege names to a broad range of privileges. These may require some future tweaking. Sponsored by: nCircle Network Security, Inc. Obtained from: TrustedBSD Project Discussed on: arch@ Reviewed (at least in part) by: mlaier, jmg, pjd, bde, ceri, Alex Lyashkov <umka at sevcity dot net>, Skip Ford <skip dot ford at verizon dot net>, Antoine Brodin <antoine dot brodin at laposte dot net>
2006-11-06 13:42:10 +00:00
if (rtm->rtm_type != RTM_GET) {
error = priv_check(curthread, PRIV_NET_ROUTE);
if (error)
senderr(error);
}
Synch with NetBSD and OpenBSD. Allow non-superuser to open, listen to, and send safe commands on the routing socket. Superuser priviledge is required for all commands but RTM_GET. Lose `setuid root' bit of route(8). Reviewed by: wollman, dd
2001-08-31 12:31:09 +00:00
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
switch (rtm->rtm_type) {
cleanups prior to adding locking (and in some cases to eliminate locking): o move route_cb to be private to rtsock.c o replace global static route_proto by locals o eliminate global #define shorthands for info references o remove some register decls o ansi-fy function decls o move items to be close in scope to their usage o add rt_dispatch function for dispatching the actual message o cleanup tangled logic for doing all-but-me msg send Support by: FreeBSD Foundation
2003-10-03 18:15:54 +00:00
struct rtentry *saved_nrt;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
case RTM_ADD:
Minor changes to improve code readability (no actual code changes): + replace 0 with NULL where appropriate (not complete) + remove register declaration while there + add argument names to function prototypes to have a better idea of what they are used for + add 'const' qualifiers in 3 places
2004-04-18 00:56:44 +00:00
if (info.rti_info[RTAX_GATEWAY] == NULL)
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
senderr(EINVAL);
Minor changes to improve code readability (no actual code changes): + replace 0 with NULL where appropriate (not complete) + remove register declaration while there + add argument names to function prototypes to have a better idea of what they are used for + add 'const' qualifiers in 3 places
2004-04-18 00:56:44 +00:00
saved_nrt = NULL;
Pull post-4.4BSD change to sys/net/route.c from BSD/OS 4.2. Have sys/net/route.c:rtrequest1(), which takes ``rt_addrinfo *'' as the argument. Pass rt_addrinfo all the way down to rtrequest1 and ifa->ifa_rtrequest. 3rd argument of ifa->ifa_rtrequest is now ``rt_addrinfo *'' instead of ``sockaddr *'' (almost noone is using it anyways). Benefit: the following command now works. Previously we needed two route(8) invocations, "add" then "change". # route add -inet6 default ::1 -ifp gif0 Remove unsafe typecast in rtrequest(), from ``rtentry *'' to ``sockaddr *''. It was introduced by 4.3BSD-Reno and never corrected. Obtained from: BSD/OS, NetBSD MFC after: 1 month PR: kern/28360
2001-10-17 18:07:05 +00:00
error = rtrequest1(RTM_ADD, &info, &saved_nrt);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
if (error == 0 && saved_nrt) {
Locking for updates to routing table entries. Each rtentry gets a mutex that covers updates to the contents. Note this is separate from holding a reference and/or locking the routing table itself. Other/related changes: o rtredirect loses the final parameter by which an rtentry reference may be returned; this was never used and added unwarranted complexity for locking. o minor style cleanups to routing code (e.g. ansi-fy function decls) o remove the logic to bump the refcnt on the parent of cloned routes, we assume the parent will remain as long as the clone; doing this avoids a circularity in locking during delete o convert some timeouts to MPSAFE callouts Notes: 1. rt_mtx in struct rtentry is guarded by #ifdef _KERNEL as user-level applications cannot/do-no know about mutex's. Doing this requires that the mutex be the last element in the structure. A better solution is to introduce an externalized version of struct rtentry but this is a major task because of the intertwining of rtentry and other data structures that are visible to user applications. 2. There are known LOR's that are expected to go away with forthcoming work to eliminate many held references. If not these will be resolved prior to release. 3. ATM changes are untested. Sponsored by: FreeBSD Foundation Obtained from: BSD/OS (partly)
2003-10-04 03:44:50 +00:00
RT_LOCK(saved_nrt);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
rt_setmetrics(rtm->rtm_inits,
&rtm->rtm_rmx, &saved_nrt->rt_rmx);
- Fill in the correct rtm_index for RTM_ADD and RTM_CHANGE messages. - Allow RTM_CHANGE to change a number of route flags as specified by RTF_FMASK. - The unused rtm_use field in struct rt_msghdr is redesignated as rtm_fmask field to communicate route flag changes in RTM_CHANGE messages from userland. The use count of a route was moved to rtm_rmx a long time ago. For source code compatibility reasons a define of rtm_use to rtm_fmask is provided. These changes faciliate running of multiple cooperating routing daemons at the same time without causing undesired interference. Open[BGP|OSPF]D make use of these features to have IGP routes override EGP ones. Obtained from: OpenBSD (claudio@) MFC after: 3 days
2006-03-15 19:39:09 +00:00
rtm->rtm_index = saved_nrt->rt_ifp->if_index;
replace explicit changes to rt_refcnt by RT_ADDREF and RT_REMREF macros that expand to include assertions when the system is built with INVARIANTS Supported by: FreeBSD Foundation
2003-11-08 23:36:32 +00:00
RT_REMREF(saved_nrt);
cleanups prior to adding locking (and in some cases to eliminate locking): o move route_cb to be private to rtsock.c o replace global static route_proto by locals o eliminate global #define shorthands for info references o remove some register decls o ansi-fy function decls o move items to be close in scope to their usage o add rt_dispatch function for dispatching the actual message o cleanup tangled logic for doing all-but-me msg send Support by: FreeBSD Foundation
2003-10-03 18:15:54 +00:00
saved_nrt->rt_genmask = info.rti_info[RTAX_GENMASK];
Locking for updates to routing table entries. Each rtentry gets a mutex that covers updates to the contents. Note this is separate from holding a reference and/or locking the routing table itself. Other/related changes: o rtredirect loses the final parameter by which an rtentry reference may be returned; this was never used and added unwarranted complexity for locking. o minor style cleanups to routing code (e.g. ansi-fy function decls) o remove the logic to bump the refcnt on the parent of cloned routes, we assume the parent will remain as long as the clone; doing this avoids a circularity in locking during delete o convert some timeouts to MPSAFE callouts Notes: 1. rt_mtx in struct rtentry is guarded by #ifdef _KERNEL as user-level applications cannot/do-no know about mutex's. Doing this requires that the mutex be the last element in the structure. A better solution is to introduce an externalized version of struct rtentry but this is a major task because of the intertwining of rtentry and other data structures that are visible to user applications. 2. There are known LOR's that are expected to go away with forthcoming work to eliminate many held references. If not these will be resolved prior to release. 3. ATM changes are untested. Sponsored by: FreeBSD Foundation Obtained from: BSD/OS (partly)
2003-10-04 03:44:50 +00:00
RT_UNLOCK(saved_nrt);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
break;
case RTM_DELETE:
Minor changes to improve code readability (no actual code changes): + replace 0 with NULL where appropriate (not complete) + remove register declaration while there + add argument names to function prototypes to have a better idea of what they are used for + add 'const' qualifiers in 3 places
2004-04-18 00:56:44 +00:00
saved_nrt = NULL;
Pull post-4.4BSD change to sys/net/route.c from BSD/OS 4.2. Have sys/net/route.c:rtrequest1(), which takes ``rt_addrinfo *'' as the argument. Pass rt_addrinfo all the way down to rtrequest1 and ifa->ifa_rtrequest. 3rd argument of ifa->ifa_rtrequest is now ``rt_addrinfo *'' instead of ``sockaddr *'' (almost noone is using it anyways). Benefit: the following command now works. Previously we needed two route(8) invocations, "add" then "change". # route add -inet6 default ::1 -ifp gif0 Remove unsafe typecast in rtrequest(), from ``rtentry *'' to ``sockaddr *''. It was introduced by 4.3BSD-Reno and never corrected. Obtained from: BSD/OS, NetBSD MFC after: 1 month PR: kern/28360
2001-10-17 18:07:05 +00:00
error = rtrequest1(RTM_DELETE, &info, &saved_nrt);
Updated routing-socket code from Berkeley Obtained from: Keith Bostic by way of Paul Traina
1995-05-10 16:48:52 +00:00
if (error == 0) {
Locking for updates to routing table entries. Each rtentry gets a mutex that covers updates to the contents. Note this is separate from holding a reference and/or locking the routing table itself. Other/related changes: o rtredirect loses the final parameter by which an rtentry reference may be returned; this was never used and added unwarranted complexity for locking. o minor style cleanups to routing code (e.g. ansi-fy function decls) o remove the logic to bump the refcnt on the parent of cloned routes, we assume the parent will remain as long as the clone; doing this avoids a circularity in locking during delete o convert some timeouts to MPSAFE callouts Notes: 1. rt_mtx in struct rtentry is guarded by #ifdef _KERNEL as user-level applications cannot/do-no know about mutex's. Doing this requires that the mutex be the last element in the structure. A better solution is to introduce an externalized version of struct rtentry but this is a major task because of the intertwining of rtentry and other data structures that are visible to user applications. 2. There are known LOR's that are expected to go away with forthcoming work to eliminate many held references. If not these will be resolved prior to release. 3. ATM changes are untested. Sponsored by: FreeBSD Foundation Obtained from: BSD/OS (partly)
2003-10-04 03:44:50 +00:00
RT_LOCK(saved_nrt);
If the caller of rtrequest*(RTM_DELETE, ...) asked for a copy of the entry being removed (ret_nrt != NULL), increment the entry's rt_refcnt like we do it for RTM_ADD and RTM_RESOLVE, rather than messing around with 1->0 transitions for rtfree() all over.
2002-12-25 10:21:02 +00:00
rt = saved_nrt;
Updated routing-socket code from Berkeley Obtained from: Keith Bostic by way of Paul Traina
1995-05-10 16:48:52 +00:00
goto report;
}
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
break;
case RTM_GET:
case RTM_CHANGE:
case RTM_LOCK:
cleanups prior to adding locking (and in some cases to eliminate locking): o move route_cb to be private to rtsock.c o replace global static route_proto by locals o eliminate global #define shorthands for info references o remove some register decls o ansi-fy function decls o move items to be close in scope to their usage o add rt_dispatch function for dispatching the actual message o cleanup tangled logic for doing all-but-me msg send Support by: FreeBSD Foundation
2003-10-03 18:15:54 +00:00
rnh = rt_tables[info.rti_info[RTAX_DST]->sa_family];
Minor changes to improve code readability (no actual code changes): + replace 0 with NULL where appropriate (not complete) + remove register declaration while there + add argument names to function prototypes to have a better idea of what they are used for + add 'const' qualifiers in 3 places
2004-04-18 00:56:44 +00:00
if (rnh == NULL)
Updated routing-socket code from Berkeley Obtained from: Keith Bostic by way of Paul Traina
1995-05-10 16:48:52 +00:00
senderr(EAFNOSUPPORT);
SMP locking for radix nodes.
2002-12-24 03:03:39 +00:00
RADIX_NODE_HEAD_LOCK(rnh);
cleanups prior to adding locking (and in some cases to eliminate locking): o move route_cb to be private to rtsock.c o replace global static route_proto by locals o eliminate global #define shorthands for info references o remove some register decls o ansi-fy function decls o move items to be close in scope to their usage o add rt_dispatch function for dispatching the actual message o cleanup tangled logic for doing all-but-me msg send Support by: FreeBSD Foundation
2003-10-03 18:15:54 +00:00
rt = (struct rtentry *) rnh->rnh_lookup(info.rti_info[RTAX_DST],
info.rti_info[RTAX_NETMASK], rnh);
o To prevent a race between RTM_DELETE message and arptimer() deleting stale entry, we need to lock rtentry before unlocking radix head. Reviewed by: sam
2005-08-11 08:26:31 +00:00
if (rt == NULL) { /* XXX looks bogus */
RADIX_NODE_HEAD_UNLOCK(rnh);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
senderr(ESRCH);
o To prevent a race between RTM_DELETE message and arptimer() deleting stale entry, we need to lock rtentry before unlocking radix head. Reviewed by: sam
2005-08-11 08:26:31 +00:00
}
Locking for updates to routing table entries. Each rtentry gets a mutex that covers updates to the contents. Note this is separate from holding a reference and/or locking the routing table itself. Other/related changes: o rtredirect loses the final parameter by which an rtentry reference may be returned; this was never used and added unwarranted complexity for locking. o minor style cleanups to routing code (e.g. ansi-fy function decls) o remove the logic to bump the refcnt on the parent of cloned routes, we assume the parent will remain as long as the clone; doing this avoids a circularity in locking during delete o convert some timeouts to MPSAFE callouts Notes: 1. rt_mtx in struct rtentry is guarded by #ifdef _KERNEL as user-level applications cannot/do-no know about mutex's. Doing this requires that the mutex be the last element in the structure. A better solution is to introduce an externalized version of struct rtentry but this is a major task because of the intertwining of rtentry and other data structures that are visible to user applications. 2. There are known LOR's that are expected to go away with forthcoming work to eliminate many held references. If not these will be resolved prior to release. 3. ATM changes are untested. Sponsored by: FreeBSD Foundation Obtained from: BSD/OS (partly)
2003-10-04 03:44:50 +00:00
RT_LOCK(rt);
replace explicit changes to rt_refcnt by RT_ADDREF and RT_REMREF macros that expand to include assertions when the system is built with INVARIANTS Supported by: FreeBSD Foundation
2003-11-08 23:36:32 +00:00
RT_ADDREF(rt);
o To prevent a race between RTM_DELETE message and arptimer() deleting stale entry, we need to lock rtentry before unlocking radix head. Reviewed by: sam
2005-08-11 08:26:31 +00:00
RADIX_NODE_HEAD_UNLOCK(rnh);
SMP locking for radix nodes.
2002-12-24 03:03:39 +00:00
Fix for PR 82974. We were not checking that the route looked up in the case of an RTM_CHANGE was specific, i.e. that it matched completely. This led to a route change of a non-existent route changing the default route as the radix code would simply back track to that point and hand that route back to the routing socket code. PR: 82974 Reviewed by: Tai-hwa Liang <avatar@mmlab.cse.yzu.edu.tw> Ben Kaduk <minimarmot@gmail.com> Bjoern A. Zeeb <bzeeb-lists@lists.zabbadoz.net> Obtained from: OpenBSD with modifications. MFC after: 2 weeks
2005-07-15 09:18:34 +00:00
/*
* Fix for PR: 82974
*
* RTM_CHANGE/LOCK need a perfect match, rn_lookup()
* returns a perfect match in case a netmask is
* specified. For host routes only a longest prefix
* match is returned so it is necessary to compare the
* existence of the netmask. If both have a netmask
* rnh_lookup() did a perfect match and if none of them
* have a netmask both are host routes which is also a
* perfect match.
*/
if (rtm->rtm_type != RTM_GET &&
(!rt_mask(rt) != !info.rti_info[RTAX_NETMASK])) {
RT_UNLOCK(rt);
senderr(ESRCH);
}
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
switch(rtm->rtm_type) {
case RTM_GET:
Updated routing-socket code from Berkeley Obtained from: Keith Bostic by way of Paul Traina
1995-05-10 16:48:52 +00:00
report:
Locking for updates to routing table entries. Each rtentry gets a mutex that covers updates to the contents. Note this is separate from holding a reference and/or locking the routing table itself. Other/related changes: o rtredirect loses the final parameter by which an rtentry reference may be returned; this was never used and added unwarranted complexity for locking. o minor style cleanups to routing code (e.g. ansi-fy function decls) o remove the logic to bump the refcnt on the parent of cloned routes, we assume the parent will remain as long as the clone; doing this avoids a circularity in locking during delete o convert some timeouts to MPSAFE callouts Notes: 1. rt_mtx in struct rtentry is guarded by #ifdef _KERNEL as user-level applications cannot/do-no know about mutex's. Doing this requires that the mutex be the last element in the structure. A better solution is to introduce an externalized version of struct rtentry but this is a major task because of the intertwining of rtentry and other data structures that are visible to user applications. 2. There are known LOR's that are expected to go away with forthcoming work to eliminate many held references. If not these will be resolved prior to release. 3. ATM changes are untested. Sponsored by: FreeBSD Foundation Obtained from: BSD/OS (partly)
2003-10-04 03:44:50 +00:00
RT_LOCK_ASSERT(rt);
cleanups prior to adding locking (and in some cases to eliminate locking): o move route_cb to be private to rtsock.c o replace global static route_proto by locals o eliminate global #define shorthands for info references o remove some register decls o ansi-fy function decls o move items to be close in scope to their usage o add rt_dispatch function for dispatching the actual message o cleanup tangled logic for doing all-but-me msg send Support by: FreeBSD Foundation
2003-10-03 18:15:54 +00:00
info.rti_info[RTAX_DST] = rt_key(rt);
info.rti_info[RTAX_GATEWAY] = rt->rt_gateway;
info.rti_info[RTAX_NETMASK] = rt_mask(rt);
info.rti_info[RTAX_GENMASK] = rt->rt_genmask;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
if (rtm->rtm_addrs & (RTA_IFP | RTA_IFA)) {
Cosmetics: to silence gcc -wall.
1994-10-08 22:38:27 +00:00
ifp = rt->rt_ifp;
if (ifp) {
Consistently use ifaddr_byindex() to access the link-level address of an interface. No functional change. On passing, comment a likely bug in net/rtsock.c:sysctl_ifmalist() which, if confirmed, would deserve to be fixed and MFC'ed
2004-04-16 08:14:34 +00:00
info.rti_info[RTAX_IFP] =
- Store pointer to the link-level address right in "struct ifnet" rather than in ifindex_table[]; all (except one) accesses are through ifp anyway. IF_LLADDR() works faster, and all (except one) ifaddr_byindex() users were converted to use ifp->if_addr. - Stop storing a (pointer to) Ethernet address in "struct arpcom", and drop the IFP2ENADDR() macro; all users have been converted to use IF_LLADDR() instead.
2005-11-11 16:04:59 +00:00
ifp->if_addr->ifa_addr;
Give jail(8) the feature to allow raw sockets from within a jail, which is less restrictive but allows for more flexible jail usage (for those who are willing to make the sacrifice). The default is off, but allowing raw sockets within jails can now be accomplished by tuning security.jail.allow_raw_sockets to 1. Turning this on will allow you to use things like ping(8) or traceroute(8) from within a jail. The patch being committed is not identical to the patch in the PR. The committed version is more friendly to APIs which pjd is working on, so it should integrate into his work quite nicely. This change has also been presented and addressed on the freebsd-hackers mailing list. Submitted by: Christian S.J. Peron <maneo@bsdpro.com> PR: kern/65800
2004-04-26 19:46:52 +00:00
if (jailed(so->so_cred)) {
Be consistent and use bzero() instead of memset().
2004-07-06 03:29:41 +00:00
bzero(&jail, sizeof(jail));
Give jail(8) the feature to allow raw sockets from within a jail, which is less restrictive but allows for more flexible jail usage (for those who are willing to make the sacrifice). The default is off, but allowing raw sockets within jails can now be accomplished by tuning security.jail.allow_raw_sockets to 1. Turning this on will allow you to use things like ping(8) or traceroute(8) from within a jail. The patch being committed is not identical to the patch in the PR. The committed version is more friendly to APIs which pjd is working on, so it should integrate into his work quite nicely. This change has also been presented and addressed on the freebsd-hackers mailing list. Submitted by: Christian S.J. Peron <maneo@bsdpro.com> PR: kern/65800
2004-04-26 19:46:52 +00:00
jail.sin_family = PF_INET;
jail.sin_len = sizeof(jail);
jail.sin_addr.s_addr =
htonl(prison_getip(so->so_cred));
info.rti_info[RTAX_IFA] =
o Fix misindentation in the previous commit.
2004-05-03 17:15:34 +00:00
(struct sockaddr *)&jail;
Give jail(8) the feature to allow raw sockets from within a jail, which is less restrictive but allows for more flexible jail usage (for those who are willing to make the sacrifice). The default is off, but allowing raw sockets within jails can now be accomplished by tuning security.jail.allow_raw_sockets to 1. Turning this on will allow you to use things like ping(8) or traceroute(8) from within a jail. The patch being committed is not identical to the patch in the PR. The committed version is more friendly to APIs which pjd is working on, so it should integrate into his work quite nicely. This change has also been presented and addressed on the freebsd-hackers mailing list. Submitted by: Christian S.J. Peron <maneo@bsdpro.com> PR: kern/65800
2004-04-26 19:46:52 +00:00
} else
info.rti_info[RTAX_IFA] =
o Fix misindentation in the previous commit.
2004-05-03 17:15:34 +00:00
rt->rt_ifa->ifa_addr;
Bring in latest CSRG revisions to this file: - Report destination address of a P2P link when servicing routing socket messages. - Report interface name, address, and destination address of a P2P link when servicing NET_RT_{DUMP,FLAGS} sysctls. Part of CSRG revision 8.6 coresponds to revision 1.12. CSRG revision 8.7 corresponds to revision 1.15.
2001-10-17 11:23:59 +00:00
if (ifp->if_flags & IFF_POINTOPOINT)
o Fix misindentation in the previous commit.
2004-05-03 17:15:34 +00:00
info.rti_info[RTAX_BRD] =
rt->rt_ifa->ifa_dstaddr;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
rtm->rtm_index = ifp->if_index;
} else {
Minor changes to improve code readability (no actual code changes): + replace 0 with NULL where appropriate (not complete) + remove register declaration while there + add argument names to function prototypes to have a better idea of what they are used for + add 'const' qualifiers in 3 places
2004-04-18 00:56:44 +00:00
info.rti_info[RTAX_IFP] = NULL;
info.rti_info[RTAX_IFA] = NULL;
Fix some harmless mis-indents. Spotted by: FlexeLint
2002-10-01 15:48:31 +00:00
}
When returing an RTM_GET message through the routing socket fill in the rtm_index field whenever we have an interface pointer. This is consistent with the RTM_GET messages returned by sysctl().
2005-06-09 12:20:50 +00:00
} else if ((ifp = rt->rt_ifp) != NULL) {
rtm->rtm_index = ifp->if_index;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
misc cleanup in sysctl_ifmalist(): + remove a partly incorrect comment that i introduced in the last commit; + deal with the correct part of the above comment by cleaning up the updates of 'info' -- rti_addrs needd not to be updated, rti_info[RTAX_IFP] can be set once outside the loop. While at it, correct a few misspelling of NULL as 0, but there are way too many in this file, and i did not want to clutter the important part of this commit.
2004-04-17 15:09:36 +00:00
len = rt_msg2(rtm->rtm_type, &info, NULL, NULL);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
if (len > rtm->rtm_msglen) {
struct rt_msghdr *new_rtm;
R_Malloc(new_rtm, struct rt_msghdr *, len);
Minor changes to improve code readability (no actual code changes): + replace 0 with NULL where appropriate (not complete) + remove register declaration while there + add argument names to function prototypes to have a better idea of what they are used for + add 'const' qualifiers in 3 places
2004-04-18 00:56:44 +00:00
if (new_rtm == NULL) {
Locking for updates to routing table entries. Each rtentry gets a mutex that covers updates to the contents. Note this is separate from holding a reference and/or locking the routing table itself. Other/related changes: o rtredirect loses the final parameter by which an rtentry reference may be returned; this was never used and added unwarranted complexity for locking. o minor style cleanups to routing code (e.g. ansi-fy function decls) o remove the logic to bump the refcnt on the parent of cloned routes, we assume the parent will remain as long as the clone; doing this avoids a circularity in locking during delete o convert some timeouts to MPSAFE callouts Notes: 1. rt_mtx in struct rtentry is guarded by #ifdef _KERNEL as user-level applications cannot/do-no know about mutex's. Doing this requires that the mutex be the last element in the structure. A better solution is to introduce an externalized version of struct rtentry but this is a major task because of the intertwining of rtentry and other data structures that are visible to user applications. 2. There are known LOR's that are expected to go away with forthcoming work to eliminate many held references. If not these will be resolved prior to release. 3. ATM changes are untested. Sponsored by: FreeBSD Foundation Obtained from: BSD/OS (partly)
2003-10-04 03:44:50 +00:00
RT_UNLOCK(rt);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
senderr(ENOBUFS);
cleanups prior to adding locking (and in some cases to eliminate locking): o move route_cb to be private to rtsock.c o replace global static route_proto by locals o eliminate global #define shorthands for info references o remove some register decls o ansi-fy function decls o move items to be close in scope to their usage o add rt_dispatch function for dispatching the actual message o cleanup tangled logic for doing all-but-me msg send Support by: FreeBSD Foundation
2003-10-03 18:15:54 +00:00
}
+ replace Bcmp/Bzero with 'the real thing' as in the rest of the file. + remember to check and fix or explain a strange cast in route_output()
2004-04-18 11:47:04 +00:00
bcopy(rtm, new_rtm, rtm->rtm_msglen);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
Free(rtm); rtm = new_rtm;
}
misc cleanup in sysctl_ifmalist(): + remove a partly incorrect comment that i introduced in the last commit; + deal with the correct part of the above comment by cleaning up the updates of 'info' -- rti_addrs needd not to be updated, rti_info[RTAX_IFP] can be set once outside the loop. While at it, correct a few misspelling of NULL as 0, but there are way too many in this file, and i did not want to clutter the important part of this commit.
2004-04-17 15:09:36 +00:00
(void)rt_msg2(rtm->rtm_type, &info, (caddr_t)rtm, NULL);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
rtm->rtm_flags = rt->rt_flags;
- Fill in the correct rtm_index for RTM_ADD and RTM_CHANGE messages. - Allow RTM_CHANGE to change a number of route flags as specified by RTF_FMASK. - The unused rtm_use field in struct rt_msghdr is redesignated as rtm_fmask field to communicate route flag changes in RTM_CHANGE messages from userland. The use count of a route was moved to rtm_rmx a long time ago. For source code compatibility reasons a define of rtm_use to rtm_fmask is provided. These changes faciliate running of multiple cooperating routing daemons at the same time without causing undesired interference. Open[BGP|OSPF]D make use of these features to have IGP routes override EGP ones. Obtained from: OpenBSD (claudio@) MFC after: 3 days
2006-03-15 19:39:09 +00:00
rtm->rtm_use = 0;
Introduce tcp_hostcache and remove the tcp specific metrics from the routing table. Move all usage and references in the tcp stack from the routing table metrics to the tcp hostcache. It caches measured parameters of past tcp sessions to provide better initial start values for following connections from or to the same source or destination. Depending on the network parameters to/from the remote host this can lead to significant speedups for new tcp connections after the first one because they inherit and shortcut the learning curve. tcp_hostcache is designed for multiple concurrent access in SMP environments with high contention and is hash indexed by remote ip address. It removes significant locking requirements from the tcp stack with regard to the routing table. Reviewed by: sam (mentor), bms Reviewed by: -net, -current, core@kame.net (IPv6 parts) Approved by: re (scottl)
2003-11-20 20:07:39 +00:00
rt_getmetrics(&rt->rt_rmx, &rtm->rtm_rmx);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
rtm->rtm_addrs = info.rti_addrs;
break;
case RTM_CHANGE:
cleanups prior to adding locking (and in some cases to eliminate locking): o move route_cb to be private to rtsock.c o replace global static route_proto by locals o eliminate global #define shorthands for info references o remove some register decls o ansi-fy function decls o move items to be close in scope to their usage o add rt_dispatch function for dispatching the actual message o cleanup tangled logic for doing all-but-me msg send Support by: FreeBSD Foundation
2003-10-03 18:15:54 +00:00
/*
* New gateway could require new ifaddr, ifp;
* flags may also be different; ifp may be specified
* by ll sockaddr when protocol address is ambiguous
*/
if (((rt->rt_flags & RTF_GATEWAY) &&
info.rti_info[RTAX_GATEWAY] != NULL) ||
info.rti_info[RTAX_IFP] != NULL ||
(info.rti_info[RTAX_IFA] != NULL &&
!sa_equal(info.rti_info[RTAX_IFA],
rt->rt_ifa->ifa_addr))) {
Drop current rtentry lock before calling rt_getifa(). This fixes a LOR and a possible recursive use of rtentry mutex. PR: kern/69356 Reviewed by: sam
2005-09-19 16:27:22 +00:00
RT_UNLOCK(rt);
if ((error = rt_getifa(&info)) != 0)
Only call rt_getifa() if we've either been passed a gateway or if we've been given an RTA_IFP or changed RTA_IFA sockaddr. This fixes the following bug: >/dev/tun100 >/dev/tun101 ifconfig tun100 1.2.3.4 5.6.7.8 ifconfig tun101 1.2.3.4 6.7.8.9 route change 6.7.8.9 -ifa 1.2.3.4 -iface -mtu 500 which erroneously changed tun101's host route to have an ifp of tun100 (rt_getifa() sets the ifp after calling ifa_ifwithnet(1.2.3.4)) This incarnation submitted by: ru
2001-12-19 16:03:27 +00:00
senderr(error);
Drop current rtentry lock before calling rt_getifa(). This fixes a LOR and a possible recursive use of rtentry mutex. PR: kern/69356 Reviewed by: sam
2005-09-19 16:27:22 +00:00
RT_LOCK(rt);
Only call rt_getifa() if we've either been passed a gateway or if we've been given an RTA_IFP or changed RTA_IFA sockaddr. This fixes the following bug: >/dev/tun100 >/dev/tun101 ifconfig tun100 1.2.3.4 5.6.7.8 ifconfig tun101 1.2.3.4 6.7.8.9 route change 6.7.8.9 -ifa 1.2.3.4 -iface -mtu 500 which erroneously changed tun101's host route to have an ifp of tun100 (rt_getifa() sets the ifp after calling ifa_ifwithnet(1.2.3.4)) This incarnation submitted by: ru
2001-12-19 16:03:27 +00:00
}
Fix regression in rev. 1.140. Reported by: Yuriy Tsibizov <Yuriy.Tsibizov gfk.ru>, bsam
2007-03-27 19:36:12 +00:00
if (info.rti_ifa != NULL &&
info.rti_ifa != rt->rt_ifa &&
rt->rt_ifa != NULL &&
When working on an RTM_CHANGE do the route editing in the following sequence. First, if rt_ifa is going to be changed, then call ifa_rtrequest(RTM_DELETE). Second, if gateway is going to be changed, then call rt_setgate(). Third, change rt_ifa. With this change we are able to change a link level route to a gateway one, that wasn't possible before: # ifconfig em0 192.168.22.1/24 # arp -s 192.168.22.99 00:11:22:33:44:55 # route change 192.168.22.99 192.168.22.199 # ping 192.168.22.99 db> Reported by: avatar
2007-03-22 10:51:03 +00:00
rt->rt_ifa->ifa_rtrequest != NULL) {
rt->rt_ifa->ifa_rtrequest(RTM_DELETE, rt,
&info);
IFAFREE(rt->rt_ifa);
}
if (info.rti_info[RTAX_GATEWAY] != NULL) {
if ((error = rt_setgate(rt, rt_key(rt),
cleanups prior to adding locking (and in some cases to eliminate locking): o move route_cb to be private to rtsock.c o replace global static route_proto by locals o eliminate global #define shorthands for info references o remove some register decls o ansi-fy function decls o move items to be close in scope to their usage o add rt_dispatch function for dispatching the actual message o cleanup tangled logic for doing all-but-me msg send Support by: FreeBSD Foundation
2003-10-03 18:15:54 +00:00
info.rti_info[RTAX_GATEWAY])) != 0) {
When working on an RTM_CHANGE do the route editing in the following sequence. First, if rt_ifa is going to be changed, then call ifa_rtrequest(RTM_DELETE). Second, if gateway is going to be changed, then call rt_setgate(). Third, change rt_ifa. With this change we are able to change a link level route to a gateway one, that wasn't possible before: # ifconfig em0 192.168.22.1/24 # arp -s 192.168.22.99 00:11:22:33:44:55 # route change 192.168.22.99 192.168.22.199 # ping 192.168.22.99 db> Reported by: avatar
2007-03-22 10:51:03 +00:00
RT_UNLOCK(rt);
senderr(error);
}
rt->rt_flags |= RTF_GATEWAY;
cleanups prior to adding locking (and in some cases to eliminate locking): o move route_cb to be private to rtsock.c o replace global static route_proto by locals o eliminate global #define shorthands for info references o remove some register decls o ansi-fy function decls o move items to be close in scope to their usage o add rt_dispatch function for dispatching the actual message o cleanup tangled logic for doing all-but-me msg send Support by: FreeBSD Foundation
2003-10-03 18:15:54 +00:00
}
Fix regression in rev. 1.140. Reported by: Yuriy Tsibizov <Yuriy.Tsibizov gfk.ru>, bsam
2007-03-27 19:36:12 +00:00
if (info.rti_ifa != NULL &&
info.rti_ifa != rt->rt_ifa) {
IFAREF(info.rti_ifa);
When working on an RTM_CHANGE do the route editing in the following sequence. First, if rt_ifa is going to be changed, then call ifa_rtrequest(RTM_DELETE). Second, if gateway is going to be changed, then call rt_setgate(). Third, change rt_ifa. With this change we are able to change a link level route to a gateway one, that wasn't possible before: # ifconfig em0 192.168.22.1/24 # arp -s 192.168.22.99 00:11:22:33:44:55 # route change 192.168.22.99 192.168.22.199 # ping 192.168.22.99 db> Reported by: avatar
2007-03-22 10:51:03 +00:00
rt->rt_ifa = info.rti_ifa;
Fix regression in rev. 1.140. Reported by: Yuriy Tsibizov <Yuriy.Tsibizov gfk.ru>, bsam
2007-03-27 19:36:12 +00:00
rt->rt_ifp = info.rti_ifp;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
- Fill in the correct rtm_index for RTM_ADD and RTM_CHANGE messages. - Allow RTM_CHANGE to change a number of route flags as specified by RTF_FMASK. - The unused rtm_use field in struct rt_msghdr is redesignated as rtm_fmask field to communicate route flag changes in RTM_CHANGE messages from userland. The use count of a route was moved to rtm_rmx a long time ago. For source code compatibility reasons a define of rtm_use to rtm_fmask is provided. These changes faciliate running of multiple cooperating routing daemons at the same time without causing undesired interference. Open[BGP|OSPF]D make use of these features to have IGP routes override EGP ones. Obtained from: OpenBSD (claudio@) MFC after: 3 days
2006-03-15 19:39:09 +00:00
/* Allow some flags to be toggled on change. */
if (rtm->rtm_fmask & RTF_FMASK)
rt->rt_flags = (rt->rt_flags &
~rtm->rtm_fmask) |
(rtm->rtm_flags & rtm->rtm_fmask);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
rt_setmetrics(rtm->rtm_inits, &rtm->rtm_rmx,
&rt->rt_rmx);
- Fill in the correct rtm_index for RTM_ADD and RTM_CHANGE messages. - Allow RTM_CHANGE to change a number of route flags as specified by RTF_FMASK. - The unused rtm_use field in struct rt_msghdr is redesignated as rtm_fmask field to communicate route flag changes in RTM_CHANGE messages from userland. The use count of a route was moved to rtm_rmx a long time ago. For source code compatibility reasons a define of rtm_use to rtm_fmask is provided. These changes faciliate running of multiple cooperating routing daemons at the same time without causing undesired interference. Open[BGP|OSPF]D make use of these features to have IGP routes override EGP ones. Obtained from: OpenBSD (claudio@) MFC after: 3 days
2006-03-15 19:39:09 +00:00
rtm->rtm_index = rt->rt_ifp->if_index;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
if (rt->rt_ifa && rt->rt_ifa->ifa_rtrequest)
Pull post-4.4BSD change to sys/net/route.c from BSD/OS 4.2. Have sys/net/route.c:rtrequest1(), which takes ``rt_addrinfo *'' as the argument. Pass rt_addrinfo all the way down to rtrequest1 and ifa->ifa_rtrequest. 3rd argument of ifa->ifa_rtrequest is now ``rt_addrinfo *'' instead of ``sockaddr *'' (almost noone is using it anyways). Benefit: the following command now works. Previously we needed two route(8) invocations, "add" then "change". # route add -inet6 default ::1 -ifp gif0 Remove unsafe typecast in rtrequest(), from ``rtentry *'' to ``sockaddr *''. It was introduced by 4.3BSD-Reno and never corrected. Obtained from: BSD/OS, NetBSD MFC after: 1 month PR: kern/28360
2001-10-17 18:07:05 +00:00
rt->rt_ifa->ifa_rtrequest(RTM_ADD, rt, &info);
cleanups prior to adding locking (and in some cases to eliminate locking): o move route_cb to be private to rtsock.c o replace global static route_proto by locals o eliminate global #define shorthands for info references o remove some register decls o ansi-fy function decls o move items to be close in scope to their usage o add rt_dispatch function for dispatching the actual message o cleanup tangled logic for doing all-but-me msg send Support by: FreeBSD Foundation
2003-10-03 18:15:54 +00:00
if (info.rti_info[RTAX_GENMASK])
rt->rt_genmask = info.rti_info[RTAX_GENMASK];
Replace various spelling with FALLTHROUGH which is lint()able
2002-08-25 13:23:09 +00:00
/* FALLTHROUGH */
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
case RTM_LOCK:
Introduce tcp_hostcache and remove the tcp specific metrics from the routing table. Move all usage and references in the tcp stack from the routing table metrics to the tcp hostcache. It caches measured parameters of past tcp sessions to provide better initial start values for following connections from or to the same source or destination. Depending on the network parameters to/from the remote host this can lead to significant speedups for new tcp connections after the first one because they inherit and shortcut the learning curve. tcp_hostcache is designed for multiple concurrent access in SMP environments with high contention and is hash indexed by remote ip address. It removes significant locking requirements from the tcp stack with regard to the routing table. Reviewed by: sam (mentor), bms Reviewed by: -net, -current, core@kame.net (IPv6 parts) Approved by: re (scottl)
2003-11-20 20:07:39 +00:00
/* We don't support locks anymore */
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
break;
}
Locking for updates to routing table entries. Each rtentry gets a mutex that covers updates to the contents. Note this is separate from holding a reference and/or locking the routing table itself. Other/related changes: o rtredirect loses the final parameter by which an rtentry reference may be returned; this was never used and added unwarranted complexity for locking. o minor style cleanups to routing code (e.g. ansi-fy function decls) o remove the logic to bump the refcnt on the parent of cloned routes, we assume the parent will remain as long as the clone; doing this avoids a circularity in locking during delete o convert some timeouts to MPSAFE callouts Notes: 1. rt_mtx in struct rtentry is guarded by #ifdef _KERNEL as user-level applications cannot/do-no know about mutex's. Doing this requires that the mutex be the last element in the structure. A better solution is to introduce an externalized version of struct rtentry but this is a major task because of the intertwining of rtentry and other data structures that are visible to user applications. 2. There are known LOR's that are expected to go away with forthcoming work to eliminate many held references. If not these will be resolved prior to release. 3. ATM changes are untested. Sponsored by: FreeBSD Foundation Obtained from: BSD/OS (partly)
2003-10-04 03:44:50 +00:00
RT_UNLOCK(rt);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
break;
default:
senderr(EOPNOTSUPP);
}
flush:
if (rtm) {
if (error)
rtm->rtm_errno = error;
Remove trailing whitespace.
1995-05-30 08:16:23 +00:00
else
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
rtm->rtm_flags |= RTF_DONE;
}
cleanups prior to adding locking (and in some cases to eliminate locking): o move route_cb to be private to rtsock.c o replace global static route_proto by locals o eliminate global #define shorthands for info references o remove some register decls o ansi-fy function decls o move items to be close in scope to their usage o add rt_dispatch function for dispatching the actual message o cleanup tangled logic for doing all-but-me msg send Support by: FreeBSD Foundation
2003-10-03 18:15:54 +00:00
if (rt) /* XXX can this be true? */
RTFREE(rt);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
{
Minor changes to improve code readability (no actual code changes): + replace 0 with NULL where appropriate (not complete) + remove register declaration while there + add argument names to function prototypes to have a better idea of what they are used for + add 'const' qualifiers in 3 places
2004-04-18 00:56:44 +00:00
struct rawcb *rp = NULL;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
/*
* Check to see if we don't want our own messages.
*/
if ((so->so_options & SO_USELOOPBACK) == 0) {
if (route_cb.any_count <= 1) {
if (rtm)
Free(rtm);
m_freem(m);
return (error);
}
/* There is another listener, so construct message */
rp = sotorawcb(so);
Back out my lats commit of locking down a socket, it conflicts with hsu's work. Requested by: hsu
2002-05-31 11:52:35 +00:00
}
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
if (rtm) {
m_copyback(m, 0, rtm->rtm_msglen, (caddr_t)rtm);
adjust mbuf length right in route_output(). Obtained from: KAME MFC after: 1 week
2001-07-04 11:29:55 +00:00
if (m->m_pkthdr.len < rtm->rtm_msglen) {
m_freem(m);
m = NULL;
} else if (m->m_pkthdr.len > rtm->rtm_msglen)
m_adj(m, rtm->rtm_msglen - m->m_pkthdr.len);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
Free(rtm);
}
cleanups prior to adding locking (and in some cases to eliminate locking): o move route_cb to be private to rtsock.c o replace global static route_proto by locals o eliminate global #define shorthands for info references o remove some register decls o ansi-fy function decls o move items to be close in scope to their usage o add rt_dispatch function for dispatching the actual message o cleanup tangled logic for doing all-but-me msg send Support by: FreeBSD Foundation
2003-10-03 18:15:54 +00:00
if (m) {
if (rp) {
/*
* XXX insure we don't get a copy by
* invalidating our protocol
*/
unsigned short family = rp->rcb_proto.sp_family;
rp->rcb_proto.sp_family = 0;
rt_dispatch(m, info.rti_info[RTAX_DST]);
rp->rcb_proto.sp_family = family;
} else
rt_dispatch(m, info.rti_info[RTAX_DST]);
}
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
return (error);
cleanups prior to adding locking (and in some cases to eliminate locking): o move route_cb to be private to rtsock.c o replace global static route_proto by locals o eliminate global #define shorthands for info references o remove some register decls o ansi-fy function decls o move items to be close in scope to their usage o add rt_dispatch function for dispatching the actual message o cleanup tangled logic for doing all-but-me msg send Support by: FreeBSD Foundation
2003-10-03 18:15:54 +00:00
#undef sa_equal
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
All net.* sysctl converted now.
1995-11-16 19:00:27 +00:00
static void
Minor changes to improve code readability (no actual code changes): + replace 0 with NULL where appropriate (not complete) + remove register declaration while there + add argument names to function prototypes to have a better idea of what they are used for + add 'const' qualifiers in 3 places
2004-04-18 00:56:44 +00:00
rt_setmetrics(u_long which, const struct rt_metrics *in,
struct rt_metrics_lite *out)
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
{
#define metric(f, e) if (which & (f)) out->e = in->e;
Introduce tcp_hostcache and remove the tcp specific metrics from the routing table. Move all usage and references in the tcp stack from the routing table metrics to the tcp hostcache. It caches measured parameters of past tcp sessions to provide better initial start values for following connections from or to the same source or destination. Depending on the network parameters to/from the remote host this can lead to significant speedups for new tcp connections after the first one because they inherit and shortcut the learning curve. tcp_hostcache is designed for multiple concurrent access in SMP environments with high contention and is hash indexed by remote ip address. It removes significant locking requirements from the tcp stack with regard to the routing table. Reviewed by: sam (mentor), bms Reviewed by: -net, -current, core@kame.net (IPv6 parts) Approved by: re (scottl)
2003-11-20 20:07:39 +00:00
/*
* Only these are stored in the routing entry since introduction
* of tcp hostcache. The rest is ignored.
*/
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
metric(RTV_MTU, rmx_mtu);
Adjust rt_(set|get)metrics() to do kernel <-> userland timebase conversion. We need it since kernel timebase has changed (time_second -> time_uptime). Approved by: glebius (mentor)
2006-07-06 00:24:36 +00:00
/* Userland -> kernel timebase conversion. */
if (which & RTV_EXPIRE)
out->rmx_expire = in->rmx_expire ?
in->rmx_expire - time_second + time_uptime : 0;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
#undef metric
}
Introduce tcp_hostcache and remove the tcp specific metrics from the routing table. Move all usage and references in the tcp stack from the routing table metrics to the tcp hostcache. It caches measured parameters of past tcp sessions to provide better initial start values for following connections from or to the same source or destination. Depending on the network parameters to/from the remote host this can lead to significant speedups for new tcp connections after the first one because they inherit and shortcut the learning curve. tcp_hostcache is designed for multiple concurrent access in SMP environments with high contention and is hash indexed by remote ip address. It removes significant locking requirements from the tcp stack with regard to the routing table. Reviewed by: sam (mentor), bms Reviewed by: -net, -current, core@kame.net (IPv6 parts) Approved by: re (scottl)
2003-11-20 20:07:39 +00:00
static void
Minor changes to improve code readability (no actual code changes): + replace 0 with NULL where appropriate (not complete) + remove register declaration while there + add argument names to function prototypes to have a better idea of what they are used for + add 'const' qualifiers in 3 places
2004-04-18 00:56:44 +00:00
rt_getmetrics(const struct rt_metrics_lite *in, struct rt_metrics *out)
Introduce tcp_hostcache and remove the tcp specific metrics from the routing table. Move all usage and references in the tcp stack from the routing table metrics to the tcp hostcache. It caches measured parameters of past tcp sessions to provide better initial start values for following connections from or to the same source or destination. Depending on the network parameters to/from the remote host this can lead to significant speedups for new tcp connections after the first one because they inherit and shortcut the learning curve. tcp_hostcache is designed for multiple concurrent access in SMP environments with high contention and is hash indexed by remote ip address. It removes significant locking requirements from the tcp stack with regard to the routing table. Reviewed by: sam (mentor), bms Reviewed by: -net, -current, core@kame.net (IPv6 parts) Approved by: re (scottl)
2003-11-20 20:07:39 +00:00
{
#define metric(e) out->e = in->e;
bzero(out, sizeof(*out));
metric(rmx_mtu);
Adjust rt_(set|get)metrics() to do kernel <-> userland timebase conversion. We need it since kernel timebase has changed (time_second -> time_uptime). Approved by: glebius (mentor)
2006-07-06 00:24:36 +00:00
/* Kernel -> userland timebase conversion. */
out->rmx_expire = in->rmx_expire ?
in->rmx_expire - time_uptime + time_second : 0;
Introduce tcp_hostcache and remove the tcp specific metrics from the routing table. Move all usage and references in the tcp stack from the routing table metrics to the tcp hostcache. It caches measured parameters of past tcp sessions to provide better initial start values for following connections from or to the same source or destination. Depending on the network parameters to/from the remote host this can lead to significant speedups for new tcp connections after the first one because they inherit and shortcut the learning curve. tcp_hostcache is designed for multiple concurrent access in SMP environments with high contention and is hash indexed by remote ip address. It removes significant locking requirements from the tcp stack with regard to the routing table. Reviewed by: sam (mentor), bms Reviewed by: -net, -current, core@kame.net (IPv6 parts) Approved by: re (scottl)
2003-11-20 20:07:39 +00:00
#undef metric
}
Finally track down the reason for some of my occasional kernel crashes. Route(1) has a bug that sends a bad message to the kernel. The kernel trusts it and crashes. Add some sanity checks so that we don't trust the user quite as much any more. (also add a comment in if_ethersubr.c)
1997-07-15 23:25:32 +00:00
/*
* Extract the addresses of the passed sockaddrs.
* Do a little sanity checking so as to avoid bad memory references.
An actual fix for the routing default crashes that 1/ is compatible with the old route(1) in case needed. 2/ actually fixes the problem while vetting bad user input. note: I have already fixed route(1) so the problem shouldn't occur. if it does. use 0.0.0.0/0 instead of the word 'default' :)
1997-07-18 11:44:24 +00:00
* This data is derived straight from userland.
Finally track down the reason for some of my occasional kernel crashes. Route(1) has a bug that sends a bad message to the kernel. The kernel trusts it and crashes. Add some sanity checks so that we don't trust the user quite as much any more. (also add a comment in if_ethersubr.c)
1997-07-15 23:25:32 +00:00
*/
An actual fix for the routing default crashes that 1/ is compatible with the old route(1) in case needed. 2/ actually fixes the problem while vetting bad user input. note: I have already fixed route(1) so the problem shouldn't occur. if it does. use 0.0.0.0/0 instead of the word 'default' :)
1997-07-18 11:44:24 +00:00
static int
cleanups prior to adding locking (and in some cases to eliminate locking): o move route_cb to be private to rtsock.c o replace global static route_proto by locals o eliminate global #define shorthands for info references o remove some register decls o ansi-fy function decls o move items to be close in scope to their usage o add rt_dispatch function for dispatching the actual message o cleanup tangled logic for doing all-but-me msg send Support by: FreeBSD Foundation
2003-10-03 18:15:54 +00:00
rt_xaddrs(caddr_t cp, caddr_t cplim, struct rt_addrinfo *rtinfo)
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
{
route.h: introduce a macro, SA_SIZE(struct sockaddr *) which returns the space occupied by a struct sockaddr when passed through a routing socket. Use it to replace the macro ROUNDUP(int), that does the same but is redefined by every file which uses it, courtesy of the School of Cut'n'Paste Programming(TM). (partial) userland changes to follow.
2004-04-13 11:22:22 +00:00
struct sockaddr *sa;
int i;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
cleanups prior to adding locking (and in some cases to eliminate locking): o move route_cb to be private to rtsock.c o replace global static route_proto by locals o eliminate global #define shorthands for info references o remove some register decls o ansi-fy function decls o move items to be close in scope to their usage o add rt_dispatch function for dispatching the actual message o cleanup tangled logic for doing all-but-me msg send Support by: FreeBSD Foundation
2003-10-03 18:15:54 +00:00
for (i = 0; i < RTAX_MAX && cp < cplim; i++) {
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
if ((rtinfo->rti_addrs & (1 << i)) == 0)
continue;
Bungled cut/paste leaves kernel with page faults.. (read all about it!)
1997-07-16 14:55:14 +00:00
sa = (struct sockaddr *)cp;
Finally track down the reason for some of my occasional kernel crashes. Route(1) has a bug that sends a bad message to the kernel. The kernel trusts it and crashes. Add some sanity checks so that we don't trust the user quite as much any more. (also add a comment in if_ethersubr.c)
1997-07-15 23:25:32 +00:00
/*
An actual fix for the routing default crashes that 1/ is compatible with the old route(1) in case needed. 2/ actually fixes the problem while vetting bad user input. note: I have already fixed route(1) so the problem shouldn't occur. if it does. use 0.0.0.0/0 instead of the word 'default' :)
1997-07-18 11:44:24 +00:00
* It won't fit.
Finally track down the reason for some of my occasional kernel crashes. Route(1) has a bug that sends a bad message to the kernel. The kernel trusts it and crashes. Add some sanity checks so that we don't trust the user quite as much any more. (also add a comment in if_ethersubr.c)
1997-07-15 23:25:32 +00:00
*/
cleanups prior to adding locking (and in some cases to eliminate locking): o move route_cb to be private to rtsock.c o replace global static route_proto by locals o eliminate global #define shorthands for info references o remove some register decls o ansi-fy function decls o move items to be close in scope to their usage o add rt_dispatch function for dispatching the actual message o cleanup tangled logic for doing all-but-me msg send Support by: FreeBSD Foundation
2003-10-03 18:15:54 +00:00
if (cp + sa->sa_len > cplim)
An actual fix for the routing default crashes that 1/ is compatible with the old route(1) in case needed. 2/ actually fixes the problem while vetting bad user input. note: I have already fixed route(1) so the problem shouldn't occur. if it does. use 0.0.0.0/0 instead of the word 'default' :)
1997-07-18 11:44:24 +00:00
return (EINVAL);
Finally track down the reason for some of my occasional kernel crashes. Route(1) has a bug that sends a bad message to the kernel. The kernel trusts it and crashes. Add some sanity checks so that we don't trust the user quite as much any more. (also add a comment in if_ethersubr.c)
1997-07-15 23:25:32 +00:00
/*
* there are no more.. quit now
* If there are more bits, they are in error.
* I've seen this. route(1) can evidently generate these.
* This causes kernel to core dump.
An actual fix for the routing default crashes that 1/ is compatible with the old route(1) in case needed. 2/ actually fixes the problem while vetting bad user input. note: I have already fixed route(1) so the problem shouldn't occur. if it does. use 0.0.0.0/0 instead of the word 'default' :)
1997-07-18 11:44:24 +00:00
* for compatibility, If we see this, point to a safe address.
Finally track down the reason for some of my occasional kernel crashes. Route(1) has a bug that sends a bad message to the kernel. The kernel trusts it and crashes. Add some sanity checks so that we don't trust the user quite as much any more. (also add a comment in if_ethersubr.c)
1997-07-15 23:25:32 +00:00
*/
An actual fix for the routing default crashes that 1/ is compatible with the old route(1) in case needed. 2/ actually fixes the problem while vetting bad user input. note: I have already fixed route(1) so the problem shouldn't occur. if it does. use 0.0.0.0/0 instead of the word 'default' :)
1997-07-18 11:44:24 +00:00
if (sa->sa_len == 0) {
rtinfo->rti_info[i] = &sa_zero;
return (0); /* should be EINVAL but for compat */
}
/* accept it */
rtinfo->rti_info[i] = sa;
route.h: introduce a macro, SA_SIZE(struct sockaddr *) which returns the space occupied by a struct sockaddr when passed through a routing socket. Use it to replace the macro ROUNDUP(int), that does the same but is redefined by every file which uses it, courtesy of the School of Cut'n'Paste Programming(TM). (partial) userland changes to follow.
2004-04-13 11:22:22 +00:00
cp += SA_SIZE(sa);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
An actual fix for the routing default crashes that 1/ is compatible with the old route(1) in case needed. 2/ actually fixes the problem while vetting bad user input. note: I have already fixed route(1) so the problem shouldn't occur. if it does. use 0.0.0.0/0 instead of the word 'default' :)
1997-07-18 11:44:24 +00:00
return (0);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
static struct mbuf *
cleanups prior to adding locking (and in some cases to eliminate locking): o move route_cb to be private to rtsock.c o replace global static route_proto by locals o eliminate global #define shorthands for info references o remove some register decls o ansi-fy function decls o move items to be close in scope to their usage o add rt_dispatch function for dispatching the actual message o cleanup tangled logic for doing all-but-me msg send Support by: FreeBSD Foundation
2003-10-03 18:15:54 +00:00
rt_msg1(int type, struct rt_addrinfo *rtinfo)
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
{
Minor changes to improve code readability (no actual code changes): + replace 0 with NULL where appropriate (not complete) + remove register declaration while there + add argument names to function prototypes to have a better idea of what they are used for + add 'const' qualifiers in 3 places
2004-04-18 00:56:44 +00:00
struct rt_msghdr *rtm;
struct mbuf *m;
int i;
struct sockaddr *sa;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
int len, dlen;
switch (type) {
case RTM_DELADDR:
case RTM_NEWADDR:
len = sizeof(struct ifa_msghdr);
break;
Use the new if_multiaddrs list for multicast addresses rather than the previous hackery involving struct in_ifaddr and arpcom. Get rid of the abominable multi_kludge. Update all network interfaces to use the new machanism. Distressingly few Ethernet drivers program the multicast filter properly (assuming the hardware has one, which it usually does).
1997-01-13 21:26:53 +00:00
case RTM_DELMADDR:
case RTM_NEWMADDR:
len = sizeof(struct ifma_msghdr);
break;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
case RTM_IFINFO:
len = sizeof(struct if_msghdr);
break;
Introduce an interface announcement message for the routing socket so that routing daemons and other interested parties know when an interface is attached/detached. PR: kern/33747 Obtained from: NetBSD MFC after: 2 weeks
2002-01-18 14:33:04 +00:00
case RTM_IFANNOUNCE:
Add 802.11-specific events that are dispatched through the routing socket. This really doesn't belong here but is preferred (for the moment) over adding yet another mechanism for sending msgs from the kernel to user apps. Reviewed by: imp
2004-10-05 19:48:33 +00:00
case RTM_IEEE80211:
Introduce an interface announcement message for the routing socket so that routing daemons and other interested parties know when an interface is attached/detached. PR: kern/33747 Obtained from: NetBSD MFC after: 2 weeks
2002-01-18 14:33:04 +00:00
len = sizeof(struct if_announcemsghdr);
break;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
default:
len = sizeof(struct rt_msghdr);
}
Sync with recent KAME. This work was based on kame-20010528-freebsd43-snap.tgz and some critical problem after the snap was out were fixed. There are many many changes since last KAME merge. TODO: - The definitions of SADB_* in sys/net/pfkeyv2.h are still different from RFC2407/IANA assignment because of binary compatibility issue. It should be fixed under 5-CURRENT. - ip6po_m member of struct ip6_pktopts is no longer used. But, it is still there because of binary compatibility issue. It should be removed under 5-CURRENT. Reviewed by: itojun Obtained from: KAME MFC after: 3 weeks
2001-06-11 12:39:29 +00:00
if (len > MCLBYTES)
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
panic("rt_msg1");
Back out M_* changes, per decision of the TRB. Approved by: trb
2003-02-19 05:47:46 +00:00
m = m_gethdr(M_DONTWAIT, MT_DATA);
Sync with recent KAME. This work was based on kame-20010528-freebsd43-snap.tgz and some critical problem after the snap was out were fixed. There are many many changes since last KAME merge. TODO: - The definitions of SADB_* in sys/net/pfkeyv2.h are still different from RFC2407/IANA assignment because of binary compatibility issue. It should be fixed under 5-CURRENT. - ip6po_m member of struct ip6_pktopts is no longer used. But, it is still there because of binary compatibility issue. It should be removed under 5-CURRENT. Reviewed by: itojun Obtained from: KAME MFC after: 3 weeks
2001-06-11 12:39:29 +00:00
if (m && len > MHLEN) {
Back out M_* changes, per decision of the TRB. Approved by: trb
2003-02-19 05:47:46 +00:00
MCLGET(m, M_DONTWAIT);
Sync with recent KAME. This work was based on kame-20010528-freebsd43-snap.tgz and some critical problem after the snap was out were fixed. There are many many changes since last KAME merge. TODO: - The definitions of SADB_* in sys/net/pfkeyv2.h are still different from RFC2407/IANA assignment because of binary compatibility issue. It should be fixed under 5-CURRENT. - ip6po_m member of struct ip6_pktopts is no longer used. But, it is still there because of binary compatibility issue. It should be removed under 5-CURRENT. Reviewed by: itojun Obtained from: KAME MFC after: 3 weeks
2001-06-11 12:39:29 +00:00
if ((m->m_flags & M_EXT) == 0) {
m_free(m);
m = NULL;
}
}
Minor changes to improve code readability (no actual code changes): + replace 0 with NULL where appropriate (not complete) + remove register declaration while there + add argument names to function prototypes to have a better idea of what they are used for + add 'const' qualifiers in 3 places
2004-04-18 00:56:44 +00:00
if (m == NULL)
Sync with recent KAME. This work was based on kame-20010528-freebsd43-snap.tgz and some critical problem after the snap was out were fixed. There are many many changes since last KAME merge. TODO: - The definitions of SADB_* in sys/net/pfkeyv2.h are still different from RFC2407/IANA assignment because of binary compatibility issue. It should be fixed under 5-CURRENT. - ip6po_m member of struct ip6_pktopts is no longer used. But, it is still there because of binary compatibility issue. It should be removed under 5-CURRENT. Reviewed by: itojun Obtained from: KAME MFC after: 3 weeks
2001-06-11 12:39:29 +00:00
return (m);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
m->m_pkthdr.len = m->m_len = len;
Minor changes to improve code readability (no actual code changes): + replace 0 with NULL where appropriate (not complete) + remove register declaration while there + add argument names to function prototypes to have a better idea of what they are used for + add 'const' qualifiers in 3 places
2004-04-18 00:56:44 +00:00
m->m_pkthdr.rcvif = NULL;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
rtm = mtod(m, struct rt_msghdr *);
bzero((caddr_t)rtm, len);
for (i = 0; i < RTAX_MAX; i++) {
if ((sa = rtinfo->rti_info[i]) == NULL)
continue;
rtinfo->rti_addrs |= (1 << i);
route.h: introduce a macro, SA_SIZE(struct sockaddr *) which returns the space occupied by a struct sockaddr when passed through a routing socket. Use it to replace the macro ROUNDUP(int), that does the same but is redefined by every file which uses it, courtesy of the School of Cut'n'Paste Programming(TM). (partial) userland changes to follow.
2004-04-13 11:22:22 +00:00
dlen = SA_SIZE(sa);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
m_copyback(m, len, dlen, (caddr_t)sa);
len += dlen;
}
if (m->m_pkthdr.len != len) {
m_freem(m);
return (NULL);
}
rtm->rtm_msglen = len;
rtm->rtm_version = RTM_VERSION;
rtm->rtm_type = type;
return (m);
}
static int
cleanups prior to adding locking (and in some cases to eliminate locking): o move route_cb to be private to rtsock.c o replace global static route_proto by locals o eliminate global #define shorthands for info references o remove some register decls o ansi-fy function decls o move items to be close in scope to their usage o add rt_dispatch function for dispatching the actual message o cleanup tangled logic for doing all-but-me msg send Support by: FreeBSD Foundation
2003-10-03 18:15:54 +00:00
rt_msg2(int type, struct rt_addrinfo *rtinfo, caddr_t cp, struct walkarg *w)
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
{
Minor changes to improve code readability (no actual code changes): + replace 0 with NULL where appropriate (not complete) + remove register declaration while there + add argument names to function prototypes to have a better idea of what they are used for + add 'const' qualifiers in 3 places
2004-04-18 00:56:44 +00:00
int i;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
int len, dlen, second_time = 0;
caddr_t cp0;
rtinfo->rti_addrs = 0;
again:
switch (type) {
case RTM_DELADDR:
case RTM_NEWADDR:
len = sizeof(struct ifa_msghdr);
break;
case RTM_IFINFO:
len = sizeof(struct if_msghdr);
break;
Add a sysctl MIB, NET_RT_IFMALIST, to retrieve multicast group memberships in a protocol-independent way. Submitted by: harti
2003-11-14 18:48:15 +00:00
case RTM_NEWMADDR:
len = sizeof(struct ifma_msghdr);
break;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
default:
len = sizeof(struct rt_msghdr);
}
Cosmetics: to silence gcc -wall.
1994-10-08 22:38:27 +00:00
cp0 = cp;
if (cp0)
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
cp += len;
for (i = 0; i < RTAX_MAX; i++) {
Minor changes to improve code readability (no actual code changes): + replace 0 with NULL where appropriate (not complete) + remove register declaration while there + add argument names to function prototypes to have a better idea of what they are used for + add 'const' qualifiers in 3 places
2004-04-18 00:56:44 +00:00
struct sockaddr *sa;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
Minor changes to improve code readability (no actual code changes): + replace 0 with NULL where appropriate (not complete) + remove register declaration while there + add argument names to function prototypes to have a better idea of what they are used for + add 'const' qualifiers in 3 places
2004-04-18 00:56:44 +00:00
if ((sa = rtinfo->rti_info[i]) == NULL)
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
continue;
rtinfo->rti_addrs |= (1 << i);
route.h: introduce a macro, SA_SIZE(struct sockaddr *) which returns the space occupied by a struct sockaddr when passed through a routing socket. Use it to replace the macro ROUNDUP(int), that does the same but is redefined by every file which uses it, courtesy of the School of Cut'n'Paste Programming(TM). (partial) userland changes to follow.
2004-04-13 11:22:22 +00:00
dlen = SA_SIZE(sa);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
if (cp) {
bcopy((caddr_t)sa, cp, (unsigned)dlen);
cp += dlen;
}
len += dlen;
}
Prevent the kernel from generating an unaligned sysctl data buffer on 64-bit platforms. The unaligned access is caused by struct ifa_msghdr not being a multiple of 8-bytes in size. If an interface has an odd number of addresses, this causes the next interface to generate an unaligned access in the user-level app walking the interfaces (ifconfig). Submitted by: Bernd Walter <ticso@cicely8.cicely.de>
2002-01-27 20:39:01 +00:00
len = ALIGN(len);
Minor changes to improve code readability (no actual code changes): + replace 0 with NULL where appropriate (not complete) + remove register declaration while there + add argument names to function prototypes to have a better idea of what they are used for + add 'const' qualifiers in 3 places
2004-04-18 00:56:44 +00:00
if (cp == NULL && w != NULL && !second_time) {
struct walkarg *rw = w;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
All net.* sysctl converted now.
1995-11-16 19:00:27 +00:00
if (rw->w_req) {
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
if (rw->w_tmemsize < len) {
if (rw->w_tmem)
free(rw->w_tmem, M_RTABLE);
Cosmetics: to silence gcc -wall.
1994-10-08 22:38:27 +00:00
rw->w_tmem = (caddr_t)
malloc(len, M_RTABLE, M_NOWAIT);
if (rw->w_tmem)
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
rw->w_tmemsize = len;
}
if (rw->w_tmem) {
cp = rw->w_tmem;
second_time = 1;
goto again;
All net.* sysctl converted now.
1995-11-16 19:00:27 +00:00
}
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
}
if (cp) {
Minor changes to improve code readability (no actual code changes): + replace 0 with NULL where appropriate (not complete) + remove register declaration while there + add argument names to function prototypes to have a better idea of what they are used for + add 'const' qualifiers in 3 places
2004-04-18 00:56:44 +00:00
struct rt_msghdr *rtm = (struct rt_msghdr *)cp0;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
rtm->rtm_version = RTM_VERSION;
rtm->rtm_type = type;
rtm->rtm_msglen = len;
}
return (len);
}
/*
* This routine is called to generate a message from the routing
* socket indicating that a redirect has occured, a routing lookup
* has failed, or that a protocol has detected timeouts to a particular
* destination.
*/
void
cleanups prior to adding locking (and in some cases to eliminate locking): o move route_cb to be private to rtsock.c o replace global static route_proto by locals o eliminate global #define shorthands for info references o remove some register decls o ansi-fy function decls o move items to be close in scope to their usage o add rt_dispatch function for dispatching the actual message o cleanup tangled logic for doing all-but-me msg send Support by: FreeBSD Foundation
2003-10-03 18:15:54 +00:00
rt_missmsg(int type, struct rt_addrinfo *rtinfo, int flags, int error)
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
{
cleanups prior to adding locking (and in some cases to eliminate locking): o move route_cb to be private to rtsock.c o replace global static route_proto by locals o eliminate global #define shorthands for info references o remove some register decls o ansi-fy function decls o move items to be close in scope to their usage o add rt_dispatch function for dispatching the actual message o cleanup tangled logic for doing all-but-me msg send Support by: FreeBSD Foundation
2003-10-03 18:15:54 +00:00
struct rt_msghdr *rtm;
struct mbuf *m;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
struct sockaddr *sa = rtinfo->rti_info[RTAX_DST];
if (route_cb.any_count == 0)
return;
m = rt_msg1(type, rtinfo);
Minor changes to improve code readability (no actual code changes): + replace 0 with NULL where appropriate (not complete) + remove register declaration while there + add argument names to function prototypes to have a better idea of what they are used for + add 'const' qualifiers in 3 places
2004-04-18 00:56:44 +00:00
if (m == NULL)
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
return;
rtm = mtod(m, struct rt_msghdr *);
rtm->rtm_flags = RTF_DONE | flags;
rtm->rtm_errno = error;
rtm->rtm_addrs = rtinfo->rti_addrs;
cleanups prior to adding locking (and in some cases to eliminate locking): o move route_cb to be private to rtsock.c o replace global static route_proto by locals o eliminate global #define shorthands for info references o remove some register decls o ansi-fy function decls o move items to be close in scope to their usage o add rt_dispatch function for dispatching the actual message o cleanup tangled logic for doing all-but-me msg send Support by: FreeBSD Foundation
2003-10-03 18:15:54 +00:00
rt_dispatch(m, sa);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
/*
* This routine is called to generate a message from the routing
* socket indicating that the status of a network interface has changed.
*/
void
cleanups prior to adding locking (and in some cases to eliminate locking): o move route_cb to be private to rtsock.c o replace global static route_proto by locals o eliminate global #define shorthands for info references o remove some register decls o ansi-fy function decls o move items to be close in scope to their usage o add rt_dispatch function for dispatching the actual message o cleanup tangled logic for doing all-but-me msg send Support by: FreeBSD Foundation
2003-10-03 18:15:54 +00:00
rt_ifmsg(struct ifnet *ifp)
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
{
cleanups prior to adding locking (and in some cases to eliminate locking): o move route_cb to be private to rtsock.c o replace global static route_proto by locals o eliminate global #define shorthands for info references o remove some register decls o ansi-fy function decls o move items to be close in scope to their usage o add rt_dispatch function for dispatching the actual message o cleanup tangled logic for doing all-but-me msg send Support by: FreeBSD Foundation
2003-10-03 18:15:54 +00:00
struct if_msghdr *ifm;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
struct mbuf *m;
struct rt_addrinfo info;
if (route_cb.any_count == 0)
return;
bzero((caddr_t)&info, sizeof(info));
m = rt_msg1(RTM_IFINFO, &info);
Minor changes to improve code readability (no actual code changes): + replace 0 with NULL where appropriate (not complete) + remove register declaration while there + add argument names to function prototypes to have a better idea of what they are used for + add 'const' qualifiers in 3 places
2004-04-18 00:56:44 +00:00
if (m == NULL)
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
return;
ifm = mtod(m, struct if_msghdr *);
ifm->ifm_index = ifp->if_index;
Rename IFF_RUNNING to IFF_DRV_RUNNING, IFF_OACTIVE to IFF_DRV_OACTIVE, and move both flags from ifnet.if_flags to ifnet.if_drv_flags, making and documenting the locking of these flags the responsibility of the device driver, not the network stack. The flags for these two fields will be mutually exclusive so that they can be exposed to user space as though they were stored in the same variable. Provide #defines to provide the old names #ifndef _KERNEL, so that user applications (such as ifconfig) can use the old flag names. Using the old names in a device driver will result in a compile error in order to help device driver writers adopt the new model. When exposing the interface flags to user space, via interface ioctls or routing sockets, or the two fields together. Since the driver flags cannot currently be set for user space, no new logic is currently required to handle this case. Add some assertions that general purpose network stack routines, such as if_setflags(), are not improperly used on driver-owned flags. With this change, a large number of very minor network stack races are closed, subject to correct device driver locking. Most were likely never triggered. Driver sweep to follow; many thanks to pjd and bz for the line-by-line review they gave this patch. Reviewed by: pjd, bz MFC after: 7 days
2005-08-09 10:16:17 +00:00
ifm->ifm_flags = ifp->if_flags | ifp->if_drv_flags;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
ifm->ifm_data = ifp->if_data;
ifm->ifm_addrs = 0;
cleanups prior to adding locking (and in some cases to eliminate locking): o move route_cb to be private to rtsock.c o replace global static route_proto by locals o eliminate global #define shorthands for info references o remove some register decls o ansi-fy function decls o move items to be close in scope to their usage o add rt_dispatch function for dispatching the actual message o cleanup tangled logic for doing all-but-me msg send Support by: FreeBSD Foundation
2003-10-03 18:15:54 +00:00
rt_dispatch(m, NULL);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
/*
* This is called to generate messages from the routing socket
* indicating a network interface has had addresses associated with it.
* if we ever reverse the logic and replace messages TO the routing
* socket indicate a request to configure interfaces, then it will
* be unnecessary as the routing socket will automatically generate
* copies of it.
*/
void
cleanups prior to adding locking (and in some cases to eliminate locking): o move route_cb to be private to rtsock.c o replace global static route_proto by locals o eliminate global #define shorthands for info references o remove some register decls o ansi-fy function decls o move items to be close in scope to their usage o add rt_dispatch function for dispatching the actual message o cleanup tangled logic for doing all-but-me msg send Support by: FreeBSD Foundation
2003-10-03 18:15:54 +00:00
rt_newaddrmsg(int cmd, struct ifaddr *ifa, int error, struct rtentry *rt)
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
{
struct rt_addrinfo info;
Minor changes to improve code readability (no actual code changes): + replace 0 with NULL where appropriate (not complete) + remove register declaration while there + add argument names to function prototypes to have a better idea of what they are used for + add 'const' qualifiers in 3 places
2004-04-18 00:56:44 +00:00
struct sockaddr *sa = NULL;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
int pass;
Minor changes to improve code readability (no actual code changes): + replace 0 with NULL where appropriate (not complete) + remove register declaration while there + add argument names to function prototypes to have a better idea of what they are used for + add 'const' qualifiers in 3 places
2004-04-18 00:56:44 +00:00
struct mbuf *m = NULL;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
struct ifnet *ifp = ifa->ifa_ifp;
rt_newaddrmsg will blow up if given something other than RTM_ADD or RTM_DELETE; add an assertion, may want to do something more heavyhanded in the future Noticed by: Coverity Prevent analysis tool Reviewed by: mdodd
2005-03-26 21:49:43 +00:00
KASSERT(cmd == RTM_ADD || cmd == RTM_DELETE,
("unexpected cmd %u", cmd));
Ok, here it is, we finally add SCTP to current. Note that this work is not just mine, but it is also the works of Peter Lei and Michael Tuexen. They both are my two key other developers working on the project.. and they need ata-boy's too: **** peterlei@cisco.com tuexen@fh-muenster.de **** I did do a make sysent which updated the syscall's and sysproto.. I hope that is correct... without it you don't build since we have new syscalls for SCTP :-0 So go out and look at the NOTES, add option SCTP (make sure inet and inet6 are present too) and play with SCTP. I will see about comitting some test tools I have after I figure out where I should place them. I also have a lib (libsctp.a) that adds some of the missing socketapi functions that I need to put into lib's.. I will talk to George about this :-) There may still be some 64 bit issues in here, none of us have a 64 bit processor to test with yet.. Michael may have a MAC but thats another beast too.. If you have a mac and want to use SCTP contact Michael he maintains a web site with a loadable module with this code :-) Reviewed by: gnn Approved by: gnn
2006-11-03 15:23:16 +00:00
#ifdef SCTP
/*
* notify the SCTP stack
* this will only get called when an address is added/deleted
* XXX pass the ifaddr struct instead if ifa->ifa_addr...
*/
sctp_addr_change(ifa, cmd);
#endif /* SCTP */
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
if (route_cb.any_count == 0)
return;
for (pass = 1; pass < 3; pass++) {
bzero((caddr_t)&info, sizeof(info));
if ((cmd == RTM_ADD && pass == 1) ||
(cmd == RTM_DELETE && pass == 2)) {
Minor changes to improve code readability (no actual code changes): + replace 0 with NULL where appropriate (not complete) + remove register declaration while there + add argument names to function prototypes to have a better idea of what they are used for + add 'const' qualifiers in 3 places
2004-04-18 00:56:44 +00:00
struct ifa_msghdr *ifam;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
int ncmd = cmd == RTM_ADD ? RTM_NEWADDR : RTM_DELADDR;
cleanups prior to adding locking (and in some cases to eliminate locking): o move route_cb to be private to rtsock.c o replace global static route_proto by locals o eliminate global #define shorthands for info references o remove some register decls o ansi-fy function decls o move items to be close in scope to their usage o add rt_dispatch function for dispatching the actual message o cleanup tangled logic for doing all-but-me msg send Support by: FreeBSD Foundation
2003-10-03 18:15:54 +00:00
info.rti_info[RTAX_IFA] = sa = ifa->ifa_addr;
- Store pointer to the link-level address right in "struct ifnet" rather than in ifindex_table[]; all (except one) accesses are through ifp anyway. IF_LLADDR() works faster, and all (except one) ifaddr_byindex() users were converted to use ifp->if_addr. - Stop storing a (pointer to) Ethernet address in "struct arpcom", and drop the IFP2ENADDR() macro; all users have been converted to use IF_LLADDR() instead.
2005-11-11 16:04:59 +00:00
info.rti_info[RTAX_IFP] = ifp->if_addr->ifa_addr;
cleanups prior to adding locking (and in some cases to eliminate locking): o move route_cb to be private to rtsock.c o replace global static route_proto by locals o eliminate global #define shorthands for info references o remove some register decls o ansi-fy function decls o move items to be close in scope to their usage o add rt_dispatch function for dispatching the actual message o cleanup tangled logic for doing all-but-me msg send Support by: FreeBSD Foundation
2003-10-03 18:15:54 +00:00
info.rti_info[RTAX_NETMASK] = ifa->ifa_netmask;
info.rti_info[RTAX_BRD] = ifa->ifa_dstaddr;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
if ((m = rt_msg1(ncmd, &info)) == NULL)
continue;
ifam = mtod(m, struct ifa_msghdr *);
ifam->ifam_index = ifp->if_index;
ifam->ifam_metric = ifa->ifa_metric;
ifam->ifam_flags = ifa->ifa_flags;
ifam->ifam_addrs = info.rti_addrs;
}
if ((cmd == RTM_ADD && pass == 2) ||
(cmd == RTM_DELETE && pass == 1)) {
Minor changes to improve code readability (no actual code changes): + replace 0 with NULL where appropriate (not complete) + remove register declaration while there + add argument names to function prototypes to have a better idea of what they are used for + add 'const' qualifiers in 3 places
2004-04-18 00:56:44 +00:00
struct rt_msghdr *rtm;
Remove trailing whitespace.
1995-05-30 08:16:23 +00:00
Minor changes to improve code readability (no actual code changes): + replace 0 with NULL where appropriate (not complete) + remove register declaration while there + add argument names to function prototypes to have a better idea of what they are used for + add 'const' qualifiers in 3 places
2004-04-18 00:56:44 +00:00
if (rt == NULL)
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
continue;
cleanups prior to adding locking (and in some cases to eliminate locking): o move route_cb to be private to rtsock.c o replace global static route_proto by locals o eliminate global #define shorthands for info references o remove some register decls o ansi-fy function decls o move items to be close in scope to their usage o add rt_dispatch function for dispatching the actual message o cleanup tangled logic for doing all-but-me msg send Support by: FreeBSD Foundation
2003-10-03 18:15:54 +00:00
info.rti_info[RTAX_NETMASK] = rt_mask(rt);
info.rti_info[RTAX_DST] = sa = rt_key(rt);
info.rti_info[RTAX_GATEWAY] = rt->rt_gateway;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
if ((m = rt_msg1(cmd, &info)) == NULL)
continue;
rtm = mtod(m, struct rt_msghdr *);
rtm->rtm_index = ifp->if_index;
rtm->rtm_flags |= rt->rt_flags;
rtm->rtm_errno = error;
rtm->rtm_addrs = info.rti_addrs;
}
cleanups prior to adding locking (and in some cases to eliminate locking): o move route_cb to be private to rtsock.c o replace global static route_proto by locals o eliminate global #define shorthands for info references o remove some register decls o ansi-fy function decls o move items to be close in scope to their usage o add rt_dispatch function for dispatching the actual message o cleanup tangled logic for doing all-but-me msg send Support by: FreeBSD Foundation
2003-10-03 18:15:54 +00:00
rt_dispatch(m, sa);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
}
Use the new if_multiaddrs list for multicast addresses rather than the previous hackery involving struct in_ifaddr and arpcom. Get rid of the abominable multi_kludge. Update all network interfaces to use the new machanism. Distressingly few Ethernet drivers program the multicast filter properly (assuming the hardware has one, which it usually does).
1997-01-13 21:26:53 +00:00
/*
* This is the analogue to the rt_newaddrmsg which performs the same
* function but for multicast group memberhips. This is easier since
* there is no route state to worry about.
*/
void
cleanups prior to adding locking (and in some cases to eliminate locking): o move route_cb to be private to rtsock.c o replace global static route_proto by locals o eliminate global #define shorthands for info references o remove some register decls o ansi-fy function decls o move items to be close in scope to their usage o add rt_dispatch function for dispatching the actual message o cleanup tangled logic for doing all-but-me msg send Support by: FreeBSD Foundation
2003-10-03 18:15:54 +00:00
rt_newmaddrmsg(int cmd, struct ifmultiaddr *ifma)
Use the new if_multiaddrs list for multicast addresses rather than the previous hackery involving struct in_ifaddr and arpcom. Get rid of the abominable multi_kludge. Update all network interfaces to use the new machanism. Distressingly few Ethernet drivers program the multicast filter properly (assuming the hardware has one, which it usually does).
1997-01-13 21:26:53 +00:00
{
struct rt_addrinfo info;
Minor changes to improve code readability (no actual code changes): + replace 0 with NULL where appropriate (not complete) + remove register declaration while there + add argument names to function prototypes to have a better idea of what they are used for + add 'const' qualifiers in 3 places
2004-04-18 00:56:44 +00:00
struct mbuf *m = NULL;
Use the new if_multiaddrs list for multicast addresses rather than the previous hackery involving struct in_ifaddr and arpcom. Get rid of the abominable multi_kludge. Update all network interfaces to use the new machanism. Distressingly few Ethernet drivers program the multicast filter properly (assuming the hardware has one, which it usually does).
1997-01-13 21:26:53 +00:00
struct ifnet *ifp = ifma->ifma_ifp;
struct ifma_msghdr *ifmam;
if (route_cb.any_count == 0)
return;
bzero((caddr_t)&info, sizeof(info));
cleanups prior to adding locking (and in some cases to eliminate locking): o move route_cb to be private to rtsock.c o replace global static route_proto by locals o eliminate global #define shorthands for info references o remove some register decls o ansi-fy function decls o move items to be close in scope to their usage o add rt_dispatch function for dispatching the actual message o cleanup tangled logic for doing all-but-me msg send Support by: FreeBSD Foundation
2003-10-03 18:15:54 +00:00
info.rti_info[RTAX_IFA] = ifma->ifma_addr;
- Store pointer to the link-level address right in "struct ifnet" rather than in ifindex_table[]; all (except one) accesses are through ifp anyway. IF_LLADDR() works faster, and all (except one) ifaddr_byindex() users were converted to use ifp->if_addr. - Stop storing a (pointer to) Ethernet address in "struct arpcom", and drop the IFP2ENADDR() macro; all users have been converted to use IF_LLADDR() instead.
2005-11-11 16:04:59 +00:00
info.rti_info[RTAX_IFP] = ifp ? ifp->if_addr->ifa_addr : NULL;
Use the new if_multiaddrs list for multicast addresses rather than the previous hackery involving struct in_ifaddr and arpcom. Get rid of the abominable multi_kludge. Update all network interfaces to use the new machanism. Distressingly few Ethernet drivers program the multicast filter properly (assuming the hardware has one, which it usually does).
1997-01-13 21:26:53 +00:00
/*
* If a link-layer address is present, present it as a ``gateway''
* (similarly to how ARP entries, e.g., are presented).
*/
cleanups prior to adding locking (and in some cases to eliminate locking): o move route_cb to be private to rtsock.c o replace global static route_proto by locals o eliminate global #define shorthands for info references o remove some register decls o ansi-fy function decls o move items to be close in scope to their usage o add rt_dispatch function for dispatching the actual message o cleanup tangled logic for doing all-but-me msg send Support by: FreeBSD Foundation
2003-10-03 18:15:54 +00:00
info.rti_info[RTAX_GATEWAY] = ifma->ifma_lladdr;
m = rt_msg1(cmd, &info);
if (m == NULL)
Use the new if_multiaddrs list for multicast addresses rather than the previous hackery involving struct in_ifaddr and arpcom. Get rid of the abominable multi_kludge. Update all network interfaces to use the new machanism. Distressingly few Ethernet drivers program the multicast filter properly (assuming the hardware has one, which it usually does).
1997-01-13 21:26:53 +00:00
return;
ifmam = mtod(m, struct ifma_msghdr *);
Fix a case where hardware removal of an interface caused an attempt to announce an ll_ifma which has gone away. Add a KASSERT to catch regressions. Bug found by: Tom Uffner
2007-03-27 16:11:28 +00:00
KASSERT(ifp != NULL, ("%s: link-layer multicast address w/o ifp\n",
__func__));
Use the new if_multiaddrs list for multicast addresses rather than the previous hackery involving struct in_ifaddr and arpcom. Get rid of the abominable multi_kludge. Update all network interfaces to use the new machanism. Distressingly few Ethernet drivers program the multicast filter properly (assuming the hardware has one, which it usually does).
1997-01-13 21:26:53 +00:00
ifmam->ifmam_index = ifp->if_index;
ifmam->ifmam_addrs = info.rti_addrs;
cleanups prior to adding locking (and in some cases to eliminate locking): o move route_cb to be private to rtsock.c o replace global static route_proto by locals o eliminate global #define shorthands for info references o remove some register decls o ansi-fy function decls o move items to be close in scope to their usage o add rt_dispatch function for dispatching the actual message o cleanup tangled logic for doing all-but-me msg send Support by: FreeBSD Foundation
2003-10-03 18:15:54 +00:00
rt_dispatch(m, ifma->ifma_addr);
Use the new if_multiaddrs list for multicast addresses rather than the previous hackery involving struct in_ifaddr and arpcom. Get rid of the abominable multi_kludge. Update all network interfaces to use the new machanism. Distressingly few Ethernet drivers program the multicast filter properly (assuming the hardware has one, which it usually does).
1997-01-13 21:26:53 +00:00
}
All net.* sysctl converted now.
1995-11-16 19:00:27 +00:00
Add 802.11-specific events that are dispatched through the routing socket. This really doesn't belong here but is preferred (for the moment) over adding yet another mechanism for sending msgs from the kernel to user apps. Reviewed by: imp
2004-10-05 19:48:33 +00:00
static struct mbuf *
rt_makeifannouncemsg(struct ifnet *ifp, int type, int what,
struct rt_addrinfo *info)
{
struct if_announcemsghdr *ifan;
struct mbuf *m;
if (route_cb.any_count == 0)
return NULL;
bzero((caddr_t)info, sizeof(*info));
m = rt_msg1(type, info);
if (m != NULL) {
ifan = mtod(m, struct if_announcemsghdr *);
ifan->ifan_index = ifp->if_index;
strlcpy(ifan->ifan_name, ifp->if_xname,
sizeof(ifan->ifan_name));
ifan->ifan_what = what;
}
return m;
}
/*
* This is called to generate routing socket messages indicating
* IEEE80211 wireless events.
* XXX we piggyback on the RTM_IFANNOUNCE msg format in a clumsy way.
*/
void
rt_ieee80211msg(struct ifnet *ifp, int what, void *data, size_t data_len)
{
struct mbuf *m;
struct rt_addrinfo info;
m = rt_makeifannouncemsg(ifp, RTM_IEEE80211, what, &info);
if (m != NULL) {
/*
* Append the ieee80211 data. Try to stick it in the
* mbuf containing the ifannounce msg; otherwise allocate
* a new mbuf and append.
*
* NB: we assume m is a single mbuf.
*/
if (data_len > M_TRAILINGSPACE(m)) {
struct mbuf *n = m_get(M_NOWAIT, MT_DATA);
if (n == NULL) {
m_freem(m);
return;
}
bcopy(data, mtod(n, void *), data_len);
n->m_len = data_len;
m->m_next = n;
} else if (data_len > 0) {
bcopy(data, mtod(m, u_int8_t *) + m->m_len, data_len);
m->m_len += data_len;
}
if (m->m_flags & M_PKTHDR)
m->m_pkthdr.len += data_len;
mtod(m, struct if_announcemsghdr *)->ifan_msglen += data_len;
rt_dispatch(m, NULL);
}
}
Introduce an interface announcement message for the routing socket so that routing daemons and other interested parties know when an interface is attached/detached. PR: kern/33747 Obtained from: NetBSD MFC after: 2 weeks
2002-01-18 14:33:04 +00:00
/*
* This is called to generate routing socket messages indicating
* network interface arrival and departure.
*/
void
cleanups prior to adding locking (and in some cases to eliminate locking): o move route_cb to be private to rtsock.c o replace global static route_proto by locals o eliminate global #define shorthands for info references o remove some register decls o ansi-fy function decls o move items to be close in scope to their usage o add rt_dispatch function for dispatching the actual message o cleanup tangled logic for doing all-but-me msg send Support by: FreeBSD Foundation
2003-10-03 18:15:54 +00:00
rt_ifannouncemsg(struct ifnet *ifp, int what)
Introduce an interface announcement message for the routing socket so that routing daemons and other interested parties know when an interface is attached/detached. PR: kern/33747 Obtained from: NetBSD MFC after: 2 weeks
2002-01-18 14:33:04 +00:00
{
struct mbuf *m;
struct rt_addrinfo info;
Add 802.11-specific events that are dispatched through the routing socket. This really doesn't belong here but is preferred (for the moment) over adding yet another mechanism for sending msgs from the kernel to user apps. Reviewed by: imp
2004-10-05 19:48:33 +00:00
m = rt_makeifannouncemsg(ifp, RTM_IFANNOUNCE, what, &info);
if (m != NULL)
rt_dispatch(m, NULL);
}
Introduce an interface announcement message for the routing socket so that routing daemons and other interested parties know when an interface is attached/detached. PR: kern/33747 Obtained from: NetBSD MFC after: 2 weeks
2002-01-18 14:33:04 +00:00
cleanups prior to adding locking (and in some cases to eliminate locking): o move route_cb to be private to rtsock.c o replace global static route_proto by locals o eliminate global #define shorthands for info references o remove some register decls o ansi-fy function decls o move items to be close in scope to their usage o add rt_dispatch function for dispatching the actual message o cleanup tangled logic for doing all-but-me msg send Support by: FreeBSD Foundation
2003-10-03 18:15:54 +00:00
static void
Minor changes to improve code readability (no actual code changes): + replace 0 with NULL where appropriate (not complete) + remove register declaration while there + add argument names to function prototypes to have a better idea of what they are used for + add 'const' qualifiers in 3 places
2004-04-18 00:56:44 +00:00
rt_dispatch(struct mbuf *m, const struct sockaddr *sa)
cleanups prior to adding locking (and in some cases to eliminate locking): o move route_cb to be private to rtsock.c o replace global static route_proto by locals o eliminate global #define shorthands for info references o remove some register decls o ansi-fy function decls o move items to be close in scope to their usage o add rt_dispatch function for dispatching the actual message o cleanup tangled logic for doing all-but-me msg send Support by: FreeBSD Foundation
2003-10-03 18:15:54 +00:00
{
Introduce a netisr to deliver kernel-generated routing, avoiding recursive entering of the socket code from the routing code: - Modify rt_dispatch() to bundle up the sockaddr family, if any, associated with a pending mbuf to dispatch to routing sockets, in an m_tag on the mbuf. - Allocate NETISR_ROUTE for use by routing sockets. - Introduce rtsintrq, an ifqueue to be used by the netisr, and introduce rts_input(), a function to unbundle the tagged sockaddr and inject the mbuf and address into raw_input(), which previously occurred in rt_dispatch(). - Introduce rts_init() to initialize rtsintrq, its mutex, and register the netisr. Perform this at the same point in system initialization as setup of the domains. This change introduces asynchrony between the generation of a pending routing socket message and delivery to sockets for use by userspace. It avoids socket->routing->rtsock->socket use and helps to avoid lock order reversals between the routing code and socket code (in particular, raw socket control blocks), as route locks are held over calls to rt_dispatch(). Reviewed by: "George V.Neville-Neil" <gnn@neville-neil.com> Conceptual head nod by: sam
2004-06-09 02:48:23 +00:00
struct m_tag *tag;
cleanups prior to adding locking (and in some cases to eliminate locking): o move route_cb to be private to rtsock.c o replace global static route_proto by locals o eliminate global #define shorthands for info references o remove some register decls o ansi-fy function decls o move items to be close in scope to their usage o add rt_dispatch function for dispatching the actual message o cleanup tangled logic for doing all-but-me msg send Support by: FreeBSD Foundation
2003-10-03 18:15:54 +00:00
Introduce a netisr to deliver kernel-generated routing, avoiding recursive entering of the socket code from the routing code: - Modify rt_dispatch() to bundle up the sockaddr family, if any, associated with a pending mbuf to dispatch to routing sockets, in an m_tag on the mbuf. - Allocate NETISR_ROUTE for use by routing sockets. - Introduce rtsintrq, an ifqueue to be used by the netisr, and introduce rts_input(), a function to unbundle the tagged sockaddr and inject the mbuf and address into raw_input(), which previously occurred in rt_dispatch(). - Introduce rts_init() to initialize rtsintrq, its mutex, and register the netisr. Perform this at the same point in system initialization as setup of the domains. This change introduces asynchrony between the generation of a pending routing socket message and delivery to sockets for use by userspace. It avoids socket->routing->rtsock->socket use and helps to avoid lock order reversals between the routing code and socket code (in particular, raw socket control blocks), as route locks are held over calls to rt_dispatch(). Reviewed by: "George V.Neville-Neil" <gnn@neville-neil.com> Conceptual head nod by: sam
2004-06-09 02:48:23 +00:00
/*
* Preserve the family from the sockaddr, if any, in an m_tag for
* use when injecting the mbuf into the routing socket buffer from
* the netisr.
*/
if (sa != NULL) {
tag = m_tag_get(PACKET_TAG_RTSOCKFAM, sizeof(unsigned short),
M_NOWAIT);
if (tag == NULL) {
m_freem(m);
return;
}
eliminate dead code and collapse the remainder Noticed by: Coverity Prevent analysis tool Reviewed by: rwatson
2005-02-23 22:50:19 +00:00
*(unsigned short *)(tag + 1) = sa->sa_family;
Introduce a netisr to deliver kernel-generated routing, avoiding recursive entering of the socket code from the routing code: - Modify rt_dispatch() to bundle up the sockaddr family, if any, associated with a pending mbuf to dispatch to routing sockets, in an m_tag on the mbuf. - Allocate NETISR_ROUTE for use by routing sockets. - Introduce rtsintrq, an ifqueue to be used by the netisr, and introduce rts_input(), a function to unbundle the tagged sockaddr and inject the mbuf and address into raw_input(), which previously occurred in rt_dispatch(). - Introduce rts_init() to initialize rtsintrq, its mutex, and register the netisr. Perform this at the same point in system initialization as setup of the domains. This change introduces asynchrony between the generation of a pending routing socket message and delivery to sockets for use by userspace. It avoids socket->routing->rtsock->socket use and helps to avoid lock order reversals between the routing code and socket code (in particular, raw socket control blocks), as route locks are held over calls to rt_dispatch(). Reviewed by: "George V.Neville-Neil" <gnn@neville-neil.com> Conceptual head nod by: sam
2004-06-09 02:48:23 +00:00
m_tag_prepend(m, tag);
}
Apply error and success logic consistently to the function netisr_queue() and its users. netisr_queue() now returns (0) on success and ERRNO on failure. At the moment ENXIO (netisr queue not functional) and ENOBUFS (netisr queue full) are supported. Previously it would return (1) on success but the return value of IF_HANDOFF() was interpreted wrongly and (0) was actually returned on success. Due to this schednetisr() was never called to kick the scheduling of the isr. However this was masked by other normal packets coming through netisr_dispatch() causing the dequeueing of waiting packets. PR: kern/70988 Found by: MOROHOSHI Akihiko <moro@remus.dti.ne.jp> MFC after: 3 days
2004-08-27 18:33:08 +00:00
netisr_queue(NETISR_ROUTE, m); /* mbuf is free'd on failure. */
cleanups prior to adding locking (and in some cases to eliminate locking): o move route_cb to be private to rtsock.c o replace global static route_proto by locals o eliminate global #define shorthands for info references o remove some register decls o ansi-fy function decls o move items to be close in scope to their usage o add rt_dispatch function for dispatching the actual message o cleanup tangled logic for doing all-but-me msg send Support by: FreeBSD Foundation
2003-10-03 18:15:54 +00:00
}
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
/*
* This is used in dumping the kernel table via sysctl().
*/
Be consistent about "static" functions: if the function is marked static in its prototype, mark it static at the definition too. Inspired by: FlexeLint warning #512
2002-09-28 17:15:38 +00:00
static int
cleanups prior to adding locking (and in some cases to eliminate locking): o move route_cb to be private to rtsock.c o replace global static route_proto by locals o eliminate global #define shorthands for info references o remove some register decls o ansi-fy function decls o move items to be close in scope to their usage o add rt_dispatch function for dispatching the actual message o cleanup tangled logic for doing all-but-me msg send Support by: FreeBSD Foundation
2003-10-03 18:15:54 +00:00
sysctl_dumpentry(struct radix_node *rn, void *vw)
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
{
cleanups prior to adding locking (and in some cases to eliminate locking): o move route_cb to be private to rtsock.c o replace global static route_proto by locals o eliminate global #define shorthands for info references o remove some register decls o ansi-fy function decls o move items to be close in scope to their usage o add rt_dispatch function for dispatching the actual message o cleanup tangled logic for doing all-but-me msg send Support by: FreeBSD Foundation
2003-10-03 18:15:54 +00:00
struct walkarg *w = vw;
struct rtentry *rt = (struct rtentry *)rn;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
int error = 0, size;
struct rt_addrinfo info;
if (w->w_op == NET_RT_FLAGS && !(rt->rt_flags & w->w_arg))
return 0;
bzero((caddr_t)&info, sizeof(info));
cleanups prior to adding locking (and in some cases to eliminate locking): o move route_cb to be private to rtsock.c o replace global static route_proto by locals o eliminate global #define shorthands for info references o remove some register decls o ansi-fy function decls o move items to be close in scope to their usage o add rt_dispatch function for dispatching the actual message o cleanup tangled logic for doing all-but-me msg send Support by: FreeBSD Foundation
2003-10-03 18:15:54 +00:00
info.rti_info[RTAX_DST] = rt_key(rt);
info.rti_info[RTAX_GATEWAY] = rt->rt_gateway;
info.rti_info[RTAX_NETMASK] = rt_mask(rt);
info.rti_info[RTAX_GENMASK] = rt->rt_genmask;
Bring in latest CSRG revisions to this file: - Report destination address of a P2P link when servicing routing socket messages. - Report interface name, address, and destination address of a P2P link when servicing NET_RT_{DUMP,FLAGS} sysctls. Part of CSRG revision 8.6 coresponds to revision 1.12. CSRG revision 8.7 corresponds to revision 1.15.
2001-10-17 11:23:59 +00:00
if (rt->rt_ifp) {
- Store pointer to the link-level address right in "struct ifnet" rather than in ifindex_table[]; all (except one) accesses are through ifp anyway. IF_LLADDR() works faster, and all (except one) ifaddr_byindex() users were converted to use ifp->if_addr. - Stop storing a (pointer to) Ethernet address in "struct arpcom", and drop the IFP2ENADDR() macro; all users have been converted to use IF_LLADDR() instead.
2005-11-11 16:04:59 +00:00
info.rti_info[RTAX_IFP] = rt->rt_ifp->if_addr->ifa_addr;
cleanups prior to adding locking (and in some cases to eliminate locking): o move route_cb to be private to rtsock.c o replace global static route_proto by locals o eliminate global #define shorthands for info references o remove some register decls o ansi-fy function decls o move items to be close in scope to their usage o add rt_dispatch function for dispatching the actual message o cleanup tangled logic for doing all-but-me msg send Support by: FreeBSD Foundation
2003-10-03 18:15:54 +00:00
info.rti_info[RTAX_IFA] = rt->rt_ifa->ifa_addr;
Bring in latest CSRG revisions to this file: - Report destination address of a P2P link when servicing routing socket messages. - Report interface name, address, and destination address of a P2P link when servicing NET_RT_{DUMP,FLAGS} sysctls. Part of CSRG revision 8.6 coresponds to revision 1.12. CSRG revision 8.7 corresponds to revision 1.15.
2001-10-17 11:23:59 +00:00
if (rt->rt_ifp->if_flags & IFF_POINTOPOINT)
cleanups prior to adding locking (and in some cases to eliminate locking): o move route_cb to be private to rtsock.c o replace global static route_proto by locals o eliminate global #define shorthands for info references o remove some register decls o ansi-fy function decls o move items to be close in scope to their usage o add rt_dispatch function for dispatching the actual message o cleanup tangled logic for doing all-but-me msg send Support by: FreeBSD Foundation
2003-10-03 18:15:54 +00:00
info.rti_info[RTAX_BRD] = rt->rt_ifa->ifa_dstaddr;
Bring in latest CSRG revisions to this file: - Report destination address of a P2P link when servicing routing socket messages. - Report interface name, address, and destination address of a P2P link when servicing NET_RT_{DUMP,FLAGS} sysctls. Part of CSRG revision 8.6 coresponds to revision 1.12. CSRG revision 8.7 corresponds to revision 1.15.
2001-10-17 11:23:59 +00:00
}
misc cleanup in sysctl_ifmalist(): + remove a partly incorrect comment that i introduced in the last commit; + deal with the correct part of the above comment by cleaning up the updates of 'info' -- rti_addrs needd not to be updated, rti_info[RTAX_IFP] can be set once outside the loop. While at it, correct a few misspelling of NULL as 0, but there are way too many in this file, and i did not want to clutter the important part of this commit.
2004-04-17 15:09:36 +00:00
size = rt_msg2(RTM_GET, &info, NULL, w);
All net.* sysctl converted now.
1995-11-16 19:00:27 +00:00
if (w->w_req && w->w_tmem) {
cleanups prior to adding locking (and in some cases to eliminate locking): o move route_cb to be private to rtsock.c o replace global static route_proto by locals o eliminate global #define shorthands for info references o remove some register decls o ansi-fy function decls o move items to be close in scope to their usage o add rt_dispatch function for dispatching the actual message o cleanup tangled logic for doing all-but-me msg send Support by: FreeBSD Foundation
2003-10-03 18:15:54 +00:00
struct rt_msghdr *rtm = (struct rt_msghdr *)w->w_tmem;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
rtm->rtm_flags = rt->rt_flags;
Introduce tcp_hostcache and remove the tcp specific metrics from the routing table. Move all usage and references in the tcp stack from the routing table metrics to the tcp hostcache. It caches measured parameters of past tcp sessions to provide better initial start values for following connections from or to the same source or destination. Depending on the network parameters to/from the remote host this can lead to significant speedups for new tcp connections after the first one because they inherit and shortcut the learning curve. tcp_hostcache is designed for multiple concurrent access in SMP environments with high contention and is hash indexed by remote ip address. It removes significant locking requirements from the tcp stack with regard to the routing table. Reviewed by: sam (mentor), bms Reviewed by: -net, -current, core@kame.net (IPv6 parts) Approved by: re (scottl)
2003-11-20 20:07:39 +00:00
rtm->rtm_use = rt->rt_rmx.rmx_pksent;
rt_getmetrics(&rt->rt_rmx, &rtm->rtm_rmx);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
rtm->rtm_index = rt->rt_ifp->if_index;
rtm->rtm_errno = rtm->rtm_pid = rtm->rtm_seq = 0;
rtm->rtm_addrs = info.rti_addrs;
All net.* sysctl converted now.
1995-11-16 19:00:27 +00:00
error = SYSCTL_OUT(w->w_req, (caddr_t)rtm, size);
return (error);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
return (error);
}
Be consistent about "static" functions: if the function is marked static in its prototype, mark it static at the definition too. Inspired by: FlexeLint warning #512
2002-09-28 17:15:38 +00:00
static int
cleanups prior to adding locking (and in some cases to eliminate locking): o move route_cb to be private to rtsock.c o replace global static route_proto by locals o eliminate global #define shorthands for info references o remove some register decls o ansi-fy function decls o move items to be close in scope to their usage o add rt_dispatch function for dispatching the actual message o cleanup tangled logic for doing all-but-me msg send Support by: FreeBSD Foundation
2003-10-03 18:15:54 +00:00
sysctl_iflist(int af, struct walkarg *w)
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
{
cleanups prior to adding locking (and in some cases to eliminate locking): o move route_cb to be private to rtsock.c o replace global static route_proto by locals o eliminate global #define shorthands for info references o remove some register decls o ansi-fy function decls o move items to be close in scope to their usage o add rt_dispatch function for dispatching the actual message o cleanup tangled logic for doing all-but-me msg send Support by: FreeBSD Foundation
2003-10-03 18:15:54 +00:00
struct ifnet *ifp;
struct ifaddr *ifa;
struct rt_addrinfo info;
int len, error = 0;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
bzero((caddr_t)&info, sizeof(info));
Protect interface and address lists using the appropriate mutex. These locks were not aquired because the user buffers were not wired, thus it was possible that that SYSCTL_OUT could sleep, causing a number of different problems such as lock ordering issues and dead locks. -Wire user supplied buffer to ensure SYSCTL_OUT will not sleep. -Pickup ifnet locks to protect the list. -Where applicable pickup address locks. -Pickup radix node head locks. -Remove splnet stubs -Remove various comments about locking here, because they are no longer needed. It is the hope that these changes will make sysctl_rtsock MP safe. MFC after: 3 weeks
2005-09-10 15:12:24 +00:00
IFNET_RLOCK();
Mechanical change to use <sys/queue.h> macro API instead of fondling implementation details. Created with: sed(1) Reviewed by: md5(1)
2001-02-04 13:13:25 +00:00
TAILQ_FOREACH(ifp, &ifnet, if_link) {
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
if (w->w_arg && w->w_arg != ifp->if_index)
continue;
- Store pointer to the link-level address right in "struct ifnet" rather than in ifindex_table[]; all (except one) accesses are through ifp anyway. IF_LLADDR() works faster, and all (except one) ifaddr_byindex() users were converted to use ifp->if_addr. - Stop storing a (pointer to) Ethernet address in "struct arpcom", and drop the IFP2ENADDR() macro; all users have been converted to use IF_LLADDR() instead.
2005-11-11 16:04:59 +00:00
ifa = ifp->if_addr;
cleanups prior to adding locking (and in some cases to eliminate locking): o move route_cb to be private to rtsock.c o replace global static route_proto by locals o eliminate global #define shorthands for info references o remove some register decls o ansi-fy function decls o move items to be close in scope to their usage o add rt_dispatch function for dispatching the actual message o cleanup tangled logic for doing all-but-me msg send Support by: FreeBSD Foundation
2003-10-03 18:15:54 +00:00
info.rti_info[RTAX_IFP] = ifa->ifa_addr;
misc cleanup in sysctl_ifmalist(): + remove a partly incorrect comment that i introduced in the last commit; + deal with the correct part of the above comment by cleaning up the updates of 'info' -- rti_addrs needd not to be updated, rti_info[RTAX_IFP] can be set once outside the loop. While at it, correct a few misspelling of NULL as 0, but there are way too many in this file, and i did not want to clutter the important part of this commit.
2004-04-17 15:09:36 +00:00
len = rt_msg2(RTM_IFINFO, &info, NULL, w);
Minor changes to improve code readability (no actual code changes): + replace 0 with NULL where appropriate (not complete) + remove register declaration while there + add argument names to function prototypes to have a better idea of what they are used for + add 'const' qualifiers in 3 places
2004-04-18 00:56:44 +00:00
info.rti_info[RTAX_IFP] = NULL;
All net.* sysctl converted now.
1995-11-16 19:00:27 +00:00
if (w->w_req && w->w_tmem) {
cleanups prior to adding locking (and in some cases to eliminate locking): o move route_cb to be private to rtsock.c o replace global static route_proto by locals o eliminate global #define shorthands for info references o remove some register decls o ansi-fy function decls o move items to be close in scope to their usage o add rt_dispatch function for dispatching the actual message o cleanup tangled logic for doing all-but-me msg send Support by: FreeBSD Foundation
2003-10-03 18:15:54 +00:00
struct if_msghdr *ifm;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
ifm = (struct if_msghdr *)w->w_tmem;
ifm->ifm_index = ifp->if_index;
Rename IFF_RUNNING to IFF_DRV_RUNNING, IFF_OACTIVE to IFF_DRV_OACTIVE, and move both flags from ifnet.if_flags to ifnet.if_drv_flags, making and documenting the locking of these flags the responsibility of the device driver, not the network stack. The flags for these two fields will be mutually exclusive so that they can be exposed to user space as though they were stored in the same variable. Provide #defines to provide the old names #ifndef _KERNEL, so that user applications (such as ifconfig) can use the old flag names. Using the old names in a device driver will result in a compile error in order to help device driver writers adopt the new model. When exposing the interface flags to user space, via interface ioctls or routing sockets, or the two fields together. Since the driver flags cannot currently be set for user space, no new logic is currently required to handle this case. Add some assertions that general purpose network stack routines, such as if_setflags(), are not improperly used on driver-owned flags. With this change, a large number of very minor network stack races are closed, subject to correct device driver locking. Most were likely never triggered. Driver sweep to follow; many thanks to pjd and bz for the line-by-line review they gave this patch. Reviewed by: pjd, bz MFC after: 7 days
2005-08-09 10:16:17 +00:00
ifm->ifm_flags = ifp->if_flags | ifp->if_drv_flags;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
ifm->ifm_data = ifp->if_data;
ifm->ifm_addrs = info.rti_addrs;
All net.* sysctl converted now.
1995-11-16 19:00:27 +00:00
error = SYSCTL_OUT(w->w_req,(caddr_t)ifm, len);
Cosmetics: to silence gcc -wall.
1994-10-08 22:38:27 +00:00
if (error)
Change sysctl_iflist() so it has a single point of return. This will assist any future locking efforts.
2001-09-29 05:08:04 +00:00
goto done;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
Minor changes to improve code readability (no actual code changes): + replace 0 with NULL where appropriate (not complete) + remove register declaration while there + add argument names to function prototypes to have a better idea of what they are used for + add 'const' qualifiers in 3 places
2004-04-18 00:56:44 +00:00
while ((ifa = TAILQ_NEXT(ifa, ifa_link)) != NULL) {
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
if (af && af != ifa->ifa_addr->sa_family)
continue;
Simple p_ucred -> td_ucred changes to start using the per-thread ucred reference.
2002-02-27 18:32:23 +00:00
if (jailed(curthread->td_ucred) &&
prison_if(curthread->td_ucred, ifa->ifa_addr))
This Implements the mumbled about "Jail" feature. This is a seriously beefed up chroot kind of thing. The process is jailed along the same lines as a chroot does it, but with additional tough restrictions imposed on what the superuser can do. For all I know, it is safe to hand over the root bit inside a prison to the customer living in that prison, this is what it was developed for in fact: "real virtual servers". Each prison has an ip number associated with it, which all IP communications will be coerced to use and each prison has its own hostname. Needless to say, you need more RAM this way, but the advantage is that each customer can run their own particular version of apache and not stomp on the toes of their neighbors. It generally does what one would expect, but setting up a jail still takes a little knowledge. A few notes: I have no scripts for setting up a jail, don't ask me for them. The IP number should be an alias on one of the interfaces. mount a /proc in each jail, it will make ps more useable. /proc/<pid>/status tells the hostname of the prison for jailed processes. Quotas are only sensible if you have a mountpoint per prison. There are no privisions for stopping resource-hogging. Some "#ifdef INET" and similar may be missing (send patches!) If somebody wants to take it from here and develop it into more of a "virtual machine" they should be most welcome! Tools, comments, patches & documentation most welcome. Have fun... Sponsored by: http://www.rndassociates.com/ Run for almost a year by: http://www.servetheweb.com/
1999-04-28 11:38:52 +00:00
continue;
cleanups prior to adding locking (and in some cases to eliminate locking): o move route_cb to be private to rtsock.c o replace global static route_proto by locals o eliminate global #define shorthands for info references o remove some register decls o ansi-fy function decls o move items to be close in scope to their usage o add rt_dispatch function for dispatching the actual message o cleanup tangled logic for doing all-but-me msg send Support by: FreeBSD Foundation
2003-10-03 18:15:54 +00:00
info.rti_info[RTAX_IFA] = ifa->ifa_addr;
info.rti_info[RTAX_NETMASK] = ifa->ifa_netmask;
info.rti_info[RTAX_BRD] = ifa->ifa_dstaddr;
misc cleanup in sysctl_ifmalist(): + remove a partly incorrect comment that i introduced in the last commit; + deal with the correct part of the above comment by cleaning up the updates of 'info' -- rti_addrs needd not to be updated, rti_info[RTAX_IFP] can be set once outside the loop. While at it, correct a few misspelling of NULL as 0, but there are way too many in this file, and i did not want to clutter the important part of this commit.
2004-04-17 15:09:36 +00:00
len = rt_msg2(RTM_NEWADDR, &info, NULL, w);
All net.* sysctl converted now.
1995-11-16 19:00:27 +00:00
if (w->w_req && w->w_tmem) {
cleanups prior to adding locking (and in some cases to eliminate locking): o move route_cb to be private to rtsock.c o replace global static route_proto by locals o eliminate global #define shorthands for info references o remove some register decls o ansi-fy function decls o move items to be close in scope to their usage o add rt_dispatch function for dispatching the actual message o cleanup tangled logic for doing all-but-me msg send Support by: FreeBSD Foundation
2003-10-03 18:15:54 +00:00
struct ifa_msghdr *ifam;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
ifam = (struct ifa_msghdr *)w->w_tmem;
ifam->ifam_index = ifa->ifa_ifp->if_index;
ifam->ifam_flags = ifa->ifa_flags;
ifam->ifam_metric = ifa->ifa_metric;
ifam->ifam_addrs = info.rti_addrs;
All net.* sysctl converted now.
1995-11-16 19:00:27 +00:00
error = SYSCTL_OUT(w->w_req, w->w_tmem, len);
Cosmetics: to silence gcc -wall.
1994-10-08 22:38:27 +00:00
if (error)
Change sysctl_iflist() so it has a single point of return. This will assist any future locking efforts.
2001-09-29 05:08:04 +00:00
goto done;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
}
cleanups prior to adding locking (and in some cases to eliminate locking): o move route_cb to be private to rtsock.c o replace global static route_proto by locals o eliminate global #define shorthands for info references o remove some register decls o ansi-fy function decls o move items to be close in scope to their usage o add rt_dispatch function for dispatching the actual message o cleanup tangled logic for doing all-but-me msg send Support by: FreeBSD Foundation
2003-10-03 18:15:54 +00:00
info.rti_info[RTAX_IFA] = info.rti_info[RTAX_NETMASK] =
Minor changes to improve code readability (no actual code changes): + replace 0 with NULL where appropriate (not complete) + remove register declaration while there + add argument names to function prototypes to have a better idea of what they are used for + add 'const' qualifiers in 3 places
2004-04-18 00:56:44 +00:00
info.rti_info[RTAX_BRD] = NULL;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
Change sysctl_iflist() so it has a single point of return. This will assist any future locking efforts.
2001-09-29 05:08:04 +00:00
done:
Protect interface and address lists using the appropriate mutex. These locks were not aquired because the user buffers were not wired, thus it was possible that that SYSCTL_OUT could sleep, causing a number of different problems such as lock ordering issues and dead locks. -Wire user supplied buffer to ensure SYSCTL_OUT will not sleep. -Pickup ifnet locks to protect the list. -Where applicable pickup address locks. -Pickup radix node head locks. -Remove splnet stubs -Remove various comments about locking here, because they are no longer needed. It is the hope that these changes will make sysctl_rtsock MP safe. MFC after: 3 weeks
2005-09-10 15:12:24 +00:00
IFNET_RUNLOCK();
Change sysctl_iflist() so it has a single point of return. This will assist any future locking efforts.
2001-09-29 05:08:04 +00:00
return (error);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
Add a sysctl MIB, NET_RT_IFMALIST, to retrieve multicast group memberships in a protocol-independent way. Submitted by: harti
2003-11-14 18:48:15 +00:00
int
Consistently use ifaddr_byindex() to access the link-level address of an interface. No functional change. On passing, comment a likely bug in net/rtsock.c:sysctl_ifmalist() which, if confirmed, would deserve to be fixed and MFC'ed
2004-04-16 08:14:34 +00:00
sysctl_ifmalist(int af, struct walkarg *w)
Add a sysctl MIB, NET_RT_IFMALIST, to retrieve multicast group memberships in a protocol-independent way. Submitted by: harti
2003-11-14 18:48:15 +00:00
{
Consistently use ifaddr_byindex() to access the link-level address of an interface. No functional change. On passing, comment a likely bug in net/rtsock.c:sysctl_ifmalist() which, if confirmed, would deserve to be fixed and MFC'ed
2004-04-16 08:14:34 +00:00
struct ifnet *ifp;
Add a sysctl MIB, NET_RT_IFMALIST, to retrieve multicast group memberships in a protocol-independent way. Submitted by: harti
2003-11-14 18:48:15 +00:00
struct ifmultiaddr *ifma;
struct rt_addrinfo info;
int len, error = 0;
Consistently use ifaddr_byindex() to access the link-level address of an interface. No functional change. On passing, comment a likely bug in net/rtsock.c:sysctl_ifmalist() which, if confirmed, would deserve to be fixed and MFC'ed
2004-04-16 08:14:34 +00:00
struct ifaddr *ifa;
Add a sysctl MIB, NET_RT_IFMALIST, to retrieve multicast group memberships in a protocol-independent way. Submitted by: harti
2003-11-14 18:48:15 +00:00
bzero((caddr_t)&info, sizeof(info));
Protect interface and address lists using the appropriate mutex. These locks were not aquired because the user buffers were not wired, thus it was possible that that SYSCTL_OUT could sleep, causing a number of different problems such as lock ordering issues and dead locks. -Wire user supplied buffer to ensure SYSCTL_OUT will not sleep. -Pickup ifnet locks to protect the list. -Where applicable pickup address locks. -Pickup radix node head locks. -Remove splnet stubs -Remove various comments about locking here, because they are no longer needed. It is the hope that these changes will make sysctl_rtsock MP safe. MFC after: 3 weeks
2005-09-10 15:12:24 +00:00
IFNET_RLOCK();
Add a sysctl MIB, NET_RT_IFMALIST, to retrieve multicast group memberships in a protocol-independent way. Submitted by: harti
2003-11-14 18:48:15 +00:00
TAILQ_FOREACH(ifp, &ifnet, if_link) {
if (w->w_arg && w->w_arg != ifp->if_index)
continue;
- Store pointer to the link-level address right in "struct ifnet" rather than in ifindex_table[]; all (except one) accesses are through ifp anyway. IF_LLADDR() works faster, and all (except one) ifaddr_byindex() users were converted to use ifp->if_addr. - Stop storing a (pointer to) Ethernet address in "struct arpcom", and drop the IFP2ENADDR() macro; all users have been converted to use IF_LLADDR() instead.
2005-11-11 16:04:59 +00:00
ifa = ifp->if_addr;
misc cleanup in sysctl_ifmalist(): + remove a partly incorrect comment that i introduced in the last commit; + deal with the correct part of the above comment by cleaning up the updates of 'info' -- rti_addrs needd not to be updated, rti_info[RTAX_IFP] can be set once outside the loop. While at it, correct a few misspelling of NULL as 0, but there are way too many in this file, and i did not want to clutter the important part of this commit.
2004-04-17 15:09:36 +00:00
info.rti_info[RTAX_IFP] = ifa ? ifa->ifa_addr : NULL;
Protect interface and address lists using the appropriate mutex. These locks were not aquired because the user buffers were not wired, thus it was possible that that SYSCTL_OUT could sleep, causing a number of different problems such as lock ordering issues and dead locks. -Wire user supplied buffer to ensure SYSCTL_OUT will not sleep. -Pickup ifnet locks to protect the list. -Where applicable pickup address locks. -Pickup radix node head locks. -Remove splnet stubs -Remove various comments about locking here, because they are no longer needed. It is the hope that these changes will make sysctl_rtsock MP safe. MFC after: 3 weeks
2005-09-10 15:12:24 +00:00
IF_ADDR_LOCK(ifp);
Add a sysctl MIB, NET_RT_IFMALIST, to retrieve multicast group memberships in a protocol-independent way. Submitted by: harti
2003-11-14 18:48:15 +00:00
TAILQ_FOREACH(ifma, &ifp->if_multiaddrs, ifma_link) {
if (af && af != ifma->ifma_addr->sa_family)
continue;
if (jailed(curproc->p_ucred) &&
prison_if(curproc->p_ucred, ifma->ifma_addr))
continue;
info.rti_info[RTAX_IFA] = ifma->ifma_addr;
misc cleanup in sysctl_ifmalist(): + remove a partly incorrect comment that i introduced in the last commit; + deal with the correct part of the above comment by cleaning up the updates of 'info' -- rti_addrs needd not to be updated, rti_info[RTAX_IFP] can be set once outside the loop. While at it, correct a few misspelling of NULL as 0, but there are way too many in this file, and i did not want to clutter the important part of this commit.
2004-04-17 15:09:36 +00:00
info.rti_info[RTAX_GATEWAY] =
(ifma->ifma_addr->sa_family != AF_LINK) ?
ifma->ifma_lladdr : NULL;
len = rt_msg2(RTM_NEWMADDR, &info, NULL, w);
Add a sysctl MIB, NET_RT_IFMALIST, to retrieve multicast group memberships in a protocol-independent way. Submitted by: harti
2003-11-14 18:48:15 +00:00
if (w->w_req && w->w_tmem) {
Consistently use ifaddr_byindex() to access the link-level address of an interface. No functional change. On passing, comment a likely bug in net/rtsock.c:sysctl_ifmalist() which, if confirmed, would deserve to be fixed and MFC'ed
2004-04-16 08:14:34 +00:00
struct ifma_msghdr *ifmam;
Add a sysctl MIB, NET_RT_IFMALIST, to retrieve multicast group memberships in a protocol-independent way. Submitted by: harti
2003-11-14 18:48:15 +00:00
ifmam = (struct ifma_msghdr *)w->w_tmem;
ifmam->ifmam_index = ifma->ifma_ifp->if_index;
ifmam->ifmam_flags = 0;
ifmam->ifmam_addrs = info.rti_addrs;
error = SYSCTL_OUT(w->w_req, w->w_tmem, len);
Protect interface and address lists using the appropriate mutex. These locks were not aquired because the user buffers were not wired, thus it was possible that that SYSCTL_OUT could sleep, causing a number of different problems such as lock ordering issues and dead locks. -Wire user supplied buffer to ensure SYSCTL_OUT will not sleep. -Pickup ifnet locks to protect the list. -Where applicable pickup address locks. -Pickup radix node head locks. -Remove splnet stubs -Remove various comments about locking here, because they are no longer needed. It is the hope that these changes will make sysctl_rtsock MP safe. MFC after: 3 weeks
2005-09-10 15:12:24 +00:00
if (error) {
IF_ADDR_UNLOCK(ifp);
Add a sysctl MIB, NET_RT_IFMALIST, to retrieve multicast group memberships in a protocol-independent way. Submitted by: harti
2003-11-14 18:48:15 +00:00
goto done;
Protect interface and address lists using the appropriate mutex. These locks were not aquired because the user buffers were not wired, thus it was possible that that SYSCTL_OUT could sleep, causing a number of different problems such as lock ordering issues and dead locks. -Wire user supplied buffer to ensure SYSCTL_OUT will not sleep. -Pickup ifnet locks to protect the list. -Where applicable pickup address locks. -Pickup radix node head locks. -Remove splnet stubs -Remove various comments about locking here, because they are no longer needed. It is the hope that these changes will make sysctl_rtsock MP safe. MFC after: 3 weeks
2005-09-10 15:12:24 +00:00
}
Add a sysctl MIB, NET_RT_IFMALIST, to retrieve multicast group memberships in a protocol-independent way. Submitted by: harti
2003-11-14 18:48:15 +00:00
}
}
Protect interface and address lists using the appropriate mutex. These locks were not aquired because the user buffers were not wired, thus it was possible that that SYSCTL_OUT could sleep, causing a number of different problems such as lock ordering issues and dead locks. -Wire user supplied buffer to ensure SYSCTL_OUT will not sleep. -Pickup ifnet locks to protect the list. -Where applicable pickup address locks. -Pickup radix node head locks. -Remove splnet stubs -Remove various comments about locking here, because they are no longer needed. It is the hope that these changes will make sysctl_rtsock MP safe. MFC after: 3 weeks
2005-09-10 15:12:24 +00:00
IF_ADDR_UNLOCK(ifp);
Add a sysctl MIB, NET_RT_IFMALIST, to retrieve multicast group memberships in a protocol-independent way. Submitted by: harti
2003-11-14 18:48:15 +00:00
}
done:
Protect interface and address lists using the appropriate mutex. These locks were not aquired because the user buffers were not wired, thus it was possible that that SYSCTL_OUT could sleep, causing a number of different problems such as lock ordering issues and dead locks. -Wire user supplied buffer to ensure SYSCTL_OUT will not sleep. -Pickup ifnet locks to protect the list. -Where applicable pickup address locks. -Pickup radix node head locks. -Remove splnet stubs -Remove various comments about locking here, because they are no longer needed. It is the hope that these changes will make sysctl_rtsock MP safe. MFC after: 3 weeks
2005-09-10 15:12:24 +00:00
IFNET_RUNLOCK();
Add a sysctl MIB, NET_RT_IFMALIST, to retrieve multicast group memberships in a protocol-independent way. Submitted by: harti
2003-11-14 18:48:15 +00:00
return (error);
}
All net.* sysctl converted now.
1995-11-16 19:00:27 +00:00
static int
Previous commit changing SYSCTL_HANDLER_ARGS violated KNF. Pointed out by: bde
2000-07-04 11:25:35 +00:00
sysctl_rtsock(SYSCTL_HANDLER_ARGS)
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
{
All net.* sysctl converted now.
1995-11-16 19:00:27 +00:00
int *name = (int *)arg1;
u_int namelen = arg2;
cleanups prior to adding locking (and in some cases to eliminate locking): o move route_cb to be private to rtsock.c o replace global static route_proto by locals o eliminate global #define shorthands for info references o remove some register decls o ansi-fy function decls o move items to be close in scope to their usage o add rt_dispatch function for dispatching the actual message o cleanup tangled logic for doing all-but-me msg send Support by: FreeBSD Foundation
2003-10-03 18:15:54 +00:00
struct radix_node_head *rnh;
Protect interface and address lists using the appropriate mutex. These locks were not aquired because the user buffers were not wired, thus it was possible that that SYSCTL_OUT could sleep, causing a number of different problems such as lock ordering issues and dead locks. -Wire user supplied buffer to ensure SYSCTL_OUT will not sleep. -Pickup ifnet locks to protect the list. -Where applicable pickup address locks. -Pickup radix node head locks. -Remove splnet stubs -Remove various comments about locking here, because they are no longer needed. It is the hope that these changes will make sysctl_rtsock MP safe. MFC after: 3 weeks
2005-09-10 15:12:24 +00:00
int i, lim, error = EINVAL;
remove an almost-duplicate piece of code by setting the loop limits appropriately.
2004-04-12 20:26:01 +00:00
u_char af;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
struct walkarg w;
All net.* sysctl converted now.
1995-11-16 19:00:27 +00:00
name ++;
namelen--;
if (req->newptr)
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
return (EPERM);
if (namelen != 3)
Have sysctl() return the correct errno(2) as documented in the sysctl(3) manpage. Submitted by: ru Obtained from: BSD/OS
2002-01-25 00:50:49 +00:00
return ((namelen < 3) ? EISDIR : ENOTDIR);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
af = name[0];
Range-check the address family parameter passed in to the sysctl handler. Submitted by: ru
2002-12-25 10:51:20 +00:00
if (af > AF_MAX)
return (EINVAL);
+ replace Bcmp/Bzero with 'the real thing' as in the rest of the file. + remember to check and fix or explain a strange cast in route_output()
2004-04-18 11:47:04 +00:00
bzero(&w, sizeof(w));
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
w.w_op = name[1];
w.w_arg = name[2];
All net.* sysctl converted now.
1995-11-16 19:00:27 +00:00
w.w_req = req;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
Protect interface and address lists using the appropriate mutex. These locks were not aquired because the user buffers were not wired, thus it was possible that that SYSCTL_OUT could sleep, causing a number of different problems such as lock ordering issues and dead locks. -Wire user supplied buffer to ensure SYSCTL_OUT will not sleep. -Pickup ifnet locks to protect the list. -Where applicable pickup address locks. -Pickup radix node head locks. -Remove splnet stubs -Remove various comments about locking here, because they are no longer needed. It is the hope that these changes will make sysctl_rtsock MP safe. MFC after: 3 weeks
2005-09-10 15:12:24 +00:00
error = sysctl_wire_old_buffer(req, 0);
if (error)
return (error);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
switch (w.w_op) {
case NET_RT_DUMP:
case NET_RT_FLAGS:
remove an almost-duplicate piece of code by setting the loop limits appropriately.
2004-04-12 20:26:01 +00:00
if (af == 0) { /* dump all tables */
i = 1;
lim = AF_MAX;
} else /* dump only one table */
i = lim = af;
for (error = 0; error == 0 && i <= lim; i++)
if ((rnh = rt_tables[i]) != NULL) {
Protect interface and address lists using the appropriate mutex. These locks were not aquired because the user buffers were not wired, thus it was possible that that SYSCTL_OUT could sleep, causing a number of different problems such as lock ordering issues and dead locks. -Wire user supplied buffer to ensure SYSCTL_OUT will not sleep. -Pickup ifnet locks to protect the list. -Where applicable pickup address locks. -Pickup radix node head locks. -Remove splnet stubs -Remove various comments about locking here, because they are no longer needed. It is the hope that these changes will make sysctl_rtsock MP safe. MFC after: 3 weeks
2005-09-10 15:12:24 +00:00
RADIX_NODE_HEAD_LOCK(rnh);
SMP locking for radix nodes.
2002-12-24 03:03:39 +00:00
error = rnh->rnh_walktree(rnh,
Protect interface and address lists using the appropriate mutex. These locks were not aquired because the user buffers were not wired, thus it was possible that that SYSCTL_OUT could sleep, causing a number of different problems such as lock ordering issues and dead locks. -Wire user supplied buffer to ensure SYSCTL_OUT will not sleep. -Pickup ifnet locks to protect the list. -Where applicable pickup address locks. -Pickup radix node head locks. -Remove splnet stubs -Remove various comments about locking here, because they are no longer needed. It is the hope that these changes will make sysctl_rtsock MP safe. MFC after: 3 weeks
2005-09-10 15:12:24 +00:00
sysctl_dumpentry, &w);
RADIX_NODE_HEAD_UNLOCK(rnh);
remove an almost-duplicate piece of code by setting the loop limits appropriately.
2004-04-12 20:26:01 +00:00
} else if (af != 0)
SMP locking for radix nodes.
2002-12-24 03:03:39 +00:00
error = EAFNOSUPPORT;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
break;
case NET_RT_IFLIST:
error = sysctl_iflist(af, &w);
Add a sysctl MIB, NET_RT_IFMALIST, to retrieve multicast group memberships in a protocol-independent way. Submitted by: harti
2003-11-14 18:48:15 +00:00
break;
case NET_RT_IFMALIST:
error = sysctl_ifmalist(af, &w);
break;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
if (w.w_tmem)
free(w.w_tmem, M_RTABLE);
return (error);
}
The long-awaited mega-massive-network-code- cleanup. Part I. This commit includes the following changes: 1) Old-style (pr_usrreq()) protocols are no longer supported, the compatibility glue for them is deleted, and the kernel will panic on boot if any are compiled in. 2) Certain protocol entry points are modified to take a process structure, so they they can easily tell whether or not it is possible to sleep, and also to access credentials. 3) SS_PRIV is no more, and with it goes the SO_PRIVSTATE setsockopt() call. Protocols should use the process pointer they are now passed. 4) The PF_LOCAL and PF_ROUTE families have been updated to use the new style, as has the `raw' skeleton family. 5) PF_LOCAL sockets now obey the process's umask when creating a socket in the filesystem. As a result, LINT is now broken. I'm hoping that some enterprising hacker with a bit more time will either make the broken bits work (should be easy for netipx) or dike them out.
1997-04-27 20:01:29 +00:00
SYSCTL_NODE(_net, PF_ROUTE, routetable, CTLFLAG_RD, sysctl_rtsock, "");
All net.* sysctl converted now.
1995-11-16 19:00:27 +00:00
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
/*
* Definitions of protocols supported in the ROUTE domain.
*/
Forward declaring static variables as extern is invalid ISO-C. Now that GCC can properly handle forward static declarations, do this properly.
2005-09-07 10:06:14 +00:00
static struct domain routedomain; /* or at least forward */
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
All net.* sysctl converted now.
1995-11-16 19:00:27 +00:00
static struct protosw routesw[] = {
Use sparse initializers for "struct domain" and "struct protosw", so they are easier to follow for the human being.
2005-11-09 13:29:16 +00:00
{
.pr_type = SOCK_RAW,
.pr_domain = &routedomain,
.pr_flags = PR_ATOMIC|PR_ADDR,
.pr_output = route_output,
.pr_ctlinput = raw_ctlinput,
.pr_init = raw_init,
.pr_usrreqs = &route_usrreqs
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
};
Use sparse initializers for "struct domain" and "struct protosw", so they are easier to follow for the human being.
2005-11-09 13:29:16 +00:00
static struct domain routedomain = {
.dom_family = PF_ROUTE,
.dom_name = "route",
.dom_protosw = routesw,
.dom_protoswNPROTOSW = &routesw[sizeof(routesw)/sizeof(routesw[0])]
};
Updated routing-socket code from Berkeley Obtained from: Keith Bostic by way of Paul Traina
1995-05-10 16:48:52 +00:00
Make networking domains drop-ins, through the magic of GNU ld. (Some day, there may even be LKMs.) Also, change the internal name of `unixdomain' to `localdomain' since AF_LOCAL is now the preferred name of this family. Declare netisr correctly and in the right place.
1995-05-11 00:13:26 +00:00
DOMAIN_SET(route);
Reference in New Issue Copy Permalink