d/freebsd-dev
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
e606e5d157
freebsd-dev/sys/net/if_bridge.c

3771 lines
92 KiB
C
Raw Normal View History

Sync if_bridge to NetBSD r1.31 Rename conflicting variables when handling SNAP Ethernet frames. Obtained from: NetBSD Approved by: mlaier (mentor) Approved by: re (blanket)
2005-06-29 19:23:32 +00:00
/* $NetBSD: if_bridge.c,v 1.31 2005/06/01 19:45:34 jdc Exp $ */
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
sys: general adoption of SPDX licensing ID tags. Mainly focus on files that use BSD 2-Clause license, however the tool I was using misidentified many licenses so this was mostly a manual - error prone - task. The Software Package Data Exchange (SPDX) group provides a specification to make it easier for automated tools to detect and summarize well known opensource licenses. We are gradually adopting the specification, noting that the tags are considered only advisory and do not, in any way, superceed or replace the license texts. No functional change intended.
2017-11-27 15:23:17 +00:00
/*-
* SPDX-License-Identifier: BSD-4-Clause
*
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
* Copyright 2001 Wasabi Systems, Inc.
* All rights reserved.
*
* Written by Jason R. Thorpe for Wasabi Systems, Inc.
*
* 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.
* 3. All advertising materials mentioning features or use of this software
* must display the following acknowledgement:
* This product includes software developed for the NetBSD Project by
* Wasabi Systems, Inc.
* 4. The name of Wasabi Systems, Inc. may not be used to endorse
* or promote products derived from this software without specific prior
* written permission.
*
* THIS SOFTWARE IS PROVIDED BY WASABI SYSTEMS, INC. ``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 WASABI SYSTEMS, INC
* 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.
*/
/*
* Copyright (c) 1999, 2000 Jason L. Wright (jason@thought.net)
* 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.
*
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``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 AUTHOR 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.
*
* OpenBSD: if_bridge.c,v 1.60 2001/06/15 03:38:33 itojun Exp
*/
/*
* Network interface bridge support.
*
* TODO:
*
* - Currently only supports Ethernet-like interfaces (Ethernet,
* 802.11, VLANs on Ethernet, etc.) Figure out a nice way
Remove support for FDDI networks. Defines in net/if_media.h remain in case code copied from ifconfig is in use elsewere (supporting non-existant media type is harmless). Reviewed by: kib, jhb Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D15017
2018-04-11 17:28:24 +00:00
* to bridge other types of interfaces (maybe consider
* heterogeneous bridges).
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
*/
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
#include "opt_inet.h"
#include "opt_inet6.h"
#include <sys/param.h>
Include necessary headers that now are available due to pollution via if_var.h. Sponsored by: Netflix Sponsored by: Nginx, Inc.
2013-10-28 07:29:16 +00:00
#include <sys/eventhandler.h>
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
#include <sys/mbuf.h>
#include <sys/malloc.h>
#include <sys/protosw.h>
#include <sys/systm.h>
On the first loop for generating a bridge MAC address use the local hostid, this gives a good chance of keeping the same address over reboots. This is intended to help IPV6 and similar which generate their addresses from the mac. PR: kern/160300 Submitted by: mdodd Approved by: re (kib)
2011-09-04 22:06:32 +00:00
#include <sys/jail.h>
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
#include <sys/time.h>
#include <sys/socket.h> /* for net/if.h */
#include <sys/sockio.h>
#include <sys/ctype.h> /* string functions */
#include <sys/kernel.h>
#include <sys/random.h>
- Use the new bridgestp callback to once again flush our bridge routes when an interface is disabled. - Log port changes to syslog, defaulting to off
2006-08-02 03:54:28 +00:00
#include <sys/syslog.h>
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
#include <sys/sysctl.h>
#include <vm/uma.h>
#include <sys/module.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>
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
#include <sys/proc.h>
#include <sys/lock.h>
#include <sys/mutex.h>
#include <net/bpf.h>
#include <net/if.h>
#include <net/if_clone.h>
#include <net/if_dl.h>
#include <net/if_types.h>
#include <net/if_var.h>
#include <net/pfil.h>
Virtualize the pfil hooks so that different jails may chose different packet filters. ALso allows ipfw to be enabled on on ejail and disabled on another. In 8.0 it's a global setting. Sitting aroung in tree waiting to commit for: 2 months MFC after: 2 months
2009-10-11 05:59:43 +00:00
#include <net/vnet.h>
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
Remove struct arpcom. It is unused by most interface types, that allocate it, except Ethernet, where it carried ng_ether(4) pointer. For now carry the pointer in if_l2com directly. Sponsored by: Netflix Sponsored by: Nginx, Inc.
2014-11-07 15:14:10 +00:00
#include <netinet/in.h>
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
#include <netinet/in_systm.h>
#include <netinet/in_var.h>
#include <netinet/ip.h>
#include <netinet/ip_var.h>
#ifdef INET6
#include <netinet/ip6.h>
#include <netinet6/ip6_var.h>
- Allow ND6_IFF_AUTO_LINKLOCAL for IFT_BRIDGE. An interface with IFT_BRIDGE is initialized with !ND6_IFF_AUTO_LINKLOCAL && !ND6_IFF_ACCEPT_RTADV regardless of net.inet6.ip6.accept_rtadv and net.inet6.ip6.auto_linklocal. To configure an autoconfigured link-local address (RFC 4862), the following rc.conf(5) configuration can be used: ifconfig_bridge0_ipv6="inet6 auto_linklocal" - if_bridge(4) now removes IPv6 addresses on a member interface to be added when the parent interface or one of the existing member interfaces has an IPv6 address. if_bridge(4) merges each link-local scope zone which the member interfaces form respectively, so it causes address scope violation. Removal of the IPv6 addresses prevents it. - if_lagg(4) now removes IPv6 addresses on a member interfaces unconditionally. - Set reasonable flags to non-IPv6-capable interfaces. [*] Submitted by: rpaulo [*] MFC after: 1 week
2013-07-02 16:58:15 +00:00
#include <netinet6/in6_ifattach.h>
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
#endif
carp(4) allows people to share a set of IP addresses and can only use IPv4/v6 for inter-node communication (according to my reading). Properly wrap the carp callouts in INET || INET6 and refelect this in sys/conf/files as well. While in theory this should be ok, it might be a bit optimistic to think that carp could build with inet6 only[1]. Discussed with: mlaier [1]
2009-06-11 10:26:38 +00:00
#if defined(INET) || defined(INET6)
Allow bridge and carp to play nicely together by returning the packet if its destined for a carp interface. Obtained from: OpenBSD MFC after: 2 weeks
2006-06-08 23:40:16 +00:00
#include <netinet/ip_carp.h>
#endif
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
#include <machine/in_cksum.h>
Remove struct arpcom. It is unused by most interface types, that allocate it, except Ethernet, where it carried ng_ether(4) pointer. For now carry the pointer in if_l2com directly. Sponsored by: Netflix Sponsored by: Nginx, Inc.
2014-11-07 15:14:10 +00:00
#include <netinet/if_ether.h>
/tmp/cvsuusTrc
2006-07-26 10:43:02 +00:00
#include <net/bridgestp.h>
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
#include <net/if_bridgevar.h>
#include <net/if_llc.h>
Add the vlan tag to the bridge route table. This allows a vlan trunk to be bridged, previously legitimate traffic was not passed as the bridge could not tell that it was on a different Ethernet segment. All non-tagged traffic is treated as vlan1 as per IEEE 802.1Q-2003
2007-06-13 18:58:04 +00:00
#include <net/if_vlan_var.h>
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
#include <net/route.h>
Update route MTUs for bridge, lagg and vlan interfaces. Those interfaces may implicitly change their MTU on addition of parent interface in addition to normal SIOCSIFMTU ioctl path, where the route MTUs are updated normally. MFC after: 2 weeks Sponsored by: iXsystems, Inc.
2020-01-22 20:36:45 +00:00
#ifdef INET6
/*
* XXX: declare here to avoid to include many inet6 related files..
* should be more generalized?
*/
extern void nd6_setmtu(struct ifnet *);
#endif
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
/*
* Size of the route hash table. Must be a power of two.
*/
#ifndef BRIDGE_RTHASH_SIZE
#define BRIDGE_RTHASH_SIZE 1024
#endif
#define BRIDGE_RTHASH_MASK (BRIDGE_RTHASH_SIZE - 1)
/*
Use a more appropriate default for the maximum number of addresses in the bridge forwarding table. PR: docs/164564 Discussed with: brueffer
2012-02-29 20:58:21 +00:00
* Default maximum number of addresses to cache.
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
*/
#ifndef BRIDGE_RTABLE_MAX
Use a more appropriate default for the maximum number of addresses in the bridge forwarding table. PR: docs/164564 Discussed with: brueffer
2012-02-29 20:58:21 +00:00
#define BRIDGE_RTABLE_MAX 2000
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
#endif
/*
* Timeout (in seconds) for entries learned dynamically.
*/
#ifndef BRIDGE_RTABLE_TIMEOUT
#define BRIDGE_RTABLE_TIMEOUT (20 * 60) /* same as ARP */
#endif
/*
* Number of seconds between walks of the route list.
*/
#ifndef BRIDGE_RTABLE_PRUNE_PERIOD
#define BRIDGE_RTABLE_PRUNE_PERIOD (5 * 60)
#endif
Add code that clears certain capabilities from the member interface, these are restored when its removed from the bridge. At the moment we only clear IFCAP_TXCSUM. Since a locally generated packet on the bridge may be sent out any one or more interfaces it cant be assumed that every card does hardware csums. Most bridges don't generate a lot of traffic themselves so turning off offloading won't hurt, bridged packets are unaffected. Tested by: Bruce Walker (bmw borderware.com) MFC after: 5 days
2006-01-14 03:51:31 +00:00
/*
Be smarter about disabling interface capabilities. TOE/TSO/TXCSUM will only be disabled if one (or more) of the member interfaces does not support it. Always turn off LRO since we can not bridge a combined frame. Tested by: Stefan Lambrev
2008-07-03 15:58:30 +00:00
* List of capabilities to possibly mask on the member interface.
Add code that clears certain capabilities from the member interface, these are restored when its removed from the bridge. At the moment we only clear IFCAP_TXCSUM. Since a locally generated packet on the bridge may be sent out any one or more interfaces it cant be assumed that every card does hardware csums. Most bridges don't generate a lot of traffic themselves so turning off offloading won't hurt, bridged packets are unaffected. Tested by: Bruce Walker (bmw borderware.com) MFC after: 5 days
2006-01-14 03:51:31 +00:00
*/
Negotiate/disable TXCSUM_IPV6 same as TXCSUM.
2016-07-18 16:58:47 +00:00
#define BRIDGE_IFCAPS_MASK (IFCAP_TOE|IFCAP_TSO|IFCAP_TXCSUM|\
IFCAP_TXCSUM_IPV6)
Add code that clears certain capabilities from the member interface, these are restored when its removed from the bridge. At the moment we only clear IFCAP_TXCSUM. Since a locally generated packet on the bridge may be sent out any one or more interfaces it cant be assumed that every card does hardware csums. Most bridges don't generate a lot of traffic themselves so turning off offloading won't hurt, bridged packets are unaffected. Tested by: Bruce Walker (bmw borderware.com) MFC after: 5 days
2006-01-14 03:51:31 +00:00
When bridging LRO is causing a problem, the believe that it would work as long as all interfaces have TSO seems to be false, until the matter gets sorted out just disable LRO completely.
2009-08-24 21:04:51 +00:00
/*
* List of capabilities to strip
*/
#define BRIDGE_IFCAPS_STRIP IFCAP_LRO
bridge: Move locking defines into if_bridge.c The locking defines for if_bridge used to live in if_bridgevar.h, but they're only ever used by the bridge implementation itself (in if_bridge.c). Moving them into the .c file. Reported by: philip, emaste Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D23808
2020-02-26 08:47:18 +00:00
/*
* Bridge locking
bridge: call member interface ioctl() without NET_EPOCH We're not allowed to hold NET_EPOCH while sleeping, so when we call ioctl() handlers for member interfaces we cannot be in NET_EPOCH. We still need some protection of our CK_LISTs, so hold BRIDGE_LOCK instead. That requires changing BRIDGE_LOCK into a sleepable lock, and separating the BRIDGE_RT_LOCK, to protect bridge_rtnode lists. That lock is taken in the data path (while in NET_EPOCH), so it cannot be a sleepable lock. While here document the locking strategy. MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D26418
2020-10-06 19:19:56 +00:00
*
* The bridge relies heavily on the epoch(9) system to protect its data
* structures. This means we can safely use CK_LISTs while in NET_EPOCH, but we
* must ensure there is only one writer at a time.
*
* That is: for read accesses we only need to be in NET_EPOCH, but for write
* accesses we must hold:
*
* - BRIDGE_RT_LOCK, for any change to bridge_rtnodes
* - BRIDGE_LOCK, for any other change
*
* The BRIDGE_LOCK is a sleepable lock, because it is held accross ioctl()
* calls to bridge member interfaces and these ioctl()s can sleep.
* The BRIDGE_RT_LOCK is a non-sleepable mutex, because it is sometimes
* required while we're in NET_EPOCH and then we're not allowed to sleep.
bridge: Move locking defines into if_bridge.c The locking defines for if_bridge used to live in if_bridgevar.h, but they're only ever used by the bridge implementation itself (in if_bridge.c). Moving them into the .c file. Reported by: philip, emaste Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D23808
2020-02-26 08:47:18 +00:00
*/
#define BRIDGE_LOCK_INIT(_sc) do { \
bridge: call member interface ioctl() without NET_EPOCH We're not allowed to hold NET_EPOCH while sleeping, so when we call ioctl() handlers for member interfaces we cannot be in NET_EPOCH. We still need some protection of our CK_LISTs, so hold BRIDGE_LOCK instead. That requires changing BRIDGE_LOCK into a sleepable lock, and separating the BRIDGE_RT_LOCK, to protect bridge_rtnode lists. That lock is taken in the data path (while in NET_EPOCH), so it cannot be a sleepable lock. While here document the locking strategy. MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D26418
2020-10-06 19:19:56 +00:00
sx_init(&(_sc)->sc_sx, "if_bridge"); \
mtx_init(&(_sc)->sc_rt_mtx, "if_bridge rt", NULL, MTX_DEF); \
bridge: Move locking defines into if_bridge.c The locking defines for if_bridge used to live in if_bridgevar.h, but they're only ever used by the bridge implementation itself (in if_bridge.c). Moving them into the .c file. Reported by: philip, emaste Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D23808
2020-02-26 08:47:18 +00:00
} while (0)
#define BRIDGE_LOCK_DESTROY(_sc) do { \
bridge: call member interface ioctl() without NET_EPOCH We're not allowed to hold NET_EPOCH while sleeping, so when we call ioctl() handlers for member interfaces we cannot be in NET_EPOCH. We still need some protection of our CK_LISTs, so hold BRIDGE_LOCK instead. That requires changing BRIDGE_LOCK into a sleepable lock, and separating the BRIDGE_RT_LOCK, to protect bridge_rtnode lists. That lock is taken in the data path (while in NET_EPOCH), so it cannot be a sleepable lock. While here document the locking strategy. MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D26418
2020-10-06 19:19:56 +00:00
sx_destroy(&(_sc)->sc_sx); \
mtx_destroy(&(_sc)->sc_rt_mtx); \
bridge: Move locking defines into if_bridge.c The locking defines for if_bridge used to live in if_bridgevar.h, but they're only ever used by the bridge implementation itself (in if_bridge.c). Moving them into the .c file. Reported by: philip, emaste Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D23808
2020-02-26 08:47:18 +00:00
} while (0)
bridge: call member interface ioctl() without NET_EPOCH We're not allowed to hold NET_EPOCH while sleeping, so when we call ioctl() handlers for member interfaces we cannot be in NET_EPOCH. We still need some protection of our CK_LISTs, so hold BRIDGE_LOCK instead. That requires changing BRIDGE_LOCK into a sleepable lock, and separating the BRIDGE_RT_LOCK, to protect bridge_rtnode lists. That lock is taken in the data path (while in NET_EPOCH), so it cannot be a sleepable lock. While here document the locking strategy. MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D26418
2020-10-06 19:19:56 +00:00
#define BRIDGE_LOCK(_sc) sx_xlock(&(_sc)->sc_sx)
#define BRIDGE_UNLOCK(_sc) sx_xunlock(&(_sc)->sc_sx)
#define BRIDGE_LOCK_ASSERT(_sc) sx_assert(&(_sc)->sc_sx, SX_XLOCKED)
#define BRIDGE_LOCK_OR_NET_EPOCH_ASSERT(_sc) \
MPASS(in_epoch(net_epoch_preempt) || sx_xlocked(&(_sc)->sc_sx))
#define BRIDGE_UNLOCK_ASSERT(_sc) sx_assert(&(_sc)->sc_sx, SX_UNLOCKED)
#define BRIDGE_RT_LOCK(_sc) mtx_lock(&(_sc)->sc_rt_mtx)
#define BRIDGE_RT_UNLOCK(_sc) mtx_unlock(&(_sc)->sc_rt_mtx)
#define BRIDGE_RT_LOCK_ASSERT(_sc) mtx_assert(&(_sc)->sc_rt_mtx, MA_OWNED)
#define BRIDGE_RT_LOCK_OR_NET_EPOCH_ASSERT(_sc) \
MPASS(in_epoch(net_epoch_preempt) || mtx_owned(&(_sc)->sc_rt_mtx))
bridge: Move locking defines into if_bridge.c The locking defines for if_bridge used to live in if_bridgevar.h, but they're only ever used by the bridge implementation itself (in if_bridge.c). Moving them into the .c file. Reported by: philip, emaste Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D23808
2020-02-26 08:47:18 +00:00
Remove the dependency of bridgestp.h on if_bridgevar.h by moving a couple of private structures to if_bridge.c.
2006-07-27 21:01:48 +00:00
/*
* Bridge interface list entry.
*/
struct bridge_iflist {
bridge: Change lists to CK_LIST as a peparation for epochification Prepare the ground for a rework of the bridge locking approach. We will use an epoch-based approach in the datapath and making it safe to iterate over the interface, span and rtnode lists without holding the BRIDGE_LOCK. Replace the relevant lists by their ConcurrencyKit equivalents. No functional change in this commit. Reviewed by: emaste, ae, philip (previous version) Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24249
2020-04-05 17:15:20 +00:00
CK_LIST_ENTRY(bridge_iflist) bif_next;
Remove the dependency of bridgestp.h on if_bridgevar.h by moving a couple of private structures to if_bridge.c.
2006-07-27 21:01:48 +00:00
struct ifnet *bif_ifp; /* member if */
struct bstp_port bif_stp; /* STP state */
uint32_t bif_flags; /* member if flags */
Be smarter about disabling interface capabilities. TOE/TSO/TXCSUM will only be disabled if one (or more) of the member interfaces does not support it. Always turn off LRO since we can not bridge a combined frame. Tested by: Stefan Lambrev
2008-07-03 15:58:30 +00:00
int bif_savedcaps; /* saved capabilities */
Add an option to limit the number of source MACs that can be behind a bridge interface. Once the limit is reached packets with unknown source addresses are dropped until an existing host cache entry expires or is removed. Useful to use with the STICKY cache option. Sponsored by: miniSuperHappyDevHouse NZ
2007-11-04 08:32:27 +00:00
uint32_t bif_addrmax; /* max # of addresses */
uint32_t bif_addrcnt; /* cur. # of addresses */
uint32_t bif_addrexceeded;/* # of address violations */
bridge: epoch-ification Run the bridge datapath under epoch, rather than under the BRIDGE_LOCK(). We still take the BRIDGE_LOCK() whenever we insert or delete items in the relevant lists, but we use epoch callbacks to free items so that it's safe to iterate the lists without the BRIDGE_LOCK. Tests on mercat5/6 shows this increases bridge throughput significantly, from 3.7Mpps to 18.6Mpps. Reviewed by: emaste, philip, melifaro MFC after: 2 months Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24250
2020-04-26 16:22:35 +00:00
struct epoch_context bif_epoch_ctx;
Remove the dependency of bridgestp.h on if_bridgevar.h by moving a couple of private structures to if_bridge.c.
2006-07-27 21:01:48 +00:00
};
/*
* Bridge route node.
*/
struct bridge_rtnode {
bridge: Change lists to CK_LIST as a peparation for epochification Prepare the ground for a rework of the bridge locking approach. We will use an epoch-based approach in the datapath and making it safe to iterate over the interface, span and rtnode lists without holding the BRIDGE_LOCK. Replace the relevant lists by their ConcurrencyKit equivalents. No functional change in this commit. Reviewed by: emaste, ae, philip (previous version) Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24249
2020-04-05 17:15:20 +00:00
CK_LIST_ENTRY(bridge_rtnode) brt_hash; /* hash table linkage */
CK_LIST_ENTRY(bridge_rtnode) brt_list; /* list linkage */
Add an option to limit the number of source MACs that can be behind a bridge interface. Once the limit is reached packets with unknown source addresses are dropped until an existing host cache entry expires or is removed. Useful to use with the STICKY cache option. Sponsored by: miniSuperHappyDevHouse NZ
2007-11-04 08:32:27 +00:00
struct bridge_iflist *brt_dst; /* destination if */
Remove the dependency of bridgestp.h on if_bridgevar.h by moving a couple of private structures to if_bridge.c.
2006-07-27 21:01:48 +00:00
unsigned long brt_expire; /* expiration time */
uint8_t brt_flags; /* address flags */
uint8_t brt_addr[ETHER_ADDR_LEN];
Add the vlan tag to the bridge route table. This allows a vlan trunk to be bridged, previously legitimate traffic was not passed as the bridge could not tell that it was on a different Ethernet segment. All non-tagged traffic is treated as vlan1 as per IEEE 802.1Q-2003
2007-06-13 18:58:04 +00:00
uint16_t brt_vlan; /* vlan id */
bridge: epoch-ification Run the bridge datapath under epoch, rather than under the BRIDGE_LOCK(). We still take the BRIDGE_LOCK() whenever we insert or delete items in the relevant lists, but we use epoch callbacks to free items so that it's safe to iterate the lists without the BRIDGE_LOCK. Tests on mercat5/6 shows this increases bridge throughput significantly, from 3.7Mpps to 18.6Mpps. Reviewed by: emaste, philip, melifaro MFC after: 2 months Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24250
2020-04-26 16:22:35 +00:00
struct vnet *brt_vnet;
struct epoch_context brt_epoch_ctx;
Remove the dependency of bridgestp.h on if_bridgevar.h by moving a couple of private structures to if_bridge.c.
2006-07-27 21:01:48 +00:00
};
Add an option to limit the number of source MACs that can be behind a bridge interface. Once the limit is reached packets with unknown source addresses are dropped until an existing host cache entry expires or is removed. Useful to use with the STICKY cache option. Sponsored by: miniSuperHappyDevHouse NZ
2007-11-04 08:32:27 +00:00
#define brt_ifp brt_dst->bif_ifp
Remove the dependency of bridgestp.h on if_bridgevar.h by moving a couple of private structures to if_bridge.c.
2006-07-27 21:01:48 +00:00
/*
* Software state for each bridge.
*/
struct bridge_softc {
struct ifnet *sc_ifp; /* make this an interface */
LIST_ENTRY(bridge_softc) sc_list;
bridge: call member interface ioctl() without NET_EPOCH We're not allowed to hold NET_EPOCH while sleeping, so when we call ioctl() handlers for member interfaces we cannot be in NET_EPOCH. We still need some protection of our CK_LISTs, so hold BRIDGE_LOCK instead. That requires changing BRIDGE_LOCK into a sleepable lock, and separating the BRIDGE_RT_LOCK, to protect bridge_rtnode lists. That lock is taken in the data path (while in NET_EPOCH), so it cannot be a sleepable lock. While here document the locking strategy. MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D26418
2020-10-06 19:19:56 +00:00
struct sx sc_sx;
struct mtx sc_rt_mtx;
Remove the dependency of bridgestp.h on if_bridgevar.h by moving a couple of private structures to if_bridge.c.
2006-07-27 21:01:48 +00:00
uint32_t sc_brtmax; /* max # of addresses */
uint32_t sc_brtcnt; /* cur. # of addresses */
uint32_t sc_brttimeout; /* rt timeout in seconds */
struct callout sc_brcallout; /* bridge callout */
bridge: Change lists to CK_LIST as a peparation for epochification Prepare the ground for a rework of the bridge locking approach. We will use an epoch-based approach in the datapath and making it safe to iterate over the interface, span and rtnode lists without holding the BRIDGE_LOCK. Replace the relevant lists by their ConcurrencyKit equivalents. No functional change in this commit. Reviewed by: emaste, ae, philip (previous version) Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24249
2020-04-05 17:15:20 +00:00
CK_LIST_HEAD(, bridge_iflist) sc_iflist; /* member interface list */
CK_LIST_HEAD(, bridge_rtnode) *sc_rthash; /* our forwarding table */
CK_LIST_HEAD(, bridge_rtnode) sc_rtlist; /* list version of above */
Remove the dependency of bridgestp.h on if_bridgevar.h by moving a couple of private structures to if_bridge.c.
2006-07-27 21:01:48 +00:00
uint32_t sc_rthash_key; /* key for hash */
bridge: Change lists to CK_LIST as a peparation for epochification Prepare the ground for a rework of the bridge locking approach. We will use an epoch-based approach in the datapath and making it safe to iterate over the interface, span and rtnode lists without holding the BRIDGE_LOCK. Replace the relevant lists by their ConcurrencyKit equivalents. No functional change in this commit. Reviewed by: emaste, ae, philip (previous version) Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24249
2020-04-05 17:15:20 +00:00
CK_LIST_HEAD(, bridge_iflist) sc_spanlist; /* span ports list */
Remove the dependency of bridgestp.h on if_bridgevar.h by moving a couple of private structures to if_bridge.c.
2006-07-27 21:01:48 +00:00
struct bstp_state sc_stp; /* STP state */
Add some statistics that are needed to support RFC4188 as part of the SoC2006 work on a bridge monitoring module for BSNMP. Submitted by: shteryana (SoC 2006)
2006-07-31 20:24:46 +00:00
uint32_t sc_brtexceeded; /* # of cache drops */
bridge_delete_member is called via the event handler from if_detach after the LLADDR is reclaimed which causes a null pointer deref with inherit_mac enabled. Record the ifnet pointer of the interface and then compare that to find when to re-assign the bridge address. Submitted by: sam
2009-02-13 19:20:25 +00:00
struct ifnet *sc_ifaddr; /* member mac copied from */
ether: centralize fake hwaddr generation We currently have two places with identical fake hwaddr generation -- if_vxlan and if_bridge. Lift it into if_ethersubr for reuse in other interfaces that may also need a fake addr. Reviewed by: bryanv, kp, philip Differential Revision: https://reviews.freebsd.org/D19573
2019-03-14 17:18:00 +00:00
struct ether_addr sc_defaddr; /* Default MAC address */
bridge: epoch-ification Run the bridge datapath under epoch, rather than under the BRIDGE_LOCK(). We still take the BRIDGE_LOCK() whenever we insert or delete items in the relevant lists, but we use epoch callbacks to free items so that it's safe to iterate the lists without the BRIDGE_LOCK. Tests on mercat5/6 shows this increases bridge throughput significantly, from 3.7Mpps to 18.6Mpps. Reviewed by: emaste, philip, melifaro MFC after: 2 months Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24250
2020-04-26 16:22:35 +00:00
struct epoch_context sc_epoch_ctx;
Remove the dependency of bridgestp.h on if_bridgevar.h by moving a couple of private structures to if_bridge.c.
2006-07-27 21:01:48 +00:00
};
bridge: call member interface ioctl() without NET_EPOCH We're not allowed to hold NET_EPOCH while sleeping, so when we call ioctl() handlers for member interfaces we cannot be in NET_EPOCH. We still need some protection of our CK_LISTs, so hold BRIDGE_LOCK instead. That requires changing BRIDGE_LOCK into a sleepable lock, and separating the BRIDGE_RT_LOCK, to protect bridge_rtnode lists. That lock is taken in the data path (while in NET_EPOCH), so it cannot be a sleepable lock. While here document the locking strategy. MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D26418
2020-10-06 19:19:56 +00:00
VNET_DEFINE_STATIC(struct sx, bridge_list_sx);
#define V_bridge_list_sx VNET(bridge_list_sx)
Fix a panic when tearing down a vnet on a VIMAGE-enabled kernel. There was a race that bridge_ifdetach() could be called via ifnet_departure event handler after vnet_bridge_uninit(). PR: 195859 Reported by: Danilo Egea Gondolfo
2015-02-14 18:15:14 +00:00
static eventhandler_tag bridge_detach_cookie;
Change from a callback in if_ethersubr to using EVENTHANDLER in order to detach span ports when they disappear. The span port does not have a pointer to the softc so revert r1.31 and bring back the softc linked-list. MFC after: 2 weeks
2005-12-17 06:33:51 +00:00
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
int bridge_rtable_prune_period = BRIDGE_RTABLE_PRUNE_PERIOD;
if_bridge(4): Fix module teardown bridge_rtnode_zone still has outstanding allocations at the time of destruction in the current model because all of the interface teardown happens in a VNET_SYSUNINIT, -after- the MOD_UNLOAD has already been processed. The SYSUNINIT triggers destruction of the interfaces, which then attempts to free the memory from the zone that's already been destroyed, and we hit a panic. Solve this by virtualizing the uma_zone we allocate the rtnodes from to fix the ordering. bridge_rtable_fini should also take care to flush any remaining routes that weren't taken care of when dynamic routes were flushed in bridge_stop. Reviewed by: kp Differential Revision: https://reviews.freebsd.org/D19578
2019-03-15 13:19:52 +00:00
VNET_DEFINE_STATIC(uma_zone_t, bridge_rtnode_zone);
#define V_bridge_rtnode_zone VNET(bridge_rtnode_zone)
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
Revise network interface cloning to take an optional opaque parameter that can specify configuration parameters: o rev cloner api's to add optional parameter block o add SIOCCREATE2 that accepts parameter data o rev vlan support to use new api (maintain old code) Reviewed by: arch@
2006-07-09 06:04:01 +00:00
static int bridge_clone_create(struct if_clone *, int, caddr_t);
Change most of the bridge and stp funtions to static. This has highlighted that the following funtions are not used, wrap in '#ifdef noused' for the moment. bstp_enable_change_detection bstp_disable_change_detection bstp_set_bridge_priority bstp_set_port_priority bstp_set_path_cost
2005-10-14 10:38:12 +00:00
static void bridge_clone_destroy(struct ifnet *);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
Change most of the bridge and stp funtions to static. This has highlighted that the following funtions are not used, wrap in '#ifdef noused' for the moment. bstp_enable_change_detection bstp_disable_change_detection bstp_set_bridge_priority bstp_set_port_priority bstp_set_path_cost
2005-10-14 10:38:12 +00:00
static int bridge_ioctl(struct ifnet *, u_long, caddr_t);
Be smarter about disabling interface capabilities. TOE/TSO/TXCSUM will only be disabled if one (or more) of the member interfaces does not support it. Always turn off LRO since we can not bridge a combined frame. Tested by: Stefan Lambrev
2008-07-03 15:58:30 +00:00
static void bridge_mutecaps(struct bridge_softc *);
static void bridge_set_ifcap(struct bridge_softc *, struct bridge_iflist *,
int);
Change from a callback in if_ethersubr to using EVENTHANDLER in order to detach span ports when they disappear. The span port does not have a pointer to the softc so revert r1.31 and bring back the softc linked-list. MFC after: 2 weeks
2005-12-17 06:33:51 +00:00
static void bridge_ifdetach(void *arg __unused, struct ifnet *);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
static void bridge_init(void *);
Change most of the bridge and stp funtions to static. This has highlighted that the following funtions are not used, wrap in '#ifdef noused' for the moment. bstp_enable_change_detection bstp_disable_change_detection bstp_set_bridge_priority bstp_set_port_priority bstp_set_path_cost
2005-10-14 10:38:12 +00:00
static void bridge_dummynet(struct mbuf *, struct ifnet *);
static void bridge_stop(struct ifnet *, int);
Change bridge(4) to use if_transmit for forwarding packets to underlying interfaces instead of queueing. Tested by: ray
2012-09-03 10:08:20 +00:00
static int bridge_transmit(struct ifnet *, struct mbuf *);
if_bridge: add ALTQ support Similar to the recent addition of ALTQ support to if_vlan. Reviewed by: donner Obtained from: pfsense MFC after: 1 week Sponsored by: Rubicon Communications, LLC ("Netgate") Differential Revision: https://reviews.freebsd.org/D31675
2021-08-25 09:07:44 +00:00
#ifdef ALTQ
static void bridge_altq_start(if_t);
static int bridge_altq_transmit(if_t, struct mbuf *);
#endif
Change bridge(4) to use if_transmit for forwarding packets to underlying interfaces instead of queueing. Tested by: ray
2012-09-03 10:08:20 +00:00
static void bridge_qflush(struct ifnet *);
Change most of the bridge and stp funtions to static. This has highlighted that the following funtions are not used, wrap in '#ifdef noused' for the moment. bstp_enable_change_detection bstp_disable_change_detection bstp_set_bridge_priority bstp_set_port_priority bstp_set_path_cost
2005-10-14 10:38:12 +00:00
static struct mbuf *bridge_input(struct ifnet *, struct mbuf *);
static int bridge_output(struct ifnet *, struct mbuf *, struct sockaddr *,
Unbreak build by reverting if_bridge part of r360047. Pointy hat to: melifaro
2020-04-17 18:22:37 +00:00
struct rtentry *);
Change bridge(4) to use if_transmit for forwarding packets to underlying interfaces instead of queueing. Tested by: ray
2012-09-03 10:08:20 +00:00
static int bridge_enqueue(struct bridge_softc *, struct ifnet *,
Remove the dependency of bridgestp.h on if_bridgevar.h by moving a couple of private structures to if_bridge.c.
2006-07-27 21:01:48 +00:00
struct mbuf *);
static void bridge_rtdelete(struct bridge_softc *, struct ifnet *ifp, int);
Change most of the bridge and stp funtions to static. This has highlighted that the following funtions are not used, wrap in '#ifdef noused' for the moment. bstp_enable_change_detection bstp_disable_change_detection bstp_set_bridge_priority bstp_set_port_priority bstp_set_path_cost
2005-10-14 10:38:12 +00:00
Add a bridge interface flag called PRIVATE where any private port can not communicate with another private port. All unicast/broadcast/multicast layer2 traffic is blocked so it works much the same way as using firewall rules but scales better and is generally easier as firewall packages usually do not allow ARP blocking. An example usage would be having a number of customers on separate vlans bridged with a server network. All the vlans are marked private, they can all communicate with the server network unhindered, but can not exchange any traffic whatsoever with each other. Approved by: re (rwatson)
2007-08-01 00:33:52 +00:00
static void bridge_forward(struct bridge_softc *, struct bridge_iflist *,
struct mbuf *m);
Change most of the bridge and stp funtions to static. This has highlighted that the following funtions are not used, wrap in '#ifdef noused' for the moment. bstp_enable_change_detection bstp_disable_change_detection bstp_set_bridge_priority bstp_set_port_priority bstp_set_path_cost
2005-10-14 10:38:12 +00:00
static void bridge_timer(void *);
static void bridge_broadcast(struct bridge_softc *, struct ifnet *,
struct mbuf *, int);
Add support for creating span ports so that one can snoop bridged traffic from another interface/machine/network. Obtained from: OpenBSD MFC after: 2 weeks
2005-12-14 02:52:13 +00:00
static void bridge_span(struct bridge_softc *, struct mbuf *);
Change most of the bridge and stp funtions to static. This has highlighted that the following funtions are not used, wrap in '#ifdef noused' for the moment. bstp_enable_change_detection bstp_disable_change_detection bstp_set_bridge_priority bstp_set_port_priority bstp_set_path_cost
2005-10-14 10:38:12 +00:00
static int bridge_rtupdate(struct bridge_softc *, const uint8_t *,
Add the vlan tag to the bridge route table. This allows a vlan trunk to be bridged, previously legitimate traffic was not passed as the bridge could not tell that it was on a different Ethernet segment. All non-tagged traffic is treated as vlan1 as per IEEE 802.1Q-2003
2007-06-13 18:58:04 +00:00
uint16_t, struct bridge_iflist *, int, uint8_t);
static struct ifnet *bridge_rtlookup(struct bridge_softc *, const uint8_t *,
uint16_t);
Change most of the bridge and stp funtions to static. This has highlighted that the following funtions are not used, wrap in '#ifdef noused' for the moment. bstp_enable_change_detection bstp_disable_change_detection bstp_set_bridge_priority bstp_set_port_priority bstp_set_path_cost
2005-10-14 10:38:12 +00:00
static void bridge_rttrim(struct bridge_softc *);
static void bridge_rtage(struct bridge_softc *);
static void bridge_rtflush(struct bridge_softc *, int);
Add the vlan tag to the bridge route table. This allows a vlan trunk to be bridged, previously legitimate traffic was not passed as the bridge could not tell that it was on a different Ethernet segment. All non-tagged traffic is treated as vlan1 as per IEEE 802.1Q-2003
2007-06-13 18:58:04 +00:00
static int bridge_rtdaddr(struct bridge_softc *, const uint8_t *,
uint16_t);
Change most of the bridge and stp funtions to static. This has highlighted that the following funtions are not used, wrap in '#ifdef noused' for the moment. bstp_enable_change_detection bstp_disable_change_detection bstp_set_bridge_priority bstp_set_port_priority bstp_set_path_cost
2005-10-14 10:38:12 +00:00
Remove the M_NOWAIT from bridge_rtable_init as it isn't needed. The function return value is not even checked and could lead to a panic on a null sc_rthash. MFC after: 2 weeks
2012-10-04 07:40:55 +00:00
static void bridge_rtable_init(struct bridge_softc *);
Change most of the bridge and stp funtions to static. This has highlighted that the following funtions are not used, wrap in '#ifdef noused' for the moment. bstp_enable_change_detection bstp_disable_change_detection bstp_set_bridge_priority bstp_set_port_priority bstp_set_path_cost
2005-10-14 10:38:12 +00:00
static void bridge_rtable_fini(struct bridge_softc *);
Fix two bugs with the bridge - code expects memcmp() to return a signed value, our memcmp() returns 0 if args are equal and > 0 if not. - It's possible to hijack interface for static entry. If bridge recieves packet from interface marked as learning it will replace the bridge_rtnode entry for the source address even if such entry marked as static. Submitted by: Gleb Kurtsov <k-gleb yandex.ru> MFC after: 3 days
2006-01-31 21:21:28 +00:00
static int bridge_rtnode_addr_cmp(const uint8_t *, const uint8_t *);
Change most of the bridge and stp funtions to static. This has highlighted that the following funtions are not used, wrap in '#ifdef noused' for the moment. bstp_enable_change_detection bstp_disable_change_detection bstp_set_bridge_priority bstp_set_port_priority bstp_set_path_cost
2005-10-14 10:38:12 +00:00
static struct bridge_rtnode *bridge_rtnode_lookup(struct bridge_softc *,
Add the vlan tag to the bridge route table. This allows a vlan trunk to be bridged, previously legitimate traffic was not passed as the bridge could not tell that it was on a different Ethernet segment. All non-tagged traffic is treated as vlan1 as per IEEE 802.1Q-2003
2007-06-13 18:58:04 +00:00
const uint8_t *, uint16_t);
Change most of the bridge and stp funtions to static. This has highlighted that the following funtions are not used, wrap in '#ifdef noused' for the moment. bstp_enable_change_detection bstp_disable_change_detection bstp_set_bridge_priority bstp_set_port_priority bstp_set_path_cost
2005-10-14 10:38:12 +00:00
static int bridge_rtnode_insert(struct bridge_softc *,
struct bridge_rtnode *);
static void bridge_rtnode_destroy(struct bridge_softc *,
struct bridge_rtnode *);
Bring in support for the Rapid Spanning Tree Protocol (802.1w). RSTP provides faster spanning tree convergence, the protocol will exchange information with neighboring switches to quickly transition to forwarding without creating loops. The code will default to RSTP mode but will downgrade any port connected to a legacy STP network so is fully backward compatible. Reviewed by: syrinx Tested by: syrinx
2006-11-01 09:07:47 +00:00
static void bridge_rtable_expire(struct ifnet *, int);
- Use the new bridgestp callback to once again flush our bridge routes when an interface is disabled. - Log port changes to syslog, defaulting to off
2006-08-02 03:54:28 +00:00
static void bridge_state_change(struct ifnet *, int);
Change most of the bridge and stp funtions to static. This has highlighted that the following funtions are not used, wrap in '#ifdef noused' for the moment. bstp_enable_change_detection bstp_disable_change_detection bstp_set_bridge_priority bstp_set_port_priority bstp_set_path_cost
2005-10-14 10:38:12 +00:00
static struct bridge_iflist *bridge_lookup_member(struct bridge_softc *,
const char *name);
static struct bridge_iflist *bridge_lookup_member_if(struct bridge_softc *,
struct ifnet *ifp);
static void bridge_delete_member(struct bridge_softc *,
If we have been called from ether_ifdetach() then do not try and clear the promisc flag from the member interface, this is a no-op anyway since the interface is disappearing. The driver may have already released its resources such as miibus and this is likely to panic the kernel. Submitted and tested by: Wojciech A. Koszek MFC after: 2 weeks
2005-10-23 22:30:07 +00:00
struct bridge_iflist *, int);
Change from a callback in if_ethersubr to using EVENTHANDLER in order to detach span ports when they disappear. The span port does not have a pointer to the softc so revert r1.31 and bring back the softc linked-list. MFC after: 2 weeks
2005-12-17 06:33:51 +00:00
static void bridge_delete_span(struct bridge_softc *,
struct bridge_iflist *);
Change most of the bridge and stp funtions to static. This has highlighted that the following funtions are not used, wrap in '#ifdef noused' for the moment. bstp_enable_change_detection bstp_disable_change_detection bstp_set_bridge_priority bstp_set_port_priority bstp_set_path_cost
2005-10-14 10:38:12 +00:00
static int bridge_ioctl_add(struct bridge_softc *, void *);
static int bridge_ioctl_del(struct bridge_softc *, void *);
static int bridge_ioctl_gifflags(struct bridge_softc *, void *);
static int bridge_ioctl_sifflags(struct bridge_softc *, void *);
static int bridge_ioctl_scache(struct bridge_softc *, void *);
static int bridge_ioctl_gcache(struct bridge_softc *, void *);
static int bridge_ioctl_gifs(struct bridge_softc *, void *);
static int bridge_ioctl_rts(struct bridge_softc *, void *);
static int bridge_ioctl_saddr(struct bridge_softc *, void *);
static int bridge_ioctl_sto(struct bridge_softc *, void *);
static int bridge_ioctl_gto(struct bridge_softc *, void *);
static int bridge_ioctl_daddr(struct bridge_softc *, void *);
static int bridge_ioctl_flush(struct bridge_softc *, void *);
static int bridge_ioctl_gpri(struct bridge_softc *, void *);
static int bridge_ioctl_spri(struct bridge_softc *, void *);
static int bridge_ioctl_ght(struct bridge_softc *, void *);
static int bridge_ioctl_sht(struct bridge_softc *, void *);
static int bridge_ioctl_gfd(struct bridge_softc *, void *);
static int bridge_ioctl_sfd(struct bridge_softc *, void *);
static int bridge_ioctl_gma(struct bridge_softc *, void *);
static int bridge_ioctl_sma(struct bridge_softc *, void *);
static int bridge_ioctl_sifprio(struct bridge_softc *, void *);
static int bridge_ioctl_sifcost(struct bridge_softc *, void *);
Add an option to limit the number of source MACs that can be behind a bridge interface. Once the limit is reached packets with unknown source addresses are dropped until an existing host cache entry expires or is removed. Useful to use with the STICKY cache option. Sponsored by: miniSuperHappyDevHouse NZ
2007-11-04 08:32:27 +00:00
static int bridge_ioctl_sifmaxaddr(struct bridge_softc *, void *);
Add support for creating span ports so that one can snoop bridged traffic from another interface/machine/network. Obtained from: OpenBSD MFC after: 2 weeks
2005-12-14 02:52:13 +00:00
static int bridge_ioctl_addspan(struct bridge_softc *, void *);
static int bridge_ioctl_delspan(struct bridge_softc *, void *);
Add some statistics that are needed to support RFC4188 as part of the SoC2006 work on a bridge monitoring module for BSNMP. Submitted by: shteryana (SoC 2006)
2006-07-31 20:24:46 +00:00
static int bridge_ioctl_gbparam(struct bridge_softc *, void *);
static int bridge_ioctl_grte(struct bridge_softc *, void *);
static int bridge_ioctl_gifsstp(struct bridge_softc *, void *);
Bring in support for the Rapid Spanning Tree Protocol (802.1w). RSTP provides faster spanning tree convergence, the protocol will exchange information with neighboring switches to quickly transition to forwarding without creating loops. The code will default to RSTP mode but will downgrade any port connected to a legacy STP network so is fully backward compatible. Reviewed by: syrinx Tested by: syrinx
2006-11-01 09:07:47 +00:00
static int bridge_ioctl_sproto(struct bridge_softc *, void *);
static int bridge_ioctl_stxhc(struct bridge_softc *, void *);
Change most of the bridge and stp funtions to static. This has highlighted that the following funtions are not used, wrap in '#ifdef noused' for the moment. bstp_enable_change_detection bstp_disable_change_detection bstp_set_bridge_priority bstp_set_port_priority bstp_set_path_cost
2005-10-14 10:38:12 +00:00
static int bridge_pfil(struct mbuf **, struct ifnet *, struct ifnet *,
int);
static int bridge_ip_checkbasic(struct mbuf **mp);
Fix style(9) nits, whitespace and parentheses.
2006-05-16 22:50:41 +00:00
#ifdef INET6
Change most of the bridge and stp funtions to static. This has highlighted that the following funtions are not used, wrap in '#ifdef noused' for the moment. bstp_enable_change_detection bstp_disable_change_detection bstp_set_bridge_priority bstp_set_port_priority bstp_set_path_cost
2005-10-14 10:38:12 +00:00
static int bridge_ip6_checkbasic(struct mbuf **mp);
Fix style(9) nits, whitespace and parentheses.
2006-05-16 22:50:41 +00:00
#endif /* INET6 */
bridge: Fix fragment handling and memory leak Fragmented UDP and ICMP packets were corrupted if a firewall with reassembling feature (like pf'scrub) is enabled on the bridge. This patch fixes corrupted packet problem and the panic (triggered easly with low RAM) as explain in PR 185633. bridge_pfil and bridge_fragment relationship: bridge_pfil() receive (IN direction) packets and sent it to the firewall The firewall can be configured for reassembling fragmented packet (like pf'scrubing) in one mbuf chain when bridge_pfil() need to send this reassembled packet to the outgoing interface, it needs to re-fragment it by using bridge_fragment() bridge_fragment() had to split this mbuf (using ip_fragment) first then had to M_PREPEND each packet in the mbuf chain for adding Ethernet header. But M_PREPEND can sometime create a new mbuf on the begining of the mbuf chain, then the "main" pointer of this mbuf chain should be updated and this case is tottaly forgotten. The original bridge_fragment code (Revision 158140, 2006 April 29) came from OpenBSD, and the call to bridge_enqueue was embedded. But on FreeBSD, bridge_enqueue() is done after bridge_fragment(), then the original OpenBSD code can't work as-it of FreeBSD. PR: 185633 Submitted by: Olivier Cochard-Labbé Differential Revision: https://reviews.freebsd.org/D7780
2016-09-24 07:09:43 +00:00
static int bridge_fragment(struct ifnet *, struct mbuf **mp,
Add support for fragmenting ipv4 packets. The packet filter may reassemble the ip fragments and return a packet that is larger than the MTU of the sending interface. There is no check for DF or icmp replies as we can only get a large packet to fragment by reassembling a previous fragment, and this only happens after a call to pfil(9). Obtained from: OpenBSD (mostly) Glanced at by: mlaier MFC after: 1 month
2006-04-29 05:37:25 +00:00
struct ether_header *, int, struct llc *);
Add linkstate to bridge(4), set the link to up when at least one underlying interface is up, otherwise the link is down. This, among other things, allows carp to work on a bridge. Prodded by: glebius Tested by: Alexander Lunev
2012-04-20 09:55:50 +00:00
static void bridge_linkstate(struct ifnet *ifp);
Fix a panic I introduced in r234487, the bridge softc pointer is set to null early in the detach so rearrange things not to explode. Reported by: David Roffiaen, Gustau Perez Querol Tested by: David Roffiaen MFC after: 3 days
2012-06-11 20:12:13 +00:00
static void bridge_linkcheck(struct bridge_softc *sc);
Add linkstate to bridge(4), set the link to up when at least one underlying interface is up, otherwise the link is down. This, among other things, allows carp to work on a bridge. Prodded by: glebius Tested by: Alexander Lunev
2012-04-20 09:55:50 +00:00
Add the vlan tag to the bridge route table. This allows a vlan trunk to be bridged, previously legitimate traffic was not passed as the bridge could not tell that it was on a different Ethernet segment. All non-tagged traffic is treated as vlan1 as per IEEE 802.1Q-2003
2007-06-13 18:58:04 +00:00
/* The default bridge vlan is 1 (IEEE 802.1Q-2003 Table 9-2) */
#define VLANTAGOF(_m) \
(_m->m_flags & M_VLANTAG) ? EVL_VLANOFTAG(_m->m_pkthdr.ether_vtag) : 1
Change the passing of callbacks to a struct in case this needs to be extended in the future.
2007-03-09 19:34:55 +00:00
static struct bstp_cb_ops bridge_ops = {
.bcb_state = bridge_state_change,
.bcb_rtage = bridge_rtable_expire
};
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
SYSCTL_DECL(_net_link);
Mark more nodes as CTLFLAG_MPSAFE or CTLFLAG_NEEDGIANT (17 of many) r357614 added CTLFLAG_NEEDGIANT to make it easier to find nodes that are still not MPSAFE (or already are but aren’t properly marked). Use it in preparation for a general review of all nodes. This is non-functional change that adds annotations to SYSCTL_NODE and SYSCTL_PROC nodes using one of the soon-to-be-required flags. Mark all obvious cases as MPSAFE. All entries that haven't been marked as MPSAFE before are by default marked as NEEDGIANT Approved by: kib (mentor, blanket) Commented by: kib, gallatin, melifaro Differential Revision: https://reviews.freebsd.org/D23718
2020-02-26 14:26:36 +00:00
static SYSCTL_NODE(_net_link, IFT_BRIDGE, bridge, CTLFLAG_RW | CTLFLAG_MPSAFE, 0,
"Bridge");
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
Virtualize if_bridge(4) cloner.
2014-10-05 19:43:37 +00:00
/* only pass IP[46] packets when pfil is enabled */
Use the new VNET_DEFINE_STATIC macro when we are defining static VNET variables. Reviewed by: bz Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D16147
2018-07-24 16:35:52 +00:00
VNET_DEFINE_STATIC(int, pfil_onlyip) = 1;
Virtualize if_bridge(4) cloner.
2014-10-05 19:43:37 +00:00
#define V_pfil_onlyip VNET(pfil_onlyip)
SYSCTL_INT(_net_link_bridge, OID_AUTO, pfil_onlyip,
CTLFLAG_RWTUN | CTLFLAG_VNET, &VNET_NAME(pfil_onlyip), 0,
"Only pass IP packets when pfil is enabled");
/* run pfil hooks on the bridge interface */
Use the new VNET_DEFINE_STATIC macro when we are defining static VNET variables. Reviewed by: bz Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D16147
2018-07-24 16:35:52 +00:00
VNET_DEFINE_STATIC(int, pfil_bridge) = 1;
Virtualize if_bridge(4) cloner.
2014-10-05 19:43:37 +00:00
#define V_pfil_bridge VNET(pfil_bridge)
SYSCTL_INT(_net_link_bridge, OID_AUTO, pfil_bridge,
CTLFLAG_RWTUN | CTLFLAG_VNET, &VNET_NAME(pfil_bridge), 0,
"Packet filter on the bridge interface");
/* layer2 filter with ipfw */
Use the new VNET_DEFINE_STATIC macro when we are defining static VNET variables. Reviewed by: bz Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D16147
2018-07-24 16:35:52 +00:00
VNET_DEFINE_STATIC(int, pfil_ipfw);
Virtualize if_bridge(4) cloner.
2014-10-05 19:43:37 +00:00
#define V_pfil_ipfw VNET(pfil_ipfw)
/* layer2 ARP filter with ipfw */
Use the new VNET_DEFINE_STATIC macro when we are defining static VNET variables. Reviewed by: bz Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D16147
2018-07-24 16:35:52 +00:00
VNET_DEFINE_STATIC(int, pfil_ipfw_arp);
Virtualize if_bridge(4) cloner.
2014-10-05 19:43:37 +00:00
#define V_pfil_ipfw_arp VNET(pfil_ipfw_arp)
SYSCTL_INT(_net_link_bridge, OID_AUTO, ipfw_arp,
CTLFLAG_RWTUN | CTLFLAG_VNET, &VNET_NAME(pfil_ipfw_arp), 0,
"Filter ARP packets through IPFW layer2");
/* run pfil hooks on the member interface */
Use the new VNET_DEFINE_STATIC macro when we are defining static VNET variables. Reviewed by: bz Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D16147
2018-07-24 16:35:52 +00:00
VNET_DEFINE_STATIC(int, pfil_member) = 1;
Virtualize if_bridge(4) cloner.
2014-10-05 19:43:37 +00:00
#define V_pfil_member VNET(pfil_member)
SYSCTL_INT(_net_link_bridge, OID_AUTO, pfil_member,
CTLFLAG_RWTUN | CTLFLAG_VNET, &VNET_NAME(pfil_member), 0,
"Packet filter on the member interface");
/* run pfil hooks on the physical interface for locally destined packets */
Use the new VNET_DEFINE_STATIC macro when we are defining static VNET variables. Reviewed by: bz Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D16147
2018-07-24 16:35:52 +00:00
VNET_DEFINE_STATIC(int, pfil_local_phys);
Virtualize if_bridge(4) cloner.
2014-10-05 19:43:37 +00:00
#define V_pfil_local_phys VNET(pfil_local_phys)
SYSCTL_INT(_net_link_bridge, OID_AUTO, pfil_local_phys,
CTLFLAG_RWTUN | CTLFLAG_VNET, &VNET_NAME(pfil_local_phys), 0,
Allow additional packet filtering on the physical interface for locally destined packets, disabled by default. PR: kern/116051 Submitted by: Eygene Ryabinkin Approved by: re (bmah) MFC after: 2 weeks
2007-09-16 21:09:15 +00:00
"Packet filter on the physical interface for locally destined packets");
Virtualize if_bridge(4) cloner.
2014-10-05 19:43:37 +00:00
/* log STP state changes */
Use the new VNET_DEFINE_STATIC macro when we are defining static VNET variables. Reviewed by: bz Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D16147
2018-07-24 16:35:52 +00:00
VNET_DEFINE_STATIC(int, log_stp);
Virtualize if_bridge(4) cloner.
2014-10-05 19:43:37 +00:00
#define V_log_stp VNET(log_stp)
SYSCTL_INT(_net_link_bridge, OID_AUTO, log_stp,
CTLFLAG_RWTUN | CTLFLAG_VNET, &VNET_NAME(log_stp), 0,
"Log STP state changes");
/* share MAC with first bridge member */
Use the new VNET_DEFINE_STATIC macro when we are defining static VNET variables. Reviewed by: bz Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D16147
2018-07-24 16:35:52 +00:00
VNET_DEFINE_STATIC(int, bridge_inherit_mac);
Virtualize if_bridge(4) cloner.
2014-10-05 19:43:37 +00:00
#define V_bridge_inherit_mac VNET(bridge_inherit_mac)
SYSCTL_INT(_net_link_bridge, OID_AUTO, inherit_mac,
CTLFLAG_RWTUN | CTLFLAG_VNET, &VNET_NAME(bridge_inherit_mac), 0,
Put the bridge mac inheritance behind a sysctl with the default off as this still needs all the edge cases fixed. Submitted by: Eygene Ryabinkin
2008-09-08 03:28:26 +00:00
"Inherit MAC address from the first bridge member");
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
Use the new VNET_DEFINE_STATIC macro when we are defining static VNET variables. Reviewed by: bz Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D16147
2018-07-24 16:35:52 +00:00
VNET_DEFINE_STATIC(int, allow_llz_overlap) = 0;
- Relax the restriction on the member interfaces with LLAs. Two or more LLAs on the member interfaces are actually harmless when the parent interface does not have a LLA. - Add net.link.bridge.allow_llz_overlap. This is a knob to allow LLAs on a bridge and the member interfaces at the same time. The default is 0. Pointed out by: ume MFC after: 3 days
2013-07-28 19:49:39 +00:00
#define V_allow_llz_overlap VNET(allow_llz_overlap)
Remove SYSCTL_VNET_* macros, and simply put CTLFLAG_VNET where needed. Sponsored by: Nginx, Inc.
2014-11-07 09:39:05 +00:00
SYSCTL_INT(_net_link_bridge, OID_AUTO, allow_llz_overlap,
Remove excess CTLFLAG_VNET Sponsored by: Dell EMC Isilon
2016-10-26 23:40:07 +00:00
CTLFLAG_RW | CTLFLAG_VNET, &VNET_NAME(allow_llz_overlap), 0,
Virtualize if_bridge(4) cloner.
2014-10-05 19:43:37 +00:00
"Allow overlap of link-local scope "
- Relax the restriction on the member interfaces with LLAs. Two or more LLAs on the member interfaces are actually harmless when the parent interface does not have a LLA. - Add net.link.bridge.allow_llz_overlap. This is a knob to allow LLAs on a bridge and the member interfaces at the same time. The default is 0. Pointed out by: ume MFC after: 3 days
2013-07-28 19:49:39 +00:00
"zones of a bridge interface and the member interfaces");
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
struct bridge_control {
int (*bc_func)(struct bridge_softc *, void *);
int bc_argsize;
int bc_flags;
};
#define BC_F_COPYIN 0x01 /* copy arguments in */
#define BC_F_COPYOUT 0x02 /* copy arguments out */
#define BC_F_SUSER 0x04 /* do super-user check */
const struct bridge_control bridge_control_table[] = {
{ bridge_ioctl_add, sizeof(struct ifbreq),
BC_F_COPYIN|BC_F_SUSER },
{ bridge_ioctl_del, sizeof(struct ifbreq),
BC_F_COPYIN|BC_F_SUSER },
{ bridge_ioctl_gifflags, sizeof(struct ifbreq),
BC_F_COPYIN|BC_F_COPYOUT },
{ bridge_ioctl_sifflags, sizeof(struct ifbreq),
BC_F_COPYIN|BC_F_SUSER },
{ bridge_ioctl_scache, sizeof(struct ifbrparam),
BC_F_COPYIN|BC_F_SUSER },
{ bridge_ioctl_gcache, sizeof(struct ifbrparam),
BC_F_COPYOUT },
{ bridge_ioctl_gifs, sizeof(struct ifbifconf),
BC_F_COPYIN|BC_F_COPYOUT },
{ bridge_ioctl_rts, sizeof(struct ifbaconf),
BC_F_COPYIN|BC_F_COPYOUT },
{ bridge_ioctl_saddr, sizeof(struct ifbareq),
BC_F_COPYIN|BC_F_SUSER },
{ bridge_ioctl_sto, sizeof(struct ifbrparam),
BC_F_COPYIN|BC_F_SUSER },
{ bridge_ioctl_gto, sizeof(struct ifbrparam),
BC_F_COPYOUT },
{ bridge_ioctl_daddr, sizeof(struct ifbareq),
BC_F_COPYIN|BC_F_SUSER },
{ bridge_ioctl_flush, sizeof(struct ifbreq),
BC_F_COPYIN|BC_F_SUSER },
{ bridge_ioctl_gpri, sizeof(struct ifbrparam),
BC_F_COPYOUT },
{ bridge_ioctl_spri, sizeof(struct ifbrparam),
BC_F_COPYIN|BC_F_SUSER },
{ bridge_ioctl_ght, sizeof(struct ifbrparam),
BC_F_COPYOUT },
{ bridge_ioctl_sht, sizeof(struct ifbrparam),
BC_F_COPYIN|BC_F_SUSER },
{ bridge_ioctl_gfd, sizeof(struct ifbrparam),
BC_F_COPYOUT },
{ bridge_ioctl_sfd, sizeof(struct ifbrparam),
BC_F_COPYIN|BC_F_SUSER },
{ bridge_ioctl_gma, sizeof(struct ifbrparam),
BC_F_COPYOUT },
{ bridge_ioctl_sma, sizeof(struct ifbrparam),
BC_F_COPYIN|BC_F_SUSER },
{ bridge_ioctl_sifprio, sizeof(struct ifbreq),
BC_F_COPYIN|BC_F_SUSER },
{ bridge_ioctl_sifcost, sizeof(struct ifbreq),
BC_F_COPYIN|BC_F_SUSER },
Add support for creating span ports so that one can snoop bridged traffic from another interface/machine/network. Obtained from: OpenBSD MFC after: 2 weeks
2005-12-14 02:52:13 +00:00
{ bridge_ioctl_addspan, sizeof(struct ifbreq),
BC_F_COPYIN|BC_F_SUSER },
{ bridge_ioctl_delspan, sizeof(struct ifbreq),
BC_F_COPYIN|BC_F_SUSER },
Add some statistics that are needed to support RFC4188 as part of the SoC2006 work on a bridge monitoring module for BSNMP. Submitted by: shteryana (SoC 2006)
2006-07-31 20:24:46 +00:00
{ bridge_ioctl_gbparam, sizeof(struct ifbropreq),
BC_F_COPYOUT },
{ bridge_ioctl_grte, sizeof(struct ifbrparam),
BC_F_COPYOUT },
{ bridge_ioctl_gifsstp, sizeof(struct ifbpstpconf),
Fix SIOCGDRVSPEC/BRDGGIFSSTP ioctl: make it copyin() the user provided buffer length before trying to use it. Reviewed by: thompsa Approved by: bz (mentor) MFC after: 3 days
2006-12-03 21:50:57 +00:00
BC_F_COPYIN|BC_F_COPYOUT },
Bring in support for the Rapid Spanning Tree Protocol (802.1w). RSTP provides faster spanning tree convergence, the protocol will exchange information with neighboring switches to quickly transition to forwarding without creating loops. The code will default to RSTP mode but will downgrade any port connected to a legacy STP network so is fully backward compatible. Reviewed by: syrinx Tested by: syrinx
2006-11-01 09:07:47 +00:00
{ bridge_ioctl_sproto, sizeof(struct ifbrparam),
BC_F_COPYIN|BC_F_SUSER },
{ bridge_ioctl_stxhc, sizeof(struct ifbrparam),
BC_F_COPYIN|BC_F_SUSER },
Add an option to limit the number of source MACs that can be behind a bridge interface. Once the limit is reached packets with unknown source addresses are dropped until an existing host cache entry expires or is removed. Useful to use with the STICKY cache option. Sponsored by: miniSuperHappyDevHouse NZ
2007-11-04 08:32:27 +00:00
{ bridge_ioctl_sifmaxaddr, sizeof(struct ifbreq),
BC_F_COPYIN|BC_F_SUSER },
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
};
Virtualize if_bridge(4) cloner.
2014-10-05 19:43:37 +00:00
const int bridge_control_table_size = nitems(bridge_control_table);
Use the new VNET_DEFINE_STATIC macro when we are defining static VNET variables. Reviewed by: bz Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D16147
2018-07-24 16:35:52 +00:00
VNET_DEFINE_STATIC(LIST_HEAD(, bridge_softc), bridge_list);
Virtualize if_bridge(4) cloner.
2014-10-05 19:43:37 +00:00
#define V_bridge_list VNET(bridge_list)
bridge: call member interface ioctl() without NET_EPOCH We're not allowed to hold NET_EPOCH while sleeping, so when we call ioctl() handlers for member interfaces we cannot be in NET_EPOCH. We still need some protection of our CK_LISTs, so hold BRIDGE_LOCK instead. That requires changing BRIDGE_LOCK into a sleepable lock, and separating the BRIDGE_RT_LOCK, to protect bridge_rtnode lists. That lock is taken in the data path (while in NET_EPOCH), so it cannot be a sleepable lock. While here document the locking strategy. MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D26418
2020-10-06 19:19:56 +00:00
#define BRIDGE_LIST_LOCK_INIT(x) sx_init(&V_bridge_list_sx, \
"if_bridge list")
#define BRIDGE_LIST_LOCK_DESTROY(x) sx_destroy(&V_bridge_list_sx)
#define BRIDGE_LIST_LOCK(x) sx_xlock(&V_bridge_list_sx)
#define BRIDGE_LIST_UNLOCK(x) sx_xunlock(&V_bridge_list_sx)
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
Use the new VNET_DEFINE_STATIC macro when we are defining static VNET variables. Reviewed by: bz Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D16147
2018-07-24 16:35:52 +00:00
VNET_DEFINE_STATIC(struct if_clone *, bridge_cloner);
Virtualize if_bridge(4) cloner.
2014-10-05 19:43:37 +00:00
#define V_bridge_cloner VNET(bridge_cloner)
Change from a callback in if_ethersubr to using EVENTHANDLER in order to detach span ports when they disappear. The span port does not have a pointer to the softc so revert r1.31 and bring back the softc linked-list. MFC after: 2 weeks
2005-12-17 06:33:51 +00:00
Make the "struct if_clone" opaque to users of the cloning API. Users now use function calls: if_clone_simple() if_clone_advanced() to initialize a cloner, instead of macros that initialize if_clone structure. Discussed with: brooks, bz, 1 year ago
2012-10-16 13:37:54 +00:00
static const char bridge_name[] = "bridge";
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
Virtualize if_bridge(4) cloner.
2014-10-05 19:43:37 +00:00
static void
vnet_bridge_init(const void *unused __unused)
{
if_bridge(4): Fix module teardown bridge_rtnode_zone still has outstanding allocations at the time of destruction in the current model because all of the interface teardown happens in a VNET_SYSUNINIT, -after- the MOD_UNLOAD has already been processed. The SYSUNINIT triggers destruction of the interfaces, which then attempts to free the memory from the zone that's already been destroyed, and we hit a panic. Solve this by virtualizing the uma_zone we allocate the rtnodes from to fix the ordering. bridge_rtable_fini should also take care to flush any remaining routes that weren't taken care of when dynamic routes were flushed in bridge_stop. Reviewed by: kp Differential Revision: https://reviews.freebsd.org/D19578
2019-03-15 13:19:52 +00:00
V_bridge_rtnode_zone = uma_zcreate("bridge_rtnode",
sizeof(struct bridge_rtnode), NULL, NULL, NULL, NULL,
UMA_ALIGN_PTR, 0);
Virtualize if_bridge(4) cloner.
2014-10-05 19:43:37 +00:00
BRIDGE_LIST_LOCK_INIT();
LIST_INIT(&V_bridge_list);
V_bridge_cloner = if_clone_simple(bridge_name,
bridge_clone_create, bridge_clone_destroy, 0);
}
VNET_SYSINIT(vnet_bridge_init, SI_SUB_PROTO_IFATTACHDOMAIN, SI_ORDER_ANY,
vnet_bridge_init, NULL);
static void
vnet_bridge_uninit(const void *unused __unused)
{
if_clone_detach(V_bridge_cloner);
Fix a panic when tearing down a vnet on a VIMAGE-enabled kernel. There was a race that bridge_ifdetach() could be called via ifnet_departure event handler after vnet_bridge_uninit(). PR: 195859 Reported by: Danilo Egea Gondolfo
2015-02-14 18:15:14 +00:00
V_bridge_cloner = NULL;
Virtualize if_bridge(4) cloner.
2014-10-05 19:43:37 +00:00
BRIDGE_LIST_LOCK_DESTROY();
bridge: epoch-ification Run the bridge datapath under epoch, rather than under the BRIDGE_LOCK(). We still take the BRIDGE_LOCK() whenever we insert or delete items in the relevant lists, but we use epoch callbacks to free items so that it's safe to iterate the lists without the BRIDGE_LOCK. Tests on mercat5/6 shows this increases bridge throughput significantly, from 3.7Mpps to 18.6Mpps. Reviewed by: emaste, philip, melifaro MFC after: 2 months Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24250
2020-04-26 16:22:35 +00:00
/* Callbacks may use the UMA zone. */
epoch_drain_callbacks(net_epoch_preempt);
if_bridge(4): Fix module teardown bridge_rtnode_zone still has outstanding allocations at the time of destruction in the current model because all of the interface teardown happens in a VNET_SYSUNINIT, -after- the MOD_UNLOAD has already been processed. The SYSUNINIT triggers destruction of the interfaces, which then attempts to free the memory from the zone that's already been destroyed, and we hit a panic. Solve this by virtualizing the uma_zone we allocate the rtnodes from to fix the ordering. bridge_rtable_fini should also take care to flush any remaining routes that weren't taken care of when dynamic routes were flushed in bridge_stop. Reviewed by: kp Differential Revision: https://reviews.freebsd.org/D19578
2019-03-15 13:19:52 +00:00
uma_zdestroy(V_bridge_rtnode_zone);
Virtualize if_bridge(4) cloner.
2014-10-05 19:43:37 +00:00
}
Get closer to a VIMAGE network stack teardown from top to bottom rather than removing the network interfaces first. This change is rather larger and convoluted as the ordering requirements cannot be separated. Move the pfil(9) framework to SI_SUB_PROTO_PFIL, move Firewalls and related modules to their own SI_SUB_PROTO_FIREWALL. Move initialization of "physical" interfaces to SI_SUB_DRIVERS, move virtual (cloned) interfaces to SI_SUB_PSEUDO. Move Multicast to SI_SUB_PROTO_MC. Re-work parts of multicast initialisation and teardown, not taking the huge amount of memory into account if used as a module yet. For interface teardown we try to do as many of them as we can on SI_SUB_INIT_IF, but for some this makes no sense, e.g., when tunnelling over a higher layer protocol such as IP. In that case the interface has to go along (or before) the higher layer protocol is shutdown. Kernel hhooks need to go last on teardown as they may be used at various higher layers and we cannot remove them before we cleaned up the higher layers. For interface teardown there are multiple paths: (a) a cloned interface is destroyed (inside a VIMAGE or in the base system), (b) any interface is moved from a virtual network stack to a different network stack ("vmove"), or (c) a virtual network stack is being shut down. All code paths go through if_detach_internal() where we, depending on the vmove flag or the vnet state, make a decision on how much to shut down; in case we are destroying a VNET the individual protocol layers will cleanup their own parts thus we cannot do so again for each interface as we end up with, e.g., double-frees, destroying locks twice or acquiring already destroyed locks. When calling into protocol cleanups we equally have to tell them whether they need to detach upper layer protocols ("ulp") or not (e.g., in6_ifdetach()). Provide or enahnce helper functions to do proper cleanup at a protocol rather than at an interface level. Approved by: re (hrs) Obtained from: projects/vnet Reviewed by: gnn, jhb Sponsored by: The FreeBSD Foundation MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6747
2016-06-21 13:48:49 +00:00
VNET_SYSUNINIT(vnet_bridge_uninit, SI_SUB_PSEUDO, SI_ORDER_ANY,
Virtualize if_bridge(4) cloner.
2014-10-05 19:43:37 +00:00
vnet_bridge_uninit, NULL);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
static int
bridge_modevent(module_t mod, int type, void *data)
{
switch (type) {
case MOD_LOAD:
Add dummynet(4) support to if_bridge, this code is largely based on bridge.c. This is the final piece to match bridge.c in functionality, we can now be a drop-in replacement. Approved by: mlaier (mentor)
2005-06-10 01:25:22 +00:00
bridge_dn_p = bridge_dummynet;
Change from a callback in if_ethersubr to using EVENTHANDLER in order to detach span ports when they disappear. The span port does not have a pointer to the softc so revert r1.31 and bring back the softc linked-list. MFC after: 2 weeks
2005-12-17 06:33:51 +00:00
bridge_detach_cookie = EVENTHANDLER_REGISTER(
ifnet_departure_event, bridge_ifdetach, NULL,
EVENTHANDLER_PRI_ANY);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
break;
case MOD_UNLOAD:
Change from a callback in if_ethersubr to using EVENTHANDLER in order to detach span ports when they disappear. The span port does not have a pointer to the softc so revert r1.31 and bring back the softc linked-list. MFC after: 2 weeks
2005-12-17 06:33:51 +00:00
EVENTHANDLER_DEREGISTER(ifnet_departure_event,
bridge_detach_cookie);
Add dummynet(4) support to if_bridge, this code is largely based on bridge.c. This is the final piece to match bridge.c in functionality, we can now be a drop-in replacement. Approved by: mlaier (mentor)
2005-06-10 01:25:22 +00:00
bridge_dn_p = NULL;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
break;
default:
Fix style(9) nits, whitespace and parentheses.
2006-05-16 22:50:41 +00:00
return (EOPNOTSUPP);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
}
Fix style(9) nits, whitespace and parentheses.
2006-05-16 22:50:41 +00:00
return (0);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
}
static moduledata_t bridge_mod = {
Minor whitespace cleanup.
2005-12-17 10:03:48 +00:00
"if_bridge",
bridge_modevent,
Revert previous commit... Pointyhat to: kevlo (myself)
2012-10-10 08:36:38 +00:00
0
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
};
DECLARE_MODULE(if_bridge, bridge_mod, SI_SUB_PSEUDO, SI_ORDER_ANY);
bridge: Set module version This ensures that the loader will not load the module if it's also built in to the kernel. PR: 220860 Submitted by: Eugene Grosbein <eugen@freebsd.org> Reported by: Marie Helene Kvello-Aune <marieheleneka@gmail.com>
2017-09-21 14:14:01 +00:00
MODULE_VERSION(if_bridge, 1);
bridgestp is now a seperate module.
2006-07-26 22:15:15 +00:00
MODULE_DEPEND(if_bridge, bridgestp, 1, 1, 1);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
Bring in IPFW layer2 filtering from bridge.c, this allows Ethernet filtering using the layer2, mac and mac-type keywords. This is one of the last features that bridge.c has over if_bridge and gets us very close to a full functional replacement. Approved by: mlaier (mentor)
2005-06-07 21:20:18 +00:00
/*
Fix typo in comment. Reviewed by: thompsa
2012-12-18 06:37:23 +00:00
* handler for net.link.bridge.ipfw
Bring in IPFW layer2 filtering from bridge.c, this allows Ethernet filtering using the layer2, mac and mac-type keywords. This is one of the last features that bridge.c has over if_bridge and gets us very close to a full functional replacement. Approved by: mlaier (mentor)
2005-06-07 21:20:18 +00:00
*/
static int
sysctl_pfil_ipfw(SYSCTL_HANDLER_ARGS)
{
Virtualize if_bridge(4) cloner.
2014-10-05 19:43:37 +00:00
int enable = V_pfil_ipfw;
Minor whitespace cleanup.
2005-12-17 10:03:48 +00:00
int error;
Bring in IPFW layer2 filtering from bridge.c, this allows Ethernet filtering using the layer2, mac and mac-type keywords. This is one of the last features that bridge.c has over if_bridge and gets us very close to a full functional replacement. Approved by: mlaier (mentor)
2005-06-07 21:20:18 +00:00
Minor whitespace cleanup.
2005-12-17 10:03:48 +00:00
error = sysctl_handle_int(oidp, &enable, 0, req);
Virtualize if_bridge(4) cloner.
2014-10-05 19:43:37 +00:00
enable &= 1;
Bring in IPFW layer2 filtering from bridge.c, this allows Ethernet filtering using the layer2, mac and mac-type keywords. This is one of the last features that bridge.c has over if_bridge and gets us very close to a full functional replacement. Approved by: mlaier (mentor)
2005-06-07 21:20:18 +00:00
Virtualize if_bridge(4) cloner.
2014-10-05 19:43:37 +00:00
if (enable != V_pfil_ipfw) {
V_pfil_ipfw = enable;
Bring in IPFW layer2 filtering from bridge.c, this allows Ethernet filtering using the layer2, mac and mac-type keywords. This is one of the last features that bridge.c has over if_bridge and gets us very close to a full functional replacement. Approved by: mlaier (mentor)
2005-06-07 21:20:18 +00:00
Minor whitespace cleanup.
2005-12-17 10:03:48 +00:00
/*
* Disable pfil so that ipfw doesnt run twice, if the user
* really wants both then they can re-enable pfil_bridge and/or
When pfil(9) is enabled the bridge only considers ETHERTYPE_ARP, ETHERTYPE_IP and ETHERTYPE_IPV6 frames. Change this to be a sysctl knob so that is able to still bridge non-IP packets if desired. Also return early if all pfil_* sysctls are turned off, the user obviously does not want to filter on the bridge.
2005-12-29 09:39:15 +00:00
* pfil_member. Also allow non-ip packets as ipfw can filter by
* layer2 type.
Minor whitespace cleanup.
2005-12-17 10:03:48 +00:00
*/
Virtualize if_bridge(4) cloner.
2014-10-05 19:43:37 +00:00
if (V_pfil_ipfw) {
V_pfil_onlyip = 0;
V_pfil_bridge = 0;
V_pfil_member = 0;
Minor whitespace cleanup.
2005-12-17 10:03:48 +00:00
}
Bring in IPFW layer2 filtering from bridge.c, this allows Ethernet filtering using the layer2, mac and mac-type keywords. This is one of the last features that bridge.c has over if_bridge and gets us very close to a full functional replacement. Approved by: mlaier (mentor)
2005-06-07 21:20:18 +00:00
}
Fix style(9) nits, whitespace and parentheses.
2006-05-16 22:50:41 +00:00
return (error);
Bring in IPFW layer2 filtering from bridge.c, this allows Ethernet filtering using the layer2, mac and mac-type keywords. This is one of the last features that bridge.c has over if_bridge and gets us very close to a full functional replacement. Approved by: mlaier (mentor)
2005-06-07 21:20:18 +00:00
}
Virtualize if_bridge(4) cloner.
2014-10-05 19:43:37 +00:00
SYSCTL_PROC(_net_link_bridge, OID_AUTO, ipfw,
Mark more nodes as CTLFLAG_MPSAFE or CTLFLAG_NEEDGIANT (17 of many) r357614 added CTLFLAG_NEEDGIANT to make it easier to find nodes that are still not MPSAFE (or already are but aren’t properly marked). Use it in preparation for a general review of all nodes. This is non-functional change that adds annotations to SYSCTL_NODE and SYSCTL_PROC nodes using one of the soon-to-be-required flags. Mark all obvious cases as MPSAFE. All entries that haven't been marked as MPSAFE before are by default marked as NEEDGIANT Approved by: kib (mentor, blanket) Commented by: kib, gallatin, melifaro Differential Revision: https://reviews.freebsd.org/D23718
2020-02-26 14:26:36 +00:00
CTLTYPE_INT | CTLFLAG_RW | CTLFLAG_VNET | CTLFLAG_NEEDGIANT,
Virtualize if_bridge(4) cloner.
2014-10-05 19:43:37 +00:00
&VNET_NAME(pfil_ipfw), 0, &sysctl_pfil_ipfw, "I",
"Layer2 filter with IPFW");
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge: Remove members when assigned to a new vnet When the bridge is moved to a different vnet we must remove all of its member interfaces (and span interfaces), because we don't know if those will be moved along with it. We don't want to hold references to interfaces not in our vnet. Reviewed by: donner@ MFC after: 1 week Sponsored by: Orange Business Services Differential Revision: https://reviews.freebsd.org/D28859
2021-02-21 20:20:32 +00:00
#ifdef VIMAGE
static void
bridge_reassign(struct ifnet *ifp, struct vnet *newvnet, char *arg)
{
struct bridge_softc *sc = ifp->if_softc;
struct bridge_iflist *bif;
BRIDGE_LOCK(sc);
while ((bif = CK_LIST_FIRST(&sc->sc_iflist)) != NULL)
bridge_delete_member(sc, bif, 0);
while ((bif = CK_LIST_FIRST(&sc->sc_spanlist)) != NULL) {
bridge_delete_span(sc, bif);
}
BRIDGE_UNLOCK(sc);
ether_reassign(ifp, newvnet, arg);
}
#endif
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
/*
* bridge_clone_create:
*
* Create a new bridge instance.
*/
Change most of the bridge and stp funtions to static. This has highlighted that the following funtions are not used, wrap in '#ifdef noused' for the moment. bstp_enable_change_detection bstp_disable_change_detection bstp_set_bridge_priority bstp_set_port_priority bstp_set_path_cost
2005-10-14 10:38:12 +00:00
static int
Revise network interface cloning to take an optional opaque parameter that can specify configuration parameters: o rev cloner api's to add optional parameter block o add SIOCCREATE2 that accepts parameter data o rev vlan support to use new api (maintain old code) Reviewed by: arch@
2006-07-09 06:04:01 +00:00
bridge_clone_create(struct if_clone *ifc, int unit, caddr_t params)
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
{
bridge: Simplify mac address generation Unconditionally use ether_gen_addr() to generate bridge mac addresses. This function is now less likely to generate duplicate mac addresses across jails. The old hand rolled hostid based code adds no value. Reviewed by: bz Differential Revision: https://reviews.freebsd.org/D24432
2020-04-18 08:00:58 +00:00
struct bridge_softc *sc;
struct ifnet *ifp;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
sc = malloc(sizeof(*sc), M_DEVBUF, M_WAITOK|M_ZERO);
Catch up with the struct ifnet changes and use if_alloc(). Reviewed by: brooks Approved by: mlaier (mentor)
2005-06-10 23:52:01 +00:00
ifp = sc->sc_ifp = if_alloc(IFT_ETHER);
if (ifp == NULL) {
free(sc, M_DEVBUF);
return (ENOSPC);
}
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
Move the lock init until after if_alloc in case the allocation fails and we free the softc and return. MFC after: 3 days
2007-02-23 19:37:55 +00:00
BRIDGE_LOCK_INIT(sc);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
sc->sc_brtmax = BRIDGE_RTABLE_MAX;
sc->sc_brttimeout = BRIDGE_RTABLE_TIMEOUT;
/* Initialize our routing table. */
bridge_rtable_init(sc);
bridge: call member interface ioctl() without NET_EPOCH We're not allowed to hold NET_EPOCH while sleeping, so when we call ioctl() handlers for member interfaces we cannot be in NET_EPOCH. We still need some protection of our CK_LISTs, so hold BRIDGE_LOCK instead. That requires changing BRIDGE_LOCK into a sleepable lock, and separating the BRIDGE_RT_LOCK, to protect bridge_rtnode lists. That lock is taken in the data path (while in NET_EPOCH), so it cannot be a sleepable lock. While here document the locking strategy. MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D26418
2020-10-06 19:19:56 +00:00
callout_init_mtx(&sc->sc_brcallout, &sc->sc_rt_mtx, 0);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge: Change lists to CK_LIST as a peparation for epochification Prepare the ground for a rework of the bridge locking approach. We will use an epoch-based approach in the datapath and making it safe to iterate over the interface, span and rtnode lists without holding the BRIDGE_LOCK. Replace the relevant lists by their ConcurrencyKit equivalents. No functional change in this commit. Reviewed by: emaste, ae, philip (previous version) Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24249
2020-04-05 17:15:20 +00:00
CK_LIST_INIT(&sc->sc_iflist);
CK_LIST_INIT(&sc->sc_spanlist);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
ifp->if_softc = sc;
Make the "struct if_clone" opaque to users of the cloning API. Users now use function calls: if_clone_simple() if_clone_advanced() to initialize a cloner, instead of macros that initialize if_clone structure. Discussed with: brooks, bz, 1 year ago
2012-10-16 13:37:54 +00:00
if_initname(ifp, bridge_name, unit);
The bridge cant hear its own transmissions so set IFF_SIMPLEX. PR: kern/102361 Tested by: Radim Kolar <hsn@netmag.cz> MFC after: 3 days
2006-08-25 20:11:56 +00:00
ifp->if_flags = IFF_BROADCAST | IFF_SIMPLEX | IFF_MULTICAST;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
ifp->if_ioctl = bridge_ioctl;
if_bridge: add ALTQ support Similar to the recent addition of ALTQ support to if_vlan. Reviewed by: donner Obtained from: pfsense MFC after: 1 week Sponsored by: Rubicon Communications, LLC ("Netgate") Differential Revision: https://reviews.freebsd.org/D31675
2021-08-25 09:07:44 +00:00
#ifdef ALTQ
ifp->if_start = bridge_altq_start;
ifp->if_transmit = bridge_altq_transmit;
IFQ_SET_MAXLEN(&ifp->if_snd, ifqmaxlen);
ifp->if_snd.ifq_drv_maxlen = 0;
IFQ_SET_READY(&ifp->if_snd);
#else
Change bridge(4) to use if_transmit for forwarding packets to underlying interfaces instead of queueing. Tested by: ray
2012-09-03 10:08:20 +00:00
ifp->if_transmit = bridge_transmit;
if_bridge: add ALTQ support Similar to the recent addition of ALTQ support to if_vlan. Reviewed by: donner Obtained from: pfsense MFC after: 1 week Sponsored by: Rubicon Communications, LLC ("Netgate") Differential Revision: https://reviews.freebsd.org/D31675
2021-08-25 09:07:44 +00:00
#endif
Change bridge(4) to use if_transmit for forwarding packets to underlying interfaces instead of queueing. Tested by: ray
2012-09-03 10:08:20 +00:00
ifp->if_qflush = bridge_qflush;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
ifp->if_init = bridge_init;
ifp->if_type = IFT_BRIDGE;
bridge: Simplify mac address generation Unconditionally use ether_gen_addr() to generate bridge mac addresses. This function is now less likely to generate duplicate mac addresses across jails. The old hand rolled hostid based code adds no value. Reviewed by: bz Differential Revision: https://reviews.freebsd.org/D24432
2020-04-18 08:00:58 +00:00
ether_gen_addr(ifp, &sc->sc_defaddr);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
Change the passing of callbacks to a struct in case this needs to be extended in the future.
2007-03-09 19:34:55 +00:00
bstp_attach(&sc->sc_stp, &bridge_ops);
ether: centralize fake hwaddr generation We currently have two places with identical fake hwaddr generation -- if_vxlan and if_bridge. Lift it into if_ethersubr for reuse in other interfaces that may also need a fake addr. Reviewed by: bryanv, kp, philip Differential Revision: https://reviews.freebsd.org/D19573
2019-03-14 17:18:00 +00:00
ether_ifattach(ifp, sc->sc_defaddr.octet);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
/* Now undo some of the damage... */
ifp->if_baudrate = 0;
ifp->if_type = IFT_BRIDGE;
bridge: Remove members when assigned to a new vnet When the bridge is moved to a different vnet we must remove all of its member interfaces (and span interfaces), because we don't know if those will be moved along with it. We don't want to hold references to interfaces not in our vnet. Reviewed by: donner@ MFC after: 1 week Sponsored by: Orange Business Services Differential Revision: https://reviews.freebsd.org/D28859
2021-02-21 20:20:32 +00:00
#ifdef VIMAGE
ifp->if_reassign = bridge_reassign;
#endif
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
Virtualize if_bridge(4) cloner.
2014-10-05 19:43:37 +00:00
BRIDGE_LIST_LOCK();
LIST_INSERT_HEAD(&V_bridge_list, sc, sc_list);
BRIDGE_LIST_UNLOCK();
Change from a callback in if_ethersubr to using EVENTHANDLER in order to detach span ports when they disappear. The span port does not have a pointer to the softc so revert r1.31 and bring back the softc linked-list. MFC after: 2 weeks
2005-12-17 06:33:51 +00:00
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
return (0);
}
bridge: epoch-ification Run the bridge datapath under epoch, rather than under the BRIDGE_LOCK(). We still take the BRIDGE_LOCK() whenever we insert or delete items in the relevant lists, but we use epoch callbacks to free items so that it's safe to iterate the lists without the BRIDGE_LOCK. Tests on mercat5/6 shows this increases bridge throughput significantly, from 3.7Mpps to 18.6Mpps. Reviewed by: emaste, philip, melifaro MFC after: 2 months Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24250
2020-04-26 16:22:35 +00:00
static void
bridge_clone_destroy_cb(struct epoch_context *ctx)
{
struct bridge_softc *sc;
sc = __containerof(ctx, struct bridge_softc, sc_epoch_ctx);
BRIDGE_LOCK_DESTROY(sc);
free(sc, M_DEVBUF);
}
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
/*
* bridge_clone_destroy:
*
* Destroy a bridge instance.
*/
Change most of the bridge and stp funtions to static. This has highlighted that the following funtions are not used, wrap in '#ifdef noused' for the moment. bstp_enable_change_detection bstp_disable_change_detection bstp_set_bridge_priority bstp_set_port_priority bstp_set_path_cost
2005-10-14 10:38:12 +00:00
static void
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge_clone_destroy(struct ifnet *ifp)
{
struct bridge_softc *sc = ifp->if_softc;
struct bridge_iflist *bif;
bridge: epoch-ification Run the bridge datapath under epoch, rather than under the BRIDGE_LOCK(). We still take the BRIDGE_LOCK() whenever we insert or delete items in the relevant lists, but we use epoch callbacks to free items so that it's safe to iterate the lists without the BRIDGE_LOCK. Tests on mercat5/6 shows this increases bridge throughput significantly, from 3.7Mpps to 18.6Mpps. Reviewed by: emaste, philip, melifaro MFC after: 2 months Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24250
2020-04-26 16:22:35 +00:00
struct epoch_tracker et;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
BRIDGE_LOCK(sc);
bridge_stop(ifp, 1);
Fix a panic in softclock() if the interface is destroyed with a bpf consumer attached. This is caused by bpf_detachd clearing IFF_PROMISC on the interface which does a SIOCSIFFLAGS ioctl. The problem here is that while the interface has been stopped, IFF_UP has not been cleared so IFF_UP != IFF_DRV_RUNNING, this causes the ioctl function to init() the interface which resets the callouts. The destroy then completes and frees the softc but softclock will panic on a dead callout pointer. Ensure ifp->if_flags matches reality by clearing IFF_UP when we destroy. Silence from: rwatson Approved by: mlaier (mentor) MFC after: 3 days
2005-08-27 01:17:42 +00:00
ifp->if_flags &= ~IFF_UP;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge: Change lists to CK_LIST as a peparation for epochification Prepare the ground for a rework of the bridge locking approach. We will use an epoch-based approach in the datapath and making it safe to iterate over the interface, span and rtnode lists without holding the BRIDGE_LOCK. Replace the relevant lists by their ConcurrencyKit equivalents. No functional change in this commit. Reviewed by: emaste, ae, philip (previous version) Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24249
2020-04-05 17:15:20 +00:00
while ((bif = CK_LIST_FIRST(&sc->sc_iflist)) != NULL)
If we have been called from ether_ifdetach() then do not try and clear the promisc flag from the member interface, this is a no-op anyway since the interface is disappearing. The driver may have already released its resources such as miibus and this is likely to panic the kernel. Submitted and tested by: Wojciech A. Koszek MFC after: 2 weeks
2005-10-23 22:30:07 +00:00
bridge_delete_member(sc, bif, 0);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge: Change lists to CK_LIST as a peparation for epochification Prepare the ground for a rework of the bridge locking approach. We will use an epoch-based approach in the datapath and making it safe to iterate over the interface, span and rtnode lists without holding the BRIDGE_LOCK. Replace the relevant lists by their ConcurrencyKit equivalents. No functional change in this commit. Reviewed by: emaste, ae, philip (previous version) Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24249
2020-04-05 17:15:20 +00:00
while ((bif = CK_LIST_FIRST(&sc->sc_spanlist)) != NULL) {
Change from a callback in if_ethersubr to using EVENTHANDLER in order to detach span ports when they disappear. The span port does not have a pointer to the softc so revert r1.31 and bring back the softc linked-list. MFC after: 2 weeks
2005-12-17 06:33:51 +00:00
bridge_delete_span(sc, bif);
Add support for creating span ports so that one can snoop bridged traffic from another interface/machine/network. Obtained from: OpenBSD MFC after: 2 weeks
2005-12-14 02:52:13 +00:00
}
if_bridge(4): Fix module teardown bridge_rtnode_zone still has outstanding allocations at the time of destruction in the current model because all of the interface teardown happens in a VNET_SYSUNINIT, -after- the MOD_UNLOAD has already been processed. The SYSUNINIT triggers destruction of the interfaces, which then attempts to free the memory from the zone that's already been destroyed, and we hit a panic. Solve this by virtualizing the uma_zone we allocate the rtnodes from to fix the ordering. bridge_rtable_fini should also take care to flush any remaining routes that weren't taken care of when dynamic routes were flushed in bridge_stop. Reviewed by: kp Differential Revision: https://reviews.freebsd.org/D19578
2019-03-15 13:19:52 +00:00
/* Tear down the routing table. */
bridge_rtable_fini(sc);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
BRIDGE_UNLOCK(sc);
bridge: call member interface ioctl() without NET_EPOCH We're not allowed to hold NET_EPOCH while sleeping, so when we call ioctl() handlers for member interfaces we cannot be in NET_EPOCH. We still need some protection of our CK_LISTs, so hold BRIDGE_LOCK instead. That requires changing BRIDGE_LOCK into a sleepable lock, and separating the BRIDGE_RT_LOCK, to protect bridge_rtnode lists. That lock is taken in the data path (while in NET_EPOCH), so it cannot be a sleepable lock. While here document the locking strategy. MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D26418
2020-10-06 19:19:56 +00:00
NET_EPOCH_ENTER(et);
Mark the callouts as MPSAFE as if_bridge has been giant-free since day 1. Use the SMP friendly callout_init_mtx() while we are here. Approved by: mlaier (mentor) MFC after: 3 days
2005-08-18 20:17:00 +00:00
callout_drain(&sc->sc_brcallout);
Virtualize if_bridge(4) cloner.
2014-10-05 19:43:37 +00:00
BRIDGE_LIST_LOCK();
Change from a callback in if_ethersubr to using EVENTHANDLER in order to detach span ports when they disappear. The span port does not have a pointer to the softc so revert r1.31 and bring back the softc linked-list. MFC after: 2 weeks
2005-12-17 06:33:51 +00:00
LIST_REMOVE(sc, sc_list);
Virtualize if_bridge(4) cloner.
2014-10-05 19:43:37 +00:00
BRIDGE_LIST_UNLOCK();
Change from a callback in if_ethersubr to using EVENTHANDLER in order to detach span ports when they disappear. The span port does not have a pointer to the softc so revert r1.31 and bring back the softc linked-list. MFC after: 2 weeks
2005-12-17 06:33:51 +00:00
/tmp/cvsuusTrc
2006-07-26 10:43:02 +00:00
bstp_detach(&sc->sc_stp);
if_bridge: add ALTQ support Similar to the recent addition of ALTQ support to if_vlan. Reviewed by: donner Obtained from: pfsense MFC after: 1 week Sponsored by: Rubicon Communications, LLC ("Netgate") Differential Revision: https://reviews.freebsd.org/D31675
2021-08-25 09:07:44 +00:00
#ifdef ALTQ
IFQ_PURGE(&ifp->if_snd);
#endif
bridge: epoch-ification Run the bridge datapath under epoch, rather than under the BRIDGE_LOCK(). We still take the BRIDGE_LOCK() whenever we insert or delete items in the relevant lists, but we use epoch callbacks to free items so that it's safe to iterate the lists without the BRIDGE_LOCK. Tests on mercat5/6 shows this increases bridge throughput significantly, from 3.7Mpps to 18.6Mpps. Reviewed by: emaste, philip, melifaro MFC after: 2 months Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24250
2020-04-26 16:22:35 +00:00
NET_EPOCH_EXIT(et);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
ether_ifdetach(ifp);
In r191367 the need for if_free_type() was removed and a new member if_alloctype was used to store the origional interface type. Take advantage of this change by removing all existing uses of if_free_type() in favor of if_free(). MFC after: 1 Month
2011-11-11 22:57:52 +00:00
if_free(ifp);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge: epoch-ification Run the bridge datapath under epoch, rather than under the BRIDGE_LOCK(). We still take the BRIDGE_LOCK() whenever we insert or delete items in the relevant lists, but we use epoch callbacks to free items so that it's safe to iterate the lists without the BRIDGE_LOCK. Tests on mercat5/6 shows this increases bridge throughput significantly, from 3.7Mpps to 18.6Mpps. Reviewed by: emaste, philip, melifaro MFC after: 2 months Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24250
2020-04-26 16:22:35 +00:00
NET_EPOCH_CALL(bridge_clone_destroy_cb, &sc->sc_epoch_ctx);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
}
/*
* bridge_ioctl:
*
* Handle a control request from the operator.
*/
Change most of the bridge and stp funtions to static. This has highlighted that the following funtions are not used, wrap in '#ifdef noused' for the moment. bstp_enable_change_detection bstp_disable_change_detection bstp_set_bridge_priority bstp_set_port_priority bstp_set_path_cost
2005-10-14 10:38:12 +00:00
static int
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge_ioctl(struct ifnet *ifp, u_long cmd, caddr_t data)
{
struct bridge_softc *sc = ifp->if_softc;
- Check if_type of "addm <interface>" before setting the interface's MTU to the if_bridge(4) interface. This fixes a bug that MTU value of "addm <interface>" is used even when it is invalid for the if_bridge(4) member: # ifconfig bridge0 create # ifconfig bridge0 bridge0: flags=8802<BROADCAST,SIMPLEX,MULTICAST> metric 0 mtu 1500 ... # ifconfig bridge0 addm lo0 ifconfig: BRDGADD lo0: Invalid argument # ifconfig bridge0 bridge0: flags=8802<BROADCAST,SIMPLEX,MULTICAST> metric 0 mtu 16384 ... - Do not ignore MTU value of an interface even when if_type == IFT_GIF. This fixes MTU mismatch when an if_bridge(4) interface has a gif(4) interface and no other interface as the member, and it is directly used for L2 communication with EtherIP tunneling enabled. - Implement SIOCSIFMTU ioctl. Changing the MTU is allowed only when all members have the same MTU value.
2010-01-31 08:16:37 +00:00
struct ifreq *ifr = (struct ifreq *)data;
struct bridge_iflist *bif;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
struct thread *td = curthread;
union {
struct ifbreq ifbreq;
struct ifbifconf ifbifconf;
struct ifbareq ifbareq;
struct ifbaconf ifbaconf;
struct ifbrparam ifbrparam;
Sync with the OpenBSD port of RSTP - use flags rather than sperate ioctls for edge, p2p - implement p2p and autop2p flags - define large pathcost constant as ULL - show bridgeid and rootid in ifconfig Obtained from: Reyk Floeter <reyk@openbsd.org>
2006-11-27 00:35:40 +00:00
struct ifbropreq ifbropreq;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
} args;
struct ifdrv *ifd = (struct ifdrv *) data;
const struct bridge_control *bc;
Update route MTUs for bridge, lagg and vlan interfaces. Those interfaces may implicitly change their MTU on addition of parent interface in addition to normal SIOCSIFMTU ioctl path, where the route MTUs are updated normally. MFC after: 2 weeks Sponsored by: iXsystems, Inc.
2020-01-22 20:36:45 +00:00
int error = 0, oldmtu;
bridge: epoch-ification Run the bridge datapath under epoch, rather than under the BRIDGE_LOCK(). We still take the BRIDGE_LOCK() whenever we insert or delete items in the relevant lists, but we use epoch callbacks to free items so that it's safe to iterate the lists without the BRIDGE_LOCK. Tests on mercat5/6 shows this increases bridge throughput significantly, from 3.7Mpps to 18.6Mpps. Reviewed by: emaste, philip, melifaro MFC after: 2 months Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24250
2020-04-26 16:22:35 +00:00
bridge: call member interface ioctl() without NET_EPOCH We're not allowed to hold NET_EPOCH while sleeping, so when we call ioctl() handlers for member interfaces we cannot be in NET_EPOCH. We still need some protection of our CK_LISTs, so hold BRIDGE_LOCK instead. That requires changing BRIDGE_LOCK into a sleepable lock, and separating the BRIDGE_RT_LOCK, to protect bridge_rtnode lists. That lock is taken in the data path (while in NET_EPOCH), so it cannot be a sleepable lock. While here document the locking strategy. MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D26418
2020-10-06 19:19:56 +00:00
BRIDGE_LOCK(sc);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
switch (cmd) {
Add support for multicast to the bridge and allow inet6 addresses to be assigned to the interface. IPv6 auto-configuration is disabled. An IPv6 link-local address has a link-local scope within one link, the spec is unclear for the bridge case and it may cause scope violation. An address can be assigned in the usual way; ifconfig bridge0 inet6 xxxx:... Tested by: bmah Reviewed by: ume (netinet6) Approved by: mlaier (mentor) MFC after: 1 week
2005-09-06 21:11:59 +00:00
case SIOCADDMULTI:
case SIOCDELMULTI:
break;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
case SIOCGDRVSPEC:
case SIOCSDRVSPEC:
if (ifd->ifd_cmd >= bridge_control_table_size) {
error = EINVAL;
break;
}
bc = &bridge_control_table[ifd->ifd_cmd];
if (cmd == SIOCGDRVSPEC &&
(bc->bc_flags & BC_F_COPYOUT) == 0) {
error = EINVAL;
break;
}
else if (cmd == SIOCSDRVSPEC &&
(bc->bc_flags & BC_F_COPYOUT) != 0) {
error = EINVAL;
break;
}
if (bc->bc_flags & BC_F_SUSER) {
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
error = priv_check(td, PRIV_NET_BRIDGE);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
if (error)
break;
}
if (ifd->ifd_len != bc->bc_argsize ||
ifd->ifd_len > sizeof(args)) {
error = EINVAL;
break;
}
Fix style(9) nits, whitespace and parentheses.
2006-05-16 22:50:41 +00:00
bzero(&args, sizeof(args));
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
if (bc->bc_flags & BC_F_COPYIN) {
error = copyin(ifd->ifd_data, &args, ifd->ifd_len);
if (error)
break;
}
Update route MTUs for bridge, lagg and vlan interfaces. Those interfaces may implicitly change their MTU on addition of parent interface in addition to normal SIOCSIFMTU ioctl path, where the route MTUs are updated normally. MFC after: 2 weeks Sponsored by: iXsystems, Inc.
2020-01-22 20:36:45 +00:00
oldmtu = ifp->if_mtu;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
error = (*bc->bc_func)(sc, &args);
if (error)
break;
Update route MTUs for bridge, lagg and vlan interfaces. Those interfaces may implicitly change their MTU on addition of parent interface in addition to normal SIOCSIFMTU ioctl path, where the route MTUs are updated normally. MFC after: 2 weeks Sponsored by: iXsystems, Inc.
2020-01-22 20:36:45 +00:00
/*
* Bridge MTU may change during addition of the first port.
* If it did, do network layer specific procedure.
*/
if (ifp->if_mtu != oldmtu) {
#ifdef INET6
nd6_setmtu(ifp);
#endif
rt_updatemtu(ifp);
}
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
if (bc->bc_flags & BC_F_COPYOUT)
error = copyout(&args, ifd->ifd_data, ifd->ifd_len);
break;
case SIOCSIFFLAGS:
Propagate rename of IFF_OACTIVE and IFF_RUNNING to IFF_DRV_OACTIVE and IFF_DRV_RUNNING, as well as the move from ifnet.if_flags to ifnet.if_drv_flags. Device drivers are now responsible for synchronizing access to these flags, as they are in if_drv_flags. This helps prevent races between the network stack and device driver in maintaining the interface flags field. Many __FreeBSD__ and __FreeBSD_version checks maintained and continued; some less so. Reviewed by: pjd, bz MFC after: 7 days
2005-08-09 10:20:02 +00:00
if (!(ifp->if_flags & IFF_UP) &&
(ifp->if_drv_flags & IFF_DRV_RUNNING)) {
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
/*
* If interface is marked down and it is running,
* then stop and disable it.
*/
bridge_stop(ifp, 1);
Propagate rename of IFF_OACTIVE and IFF_RUNNING to IFF_DRV_OACTIVE and IFF_DRV_RUNNING, as well as the move from ifnet.if_flags to ifnet.if_drv_flags. Device drivers are now responsible for synchronizing access to these flags, as they are in if_drv_flags. This helps prevent races between the network stack and device driver in maintaining the interface flags field. Many __FreeBSD__ and __FreeBSD_version checks maintained and continued; some less so. Reviewed by: pjd, bz MFC after: 7 days
2005-08-09 10:20:02 +00:00
} else if ((ifp->if_flags & IFF_UP) &&
!(ifp->if_drv_flags & IFF_DRV_RUNNING)) {
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
/*
* If interface is marked up and it is stopped, then
* start it.
*/
bridge: call member interface ioctl() without NET_EPOCH We're not allowed to hold NET_EPOCH while sleeping, so when we call ioctl() handlers for member interfaces we cannot be in NET_EPOCH. We still need some protection of our CK_LISTs, so hold BRIDGE_LOCK instead. That requires changing BRIDGE_LOCK into a sleepable lock, and separating the BRIDGE_RT_LOCK, to protect bridge_rtnode lists. That lock is taken in the data path (while in NET_EPOCH), so it cannot be a sleepable lock. While here document the locking strategy. MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D26418
2020-10-06 19:19:56 +00:00
BRIDGE_UNLOCK(sc);
Fix a panic when bringing up the bridge interface. We were casting a ifnet pointer to a softc which is no longer valid since the ifnet struct was split out from the softc. Approved by: mlaier (mentor) Approved by: re (blanket)
2005-06-27 21:58:12 +00:00
(*ifp->if_init)(sc);
bridge: call member interface ioctl() without NET_EPOCH We're not allowed to hold NET_EPOCH while sleeping, so when we call ioctl() handlers for member interfaces we cannot be in NET_EPOCH. We still need some protection of our CK_LISTs, so hold BRIDGE_LOCK instead. That requires changing BRIDGE_LOCK into a sleepable lock, and separating the BRIDGE_RT_LOCK, to protect bridge_rtnode lists. That lock is taken in the data path (while in NET_EPOCH), so it cannot be a sleepable lock. While here document the locking strategy. MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D26418
2020-10-06 19:19:56 +00:00
BRIDGE_LOCK(sc);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
}
break;
case SIOCSIFMTU:
if_bridge: allow MTU changes if_bridge used to only allow MTU changes if the new MTU matched that of all member interfaces. This doesn't really make much sense, in that we really shouldn't be allowed to change the MTU of bridge member in the first place. Instead we now change the MTU of all member interfaces. If one fails we revert all interfaces back to the original MTU. We do not address the issue where bridge member interface MTUs can be changed here. Reviewed by: donner Sponsored by: Rubicon Communications, LLC ("Netgate") Differential Revision: https://reviews.freebsd.org/D31288
2021-07-23 15:22:18 +00:00
oldmtu = sc->sc_ifp->if_mtu;
- Check if_type of "addm <interface>" before setting the interface's MTU to the if_bridge(4) interface. This fixes a bug that MTU value of "addm <interface>" is used even when it is invalid for the if_bridge(4) member: # ifconfig bridge0 create # ifconfig bridge0 bridge0: flags=8802<BROADCAST,SIMPLEX,MULTICAST> metric 0 mtu 1500 ... # ifconfig bridge0 addm lo0 ifconfig: BRDGADD lo0: Invalid argument # ifconfig bridge0 bridge0: flags=8802<BROADCAST,SIMPLEX,MULTICAST> metric 0 mtu 16384 ... - Do not ignore MTU value of an interface even when if_type == IFT_GIF. This fixes MTU mismatch when an if_bridge(4) interface has a gif(4) interface and no other interface as the member, and it is directly used for L2 communication with EtherIP tunneling enabled. - Implement SIOCSIFMTU ioctl. Changing the MTU is allowed only when all members have the same MTU value.
2010-01-31 08:16:37 +00:00
if (ifr->ifr_mtu < 576) {
error = EINVAL;
break;
}
bridge: Change lists to CK_LIST as a peparation for epochification Prepare the ground for a rework of the bridge locking approach. We will use an epoch-based approach in the datapath and making it safe to iterate over the interface, span and rtnode lists without holding the BRIDGE_LOCK. Replace the relevant lists by their ConcurrencyKit equivalents. No functional change in this commit. Reviewed by: emaste, ae, philip (previous version) Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24249
2020-04-05 17:15:20 +00:00
if (CK_LIST_EMPTY(&sc->sc_iflist)) {
- Check if_type of "addm <interface>" before setting the interface's MTU to the if_bridge(4) interface. This fixes a bug that MTU value of "addm <interface>" is used even when it is invalid for the if_bridge(4) member: # ifconfig bridge0 create # ifconfig bridge0 bridge0: flags=8802<BROADCAST,SIMPLEX,MULTICAST> metric 0 mtu 1500 ... # ifconfig bridge0 addm lo0 ifconfig: BRDGADD lo0: Invalid argument # ifconfig bridge0 bridge0: flags=8802<BROADCAST,SIMPLEX,MULTICAST> metric 0 mtu 16384 ... - Do not ignore MTU value of an interface even when if_type == IFT_GIF. This fixes MTU mismatch when an if_bridge(4) interface has a gif(4) interface and no other interface as the member, and it is directly used for L2 communication with EtherIP tunneling enabled. - Implement SIOCSIFMTU ioctl. Changing the MTU is allowed only when all members have the same MTU value.
2010-01-31 08:16:37 +00:00
sc->sc_ifp->if_mtu = ifr->ifr_mtu;
break;
}
bridge: Change lists to CK_LIST as a peparation for epochification Prepare the ground for a rework of the bridge locking approach. We will use an epoch-based approach in the datapath and making it safe to iterate over the interface, span and rtnode lists without holding the BRIDGE_LOCK. Replace the relevant lists by their ConcurrencyKit equivalents. No functional change in this commit. Reviewed by: emaste, ae, philip (previous version) Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24249
2020-04-05 17:15:20 +00:00
CK_LIST_FOREACH(bif, &sc->sc_iflist, bif_next) {
if_bridge: allow MTU changes if_bridge used to only allow MTU changes if the new MTU matched that of all member interfaces. This doesn't really make much sense, in that we really shouldn't be allowed to change the MTU of bridge member in the first place. Instead we now change the MTU of all member interfaces. If one fails we revert all interfaces back to the original MTU. We do not address the issue where bridge member interface MTUs can be changed here. Reviewed by: donner Sponsored by: Rubicon Communications, LLC ("Netgate") Differential Revision: https://reviews.freebsd.org/D31288
2021-07-23 15:22:18 +00:00
error = (*bif->bif_ifp->if_ioctl)(bif->bif_ifp,
SIOCSIFMTU, (caddr_t)ifr);
if (error != 0) {
log(LOG_NOTICE, "%s: invalid MTU: %u for"
" member %s\n", sc->sc_ifp->if_xname,
ifr->ifr_mtu,
bif->bif_ifp->if_xname);
- Check if_type of "addm <interface>" before setting the interface's MTU to the if_bridge(4) interface. This fixes a bug that MTU value of "addm <interface>" is used even when it is invalid for the if_bridge(4) member: # ifconfig bridge0 create # ifconfig bridge0 bridge0: flags=8802<BROADCAST,SIMPLEX,MULTICAST> metric 0 mtu 1500 ... # ifconfig bridge0 addm lo0 ifconfig: BRDGADD lo0: Invalid argument # ifconfig bridge0 bridge0: flags=8802<BROADCAST,SIMPLEX,MULTICAST> metric 0 mtu 16384 ... - Do not ignore MTU value of an interface even when if_type == IFT_GIF. This fixes MTU mismatch when an if_bridge(4) interface has a gif(4) interface and no other interface as the member, and it is directly used for L2 communication with EtherIP tunneling enabled. - Implement SIOCSIFMTU ioctl. Changing the MTU is allowed only when all members have the same MTU value.
2010-01-31 08:16:37 +00:00
error = EINVAL;
break;
}
}
if_bridge: allow MTU changes if_bridge used to only allow MTU changes if the new MTU matched that of all member interfaces. This doesn't really make much sense, in that we really shouldn't be allowed to change the MTU of bridge member in the first place. Instead we now change the MTU of all member interfaces. If one fails we revert all interfaces back to the original MTU. We do not address the issue where bridge member interface MTUs can be changed here. Reviewed by: donner Sponsored by: Rubicon Communications, LLC ("Netgate") Differential Revision: https://reviews.freebsd.org/D31288
2021-07-23 15:22:18 +00:00
if (error) {
/* Restore the previous MTU on all member interfaces. */
ifr->ifr_mtu = oldmtu;
CK_LIST_FOREACH(bif, &sc->sc_iflist, bif_next) {
(*bif->bif_ifp->if_ioctl)(bif->bif_ifp,
SIOCSIFMTU, (caddr_t)ifr);
}
} else {
- Check if_type of "addm <interface>" before setting the interface's MTU to the if_bridge(4) interface. This fixes a bug that MTU value of "addm <interface>" is used even when it is invalid for the if_bridge(4) member: # ifconfig bridge0 create # ifconfig bridge0 bridge0: flags=8802<BROADCAST,SIMPLEX,MULTICAST> metric 0 mtu 1500 ... # ifconfig bridge0 addm lo0 ifconfig: BRDGADD lo0: Invalid argument # ifconfig bridge0 bridge0: flags=8802<BROADCAST,SIMPLEX,MULTICAST> metric 0 mtu 16384 ... - Do not ignore MTU value of an interface even when if_type == IFT_GIF. This fixes MTU mismatch when an if_bridge(4) interface has a gif(4) interface and no other interface as the member, and it is directly used for L2 communication with EtherIP tunneling enabled. - Implement SIOCSIFMTU ioctl. Changing the MTU is allowed only when all members have the same MTU value.
2010-01-31 08:16:37 +00:00
sc->sc_ifp->if_mtu = ifr->ifr_mtu;
if_bridge: allow MTU changes if_bridge used to only allow MTU changes if the new MTU matched that of all member interfaces. This doesn't really make much sense, in that we really shouldn't be allowed to change the MTU of bridge member in the first place. Instead we now change the MTU of all member interfaces. If one fails we revert all interfaces back to the original MTU. We do not address the issue where bridge member interface MTUs can be changed here. Reviewed by: donner Sponsored by: Rubicon Communications, LLC ("Netgate") Differential Revision: https://reviews.freebsd.org/D31288
2021-07-23 15:22:18 +00:00
}
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
break;
default:
Minor whitespace cleanup.
2005-12-17 10:03:48 +00:00
/*
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
* drop the lock as ether_ioctl() will call bridge_start() and
* cause the lock to be recursed.
*/
bridge: call member interface ioctl() without NET_EPOCH We're not allowed to hold NET_EPOCH while sleeping, so when we call ioctl() handlers for member interfaces we cannot be in NET_EPOCH. We still need some protection of our CK_LISTs, so hold BRIDGE_LOCK instead. That requires changing BRIDGE_LOCK into a sleepable lock, and separating the BRIDGE_RT_LOCK, to protect bridge_rtnode lists. That lock is taken in the data path (while in NET_EPOCH), so it cannot be a sleepable lock. While here document the locking strategy. MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D26418
2020-10-06 19:19:56 +00:00
BRIDGE_UNLOCK(sc);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
error = ether_ioctl(ifp, cmd, data);
bridge: call member interface ioctl() without NET_EPOCH We're not allowed to hold NET_EPOCH while sleeping, so when we call ioctl() handlers for member interfaces we cannot be in NET_EPOCH. We still need some protection of our CK_LISTs, so hold BRIDGE_LOCK instead. That requires changing BRIDGE_LOCK into a sleepable lock, and separating the BRIDGE_RT_LOCK, to protect bridge_rtnode lists. That lock is taken in the data path (while in NET_EPOCH), so it cannot be a sleepable lock. While here document the locking strategy. MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D26418
2020-10-06 19:19:56 +00:00
BRIDGE_LOCK(sc);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
break;
}
bridge: call member interface ioctl() without NET_EPOCH We're not allowed to hold NET_EPOCH while sleeping, so when we call ioctl() handlers for member interfaces we cannot be in NET_EPOCH. We still need some protection of our CK_LISTs, so hold BRIDGE_LOCK instead. That requires changing BRIDGE_LOCK into a sleepable lock, and separating the BRIDGE_RT_LOCK, to protect bridge_rtnode lists. That lock is taken in the data path (while in NET_EPOCH), so it cannot be a sleepable lock. While here document the locking strategy. MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D26418
2020-10-06 19:19:56 +00:00
BRIDGE_UNLOCK(sc);
bridge: epoch-ification Run the bridge datapath under epoch, rather than under the BRIDGE_LOCK(). We still take the BRIDGE_LOCK() whenever we insert or delete items in the relevant lists, but we use epoch callbacks to free items so that it's safe to iterate the lists without the BRIDGE_LOCK. Tests on mercat5/6 shows this increases bridge throughput significantly, from 3.7Mpps to 18.6Mpps. Reviewed by: emaste, philip, melifaro MFC after: 2 months Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24250
2020-04-26 16:22:35 +00:00
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
return (error);
}
Add code that clears certain capabilities from the member interface, these are restored when its removed from the bridge. At the moment we only clear IFCAP_TXCSUM. Since a locally generated packet on the bridge may be sent out any one or more interfaces it cant be assumed that every card does hardware csums. Most bridges don't generate a lot of traffic themselves so turning off offloading won't hurt, bridged packets are unaffected. Tested by: Bruce Walker (bmw borderware.com) MFC after: 5 days
2006-01-14 03:51:31 +00:00
/*
* bridge_mutecaps:
*
* Clear or restore unwanted capabilities on the member interface
*/
static void
Be smarter about disabling interface capabilities. TOE/TSO/TXCSUM will only be disabled if one (or more) of the member interfaces does not support it. Always turn off LRO since we can not bridge a combined frame. Tested by: Stefan Lambrev
2008-07-03 15:58:30 +00:00
bridge_mutecaps(struct bridge_softc *sc)
{
struct bridge_iflist *bif;
int enabled, mask;
bridge: epoch-ification Run the bridge datapath under epoch, rather than under the BRIDGE_LOCK(). We still take the BRIDGE_LOCK() whenever we insert or delete items in the relevant lists, but we use epoch callbacks to free items so that it's safe to iterate the lists without the BRIDGE_LOCK. Tests on mercat5/6 shows this increases bridge throughput significantly, from 3.7Mpps to 18.6Mpps. Reviewed by: emaste, philip, melifaro MFC after: 2 months Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24250
2020-04-26 16:22:35 +00:00
BRIDGE_LOCK_ASSERT(sc);
Be smarter about disabling interface capabilities. TOE/TSO/TXCSUM will only be disabled if one (or more) of the member interfaces does not support it. Always turn off LRO since we can not bridge a combined frame. Tested by: Stefan Lambrev
2008-07-03 15:58:30 +00:00
/* Initial bitmask of capabilities to test */
mask = BRIDGE_IFCAPS_MASK;
bridge: Change lists to CK_LIST as a peparation for epochification Prepare the ground for a rework of the bridge locking approach. We will use an epoch-based approach in the datapath and making it safe to iterate over the interface, span and rtnode lists without holding the BRIDGE_LOCK. Replace the relevant lists by their ConcurrencyKit equivalents. No functional change in this commit. Reviewed by: emaste, ae, philip (previous version) Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24249
2020-04-05 17:15:20 +00:00
CK_LIST_FOREACH(bif, &sc->sc_iflist, bif_next) {
Be smarter about disabling interface capabilities. TOE/TSO/TXCSUM will only be disabled if one (or more) of the member interfaces does not support it. Always turn off LRO since we can not bridge a combined frame. Tested by: Stefan Lambrev
2008-07-03 15:58:30 +00:00
/* Every member must support it or its disabled */
mask &= bif->bif_savedcaps;
}
bridge: Change lists to CK_LIST as a peparation for epochification Prepare the ground for a rework of the bridge locking approach. We will use an epoch-based approach in the datapath and making it safe to iterate over the interface, span and rtnode lists without holding the BRIDGE_LOCK. Replace the relevant lists by their ConcurrencyKit equivalents. No functional change in this commit. Reviewed by: emaste, ae, philip (previous version) Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24249
2020-04-05 17:15:20 +00:00
CK_LIST_FOREACH(bif, &sc->sc_iflist, bif_next) {
Be smarter about disabling interface capabilities. TOE/TSO/TXCSUM will only be disabled if one (or more) of the member interfaces does not support it. Always turn off LRO since we can not bridge a combined frame. Tested by: Stefan Lambrev
2008-07-03 15:58:30 +00:00
enabled = bif->bif_ifp->if_capenable;
When bridging LRO is causing a problem, the believe that it would work as long as all interfaces have TSO seems to be false, until the matter gets sorted out just disable LRO completely.
2009-08-24 21:04:51 +00:00
enabled &= ~BRIDGE_IFCAPS_STRIP;
Be smarter about disabling interface capabilities. TOE/TSO/TXCSUM will only be disabled if one (or more) of the member interfaces does not support it. Always turn off LRO since we can not bridge a combined frame. Tested by: Stefan Lambrev
2008-07-03 15:58:30 +00:00
/* strip off mask bits and enable them again if allowed */
enabled &= ~BRIDGE_IFCAPS_MASK;
enabled |= mask;
bridge_set_ifcap(sc, bif, enabled);
}
}
static void
bridge_set_ifcap(struct bridge_softc *sc, struct bridge_iflist *bif, int set)
Add code that clears certain capabilities from the member interface, these are restored when its removed from the bridge. At the moment we only clear IFCAP_TXCSUM. Since a locally generated packet on the bridge may be sent out any one or more interfaces it cant be assumed that every card does hardware csums. Most bridges don't generate a lot of traffic themselves so turning off offloading won't hurt, bridged packets are unaffected. Tested by: Bruce Walker (bmw borderware.com) MFC after: 5 days
2006-01-14 03:51:31 +00:00
{
struct ifnet *ifp = bif->bif_ifp;
struct ifreq ifr;
bridge: Fix spurious warnings about capabilities Mask off the bits we don't care about when checking that capabilities of the member interfaces have been disabled as intended. Submitted by: Ryan Moeller <ryan@ixsystems.com> Reviewed by: kristof, mav MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D18924
2019-03-04 22:01:09 +00:00
int error, mask, stuck;
Add code that clears certain capabilities from the member interface, these are restored when its removed from the bridge. At the moment we only clear IFCAP_TXCSUM. Since a locally generated packet on the bridge may be sent out any one or more interfaces it cant be assumed that every card does hardware csums. Most bridges don't generate a lot of traffic themselves so turning off offloading won't hurt, bridged packets are unaffected. Tested by: Bruce Walker (bmw borderware.com) MFC after: 5 days
2006-01-14 03:51:31 +00:00
Fix style(9) nits, whitespace and parentheses.
2006-05-16 22:50:41 +00:00
bzero(&ifr, sizeof(ifr));
Be smarter about disabling interface capabilities. TOE/TSO/TXCSUM will only be disabled if one (or more) of the member interfaces does not support it. Always turn off LRO since we can not bridge a combined frame. Tested by: Stefan Lambrev
2008-07-03 15:58:30 +00:00
ifr.ifr_reqcap = set;
Add code that clears certain capabilities from the member interface, these are restored when its removed from the bridge. At the moment we only clear IFCAP_TXCSUM. Since a locally generated packet on the bridge may be sent out any one or more interfaces it cant be assumed that every card does hardware csums. Most bridges don't generate a lot of traffic themselves so turning off offloading won't hurt, bridged packets are unaffected. Tested by: Bruce Walker (bmw borderware.com) MFC after: 5 days
2006-01-14 03:51:31 +00:00
Be smarter about disabling interface capabilities. TOE/TSO/TXCSUM will only be disabled if one (or more) of the member interfaces does not support it. Always turn off LRO since we can not bridge a combined frame. Tested by: Stefan Lambrev
2008-07-03 15:58:30 +00:00
if (ifp->if_capenable != set) {
Add code that clears certain capabilities from the member interface, these are restored when its removed from the bridge. At the moment we only clear IFCAP_TXCSUM. Since a locally generated packet on the bridge may be sent out any one or more interfaces it cant be assumed that every card does hardware csums. Most bridges don't generate a lot of traffic themselves so turning off offloading won't hurt, bridged packets are unaffected. Tested by: Bruce Walker (bmw borderware.com) MFC after: 5 days
2006-01-14 03:51:31 +00:00
error = (*ifp->if_ioctl)(ifp, SIOCSIFCAP, (caddr_t)&ifr);
Be smarter about disabling interface capabilities. TOE/TSO/TXCSUM will only be disabled if one (or more) of the member interfaces does not support it. Always turn off LRO since we can not bridge a combined frame. Tested by: Stefan Lambrev
2008-07-03 15:58:30 +00:00
if (error)
if_printf(sc->sc_ifp,
Make if_bridge complain if it can't disable some capabilities. MFC after: 2 weeks Sponsored by: iXsystems, Inc.
2017-04-29 08:52:07 +00:00
"error setting capabilities on %s: %d\n",
ifp->if_xname, error);
bridge: Fix spurious warnings about capabilities Mask off the bits we don't care about when checking that capabilities of the member interfaces have been disabled as intended. Submitted by: Ryan Moeller <ryan@ixsystems.com> Reviewed by: kristof, mav MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D18924
2019-03-04 22:01:09 +00:00
mask = BRIDGE_IFCAPS_MASK | BRIDGE_IFCAPS_STRIP;
stuck = ifp->if_capenable & mask & ~set;
if (stuck != 0)
Make if_bridge complain if it can't disable some capabilities. MFC after: 2 weeks Sponsored by: iXsystems, Inc.
2017-04-29 08:52:07 +00:00
if_printf(sc->sc_ifp,
"can't disable some capabilities on %s: 0x%x\n",
bridge: Fix spurious warnings about capabilities Mask off the bits we don't care about when checking that capabilities of the member interfaces have been disabled as intended. Submitted by: Ryan Moeller <ryan@ixsystems.com> Reviewed by: kristof, mav MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D18924
2019-03-04 22:01:09 +00:00
ifp->if_xname, stuck);
Add code that clears certain capabilities from the member interface, these are restored when its removed from the bridge. At the moment we only clear IFCAP_TXCSUM. Since a locally generated packet on the bridge may be sent out any one or more interfaces it cant be assumed that every card does hardware csums. Most bridges don't generate a lot of traffic themselves so turning off offloading won't hurt, bridged packets are unaffected. Tested by: Bruce Walker (bmw borderware.com) MFC after: 5 days
2006-01-14 03:51:31 +00:00
}
}
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
/*
* bridge_lookup_member:
*
* Lookup a bridge member interface.
*/
Make four more functions static that were missed in the last commit.
2005-10-14 20:57:02 +00:00
static struct bridge_iflist *
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge_lookup_member(struct bridge_softc *sc, const char *name)
{
struct bridge_iflist *bif;
struct ifnet *ifp;
bridge: call member interface ioctl() without NET_EPOCH We're not allowed to hold NET_EPOCH while sleeping, so when we call ioctl() handlers for member interfaces we cannot be in NET_EPOCH. We still need some protection of our CK_LISTs, so hold BRIDGE_LOCK instead. That requires changing BRIDGE_LOCK into a sleepable lock, and separating the BRIDGE_RT_LOCK, to protect bridge_rtnode lists. That lock is taken in the data path (while in NET_EPOCH), so it cannot be a sleepable lock. While here document the locking strategy. MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D26418
2020-10-06 19:19:56 +00:00
BRIDGE_LOCK_OR_NET_EPOCH_ASSERT(sc);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge: Change lists to CK_LIST as a peparation for epochification Prepare the ground for a rework of the bridge locking approach. We will use an epoch-based approach in the datapath and making it safe to iterate over the interface, span and rtnode lists without holding the BRIDGE_LOCK. Replace the relevant lists by their ConcurrencyKit equivalents. No functional change in this commit. Reviewed by: emaste, ae, philip (previous version) Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24249
2020-04-05 17:15:20 +00:00
CK_LIST_FOREACH(bif, &sc->sc_iflist, bif_next) {
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
ifp = bif->bif_ifp;
if (strcmp(ifp->if_xname, name) == 0)
return (bif);
}
return (NULL);
}
/*
* bridge_lookup_member_if:
*
* Lookup a bridge member interface by ifnet*.
*/
Make four more functions static that were missed in the last commit.
2005-10-14 20:57:02 +00:00
static struct bridge_iflist *
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge_lookup_member_if(struct bridge_softc *sc, struct ifnet *member_ifp)
{
struct bridge_iflist *bif;
bridge: call member interface ioctl() without NET_EPOCH We're not allowed to hold NET_EPOCH while sleeping, so when we call ioctl() handlers for member interfaces we cannot be in NET_EPOCH. We still need some protection of our CK_LISTs, so hold BRIDGE_LOCK instead. That requires changing BRIDGE_LOCK into a sleepable lock, and separating the BRIDGE_RT_LOCK, to protect bridge_rtnode lists. That lock is taken in the data path (while in NET_EPOCH), so it cannot be a sleepable lock. While here document the locking strategy. MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D26418
2020-10-06 19:19:56 +00:00
BRIDGE_LOCK_OR_NET_EPOCH_ASSERT(sc);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge: Change lists to CK_LIST as a peparation for epochification Prepare the ground for a rework of the bridge locking approach. We will use an epoch-based approach in the datapath and making it safe to iterate over the interface, span and rtnode lists without holding the BRIDGE_LOCK. Replace the relevant lists by their ConcurrencyKit equivalents. No functional change in this commit. Reviewed by: emaste, ae, philip (previous version) Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24249
2020-04-05 17:15:20 +00:00
CK_LIST_FOREACH(bif, &sc->sc_iflist, bif_next) {
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
if (bif->bif_ifp == member_ifp)
return (bif);
}
return (NULL);
}
bridge: epoch-ification Run the bridge datapath under epoch, rather than under the BRIDGE_LOCK(). We still take the BRIDGE_LOCK() whenever we insert or delete items in the relevant lists, but we use epoch callbacks to free items so that it's safe to iterate the lists without the BRIDGE_LOCK. Tests on mercat5/6 shows this increases bridge throughput significantly, from 3.7Mpps to 18.6Mpps. Reviewed by: emaste, philip, melifaro MFC after: 2 months Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24250
2020-04-26 16:22:35 +00:00
static void
bridge_delete_member_cb(struct epoch_context *ctx)
{
struct bridge_iflist *bif;
bif = __containerof(ctx, struct bridge_iflist, bif_epoch_ctx);
free(bif, M_DEVBUF);
}
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
/*
* bridge_delete_member:
*
* Delete the specified member interface.
*/
Change most of the bridge and stp funtions to static. This has highlighted that the following funtions are not used, wrap in '#ifdef noused' for the moment. bstp_enable_change_detection bstp_disable_change_detection bstp_set_bridge_priority bstp_set_port_priority bstp_set_path_cost
2005-10-14 10:38:12 +00:00
static void
If we have been called from ether_ifdetach() then do not try and clear the promisc flag from the member interface, this is a no-op anyway since the interface is disappearing. The driver may have already released its resources such as miibus and this is likely to panic the kernel. Submitted and tested by: Wojciech A. Koszek MFC after: 2 weeks
2005-10-23 22:30:07 +00:00
bridge_delete_member(struct bridge_softc *sc, struct bridge_iflist *bif,
int gone)
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
{
struct ifnet *ifs = bif->bif_ifp;
Set bridge MAC addresses to the MAC address of their first interface unless locally configured. This is more in line with the behaviour of other popular bridging implementations and makes bridges more predictable after reboots for example. Reviewed by: thompsa MFC after: 1 week
2008-07-01 08:14:58 +00:00
struct ifnet *fif = NULL;
bridge: Change lists to CK_LIST as a peparation for epochification Prepare the ground for a rework of the bridge locking approach. We will use an epoch-based approach in the datapath and making it safe to iterate over the interface, span and rtnode lists without holding the BRIDGE_LOCK. Replace the relevant lists by their ConcurrencyKit equivalents. No functional change in this commit. Reviewed by: emaste, ae, philip (previous version) Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24249
2020-04-05 17:15:20 +00:00
struct bridge_iflist *bifl;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
BRIDGE_LOCK_ASSERT(sc);
/tmp/cvsuusTrc
2006-07-26 10:43:02 +00:00
if (bif->bif_flags & IFBIF_STP)
use two stage creation of stp ports, this means that the stp variables can be set before the port is marked STP and they will no longer be overwrittten
2006-11-26 18:43:48 +00:00
bstp_disable(&bif->bif_stp);
/tmp/cvsuusTrc
2006-07-26 10:43:02 +00:00
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
ifs->if_bridge = NULL;
bridge: Change lists to CK_LIST as a peparation for epochification Prepare the ground for a rework of the bridge locking approach. We will use an epoch-based approach in the datapath and making it safe to iterate over the interface, span and rtnode lists without holding the BRIDGE_LOCK. Replace the relevant lists by their ConcurrencyKit equivalents. No functional change in this commit. Reviewed by: emaste, ae, philip (previous version) Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24249
2020-04-05 17:15:20 +00:00
CK_LIST_REMOVE(bif, bif_next);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
Set bridge MAC addresses to the MAC address of their first interface unless locally configured. This is more in line with the behaviour of other popular bridging implementations and makes bridges more predictable after reboots for example. Reviewed by: thompsa MFC after: 1 week
2008-07-01 08:14:58 +00:00
/*
* If removing the interface that gave the bridge its mac address, set
* the mac address of the bridge to the address of the next member, or
* to its default address if no members are left.
*/
Virtualize if_bridge(4) cloner.
2014-10-05 19:43:37 +00:00
if (V_bridge_inherit_mac && sc->sc_ifaddr == ifs) {
bridge: Change lists to CK_LIST as a peparation for epochification Prepare the ground for a rework of the bridge locking approach. We will use an epoch-based approach in the datapath and making it safe to iterate over the interface, span and rtnode lists without holding the BRIDGE_LOCK. Replace the relevant lists by their ConcurrencyKit equivalents. No functional change in this commit. Reviewed by: emaste, ae, philip (previous version) Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24249
2020-04-05 17:15:20 +00:00
if (CK_LIST_EMPTY(&sc->sc_iflist)) {
ether: centralize fake hwaddr generation We currently have two places with identical fake hwaddr generation -- if_vxlan and if_bridge. Lift it into if_ethersubr for reuse in other interfaces that may also need a fake addr. Reviewed by: bryanv, kp, philip Differential Revision: https://reviews.freebsd.org/D19573
2019-03-14 17:18:00 +00:00
bcopy(&sc->sc_defaddr,
Set bridge MAC addresses to the MAC address of their first interface unless locally configured. This is more in line with the behaviour of other popular bridging implementations and makes bridges more predictable after reboots for example. Reviewed by: thompsa MFC after: 1 week
2008-07-01 08:14:58 +00:00
IF_LLADDR(sc->sc_ifp), ETHER_ADDR_LEN);
bridge_delete_member is called via the event handler from if_detach after the LLADDR is reclaimed which causes a null pointer deref with inherit_mac enabled. Record the ifnet pointer of the interface and then compare that to find when to re-assign the bridge address. Submitted by: sam
2009-02-13 19:20:25 +00:00
sc->sc_ifaddr = NULL;
} else {
bridge: Change lists to CK_LIST as a peparation for epochification Prepare the ground for a rework of the bridge locking approach. We will use an epoch-based approach in the datapath and making it safe to iterate over the interface, span and rtnode lists without holding the BRIDGE_LOCK. Replace the relevant lists by their ConcurrencyKit equivalents. No functional change in this commit. Reviewed by: emaste, ae, philip (previous version) Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24249
2020-04-05 17:15:20 +00:00
bifl = CK_LIST_FIRST(&sc->sc_iflist);
fif = bifl->bif_ifp;
Set bridge MAC addresses to the MAC address of their first interface unless locally configured. This is more in line with the behaviour of other popular bridging implementations and makes bridges more predictable after reboots for example. Reviewed by: thompsa MFC after: 1 week
2008-07-01 08:14:58 +00:00
bcopy(IF_LLADDR(fif),
IF_LLADDR(sc->sc_ifp), ETHER_ADDR_LEN);
bridge_delete_member is called via the event handler from if_detach after the LLADDR is reclaimed which causes a null pointer deref with inherit_mac enabled. Record the ifnet pointer of the interface and then compare that to find when to re-assign the bridge address. Submitted by: sam
2009-02-13 19:20:25 +00:00
sc->sc_ifaddr = fif;
Set bridge MAC addresses to the MAC address of their first interface unless locally configured. This is more in line with the behaviour of other popular bridging implementations and makes bridges more predictable after reboots for example. Reviewed by: thompsa MFC after: 1 week
2008-07-01 08:14:58 +00:00
}
Declare a new EVENTHANDLER called iflladdr_event which signals that the L2 address on an interface has changed. This lets stacked interfaces such as vlan(4) detect that their lower interface has changed and adjust things in order to keep working. Previously this situation broke at least vlan(4) and lagg(4) configurations. The EVENTHANDLER_INVOKE call was not placed within if_setlladdr() due to the risk of a loop. PR: kern/142927 Submitted by: Nikolay Denev
2010-01-18 20:34:00 +00:00
EVENTHANDLER_INVOKE(iflladdr_event, sc->sc_ifp);
Set bridge MAC addresses to the MAC address of their first interface unless locally configured. This is more in line with the behaviour of other popular bridging implementations and makes bridges more predictable after reboots for example. Reviewed by: thompsa MFC after: 1 week
2008-07-01 08:14:58 +00:00
}
Fix a panic I introduced in r234487, the bridge softc pointer is set to null early in the detach so rearrange things not to explode. Reported by: David Roffiaen, Gustau Perez Querol Tested by: David Roffiaen MFC after: 3 days
2012-06-11 20:12:13 +00:00
bridge_linkcheck(sc);
Be smarter about disabling interface capabilities. TOE/TSO/TXCSUM will only be disabled if one (or more) of the member interfaces does not support it. Always turn off LRO since we can not bridge a combined frame. Tested by: Stefan Lambrev
2008-07-03 15:58:30 +00:00
bridge_mutecaps(sc); /* recalcuate now this interface is removed */
bridge: call member interface ioctl() without NET_EPOCH We're not allowed to hold NET_EPOCH while sleeping, so when we call ioctl() handlers for member interfaces we cannot be in NET_EPOCH. We still need some protection of our CK_LISTs, so hold BRIDGE_LOCK instead. That requires changing BRIDGE_LOCK into a sleepable lock, and separating the BRIDGE_RT_LOCK, to protect bridge_rtnode lists. That lock is taken in the data path (while in NET_EPOCH), so it cannot be a sleepable lock. While here document the locking strategy. MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D26418
2020-10-06 19:19:56 +00:00
BRIDGE_RT_LOCK(sc);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge_rtdelete(sc, ifs, IFBF_FLUSHALL);
bridge: call member interface ioctl() without NET_EPOCH We're not allowed to hold NET_EPOCH while sleeping, so when we call ioctl() handlers for member interfaces we cannot be in NET_EPOCH. We still need some protection of our CK_LISTs, so hold BRIDGE_LOCK instead. That requires changing BRIDGE_LOCK into a sleepable lock, and separating the BRIDGE_RT_LOCK, to protect bridge_rtnode lists. That lock is taken in the data path (while in NET_EPOCH), so it cannot be a sleepable lock. While here document the locking strategy. MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D26418
2020-10-06 19:19:56 +00:00
BRIDGE_RT_UNLOCK(sc);
Add an option to limit the number of source MACs that can be behind a bridge interface. Once the limit is reached packets with unknown source addresses are dropped until an existing host cache entry expires or is removed. Useful to use with the STICKY cache option. Sponsored by: miniSuperHappyDevHouse NZ
2007-11-04 08:32:27 +00:00
KASSERT(bif->bif_addrcnt == 0,
("%s: %d bridge routes referenced", __func__, bif->bif_addrcnt));
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
Allow different bridge types to coexist if_bridge has a lot of limitations that make it scale poorly to higher data rates. In my projects/VPC branch I leverage the bridge interface between layers for my high speed soft switch as well as for purposes of stacking in general. Reviewed by: sbruno@ Approved by: sbruno@ Differential Revision: https://reviews.freebsd.org/D15344
2018-05-11 05:00:40 +00:00
ifs->if_bridge_output = NULL;
ifs->if_bridge_input = NULL;
ifs->if_bridge_linkstate = NULL;
Reorder the bridge add and delete routines to avoid calling ifpromisc() with the bridge lock held.
2009-05-01 19:46:42 +00:00
if (!gone) {
switch (ifs->if_type) {
case IFT_ETHER:
case IFT_L2VLAN:
/*
Fix panic when adding vtnet interfaces to a bridge vtnet interfaces are always in promiscuous mode (at least if the VIRTIO_NET_F_CTRL_RX feature is not negotiated with the host). if_promisc() on a vtnet interface returned ENOTSUP although it has IFF_PROMISC set. This confused the bridge code. Instead we now accept all enable/disable promiscuous commands (and always keep IFF_PROMISC set). There are also two issues with the if_bridge error handling. If if_promisc() fails it uses bridge_delete_member() to clean up. This tries to disable promiscuous mode on the interface. That runs into an assert, because promiscuous mode was never set in the first place. (That's the panic reported in PR 200210.) We can only unset promiscuous mode if the interface actually is promiscuous. This goes against the reference counting done by if_promisc(), but only the first/last if_promic() calls can actually fail, so this is safe. A second issue is a double free of bif. It's already freed by bridge_delete_member(). PR: 200210 Differential Revision: https://reviews.freebsd.org/D2804 Reviewed by: philip (mentor)
2015-06-13 19:39:21 +00:00
* Take the interface out of promiscuous mode, but only
* if it was promiscuous in the first place. It might
* not be if we're in the bridge_ioctl_add() error path.
Reorder the bridge add and delete routines to avoid calling ifpromisc() with the bridge lock held.
2009-05-01 19:46:42 +00:00
*/
Fix panic when adding vtnet interfaces to a bridge vtnet interfaces are always in promiscuous mode (at least if the VIRTIO_NET_F_CTRL_RX feature is not negotiated with the host). if_promisc() on a vtnet interface returned ENOTSUP although it has IFF_PROMISC set. This confused the bridge code. Instead we now accept all enable/disable promiscuous commands (and always keep IFF_PROMISC set). There are also two issues with the if_bridge error handling. If if_promisc() fails it uses bridge_delete_member() to clean up. This tries to disable promiscuous mode on the interface. That runs into an assert, because promiscuous mode was never set in the first place. (That's the panic reported in PR 200210.) We can only unset promiscuous mode if the interface actually is promiscuous. This goes against the reference counting done by if_promisc(), but only the first/last if_promic() calls can actually fail, so this is safe. A second issue is a double free of bif. It's already freed by bridge_delete_member(). PR: 200210 Differential Revision: https://reviews.freebsd.org/D2804 Reviewed by: philip (mentor)
2015-06-13 19:39:21 +00:00
if (ifs->if_flags & IFF_PROMISC)
(void) ifpromisc(ifs, 0);
Reorder the bridge add and delete routines to avoid calling ifpromisc() with the bridge lock held.
2009-05-01 19:46:42 +00:00
break;
case IFT_GIF:
break;
default:
#ifdef DIAGNOSTIC
panic("bridge_delete_member: impossible");
#endif
break;
}
/* reneable any interface capabilities */
bridge_set_ifcap(sc, bif, bif->bif_savedcaps);
}
use two stage creation of stp ports, this means that the stp variables can be set before the port is marked STP and they will no longer be overwrittten
2006-11-26 18:43:48 +00:00
bstp_destroy(&bif->bif_stp); /* prepare to free */
bridge: epoch-ification Run the bridge datapath under epoch, rather than under the BRIDGE_LOCK(). We still take the BRIDGE_LOCK() whenever we insert or delete items in the relevant lists, but we use epoch callbacks to free items so that it's safe to iterate the lists without the BRIDGE_LOCK. Tests on mercat5/6 shows this increases bridge throughput significantly, from 3.7Mpps to 18.6Mpps. Reviewed by: emaste, philip, melifaro MFC after: 2 months Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24250
2020-04-26 16:22:35 +00:00
NET_EPOCH_CALL(bridge_delete_member_cb, &bif->bif_epoch_ctx);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
}
Change from a callback in if_ethersubr to using EVENTHANDLER in order to detach span ports when they disappear. The span port does not have a pointer to the softc so revert r1.31 and bring back the softc linked-list. MFC after: 2 weeks
2005-12-17 06:33:51 +00:00
/*
* bridge_delete_span:
*
* Delete the specified span interface.
*/
static void
bridge_delete_span(struct bridge_softc *sc, struct bridge_iflist *bif)
{
BRIDGE_LOCK_ASSERT(sc);
KASSERT(bif->bif_ifp->if_bridge == NULL,
("%s: not a span interface", __func__));
bridge: Change lists to CK_LIST as a peparation for epochification Prepare the ground for a rework of the bridge locking approach. We will use an epoch-based approach in the datapath and making it safe to iterate over the interface, span and rtnode lists without holding the BRIDGE_LOCK. Replace the relevant lists by their ConcurrencyKit equivalents. No functional change in this commit. Reviewed by: emaste, ae, philip (previous version) Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24249
2020-04-05 17:15:20 +00:00
CK_LIST_REMOVE(bif, bif_next);
bridge: epoch-ification Run the bridge datapath under epoch, rather than under the BRIDGE_LOCK(). We still take the BRIDGE_LOCK() whenever we insert or delete items in the relevant lists, but we use epoch callbacks to free items so that it's safe to iterate the lists without the BRIDGE_LOCK. Tests on mercat5/6 shows this increases bridge throughput significantly, from 3.7Mpps to 18.6Mpps. Reviewed by: emaste, philip, melifaro MFC after: 2 months Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24250
2020-04-26 16:22:35 +00:00
NET_EPOCH_CALL(bridge_delete_member_cb, &bif->bif_epoch_ctx);
Change from a callback in if_ethersubr to using EVENTHANDLER in order to detach span ports when they disappear. The span port does not have a pointer to the softc so revert r1.31 and bring back the softc linked-list. MFC after: 2 weeks
2005-12-17 06:33:51 +00:00
}
Change most of the bridge and stp funtions to static. This has highlighted that the following funtions are not used, wrap in '#ifdef noused' for the moment. bstp_enable_change_detection bstp_disable_change_detection bstp_set_bridge_priority bstp_set_port_priority bstp_set_path_cost
2005-10-14 10:38:12 +00:00
static int
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge_ioctl_add(struct bridge_softc *sc, void *arg)
{
struct ifbreq *req = arg;
struct bridge_iflist *bif = NULL;
struct ifnet *ifs;
int error = 0;
ifs = ifunit(req->ifbr_ifsname);
if (ifs == NULL)
return (ENOENT);
Be smarter about disabling interface capabilities. TOE/TSO/TXCSUM will only be disabled if one (or more) of the member interfaces does not support it. Always turn off LRO since we can not bridge a combined frame. Tested by: Stefan Lambrev
2008-07-03 15:58:30 +00:00
if (ifs->if_ioctl == NULL) /* must be supported */
return (EINVAL);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
Add support for creating span ports so that one can snoop bridged traffic from another interface/machine/network. Obtained from: OpenBSD MFC after: 2 weeks
2005-12-14 02:52:13 +00:00
/* If it's in the span list, it can't be a member. */
bridge: Change lists to CK_LIST as a peparation for epochification Prepare the ground for a rework of the bridge locking approach. We will use an epoch-based approach in the datapath and making it safe to iterate over the interface, span and rtnode lists without holding the BRIDGE_LOCK. Replace the relevant lists by their ConcurrencyKit equivalents. No functional change in this commit. Reviewed by: emaste, ae, philip (previous version) Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24249
2020-04-05 17:15:20 +00:00
CK_LIST_FOREACH(bif, &sc->sc_spanlist, bif_next)
Add support for creating span ports so that one can snoop bridged traffic from another interface/machine/network. Obtained from: OpenBSD MFC after: 2 weeks
2005-12-14 02:52:13 +00:00
if (ifs == bif->bif_ifp)
return (EBUSY);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
if (ifs->if_bridge == sc)
return (EEXIST);
if (ifs->if_bridge != NULL)
return (EBUSY);
switch (ifs->if_type) {
case IFT_ETHER:
case IFT_L2VLAN:
Add RFC 3378 EtherIP support. This change makes it possible to add gif interfaces to bridges, which will then send and receive IP protocol 97 packets. Packets are Ethernet frames with an EtherIP header prepended. Obtained from: NetBSD MFC after: 2 weeks
2005-12-21 21:29:45 +00:00
case IFT_GIF:
Reorder the bridge add and delete routines to avoid calling ifpromisc() with the bridge lock held.
2009-05-01 19:46:42 +00:00
/* permitted interface types */
Add RFC 3378 EtherIP support. This change makes it possible to add gif interfaces to bridges, which will then send and receive IP protocol 97 packets. Packets are Ethernet frames with an EtherIP header prepended. Obtained from: NetBSD MFC after: 2 weeks
2005-12-21 21:29:45 +00:00
break;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
default:
- Allow ND6_IFF_AUTO_LINKLOCAL for IFT_BRIDGE. An interface with IFT_BRIDGE is initialized with !ND6_IFF_AUTO_LINKLOCAL && !ND6_IFF_ACCEPT_RTADV regardless of net.inet6.ip6.accept_rtadv and net.inet6.ip6.auto_linklocal. To configure an autoconfigured link-local address (RFC 4862), the following rc.conf(5) configuration can be used: ifconfig_bridge0_ipv6="inet6 auto_linklocal" - if_bridge(4) now removes IPv6 addresses on a member interface to be added when the parent interface or one of the existing member interfaces has an IPv6 address. if_bridge(4) merges each link-local scope zone which the member interfaces form respectively, so it causes address scope violation. Removal of the IPv6 addresses prevents it. - if_lagg(4) now removes IPv6 addresses on a member interfaces unconditionally. - Set reasonable flags to non-IPv6-capable interfaces. [*] Submitted by: rpaulo [*] MFC after: 1 week
2013-07-02 16:58:15 +00:00
return (EINVAL);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
}
- Allow ND6_IFF_AUTO_LINKLOCAL for IFT_BRIDGE. An interface with IFT_BRIDGE is initialized with !ND6_IFF_AUTO_LINKLOCAL && !ND6_IFF_ACCEPT_RTADV regardless of net.inet6.ip6.accept_rtadv and net.inet6.ip6.auto_linklocal. To configure an autoconfigured link-local address (RFC 4862), the following rc.conf(5) configuration can be used: ifconfig_bridge0_ipv6="inet6 auto_linklocal" - if_bridge(4) now removes IPv6 addresses on a member interface to be added when the parent interface or one of the existing member interfaces has an IPv6 address. if_bridge(4) merges each link-local scope zone which the member interfaces form respectively, so it causes address scope violation. Removal of the IPv6 addresses prevents it. - if_lagg(4) now removes IPv6 addresses on a member interfaces unconditionally. - Set reasonable flags to non-IPv6-capable interfaces. [*] Submitted by: rpaulo [*] MFC after: 1 week
2013-07-02 16:58:15 +00:00
#ifdef INET6
/*
* Two valid inet6 addresses with link-local scope must not be
* on the parent interface and the member interfaces at the
* same time. This restriction is needed to prevent violation
* of link-local scope zone. Attempts to add a member
* interface which has inet6 addresses when the parent has
* inet6 triggers removal of all inet6 addresses on the member
* interface.
*/
/* Check if the parent interface has a link-local scope addr. */
- Relax the restriction on the member interfaces with LLAs. Two or more LLAs on the member interfaces are actually harmless when the parent interface does not have a LLA. - Add net.link.bridge.allow_llz_overlap. This is a knob to allow LLAs on a bridge and the member interfaces at the same time. The default is 0. Pointed out by: ume MFC after: 3 days
2013-07-28 19:49:39 +00:00
if (V_allow_llz_overlap == 0 &&
in6ifa_llaonifp(sc->sc_ifp) != NULL) {
- Allow ND6_IFF_AUTO_LINKLOCAL for IFT_BRIDGE. An interface with IFT_BRIDGE is initialized with !ND6_IFF_AUTO_LINKLOCAL && !ND6_IFF_ACCEPT_RTADV regardless of net.inet6.ip6.accept_rtadv and net.inet6.ip6.auto_linklocal. To configure an autoconfigured link-local address (RFC 4862), the following rc.conf(5) configuration can be used: ifconfig_bridge0_ipv6="inet6 auto_linklocal" - if_bridge(4) now removes IPv6 addresses on a member interface to be added when the parent interface or one of the existing member interfaces has an IPv6 address. if_bridge(4) merges each link-local scope zone which the member interfaces form respectively, so it causes address scope violation. Removal of the IPv6 addresses prevents it. - if_lagg(4) now removes IPv6 addresses on a member interfaces unconditionally. - Set reasonable flags to non-IPv6-capable interfaces. [*] Submitted by: rpaulo [*] MFC after: 1 week
2013-07-02 16:58:15 +00:00
/*
* If any, remove all inet6 addresses from the member
* interfaces.
*/
bridge: Change lists to CK_LIST as a peparation for epochification Prepare the ground for a rework of the bridge locking approach. We will use an epoch-based approach in the datapath and making it safe to iterate over the interface, span and rtnode lists without holding the BRIDGE_LOCK. Replace the relevant lists by their ConcurrencyKit equivalents. No functional change in this commit. Reviewed by: emaste, ae, philip (previous version) Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24249
2020-04-05 17:15:20 +00:00
CK_LIST_FOREACH(bif, &sc->sc_iflist, bif_next) {
- Allow ND6_IFF_AUTO_LINKLOCAL for IFT_BRIDGE. An interface with IFT_BRIDGE is initialized with !ND6_IFF_AUTO_LINKLOCAL && !ND6_IFF_ACCEPT_RTADV regardless of net.inet6.ip6.accept_rtadv and net.inet6.ip6.auto_linklocal. To configure an autoconfigured link-local address (RFC 4862), the following rc.conf(5) configuration can be used: ifconfig_bridge0_ipv6="inet6 auto_linklocal" - if_bridge(4) now removes IPv6 addresses on a member interface to be added when the parent interface or one of the existing member interfaces has an IPv6 address. if_bridge(4) merges each link-local scope zone which the member interfaces form respectively, so it causes address scope violation. Removal of the IPv6 addresses prevents it. - if_lagg(4) now removes IPv6 addresses on a member interfaces unconditionally. - Set reasonable flags to non-IPv6-capable interfaces. [*] Submitted by: rpaulo [*] MFC after: 1 week
2013-07-02 16:58:15 +00:00
if (in6ifa_llaonifp(bif->bif_ifp)) {
in6_ifdetach(bif->bif_ifp);
if_printf(sc->sc_ifp,
"IPv6 addresses on %s have been removed "
"before adding it as a member to prevent "
"IPv6 address scope violation.\n",
bif->bif_ifp->if_xname);
}
}
if (in6ifa_llaonifp(ifs)) {
in6_ifdetach(ifs);
if_printf(sc->sc_ifp,
"IPv6 addresses on %s have been removed "
"before adding it as a member to prevent "
"IPv6 address scope violation.\n",
ifs->if_xname);
}
}
#endif
- Check if_type of "addm <interface>" before setting the interface's MTU to the if_bridge(4) interface. This fixes a bug that MTU value of "addm <interface>" is used even when it is invalid for the if_bridge(4) member: # ifconfig bridge0 create # ifconfig bridge0 bridge0: flags=8802<BROADCAST,SIMPLEX,MULTICAST> metric 0 mtu 1500 ... # ifconfig bridge0 addm lo0 ifconfig: BRDGADD lo0: Invalid argument # ifconfig bridge0 bridge0: flags=8802<BROADCAST,SIMPLEX,MULTICAST> metric 0 mtu 16384 ... - Do not ignore MTU value of an interface even when if_type == IFT_GIF. This fixes MTU mismatch when an if_bridge(4) interface has a gif(4) interface and no other interface as the member, and it is directly used for L2 communication with EtherIP tunneling enabled. - Implement SIOCSIFMTU ioctl. Changing the MTU is allowed only when all members have the same MTU value.
2010-01-31 08:16:37 +00:00
/* Allow the first Ethernet member to define the MTU */
bridge: Change lists to CK_LIST as a peparation for epochification Prepare the ground for a rework of the bridge locking approach. We will use an epoch-based approach in the datapath and making it safe to iterate over the interface, span and rtnode lists without holding the BRIDGE_LOCK. Replace the relevant lists by their ConcurrencyKit equivalents. No functional change in this commit. Reviewed by: emaste, ae, philip (previous version) Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24249
2020-04-05 17:15:20 +00:00
if (CK_LIST_EMPTY(&sc->sc_iflist))
- Check if_type of "addm <interface>" before setting the interface's MTU to the if_bridge(4) interface. This fixes a bug that MTU value of "addm <interface>" is used even when it is invalid for the if_bridge(4) member: # ifconfig bridge0 create # ifconfig bridge0 bridge0: flags=8802<BROADCAST,SIMPLEX,MULTICAST> metric 0 mtu 1500 ... # ifconfig bridge0 addm lo0 ifconfig: BRDGADD lo0: Invalid argument # ifconfig bridge0 bridge0: flags=8802<BROADCAST,SIMPLEX,MULTICAST> metric 0 mtu 16384 ... - Do not ignore MTU value of an interface even when if_type == IFT_GIF. This fixes MTU mismatch when an if_bridge(4) interface has a gif(4) interface and no other interface as the member, and it is directly used for L2 communication with EtherIP tunneling enabled. - Implement SIOCSIFMTU ioctl. Changing the MTU is allowed only when all members have the same MTU value.
2010-01-31 08:16:37 +00:00
sc->sc_ifp->if_mtu = ifs->if_mtu;
else if (sc->sc_ifp->if_mtu != ifs->if_mtu) {
Since 32-bit if_baudrate isn't enough to describe a baud rate of a 10 Gbit interface, in the r241616 a crutch was provided. It didn't work well, and finally we decided that it is time to break ABI and simply make if_baudrate a 64-bit value. Meanwhile, the entire struct if_data was reviewed. o Remove the if_baudrate_pf crutch. o Make all fields of struct if_data fixed machine independent size. The notion of data (packet counters, etc) are by no means MD. And it is a bug that on amd64 we've got a 64-bit counters, while on i386 32-bit, which at modern speeds overflow within a second. This also removes quite a lot of COMPAT_FREEBSD32 code. o Give 16 bit for the ifi_datalen field. This field was provided to make future changes to if_data less ABI breaking. Unfortunately the 8 bit size of it had effectively limited sizeof if_data to 256 bytes. o Give 32 bits to ifi_mtu and ifi_metric. o Give 64 bits to the rest of fields, since they are counters. __FreeBSD_version bumped. Discussed with: emax Sponsored by: Netflix Sponsored by: Nginx, Inc.
2014-03-13 03:42:24 +00:00
if_printf(sc->sc_ifp, "invalid MTU: %u(%s) != %u\n",
- Check if_type of "addm <interface>" before setting the interface's MTU to the if_bridge(4) interface. This fixes a bug that MTU value of "addm <interface>" is used even when it is invalid for the if_bridge(4) member: # ifconfig bridge0 create # ifconfig bridge0 bridge0: flags=8802<BROADCAST,SIMPLEX,MULTICAST> metric 0 mtu 1500 ... # ifconfig bridge0 addm lo0 ifconfig: BRDGADD lo0: Invalid argument # ifconfig bridge0 bridge0: flags=8802<BROADCAST,SIMPLEX,MULTICAST> metric 0 mtu 16384 ... - Do not ignore MTU value of an interface even when if_type == IFT_GIF. This fixes MTU mismatch when an if_bridge(4) interface has a gif(4) interface and no other interface as the member, and it is directly used for L2 communication with EtherIP tunneling enabled. - Implement SIOCSIFMTU ioctl. Changing the MTU is allowed only when all members have the same MTU value.
2010-01-31 08:16:37 +00:00
ifs->if_mtu, ifs->if_xname, sc->sc_ifp->if_mtu);
- Allow ND6_IFF_AUTO_LINKLOCAL for IFT_BRIDGE. An interface with IFT_BRIDGE is initialized with !ND6_IFF_AUTO_LINKLOCAL && !ND6_IFF_ACCEPT_RTADV regardless of net.inet6.ip6.accept_rtadv and net.inet6.ip6.auto_linklocal. To configure an autoconfigured link-local address (RFC 4862), the following rc.conf(5) configuration can be used: ifconfig_bridge0_ipv6="inet6 auto_linklocal" - if_bridge(4) now removes IPv6 addresses on a member interface to be added when the parent interface or one of the existing member interfaces has an IPv6 address. if_bridge(4) merges each link-local scope zone which the member interfaces form respectively, so it causes address scope violation. Removal of the IPv6 addresses prevents it. - if_lagg(4) now removes IPv6 addresses on a member interfaces unconditionally. - Set reasonable flags to non-IPv6-capable interfaces. [*] Submitted by: rpaulo [*] MFC after: 1 week
2013-07-02 16:58:15 +00:00
return (EINVAL);
- Check if_type of "addm <interface>" before setting the interface's MTU to the if_bridge(4) interface. This fixes a bug that MTU value of "addm <interface>" is used even when it is invalid for the if_bridge(4) member: # ifconfig bridge0 create # ifconfig bridge0 bridge0: flags=8802<BROADCAST,SIMPLEX,MULTICAST> metric 0 mtu 1500 ... # ifconfig bridge0 addm lo0 ifconfig: BRDGADD lo0: Invalid argument # ifconfig bridge0 bridge0: flags=8802<BROADCAST,SIMPLEX,MULTICAST> metric 0 mtu 16384 ... - Do not ignore MTU value of an interface even when if_type == IFT_GIF. This fixes MTU mismatch when an if_bridge(4) interface has a gif(4) interface and no other interface as the member, and it is directly used for L2 communication with EtherIP tunneling enabled. - Implement SIOCSIFMTU ioctl. Changing the MTU is allowed only when all members have the same MTU value.
2010-01-31 08:16:37 +00:00
}
- Allow ND6_IFF_AUTO_LINKLOCAL for IFT_BRIDGE. An interface with IFT_BRIDGE is initialized with !ND6_IFF_AUTO_LINKLOCAL && !ND6_IFF_ACCEPT_RTADV regardless of net.inet6.ip6.accept_rtadv and net.inet6.ip6.auto_linklocal. To configure an autoconfigured link-local address (RFC 4862), the following rc.conf(5) configuration can be used: ifconfig_bridge0_ipv6="inet6 auto_linklocal" - if_bridge(4) now removes IPv6 addresses on a member interface to be added when the parent interface or one of the existing member interfaces has an IPv6 address. if_bridge(4) merges each link-local scope zone which the member interfaces form respectively, so it causes address scope violation. Removal of the IPv6 addresses prevents it. - if_lagg(4) now removes IPv6 addresses on a member interfaces unconditionally. - Set reasonable flags to non-IPv6-capable interfaces. [*] Submitted by: rpaulo [*] MFC after: 1 week
2013-07-02 16:58:15 +00:00
bif = malloc(sizeof(*bif), M_DEVBUF, M_NOWAIT|M_ZERO);
if (bif == NULL)
return (ENOMEM);
bif->bif_ifp = ifs;
bif->bif_flags = IFBIF_LEARNING | IFBIF_DISCOVER;
bif->bif_savedcaps = ifs->if_capenable;
Set bridge MAC addresses to the MAC address of their first interface unless locally configured. This is more in line with the behaviour of other popular bridging implementations and makes bridges more predictable after reboots for example. Reviewed by: thompsa MFC after: 1 week
2008-07-01 08:14:58 +00:00
/*
* Assign the interface's MAC address to the bridge if it's the first
* member and the MAC address of the bridge has not been changed from
* the default randomly generated one.
*/
bridge: Change lists to CK_LIST as a peparation for epochification Prepare the ground for a rework of the bridge locking approach. We will use an epoch-based approach in the datapath and making it safe to iterate over the interface, span and rtnode lists without holding the BRIDGE_LOCK. Replace the relevant lists by their ConcurrencyKit equivalents. No functional change in this commit. Reviewed by: emaste, ae, philip (previous version) Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24249
2020-04-05 17:15:20 +00:00
if (V_bridge_inherit_mac && CK_LIST_EMPTY(&sc->sc_iflist) &&
ether: centralize fake hwaddr generation We currently have two places with identical fake hwaddr generation -- if_vxlan and if_bridge. Lift it into if_ethersubr for reuse in other interfaces that may also need a fake addr. Reviewed by: bryanv, kp, philip Differential Revision: https://reviews.freebsd.org/D19573
2019-03-14 17:18:00 +00:00
!memcmp(IF_LLADDR(sc->sc_ifp), sc->sc_defaddr.octet, ETHER_ADDR_LEN)) {
Set bridge MAC addresses to the MAC address of their first interface unless locally configured. This is more in line with the behaviour of other popular bridging implementations and makes bridges more predictable after reboots for example. Reviewed by: thompsa MFC after: 1 week
2008-07-01 08:14:58 +00:00
bcopy(IF_LLADDR(ifs), IF_LLADDR(sc->sc_ifp), ETHER_ADDR_LEN);
bridge_delete_member is called via the event handler from if_detach after the LLADDR is reclaimed which causes a null pointer deref with inherit_mac enabled. Record the ifnet pointer of the interface and then compare that to find when to re-assign the bridge address. Submitted by: sam
2009-02-13 19:20:25 +00:00
sc->sc_ifaddr = ifs;
Declare a new EVENTHANDLER called iflladdr_event which signals that the L2 address on an interface has changed. This lets stacked interfaces such as vlan(4) detect that their lower interface has changed and adjust things in order to keep working. Previously this situation broke at least vlan(4) and lagg(4) configurations. The EVENTHANDLER_INVOKE call was not placed within if_setlladdr() due to the risk of a loop. PR: kern/142927 Submitted by: Nikolay Denev
2010-01-18 20:34:00 +00:00
EVENTHANDLER_INVOKE(iflladdr_event, sc->sc_ifp);
bridge_delete_member is called via the event handler from if_detach after the LLADDR is reclaimed which causes a null pointer deref with inherit_mac enabled. Record the ifnet pointer of the interface and then compare that to find when to re-assign the bridge address. Submitted by: sam
2009-02-13 19:20:25 +00:00
}
Set bridge MAC addresses to the MAC address of their first interface unless locally configured. This is more in line with the behaviour of other popular bridging implementations and makes bridges more predictable after reboots for example. Reviewed by: thompsa MFC after: 1 week
2008-07-01 08:14:58 +00:00
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
ifs->if_bridge = sc;
Allow different bridge types to coexist if_bridge has a lot of limitations that make it scale poorly to higher data rates. In my projects/VPC branch I leverage the bridge interface between layers for my high speed soft switch as well as for purposes of stacking in general. Reviewed by: sbruno@ Approved by: sbruno@ Differential Revision: https://reviews.freebsd.org/D15344
2018-05-11 05:00:40 +00:00
ifs->if_bridge_output = bridge_output;
ifs->if_bridge_input = bridge_input;
ifs->if_bridge_linkstate = bridge_linkstate;
use two stage creation of stp ports, this means that the stp variables can be set before the port is marked STP and they will no longer be overwrittten
2006-11-26 18:43:48 +00:00
bstp_create(&sc->sc_stp, &bif->bif_stp, bif->bif_ifp);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
/*
* XXX: XLOCK HERE!?!
*
* NOTE: insert_***HEAD*** should be safe for the traversals.
*/
bridge: Change lists to CK_LIST as a peparation for epochification Prepare the ground for a rework of the bridge locking approach. We will use an epoch-based approach in the datapath and making it safe to iterate over the interface, span and rtnode lists without holding the BRIDGE_LOCK. Replace the relevant lists by their ConcurrencyKit equivalents. No functional change in this commit. Reviewed by: emaste, ae, philip (previous version) Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24249
2020-04-05 17:15:20 +00:00
CK_LIST_INSERT_HEAD(&sc->sc_iflist, bif, bif_next);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
Be smarter about disabling interface capabilities. TOE/TSO/TXCSUM will only be disabled if one (or more) of the member interfaces does not support it. Always turn off LRO since we can not bridge a combined frame. Tested by: Stefan Lambrev
2008-07-03 15:58:30 +00:00
/* Set interface capabilities to the intersection set of all members */
bridge_mutecaps(sc);
Fix a panic I introduced in r234487, the bridge softc pointer is set to null early in the detach so rearrange things not to explode. Reported by: David Roffiaen, Gustau Perez Querol Tested by: David Roffiaen MFC after: 3 days
2012-06-11 20:12:13 +00:00
bridge_linkcheck(sc);
Reorder the bridge add and delete routines to avoid calling ifpromisc() with the bridge lock held.
2009-05-01 19:46:42 +00:00
Add linkstate to bridge(4), set the link to up when at least one underlying interface is up, otherwise the link is down. This, among other things, allows carp to work on a bridge. Prodded by: glebius Tested by: Alexander Lunev
2012-04-20 09:55:50 +00:00
/* Place the interface into promiscuous mode */
Reorder the bridge add and delete routines to avoid calling ifpromisc() with the bridge lock held.
2009-05-01 19:46:42 +00:00
switch (ifs->if_type) {
Add linkstate to bridge(4), set the link to up when at least one underlying interface is up, otherwise the link is down. This, among other things, allows carp to work on a bridge. Prodded by: glebius Tested by: Alexander Lunev
2012-04-20 09:55:50 +00:00
case IFT_ETHER:
case IFT_L2VLAN:
error = ifpromisc(ifs, 1);
break;
Reorder the bridge add and delete routines to avoid calling ifpromisc() with the bridge lock held.
2009-05-01 19:46:42 +00:00
}
- Allow ND6_IFF_AUTO_LINKLOCAL for IFT_BRIDGE. An interface with IFT_BRIDGE is initialized with !ND6_IFF_AUTO_LINKLOCAL && !ND6_IFF_ACCEPT_RTADV regardless of net.inet6.ip6.accept_rtadv and net.inet6.ip6.auto_linklocal. To configure an autoconfigured link-local address (RFC 4862), the following rc.conf(5) configuration can be used: ifconfig_bridge0_ipv6="inet6 auto_linklocal" - if_bridge(4) now removes IPv6 addresses on a member interface to be added when the parent interface or one of the existing member interfaces has an IPv6 address. if_bridge(4) merges each link-local scope zone which the member interfaces form respectively, so it causes address scope violation. Removal of the IPv6 addresses prevents it. - if_lagg(4) now removes IPv6 addresses on a member interfaces unconditionally. - Set reasonable flags to non-IPv6-capable interfaces. [*] Submitted by: rpaulo [*] MFC after: 1 week
2013-07-02 16:58:15 +00:00
Fix panic when adding vtnet interfaces to a bridge vtnet interfaces are always in promiscuous mode (at least if the VIRTIO_NET_F_CTRL_RX feature is not negotiated with the host). if_promisc() on a vtnet interface returned ENOTSUP although it has IFF_PROMISC set. This confused the bridge code. Instead we now accept all enable/disable promiscuous commands (and always keep IFF_PROMISC set). There are also two issues with the if_bridge error handling. If if_promisc() fails it uses bridge_delete_member() to clean up. This tries to disable promiscuous mode on the interface. That runs into an assert, because promiscuous mode was never set in the first place. (That's the panic reported in PR 200210.) We can only unset promiscuous mode if the interface actually is promiscuous. This goes against the reference counting done by if_promisc(), but only the first/last if_promic() calls can actually fail, so this is safe. A second issue is a double free of bif. It's already freed by bridge_delete_member(). PR: 200210 Differential Revision: https://reviews.freebsd.org/D2804 Reviewed by: philip (mentor)
2015-06-13 19:39:21 +00:00
if (error)
- Allow ND6_IFF_AUTO_LINKLOCAL for IFT_BRIDGE. An interface with IFT_BRIDGE is initialized with !ND6_IFF_AUTO_LINKLOCAL && !ND6_IFF_ACCEPT_RTADV regardless of net.inet6.ip6.accept_rtadv and net.inet6.ip6.auto_linklocal. To configure an autoconfigured link-local address (RFC 4862), the following rc.conf(5) configuration can be used: ifconfig_bridge0_ipv6="inet6 auto_linklocal" - if_bridge(4) now removes IPv6 addresses on a member interface to be added when the parent interface or one of the existing member interfaces has an IPv6 address. if_bridge(4) merges each link-local scope zone which the member interfaces form respectively, so it causes address scope violation. Removal of the IPv6 addresses prevents it. - if_lagg(4) now removes IPv6 addresses on a member interfaces unconditionally. - Set reasonable flags to non-IPv6-capable interfaces. [*] Submitted by: rpaulo [*] MFC after: 1 week
2013-07-02 16:58:15 +00:00
bridge_delete_member(sc, bif, 0);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
return (error);
}
Change most of the bridge and stp funtions to static. This has highlighted that the following funtions are not used, wrap in '#ifdef noused' for the moment. bstp_enable_change_detection bstp_disable_change_detection bstp_set_bridge_priority bstp_set_port_priority bstp_set_path_cost
2005-10-14 10:38:12 +00:00
static int
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge_ioctl_del(struct bridge_softc *sc, void *arg)
{
struct ifbreq *req = arg;
struct bridge_iflist *bif;
bif = bridge_lookup_member(sc, req->ifbr_ifsname);
if (bif == NULL)
return (ENOENT);
If we have been called from ether_ifdetach() then do not try and clear the promisc flag from the member interface, this is a no-op anyway since the interface is disappearing. The driver may have already released its resources such as miibus and this is likely to panic the kernel. Submitted and tested by: Wojciech A. Koszek MFC after: 2 weeks
2005-10-23 22:30:07 +00:00
bridge_delete_member(sc, bif, 0);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
return (0);
}
Change most of the bridge and stp funtions to static. This has highlighted that the following funtions are not used, wrap in '#ifdef noused' for the moment. bstp_enable_change_detection bstp_disable_change_detection bstp_set_bridge_priority bstp_set_port_priority bstp_set_path_cost
2005-10-14 10:38:12 +00:00
static int
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge_ioctl_gifflags(struct bridge_softc *sc, void *arg)
{
struct ifbreq *req = arg;
struct bridge_iflist *bif;
Bring in support for the Rapid Spanning Tree Protocol (802.1w). RSTP provides faster spanning tree convergence, the protocol will exchange information with neighboring switches to quickly transition to forwarding without creating loops. The code will default to RSTP mode but will downgrade any port connected to a legacy STP network so is fully backward compatible. Reviewed by: syrinx Tested by: syrinx
2006-11-01 09:07:47 +00:00
struct bstp_port *bp;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bif = bridge_lookup_member(sc, req->ifbr_ifsname);
if (bif == NULL)
return (ENOENT);
Bring in support for the Rapid Spanning Tree Protocol (802.1w). RSTP provides faster spanning tree convergence, the protocol will exchange information with neighboring switches to quickly transition to forwarding without creating loops. The code will default to RSTP mode but will downgrade any port connected to a legacy STP network so is fully backward compatible. Reviewed by: syrinx Tested by: syrinx
2006-11-01 09:07:47 +00:00
bp = &bif->bif_stp;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
req->ifbr_ifsflags = bif->bif_flags;
Bring in support for the Rapid Spanning Tree Protocol (802.1w). RSTP provides faster spanning tree convergence, the protocol will exchange information with neighboring switches to quickly transition to forwarding without creating loops. The code will default to RSTP mode but will downgrade any port connected to a legacy STP network so is fully backward compatible. Reviewed by: syrinx Tested by: syrinx
2006-11-01 09:07:47 +00:00
req->ifbr_state = bp->bp_state;
req->ifbr_priority = bp->bp_priority;
req->ifbr_path_cost = bp->bp_path_cost;
req->ifbr_portno = bif->bif_ifp->if_index & 0xfff;
req->ifbr_proto = bp->bp_protover;
req->ifbr_role = bp->bp_role;
req->ifbr_stpflags = bp->bp_flags;
Add an option to limit the number of source MACs that can be behind a bridge interface. Once the limit is reached packets with unknown source addresses are dropped until an existing host cache entry expires or is removed. Useful to use with the STICKY cache option. Sponsored by: miniSuperHappyDevHouse NZ
2007-11-04 08:32:27 +00:00
req->ifbr_addrcnt = bif->bif_addrcnt;
req->ifbr_addrmax = bif->bif_addrmax;
req->ifbr_addrexceeded = bif->bif_addrexceeded;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
Sync with the OpenBSD port of RSTP - use flags rather than sperate ioctls for edge, p2p - implement p2p and autop2p flags - define large pathcost constant as ULL - show bridgeid and rootid in ifconfig Obtained from: Reyk Floeter <reyk@openbsd.org>
2006-11-27 00:35:40 +00:00
/* Copy STP state options as flags */
if (bp->bp_operedge)
req->ifbr_ifsflags |= IFBIF_BSTP_EDGE;
if (bp->bp_flags & BSTP_PORT_AUTOEDGE)
req->ifbr_ifsflags |= IFBIF_BSTP_AUTOEDGE;
These days P2P means peer-2-peer (also well known from serveral filesharing protocols) while PointToPoint has been PtP links. Change the variables accordingly while the code is still fresh and undocumented. Requested by: bz
2006-12-11 23:46:40 +00:00
if (bp->bp_ptp_link)
req->ifbr_ifsflags |= IFBIF_BSTP_PTP;
if (bp->bp_flags & BSTP_PORT_AUTOPTP)
req->ifbr_ifsflags |= IFBIF_BSTP_AUTOPTP;
Add two new flags to if_bridge(4) indicating whether the edge flag of the bridge port and path cost have been administratively set or calculated automatically by RSTP. Make sure to transition from non-edge to edge when the port goes down and the edge flag was manually set before. This is needed to comply with the condition ((!portEnabled && AdminEdge) || ....) in the Bridge Detection State Machine (IEE802.1D-2004, p. 171). Reviewed by: thompsa Approved by: bz (mentor)
2006-12-04 14:45:02 +00:00
if (bp->bp_flags & BSTP_PORT_ADMEDGE)
req->ifbr_ifsflags |= IFBIF_BSTP_ADMEDGE;
if (bp->bp_flags & BSTP_PORT_ADMCOST)
req->ifbr_ifsflags |= IFBIF_BSTP_ADMCOST;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
return (0);
}
Change most of the bridge and stp funtions to static. This has highlighted that the following funtions are not used, wrap in '#ifdef noused' for the moment. bstp_enable_change_detection bstp_disable_change_detection bstp_set_bridge_priority bstp_set_port_priority bstp_set_path_cost
2005-10-14 10:38:12 +00:00
static int
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge_ioctl_sifflags(struct bridge_softc *sc, void *arg)
{
bridge/stp: Ensure we enter NET_EPOCH whenever we can send traffic Reviewed by: donner@ MFC after: 1 week Sponsored by: Orange Business Services Differential Revision: https://reviews.freebsd.org/D28858
2021-02-21 20:18:46 +00:00
struct epoch_tracker et;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
struct ifbreq *req = arg;
struct bridge_iflist *bif;
Sync with the OpenBSD port of RSTP - use flags rather than sperate ioctls for edge, p2p - implement p2p and autop2p flags - define large pathcost constant as ULL - show bridgeid and rootid in ifconfig Obtained from: Reyk Floeter <reyk@openbsd.org>
2006-11-27 00:35:40 +00:00
struct bstp_port *bp;
/tmp/cvsuusTrc
2006-07-26 10:43:02 +00:00
int error;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bif = bridge_lookup_member(sc, req->ifbr_ifsname);
if (bif == NULL)
return (ENOENT);
Sync with the OpenBSD port of RSTP - use flags rather than sperate ioctls for edge, p2p - implement p2p and autop2p flags - define large pathcost constant as ULL - show bridgeid and rootid in ifconfig Obtained from: Reyk Floeter <reyk@openbsd.org>
2006-11-27 00:35:40 +00:00
bp = &bif->bif_stp;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
Add support for creating span ports so that one can snoop bridged traffic from another interface/machine/network. Obtained from: OpenBSD MFC after: 2 weeks
2005-12-14 02:52:13 +00:00
if (req->ifbr_ifsflags & IFBIF_SPAN)
/* SPAN is readonly */
return (EINVAL);
bridge/stp: Ensure we enter NET_EPOCH whenever we can send traffic Reviewed by: donner@ MFC after: 1 week Sponsored by: Orange Business Services Differential Revision: https://reviews.freebsd.org/D28858
2021-02-21 20:18:46 +00:00
NET_EPOCH_ENTER(et);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
if (req->ifbr_ifsflags & IFBIF_STP) {
/tmp/cvsuusTrc
2006-07-26 10:43:02 +00:00
if ((bif->bif_flags & IFBIF_STP) == 0) {
use two stage creation of stp ports, this means that the stp variables can be set before the port is marked STP and they will no longer be overwrittten
2006-11-26 18:43:48 +00:00
error = bstp_enable(&bif->bif_stp);
bridge/stp: Ensure we enter NET_EPOCH whenever we can send traffic Reviewed by: donner@ MFC after: 1 week Sponsored by: Orange Business Services Differential Revision: https://reviews.freebsd.org/D28858
2021-02-21 20:18:46 +00:00
if (error) {
NET_EPOCH_EXIT(et);
/tmp/cvsuusTrc
2006-07-26 10:43:02 +00:00
return (error);
bridge/stp: Ensure we enter NET_EPOCH whenever we can send traffic Reviewed by: donner@ MFC after: 1 week Sponsored by: Orange Business Services Differential Revision: https://reviews.freebsd.org/D28858
2021-02-21 20:18:46 +00:00
}
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
}
/tmp/cvsuusTrc
2006-07-26 10:43:02 +00:00
} else {
if ((bif->bif_flags & IFBIF_STP) != 0)
use two stage creation of stp ports, this means that the stp variables can be set before the port is marked STP and they will no longer be overwrittten
2006-11-26 18:43:48 +00:00
bstp_disable(&bif->bif_stp);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
}
Sync with the OpenBSD port of RSTP - use flags rather than sperate ioctls for edge, p2p - implement p2p and autop2p flags - define large pathcost constant as ULL - show bridgeid and rootid in ifconfig Obtained from: Reyk Floeter <reyk@openbsd.org>
2006-11-27 00:35:40 +00:00
/* Pass on STP flags */
bstp_set_edge(bp, req->ifbr_ifsflags & IFBIF_BSTP_EDGE ? 1 : 0);
bstp_set_autoedge(bp, req->ifbr_ifsflags & IFBIF_BSTP_AUTOEDGE ? 1 : 0);
These days P2P means peer-2-peer (also well known from serveral filesharing protocols) while PointToPoint has been PtP links. Change the variables accordingly while the code is still fresh and undocumented. Requested by: bz
2006-12-11 23:46:40 +00:00
bstp_set_ptp(bp, req->ifbr_ifsflags & IFBIF_BSTP_PTP ? 1 : 0);
bstp_set_autoptp(bp, req->ifbr_ifsflags & IFBIF_BSTP_AUTOPTP ? 1 : 0);
Sync with the OpenBSD port of RSTP - use flags rather than sperate ioctls for edge, p2p - implement p2p and autop2p flags - define large pathcost constant as ULL - show bridgeid and rootid in ifconfig Obtained from: Reyk Floeter <reyk@openbsd.org>
2006-11-27 00:35:40 +00:00
/* Save the bits relating to the bridge */
bif->bif_flags = req->ifbr_ifsflags & IFBIFMASK;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge/stp: Ensure we enter NET_EPOCH whenever we can send traffic Reviewed by: donner@ MFC after: 1 week Sponsored by: Orange Business Services Differential Revision: https://reviews.freebsd.org/D28858
2021-02-21 20:18:46 +00:00
NET_EPOCH_EXIT(et);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
return (0);
}
Change most of the bridge and stp funtions to static. This has highlighted that the following funtions are not used, wrap in '#ifdef noused' for the moment. bstp_enable_change_detection bstp_disable_change_detection bstp_set_bridge_priority bstp_set_port_priority bstp_set_path_cost
2005-10-14 10:38:12 +00:00
static int
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge_ioctl_scache(struct bridge_softc *sc, void *arg)
{
struct ifbrparam *param = arg;
sc->sc_brtmax = param->ifbrp_csize;
bridge_rttrim(sc);
return (0);
}
Change most of the bridge and stp funtions to static. This has highlighted that the following funtions are not used, wrap in '#ifdef noused' for the moment. bstp_enable_change_detection bstp_disable_change_detection bstp_set_bridge_priority bstp_set_port_priority bstp_set_path_cost
2005-10-14 10:38:12 +00:00
static int
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge_ioctl_gcache(struct bridge_softc *sc, void *arg)
{
struct ifbrparam *param = arg;
param->ifbrp_csize = sc->sc_brtmax;
return (0);
}
Change most of the bridge and stp funtions to static. This has highlighted that the following funtions are not used, wrap in '#ifdef noused' for the moment. bstp_enable_change_detection bstp_disable_change_detection bstp_set_bridge_priority bstp_set_port_priority bstp_set_path_cost
2005-10-14 10:38:12 +00:00
static int
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge_ioctl_gifs(struct bridge_softc *sc, void *arg)
{
struct ifbifconf *bifc = arg;
struct bridge_iflist *bif;
struct ifbreq breq;
Avoid holding the softc lock when using copyout(). Reported by: dfr Approved by: re (rwatson)
2007-07-26 20:30:18 +00:00
char *buf, *outbuf;
int count, buflen, len, error = 0;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
count = 0;
bridge: Change lists to CK_LIST as a peparation for epochification Prepare the ground for a rework of the bridge locking approach. We will use an epoch-based approach in the datapath and making it safe to iterate over the interface, span and rtnode lists without holding the BRIDGE_LOCK. Replace the relevant lists by their ConcurrencyKit equivalents. No functional change in this commit. Reviewed by: emaste, ae, philip (previous version) Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24249
2020-04-05 17:15:20 +00:00
CK_LIST_FOREACH(bif, &sc->sc_iflist, bif_next)
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
count++;
bridge: Change lists to CK_LIST as a peparation for epochification Prepare the ground for a rework of the bridge locking approach. We will use an epoch-based approach in the datapath and making it safe to iterate over the interface, span and rtnode lists without holding the BRIDGE_LOCK. Replace the relevant lists by their ConcurrencyKit equivalents. No functional change in this commit. Reviewed by: emaste, ae, philip (previous version) Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24249
2020-04-05 17:15:20 +00:00
CK_LIST_FOREACH(bif, &sc->sc_spanlist, bif_next)
Add support for creating span ports so that one can snoop bridged traffic from another interface/machine/network. Obtained from: OpenBSD MFC after: 2 weeks
2005-12-14 02:52:13 +00:00
count++;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
Avoid holding the softc lock when using copyout(). Reported by: dfr Approved by: re (rwatson)
2007-07-26 20:30:18 +00:00
buflen = sizeof(breq) * count;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
if (bifc->ifbic_len == 0) {
Avoid holding the softc lock when using copyout(). Reported by: dfr Approved by: re (rwatson)
2007-07-26 20:30:18 +00:00
bifc->ifbic_len = buflen;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
return (0);
}
bridge: Don't sleep during epoch While it doesn't trigger INVARIANTS or WITNESS on head it does in stable/12. There's also no reason for it, as we can easily report the out of memory error to the caller (i.e. userspace). All of these can already fail. PR: 248046 MFC after: 3 days
2020-07-18 12:43:11 +00:00
outbuf = malloc(buflen, M_TEMP, M_NOWAIT | M_ZERO);
if (outbuf == NULL)
return (ENOMEM);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
count = 0;
Avoid holding the softc lock when using copyout(). Reported by: dfr Approved by: re (rwatson)
2007-07-26 20:30:18 +00:00
buf = outbuf;
len = min(bifc->ifbic_len, buflen);
Fix style(9) nits, whitespace and parentheses.
2006-05-16 22:50:41 +00:00
bzero(&breq, sizeof(breq));
bridge: Change lists to CK_LIST as a peparation for epochification Prepare the ground for a rework of the bridge locking approach. We will use an epoch-based approach in the datapath and making it safe to iterate over the interface, span and rtnode lists without holding the BRIDGE_LOCK. Replace the relevant lists by their ConcurrencyKit equivalents. No functional change in this commit. Reviewed by: emaste, ae, philip (previous version) Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24249
2020-04-05 17:15:20 +00:00
CK_LIST_FOREACH(bif, &sc->sc_iflist, bif_next) {
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
if (len < sizeof(breq))
break;
strlcpy(breq.ifbr_ifsname, bif->bif_ifp->if_xname,
sizeof(breq.ifbr_ifsname));
Sync with the OpenBSD port of RSTP - use flags rather than sperate ioctls for edge, p2p - implement p2p and autop2p flags - define large pathcost constant as ULL - show bridgeid and rootid in ifconfig Obtained from: Reyk Floeter <reyk@openbsd.org>
2006-11-27 00:35:40 +00:00
/* Fill in the ifbreq structure */
error = bridge_ioctl_gifflags(sc, &breq);
if (error)
break;
Avoid holding the softc lock when using copyout(). Reported by: dfr Approved by: re (rwatson)
2007-07-26 20:30:18 +00:00
memcpy(buf, &breq, sizeof(breq));
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
count++;
Avoid holding the softc lock when using copyout(). Reported by: dfr Approved by: re (rwatson)
2007-07-26 20:30:18 +00:00
buf += sizeof(breq);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
len -= sizeof(breq);
}
bridge: Change lists to CK_LIST as a peparation for epochification Prepare the ground for a rework of the bridge locking approach. We will use an epoch-based approach in the datapath and making it safe to iterate over the interface, span and rtnode lists without holding the BRIDGE_LOCK. Replace the relevant lists by their ConcurrencyKit equivalents. No functional change in this commit. Reviewed by: emaste, ae, philip (previous version) Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24249
2020-04-05 17:15:20 +00:00
CK_LIST_FOREACH(bif, &sc->sc_spanlist, bif_next) {
Add support for creating span ports so that one can snoop bridged traffic from another interface/machine/network. Obtained from: OpenBSD MFC after: 2 weeks
2005-12-14 02:52:13 +00:00
if (len < sizeof(breq))
break;
strlcpy(breq.ifbr_ifsname, bif->bif_ifp->if_xname,
sizeof(breq.ifbr_ifsname));
breq.ifbr_ifsflags = bif->bif_flags;
Bring in support for the Rapid Spanning Tree Protocol (802.1w). RSTP provides faster spanning tree convergence, the protocol will exchange information with neighboring switches to quickly transition to forwarding without creating loops. The code will default to RSTP mode but will downgrade any port connected to a legacy STP network so is fully backward compatible. Reviewed by: syrinx Tested by: syrinx
2006-11-01 09:07:47 +00:00
breq.ifbr_portno = bif->bif_ifp->if_index & 0xfff;
Avoid holding the softc lock when using copyout(). Reported by: dfr Approved by: re (rwatson)
2007-07-26 20:30:18 +00:00
memcpy(buf, &breq, sizeof(breq));
Add support for creating span ports so that one can snoop bridged traffic from another interface/machine/network. Obtained from: OpenBSD MFC after: 2 weeks
2005-12-14 02:52:13 +00:00
count++;
Avoid holding the softc lock when using copyout(). Reported by: dfr Approved by: re (rwatson)
2007-07-26 20:30:18 +00:00
buf += sizeof(breq);
Add support for creating span ports so that one can snoop bridged traffic from another interface/machine/network. Obtained from: OpenBSD MFC after: 2 weeks
2005-12-14 02:52:13 +00:00
len -= sizeof(breq);
}
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bifc->ifbic_len = sizeof(breq) * count;
Avoid holding the softc lock when using copyout(). Reported by: dfr Approved by: re (rwatson)
2007-07-26 20:30:18 +00:00
error = copyout(outbuf, bifc->ifbic_req, bifc->ifbic_len);
free(outbuf, M_TEMP);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
return (error);
}
Change most of the bridge and stp funtions to static. This has highlighted that the following funtions are not used, wrap in '#ifdef noused' for the moment. bstp_enable_change_detection bstp_disable_change_detection bstp_set_bridge_priority bstp_set_port_priority bstp_set_path_cost
2005-10-14 10:38:12 +00:00
static int
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge_ioctl_rts(struct bridge_softc *sc, void *arg)
{
struct ifbaconf *bac = arg;
struct bridge_rtnode *brt;
struct ifbareq bareq;
Avoid holding the softc lock when using copyout(). Reported by: dfr Approved by: re (rwatson)
2007-07-26 20:30:18 +00:00
char *buf, *outbuf;
int count, buflen, len, error = 0;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
if (bac->ifbac_len == 0)
return (0);
Avoid holding the softc lock when using copyout(). Reported by: dfr Approved by: re (rwatson)
2007-07-26 20:30:18 +00:00
count = 0;
bridge: Change lists to CK_LIST as a peparation for epochification Prepare the ground for a rework of the bridge locking approach. We will use an epoch-based approach in the datapath and making it safe to iterate over the interface, span and rtnode lists without holding the BRIDGE_LOCK. Replace the relevant lists by their ConcurrencyKit equivalents. No functional change in this commit. Reviewed by: emaste, ae, philip (previous version) Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24249
2020-04-05 17:15:20 +00:00
CK_LIST_FOREACH(brt, &sc->sc_rtlist, brt_list)
Avoid holding the softc lock when using copyout(). Reported by: dfr Approved by: re (rwatson)
2007-07-26 20:30:18 +00:00
count++;
buflen = sizeof(bareq) * count;
bridge: Don't sleep during epoch While it doesn't trigger INVARIANTS or WITNESS on head it does in stable/12. There's also no reason for it, as we can easily report the out of memory error to the caller (i.e. userspace). All of these can already fail. PR: 248046 MFC after: 3 days
2020-07-18 12:43:11 +00:00
outbuf = malloc(buflen, M_TEMP, M_NOWAIT | M_ZERO);
if (outbuf == NULL)
return (ENOMEM);
Avoid holding the softc lock when using copyout(). Reported by: dfr Approved by: re (rwatson)
2007-07-26 20:30:18 +00:00
count = 0;
buf = outbuf;
len = min(bac->ifbac_len, buflen);
Fix style(9) nits, whitespace and parentheses.
2006-05-16 22:50:41 +00:00
bzero(&bareq, sizeof(bareq));
bridge: Change lists to CK_LIST as a peparation for epochification Prepare the ground for a rework of the bridge locking approach. We will use an epoch-based approach in the datapath and making it safe to iterate over the interface, span and rtnode lists without holding the BRIDGE_LOCK. Replace the relevant lists by their ConcurrencyKit equivalents. No functional change in this commit. Reviewed by: emaste, ae, philip (previous version) Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24249
2020-04-05 17:15:20 +00:00
CK_LIST_FOREACH(brt, &sc->sc_rtlist, brt_list) {
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
if (len < sizeof(bareq))
goto out;
strlcpy(bareq.ifba_ifsname, brt->brt_ifp->if_xname,
sizeof(bareq.ifba_ifsname));
memcpy(bareq.ifba_dst, brt->brt_addr, sizeof(brt->brt_addr));
Add the vlan tag to the bridge route table. This allows a vlan trunk to be bridged, previously legitimate traffic was not passed as the bridge could not tell that it was on a different Ethernet segment. All non-tagged traffic is treated as vlan1 as per IEEE 802.1Q-2003
2007-06-13 18:58:04 +00:00
bareq.ifba_vlan = brt->brt_vlan;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
if ((brt->brt_flags & IFBAF_TYPEMASK) == IFBAF_DYNAMIC &&
It is better to use time_uptime here since it is monotonic. Pointed out by: glebius
2006-01-02 22:23:03 +00:00
time_uptime < brt->brt_expire)
bareq.ifba_expire = brt->brt_expire - time_uptime;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
else
bareq.ifba_expire = 0;
bareq.ifba_flags = brt->brt_flags;
Avoid holding the softc lock when using copyout(). Reported by: dfr Approved by: re (rwatson)
2007-07-26 20:30:18 +00:00
memcpy(buf, &bareq, sizeof(bareq));
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
count++;
Avoid holding the softc lock when using copyout(). Reported by: dfr Approved by: re (rwatson)
2007-07-26 20:30:18 +00:00
buf += sizeof(bareq);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
len -= sizeof(bareq);
}
Minor whitespace cleanup.
2005-12-17 10:03:48 +00:00
out:
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bac->ifbac_len = sizeof(bareq) * count;
Avoid holding the softc lock when using copyout(). Reported by: dfr Approved by: re (rwatson)
2007-07-26 20:30:18 +00:00
error = copyout(outbuf, bac->ifbac_req, bac->ifbac_len);
free(outbuf, M_TEMP);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
return (error);
}
Change most of the bridge and stp funtions to static. This has highlighted that the following funtions are not used, wrap in '#ifdef noused' for the moment. bstp_enable_change_detection bstp_disable_change_detection bstp_set_bridge_priority bstp_set_port_priority bstp_set_path_cost
2005-10-14 10:38:12 +00:00
static int
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge_ioctl_saddr(struct bridge_softc *sc, void *arg)
{
struct ifbareq *req = arg;
struct bridge_iflist *bif;
bridge: call member interface ioctl() without NET_EPOCH We're not allowed to hold NET_EPOCH while sleeping, so when we call ioctl() handlers for member interfaces we cannot be in NET_EPOCH. We still need some protection of our CK_LISTs, so hold BRIDGE_LOCK instead. That requires changing BRIDGE_LOCK into a sleepable lock, and separating the BRIDGE_RT_LOCK, to protect bridge_rtnode lists. That lock is taken in the data path (while in NET_EPOCH), so it cannot be a sleepable lock. While here document the locking strategy. MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D26418
2020-10-06 19:19:56 +00:00
struct epoch_tracker et;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
int error;
bridge: call member interface ioctl() without NET_EPOCH We're not allowed to hold NET_EPOCH while sleeping, so when we call ioctl() handlers for member interfaces we cannot be in NET_EPOCH. We still need some protection of our CK_LISTs, so hold BRIDGE_LOCK instead. That requires changing BRIDGE_LOCK into a sleepable lock, and separating the BRIDGE_RT_LOCK, to protect bridge_rtnode lists. That lock is taken in the data path (while in NET_EPOCH), so it cannot be a sleepable lock. While here document the locking strategy. MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D26418
2020-10-06 19:19:56 +00:00
NET_EPOCH_ENTER(et);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bif = bridge_lookup_member(sc, req->ifba_ifsname);
bridge: call member interface ioctl() without NET_EPOCH We're not allowed to hold NET_EPOCH while sleeping, so when we call ioctl() handlers for member interfaces we cannot be in NET_EPOCH. We still need some protection of our CK_LISTs, so hold BRIDGE_LOCK instead. That requires changing BRIDGE_LOCK into a sleepable lock, and separating the BRIDGE_RT_LOCK, to protect bridge_rtnode lists. That lock is taken in the data path (while in NET_EPOCH), so it cannot be a sleepable lock. While here document the locking strategy. MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D26418
2020-10-06 19:19:56 +00:00
if (bif == NULL) {
NET_EPOCH_EXIT(et);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
return (ENOENT);
bridge: call member interface ioctl() without NET_EPOCH We're not allowed to hold NET_EPOCH while sleeping, so when we call ioctl() handlers for member interfaces we cannot be in NET_EPOCH. We still need some protection of our CK_LISTs, so hold BRIDGE_LOCK instead. That requires changing BRIDGE_LOCK into a sleepable lock, and separating the BRIDGE_RT_LOCK, to protect bridge_rtnode lists. That lock is taken in the data path (while in NET_EPOCH), so it cannot be a sleepable lock. While here document the locking strategy. MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D26418
2020-10-06 19:19:56 +00:00
}
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge: epoch-ification Run the bridge datapath under epoch, rather than under the BRIDGE_LOCK(). We still take the BRIDGE_LOCK() whenever we insert or delete items in the relevant lists, but we use epoch callbacks to free items so that it's safe to iterate the lists without the BRIDGE_LOCK. Tests on mercat5/6 shows this increases bridge throughput significantly, from 3.7Mpps to 18.6Mpps. Reviewed by: emaste, philip, melifaro MFC after: 2 months Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24250
2020-04-26 16:22:35 +00:00
/* bridge_rtupdate() may acquire the lock. */
Add the vlan tag to the bridge route table. This allows a vlan trunk to be bridged, previously legitimate traffic was not passed as the bridge could not tell that it was on a different Ethernet segment. All non-tagged traffic is treated as vlan1 as per IEEE 802.1Q-2003
2007-06-13 18:58:04 +00:00
error = bridge_rtupdate(sc, req->ifba_dst, req->ifba_vlan, bif, 1,
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
req->ifba_flags);
bridge: call member interface ioctl() without NET_EPOCH We're not allowed to hold NET_EPOCH while sleeping, so when we call ioctl() handlers for member interfaces we cannot be in NET_EPOCH. We still need some protection of our CK_LISTs, so hold BRIDGE_LOCK instead. That requires changing BRIDGE_LOCK into a sleepable lock, and separating the BRIDGE_RT_LOCK, to protect bridge_rtnode lists. That lock is taken in the data path (while in NET_EPOCH), so it cannot be a sleepable lock. While here document the locking strategy. MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D26418
2020-10-06 19:19:56 +00:00
NET_EPOCH_EXIT(et);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
return (error);
}
Change most of the bridge and stp funtions to static. This has highlighted that the following funtions are not used, wrap in '#ifdef noused' for the moment. bstp_enable_change_detection bstp_disable_change_detection bstp_set_bridge_priority bstp_set_port_priority bstp_set_path_cost
2005-10-14 10:38:12 +00:00
static int
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge_ioctl_sto(struct bridge_softc *sc, void *arg)
{
struct ifbrparam *param = arg;
sc->sc_brttimeout = param->ifbrp_ctime;
return (0);
}
Change most of the bridge and stp funtions to static. This has highlighted that the following funtions are not used, wrap in '#ifdef noused' for the moment. bstp_enable_change_detection bstp_disable_change_detection bstp_set_bridge_priority bstp_set_port_priority bstp_set_path_cost
2005-10-14 10:38:12 +00:00
static int
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge_ioctl_gto(struct bridge_softc *sc, void *arg)
{
struct ifbrparam *param = arg;
param->ifbrp_ctime = sc->sc_brttimeout;
return (0);
}
Change most of the bridge and stp funtions to static. This has highlighted that the following funtions are not used, wrap in '#ifdef noused' for the moment. bstp_enable_change_detection bstp_disable_change_detection bstp_set_bridge_priority bstp_set_port_priority bstp_set_path_cost
2005-10-14 10:38:12 +00:00
static int
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge_ioctl_daddr(struct bridge_softc *sc, void *arg)
{
struct ifbareq *req = arg;
Add the vlan tag to the bridge route table. This allows a vlan trunk to be bridged, previously legitimate traffic was not passed as the bridge could not tell that it was on a different Ethernet segment. All non-tagged traffic is treated as vlan1 as per IEEE 802.1Q-2003
2007-06-13 18:58:04 +00:00
return (bridge_rtdaddr(sc, req->ifba_dst, req->ifba_vlan));
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
}
Change most of the bridge and stp funtions to static. This has highlighted that the following funtions are not used, wrap in '#ifdef noused' for the moment. bstp_enable_change_detection bstp_disable_change_detection bstp_set_bridge_priority bstp_set_port_priority bstp_set_path_cost
2005-10-14 10:38:12 +00:00
static int
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge_ioctl_flush(struct bridge_softc *sc, void *arg)
{
struct ifbreq *req = arg;
bridge: call member interface ioctl() without NET_EPOCH We're not allowed to hold NET_EPOCH while sleeping, so when we call ioctl() handlers for member interfaces we cannot be in NET_EPOCH. We still need some protection of our CK_LISTs, so hold BRIDGE_LOCK instead. That requires changing BRIDGE_LOCK into a sleepable lock, and separating the BRIDGE_RT_LOCK, to protect bridge_rtnode lists. That lock is taken in the data path (while in NET_EPOCH), so it cannot be a sleepable lock. While here document the locking strategy. MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D26418
2020-10-06 19:19:56 +00:00
BRIDGE_RT_LOCK(sc);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge_rtflush(sc, req->ifbr_ifsflags);
bridge: call member interface ioctl() without NET_EPOCH We're not allowed to hold NET_EPOCH while sleeping, so when we call ioctl() handlers for member interfaces we cannot be in NET_EPOCH. We still need some protection of our CK_LISTs, so hold BRIDGE_LOCK instead. That requires changing BRIDGE_LOCK into a sleepable lock, and separating the BRIDGE_RT_LOCK, to protect bridge_rtnode lists. That lock is taken in the data path (while in NET_EPOCH), so it cannot be a sleepable lock. While here document the locking strategy. MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D26418
2020-10-06 19:19:56 +00:00
BRIDGE_RT_UNLOCK(sc);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
return (0);
}
Change most of the bridge and stp funtions to static. This has highlighted that the following funtions are not used, wrap in '#ifdef noused' for the moment. bstp_enable_change_detection bstp_disable_change_detection bstp_set_bridge_priority bstp_set_port_priority bstp_set_path_cost
2005-10-14 10:38:12 +00:00
static int
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge_ioctl_gpri(struct bridge_softc *sc, void *arg)
{
struct ifbrparam *param = arg;
/tmp/cvsuusTrc
2006-07-26 10:43:02 +00:00
struct bstp_state *bs = &sc->sc_stp;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
/tmp/cvsuusTrc
2006-07-26 10:43:02 +00:00
param->ifbrp_prio = bs->bs_bridge_priority;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
return (0);
}
Change most of the bridge and stp funtions to static. This has highlighted that the following funtions are not used, wrap in '#ifdef noused' for the moment. bstp_enable_change_detection bstp_disable_change_detection bstp_set_bridge_priority bstp_set_port_priority bstp_set_path_cost
2005-10-14 10:38:12 +00:00
static int
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge_ioctl_spri(struct bridge_softc *sc, void *arg)
{
struct ifbrparam *param = arg;
Bring in support for the Rapid Spanning Tree Protocol (802.1w). RSTP provides faster spanning tree convergence, the protocol will exchange information with neighboring switches to quickly transition to forwarding without creating loops. The code will default to RSTP mode but will downgrade any port connected to a legacy STP network so is fully backward compatible. Reviewed by: syrinx Tested by: syrinx
2006-11-01 09:07:47 +00:00
return (bstp_set_priority(&sc->sc_stp, param->ifbrp_prio));
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
}
Change most of the bridge and stp funtions to static. This has highlighted that the following funtions are not used, wrap in '#ifdef noused' for the moment. bstp_enable_change_detection bstp_disable_change_detection bstp_set_bridge_priority bstp_set_port_priority bstp_set_path_cost
2005-10-14 10:38:12 +00:00
static int
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge_ioctl_ght(struct bridge_softc *sc, void *arg)
{
struct ifbrparam *param = arg;
/tmp/cvsuusTrc
2006-07-26 10:43:02 +00:00
struct bstp_state *bs = &sc->sc_stp;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
Bring in support for the Rapid Spanning Tree Protocol (802.1w). RSTP provides faster spanning tree convergence, the protocol will exchange information with neighboring switches to quickly transition to forwarding without creating loops. The code will default to RSTP mode but will downgrade any port connected to a legacy STP network so is fully backward compatible. Reviewed by: syrinx Tested by: syrinx
2006-11-01 09:07:47 +00:00
param->ifbrp_hellotime = bs->bs_bridge_htime >> 8;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
return (0);
}
Change most of the bridge and stp funtions to static. This has highlighted that the following funtions are not used, wrap in '#ifdef noused' for the moment. bstp_enable_change_detection bstp_disable_change_detection bstp_set_bridge_priority bstp_set_port_priority bstp_set_path_cost
2005-10-14 10:38:12 +00:00
static int
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge_ioctl_sht(struct bridge_softc *sc, void *arg)
{
struct ifbrparam *param = arg;
Bring in support for the Rapid Spanning Tree Protocol (802.1w). RSTP provides faster spanning tree convergence, the protocol will exchange information with neighboring switches to quickly transition to forwarding without creating loops. The code will default to RSTP mode but will downgrade any port connected to a legacy STP network so is fully backward compatible. Reviewed by: syrinx Tested by: syrinx
2006-11-01 09:07:47 +00:00
return (bstp_set_htime(&sc->sc_stp, param->ifbrp_hellotime));
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
}
Change most of the bridge and stp funtions to static. This has highlighted that the following funtions are not used, wrap in '#ifdef noused' for the moment. bstp_enable_change_detection bstp_disable_change_detection bstp_set_bridge_priority bstp_set_port_priority bstp_set_path_cost
2005-10-14 10:38:12 +00:00
static int
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge_ioctl_gfd(struct bridge_softc *sc, void *arg)
{
struct ifbrparam *param = arg;
/tmp/cvsuusTrc
2006-07-26 10:43:02 +00:00
struct bstp_state *bs = &sc->sc_stp;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
Bring in support for the Rapid Spanning Tree Protocol (802.1w). RSTP provides faster spanning tree convergence, the protocol will exchange information with neighboring switches to quickly transition to forwarding without creating loops. The code will default to RSTP mode but will downgrade any port connected to a legacy STP network so is fully backward compatible. Reviewed by: syrinx Tested by: syrinx
2006-11-01 09:07:47 +00:00
param->ifbrp_fwddelay = bs->bs_bridge_fdelay >> 8;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
return (0);
}
Change most of the bridge and stp funtions to static. This has highlighted that the following funtions are not used, wrap in '#ifdef noused' for the moment. bstp_enable_change_detection bstp_disable_change_detection bstp_set_bridge_priority bstp_set_port_priority bstp_set_path_cost
2005-10-14 10:38:12 +00:00
static int
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge_ioctl_sfd(struct bridge_softc *sc, void *arg)
{
struct ifbrparam *param = arg;
Bring in support for the Rapid Spanning Tree Protocol (802.1w). RSTP provides faster spanning tree convergence, the protocol will exchange information with neighboring switches to quickly transition to forwarding without creating loops. The code will default to RSTP mode but will downgrade any port connected to a legacy STP network so is fully backward compatible. Reviewed by: syrinx Tested by: syrinx
2006-11-01 09:07:47 +00:00
return (bstp_set_fdelay(&sc->sc_stp, param->ifbrp_fwddelay));
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
}
Change most of the bridge and stp funtions to static. This has highlighted that the following funtions are not used, wrap in '#ifdef noused' for the moment. bstp_enable_change_detection bstp_disable_change_detection bstp_set_bridge_priority bstp_set_port_priority bstp_set_path_cost
2005-10-14 10:38:12 +00:00
static int
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge_ioctl_gma(struct bridge_softc *sc, void *arg)
{
struct ifbrparam *param = arg;
/tmp/cvsuusTrc
2006-07-26 10:43:02 +00:00
struct bstp_state *bs = &sc->sc_stp;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
/tmp/cvsuusTrc
2006-07-26 10:43:02 +00:00
param->ifbrp_maxage = bs->bs_bridge_max_age >> 8;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
return (0);
}
Change most of the bridge and stp funtions to static. This has highlighted that the following funtions are not used, wrap in '#ifdef noused' for the moment. bstp_enable_change_detection bstp_disable_change_detection bstp_set_bridge_priority bstp_set_port_priority bstp_set_path_cost
2005-10-14 10:38:12 +00:00
static int
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge_ioctl_sma(struct bridge_softc *sc, void *arg)
{
struct ifbrparam *param = arg;
Bring in support for the Rapid Spanning Tree Protocol (802.1w). RSTP provides faster spanning tree convergence, the protocol will exchange information with neighboring switches to quickly transition to forwarding without creating loops. The code will default to RSTP mode but will downgrade any port connected to a legacy STP network so is fully backward compatible. Reviewed by: syrinx Tested by: syrinx
2006-11-01 09:07:47 +00:00
return (bstp_set_maxage(&sc->sc_stp, param->ifbrp_maxage));
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
}
Change most of the bridge and stp funtions to static. This has highlighted that the following funtions are not used, wrap in '#ifdef noused' for the moment. bstp_enable_change_detection bstp_disable_change_detection bstp_set_bridge_priority bstp_set_port_priority bstp_set_path_cost
2005-10-14 10:38:12 +00:00
static int
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge_ioctl_sifprio(struct bridge_softc *sc, void *arg)
{
struct ifbreq *req = arg;
struct bridge_iflist *bif;
bif = bridge_lookup_member(sc, req->ifbr_ifsname);
if (bif == NULL)
return (ENOENT);
Bring in support for the Rapid Spanning Tree Protocol (802.1w). RSTP provides faster spanning tree convergence, the protocol will exchange information with neighboring switches to quickly transition to forwarding without creating loops. The code will default to RSTP mode but will downgrade any port connected to a legacy STP network so is fully backward compatible. Reviewed by: syrinx Tested by: syrinx
2006-11-01 09:07:47 +00:00
return (bstp_set_port_priority(&bif->bif_stp, req->ifbr_priority));
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
}
Change most of the bridge and stp funtions to static. This has highlighted that the following funtions are not used, wrap in '#ifdef noused' for the moment. bstp_enable_change_detection bstp_disable_change_detection bstp_set_bridge_priority bstp_set_port_priority bstp_set_path_cost
2005-10-14 10:38:12 +00:00
static int
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge_ioctl_sifcost(struct bridge_softc *sc, void *arg)
{
struct ifbreq *req = arg;
struct bridge_iflist *bif;
bif = bridge_lookup_member(sc, req->ifbr_ifsname);
if (bif == NULL)
return (ENOENT);
Bring in support for the Rapid Spanning Tree Protocol (802.1w). RSTP provides faster spanning tree convergence, the protocol will exchange information with neighboring switches to quickly transition to forwarding without creating loops. The code will default to RSTP mode but will downgrade any port connected to a legacy STP network so is fully backward compatible. Reviewed by: syrinx Tested by: syrinx
2006-11-01 09:07:47 +00:00
return (bstp_set_path_cost(&bif->bif_stp, req->ifbr_path_cost));
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
}
Add an option to limit the number of source MACs that can be behind a bridge interface. Once the limit is reached packets with unknown source addresses are dropped until an existing host cache entry expires or is removed. Useful to use with the STICKY cache option. Sponsored by: miniSuperHappyDevHouse NZ
2007-11-04 08:32:27 +00:00
static int
bridge_ioctl_sifmaxaddr(struct bridge_softc *sc, void *arg)
{
struct ifbreq *req = arg;
struct bridge_iflist *bif;
bif = bridge_lookup_member(sc, req->ifbr_ifsname);
if (bif == NULL)
return (ENOENT);
bif->bif_addrmax = req->ifbr_addrmax;
return (0);
}
Add support for creating span ports so that one can snoop bridged traffic from another interface/machine/network. Obtained from: OpenBSD MFC after: 2 weeks
2005-12-14 02:52:13 +00:00
static int
bridge_ioctl_addspan(struct bridge_softc *sc, void *arg)
{
struct ifbreq *req = arg;
struct bridge_iflist *bif = NULL;
struct ifnet *ifs;
ifs = ifunit(req->ifbr_ifsname);
if (ifs == NULL)
return (ENOENT);
bridge: Change lists to CK_LIST as a peparation for epochification Prepare the ground for a rework of the bridge locking approach. We will use an epoch-based approach in the datapath and making it safe to iterate over the interface, span and rtnode lists without holding the BRIDGE_LOCK. Replace the relevant lists by their ConcurrencyKit equivalents. No functional change in this commit. Reviewed by: emaste, ae, philip (previous version) Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24249
2020-04-05 17:15:20 +00:00
CK_LIST_FOREACH(bif, &sc->sc_spanlist, bif_next)
Add support for creating span ports so that one can snoop bridged traffic from another interface/machine/network. Obtained from: OpenBSD MFC after: 2 weeks
2005-12-14 02:52:13 +00:00
if (ifs == bif->bif_ifp)
return (EBUSY);
if (ifs->if_bridge != NULL)
return (EBUSY);
switch (ifs->if_type) {
case IFT_ETHER:
Allow gif interfaces to be added as span ports, the user may want to send a copy of all packets to the other side of the world.
2006-06-20 21:28:18 +00:00
case IFT_GIF:
Add support for creating span ports so that one can snoop bridged traffic from another interface/machine/network. Obtained from: OpenBSD MFC after: 2 weeks
2005-12-14 02:52:13 +00:00
case IFT_L2VLAN:
break;
default:
return (EINVAL);
}
bif = malloc(sizeof(*bif), M_DEVBUF, M_NOWAIT|M_ZERO);
if (bif == NULL)
return (ENOMEM);
bif->bif_ifp = ifs;
bif->bif_flags = IFBIF_SPAN;
bridge: Change lists to CK_LIST as a peparation for epochification Prepare the ground for a rework of the bridge locking approach. We will use an epoch-based approach in the datapath and making it safe to iterate over the interface, span and rtnode lists without holding the BRIDGE_LOCK. Replace the relevant lists by their ConcurrencyKit equivalents. No functional change in this commit. Reviewed by: emaste, ae, philip (previous version) Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24249
2020-04-05 17:15:20 +00:00
CK_LIST_INSERT_HEAD(&sc->sc_spanlist, bif, bif_next);
Add support for creating span ports so that one can snoop bridged traffic from another interface/machine/network. Obtained from: OpenBSD MFC after: 2 weeks
2005-12-14 02:52:13 +00:00
return (0);
}
static int
bridge_ioctl_delspan(struct bridge_softc *sc, void *arg)
{
struct ifbreq *req = arg;
struct bridge_iflist *bif;
struct ifnet *ifs;
ifs = ifunit(req->ifbr_ifsname);
if (ifs == NULL)
return (ENOENT);
bridge: Change lists to CK_LIST as a peparation for epochification Prepare the ground for a rework of the bridge locking approach. We will use an epoch-based approach in the datapath and making it safe to iterate over the interface, span and rtnode lists without holding the BRIDGE_LOCK. Replace the relevant lists by their ConcurrencyKit equivalents. No functional change in this commit. Reviewed by: emaste, ae, philip (previous version) Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24249
2020-04-05 17:15:20 +00:00
CK_LIST_FOREACH(bif, &sc->sc_spanlist, bif_next)
Add support for creating span ports so that one can snoop bridged traffic from another interface/machine/network. Obtained from: OpenBSD MFC after: 2 weeks
2005-12-14 02:52:13 +00:00
if (ifs == bif->bif_ifp)
break;
if (bif == NULL)
return (ENOENT);
Change from a callback in if_ethersubr to using EVENTHANDLER in order to detach span ports when they disappear. The span port does not have a pointer to the softc so revert r1.31 and bring back the softc linked-list. MFC after: 2 weeks
2005-12-17 06:33:51 +00:00
bridge_delete_span(sc, bif);
Add support for creating span ports so that one can snoop bridged traffic from another interface/machine/network. Obtained from: OpenBSD MFC after: 2 weeks
2005-12-14 02:52:13 +00:00
return (0);
}
Add some statistics that are needed to support RFC4188 as part of the SoC2006 work on a bridge monitoring module for BSNMP. Submitted by: shteryana (SoC 2006)
2006-07-31 20:24:46 +00:00
static int
bridge_ioctl_gbparam(struct bridge_softc *sc, void *arg)
{
struct ifbropreq *req = arg;
Bring in support for the Rapid Spanning Tree Protocol (802.1w). RSTP provides faster spanning tree convergence, the protocol will exchange information with neighboring switches to quickly transition to forwarding without creating loops. The code will default to RSTP mode but will downgrade any port connected to a legacy STP network so is fully backward compatible. Reviewed by: syrinx Tested by: syrinx
2006-11-01 09:07:47 +00:00
struct bstp_state *bs = &sc->sc_stp;
Add some statistics that are needed to support RFC4188 as part of the SoC2006 work on a bridge monitoring module for BSNMP. Submitted by: shteryana (SoC 2006)
2006-07-31 20:24:46 +00:00
struct bstp_port *root_port;
Bring in support for the Rapid Spanning Tree Protocol (802.1w). RSTP provides faster spanning tree convergence, the protocol will exchange information with neighboring switches to quickly transition to forwarding without creating loops. The code will default to RSTP mode but will downgrade any port connected to a legacy STP network so is fully backward compatible. Reviewed by: syrinx Tested by: syrinx
2006-11-01 09:07:47 +00:00
req->ifbop_maxage = bs->bs_bridge_max_age >> 8;
req->ifbop_hellotime = bs->bs_bridge_htime >> 8;
req->ifbop_fwddelay = bs->bs_bridge_fdelay >> 8;
Add some statistics that are needed to support RFC4188 as part of the SoC2006 work on a bridge monitoring module for BSNMP. Submitted by: shteryana (SoC 2006)
2006-07-31 20:24:46 +00:00
Bring in support for the Rapid Spanning Tree Protocol (802.1w). RSTP provides faster spanning tree convergence, the protocol will exchange information with neighboring switches to quickly transition to forwarding without creating loops. The code will default to RSTP mode but will downgrade any port connected to a legacy STP network so is fully backward compatible. Reviewed by: syrinx Tested by: syrinx
2006-11-01 09:07:47 +00:00
root_port = bs->bs_root_port;
Add some statistics that are needed to support RFC4188 as part of the SoC2006 work on a bridge monitoring module for BSNMP. Submitted by: shteryana (SoC 2006)
2006-07-31 20:24:46 +00:00
if (root_port == NULL)
req->ifbop_root_port = 0;
else
req->ifbop_root_port = root_port->bp_ifp->if_index;
Bring in support for the Rapid Spanning Tree Protocol (802.1w). RSTP provides faster spanning tree convergence, the protocol will exchange information with neighboring switches to quickly transition to forwarding without creating loops. The code will default to RSTP mode but will downgrade any port connected to a legacy STP network so is fully backward compatible. Reviewed by: syrinx Tested by: syrinx
2006-11-01 09:07:47 +00:00
req->ifbop_holdcount = bs->bs_txholdcount;
req->ifbop_priority = bs->bs_bridge_priority;
req->ifbop_protocol = bs->bs_protover;
req->ifbop_root_path_cost = bs->bs_root_pv.pv_cost;
Sync with the OpenBSD port of RSTP - use flags rather than sperate ioctls for edge, p2p - implement p2p and autop2p flags - define large pathcost constant as ULL - show bridgeid and rootid in ifconfig Obtained from: Reyk Floeter <reyk@openbsd.org>
2006-11-27 00:35:40 +00:00
req->ifbop_bridgeid = bs->bs_bridge_pv.pv_dbridge_id;
Bring in support for the Rapid Spanning Tree Protocol (802.1w). RSTP provides faster spanning tree convergence, the protocol will exchange information with neighboring switches to quickly transition to forwarding without creating loops. The code will default to RSTP mode but will downgrade any port connected to a legacy STP network so is fully backward compatible. Reviewed by: syrinx Tested by: syrinx
2006-11-01 09:07:47 +00:00
req->ifbop_designated_root = bs->bs_root_pv.pv_root_id;
Sync with the OpenBSD port of RSTP - use flags rather than sperate ioctls for edge, p2p - implement p2p and autop2p flags - define large pathcost constant as ULL - show bridgeid and rootid in ifconfig Obtained from: Reyk Floeter <reyk@openbsd.org>
2006-11-27 00:35:40 +00:00
req->ifbop_designated_bridge = bs->bs_root_pv.pv_dbridge_id;
Bring in support for the Rapid Spanning Tree Protocol (802.1w). RSTP provides faster spanning tree convergence, the protocol will exchange information with neighboring switches to quickly transition to forwarding without creating loops. The code will default to RSTP mode but will downgrade any port connected to a legacy STP network so is fully backward compatible. Reviewed by: syrinx Tested by: syrinx
2006-11-01 09:07:47 +00:00
req->ifbop_last_tc_time.tv_sec = bs->bs_last_tc_time.tv_sec;
req->ifbop_last_tc_time.tv_usec = bs->bs_last_tc_time.tv_usec;
Add some statistics that are needed to support RFC4188 as part of the SoC2006 work on a bridge monitoring module for BSNMP. Submitted by: shteryana (SoC 2006)
2006-07-31 20:24:46 +00:00
return (0);
}
static int
bridge_ioctl_grte(struct bridge_softc *sc, void *arg)
{
struct ifbrparam *param = arg;
param->ifbrp_cexceeded = sc->sc_brtexceeded;
return (0);
}
static int
bridge_ioctl_gifsstp(struct bridge_softc *sc, void *arg)
{
struct ifbpstpconf *bifstp = arg;
struct bridge_iflist *bif;
Bring in support for the Rapid Spanning Tree Protocol (802.1w). RSTP provides faster spanning tree convergence, the protocol will exchange information with neighboring switches to quickly transition to forwarding without creating loops. The code will default to RSTP mode but will downgrade any port connected to a legacy STP network so is fully backward compatible. Reviewed by: syrinx Tested by: syrinx
2006-11-01 09:07:47 +00:00
struct bstp_port *bp;
Add some statistics that are needed to support RFC4188 as part of the SoC2006 work on a bridge monitoring module for BSNMP. Submitted by: shteryana (SoC 2006)
2006-07-31 20:24:46 +00:00
struct ifbpstpreq bpreq;
Avoid holding the softc lock when using copyout(). Reported by: dfr Approved by: re (rwatson)
2007-07-26 20:30:18 +00:00
char *buf, *outbuf;
int count, buflen, len, error = 0;
Add some statistics that are needed to support RFC4188 as part of the SoC2006 work on a bridge monitoring module for BSNMP. Submitted by: shteryana (SoC 2006)
2006-07-31 20:24:46 +00:00
count = 0;
bridge: Change lists to CK_LIST as a peparation for epochification Prepare the ground for a rework of the bridge locking approach. We will use an epoch-based approach in the datapath and making it safe to iterate over the interface, span and rtnode lists without holding the BRIDGE_LOCK. Replace the relevant lists by their ConcurrencyKit equivalents. No functional change in this commit. Reviewed by: emaste, ae, philip (previous version) Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24249
2020-04-05 17:15:20 +00:00
CK_LIST_FOREACH(bif, &sc->sc_iflist, bif_next) {
Add some statistics that are needed to support RFC4188 as part of the SoC2006 work on a bridge monitoring module for BSNMP. Submitted by: shteryana (SoC 2006)
2006-07-31 20:24:46 +00:00
if ((bif->bif_flags & IFBIF_STP) != 0)
count++;
}
Avoid holding the softc lock when using copyout(). Reported by: dfr Approved by: re (rwatson)
2007-07-26 20:30:18 +00:00
buflen = sizeof(bpreq) * count;
Add some statistics that are needed to support RFC4188 as part of the SoC2006 work on a bridge monitoring module for BSNMP. Submitted by: shteryana (SoC 2006)
2006-07-31 20:24:46 +00:00
if (bifstp->ifbpstp_len == 0) {
Avoid holding the softc lock when using copyout(). Reported by: dfr Approved by: re (rwatson)
2007-07-26 20:30:18 +00:00
bifstp->ifbpstp_len = buflen;
Add some statistics that are needed to support RFC4188 as part of the SoC2006 work on a bridge monitoring module for BSNMP. Submitted by: shteryana (SoC 2006)
2006-07-31 20:24:46 +00:00
return (0);
}
bridge: Don't sleep during epoch While it doesn't trigger INVARIANTS or WITNESS on head it does in stable/12. There's also no reason for it, as we can easily report the out of memory error to the caller (i.e. userspace). All of these can already fail. PR: 248046 MFC after: 3 days
2020-07-18 12:43:11 +00:00
outbuf = malloc(buflen, M_TEMP, M_NOWAIT | M_ZERO);
if (outbuf == NULL)
return (ENOMEM);
Avoid holding the softc lock when using copyout(). Reported by: dfr Approved by: re (rwatson)
2007-07-26 20:30:18 +00:00
Add some statistics that are needed to support RFC4188 as part of the SoC2006 work on a bridge monitoring module for BSNMP. Submitted by: shteryana (SoC 2006)
2006-07-31 20:24:46 +00:00
count = 0;
Avoid holding the softc lock when using copyout(). Reported by: dfr Approved by: re (rwatson)
2007-07-26 20:30:18 +00:00
buf = outbuf;
len = min(bifstp->ifbpstp_len, buflen);
Add some statistics that are needed to support RFC4188 as part of the SoC2006 work on a bridge monitoring module for BSNMP. Submitted by: shteryana (SoC 2006)
2006-07-31 20:24:46 +00:00
bzero(&bpreq, sizeof(bpreq));
bridge: Change lists to CK_LIST as a peparation for epochification Prepare the ground for a rework of the bridge locking approach. We will use an epoch-based approach in the datapath and making it safe to iterate over the interface, span and rtnode lists without holding the BRIDGE_LOCK. Replace the relevant lists by their ConcurrencyKit equivalents. No functional change in this commit. Reviewed by: emaste, ae, philip (previous version) Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24249
2020-04-05 17:15:20 +00:00
CK_LIST_FOREACH(bif, &sc->sc_iflist, bif_next) {
Add some statistics that are needed to support RFC4188 as part of the SoC2006 work on a bridge monitoring module for BSNMP. Submitted by: shteryana (SoC 2006)
2006-07-31 20:24:46 +00:00
if (len < sizeof(bpreq))
break;
if ((bif->bif_flags & IFBIF_STP) == 0)
continue;
Bring in support for the Rapid Spanning Tree Protocol (802.1w). RSTP provides faster spanning tree convergence, the protocol will exchange information with neighboring switches to quickly transition to forwarding without creating loops. The code will default to RSTP mode but will downgrade any port connected to a legacy STP network so is fully backward compatible. Reviewed by: syrinx Tested by: syrinx
2006-11-01 09:07:47 +00:00
bp = &bif->bif_stp;
bpreq.ifbp_portno = bif->bif_ifp->if_index & 0xfff;
bpreq.ifbp_fwd_trans = bp->bp_forward_transitions;
bpreq.ifbp_design_cost = bp->bp_desg_pv.pv_cost;
bpreq.ifbp_design_port = bp->bp_desg_pv.pv_port_id;
bpreq.ifbp_design_bridge = bp->bp_desg_pv.pv_dbridge_id;
bpreq.ifbp_design_root = bp->bp_desg_pv.pv_root_id;
Add some statistics that are needed to support RFC4188 as part of the SoC2006 work on a bridge monitoring module for BSNMP. Submitted by: shteryana (SoC 2006)
2006-07-31 20:24:46 +00:00
Avoid holding the softc lock when using copyout(). Reported by: dfr Approved by: re (rwatson)
2007-07-26 20:30:18 +00:00
memcpy(buf, &bpreq, sizeof(bpreq));
Add some statistics that are needed to support RFC4188 as part of the SoC2006 work on a bridge monitoring module for BSNMP. Submitted by: shteryana (SoC 2006)
2006-07-31 20:24:46 +00:00
count++;
Avoid holding the softc lock when using copyout(). Reported by: dfr Approved by: re (rwatson)
2007-07-26 20:30:18 +00:00
buf += sizeof(bpreq);
Add some statistics that are needed to support RFC4188 as part of the SoC2006 work on a bridge monitoring module for BSNMP. Submitted by: shteryana (SoC 2006)
2006-07-31 20:24:46 +00:00
len -= sizeof(bpreq);
}
bifstp->ifbpstp_len = sizeof(bpreq) * count;
Avoid holding the softc lock when using copyout(). Reported by: dfr Approved by: re (rwatson)
2007-07-26 20:30:18 +00:00
error = copyout(outbuf, bifstp->ifbpstp_req, bifstp->ifbpstp_len);
free(outbuf, M_TEMP);
Add some statistics that are needed to support RFC4188 as part of the SoC2006 work on a bridge monitoring module for BSNMP. Submitted by: shteryana (SoC 2006)
2006-07-31 20:24:46 +00:00
return (error);
}
Bring in support for the Rapid Spanning Tree Protocol (802.1w). RSTP provides faster spanning tree convergence, the protocol will exchange information with neighboring switches to quickly transition to forwarding without creating loops. The code will default to RSTP mode but will downgrade any port connected to a legacy STP network so is fully backward compatible. Reviewed by: syrinx Tested by: syrinx
2006-11-01 09:07:47 +00:00
static int
bridge_ioctl_sproto(struct bridge_softc *sc, void *arg)
{
struct ifbrparam *param = arg;
return (bstp_set_protocol(&sc->sc_stp, param->ifbrp_proto));
}
static int
bridge_ioctl_stxhc(struct bridge_softc *sc, void *arg)
{
struct ifbrparam *param = arg;
return (bstp_set_holdcount(&sc->sc_stp, param->ifbrp_txhc));
}
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
/*
* bridge_ifdetach:
*
* Detach an interface from a bridge. Called when a member
* interface is detaching.
*/
Change most of the bridge and stp funtions to static. This has highlighted that the following funtions are not used, wrap in '#ifdef noused' for the moment. bstp_enable_change_detection bstp_disable_change_detection bstp_set_bridge_priority bstp_set_port_priority bstp_set_path_cost
2005-10-14 10:38:12 +00:00
static void
Change from a callback in if_ethersubr to using EVENTHANDLER in order to detach span ports when they disappear. The span port does not have a pointer to the softc so revert r1.31 and bring back the softc linked-list. MFC after: 2 weeks
2005-12-17 06:33:51 +00:00
bridge_ifdetach(void *arg __unused, struct ifnet *ifp)
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
{
struct bridge_softc *sc = ifp->if_bridge;
If we have been called from ether_ifdetach() then do not try and clear the promisc flag from the member interface, this is a no-op anyway since the interface is disappearing. The driver may have already released its resources such as miibus and this is likely to panic the kernel. Submitted and tested by: Wojciech A. Koszek MFC after: 2 weeks
2005-10-23 22:30:07 +00:00
struct bridge_iflist *bif;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
Ignore interface renames instead of removing the interface from the bridge group. Reviewed by: rstone Approved by: rstone (co-mentor) Sponsored by: Sandvine Incorporated MFC after: 1 week
2013-03-28 20:37:07 +00:00
if (ifp->if_flags & IFF_RENAMING)
return;
Fix a panic when tearing down a vnet on a VIMAGE-enabled kernel. There was a race that bridge_ifdetach() could be called via ifnet_departure event handler after vnet_bridge_uninit(). PR: 195859 Reported by: Danilo Egea Gondolfo
2015-02-14 18:15:14 +00:00
if (V_bridge_cloner == NULL) {
/*
* This detach handler can be called after
* vnet_bridge_uninit(). Just return in that case.
*/
return;
}
Change from a callback in if_ethersubr to using EVENTHANDLER in order to detach span ports when they disappear. The span port does not have a pointer to the softc so revert r1.31 and bring back the softc linked-list. MFC after: 2 weeks
2005-12-17 06:33:51 +00:00
/* Check if the interface is a bridge member */
if (sc != NULL) {
BRIDGE_LOCK(sc);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
Change from a callback in if_ethersubr to using EVENTHANDLER in order to detach span ports when they disappear. The span port does not have a pointer to the softc so revert r1.31 and bring back the softc linked-list. MFC after: 2 weeks
2005-12-17 06:33:51 +00:00
bif = bridge_lookup_member_if(sc, ifp);
if (bif != NULL)
bridge_delete_member(sc, bif, 1);
BRIDGE_UNLOCK(sc);
If we have been called from ether_ifdetach() then do not try and clear the promisc flag from the member interface, this is a no-op anyway since the interface is disappearing. The driver may have already released its resources such as miibus and this is likely to panic the kernel. Submitted and tested by: Wojciech A. Koszek MFC after: 2 weeks
2005-10-23 22:30:07 +00:00
return;
Change from a callback in if_ethersubr to using EVENTHANDLER in order to detach span ports when they disappear. The span port does not have a pointer to the softc so revert r1.31 and bring back the softc linked-list. MFC after: 2 weeks
2005-12-17 06:33:51 +00:00
}
If we have been called from ether_ifdetach() then do not try and clear the promisc flag from the member interface, this is a no-op anyway since the interface is disappearing. The driver may have already released its resources such as miibus and this is likely to panic the kernel. Submitted and tested by: Wojciech A. Koszek MFC after: 2 weeks
2005-10-23 22:30:07 +00:00
Change from a callback in if_ethersubr to using EVENTHANDLER in order to detach span ports when they disappear. The span port does not have a pointer to the softc so revert r1.31 and bring back the softc linked-list. MFC after: 2 weeks
2005-12-17 06:33:51 +00:00
/* Check if the interface is a span port */
Virtualize if_bridge(4) cloner.
2014-10-05 19:43:37 +00:00
BRIDGE_LIST_LOCK();
LIST_FOREACH(sc, &V_bridge_list, sc_list) {
Change from a callback in if_ethersubr to using EVENTHANDLER in order to detach span ports when they disappear. The span port does not have a pointer to the softc so revert r1.31 and bring back the softc linked-list. MFC after: 2 weeks
2005-12-17 06:33:51 +00:00
BRIDGE_LOCK(sc);
bridge: Change lists to CK_LIST as a peparation for epochification Prepare the ground for a rework of the bridge locking approach. We will use an epoch-based approach in the datapath and making it safe to iterate over the interface, span and rtnode lists without holding the BRIDGE_LOCK. Replace the relevant lists by their ConcurrencyKit equivalents. No functional change in this commit. Reviewed by: emaste, ae, philip (previous version) Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24249
2020-04-05 17:15:20 +00:00
CK_LIST_FOREACH(bif, &sc->sc_spanlist, bif_next)
Change from a callback in if_ethersubr to using EVENTHANDLER in order to detach span ports when they disappear. The span port does not have a pointer to the softc so revert r1.31 and bring back the softc linked-list. MFC after: 2 weeks
2005-12-17 06:33:51 +00:00
if (ifp == bif->bif_ifp) {
bridge_delete_span(sc, bif);
break;
}
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
Change from a callback in if_ethersubr to using EVENTHANDLER in order to detach span ports when they disappear. The span port does not have a pointer to the softc so revert r1.31 and bring back the softc linked-list. MFC after: 2 weeks
2005-12-17 06:33:51 +00:00
BRIDGE_UNLOCK(sc);
}
Virtualize if_bridge(4) cloner.
2014-10-05 19:43:37 +00:00
BRIDGE_LIST_UNLOCK();
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
}
/*
* bridge_init:
*
* Initialize a bridge interface.
*/
static void
bridge_init(void *xsc)
{
struct bridge_softc *sc = (struct bridge_softc *)xsc;
Stop embedding struct ifnet at the top of driver softcs. Instead the struct ifnet or the layer 2 common structure it was embedded in have been replaced with a struct ifnet pointer to be filled by a call to the new function, if_alloc(). The layer 2 common structure is also allocated via if_alloc() based on the interface type. It is hung off the new struct ifnet member, if_l2com. This change removes the size of these structures from the kernel ABI and will allow us to better manage them as interfaces come and go. Other changes of note: - Struct arpcom is no longer referenced in normal interface code. Instead the Ethernet address is accessed via the IFP2ENADDR() macro. To enforce this ac_enaddr has been renamed to _ac_enaddr. - The second argument to ether_ifattach is now always the mac address from driver private storage rather than sometimes being ac_enaddr. Reviewed by: sobomax, sam
2005-06-10 16:49:24 +00:00
struct ifnet *ifp = sc->sc_ifp;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
Propagate rename of IFF_OACTIVE and IFF_RUNNING to IFF_DRV_OACTIVE and IFF_DRV_RUNNING, as well as the move from ifnet.if_flags to ifnet.if_drv_flags. Device drivers are now responsible for synchronizing access to these flags, as they are in if_drv_flags. This helps prevent races between the network stack and device driver in maintaining the interface flags field. Many __FreeBSD__ and __FreeBSD_version checks maintained and continued; some less so. Reviewed by: pjd, bz MFC after: 7 days
2005-08-09 10:20:02 +00:00
if (ifp->if_drv_flags & IFF_DRV_RUNNING)
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
return;
Mark the callouts as MPSAFE as if_bridge has been giant-free since day 1. Use the SMP friendly callout_init_mtx() while we are here. Approved by: mlaier (mentor) MFC after: 3 days
2005-08-18 20:17:00 +00:00
BRIDGE_LOCK(sc);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
callout_reset(&sc->sc_brcallout, bridge_rtable_prune_period * hz,
bridge_timer, sc);
Propagate rename of IFF_OACTIVE and IFF_RUNNING to IFF_DRV_OACTIVE and IFF_DRV_RUNNING, as well as the move from ifnet.if_flags to ifnet.if_drv_flags. Device drivers are now responsible for synchronizing access to these flags, as they are in if_drv_flags. This helps prevent races between the network stack and device driver in maintaining the interface flags field. Many __FreeBSD__ and __FreeBSD_version checks maintained and continued; some less so. Reviewed by: pjd, bz MFC after: 7 days
2005-08-09 10:20:02 +00:00
ifp->if_drv_flags |= IFF_DRV_RUNNING;
/tmp/cvsuusTrc
2006-07-26 10:43:02 +00:00
bstp_init(&sc->sc_stp); /* Initialize Spanning Tree */
Ensure that we are holding the lock when initialising the bridge interface. We could initialise while unlocked if the bridge is not up when setting the inet address, ether_ioctl() would call bridge_init. Change it so bridge_init is always called unlocked and then locks before calling bstp_initialization(). Reported by: Michal Mertl Approved by: mlaier (mentor) MFC after: 3 days
2005-08-15 02:50:13 +00:00
BRIDGE_UNLOCK(sc);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
}
/*
* bridge_stop:
*
* Stop the bridge interface.
*/
Change most of the bridge and stp funtions to static. This has highlighted that the following funtions are not used, wrap in '#ifdef noused' for the moment. bstp_enable_change_detection bstp_disable_change_detection bstp_set_bridge_priority bstp_set_port_priority bstp_set_path_cost
2005-10-14 10:38:12 +00:00
static void
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge_stop(struct ifnet *ifp, int disable)
{
struct bridge_softc *sc = ifp->if_softc;
Ensure that we are holding the lock when initialising the bridge interface. We could initialise while unlocked if the bridge is not up when setting the inet address, ether_ioctl() would call bridge_init. Change it so bridge_init is always called unlocked and then locks before calling bstp_initialization(). Reported by: Michal Mertl Approved by: mlaier (mentor) MFC after: 3 days
2005-08-15 02:50:13 +00:00
BRIDGE_LOCK_ASSERT(sc);
Propagate rename of IFF_OACTIVE and IFF_RUNNING to IFF_DRV_OACTIVE and IFF_DRV_RUNNING, as well as the move from ifnet.if_flags to ifnet.if_drv_flags. Device drivers are now responsible for synchronizing access to these flags, as they are in if_drv_flags. This helps prevent races between the network stack and device driver in maintaining the interface flags field. Many __FreeBSD__ and __FreeBSD_version checks maintained and continued; some less so. Reviewed by: pjd, bz MFC after: 7 days
2005-08-09 10:20:02 +00:00
if ((ifp->if_drv_flags & IFF_DRV_RUNNING) == 0)
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
return;
bridge: call member interface ioctl() without NET_EPOCH We're not allowed to hold NET_EPOCH while sleeping, so when we call ioctl() handlers for member interfaces we cannot be in NET_EPOCH. We still need some protection of our CK_LISTs, so hold BRIDGE_LOCK instead. That requires changing BRIDGE_LOCK into a sleepable lock, and separating the BRIDGE_RT_LOCK, to protect bridge_rtnode lists. That lock is taken in the data path (while in NET_EPOCH), so it cannot be a sleepable lock. While here document the locking strategy. MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D26418
2020-10-06 19:19:56 +00:00
BRIDGE_RT_LOCK(sc);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
callout_stop(&sc->sc_brcallout);
bridge: call member interface ioctl() without NET_EPOCH We're not allowed to hold NET_EPOCH while sleeping, so when we call ioctl() handlers for member interfaces we cannot be in NET_EPOCH. We still need some protection of our CK_LISTs, so hold BRIDGE_LOCK instead. That requires changing BRIDGE_LOCK into a sleepable lock, and separating the BRIDGE_RT_LOCK, to protect bridge_rtnode lists. That lock is taken in the data path (while in NET_EPOCH), so it cannot be a sleepable lock. While here document the locking strategy. MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D26418
2020-10-06 19:19:56 +00:00
/tmp/cvsuusTrc
2006-07-26 10:43:02 +00:00
bstp_stop(&sc->sc_stp);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge_rtflush(sc, IFBF_FLUSHDYN);
bridge: call member interface ioctl() without NET_EPOCH We're not allowed to hold NET_EPOCH while sleeping, so when we call ioctl() handlers for member interfaces we cannot be in NET_EPOCH. We still need some protection of our CK_LISTs, so hold BRIDGE_LOCK instead. That requires changing BRIDGE_LOCK into a sleepable lock, and separating the BRIDGE_RT_LOCK, to protect bridge_rtnode lists. That lock is taken in the data path (while in NET_EPOCH), so it cannot be a sleepable lock. While here document the locking strategy. MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D26418
2020-10-06 19:19:56 +00:00
BRIDGE_RT_UNLOCK(sc);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
Propagate rename of IFF_OACTIVE and IFF_RUNNING to IFF_DRV_OACTIVE and IFF_DRV_RUNNING, as well as the move from ifnet.if_flags to ifnet.if_drv_flags. Device drivers are now responsible for synchronizing access to these flags, as they are in if_drv_flags. This helps prevent races between the network stack and device driver in maintaining the interface flags field. Many __FreeBSD__ and __FreeBSD_version checks maintained and continued; some less so. Reviewed by: pjd, bz MFC after: 7 days
2005-08-09 10:20:02 +00:00
ifp->if_drv_flags &= ~IFF_DRV_RUNNING;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
}
/*
* bridge_enqueue:
*
* Enqueue a packet on a bridge member interface.
*
*/
Change bridge(4) to use if_transmit for forwarding packets to underlying interfaces instead of queueing. Tested by: ray
2012-09-03 10:08:20 +00:00
static int
- Previously when broadcasting to N number of interfaces we would run pfil hooks for each outgoing interface but also run pfil hooks _N times_ on the bridge interface. This is changed so pfil hooks are run once for the bridge interface (bridge0) and then only on the outgoing interfaces in the broadcast loop. - Simplify bridge_enqueue() by moving bridge_pfil() to the callers. - Check (inet6_pfil_hook.ph_busy_count >= 0), it may be possible to have a packet filter hooked for only ipv6 but we were only checking if ipv4 hooks were busy. - Minor optimisation for null mbuf check after bridge_pfil(), move it into the if-block as it couldnt possibly be null outside. Prodded by: mlaier Approved by: re (scottl), mlaier (mentor)
2005-07-06 01:24:45 +00:00
bridge_enqueue(struct bridge_softc *sc, struct ifnet *dst_ifp, struct mbuf *m)
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
{
Add support for fragmenting ipv4 packets. The packet filter may reassemble the ip fragments and return a packet that is larger than the MTU of the sending interface. There is no check for DF or icmp replies as we can only get a large packet to fragment by reassembling a previous fragment, and this only happens after a call to pfil(9). Obtained from: OpenBSD (mostly) Glanced at by: mlaier MFC after: 1 month
2006-04-29 05:37:25 +00:00
int len, err = 0;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
short mflags;
Add support for fragmenting ipv4 packets. The packet filter may reassemble the ip fragments and return a packet that is larger than the MTU of the sending interface. There is no check for DF or icmp replies as we can only get a large packet to fragment by reassembling a previous fragment, and this only happens after a call to pfil(9). Obtained from: OpenBSD (mostly) Glanced at by: mlaier MFC after: 1 month
2006-04-29 05:37:25 +00:00
struct mbuf *m0;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
Fix spelling mistake in comment.
2006-06-19 02:25:11 +00:00
/* We may be sending a fragment so traverse the mbuf */
Add support for fragmenting ipv4 packets. The packet filter may reassemble the ip fragments and return a packet that is larger than the MTU of the sending interface. There is no check for DF or icmp replies as we can only get a large packet to fragment by reassembling a previous fragment, and this only happens after a call to pfil(9). Obtained from: OpenBSD (mostly) Glanced at by: mlaier MFC after: 1 month
2006-04-29 05:37:25 +00:00
for (; m; m = m0) {
m0 = m->m_nextpkt;
m->m_nextpkt = NULL;
- In the bridge_enqueue() do success/error accounting for each fragment, not only once. - In the GRAB_OUR_PACKETS() macro do increase if_ibytes.
2012-09-26 20:09:48 +00:00
len = m->m_pkthdr.len;
mflags = m->m_flags;
The bridging output function puts the mbuf directly on the interfaces send queue so the output network card must support the same tagging mechanism as how the frame was input (prepended Ethernet header tag or stripped HW mflag). Now the vlan Ethernet header is _always_ stripped in ether_input and the mbuf flagged, only only network cards with VLAN_HWTAGGING enabled would properly re-tag any outgoing vlan frames. If the outgoing interface does not support hardware tagging then readd the vlan header to the front of the frame. Move the common vlan encapsulation in to ether_vlanencap(). Reported by: Erik Osterholm, Jon Otterholm MFC after: 1 week
2007-10-18 21:22:15 +00:00
/*
* If underlying interface can not do VLAN tag insertion itself
* then attach a packet tag that holds it.
*/
if ((m->m_flags & M_VLANTAG) &&
(dst_ifp->if_capenable & IFCAP_VLAN_HWTAGGING) == 0) {
m = ether_vlanencap(m, m->m_pkthdr.ether_vtag);
if (m == NULL) {
if_printf(dst_ifp,
"unable to prepend VLAN header\n");
Mechanically convert to if_inc_counter().
2014-09-19 10:39:58 +00:00
if_inc_counter(dst_ifp, IFCOUNTER_OERRORS, 1);
The bridging output function puts the mbuf directly on the interfaces send queue so the output network card must support the same tagging mechanism as how the frame was input (prepended Ethernet header tag or stripped HW mflag). Now the vlan Ethernet header is _always_ stripped in ether_input and the mbuf flagged, only only network cards with VLAN_HWTAGGING enabled would properly re-tag any outgoing vlan frames. If the outgoing interface does not support hardware tagging then readd the vlan header to the front of the frame. Move the common vlan encapsulation in to ether_vlanencap(). Reported by: Erik Osterholm, Jon Otterholm MFC after: 1 week
2007-10-18 21:22:15 +00:00
continue;
}
m->m_flags &= ~M_VLANTAG;
}
bridge: Fix fragment handling and memory leak Fragmented UDP and ICMP packets were corrupted if a firewall with reassembling feature (like pf'scrub) is enabled on the bridge. This patch fixes corrupted packet problem and the panic (triggered easly with low RAM) as explain in PR 185633. bridge_pfil and bridge_fragment relationship: bridge_pfil() receive (IN direction) packets and sent it to the firewall The firewall can be configured for reassembling fragmented packet (like pf'scrubing) in one mbuf chain when bridge_pfil() need to send this reassembled packet to the outgoing interface, it needs to re-fragment it by using bridge_fragment() bridge_fragment() had to split this mbuf (using ip_fragment) first then had to M_PREPEND each packet in the mbuf chain for adding Ethernet header. But M_PREPEND can sometime create a new mbuf on the begining of the mbuf chain, then the "main" pointer of this mbuf chain should be updated and this case is tottaly forgotten. The original bridge_fragment code (Revision 158140, 2006 April 29) came from OpenBSD, and the call to bridge_enqueue was embedded. But on FreeBSD, bridge_enqueue() is done after bridge_fragment(), then the original OpenBSD code can't work as-it of FreeBSD. PR: 185633 Submitted by: Olivier Cochard-Labbé Differential Revision: https://reviews.freebsd.org/D7780
2016-09-24 07:09:43 +00:00
M_ASSERTPKTHDR(m); /* We shouldn't transmit mbuf without pkthdr */
Simplify error case Submitted by: thompsa@
2012-07-10 20:59:35 +00:00
if ((err = dst_ifp->if_transmit(dst_ifp, m))) {
m_freem(m0);
Mechanically convert to if_inc_counter().
2014-09-19 10:39:58 +00:00
if_inc_counter(sc->sc_ifp, IFCOUNTER_OERRORS, 1);
Simplify error case Submitted by: thompsa@
2012-07-10 20:59:35 +00:00
break;
}
Add support for fragmenting ipv4 packets. The packet filter may reassemble the ip fragments and return a packet that is larger than the MTU of the sending interface. There is no check for DF or icmp replies as we can only get a large packet to fragment by reassembling a previous fragment, and this only happens after a call to pfil(9). Obtained from: OpenBSD (mostly) Glanced at by: mlaier MFC after: 1 month
2006-04-29 05:37:25 +00:00
Mechanically convert to if_inc_counter().
2014-09-19 10:39:58 +00:00
if_inc_counter(sc->sc_ifp, IFCOUNTER_OPACKETS, 1);
if_inc_counter(sc->sc_ifp, IFCOUNTER_OBYTES, len);
use if_transmit intead of direct frobbing of the if_snd q; this is no longer allowed Identified by: rwatson Reviewed by: kmacy
2009-04-27 22:06:49 +00:00
if (mflags & M_MCAST)
Mechanically convert to if_inc_counter().
2014-09-19 10:39:58 +00:00
if_inc_counter(sc->sc_ifp, IFCOUNTER_OMCASTS, 1);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
}
Change bridge(4) to use if_transmit for forwarding packets to underlying interfaces instead of queueing. Tested by: ray
2012-09-03 10:08:20 +00:00
return (err);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
}
Add dummynet(4) support to if_bridge, this code is largely based on bridge.c. This is the final piece to match bridge.c in functionality, we can now be a drop-in replacement. Approved by: mlaier (mentor)
2005-06-10 01:25:22 +00:00
/*
* bridge_dummynet:
*
* Receive a queued packet from dummynet and pass it on to the output
* interface.
*
* The mbuf has the Ethernet header already attached.
*/
Change most of the bridge and stp funtions to static. This has highlighted that the following funtions are not used, wrap in '#ifdef noused' for the moment. bstp_enable_change_detection bstp_disable_change_detection bstp_set_bridge_priority bstp_set_port_priority bstp_set_path_cost
2005-10-14 10:38:12 +00:00
static void
Add dummynet(4) support to if_bridge, this code is largely based on bridge.c. This is the final piece to match bridge.c in functionality, we can now be a drop-in replacement. Approved by: mlaier (mentor)
2005-06-10 01:25:22 +00:00
bridge_dummynet(struct mbuf *m, struct ifnet *ifp)
{
struct bridge_softc *sc;
sc = ifp->if_bridge;
/*
* The packet didnt originate from a member interface. This should only
* ever happen if a member interface is removed while packets are
* queued for it.
*/
Add missing {} in last commit.
2005-06-10 15:53:21 +00:00
if (sc == NULL) {
Add dummynet(4) support to if_bridge, this code is largely based on bridge.c. This is the final piece to match bridge.c in functionality, we can now be a drop-in replacement. Approved by: mlaier (mentor)
2005-06-10 01:25:22 +00:00
m_freem(m);
return;
Add missing {} in last commit.
2005-06-10 15:53:21 +00:00
}
Add dummynet(4) support to if_bridge, this code is largely based on bridge.c. This is the final piece to match bridge.c in functionality, we can now be a drop-in replacement. Approved by: mlaier (mentor)
2005-06-10 01:25:22 +00:00
New pfil(9) KPI together with newborn pfil API and control utility. The KPI have been reviewed and cleansed of features that were planned back 20 years ago and never implemented. The pfil(9) internals have been made opaque to protocols with only returned types and function declarations exposed. The KPI is made more strict, but at the same time more extensible, as kernel uses same command structures that userland ioctl uses. In nutshell [KA]PI is about declaring filtering points, declaring filters and linking and unlinking them together. New [KA]PI makes it possible to reconfigure pfil(9) configuration: change order of hooks, rehook filter from one filtering point to a different one, disconnect a hook on output leaving it on input only, prepend/append a filter to existing list of filters. Now it possible for a single packet filter to provide multiple rulesets that may be linked to different points. Think of per-interface ACLs in Cisco or Juniper. None of existing packet filters yet support that, however limited usage is already possible, e.g. default ruleset can be moved to single interface, as soon as interface would pride their filtering points. Another future feature is possiblity to create pfil heads, that provide not an mbuf pointer but just a memory pointer with length. That would allow filtering at very early stages of a packet lifecycle, e.g. when packet has just been received by a NIC and no mbuf was yet allocated. Differential Revision: https://reviews.freebsd.org/D18951
2019-01-31 23:01:03 +00:00
if (PFIL_HOOKED_OUT(V_inet_pfil_head)
- Previously when broadcasting to N number of interfaces we would run pfil hooks for each outgoing interface but also run pfil hooks _N times_ on the bridge interface. This is changed so pfil hooks are run once for the bridge interface (bridge0) and then only on the outgoing interfaces in the broadcast loop. - Simplify bridge_enqueue() by moving bridge_pfil() to the callers. - Check (inet6_pfil_hook.ph_busy_count >= 0), it may be possible to have a packet filter hooked for only ipv6 but we were only checking if ipv4 hooks were busy. - Minor optimisation for null mbuf check after bridge_pfil(), move it into the if-block as it couldnt possibly be null outside. Prodded by: mlaier Approved by: re (scottl), mlaier (mentor)
2005-07-06 01:24:45 +00:00
#ifdef INET6
New pfil(9) KPI together with newborn pfil API and control utility. The KPI have been reviewed and cleansed of features that were planned back 20 years ago and never implemented. The pfil(9) internals have been made opaque to protocols with only returned types and function declarations exposed. The KPI is made more strict, but at the same time more extensible, as kernel uses same command structures that userland ioctl uses. In nutshell [KA]PI is about declaring filtering points, declaring filters and linking and unlinking them together. New [KA]PI makes it possible to reconfigure pfil(9) configuration: change order of hooks, rehook filter from one filtering point to a different one, disconnect a hook on output leaving it on input only, prepend/append a filter to existing list of filters. Now it possible for a single packet filter to provide multiple rulesets that may be linked to different points. Think of per-interface ACLs in Cisco or Juniper. None of existing packet filters yet support that, however limited usage is already possible, e.g. default ruleset can be moved to single interface, as soon as interface would pride their filtering points. Another future feature is possiblity to create pfil heads, that provide not an mbuf pointer but just a memory pointer with length. That would allow filtering at very early stages of a packet lifecycle, e.g. when packet has just been received by a NIC and no mbuf was yet allocated. Differential Revision: https://reviews.freebsd.org/D18951
2019-01-31 23:01:03 +00:00
|| PFIL_HOOKED_OUT(V_inet6_pfil_head)
- Previously when broadcasting to N number of interfaces we would run pfil hooks for each outgoing interface but also run pfil hooks _N times_ on the bridge interface. This is changed so pfil hooks are run once for the bridge interface (bridge0) and then only on the outgoing interfaces in the broadcast loop. - Simplify bridge_enqueue() by moving bridge_pfil() to the callers. - Check (inet6_pfil_hook.ph_busy_count >= 0), it may be possible to have a packet filter hooked for only ipv6 but we were only checking if ipv4 hooks were busy. - Minor optimisation for null mbuf check after bridge_pfil(), move it into the if-block as it couldnt possibly be null outside. Prodded by: mlaier Approved by: re (scottl), mlaier (mentor)
2005-07-06 01:24:45 +00:00
#endif
) {
if (bridge_pfil(&m, sc->sc_ifp, ifp, PFIL_OUT) != 0)
return;
if (m == NULL)
return;
}
bridge_enqueue(sc, ifp, m);
Add dummynet(4) support to if_bridge, this code is largely based on bridge.c. This is the final piece to match bridge.c in functionality, we can now be a drop-in replacement. Approved by: mlaier (mentor)
2005-06-10 01:25:22 +00:00
}
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
/*
* bridge_output:
*
* Send output from a bridge member interface. This
* performs the bridging function for locally originated
* packets.
*
* The mbuf has the Ethernet header already attached. We must
* enqueue or free the mbuf before returning.
*/
Change most of the bridge and stp funtions to static. This has highlighted that the following funtions are not used, wrap in '#ifdef noused' for the moment. bstp_enable_change_detection bstp_disable_change_detection bstp_set_bridge_priority bstp_set_port_priority bstp_set_path_cost
2005-10-14 10:38:12 +00:00
static int
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge_output(struct ifnet *ifp, struct mbuf *m, struct sockaddr *sa,
Unbreak build by reverting if_bridge part of r360047. Pointy hat to: melifaro
2020-04-17 18:22:37 +00:00
struct rtentry *rt)
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
{
struct ether_header *eh;
if_bridge(4): Complete bpf auditing of local traffic over the bridge There were two remaining "gaps" in auditing local bridge traffic with bpf(4): Locally originated outbound traffic from a member interface is invisible to the bridge's bpf(4) interface. Inbound traffic locally destined to a member interface is invisible to the member's bpf(4) interface -- this traffic has no chance after bridge_input to otherwise pass it over, and it wasn't originally received on this interface. I call these "gaps" because they don't affect conventional bridge setups. Alas, being able to establish an audit trail of all locally destined traffic for setups that can function like this is useful in some scenarios. Reviewed by: kp MFC after: 1 week Differential Revision: https://reviews.freebsd.org/D19757
2019-05-29 01:08:30 +00:00
struct ifnet *bifp, *dst_if;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
struct bridge_softc *sc;
Add the vlan tag to the bridge route table. This allows a vlan trunk to be bridged, previously legitimate traffic was not passed as the bridge could not tell that it was on a different Ethernet segment. All non-tagged traffic is treated as vlan1 as per IEEE 802.1Q-2003
2007-06-13 18:58:04 +00:00
uint16_t vlan;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge: epoch-ification Run the bridge datapath under epoch, rather than under the BRIDGE_LOCK(). We still take the BRIDGE_LOCK() whenever we insert or delete items in the relevant lists, but we use epoch callbacks to free items so that it's safe to iterate the lists without the BRIDGE_LOCK. Tests on mercat5/6 shows this increases bridge throughput significantly, from 3.7Mpps to 18.6Mpps. Reviewed by: emaste, philip, melifaro MFC after: 2 months Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24250
2020-04-26 16:22:35 +00:00
NET_EPOCH_ASSERT();
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
if (m->m_len < ETHER_HDR_LEN) {
m = m_pullup(m, ETHER_HDR_LEN);
if (m == NULL)
return (0);
}
eh = mtod(m, struct ether_header *);
sc = ifp->if_bridge;
Add the vlan tag to the bridge route table. This allows a vlan trunk to be bridged, previously legitimate traffic was not passed as the bridge could not tell that it was on a different Ethernet segment. All non-tagged traffic is treated as vlan1 as per IEEE 802.1Q-2003
2007-06-13 18:58:04 +00:00
vlan = VLANTAGOF(m);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
if_bridge(4): Complete bpf auditing of local traffic over the bridge There were two remaining "gaps" in auditing local bridge traffic with bpf(4): Locally originated outbound traffic from a member interface is invisible to the bridge's bpf(4) interface. Inbound traffic locally destined to a member interface is invisible to the member's bpf(4) interface -- this traffic has no chance after bridge_input to otherwise pass it over, and it wasn't originally received on this interface. I call these "gaps" because they don't affect conventional bridge setups. Alas, being able to establish an audit trail of all locally destined traffic for setups that can function like this is useful in some scenarios. Reviewed by: kp MFC after: 1 week Differential Revision: https://reviews.freebsd.org/D19757
2019-05-29 01:08:30 +00:00
bifp = sc->sc_ifp;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
/*
* If bridge is down, but the original output interface is up,
* go ahead and send out that interface. Otherwise, the packet
* is dropped below.
*/
if_bridge(4): Complete bpf auditing of local traffic over the bridge There were two remaining "gaps" in auditing local bridge traffic with bpf(4): Locally originated outbound traffic from a member interface is invisible to the bridge's bpf(4) interface. Inbound traffic locally destined to a member interface is invisible to the member's bpf(4) interface -- this traffic has no chance after bridge_input to otherwise pass it over, and it wasn't originally received on this interface. I call these "gaps" because they don't affect conventional bridge setups. Alas, being able to establish an audit trail of all locally destined traffic for setups that can function like this is useful in some scenarios. Reviewed by: kp MFC after: 1 week Differential Revision: https://reviews.freebsd.org/D19757
2019-05-29 01:08:30 +00:00
if ((bifp->if_drv_flags & IFF_DRV_RUNNING) == 0) {
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
dst_if = ifp;
goto sendunicast;
}
/*
* If the packet is a multicast, or we don't know a better way to
* get there, send to all interfaces.
*/
if (ETHER_IS_MULTICAST(eh->ether_dhost))
dst_if = NULL;
else
Add the vlan tag to the bridge route table. This allows a vlan trunk to be bridged, previously legitimate traffic was not passed as the bridge could not tell that it was on a different Ethernet segment. All non-tagged traffic is treated as vlan1 as per IEEE 802.1Q-2003
2007-06-13 18:58:04 +00:00
dst_if = bridge_rtlookup(sc, eh->ether_dhost, vlan);
if_bridge(4): Complete bpf auditing of local traffic over the bridge There were two remaining "gaps" in auditing local bridge traffic with bpf(4): Locally originated outbound traffic from a member interface is invisible to the bridge's bpf(4) interface. Inbound traffic locally destined to a member interface is invisible to the member's bpf(4) interface -- this traffic has no chance after bridge_input to otherwise pass it over, and it wasn't originally received on this interface. I call these "gaps" because they don't affect conventional bridge setups. Alas, being able to establish an audit trail of all locally destined traffic for setups that can function like this is useful in some scenarios. Reviewed by: kp MFC after: 1 week Differential Revision: https://reviews.freebsd.org/D19757
2019-05-29 01:08:30 +00:00
/* Tap any traffic not passing back out the originating interface */
if (dst_if != ifp)
ETHER_BPF_MTAP(bifp, m);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
if (dst_if == NULL) {
struct bridge_iflist *bif;
struct mbuf *mc;
bridge: epoch-ification Run the bridge datapath under epoch, rather than under the BRIDGE_LOCK(). We still take the BRIDGE_LOCK() whenever we insert or delete items in the relevant lists, but we use epoch callbacks to free items so that it's safe to iterate the lists without the BRIDGE_LOCK. Tests on mercat5/6 shows this increases bridge throughput significantly, from 3.7Mpps to 18.6Mpps. Reviewed by: emaste, philip, melifaro MFC after: 2 months Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24250
2020-04-26 16:22:35 +00:00
int used = 0;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
Call bridge_span before dropping the lock. MFC after: 5 days
2006-08-17 10:18:31 +00:00
bridge_span(sc, m);
bridge: Change lists to CK_LIST as a peparation for epochification Prepare the ground for a rework of the bridge locking approach. We will use an epoch-based approach in the datapath and making it safe to iterate over the interface, span and rtnode lists without holding the BRIDGE_LOCK. Replace the relevant lists by their ConcurrencyKit equivalents. No functional change in this commit. Reviewed by: emaste, ae, philip (previous version) Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24249
2020-04-05 17:15:20 +00:00
CK_LIST_FOREACH(bif, &sc->sc_iflist, bif_next) {
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
dst_if = bif->bif_ifp;
Add RFC 3378 EtherIP support. This change makes it possible to add gif interfaces to bridges, which will then send and receive IP protocol 97 packets. Packets are Ethernet frames with an EtherIP header prepended. Obtained from: NetBSD MFC after: 2 weeks
2005-12-21 21:29:45 +00:00
if (dst_if->if_type == IFT_GIF)
continue;
Propagate rename of IFF_OACTIVE and IFF_RUNNING to IFF_DRV_OACTIVE and IFF_DRV_RUNNING, as well as the move from ifnet.if_flags to ifnet.if_drv_flags. Device drivers are now responsible for synchronizing access to these flags, as they are in if_drv_flags. This helps prevent races between the network stack and device driver in maintaining the interface flags field. Many __FreeBSD__ and __FreeBSD_version checks maintained and continued; some less so. Reviewed by: pjd, bz MFC after: 7 days
2005-08-09 10:20:02 +00:00
if ((dst_if->if_drv_flags & IFF_DRV_RUNNING) == 0)
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
continue;
/*
* If this is not the original output interface,
* and the interface is participating in spanning
* tree, make sure the port is in a state that
* allows forwarding.
*/
Bring in support for the Rapid Spanning Tree Protocol (802.1w). RSTP provides faster spanning tree convergence, the protocol will exchange information with neighboring switches to quickly transition to forwarding without creating loops. The code will default to RSTP mode but will downgrade any port connected to a legacy STP network so is fully backward compatible. Reviewed by: syrinx Tested by: syrinx
2006-11-01 09:07:47 +00:00
if (dst_if != ifp && (bif->bif_flags & IFBIF_STP) &&
bif->bif_stp.bp_state == BSTP_IFSTATE_DISCARDING)
continue;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge: Change lists to CK_LIST as a peparation for epochification Prepare the ground for a rework of the bridge locking approach. We will use an epoch-based approach in the datapath and making it safe to iterate over the interface, span and rtnode lists without holding the BRIDGE_LOCK. Replace the relevant lists by their ConcurrencyKit equivalents. No functional change in this commit. Reviewed by: emaste, ae, philip (previous version) Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24249
2020-04-05 17:15:20 +00:00
if (CK_LIST_NEXT(bif, bif_next) == NULL) {
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
used = 1;
mc = m;
} else {
bridge: Don't share broadcast packets if_bridge duplicates broadcast packets with m_copypacket(), which creates shared packets. In certain circumstances these packets can be processed by udp_usrreq.c:udp_input() first, which modifies the mbuf as part of the checksum verification. That may lead to incorrect packets being transmitted. Use m_dup() to create independent mbufs instead. Reported by: Richard Russo <toast@ruka.org> Reviewed by: donner, afedorov MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D34319
2022-02-19 15:34:31 +00:00
mc = m_dup(m, M_NOWAIT);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
if (mc == NULL) {
if_bridge(4): Complete bpf auditing of local traffic over the bridge There were two remaining "gaps" in auditing local bridge traffic with bpf(4): Locally originated outbound traffic from a member interface is invisible to the bridge's bpf(4) interface. Inbound traffic locally destined to a member interface is invisible to the member's bpf(4) interface -- this traffic has no chance after bridge_input to otherwise pass it over, and it wasn't originally received on this interface. I call these "gaps" because they don't affect conventional bridge setups. Alas, being able to establish an audit trail of all locally destined traffic for setups that can function like this is useful in some scenarios. Reviewed by: kp MFC after: 1 week Differential Revision: https://reviews.freebsd.org/D19757
2019-05-29 01:08:30 +00:00
if_inc_counter(bifp, IFCOUNTER_OERRORS, 1);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
continue;
}
}
- Previously when broadcasting to N number of interfaces we would run pfil hooks for each outgoing interface but also run pfil hooks _N times_ on the bridge interface. This is changed so pfil hooks are run once for the bridge interface (bridge0) and then only on the outgoing interfaces in the broadcast loop. - Simplify bridge_enqueue() by moving bridge_pfil() to the callers. - Check (inet6_pfil_hook.ph_busy_count >= 0), it may be possible to have a packet filter hooked for only ipv6 but we were only checking if ipv4 hooks were busy. - Minor optimisation for null mbuf check after bridge_pfil(), move it into the if-block as it couldnt possibly be null outside. Prodded by: mlaier Approved by: re (scottl), mlaier (mentor)
2005-07-06 01:24:45 +00:00
bridge_enqueue(sc, dst_if, mc);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
}
if (used == 0)
m_freem(m);
return (0);
}
Minor whitespace cleanup.
2005-12-17 10:03:48 +00:00
sendunicast:
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
/*
* XXX Spanning tree consideration here?
*/
Add support for creating span ports so that one can snoop bridged traffic from another interface/machine/network. Obtained from: OpenBSD MFC after: 2 weeks
2005-12-14 02:52:13 +00:00
bridge_span(sc, m);
Propagate rename of IFF_OACTIVE and IFF_RUNNING to IFF_DRV_OACTIVE and IFF_DRV_RUNNING, as well as the move from ifnet.if_flags to ifnet.if_drv_flags. Device drivers are now responsible for synchronizing access to these flags, as they are in if_drv_flags. This helps prevent races between the network stack and device driver in maintaining the interface flags field. Many __FreeBSD__ and __FreeBSD_version checks maintained and continued; some less so. Reviewed by: pjd, bz MFC after: 7 days
2005-08-09 10:20:02 +00:00
if ((dst_if->if_drv_flags & IFF_DRV_RUNNING) == 0) {
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
m_freem(m);
return (0);
}
- Previously when broadcasting to N number of interfaces we would run pfil hooks for each outgoing interface but also run pfil hooks _N times_ on the bridge interface. This is changed so pfil hooks are run once for the bridge interface (bridge0) and then only on the outgoing interfaces in the broadcast loop. - Simplify bridge_enqueue() by moving bridge_pfil() to the callers. - Check (inet6_pfil_hook.ph_busy_count >= 0), it may be possible to have a packet filter hooked for only ipv6 but we were only checking if ipv4 hooks were busy. - Minor optimisation for null mbuf check after bridge_pfil(), move it into the if-block as it couldnt possibly be null outside. Prodded by: mlaier Approved by: re (scottl), mlaier (mentor)
2005-07-06 01:24:45 +00:00
bridge_enqueue(sc, dst_if, m);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
return (0);
}
/*
Change bridge(4) to use if_transmit for forwarding packets to underlying interfaces instead of queueing. Tested by: ray
2012-09-03 10:08:20 +00:00
* bridge_transmit:
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
*
Change bridge(4) to use if_transmit for forwarding packets to underlying interfaces instead of queueing. Tested by: ray
2012-09-03 10:08:20 +00:00
* Do output on a bridge.
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
*
*/
Change bridge(4) to use if_transmit for forwarding packets to underlying interfaces instead of queueing. Tested by: ray
2012-09-03 10:08:20 +00:00
static int
bridge_transmit(struct ifnet *ifp, struct mbuf *m)
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
{
struct bridge_softc *sc;
Fix fallout from r240071. If destination interface lookup fails, we should broadcast a packet, not try to deliver it to NULL. Reported by: rpaulo
2012-10-24 18:33:44 +00:00
struct ether_header *eh;
struct ifnet *dst_if;
Change bridge(4) to use if_transmit for forwarding packets to underlying interfaces instead of queueing. Tested by: ray
2012-09-03 10:08:20 +00:00
int error = 0;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
sc = ifp->if_softc;
Change bridge(4) to use if_transmit for forwarding packets to underlying interfaces instead of queueing. Tested by: ray
2012-09-03 10:08:20 +00:00
ETHER_BPF_MTAP(ifp, m);
Fix fallout from r240071. If destination interface lookup fails, we should broadcast a packet, not try to deliver it to NULL. Reported by: rpaulo
2012-10-24 18:33:44 +00:00
eh = mtod(m, struct ether_header *);
Change bridge(4) to use if_transmit for forwarding packets to underlying interfaces instead of queueing. Tested by: ray
2012-09-03 10:08:20 +00:00
Fix fallout from r240071. If destination interface lookup fails, we should broadcast a packet, not try to deliver it to NULL. Reported by: rpaulo
2012-10-24 18:33:44 +00:00
if (((m->m_flags & (M_BCAST|M_MCAST)) == 0) &&
(dst_if = bridge_rtlookup(sc, eh->ether_dhost, 1)) != NULL) {
Change bridge(4) to use if_transmit for forwarding packets to underlying interfaces instead of queueing. Tested by: ray
2012-09-03 10:08:20 +00:00
error = bridge_enqueue(sc, dst_if, m);
} else
bridge_broadcast(sc, ifp, m, 0);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
Change bridge(4) to use if_transmit for forwarding packets to underlying interfaces instead of queueing. Tested by: ray
2012-09-03 10:08:20 +00:00
return (error);
}
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
if_bridge: add ALTQ support Similar to the recent addition of ALTQ support to if_vlan. Reviewed by: donner Obtained from: pfsense MFC after: 1 week Sponsored by: Rubicon Communications, LLC ("Netgate") Differential Revision: https://reviews.freebsd.org/D31675
2021-08-25 09:07:44 +00:00
#ifdef ALTQ
static void
bridge_altq_start(if_t ifp)
{
struct ifaltq *ifq = &ifp->if_snd;
struct mbuf *m;
IFQ_LOCK(ifq);
IFQ_DEQUEUE_NOLOCK(ifq, m);
while (m != NULL) {
bridge_transmit(ifp, m);
IFQ_DEQUEUE_NOLOCK(ifq, m);
}
IFQ_UNLOCK(ifq);
}
static int
bridge_altq_transmit(if_t ifp, struct mbuf *m)
{
int err;
if (ALTQ_IS_ENABLED(&ifp->if_snd)) {
IFQ_ENQUEUE(&ifp->if_snd, m, err);
if (err == 0)
bridge_altq_start(ifp);
} else
err = bridge_transmit(ifp, m);
return (err);
}
#endif /* ALTQ */
Change bridge(4) to use if_transmit for forwarding packets to underlying interfaces instead of queueing. Tested by: ray
2012-09-03 10:08:20 +00:00
/*
* The ifp->if_qflush entry point for if_bridge(4) is no-op.
*/
static void
bridge_qflush(struct ifnet *ifp __unused)
{
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
}
/*
* bridge_forward:
*
* The forwarding function of the bridge.
*
* NOTE: Releases the lock on return.
*/
Change most of the bridge and stp funtions to static. This has highlighted that the following funtions are not used, wrap in '#ifdef noused' for the moment. bstp_enable_change_detection bstp_disable_change_detection bstp_set_bridge_priority bstp_set_port_priority bstp_set_path_cost
2005-10-14 10:38:12 +00:00
static void
Add a bridge interface flag called PRIVATE where any private port can not communicate with another private port. All unicast/broadcast/multicast layer2 traffic is blocked so it works much the same way as using firewall rules but scales better and is generally easier as firewall packages usually do not allow ARP blocking. An example usage would be having a number of customers on separate vlans bridged with a server network. All the vlans are marked private, they can all communicate with the server network unhindered, but can not exchange any traffic whatsoever with each other. Approved by: re (rwatson)
2007-08-01 00:33:52 +00:00
bridge_forward(struct bridge_softc *sc, struct bridge_iflist *sbif,
struct mbuf *m)
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
{
Add a bridge interface flag called PRIVATE where any private port can not communicate with another private port. All unicast/broadcast/multicast layer2 traffic is blocked so it works much the same way as using firewall rules but scales better and is generally easier as firewall packages usually do not allow ARP blocking. An example usage would be having a number of customers on separate vlans bridged with a server network. All the vlans are marked private, they can all communicate with the server network unhindered, but can not exchange any traffic whatsoever with each other. Approved by: re (rwatson)
2007-08-01 00:33:52 +00:00
struct bridge_iflist *dbif;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
struct ifnet *src_if, *dst_if, *ifp;
struct ether_header *eh;
Add the vlan tag to the bridge route table. This allows a vlan trunk to be bridged, previously legitimate traffic was not passed as the bridge could not tell that it was on a different Ethernet segment. All non-tagged traffic is treated as vlan1 as per IEEE 802.1Q-2003
2007-06-13 18:58:04 +00:00
uint16_t vlan;
IEEE 802.1D-2004 states, frames containing any of the group MAC Addresses specified in Table 7-10 in their destination address field shall not be relayed by the Bridge. Add a check in bridge_forward() to adhere to this. PR: kern/119744
2008-01-18 00:19:10 +00:00
uint8_t *dst;
Add an option to limit the number of source MACs that can be behind a bridge interface. Once the limit is reached packets with unknown source addresses are dropped until an existing host cache entry expires or is removed. Useful to use with the STICKY cache option. Sponsored by: miniSuperHappyDevHouse NZ
2007-11-04 08:32:27 +00:00
int error;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge: epoch-ification Run the bridge datapath under epoch, rather than under the BRIDGE_LOCK(). We still take the BRIDGE_LOCK() whenever we insert or delete items in the relevant lists, but we use epoch callbacks to free items so that it's safe to iterate the lists without the BRIDGE_LOCK. Tests on mercat5/6 shows this increases bridge throughput significantly, from 3.7Mpps to 18.6Mpps. Reviewed by: emaste, philip, melifaro MFC after: 2 months Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24250
2020-04-26 16:22:35 +00:00
NET_EPOCH_ASSERT();
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
src_if = m->m_pkthdr.rcvif;
Stop embedding struct ifnet at the top of driver softcs. Instead the struct ifnet or the layer 2 common structure it was embedded in have been replaced with a struct ifnet pointer to be filled by a call to the new function, if_alloc(). The layer 2 common structure is also allocated via if_alloc() based on the interface type. It is hung off the new struct ifnet member, if_l2com. This change removes the size of these structures from the kernel ABI and will allow us to better manage them as interfaces come and go. Other changes of note: - Struct arpcom is no longer referenced in normal interface code. Instead the Ethernet address is accessed via the IFP2ENADDR() macro. To enforce this ac_enaddr has been renamed to _ac_enaddr. - The second argument to ether_ifattach is now always the mac address from driver private storage rather than sometimes being ac_enaddr. Reviewed by: sobomax, sam
2005-06-10 16:49:24 +00:00
ifp = sc->sc_ifp;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
Mechanically convert to if_inc_counter().
2014-09-19 10:39:58 +00:00
if_inc_counter(ifp, IFCOUNTER_IPACKETS, 1);
if_inc_counter(ifp, IFCOUNTER_IBYTES, m->m_pkthdr.len);
Add the vlan tag to the bridge route table. This allows a vlan trunk to be bridged, previously legitimate traffic was not passed as the bridge could not tell that it was on a different Ethernet segment. All non-tagged traffic is treated as vlan1 as per IEEE 802.1Q-2003
2007-06-13 18:58:04 +00:00
vlan = VLANTAGOF(m);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
Add a bridge interface flag called PRIVATE where any private port can not communicate with another private port. All unicast/broadcast/multicast layer2 traffic is blocked so it works much the same way as using firewall rules but scales better and is generally easier as firewall packages usually do not allow ARP blocking. An example usage would be having a number of customers on separate vlans bridged with a server network. All the vlans are marked private, they can all communicate with the server network unhindered, but can not exchange any traffic whatsoever with each other. Approved by: re (rwatson)
2007-08-01 00:33:52 +00:00
if ((sbif->bif_flags & IFBIF_STP) &&
Simplify the error handling and use the dereferenced sc->sc_ifp pointer.
2007-12-18 09:13:04 +00:00
sbif->bif_stp.bp_state == BSTP_IFSTATE_DISCARDING)
goto drop;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
eh = mtod(m, struct ether_header *);
IEEE 802.1D-2004 states, frames containing any of the group MAC Addresses specified in Table 7-10 in their destination address field shall not be relayed by the Bridge. Add a check in bridge_forward() to adhere to this. PR: kern/119744
2008-01-18 00:19:10 +00:00
dst = eh->ether_dhost;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
Add an option to limit the number of source MACs that can be behind a bridge interface. Once the limit is reached packets with unknown source addresses are dropped until an existing host cache entry expires or is removed. Useful to use with the STICKY cache option. Sponsored by: miniSuperHappyDevHouse NZ
2007-11-04 08:32:27 +00:00
/* If the interface is learning, record the address. */
if (sbif->bif_flags & IFBIF_LEARNING) {
error = bridge_rtupdate(sc, eh->ether_shost, vlan,
Add a bridge interface flag called PRIVATE where any private port can not communicate with another private port. All unicast/broadcast/multicast layer2 traffic is blocked so it works much the same way as using firewall rules but scales better and is generally easier as firewall packages usually do not allow ARP blocking. An example usage would be having a number of customers on separate vlans bridged with a server network. All the vlans are marked private, they can all communicate with the server network unhindered, but can not exchange any traffic whatsoever with each other. Approved by: re (rwatson)
2007-08-01 00:33:52 +00:00
sbif, 0, IFBAF_DYNAMIC);
Add an option to limit the number of source MACs that can be behind a bridge interface. Once the limit is reached packets with unknown source addresses are dropped until an existing host cache entry expires or is removed. Useful to use with the STICKY cache option. Sponsored by: miniSuperHappyDevHouse NZ
2007-11-04 08:32:27 +00:00
/*
* If the interface has addresses limits then deny any source
* that is not in the cache.
*/
Simplify the error handling and use the dereferenced sc->sc_ifp pointer.
2007-12-18 09:13:04 +00:00
if (error && sbif->bif_addrmax)
goto drop;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
}
Add a bridge interface flag called PRIVATE where any private port can not communicate with another private port. All unicast/broadcast/multicast layer2 traffic is blocked so it works much the same way as using firewall rules but scales better and is generally easier as firewall packages usually do not allow ARP blocking. An example usage would be having a number of customers on separate vlans bridged with a server network. All the vlans are marked private, they can all communicate with the server network unhindered, but can not exchange any traffic whatsoever with each other. Approved by: re (rwatson)
2007-08-01 00:33:52 +00:00
if ((sbif->bif_flags & IFBIF_STP) != 0 &&
Simplify the error handling and use the dereferenced sc->sc_ifp pointer.
2007-12-18 09:13:04 +00:00
sbif->bif_stp.bp_state == BSTP_IFSTATE_LEARNING)
goto drop;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
/*
* At this point, the port either doesn't participate
* in spanning tree or it is in the forwarding state.
*/
/*
* If the packet is unicast, destined for someone on
* "this" side of the bridge, drop it.
*/
if ((m->m_flags & (M_BCAST|M_MCAST)) == 0) {
IEEE 802.1D-2004 states, frames containing any of the group MAC Addresses specified in Table 7-10 in their destination address field shall not be relayed by the Bridge. Add a check in bridge_forward() to adhere to this. PR: kern/119744
2008-01-18 00:19:10 +00:00
dst_if = bridge_rtlookup(sc, dst, vlan);
Simplify the error handling and use the dereferenced sc->sc_ifp pointer.
2007-12-18 09:13:04 +00:00
if (src_if == dst_if)
goto drop;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
} else {
IEEE 802.1D-2004 states, frames containing any of the group MAC Addresses specified in Table 7-10 in their destination address field shall not be relayed by the Bridge. Add a check in bridge_forward() to adhere to this. PR: kern/119744
2008-01-18 00:19:10 +00:00
/*
* Check if its a reserved multicast address, any address
* listed in 802.1D section 7.12.6 may not be forwarded by the
* bridge.
* This is currently 01-80-C2-00-00-00 to 01-80-C2-00-00-0F
*/
if (dst[0] == 0x01 && dst[1] == 0x80 &&
dst[2] == 0xc2 && dst[3] == 0x00 &&
dst[4] == 0x00 && dst[5] <= 0x0f)
goto drop;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
/* ...forward it to all interfaces. */
Mechanically convert to if_inc_counter().
2014-09-19 10:39:58 +00:00
if_inc_counter(ifp, IFCOUNTER_IMCASTS, 1);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
dst_if = NULL;
}
Slightly re-worked bpf(4) code associated with bridging: if we have a destination interface as a member of our bridge or this is a unicast packet, push it through the bpf(4) machinery. For broadcast or multicast packets, don't bother with the bpf(4) because it will be re-injected into ether_input. We do this before we pass the packets through the pfil(9) framework, as it is possible that pfil(9) will drop the packet or possibly modify it, making it very difficult to debug firewall issues on the bridge. Further, implemented IFF_MONITOR for bridge interfaces. This does much the same thing that it does for regular network interfaces: it pushes the packet to any bpf(4) peers and then returns. This bypasses all of the bridge machinery, saving mutex acquisitions, list traversals, and other operations performed by the bridging code. This change to the bridging code is useful in situations where individuals use a bridge to multiplex RX/TX signals from two interfaces, as is required by some network taps for de-multiplexing links and transmitting the RX/TX signals out through two separate interfaces. This behaviour is quite common for network taps monitoring links, especially for certain manufacturers. Reviewed by: thompsa MFC after: 1 month Sponsored by: Seccuris Labs
2006-03-03 05:58:18 +00:00
/*
* If we have a destination interface which is a member of our bridge,
* OR this is a unicast packet, push it through the bpf(4) machinery.
* For broadcast or multicast packets, don't bother because it will
* be reinjected into ether_input. We do this before we pass the packets
* through the pfil(9) framework, as it is possible that pfil(9) will
* drop the packet, or possibly modify it, making it difficult to debug
* firewall issues on the bridge.
*/
if (dst_if != NULL || (m->m_flags & (M_BCAST | M_MCAST)) == 0)
Use ETHER_BPF_MTAP so that the vlan tags are visible to bpf(4) when bridging a vlan trunk. Discussed with: csjp MFC after: 3 days
2007-10-20 02:10:10 +00:00
ETHER_BPF_MTAP(ifp, m);
Slightly re-worked bpf(4) code associated with bridging: if we have a destination interface as a member of our bridge or this is a unicast packet, push it through the bpf(4) machinery. For broadcast or multicast packets, don't bother with the bpf(4) because it will be re-injected into ether_input. We do this before we pass the packets through the pfil(9) framework, as it is possible that pfil(9) will drop the packet or possibly modify it, making it very difficult to debug firewall issues on the bridge. Further, implemented IFF_MONITOR for bridge interfaces. This does much the same thing that it does for regular network interfaces: it pushes the packet to any bpf(4) peers and then returns. This bypasses all of the bridge machinery, saving mutex acquisitions, list traversals, and other operations performed by the bridging code. This change to the bridging code is useful in situations where individuals use a bridge to multiplex RX/TX signals from two interfaces, as is required by some network taps for de-multiplexing links and transmitting the RX/TX signals out through two separate interfaces. This behaviour is quite common for network taps monitoring links, especially for certain manufacturers. Reviewed by: thompsa MFC after: 1 month Sponsored by: Seccuris Labs
2006-03-03 05:58:18 +00:00
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
/* run the packet filter */
New pfil(9) KPI together with newborn pfil API and control utility. The KPI have been reviewed and cleansed of features that were planned back 20 years ago and never implemented. The pfil(9) internals have been made opaque to protocols with only returned types and function declarations exposed. The KPI is made more strict, but at the same time more extensible, as kernel uses same command structures that userland ioctl uses. In nutshell [KA]PI is about declaring filtering points, declaring filters and linking and unlinking them together. New [KA]PI makes it possible to reconfigure pfil(9) configuration: change order of hooks, rehook filter from one filtering point to a different one, disconnect a hook on output leaving it on input only, prepend/append a filter to existing list of filters. Now it possible for a single packet filter to provide multiple rulesets that may be linked to different points. Think of per-interface ACLs in Cisco or Juniper. None of existing packet filters yet support that, however limited usage is already possible, e.g. default ruleset can be moved to single interface, as soon as interface would pride their filtering points. Another future feature is possiblity to create pfil heads, that provide not an mbuf pointer but just a memory pointer with length. That would allow filtering at very early stages of a packet lifecycle, e.g. when packet has just been received by a NIC and no mbuf was yet allocated. Differential Revision: https://reviews.freebsd.org/D18951
2019-01-31 23:01:03 +00:00
if (PFIL_HOOKED_IN(V_inet_pfil_head)
- Previously when broadcasting to N number of interfaces we would run pfil hooks for each outgoing interface but also run pfil hooks _N times_ on the bridge interface. This is changed so pfil hooks are run once for the bridge interface (bridge0) and then only on the outgoing interfaces in the broadcast loop. - Simplify bridge_enqueue() by moving bridge_pfil() to the callers. - Check (inet6_pfil_hook.ph_busy_count >= 0), it may be possible to have a packet filter hooked for only ipv6 but we were only checking if ipv4 hooks were busy. - Minor optimisation for null mbuf check after bridge_pfil(), move it into the if-block as it couldnt possibly be null outside. Prodded by: mlaier Approved by: re (scottl), mlaier (mentor)
2005-07-06 01:24:45 +00:00
#ifdef INET6
New pfil(9) KPI together with newborn pfil API and control utility. The KPI have been reviewed and cleansed of features that were planned back 20 years ago and never implemented. The pfil(9) internals have been made opaque to protocols with only returned types and function declarations exposed. The KPI is made more strict, but at the same time more extensible, as kernel uses same command structures that userland ioctl uses. In nutshell [KA]PI is about declaring filtering points, declaring filters and linking and unlinking them together. New [KA]PI makes it possible to reconfigure pfil(9) configuration: change order of hooks, rehook filter from one filtering point to a different one, disconnect a hook on output leaving it on input only, prepend/append a filter to existing list of filters. Now it possible for a single packet filter to provide multiple rulesets that may be linked to different points. Think of per-interface ACLs in Cisco or Juniper. None of existing packet filters yet support that, however limited usage is already possible, e.g. default ruleset can be moved to single interface, as soon as interface would pride their filtering points. Another future feature is possiblity to create pfil heads, that provide not an mbuf pointer but just a memory pointer with length. That would allow filtering at very early stages of a packet lifecycle, e.g. when packet has just been received by a NIC and no mbuf was yet allocated. Differential Revision: https://reviews.freebsd.org/D18951
2019-01-31 23:01:03 +00:00
|| PFIL_HOOKED_IN(V_inet6_pfil_head)
- Previously when broadcasting to N number of interfaces we would run pfil hooks for each outgoing interface but also run pfil hooks _N times_ on the bridge interface. This is changed so pfil hooks are run once for the bridge interface (bridge0) and then only on the outgoing interfaces in the broadcast loop. - Simplify bridge_enqueue() by moving bridge_pfil() to the callers. - Check (inet6_pfil_hook.ph_busy_count >= 0), it may be possible to have a packet filter hooked for only ipv6 but we were only checking if ipv4 hooks were busy. - Minor optimisation for null mbuf check after bridge_pfil(), move it into the if-block as it couldnt possibly be null outside. Prodded by: mlaier Approved by: re (scottl), mlaier (mentor)
2005-07-06 01:24:45 +00:00
#endif
) {
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
if (bridge_pfil(&m, ifp, src_if, PFIL_IN) != 0)
return;
- Previously when broadcasting to N number of interfaces we would run pfil hooks for each outgoing interface but also run pfil hooks _N times_ on the bridge interface. This is changed so pfil hooks are run once for the bridge interface (bridge0) and then only on the outgoing interfaces in the broadcast loop. - Simplify bridge_enqueue() by moving bridge_pfil() to the callers. - Check (inet6_pfil_hook.ph_busy_count >= 0), it may be possible to have a packet filter hooked for only ipv6 but we were only checking if ipv4 hooks were busy. - Minor optimisation for null mbuf check after bridge_pfil(), move it into the if-block as it couldnt possibly be null outside. Prodded by: mlaier Approved by: re (scottl), mlaier (mentor)
2005-07-06 01:24:45 +00:00
if (m == NULL)
return;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
}
if (dst_if == NULL) {
Do not packet filter in the bridge_start() routine, locally generated packets are already filtered by the higher layers. Approved by: mlaier (mentor) MFC after: 3 days
2005-10-02 19:15:56 +00:00
bridge_broadcast(sc, src_if, m, 1);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
return;
}
/*
* At this point, we're dealing with a unicast frame
* going to a different interface.
*/
Simplify the error handling and use the dereferenced sc->sc_ifp pointer.
2007-12-18 09:13:04 +00:00
if ((dst_if->if_drv_flags & IFF_DRV_RUNNING) == 0)
goto drop;
Add a bridge interface flag called PRIVATE where any private port can not communicate with another private port. All unicast/broadcast/multicast layer2 traffic is blocked so it works much the same way as using firewall rules but scales better and is generally easier as firewall packages usually do not allow ARP blocking. An example usage would be having a number of customers on separate vlans bridged with a server network. All the vlans are marked private, they can all communicate with the server network unhindered, but can not exchange any traffic whatsoever with each other. Approved by: re (rwatson)
2007-08-01 00:33:52 +00:00
dbif = bridge_lookup_member_if(sc, dst_if);
Simplify the error handling and use the dereferenced sc->sc_ifp pointer.
2007-12-18 09:13:04 +00:00
if (dbif == NULL)
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
/* Not a member of the bridge (anymore?) */
Simplify the error handling and use the dereferenced sc->sc_ifp pointer.
2007-12-18 09:13:04 +00:00
goto drop;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
Add a bridge interface flag called PRIVATE where any private port can not communicate with another private port. All unicast/broadcast/multicast layer2 traffic is blocked so it works much the same way as using firewall rules but scales better and is generally easier as firewall packages usually do not allow ARP blocking. An example usage would be having a number of customers on separate vlans bridged with a server network. All the vlans are marked private, they can all communicate with the server network unhindered, but can not exchange any traffic whatsoever with each other. Approved by: re (rwatson)
2007-08-01 00:33:52 +00:00
/* Private segments can not talk to each other */
Simplify the error handling and use the dereferenced sc->sc_ifp pointer.
2007-12-18 09:13:04 +00:00
if (sbif->bif_flags & dbif->bif_flags & IFBIF_PRIVATE)
goto drop;
Add a bridge interface flag called PRIVATE where any private port can not communicate with another private port. All unicast/broadcast/multicast layer2 traffic is blocked so it works much the same way as using firewall rules but scales better and is generally easier as firewall packages usually do not allow ARP blocking. An example usage would be having a number of customers on separate vlans bridged with a server network. All the vlans are marked private, they can all communicate with the server network unhindered, but can not exchange any traffic whatsoever with each other. Approved by: re (rwatson)
2007-08-01 00:33:52 +00:00
if ((dbif->bif_flags & IFBIF_STP) &&
Simplify the error handling and use the dereferenced sc->sc_ifp pointer.
2007-12-18 09:13:04 +00:00
dbif->bif_stp.bp_state == BSTP_IFSTATE_DISCARDING)
goto drop;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
New pfil(9) KPI together with newborn pfil API and control utility. The KPI have been reviewed and cleansed of features that were planned back 20 years ago and never implemented. The pfil(9) internals have been made opaque to protocols with only returned types and function declarations exposed. The KPI is made more strict, but at the same time more extensible, as kernel uses same command structures that userland ioctl uses. In nutshell [KA]PI is about declaring filtering points, declaring filters and linking and unlinking them together. New [KA]PI makes it possible to reconfigure pfil(9) configuration: change order of hooks, rehook filter from one filtering point to a different one, disconnect a hook on output leaving it on input only, prepend/append a filter to existing list of filters. Now it possible for a single packet filter to provide multiple rulesets that may be linked to different points. Think of per-interface ACLs in Cisco or Juniper. None of existing packet filters yet support that, however limited usage is already possible, e.g. default ruleset can be moved to single interface, as soon as interface would pride their filtering points. Another future feature is possiblity to create pfil heads, that provide not an mbuf pointer but just a memory pointer with length. That would allow filtering at very early stages of a packet lifecycle, e.g. when packet has just been received by a NIC and no mbuf was yet allocated. Differential Revision: https://reviews.freebsd.org/D18951
2019-01-31 23:01:03 +00:00
if (PFIL_HOOKED_OUT(V_inet_pfil_head)
- Previously when broadcasting to N number of interfaces we would run pfil hooks for each outgoing interface but also run pfil hooks _N times_ on the bridge interface. This is changed so pfil hooks are run once for the bridge interface (bridge0) and then only on the outgoing interfaces in the broadcast loop. - Simplify bridge_enqueue() by moving bridge_pfil() to the callers. - Check (inet6_pfil_hook.ph_busy_count >= 0), it may be possible to have a packet filter hooked for only ipv6 but we were only checking if ipv4 hooks were busy. - Minor optimisation for null mbuf check after bridge_pfil(), move it into the if-block as it couldnt possibly be null outside. Prodded by: mlaier Approved by: re (scottl), mlaier (mentor)
2005-07-06 01:24:45 +00:00
#ifdef INET6
New pfil(9) KPI together with newborn pfil API and control utility. The KPI have been reviewed and cleansed of features that were planned back 20 years ago and never implemented. The pfil(9) internals have been made opaque to protocols with only returned types and function declarations exposed. The KPI is made more strict, but at the same time more extensible, as kernel uses same command structures that userland ioctl uses. In nutshell [KA]PI is about declaring filtering points, declaring filters and linking and unlinking them together. New [KA]PI makes it possible to reconfigure pfil(9) configuration: change order of hooks, rehook filter from one filtering point to a different one, disconnect a hook on output leaving it on input only, prepend/append a filter to existing list of filters. Now it possible for a single packet filter to provide multiple rulesets that may be linked to different points. Think of per-interface ACLs in Cisco or Juniper. None of existing packet filters yet support that, however limited usage is already possible, e.g. default ruleset can be moved to single interface, as soon as interface would pride their filtering points. Another future feature is possiblity to create pfil heads, that provide not an mbuf pointer but just a memory pointer with length. That would allow filtering at very early stages of a packet lifecycle, e.g. when packet has just been received by a NIC and no mbuf was yet allocated. Differential Revision: https://reviews.freebsd.org/D18951
2019-01-31 23:01:03 +00:00
|| PFIL_HOOKED_OUT(V_inet6_pfil_head)
- Previously when broadcasting to N number of interfaces we would run pfil hooks for each outgoing interface but also run pfil hooks _N times_ on the bridge interface. This is changed so pfil hooks are run once for the bridge interface (bridge0) and then only on the outgoing interfaces in the broadcast loop. - Simplify bridge_enqueue() by moving bridge_pfil() to the callers. - Check (inet6_pfil_hook.ph_busy_count >= 0), it may be possible to have a packet filter hooked for only ipv6 but we were only checking if ipv4 hooks were busy. - Minor optimisation for null mbuf check after bridge_pfil(), move it into the if-block as it couldnt possibly be null outside. Prodded by: mlaier Approved by: re (scottl), mlaier (mentor)
2005-07-06 01:24:45 +00:00
#endif
) {
Simplify the error handling and use the dereferenced sc->sc_ifp pointer.
2007-12-18 09:13:04 +00:00
if (bridge_pfil(&m, ifp, dst_if, PFIL_OUT) != 0)
- Previously when broadcasting to N number of interfaces we would run pfil hooks for each outgoing interface but also run pfil hooks _N times_ on the bridge interface. This is changed so pfil hooks are run once for the bridge interface (bridge0) and then only on the outgoing interfaces in the broadcast loop. - Simplify bridge_enqueue() by moving bridge_pfil() to the callers. - Check (inet6_pfil_hook.ph_busy_count >= 0), it may be possible to have a packet filter hooked for only ipv6 but we were only checking if ipv4 hooks were busy. - Minor optimisation for null mbuf check after bridge_pfil(), move it into the if-block as it couldnt possibly be null outside. Prodded by: mlaier Approved by: re (scottl), mlaier (mentor)
2005-07-06 01:24:45 +00:00
return;
if (m == NULL)
return;
}
bridge_enqueue(sc, dst_if, m);
Simplify the error handling and use the dereferenced sc->sc_ifp pointer.
2007-12-18 09:13:04 +00:00
return;
drop:
m_freem(m);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
}
/*
* bridge_input:
*
* Receive input from a member interface. Queue the packet for
* bridging if it is not for us.
*/
Make four more functions static that were missed in the last commit.
2005-10-14 20:57:02 +00:00
static struct mbuf *
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge_input(struct ifnet *ifp, struct mbuf *m)
{
struct bridge_softc *sc = ifp->if_bridge;
Add a new address cache type called sticky. On an interface marked sticky any address learned by the bridge is made permanent, the address will not age out and most importantly will not migrate to another interface. This can be used to stop mac address poisoning or clients roaming in much the same way as static entries without the hassle of preloading the table.
2006-11-09 06:32:38 +00:00
struct bridge_iflist *bif, *bif2;
Add support for multicast to the bridge and allow inet6 addresses to be assigned to the interface. IPv6 auto-configuration is disabled. An IPv6 link-local address has a link-local scope within one link, the spec is unclear for the bridge case and it may cause scope violation. An address can be assigned in the usual way; ifconfig bridge0 inet6 xxxx:... Tested by: bmah Reviewed by: ume (netinet6) Approved by: mlaier (mentor) MFC after: 1 week
2005-09-06 21:11:59 +00:00
struct ifnet *bifp;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
struct ether_header *eh;
Add support for multicast to the bridge and allow inet6 addresses to be assigned to the interface. IPv6 auto-configuration is disabled. An IPv6 link-local address has a link-local scope within one link, the spec is unclear for the bridge case and it may cause scope violation. An address can be assigned in the usual way; ifconfig bridge0 inet6 xxxx:... Tested by: bmah Reviewed by: ume (netinet6) Approved by: mlaier (mentor) MFC after: 1 week
2005-09-06 21:11:59 +00:00
struct mbuf *mc, *mc2;
Add the vlan tag to the bridge route table. This allows a vlan trunk to be bridged, previously legitimate traffic was not passed as the bridge could not tell that it was on a different Ethernet segment. All non-tagged traffic is treated as vlan1 as per IEEE 802.1Q-2003
2007-06-13 18:58:04 +00:00
uint16_t vlan;
Add an option to limit the number of source MACs that can be behind a bridge interface. Once the limit is reached packets with unknown source addresses are dropped until an existing host cache entry expires or is removed. Useful to use with the STICKY cache option. Sponsored by: miniSuperHappyDevHouse NZ
2007-11-04 08:32:27 +00:00
int error;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge: epoch-ification Run the bridge datapath under epoch, rather than under the BRIDGE_LOCK(). We still take the BRIDGE_LOCK() whenever we insert or delete items in the relevant lists, but we use epoch callbacks to free items so that it's safe to iterate the lists without the BRIDGE_LOCK. Tests on mercat5/6 shows this increases bridge throughput significantly, from 3.7Mpps to 18.6Mpps. Reviewed by: emaste, philip, melifaro MFC after: 2 months Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24250
2020-04-26 16:22:35 +00:00
NET_EPOCH_ASSERT();
Propagate rename of IFF_OACTIVE and IFF_RUNNING to IFF_DRV_OACTIVE and IFF_DRV_RUNNING, as well as the move from ifnet.if_flags to ifnet.if_drv_flags. Device drivers are now responsible for synchronizing access to these flags, as they are in if_drv_flags. This helps prevent races between the network stack and device driver in maintaining the interface flags field. Many __FreeBSD__ and __FreeBSD_version checks maintained and continued; some less so. Reviewed by: pjd, bz MFC after: 7 days
2005-08-09 10:20:02 +00:00
if ((sc->sc_ifp->if_drv_flags & IFF_DRV_RUNNING) == 0)
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
return (m);
Add support for multicast to the bridge and allow inet6 addresses to be assigned to the interface. IPv6 auto-configuration is disabled. An IPv6 link-local address has a link-local scope within one link, the spec is unclear for the bridge case and it may cause scope violation. An address can be assigned in the usual way; ifconfig bridge0 inet6 xxxx:... Tested by: bmah Reviewed by: ume (netinet6) Approved by: mlaier (mentor) MFC after: 1 week
2005-09-06 21:11:59 +00:00
bifp = sc->sc_ifp;
Add the vlan tag to the bridge route table. This allows a vlan trunk to be bridged, previously legitimate traffic was not passed as the bridge could not tell that it was on a different Ethernet segment. All non-tagged traffic is treated as vlan1 as per IEEE 802.1Q-2003
2007-06-13 18:58:04 +00:00
vlan = VLANTAGOF(m);
Add support for multicast to the bridge and allow inet6 addresses to be assigned to the interface. IPv6 auto-configuration is disabled. An IPv6 link-local address has a link-local scope within one link, the spec is unclear for the bridge case and it may cause scope violation. An address can be assigned in the usual way; ifconfig bridge0 inet6 xxxx:... Tested by: bmah Reviewed by: ume (netinet6) Approved by: mlaier (mentor) MFC after: 1 week
2005-09-06 21:11:59 +00:00
Slightly re-worked bpf(4) code associated with bridging: if we have a destination interface as a member of our bridge or this is a unicast packet, push it through the bpf(4) machinery. For broadcast or multicast packets, don't bother with the bpf(4) because it will be re-injected into ether_input. We do this before we pass the packets through the pfil(9) framework, as it is possible that pfil(9) will drop the packet or possibly modify it, making it very difficult to debug firewall issues on the bridge. Further, implemented IFF_MONITOR for bridge interfaces. This does much the same thing that it does for regular network interfaces: it pushes the packet to any bpf(4) peers and then returns. This bypasses all of the bridge machinery, saving mutex acquisitions, list traversals, and other operations performed by the bridging code. This change to the bridging code is useful in situations where individuals use a bridge to multiplex RX/TX signals from two interfaces, as is required by some network taps for de-multiplexing links and transmitting the RX/TX signals out through two separate interfaces. This behaviour is quite common for network taps monitoring links, especially for certain manufacturers. Reviewed by: thompsa MFC after: 1 month Sponsored by: Seccuris Labs
2006-03-03 05:58:18 +00:00
/*
* Implement support for bridge monitoring. If this flag has been
* set on this interface, discard the packet once we push it through
* the bpf(4) machinery, but before we do, increment the byte and
* packet counters associated with this interface.
*/
if ((bifp->if_flags & IFF_MONITOR) != 0) {
m->m_pkthdr.rcvif = bifp;
Use ETHER_BPF_MTAP so that the vlan tags are visible to bpf(4) when bridging a vlan trunk. Discussed with: csjp MFC after: 3 days
2007-10-20 02:10:10 +00:00
ETHER_BPF_MTAP(bifp, m);
Mechanically convert to if_inc_counter().
2014-09-19 10:39:58 +00:00
if_inc_counter(bifp, IFCOUNTER_IPACKETS, 1);
if_inc_counter(bifp, IFCOUNTER_IBYTES, m->m_pkthdr.len);
Fix possible leak when bridge is in monitor mode. Use m_freem() which will free the entire chain, instead of using m_free() which will free just the mbuf that was passed. Discussed with: thompsa MFC after: 3 days
2006-11-05 17:56:25 +00:00
m_freem(m);
Slightly re-worked bpf(4) code associated with bridging: if we have a destination interface as a member of our bridge or this is a unicast packet, push it through the bpf(4) machinery. For broadcast or multicast packets, don't bother with the bpf(4) because it will be re-injected into ether_input. We do this before we pass the packets through the pfil(9) framework, as it is possible that pfil(9) will drop the packet or possibly modify it, making it very difficult to debug firewall issues on the bridge. Further, implemented IFF_MONITOR for bridge interfaces. This does much the same thing that it does for regular network interfaces: it pushes the packet to any bpf(4) peers and then returns. This bypasses all of the bridge machinery, saving mutex acquisitions, list traversals, and other operations performed by the bridging code. This change to the bridging code is useful in situations where individuals use a bridge to multiplex RX/TX signals from two interfaces, as is required by some network taps for de-multiplexing links and transmitting the RX/TX signals out through two separate interfaces. This behaviour is quite common for network taps monitoring links, especially for certain manufacturers. Reviewed by: thompsa MFC after: 1 month Sponsored by: Seccuris Labs
2006-03-03 05:58:18 +00:00
return (NULL);
}
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bif = bridge_lookup_member_if(sc, ifp);
if (bif == NULL) {
return (m);
}
eh = mtod(m, struct ether_header *);
Add support for creating span ports so that one can snoop bridged traffic from another interface/machine/network. Obtained from: OpenBSD MFC after: 2 weeks
2005-12-14 02:52:13 +00:00
bridge_span(sc, m);
M_BCAST & M_MCAST are now set by ether_input before passing to the bridge.
2007-03-19 19:13:02 +00:00
if (m->m_flags & (M_BCAST|M_MCAST)) {
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
/* Tap off 802.1D packets; they do not get forwarded. */
if (memcmp(eh->ether_dhost, bstp_etheraddr,
ETHER_ADDR_LEN) == 0) {
bstp_input() always consumes the packet so remove the mbuf handling dance around it. Obtained from: OpenBSD (r1.37)
2012-02-23 00:59:21 +00:00
bstp_input(&bif->bif_stp, ifp, m); /* consumes mbuf */
return (NULL);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
}
Bring in support for the Rapid Spanning Tree Protocol (802.1w). RSTP provides faster spanning tree convergence, the protocol will exchange information with neighboring switches to quickly transition to forwarding without creating loops. The code will default to RSTP mode but will downgrade any port connected to a legacy STP network so is fully backward compatible. Reviewed by: syrinx Tested by: syrinx
2006-11-01 09:07:47 +00:00
if ((bif->bif_flags & IFBIF_STP) &&
bif->bif_stp.bp_state == BSTP_IFSTATE_DISCARDING) {
return (m);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
}
/*
* Make a deep copy of the packet and enqueue the copy
* for bridge processing; return the original packet for
* local processing.
*/
Mechanically substitute flags from historic mbuf allocator with malloc(9) flags within sys. Exceptions: - sys/contrib not touched - sys/mbuf.h edited manually
2012-12-05 08:04:20 +00:00
mc = m_dup(m, M_NOWAIT);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
if (mc == NULL) {
return (m);
}
/* Perform the bridge forwarding function with the copy. */
Add a bridge interface flag called PRIVATE where any private port can not communicate with another private port. All unicast/broadcast/multicast layer2 traffic is blocked so it works much the same way as using firewall rules but scales better and is generally easier as firewall packages usually do not allow ARP blocking. An example usage would be having a number of customers on separate vlans bridged with a server network. All the vlans are marked private, they can all communicate with the server network unhindered, but can not exchange any traffic whatsoever with each other. Approved by: re (rwatson)
2007-08-01 00:33:52 +00:00
bridge_forward(sc, bif, mc);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
Add support for multicast to the bridge and allow inet6 addresses to be assigned to the interface. IPv6 auto-configuration is disabled. An IPv6 link-local address has a link-local scope within one link, the spec is unclear for the bridge case and it may cause scope violation. An address can be assigned in the usual way; ifconfig bridge0 inet6 xxxx:... Tested by: bmah Reviewed by: ume (netinet6) Approved by: mlaier (mentor) MFC after: 1 week
2005-09-06 21:11:59 +00:00
/*
* Reinject the mbuf as arriving on the bridge so we have a
* chance at claiming multicast packets. We can not loop back
* here from ether_input as a bridge is never a member of a
* bridge.
*/
KASSERT(bifp->if_bridge == NULL,
("loop created in bridge_input"));
Mechanically substitute flags from historic mbuf allocator with malloc(9) flags within sys. Exceptions: - sys/contrib not touched - sys/mbuf.h edited manually
2012-12-05 08:04:20 +00:00
mc2 = m_dup(m, M_NOWAIT);
It is not safe to use m_copypacket() here as the returned mbuf is readonly, change to m_dup and keep the alignment on the layer3 header. MFC after: 1 week
2005-12-15 19:34:39 +00:00
if (mc2 != NULL) {
/* Keep the layer3 header aligned */
int i = min(mc2->m_pkthdr.len, max_protohdr);
mc2 = m_copyup(mc2, i, ETHER_ALIGN);
}
Add support for multicast to the bridge and allow inet6 addresses to be assigned to the interface. IPv6 auto-configuration is disabled. An IPv6 link-local address has a link-local scope within one link, the spec is unclear for the bridge case and it may cause scope violation. An address can be assigned in the usual way; ifconfig bridge0 inet6 xxxx:... Tested by: bmah Reviewed by: ume (netinet6) Approved by: mlaier (mentor) MFC after: 1 week
2005-09-06 21:11:59 +00:00
if (mc2 != NULL) {
mc2->m_pkthdr.rcvif = bifp;
(*bifp->if_input)(bifp, mc2);
}
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
/* Return the original packet for local processing. */
return (m);
}
Bring in support for the Rapid Spanning Tree Protocol (802.1w). RSTP provides faster spanning tree convergence, the protocol will exchange information with neighboring switches to quickly transition to forwarding without creating loops. The code will default to RSTP mode but will downgrade any port connected to a legacy STP network so is fully backward compatible. Reviewed by: syrinx Tested by: syrinx
2006-11-01 09:07:47 +00:00
if ((bif->bif_flags & IFBIF_STP) &&
bif->bif_stp.bp_state == BSTP_IFSTATE_DISCARDING) {
return (m);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
}
Allow carp(4) to be loaded as a kernel module. Follow precedent set by bridge(4), lagg(4) etc. and make use of function pointers and pf_proto_register() to hook carp into the network stack. Currently, because of the uncertainty about whether the unload path is free of race condition panics, unloads are disallowed by default. Compiling with CARPMOD_CAN_UNLOAD in CFLAGS removes this anti foot shooting measure. This commit requires IP6PROTOSPACER, introduced in r211115. Reviewed by: bz, simon Approved by: ken (mentor) MFC after: 2 weeks
2010-08-11 00:51:50 +00:00
#if (defined(INET) || defined(INET6))
Give a chance for packet to appear with a correct input interfaces in case of multiple interfaces with the same MAC in the same bridge. This commit do not solve the entire problem. Only case where packet arrived from such interface. PR: kern/109815 MFC after: 7 days Submitted by: Eygene Ryabinkin and rik@ Discussed with: bms@, thompsa@, yar@
2007-03-18 23:28:53 +00:00
# define OR_CARP_CHECK_WE_ARE_DST(iface) \
|| ((iface)->if_carp \
Unbreak LINT by moving all carp hooks to net/if.c / netinet/ip_carp.h, with the appropriate ifdefs. Reviewed by: bz Approved by: ken (mentor)
2010-08-11 20:18:19 +00:00
&& (*carp_forus_p)((iface), eh->ether_dhost))
Give a chance for packet to appear with a correct input interfaces in case of multiple interfaces with the same MAC in the same bridge. This commit do not solve the entire problem. Only case where packet arrived from such interface. PR: kern/109815 MFC after: 7 days Submitted by: Eygene Ryabinkin and rik@ Discussed with: bms@, thompsa@, yar@
2007-03-18 23:28:53 +00:00
# define OR_CARP_CHECK_WE_ARE_SRC(iface) \
|| ((iface)->if_carp \
Unbreak LINT by moving all carp hooks to net/if.c / netinet/ip_carp.h, with the appropriate ifdefs. Reviewed by: bz Approved by: ken (mentor)
2010-08-11 20:18:19 +00:00
&& (*carp_forus_p)((iface), eh->ether_shost))
Give a chance for packet to appear with a correct input interfaces in case of multiple interfaces with the same MAC in the same bridge. This commit do not solve the entire problem. Only case where packet arrived from such interface. PR: kern/109815 MFC after: 7 days Submitted by: Eygene Ryabinkin and rik@ Discussed with: bms@, thompsa@, yar@
2007-03-18 23:28:53 +00:00
#else
# define OR_CARP_CHECK_WE_ARE_DST(iface)
# define OR_CARP_CHECK_WE_ARE_SRC(iface)
#endif
Remove a chunk of duplicated code, test the destination address against the bridge the same way we check member interfaces.
2008-01-18 09:34:09 +00:00
#ifdef INET6
# define OR_PFIL_HOOKED_INET6 \
New pfil(9) KPI together with newborn pfil API and control utility. The KPI have been reviewed and cleansed of features that were planned back 20 years ago and never implemented. The pfil(9) internals have been made opaque to protocols with only returned types and function declarations exposed. The KPI is made more strict, but at the same time more extensible, as kernel uses same command structures that userland ioctl uses. In nutshell [KA]PI is about declaring filtering points, declaring filters and linking and unlinking them together. New [KA]PI makes it possible to reconfigure pfil(9) configuration: change order of hooks, rehook filter from one filtering point to a different one, disconnect a hook on output leaving it on input only, prepend/append a filter to existing list of filters. Now it possible for a single packet filter to provide multiple rulesets that may be linked to different points. Think of per-interface ACLs in Cisco or Juniper. None of existing packet filters yet support that, however limited usage is already possible, e.g. default ruleset can be moved to single interface, as soon as interface would pride their filtering points. Another future feature is possiblity to create pfil heads, that provide not an mbuf pointer but just a memory pointer with length. That would allow filtering at very early stages of a packet lifecycle, e.g. when packet has just been received by a NIC and no mbuf was yet allocated. Differential Revision: https://reviews.freebsd.org/D18951
2019-01-31 23:01:03 +00:00
|| PFIL_HOOKED_IN(V_inet6_pfil_head)
Remove a chunk of duplicated code, test the destination address against the bridge the same way we check member interfaces.
2008-01-18 09:34:09 +00:00
#else
# define OR_PFIL_HOOKED_INET6
#endif
Give a chance for packet to appear with a correct input interfaces in case of multiple interfaces with the same MAC in the same bridge. This commit do not solve the entire problem. Only case where packet arrived from such interface. PR: kern/109815 MFC after: 7 days Submitted by: Eygene Ryabinkin and rik@ Discussed with: bms@, thompsa@, yar@
2007-03-18 23:28:53 +00:00
#define GRAB_OUR_PACKETS(iface) \
if ((iface)->if_type == IFT_GIF) \
continue; \
/* It is destined for us. */ \
if (memcmp(IF_LLADDR((iface)), eh->ether_dhost, ETHER_ADDR_LEN) == 0 \
OR_CARP_CHECK_WE_ARE_DST((iface)) \
) { \
Add an option to limit the number of source MACs that can be behind a bridge interface. Once the limit is reached packets with unknown source addresses are dropped until an existing host cache entry expires or is removed. Useful to use with the STICKY cache option. Sponsored by: miniSuperHappyDevHouse NZ
2007-11-04 08:32:27 +00:00
if (bif->bif_flags & IFBIF_LEARNING) { \
error = bridge_rtupdate(sc, eh->ether_shost, \
Add the vlan tag to the bridge route table. This allows a vlan trunk to be bridged, previously legitimate traffic was not passed as the bridge could not tell that it was on a different Ethernet segment. All non-tagged traffic is treated as vlan1 as per IEEE 802.1Q-2003
2007-06-13 18:58:04 +00:00
vlan, bif, 0, IFBAF_DYNAMIC); \
Add an option to limit the number of source MACs that can be behind a bridge interface. Once the limit is reached packets with unknown source addresses are dropped until an existing host cache entry expires or is removed. Useful to use with the STICKY cache option. Sponsored by: miniSuperHappyDevHouse NZ
2007-11-04 08:32:27 +00:00
if (error && bif->bif_addrmax) { \
m_freem(m); \
return (NULL); \
} \
} \
Give a chance for packet to appear with a correct input interfaces in case of multiple interfaces with the same MAC in the same bridge. This commit do not solve the entire problem. Only case where packet arrived from such interface. PR: kern/109815 MFC after: 7 days Submitted by: Eygene Ryabinkin and rik@ Discussed with: bms@, thompsa@, yar@
2007-03-18 23:28:53 +00:00
m->m_pkthdr.rcvif = iface; \
if_bridge(4): ensure all traffic passing over the bridge is accounted for Consider a bridge0 with em0 and em1 members. Traffic rx'd by em0 and transmitted by bridge0 through em1 gets accounted for in IPACKETS/IBYTES and bridge0 bpf -- assuming it's not unicast traffic destined for em1. Unicast traffic destined for em1 traffic is not accounted for by any mechanism, and isn't pushed through bridge0's bpf machinery as any other packets that pass over the bridge do. Fix this and simplify GRAB_OUR_PACKETS by bailing out early if it was rx'd by the interface that it was addressed for. Everything else there is relevant for any traffic that came in from one member that's being directed at another member of the bridge. Reviewed by: kp MFC after: 1 week Differential Revision: https://reviews.freebsd.org/D19614
2019-03-28 03:31:51 +00:00
if ((iface) == ifp) { \
/* Skip bridge processing... src == dest */ \
return (m); \
} \
/* It's passing over or to the bridge, locally. */ \
ETHER_BPF_MTAP(bifp, m); \
if_inc_counter(bifp, IFCOUNTER_IPACKETS, 1); \
if_inc_counter(bifp, IFCOUNTER_IBYTES, m->m_pkthdr.len); \
/* Filter on the physical interface. */ \
if (V_pfil_local_phys && (PFIL_HOOKED_IN(V_inet_pfil_head) \
OR_PFIL_HOOKED_INET6)) { \
if (bridge_pfil(&m, NULL, ifp, \
PFIL_IN) != 0 || m == NULL) { \
return (NULL); \
} \
} \
if_bridge(4): Complete bpf auditing of local traffic over the bridge There were two remaining "gaps" in auditing local bridge traffic with bpf(4): Locally originated outbound traffic from a member interface is invisible to the bridge's bpf(4) interface. Inbound traffic locally destined to a member interface is invisible to the member's bpf(4) interface -- this traffic has no chance after bridge_input to otherwise pass it over, and it wasn't originally received on this interface. I call these "gaps" because they don't affect conventional bridge setups. Alas, being able to establish an audit trail of all locally destined traffic for setups that can function like this is useful in some scenarios. Reviewed by: kp MFC after: 1 week Differential Revision: https://reviews.freebsd.org/D19757
2019-05-29 01:08:30 +00:00
if ((iface) != bifp) \
ETHER_BPF_MTAP(iface, m); \
Give a chance for packet to appear with a correct input interfaces in case of multiple interfaces with the same MAC in the same bridge. This commit do not solve the entire problem. Only case where packet arrived from such interface. PR: kern/109815 MFC after: 7 days Submitted by: Eygene Ryabinkin and rik@ Discussed with: bms@, thompsa@, yar@
2007-03-18 23:28:53 +00:00
return (m); \
} \
\
/* We just received a packet that we sent out. */ \
if (memcmp(IF_LLADDR((iface)), eh->ether_shost, ETHER_ADDR_LEN) == 0 \
OR_CARP_CHECK_WE_ARE_SRC((iface)) \
) { \
m_freem(m); \
return (NULL); \
}
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
/*
Remove a chunk of duplicated code, test the destination address against the bridge the same way we check member interfaces.
2008-01-18 09:34:09 +00:00
* Unicast. Make sure it's not for the bridge.
*/
do { GRAB_OUR_PACKETS(bifp) } while (0);
/*
Give a chance for packet to appear with a correct input interfaces in case of multiple interfaces with the same MAC in the same bridge. This commit do not solve the entire problem. Only case where packet arrived from such interface. PR: kern/109815 MFC after: 7 days Submitted by: Eygene Ryabinkin and rik@ Discussed with: bms@, thompsa@, yar@
2007-03-18 23:28:53 +00:00
* Give a chance for ifp at first priority. This will help when the
* packet comes through the interface like VLAN's with the same MACs
* on several interfaces from the same bridge. This also will save
* some CPU cycles in case the destination interface and the input
* interface (eq ifp) are the same.
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
*/
Give a chance for packet to appear with a correct input interfaces in case of multiple interfaces with the same MAC in the same bridge. This commit do not solve the entire problem. Only case where packet arrived from such interface. PR: kern/109815 MFC after: 7 days Submitted by: Eygene Ryabinkin and rik@ Discussed with: bms@, thompsa@, yar@
2007-03-18 23:28:53 +00:00
do { GRAB_OUR_PACKETS(ifp) } while (0);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
Give a chance for packet to appear with a correct input interfaces in case of multiple interfaces with the same MAC in the same bridge. This commit do not solve the entire problem. Only case where packet arrived from such interface. PR: kern/109815 MFC after: 7 days Submitted by: Eygene Ryabinkin and rik@ Discussed with: bms@, thompsa@, yar@
2007-03-18 23:28:53 +00:00
/* Now check the all bridge members. */
bridge: Change lists to CK_LIST as a peparation for epochification Prepare the ground for a rework of the bridge locking approach. We will use an epoch-based approach in the datapath and making it safe to iterate over the interface, span and rtnode lists without holding the BRIDGE_LOCK. Replace the relevant lists by their ConcurrencyKit equivalents. No functional change in this commit. Reviewed by: emaste, ae, philip (previous version) Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24249
2020-04-05 17:15:20 +00:00
CK_LIST_FOREACH(bif2, &sc->sc_iflist, bif_next) {
Give a chance for packet to appear with a correct input interfaces in case of multiple interfaces with the same MAC in the same bridge. This commit do not solve the entire problem. Only case where packet arrived from such interface. PR: kern/109815 MFC after: 7 days Submitted by: Eygene Ryabinkin and rik@ Discussed with: bms@, thompsa@, yar@
2007-03-18 23:28:53 +00:00
GRAB_OUR_PACKETS(bif2->bif_ifp)
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
}
Give a chance for packet to appear with a correct input interfaces in case of multiple interfaces with the same MAC in the same bridge. This commit do not solve the entire problem. Only case where packet arrived from such interface. PR: kern/109815 MFC after: 7 days Submitted by: Eygene Ryabinkin and rik@ Discussed with: bms@, thompsa@, yar@
2007-03-18 23:28:53 +00:00
#undef OR_CARP_CHECK_WE_ARE_DST
#undef OR_CARP_CHECK_WE_ARE_SRC
Remove a chunk of duplicated code, test the destination address against the bridge the same way we check member interfaces.
2008-01-18 09:34:09 +00:00
#undef OR_PFIL_HOOKED_INET6
Give a chance for packet to appear with a correct input interfaces in case of multiple interfaces with the same MAC in the same bridge. This commit do not solve the entire problem. Only case where packet arrived from such interface. PR: kern/109815 MFC after: 7 days Submitted by: Eygene Ryabinkin and rik@ Discussed with: bms@, thompsa@, yar@
2007-03-18 23:28:53 +00:00
#undef GRAB_OUR_PACKETS
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
/* Perform the bridge forwarding function. */
Add a bridge interface flag called PRIVATE where any private port can not communicate with another private port. All unicast/broadcast/multicast layer2 traffic is blocked so it works much the same way as using firewall rules but scales better and is generally easier as firewall packages usually do not allow ARP blocking. An example usage would be having a number of customers on separate vlans bridged with a server network. All the vlans are marked private, they can all communicate with the server network unhindered, but can not exchange any traffic whatsoever with each other. Approved by: re (rwatson)
2007-08-01 00:33:52 +00:00
bridge_forward(sc, bif, m);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
return (NULL);
}
/*
* bridge_broadcast:
*
* Send a frame to all interfaces that are members of
* the bridge, except for the one on which the packet
* arrived.
*
* NOTE: Releases the lock on return.
*/
Change most of the bridge and stp funtions to static. This has highlighted that the following funtions are not used, wrap in '#ifdef noused' for the moment. bstp_enable_change_detection bstp_disable_change_detection bstp_set_bridge_priority bstp_set_port_priority bstp_set_path_cost
2005-10-14 10:38:12 +00:00
static void
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge_broadcast(struct bridge_softc *sc, struct ifnet *src_if,
Do not packet filter in the bridge_start() routine, locally generated packets are already filtered by the higher layers. Approved by: mlaier (mentor) MFC after: 3 days
2005-10-02 19:15:56 +00:00
struct mbuf *m, int runfilt)
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
{
Add a bridge interface flag called PRIVATE where any private port can not communicate with another private port. All unicast/broadcast/multicast layer2 traffic is blocked so it works much the same way as using firewall rules but scales better and is generally easier as firewall packages usually do not allow ARP blocking. An example usage would be having a number of customers on separate vlans bridged with a server network. All the vlans are marked private, they can all communicate with the server network unhindered, but can not exchange any traffic whatsoever with each other. Approved by: re (rwatson)
2007-08-01 00:33:52 +00:00
struct bridge_iflist *dbif, *sbif;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
struct mbuf *mc;
struct ifnet *dst_if;
bridge: epoch-ification Run the bridge datapath under epoch, rather than under the BRIDGE_LOCK(). We still take the BRIDGE_LOCK() whenever we insert or delete items in the relevant lists, but we use epoch callbacks to free items so that it's safe to iterate the lists without the BRIDGE_LOCK. Tests on mercat5/6 shows this increases bridge throughput significantly, from 3.7Mpps to 18.6Mpps. Reviewed by: emaste, philip, melifaro MFC after: 2 months Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24250
2020-04-26 16:22:35 +00:00
int used = 0, i;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge: epoch-ification Run the bridge datapath under epoch, rather than under the BRIDGE_LOCK(). We still take the BRIDGE_LOCK() whenever we insert or delete items in the relevant lists, but we use epoch callbacks to free items so that it's safe to iterate the lists without the BRIDGE_LOCK. Tests on mercat5/6 shows this increases bridge throughput significantly, from 3.7Mpps to 18.6Mpps. Reviewed by: emaste, philip, melifaro MFC after: 2 months Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24250
2020-04-26 16:22:35 +00:00
NET_EPOCH_ASSERT();
Add a bridge interface flag called PRIVATE where any private port can not communicate with another private port. All unicast/broadcast/multicast layer2 traffic is blocked so it works much the same way as using firewall rules but scales better and is generally easier as firewall packages usually do not allow ARP blocking. An example usage would be having a number of customers on separate vlans bridged with a server network. All the vlans are marked private, they can all communicate with the server network unhindered, but can not exchange any traffic whatsoever with each other. Approved by: re (rwatson)
2007-08-01 00:33:52 +00:00
bridge: epoch-ification Run the bridge datapath under epoch, rather than under the BRIDGE_LOCK(). We still take the BRIDGE_LOCK() whenever we insert or delete items in the relevant lists, but we use epoch callbacks to free items so that it's safe to iterate the lists without the BRIDGE_LOCK. Tests on mercat5/6 shows this increases bridge throughput significantly, from 3.7Mpps to 18.6Mpps. Reviewed by: emaste, philip, melifaro MFC after: 2 months Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24250
2020-04-26 16:22:35 +00:00
sbif = bridge_lookup_member_if(sc, src_if);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
- Previously when broadcasting to N number of interfaces we would run pfil hooks for each outgoing interface but also run pfil hooks _N times_ on the bridge interface. This is changed so pfil hooks are run once for the bridge interface (bridge0) and then only on the outgoing interfaces in the broadcast loop. - Simplify bridge_enqueue() by moving bridge_pfil() to the callers. - Check (inet6_pfil_hook.ph_busy_count >= 0), it may be possible to have a packet filter hooked for only ipv6 but we were only checking if ipv4 hooks were busy. - Minor optimisation for null mbuf check after bridge_pfil(), move it into the if-block as it couldnt possibly be null outside. Prodded by: mlaier Approved by: re (scottl), mlaier (mentor)
2005-07-06 01:24:45 +00:00
/* Filter on the bridge interface before broadcasting */
New pfil(9) KPI together with newborn pfil API and control utility. The KPI have been reviewed and cleansed of features that were planned back 20 years ago and never implemented. The pfil(9) internals have been made opaque to protocols with only returned types and function declarations exposed. The KPI is made more strict, but at the same time more extensible, as kernel uses same command structures that userland ioctl uses. In nutshell [KA]PI is about declaring filtering points, declaring filters and linking and unlinking them together. New [KA]PI makes it possible to reconfigure pfil(9) configuration: change order of hooks, rehook filter from one filtering point to a different one, disconnect a hook on output leaving it on input only, prepend/append a filter to existing list of filters. Now it possible for a single packet filter to provide multiple rulesets that may be linked to different points. Think of per-interface ACLs in Cisco or Juniper. None of existing packet filters yet support that, however limited usage is already possible, e.g. default ruleset can be moved to single interface, as soon as interface would pride their filtering points. Another future feature is possiblity to create pfil heads, that provide not an mbuf pointer but just a memory pointer with length. That would allow filtering at very early stages of a packet lifecycle, e.g. when packet has just been received by a NIC and no mbuf was yet allocated. Differential Revision: https://reviews.freebsd.org/D18951
2019-01-31 23:01:03 +00:00
if (runfilt && (PFIL_HOOKED_OUT(V_inet_pfil_head)
- Previously when broadcasting to N number of interfaces we would run pfil hooks for each outgoing interface but also run pfil hooks _N times_ on the bridge interface. This is changed so pfil hooks are run once for the bridge interface (bridge0) and then only on the outgoing interfaces in the broadcast loop. - Simplify bridge_enqueue() by moving bridge_pfil() to the callers. - Check (inet6_pfil_hook.ph_busy_count >= 0), it may be possible to have a packet filter hooked for only ipv6 but we were only checking if ipv4 hooks were busy. - Minor optimisation for null mbuf check after bridge_pfil(), move it into the if-block as it couldnt possibly be null outside. Prodded by: mlaier Approved by: re (scottl), mlaier (mentor)
2005-07-06 01:24:45 +00:00
#ifdef INET6
New pfil(9) KPI together with newborn pfil API and control utility. The KPI have been reviewed and cleansed of features that were planned back 20 years ago and never implemented. The pfil(9) internals have been made opaque to protocols with only returned types and function declarations exposed. The KPI is made more strict, but at the same time more extensible, as kernel uses same command structures that userland ioctl uses. In nutshell [KA]PI is about declaring filtering points, declaring filters and linking and unlinking them together. New [KA]PI makes it possible to reconfigure pfil(9) configuration: change order of hooks, rehook filter from one filtering point to a different one, disconnect a hook on output leaving it on input only, prepend/append a filter to existing list of filters. Now it possible for a single packet filter to provide multiple rulesets that may be linked to different points. Think of per-interface ACLs in Cisco or Juniper. None of existing packet filters yet support that, however limited usage is already possible, e.g. default ruleset can be moved to single interface, as soon as interface would pride their filtering points. Another future feature is possiblity to create pfil heads, that provide not an mbuf pointer but just a memory pointer with length. That would allow filtering at very early stages of a packet lifecycle, e.g. when packet has just been received by a NIC and no mbuf was yet allocated. Differential Revision: https://reviews.freebsd.org/D18951
2019-01-31 23:01:03 +00:00
|| PFIL_HOOKED_OUT(V_inet6_pfil_head)
- Previously when broadcasting to N number of interfaces we would run pfil hooks for each outgoing interface but also run pfil hooks _N times_ on the bridge interface. This is changed so pfil hooks are run once for the bridge interface (bridge0) and then only on the outgoing interfaces in the broadcast loop. - Simplify bridge_enqueue() by moving bridge_pfil() to the callers. - Check (inet6_pfil_hook.ph_busy_count >= 0), it may be possible to have a packet filter hooked for only ipv6 but we were only checking if ipv4 hooks were busy. - Minor optimisation for null mbuf check after bridge_pfil(), move it into the if-block as it couldnt possibly be null outside. Prodded by: mlaier Approved by: re (scottl), mlaier (mentor)
2005-07-06 01:24:45 +00:00
#endif
Do not packet filter in the bridge_start() routine, locally generated packets are already filtered by the higher layers. Approved by: mlaier (mentor) MFC after: 3 days
2005-10-02 19:15:56 +00:00
)) {
- Previously when broadcasting to N number of interfaces we would run pfil hooks for each outgoing interface but also run pfil hooks _N times_ on the bridge interface. This is changed so pfil hooks are run once for the bridge interface (bridge0) and then only on the outgoing interfaces in the broadcast loop. - Simplify bridge_enqueue() by moving bridge_pfil() to the callers. - Check (inet6_pfil_hook.ph_busy_count >= 0), it may be possible to have a packet filter hooked for only ipv6 but we were only checking if ipv4 hooks were busy. - Minor optimisation for null mbuf check after bridge_pfil(), move it into the if-block as it couldnt possibly be null outside. Prodded by: mlaier Approved by: re (scottl), mlaier (mentor)
2005-07-06 01:24:45 +00:00
if (bridge_pfil(&m, sc->sc_ifp, NULL, PFIL_OUT) != 0)
bridge: epoch-ification Run the bridge datapath under epoch, rather than under the BRIDGE_LOCK(). We still take the BRIDGE_LOCK() whenever we insert or delete items in the relevant lists, but we use epoch callbacks to free items so that it's safe to iterate the lists without the BRIDGE_LOCK. Tests on mercat5/6 shows this increases bridge throughput significantly, from 3.7Mpps to 18.6Mpps. Reviewed by: emaste, philip, melifaro MFC after: 2 months Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24250
2020-04-26 16:22:35 +00:00
return;
- Previously when broadcasting to N number of interfaces we would run pfil hooks for each outgoing interface but also run pfil hooks _N times_ on the bridge interface. This is changed so pfil hooks are run once for the bridge interface (bridge0) and then only on the outgoing interfaces in the broadcast loop. - Simplify bridge_enqueue() by moving bridge_pfil() to the callers. - Check (inet6_pfil_hook.ph_busy_count >= 0), it may be possible to have a packet filter hooked for only ipv6 but we were only checking if ipv4 hooks were busy. - Minor optimisation for null mbuf check after bridge_pfil(), move it into the if-block as it couldnt possibly be null outside. Prodded by: mlaier Approved by: re (scottl), mlaier (mentor)
2005-07-06 01:24:45 +00:00
if (m == NULL)
bridge: epoch-ification Run the bridge datapath under epoch, rather than under the BRIDGE_LOCK(). We still take the BRIDGE_LOCK() whenever we insert or delete items in the relevant lists, but we use epoch callbacks to free items so that it's safe to iterate the lists without the BRIDGE_LOCK. Tests on mercat5/6 shows this increases bridge throughput significantly, from 3.7Mpps to 18.6Mpps. Reviewed by: emaste, philip, melifaro MFC after: 2 months Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24250
2020-04-26 16:22:35 +00:00
return;
- Previously when broadcasting to N number of interfaces we would run pfil hooks for each outgoing interface but also run pfil hooks _N times_ on the bridge interface. This is changed so pfil hooks are run once for the bridge interface (bridge0) and then only on the outgoing interfaces in the broadcast loop. - Simplify bridge_enqueue() by moving bridge_pfil() to the callers. - Check (inet6_pfil_hook.ph_busy_count >= 0), it may be possible to have a packet filter hooked for only ipv6 but we were only checking if ipv4 hooks were busy. - Minor optimisation for null mbuf check after bridge_pfil(), move it into the if-block as it couldnt possibly be null outside. Prodded by: mlaier Approved by: re (scottl), mlaier (mentor)
2005-07-06 01:24:45 +00:00
}
bridge: Change lists to CK_LIST as a peparation for epochification Prepare the ground for a rework of the bridge locking approach. We will use an epoch-based approach in the datapath and making it safe to iterate over the interface, span and rtnode lists without holding the BRIDGE_LOCK. Replace the relevant lists by their ConcurrencyKit equivalents. No functional change in this commit. Reviewed by: emaste, ae, philip (previous version) Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24249
2020-04-05 17:15:20 +00:00
CK_LIST_FOREACH(dbif, &sc->sc_iflist, bif_next) {
Add a bridge interface flag called PRIVATE where any private port can not communicate with another private port. All unicast/broadcast/multicast layer2 traffic is blocked so it works much the same way as using firewall rules but scales better and is generally easier as firewall packages usually do not allow ARP blocking. An example usage would be having a number of customers on separate vlans bridged with a server network. All the vlans are marked private, they can all communicate with the server network unhindered, but can not exchange any traffic whatsoever with each other. Approved by: re (rwatson)
2007-08-01 00:33:52 +00:00
dst_if = dbif->bif_ifp;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
if (dst_if == src_if)
continue;
Add a bridge interface flag called PRIVATE where any private port can not communicate with another private port. All unicast/broadcast/multicast layer2 traffic is blocked so it works much the same way as using firewall rules but scales better and is generally easier as firewall packages usually do not allow ARP blocking. An example usage would be having a number of customers on separate vlans bridged with a server network. All the vlans are marked private, they can all communicate with the server network unhindered, but can not exchange any traffic whatsoever with each other. Approved by: re (rwatson)
2007-08-01 00:33:52 +00:00
/* Private segments can not talk to each other */
if (sbif && (sbif->bif_flags & dbif->bif_flags & IFBIF_PRIVATE))
continue;
if ((dbif->bif_flags & IFBIF_STP) &&
dbif->bif_stp.bp_state == BSTP_IFSTATE_DISCARDING)
Bring in support for the Rapid Spanning Tree Protocol (802.1w). RSTP provides faster spanning tree convergence, the protocol will exchange information with neighboring switches to quickly transition to forwarding without creating loops. The code will default to RSTP mode but will downgrade any port connected to a legacy STP network so is fully backward compatible. Reviewed by: syrinx Tested by: syrinx
2006-11-01 09:07:47 +00:00
continue;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
Add a bridge interface flag called PRIVATE where any private port can not communicate with another private port. All unicast/broadcast/multicast layer2 traffic is blocked so it works much the same way as using firewall rules but scales better and is generally easier as firewall packages usually do not allow ARP blocking. An example usage would be having a number of customers on separate vlans bridged with a server network. All the vlans are marked private, they can all communicate with the server network unhindered, but can not exchange any traffic whatsoever with each other. Approved by: re (rwatson)
2007-08-01 00:33:52 +00:00
if ((dbif->bif_flags & IFBIF_DISCOVER) == 0 &&
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
(m->m_flags & (M_BCAST|M_MCAST)) == 0)
continue;
Propagate rename of IFF_OACTIVE and IFF_RUNNING to IFF_DRV_OACTIVE and IFF_DRV_RUNNING, as well as the move from ifnet.if_flags to ifnet.if_drv_flags. Device drivers are now responsible for synchronizing access to these flags, as they are in if_drv_flags. This helps prevent races between the network stack and device driver in maintaining the interface flags field. Many __FreeBSD__ and __FreeBSD_version checks maintained and continued; some less so. Reviewed by: pjd, bz MFC after: 7 days
2005-08-09 10:20:02 +00:00
if ((dst_if->if_drv_flags & IFF_DRV_RUNNING) == 0)
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
continue;
bridge: Change lists to CK_LIST as a peparation for epochification Prepare the ground for a rework of the bridge locking approach. We will use an epoch-based approach in the datapath and making it safe to iterate over the interface, span and rtnode lists without holding the BRIDGE_LOCK. Replace the relevant lists by their ConcurrencyKit equivalents. No functional change in this commit. Reviewed by: emaste, ae, philip (previous version) Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24249
2020-04-05 17:15:20 +00:00
if (CK_LIST_NEXT(dbif, bif_next) == NULL) {
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
mc = m;
used = 1;
} else {
Mechanically substitute flags from historic mbuf allocator with malloc(9) flags within sys. Exceptions: - sys/contrib not touched - sys/mbuf.h edited manually
2012-12-05 08:04:20 +00:00
mc = m_dup(m, M_NOWAIT);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
if (mc == NULL) {
Mechanically convert to if_inc_counter().
2014-09-19 10:39:58 +00:00
if_inc_counter(sc->sc_ifp, IFCOUNTER_OERRORS, 1);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
continue;
}
}
- Previously when broadcasting to N number of interfaces we would run pfil hooks for each outgoing interface but also run pfil hooks _N times_ on the bridge interface. This is changed so pfil hooks are run once for the bridge interface (bridge0) and then only on the outgoing interfaces in the broadcast loop. - Simplify bridge_enqueue() by moving bridge_pfil() to the callers. - Check (inet6_pfil_hook.ph_busy_count >= 0), it may be possible to have a packet filter hooked for only ipv6 but we were only checking if ipv4 hooks were busy. - Minor optimisation for null mbuf check after bridge_pfil(), move it into the if-block as it couldnt possibly be null outside. Prodded by: mlaier Approved by: re (scottl), mlaier (mentor)
2005-07-06 01:24:45 +00:00
/*
* Filter on the output interface. Pass a NULL bridge interface
Minor whitespace cleanup.
2005-12-17 10:03:48 +00:00
* pointer so we do not redundantly filter on the bridge for
- Previously when broadcasting to N number of interfaces we would run pfil hooks for each outgoing interface but also run pfil hooks _N times_ on the bridge interface. This is changed so pfil hooks are run once for the bridge interface (bridge0) and then only on the outgoing interfaces in the broadcast loop. - Simplify bridge_enqueue() by moving bridge_pfil() to the callers. - Check (inet6_pfil_hook.ph_busy_count >= 0), it may be possible to have a packet filter hooked for only ipv6 but we were only checking if ipv4 hooks were busy. - Minor optimisation for null mbuf check after bridge_pfil(), move it into the if-block as it couldnt possibly be null outside. Prodded by: mlaier Approved by: re (scottl), mlaier (mentor)
2005-07-06 01:24:45 +00:00
* each interface we broadcast on.
*/
New pfil(9) KPI together with newborn pfil API and control utility. The KPI have been reviewed and cleansed of features that were planned back 20 years ago and never implemented. The pfil(9) internals have been made opaque to protocols with only returned types and function declarations exposed. The KPI is made more strict, but at the same time more extensible, as kernel uses same command structures that userland ioctl uses. In nutshell [KA]PI is about declaring filtering points, declaring filters and linking and unlinking them together. New [KA]PI makes it possible to reconfigure pfil(9) configuration: change order of hooks, rehook filter from one filtering point to a different one, disconnect a hook on output leaving it on input only, prepend/append a filter to existing list of filters. Now it possible for a single packet filter to provide multiple rulesets that may be linked to different points. Think of per-interface ACLs in Cisco or Juniper. None of existing packet filters yet support that, however limited usage is already possible, e.g. default ruleset can be moved to single interface, as soon as interface would pride their filtering points. Another future feature is possiblity to create pfil heads, that provide not an mbuf pointer but just a memory pointer with length. That would allow filtering at very early stages of a packet lifecycle, e.g. when packet has just been received by a NIC and no mbuf was yet allocated. Differential Revision: https://reviews.freebsd.org/D18951
2019-01-31 23:01:03 +00:00
if (runfilt && (PFIL_HOOKED_OUT(V_inet_pfil_head)
- Previously when broadcasting to N number of interfaces we would run pfil hooks for each outgoing interface but also run pfil hooks _N times_ on the bridge interface. This is changed so pfil hooks are run once for the bridge interface (bridge0) and then only on the outgoing interfaces in the broadcast loop. - Simplify bridge_enqueue() by moving bridge_pfil() to the callers. - Check (inet6_pfil_hook.ph_busy_count >= 0), it may be possible to have a packet filter hooked for only ipv6 but we were only checking if ipv4 hooks were busy. - Minor optimisation for null mbuf check after bridge_pfil(), move it into the if-block as it couldnt possibly be null outside. Prodded by: mlaier Approved by: re (scottl), mlaier (mentor)
2005-07-06 01:24:45 +00:00
#ifdef INET6
New pfil(9) KPI together with newborn pfil API and control utility. The KPI have been reviewed and cleansed of features that were planned back 20 years ago and never implemented. The pfil(9) internals have been made opaque to protocols with only returned types and function declarations exposed. The KPI is made more strict, but at the same time more extensible, as kernel uses same command structures that userland ioctl uses. In nutshell [KA]PI is about declaring filtering points, declaring filters and linking and unlinking them together. New [KA]PI makes it possible to reconfigure pfil(9) configuration: change order of hooks, rehook filter from one filtering point to a different one, disconnect a hook on output leaving it on input only, prepend/append a filter to existing list of filters. Now it possible for a single packet filter to provide multiple rulesets that may be linked to different points. Think of per-interface ACLs in Cisco or Juniper. None of existing packet filters yet support that, however limited usage is already possible, e.g. default ruleset can be moved to single interface, as soon as interface would pride their filtering points. Another future feature is possiblity to create pfil heads, that provide not an mbuf pointer but just a memory pointer with length. That would allow filtering at very early stages of a packet lifecycle, e.g. when packet has just been received by a NIC and no mbuf was yet allocated. Differential Revision: https://reviews.freebsd.org/D18951
2019-01-31 23:01:03 +00:00
|| PFIL_HOOKED_OUT(V_inet6_pfil_head)
- Previously when broadcasting to N number of interfaces we would run pfil hooks for each outgoing interface but also run pfil hooks _N times_ on the bridge interface. This is changed so pfil hooks are run once for the bridge interface (bridge0) and then only on the outgoing interfaces in the broadcast loop. - Simplify bridge_enqueue() by moving bridge_pfil() to the callers. - Check (inet6_pfil_hook.ph_busy_count >= 0), it may be possible to have a packet filter hooked for only ipv6 but we were only checking if ipv4 hooks were busy. - Minor optimisation for null mbuf check after bridge_pfil(), move it into the if-block as it couldnt possibly be null outside. Prodded by: mlaier Approved by: re (scottl), mlaier (mentor)
2005-07-06 01:24:45 +00:00
#endif
Do not packet filter in the bridge_start() routine, locally generated packets are already filtered by the higher layers. Approved by: mlaier (mentor) MFC after: 3 days
2005-10-02 19:15:56 +00:00
)) {
m_dup () packet not m_copypacket () since we will modify it. For more details see PR kern/94448. PR: kern/94448 Original patch: Eygene A. Ryabinkin <rea-fbsd at rea dot mbslab dot kiae dot ru>Final patch: thompsa@ Tested by: thompsa@, Eygene A. Ryabinkin MFC after: 7 days
2006-03-23 22:57:10 +00:00
if (used == 0) {
/* Keep the layer3 header aligned */
i = min(mc->m_pkthdr.len, max_protohdr);
mc = m_copyup(mc, i, ETHER_ALIGN);
if (mc == NULL) {
Mechanically convert to if_inc_counter().
2014-09-19 10:39:58 +00:00
if_inc_counter(sc->sc_ifp, IFCOUNTER_OERRORS, 1);
m_dup () packet not m_copypacket () since we will modify it. For more details see PR kern/94448. PR: kern/94448 Original patch: Eygene A. Ryabinkin <rea-fbsd at rea dot mbslab dot kiae dot ru>Final patch: thompsa@ Tested by: thompsa@, Eygene A. Ryabinkin MFC after: 7 days
2006-03-23 22:57:10 +00:00
continue;
}
}
Fix a mbuf and refcnt leak in the broadcast code. If the packet is rejected from pfil(9) then continue the loop rather than returning, this means that we can still try to send it out the remaining interfaces but more importantly the mbuf is freed and refcount decremented on exit.
2005-11-13 19:36:59 +00:00
if (bridge_pfil(&mc, NULL, dst_if, PFIL_OUT) != 0)
continue;
if (mc == NULL)
continue;
- Previously when broadcasting to N number of interfaces we would run pfil hooks for each outgoing interface but also run pfil hooks _N times_ on the bridge interface. This is changed so pfil hooks are run once for the bridge interface (bridge0) and then only on the outgoing interfaces in the broadcast loop. - Simplify bridge_enqueue() by moving bridge_pfil() to the callers. - Check (inet6_pfil_hook.ph_busy_count >= 0), it may be possible to have a packet filter hooked for only ipv6 but we were only checking if ipv4 hooks were busy. - Minor optimisation for null mbuf check after bridge_pfil(), move it into the if-block as it couldnt possibly be null outside. Prodded by: mlaier Approved by: re (scottl), mlaier (mentor)
2005-07-06 01:24:45 +00:00
}
Further clean up the bridge hooks in if_ethersubr.c and ng_ether.c - move the function pointer definitions to if_bridgevar.h - move most of the logic to the new BRIDGE_INPUT and BRIDGE_OUTPUT macros - remove unneeded functions from if_bridgevar.h and sort a little.
2005-10-14 02:38:47 +00:00
- Previously when broadcasting to N number of interfaces we would run pfil hooks for each outgoing interface but also run pfil hooks _N times_ on the bridge interface. This is changed so pfil hooks are run once for the bridge interface (bridge0) and then only on the outgoing interfaces in the broadcast loop. - Simplify bridge_enqueue() by moving bridge_pfil() to the callers. - Check (inet6_pfil_hook.ph_busy_count >= 0), it may be possible to have a packet filter hooked for only ipv6 but we were only checking if ipv4 hooks were busy. - Minor optimisation for null mbuf check after bridge_pfil(), move it into the if-block as it couldnt possibly be null outside. Prodded by: mlaier Approved by: re (scottl), mlaier (mentor)
2005-07-06 01:24:45 +00:00
bridge_enqueue(sc, dst_if, mc);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
}
if (used == 0)
m_freem(m);
}
Add support for creating span ports so that one can snoop bridged traffic from another interface/machine/network. Obtained from: OpenBSD MFC after: 2 weeks
2005-12-14 02:52:13 +00:00
/*
* bridge_span:
*
* Duplicate a packet out one or more interfaces that are in span mode,
* the original mbuf is unmodified.
*/
static void
bridge_span(struct bridge_softc *sc, struct mbuf *m)
{
struct bridge_iflist *bif;
struct ifnet *dst_if;
struct mbuf *mc;
bridge: epoch-ification Run the bridge datapath under epoch, rather than under the BRIDGE_LOCK(). We still take the BRIDGE_LOCK() whenever we insert or delete items in the relevant lists, but we use epoch callbacks to free items so that it's safe to iterate the lists without the BRIDGE_LOCK. Tests on mercat5/6 shows this increases bridge throughput significantly, from 3.7Mpps to 18.6Mpps. Reviewed by: emaste, philip, melifaro MFC after: 2 months Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24250
2020-04-26 16:22:35 +00:00
NET_EPOCH_ASSERT();
bridge: Change lists to CK_LIST as a peparation for epochification Prepare the ground for a rework of the bridge locking approach. We will use an epoch-based approach in the datapath and making it safe to iterate over the interface, span and rtnode lists without holding the BRIDGE_LOCK. Replace the relevant lists by their ConcurrencyKit equivalents. No functional change in this commit. Reviewed by: emaste, ae, philip (previous version) Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24249
2020-04-05 17:15:20 +00:00
if (CK_LIST_EMPTY(&sc->sc_spanlist))
Add support for creating span ports so that one can snoop bridged traffic from another interface/machine/network. Obtained from: OpenBSD MFC after: 2 weeks
2005-12-14 02:52:13 +00:00
return;
bridge: Change lists to CK_LIST as a peparation for epochification Prepare the ground for a rework of the bridge locking approach. We will use an epoch-based approach in the datapath and making it safe to iterate over the interface, span and rtnode lists without holding the BRIDGE_LOCK. Replace the relevant lists by their ConcurrencyKit equivalents. No functional change in this commit. Reviewed by: emaste, ae, philip (previous version) Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24249
2020-04-05 17:15:20 +00:00
CK_LIST_FOREACH(bif, &sc->sc_spanlist, bif_next) {
Add support for creating span ports so that one can snoop bridged traffic from another interface/machine/network. Obtained from: OpenBSD MFC after: 2 weeks
2005-12-14 02:52:13 +00:00
dst_if = bif->bif_ifp;
Fix style(9) nits, whitespace and parentheses.
2006-05-16 22:50:41 +00:00
Add support for creating span ports so that one can snoop bridged traffic from another interface/machine/network. Obtained from: OpenBSD MFC after: 2 weeks
2005-12-14 02:52:13 +00:00
if ((dst_if->if_drv_flags & IFF_DRV_RUNNING) == 0)
continue;
bridge: Don't share broadcast packets if_bridge duplicates broadcast packets with m_copypacket(), which creates shared packets. In certain circumstances these packets can be processed by udp_usrreq.c:udp_input() first, which modifies the mbuf as part of the checksum verification. That may lead to incorrect packets being transmitted. Use m_dup() to create independent mbufs instead. Reported by: Richard Russo <toast@ruka.org> Reviewed by: donner, afedorov MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D34319
2022-02-19 15:34:31 +00:00
mc = m_dup(m, M_NOWAIT);
Add support for creating span ports so that one can snoop bridged traffic from another interface/machine/network. Obtained from: OpenBSD MFC after: 2 weeks
2005-12-14 02:52:13 +00:00
if (mc == NULL) {
Mechanically convert to if_inc_counter().
2014-09-19 10:39:58 +00:00
if_inc_counter(sc->sc_ifp, IFCOUNTER_OERRORS, 1);
Add support for creating span ports so that one can snoop bridged traffic from another interface/machine/network. Obtained from: OpenBSD MFC after: 2 weeks
2005-12-14 02:52:13 +00:00
continue;
}
bridge_enqueue(sc, dst_if, mc);
}
}
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
/*
* bridge_rtupdate:
*
* Add a bridge routing entry.
*/
Change most of the bridge and stp funtions to static. This has highlighted that the following funtions are not used, wrap in '#ifdef noused' for the moment. bstp_enable_change_detection bstp_disable_change_detection bstp_set_bridge_priority bstp_set_port_priority bstp_set_path_cost
2005-10-14 10:38:12 +00:00
static int
Add the vlan tag to the bridge route table. This allows a vlan trunk to be bridged, previously legitimate traffic was not passed as the bridge could not tell that it was on a different Ethernet segment. All non-tagged traffic is treated as vlan1 as per IEEE 802.1Q-2003
2007-06-13 18:58:04 +00:00
bridge_rtupdate(struct bridge_softc *sc, const uint8_t *dst, uint16_t vlan,
Add a new address cache type called sticky. On an interface marked sticky any address learned by the bridge is made permanent, the address will not age out and most importantly will not migrate to another interface. This can be used to stop mac address poisoning or clients roaming in much the same way as static entries without the hassle of preloading the table.
2006-11-09 06:32:38 +00:00
struct bridge_iflist *bif, int setflags, uint8_t flags)
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
{
struct bridge_rtnode *brt;
int error;
bridge: call member interface ioctl() without NET_EPOCH We're not allowed to hold NET_EPOCH while sleeping, so when we call ioctl() handlers for member interfaces we cannot be in NET_EPOCH. We still need some protection of our CK_LISTs, so hold BRIDGE_LOCK instead. That requires changing BRIDGE_LOCK into a sleepable lock, and separating the BRIDGE_RT_LOCK, to protect bridge_rtnode lists. That lock is taken in the data path (while in NET_EPOCH), so it cannot be a sleepable lock. While here document the locking strategy. MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D26418
2020-10-06 19:19:56 +00:00
BRIDGE_LOCK_OR_NET_EPOCH_ASSERT(sc);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
Add an option to limit the number of source MACs that can be behind a bridge interface. Once the limit is reached packets with unknown source addresses are dropped until an existing host cache entry expires or is removed. Useful to use with the STICKY cache option. Sponsored by: miniSuperHappyDevHouse NZ
2007-11-04 08:32:27 +00:00
/* Check the source address is valid and not multicast. */
if (ETHER_IS_MULTICAST(dst) ||
(dst[0] == 0 && dst[1] == 0 && dst[2] == 0 &&
dst[3] == 0 && dst[4] == 0 && dst[5] == 0) != 0)
return (EINVAL);
Add the vlan tag to the bridge route table. This allows a vlan trunk to be bridged, previously legitimate traffic was not passed as the bridge could not tell that it was on a different Ethernet segment. All non-tagged traffic is treated as vlan1 as per IEEE 802.1Q-2003
2007-06-13 18:58:04 +00:00
/* 802.1p frames map to vlan 1 */
if (vlan == 0)
vlan = 1;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
/*
* A route for this destination might already exist. If so,
* update it, otherwise create a new one.
*/
Add the vlan tag to the bridge route table. This allows a vlan trunk to be bridged, previously legitimate traffic was not passed as the bridge could not tell that it was on a different Ethernet segment. All non-tagged traffic is treated as vlan1 as per IEEE 802.1Q-2003
2007-06-13 18:58:04 +00:00
if ((brt = bridge_rtnode_lookup(sc, dst, vlan)) == NULL) {
bridge: call member interface ioctl() without NET_EPOCH We're not allowed to hold NET_EPOCH while sleeping, so when we call ioctl() handlers for member interfaces we cannot be in NET_EPOCH. We still need some protection of our CK_LISTs, so hold BRIDGE_LOCK instead. That requires changing BRIDGE_LOCK into a sleepable lock, and separating the BRIDGE_RT_LOCK, to protect bridge_rtnode lists. That lock is taken in the data path (while in NET_EPOCH), so it cannot be a sleepable lock. While here document the locking strategy. MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D26418
2020-10-06 19:19:56 +00:00
BRIDGE_RT_LOCK(sc);
bridge: epoch-ification Run the bridge datapath under epoch, rather than under the BRIDGE_LOCK(). We still take the BRIDGE_LOCK() whenever we insert or delete items in the relevant lists, but we use epoch callbacks to free items so that it's safe to iterate the lists without the BRIDGE_LOCK. Tests on mercat5/6 shows this increases bridge throughput significantly, from 3.7Mpps to 18.6Mpps. Reviewed by: emaste, philip, melifaro MFC after: 2 months Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24250
2020-04-26 16:22:35 +00:00
/* Check again, now that we have the lock. There could have
* been a race and we only want to insert this once. */
if ((brt = bridge_rtnode_lookup(sc, dst, vlan)) != NULL) {
bridge: call member interface ioctl() without NET_EPOCH We're not allowed to hold NET_EPOCH while sleeping, so when we call ioctl() handlers for member interfaces we cannot be in NET_EPOCH. We still need some protection of our CK_LISTs, so hold BRIDGE_LOCK instead. That requires changing BRIDGE_LOCK into a sleepable lock, and separating the BRIDGE_RT_LOCK, to protect bridge_rtnode lists. That lock is taken in the data path (while in NET_EPOCH), so it cannot be a sleepable lock. While here document the locking strategy. MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D26418
2020-10-06 19:19:56 +00:00
BRIDGE_RT_UNLOCK(sc);
bridge: epoch-ification Run the bridge datapath under epoch, rather than under the BRIDGE_LOCK(). We still take the BRIDGE_LOCK() whenever we insert or delete items in the relevant lists, but we use epoch callbacks to free items so that it's safe to iterate the lists without the BRIDGE_LOCK. Tests on mercat5/6 shows this increases bridge throughput significantly, from 3.7Mpps to 18.6Mpps. Reviewed by: emaste, philip, melifaro MFC after: 2 months Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24250
2020-04-26 16:22:35 +00:00
return (0);
}
Add some statistics that are needed to support RFC4188 as part of the SoC2006 work on a bridge monitoring module for BSNMP. Submitted by: shteryana (SoC 2006)
2006-07-31 20:24:46 +00:00
if (sc->sc_brtcnt >= sc->sc_brtmax) {
sc->sc_brtexceeded++;
bridge: call member interface ioctl() without NET_EPOCH We're not allowed to hold NET_EPOCH while sleeping, so when we call ioctl() handlers for member interfaces we cannot be in NET_EPOCH. We still need some protection of our CK_LISTs, so hold BRIDGE_LOCK instead. That requires changing BRIDGE_LOCK into a sleepable lock, and separating the BRIDGE_RT_LOCK, to protect bridge_rtnode lists. That lock is taken in the data path (while in NET_EPOCH), so it cannot be a sleepable lock. While here document the locking strategy. MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D26418
2020-10-06 19:19:56 +00:00
BRIDGE_RT_UNLOCK(sc);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
return (ENOSPC);
Add some statistics that are needed to support RFC4188 as part of the SoC2006 work on a bridge monitoring module for BSNMP. Submitted by: shteryana (SoC 2006)
2006-07-31 20:24:46 +00:00
}
Add an option to limit the number of source MACs that can be behind a bridge interface. Once the limit is reached packets with unknown source addresses are dropped until an existing host cache entry expires or is removed. Useful to use with the STICKY cache option. Sponsored by: miniSuperHappyDevHouse NZ
2007-11-04 08:32:27 +00:00
/* Check per interface address limits (if enabled) */
if (bif->bif_addrmax && bif->bif_addrcnt >= bif->bif_addrmax) {
bif->bif_addrexceeded++;
bridge: call member interface ioctl() without NET_EPOCH We're not allowed to hold NET_EPOCH while sleeping, so when we call ioctl() handlers for member interfaces we cannot be in NET_EPOCH. We still need some protection of our CK_LISTs, so hold BRIDGE_LOCK instead. That requires changing BRIDGE_LOCK into a sleepable lock, and separating the BRIDGE_RT_LOCK, to protect bridge_rtnode lists. That lock is taken in the data path (while in NET_EPOCH), so it cannot be a sleepable lock. While here document the locking strategy. MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D26418
2020-10-06 19:19:56 +00:00
BRIDGE_RT_UNLOCK(sc);
Add an option to limit the number of source MACs that can be behind a bridge interface. Once the limit is reached packets with unknown source addresses are dropped until an existing host cache entry expires or is removed. Useful to use with the STICKY cache option. Sponsored by: miniSuperHappyDevHouse NZ
2007-11-04 08:32:27 +00:00
return (ENOSPC);
}
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
/*
* Allocate a new bridge forwarding node, and
* initialize the expiration time and Ethernet
* address.
*/
if_bridge(4): Fix module teardown bridge_rtnode_zone still has outstanding allocations at the time of destruction in the current model because all of the interface teardown happens in a VNET_SYSUNINIT, -after- the MOD_UNLOAD has already been processed. The SYSUNINIT triggers destruction of the interfaces, which then attempts to free the memory from the zone that's already been destroyed, and we hit a panic. Solve this by virtualizing the uma_zone we allocate the rtnodes from to fix the ordering. bridge_rtable_fini should also take care to flush any remaining routes that weren't taken care of when dynamic routes were flushed in bridge_stop. Reviewed by: kp Differential Revision: https://reviews.freebsd.org/D19578
2019-03-15 13:19:52 +00:00
brt = uma_zalloc(V_bridge_rtnode_zone, M_NOWAIT | M_ZERO);
bridge: epoch-ification Run the bridge datapath under epoch, rather than under the BRIDGE_LOCK(). We still take the BRIDGE_LOCK() whenever we insert or delete items in the relevant lists, but we use epoch callbacks to free items so that it's safe to iterate the lists without the BRIDGE_LOCK. Tests on mercat5/6 shows this increases bridge throughput significantly, from 3.7Mpps to 18.6Mpps. Reviewed by: emaste, philip, melifaro MFC after: 2 months Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24250
2020-04-26 16:22:35 +00:00
if (brt == NULL) {
bridge: call member interface ioctl() without NET_EPOCH We're not allowed to hold NET_EPOCH while sleeping, so when we call ioctl() handlers for member interfaces we cannot be in NET_EPOCH. We still need some protection of our CK_LISTs, so hold BRIDGE_LOCK instead. That requires changing BRIDGE_LOCK into a sleepable lock, and separating the BRIDGE_RT_LOCK, to protect bridge_rtnode lists. That lock is taken in the data path (while in NET_EPOCH), so it cannot be a sleepable lock. While here document the locking strategy. MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D26418
2020-10-06 19:19:56 +00:00
BRIDGE_RT_UNLOCK(sc);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
return (ENOMEM);
bridge: epoch-ification Run the bridge datapath under epoch, rather than under the BRIDGE_LOCK(). We still take the BRIDGE_LOCK() whenever we insert or delete items in the relevant lists, but we use epoch callbacks to free items so that it's safe to iterate the lists without the BRIDGE_LOCK. Tests on mercat5/6 shows this increases bridge throughput significantly, from 3.7Mpps to 18.6Mpps. Reviewed by: emaste, philip, melifaro MFC after: 2 months Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24250
2020-04-26 16:22:35 +00:00
}
brt->brt_vnet = curvnet;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
Add a new address cache type called sticky. On an interface marked sticky any address learned by the bridge is made permanent, the address will not age out and most importantly will not migrate to another interface. This can be used to stop mac address poisoning or clients roaming in much the same way as static entries without the hassle of preloading the table.
2006-11-09 06:32:38 +00:00
if (bif->bif_flags & IFBIF_STICKY)
brt->brt_flags = IFBAF_STICKY;
else
brt->brt_flags = IFBAF_DYNAMIC;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
memcpy(brt->brt_addr, dst, ETHER_ADDR_LEN);
Add the vlan tag to the bridge route table. This allows a vlan trunk to be bridged, previously legitimate traffic was not passed as the bridge could not tell that it was on a different Ethernet segment. All non-tagged traffic is treated as vlan1 as per IEEE 802.1Q-2003
2007-06-13 18:58:04 +00:00
brt->brt_vlan = vlan;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
if ((error = bridge_rtnode_insert(sc, brt)) != 0) {
if_bridge(4): Fix module teardown bridge_rtnode_zone still has outstanding allocations at the time of destruction in the current model because all of the interface teardown happens in a VNET_SYSUNINIT, -after- the MOD_UNLOAD has already been processed. The SYSUNINIT triggers destruction of the interfaces, which then attempts to free the memory from the zone that's already been destroyed, and we hit a panic. Solve this by virtualizing the uma_zone we allocate the rtnodes from to fix the ordering. bridge_rtable_fini should also take care to flush any remaining routes that weren't taken care of when dynamic routes were flushed in bridge_stop. Reviewed by: kp Differential Revision: https://reviews.freebsd.org/D19578
2019-03-15 13:19:52 +00:00
uma_zfree(V_bridge_rtnode_zone, brt);
bridge: call member interface ioctl() without NET_EPOCH We're not allowed to hold NET_EPOCH while sleeping, so when we call ioctl() handlers for member interfaces we cannot be in NET_EPOCH. We still need some protection of our CK_LISTs, so hold BRIDGE_LOCK instead. That requires changing BRIDGE_LOCK into a sleepable lock, and separating the BRIDGE_RT_LOCK, to protect bridge_rtnode lists. That lock is taken in the data path (while in NET_EPOCH), so it cannot be a sleepable lock. While here document the locking strategy. MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D26418
2020-10-06 19:19:56 +00:00
BRIDGE_RT_UNLOCK(sc);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
return (error);
}
Add an option to limit the number of source MACs that can be behind a bridge interface. Once the limit is reached packets with unknown source addresses are dropped until an existing host cache entry expires or is removed. Useful to use with the STICKY cache option. Sponsored by: miniSuperHappyDevHouse NZ
2007-11-04 08:32:27 +00:00
brt->brt_dst = bif;
bif->bif_addrcnt++;
bridge: epoch-ification Run the bridge datapath under epoch, rather than under the BRIDGE_LOCK(). We still take the BRIDGE_LOCK() whenever we insert or delete items in the relevant lists, but we use epoch callbacks to free items so that it's safe to iterate the lists without the BRIDGE_LOCK. Tests on mercat5/6 shows this increases bridge throughput significantly, from 3.7Mpps to 18.6Mpps. Reviewed by: emaste, philip, melifaro MFC after: 2 months Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24250
2020-04-26 16:22:35 +00:00
bridge: call member interface ioctl() without NET_EPOCH We're not allowed to hold NET_EPOCH while sleeping, so when we call ioctl() handlers for member interfaces we cannot be in NET_EPOCH. We still need some protection of our CK_LISTs, so hold BRIDGE_LOCK instead. That requires changing BRIDGE_LOCK into a sleepable lock, and separating the BRIDGE_RT_LOCK, to protect bridge_rtnode lists. That lock is taken in the data path (while in NET_EPOCH), so it cannot be a sleepable lock. While here document the locking strategy. MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D26418
2020-10-06 19:19:56 +00:00
BRIDGE_RT_UNLOCK(sc);
Add an option to limit the number of source MACs that can be behind a bridge interface. Once the limit is reached packets with unknown source addresses are dropped until an existing host cache entry expires or is removed. Useful to use with the STICKY cache option. Sponsored by: miniSuperHappyDevHouse NZ
2007-11-04 08:32:27 +00:00
}
if ((brt->brt_flags & IFBAF_TYPEMASK) == IFBAF_DYNAMIC &&
brt->brt_dst != bif) {
bridge: call member interface ioctl() without NET_EPOCH We're not allowed to hold NET_EPOCH while sleeping, so when we call ioctl() handlers for member interfaces we cannot be in NET_EPOCH. We still need some protection of our CK_LISTs, so hold BRIDGE_LOCK instead. That requires changing BRIDGE_LOCK into a sleepable lock, and separating the BRIDGE_RT_LOCK, to protect bridge_rtnode lists. That lock is taken in the data path (while in NET_EPOCH), so it cannot be a sleepable lock. While here document the locking strategy. MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D26418
2020-10-06 19:19:56 +00:00
BRIDGE_RT_LOCK(sc);
Add an option to limit the number of source MACs that can be behind a bridge interface. Once the limit is reached packets with unknown source addresses are dropped until an existing host cache entry expires or is removed. Useful to use with the STICKY cache option. Sponsored by: miniSuperHappyDevHouse NZ
2007-11-04 08:32:27 +00:00
brt->brt_dst->bif_addrcnt--;
brt->brt_dst = bif;
brt->brt_dst->bif_addrcnt++;
bridge: call member interface ioctl() without NET_EPOCH We're not allowed to hold NET_EPOCH while sleeping, so when we call ioctl() handlers for member interfaces we cannot be in NET_EPOCH. We still need some protection of our CK_LISTs, so hold BRIDGE_LOCK instead. That requires changing BRIDGE_LOCK into a sleepable lock, and separating the BRIDGE_RT_LOCK, to protect bridge_rtnode lists. That lock is taken in the data path (while in NET_EPOCH), so it cannot be a sleepable lock. While here document the locking strategy. MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D26418
2020-10-06 19:19:56 +00:00
BRIDGE_RT_UNLOCK(sc);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
}
Fix a brain-o in the last commit, the conditional was always false.
2006-01-02 23:02:43 +00:00
if ((flags & IFBAF_TYPEMASK) == IFBAF_DYNAMIC)
Reorganise bridge_rtupdate slightly to reduce duplication.
2006-01-02 22:44:54 +00:00
brt->brt_expire = time_uptime + sc->sc_brttimeout;
Reset the route expiry time on each update rather than always letting them get GC'd and recreated.
2006-01-02 22:29:41 +00:00
if (setflags)
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
brt->brt_flags = flags;
return (0);
}
/*
* bridge_rtlookup:
*
* Lookup the destination interface for an address.
*/
Make four more functions static that were missed in the last commit.
2005-10-14 20:57:02 +00:00
static struct ifnet *
Add the vlan tag to the bridge route table. This allows a vlan trunk to be bridged, previously legitimate traffic was not passed as the bridge could not tell that it was on a different Ethernet segment. All non-tagged traffic is treated as vlan1 as per IEEE 802.1Q-2003
2007-06-13 18:58:04 +00:00
bridge_rtlookup(struct bridge_softc *sc, const uint8_t *addr, uint16_t vlan)
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
{
struct bridge_rtnode *brt;
bridge: epoch-ification Run the bridge datapath under epoch, rather than under the BRIDGE_LOCK(). We still take the BRIDGE_LOCK() whenever we insert or delete items in the relevant lists, but we use epoch callbacks to free items so that it's safe to iterate the lists without the BRIDGE_LOCK. Tests on mercat5/6 shows this increases bridge throughput significantly, from 3.7Mpps to 18.6Mpps. Reviewed by: emaste, philip, melifaro MFC after: 2 months Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24250
2020-04-26 16:22:35 +00:00
NET_EPOCH_ASSERT();
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
Add the vlan tag to the bridge route table. This allows a vlan trunk to be bridged, previously legitimate traffic was not passed as the bridge could not tell that it was on a different Ethernet segment. All non-tagged traffic is treated as vlan1 as per IEEE 802.1Q-2003
2007-06-13 18:58:04 +00:00
if ((brt = bridge_rtnode_lookup(sc, addr, vlan)) == NULL)
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
return (NULL);
return (brt->brt_ifp);
}
/*
* bridge_rttrim:
*
* Trim the routine table so that we have a number
* of routing entries less than or equal to the
* maximum number.
*/
Change most of the bridge and stp funtions to static. This has highlighted that the following funtions are not used, wrap in '#ifdef noused' for the moment. bstp_enable_change_detection bstp_disable_change_detection bstp_set_bridge_priority bstp_set_port_priority bstp_set_path_cost
2005-10-14 10:38:12 +00:00
static void
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge_rttrim(struct bridge_softc *sc)
{
struct bridge_rtnode *brt, *nbrt;
bridge: epoch-ification Run the bridge datapath under epoch, rather than under the BRIDGE_LOCK(). We still take the BRIDGE_LOCK() whenever we insert or delete items in the relevant lists, but we use epoch callbacks to free items so that it's safe to iterate the lists without the BRIDGE_LOCK. Tests on mercat5/6 shows this increases bridge throughput significantly, from 3.7Mpps to 18.6Mpps. Reviewed by: emaste, philip, melifaro MFC after: 2 months Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24250
2020-04-26 16:22:35 +00:00
NET_EPOCH_ASSERT();
bridge: call member interface ioctl() without NET_EPOCH We're not allowed to hold NET_EPOCH while sleeping, so when we call ioctl() handlers for member interfaces we cannot be in NET_EPOCH. We still need some protection of our CK_LISTs, so hold BRIDGE_LOCK instead. That requires changing BRIDGE_LOCK into a sleepable lock, and separating the BRIDGE_RT_LOCK, to protect bridge_rtnode lists. That lock is taken in the data path (while in NET_EPOCH), so it cannot be a sleepable lock. While here document the locking strategy. MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D26418
2020-10-06 19:19:56 +00:00
BRIDGE_RT_LOCK_ASSERT(sc);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
/* Make sure we actually need to do this. */
if (sc->sc_brtcnt <= sc->sc_brtmax)
return;
/* Force an aging cycle; this might trim enough addresses. */
bridge_rtage(sc);
if (sc->sc_brtcnt <= sc->sc_brtmax)
return;
bridge: Change lists to CK_LIST as a peparation for epochification Prepare the ground for a rework of the bridge locking approach. We will use an epoch-based approach in the datapath and making it safe to iterate over the interface, span and rtnode lists without holding the BRIDGE_LOCK. Replace the relevant lists by their ConcurrencyKit equivalents. No functional change in this commit. Reviewed by: emaste, ae, philip (previous version) Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24249
2020-04-05 17:15:20 +00:00
CK_LIST_FOREACH_SAFE(brt, &sc->sc_rtlist, brt_list, nbrt) {
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
if ((brt->brt_flags & IFBAF_TYPEMASK) == IFBAF_DYNAMIC) {
bridge_rtnode_destroy(sc, brt);
if (sc->sc_brtcnt <= sc->sc_brtmax)
return;
}
}
}
/*
* bridge_timer:
*
* Aging timer for the bridge.
*/
Change most of the bridge and stp funtions to static. This has highlighted that the following funtions are not used, wrap in '#ifdef noused' for the moment. bstp_enable_change_detection bstp_disable_change_detection bstp_set_bridge_priority bstp_set_port_priority bstp_set_path_cost
2005-10-14 10:38:12 +00:00
static void
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge_timer(void *arg)
{
struct bridge_softc *sc = arg;
bridge: call member interface ioctl() without NET_EPOCH We're not allowed to hold NET_EPOCH while sleeping, so when we call ioctl() handlers for member interfaces we cannot be in NET_EPOCH. We still need some protection of our CK_LISTs, so hold BRIDGE_LOCK instead. That requires changing BRIDGE_LOCK into a sleepable lock, and separating the BRIDGE_RT_LOCK, to protect bridge_rtnode lists. That lock is taken in the data path (while in NET_EPOCH), so it cannot be a sleepable lock. While here document the locking strategy. MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D26418
2020-10-06 19:19:56 +00:00
BRIDGE_RT_LOCK_ASSERT(sc);
Mark the callouts as MPSAFE as if_bridge has been giant-free since day 1. Use the SMP friendly callout_init_mtx() while we are here. Approved by: mlaier (mentor) MFC after: 3 days
2005-08-18 20:17:00 +00:00
if_bridge(4): Fix module teardown bridge_rtnode_zone still has outstanding allocations at the time of destruction in the current model because all of the interface teardown happens in a VNET_SYSUNINIT, -after- the MOD_UNLOAD has already been processed. The SYSUNINIT triggers destruction of the interfaces, which then attempts to free the memory from the zone that's already been destroyed, and we hit a panic. Solve this by virtualizing the uma_zone we allocate the rtnodes from to fix the ordering. bridge_rtable_fini should also take care to flush any remaining routes that weren't taken care of when dynamic routes were flushed in bridge_stop. Reviewed by: kp Differential Revision: https://reviews.freebsd.org/D19578
2019-03-15 13:19:52 +00:00
/* Destruction of rtnodes requires a proper vnet context */
CURVNET_SET(sc->sc_ifp->if_vnet);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge_rtage(sc);
Propagate rename of IFF_OACTIVE and IFF_RUNNING to IFF_DRV_OACTIVE and IFF_DRV_RUNNING, as well as the move from ifnet.if_flags to ifnet.if_drv_flags. Device drivers are now responsible for synchronizing access to these flags, as they are in if_drv_flags. This helps prevent races between the network stack and device driver in maintaining the interface flags field. Many __FreeBSD__ and __FreeBSD_version checks maintained and continued; some less so. Reviewed by: pjd, bz MFC after: 7 days
2005-08-09 10:20:02 +00:00
if (sc->sc_ifp->if_drv_flags & IFF_DRV_RUNNING)
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
callout_reset(&sc->sc_brcallout,
bridge_rtable_prune_period * hz, bridge_timer, sc);
if_bridge(4): Fix module teardown bridge_rtnode_zone still has outstanding allocations at the time of destruction in the current model because all of the interface teardown happens in a VNET_SYSUNINIT, -after- the MOD_UNLOAD has already been processed. The SYSUNINIT triggers destruction of the interfaces, which then attempts to free the memory from the zone that's already been destroyed, and we hit a panic. Solve this by virtualizing the uma_zone we allocate the rtnodes from to fix the ordering. bridge_rtable_fini should also take care to flush any remaining routes that weren't taken care of when dynamic routes were flushed in bridge_stop. Reviewed by: kp Differential Revision: https://reviews.freebsd.org/D19578
2019-03-15 13:19:52 +00:00
CURVNET_RESTORE();
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
}
/*
* bridge_rtage:
*
* Perform an aging cycle.
*/
Change most of the bridge and stp funtions to static. This has highlighted that the following funtions are not used, wrap in '#ifdef noused' for the moment. bstp_enable_change_detection bstp_disable_change_detection bstp_set_bridge_priority bstp_set_port_priority bstp_set_path_cost
2005-10-14 10:38:12 +00:00
static void
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge_rtage(struct bridge_softc *sc)
{
struct bridge_rtnode *brt, *nbrt;
bridge: call member interface ioctl() without NET_EPOCH We're not allowed to hold NET_EPOCH while sleeping, so when we call ioctl() handlers for member interfaces we cannot be in NET_EPOCH. We still need some protection of our CK_LISTs, so hold BRIDGE_LOCK instead. That requires changing BRIDGE_LOCK into a sleepable lock, and separating the BRIDGE_RT_LOCK, to protect bridge_rtnode lists. That lock is taken in the data path (while in NET_EPOCH), so it cannot be a sleepable lock. While here document the locking strategy. MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D26418
2020-10-06 19:19:56 +00:00
BRIDGE_RT_LOCK_ASSERT(sc);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge: Change lists to CK_LIST as a peparation for epochification Prepare the ground for a rework of the bridge locking approach. We will use an epoch-based approach in the datapath and making it safe to iterate over the interface, span and rtnode lists without holding the BRIDGE_LOCK. Replace the relevant lists by their ConcurrencyKit equivalents. No functional change in this commit. Reviewed by: emaste, ae, philip (previous version) Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24249
2020-04-05 17:15:20 +00:00
CK_LIST_FOREACH_SAFE(brt, &sc->sc_rtlist, brt_list, nbrt) {
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
if ((brt->brt_flags & IFBAF_TYPEMASK) == IFBAF_DYNAMIC) {
It is better to use time_uptime here since it is monotonic. Pointed out by: glebius
2006-01-02 22:23:03 +00:00
if (time_uptime >= brt->brt_expire)
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge_rtnode_destroy(sc, brt);
}
}
}
/*
* bridge_rtflush:
*
* Remove all dynamic addresses from the bridge.
*/
Change most of the bridge and stp funtions to static. This has highlighted that the following funtions are not used, wrap in '#ifdef noused' for the moment. bstp_enable_change_detection bstp_disable_change_detection bstp_set_bridge_priority bstp_set_port_priority bstp_set_path_cost
2005-10-14 10:38:12 +00:00
static void
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge_rtflush(struct bridge_softc *sc, int full)
{
struct bridge_rtnode *brt, *nbrt;
bridge: call member interface ioctl() without NET_EPOCH We're not allowed to hold NET_EPOCH while sleeping, so when we call ioctl() handlers for member interfaces we cannot be in NET_EPOCH. We still need some protection of our CK_LISTs, so hold BRIDGE_LOCK instead. That requires changing BRIDGE_LOCK into a sleepable lock, and separating the BRIDGE_RT_LOCK, to protect bridge_rtnode lists. That lock is taken in the data path (while in NET_EPOCH), so it cannot be a sleepable lock. While here document the locking strategy. MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D26418
2020-10-06 19:19:56 +00:00
BRIDGE_RT_LOCK_ASSERT(sc);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge: Change lists to CK_LIST as a peparation for epochification Prepare the ground for a rework of the bridge locking approach. We will use an epoch-based approach in the datapath and making it safe to iterate over the interface, span and rtnode lists without holding the BRIDGE_LOCK. Replace the relevant lists by their ConcurrencyKit equivalents. No functional change in this commit. Reviewed by: emaste, ae, philip (previous version) Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24249
2020-04-05 17:15:20 +00:00
CK_LIST_FOREACH_SAFE(brt, &sc->sc_rtlist, brt_list, nbrt) {
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
if (full || (brt->brt_flags & IFBAF_TYPEMASK) == IFBAF_DYNAMIC)
bridge_rtnode_destroy(sc, brt);
}
}
/*
* bridge_rtdaddr:
*
* Remove an address from the table.
*/
Change most of the bridge and stp funtions to static. This has highlighted that the following funtions are not used, wrap in '#ifdef noused' for the moment. bstp_enable_change_detection bstp_disable_change_detection bstp_set_bridge_priority bstp_set_port_priority bstp_set_path_cost
2005-10-14 10:38:12 +00:00
static int
Add the vlan tag to the bridge route table. This allows a vlan trunk to be bridged, previously legitimate traffic was not passed as the bridge could not tell that it was on a different Ethernet segment. All non-tagged traffic is treated as vlan1 as per IEEE 802.1Q-2003
2007-06-13 18:58:04 +00:00
bridge_rtdaddr(struct bridge_softc *sc, const uint8_t *addr, uint16_t vlan)
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
{
struct bridge_rtnode *brt;
Add the vlan tag to the bridge route table. This allows a vlan trunk to be bridged, previously legitimate traffic was not passed as the bridge could not tell that it was on a different Ethernet segment. All non-tagged traffic is treated as vlan1 as per IEEE 802.1Q-2003
2007-06-13 18:58:04 +00:00
int found = 0;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge: call member interface ioctl() without NET_EPOCH We're not allowed to hold NET_EPOCH while sleeping, so when we call ioctl() handlers for member interfaces we cannot be in NET_EPOCH. We still need some protection of our CK_LISTs, so hold BRIDGE_LOCK instead. That requires changing BRIDGE_LOCK into a sleepable lock, and separating the BRIDGE_RT_LOCK, to protect bridge_rtnode lists. That lock is taken in the data path (while in NET_EPOCH), so it cannot be a sleepable lock. While here document the locking strategy. MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D26418
2020-10-06 19:19:56 +00:00
BRIDGE_RT_LOCK(sc);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
Add the vlan tag to the bridge route table. This allows a vlan trunk to be bridged, previously legitimate traffic was not passed as the bridge could not tell that it was on a different Ethernet segment. All non-tagged traffic is treated as vlan1 as per IEEE 802.1Q-2003
2007-06-13 18:58:04 +00:00
/*
* If vlan is zero then we want to delete for all vlans so the lookup
* may return more than one.
*/
while ((brt = bridge_rtnode_lookup(sc, addr, vlan)) != NULL) {
bridge_rtnode_destroy(sc, brt);
found = 1;
}
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge: call member interface ioctl() without NET_EPOCH We're not allowed to hold NET_EPOCH while sleeping, so when we call ioctl() handlers for member interfaces we cannot be in NET_EPOCH. We still need some protection of our CK_LISTs, so hold BRIDGE_LOCK instead. That requires changing BRIDGE_LOCK into a sleepable lock, and separating the BRIDGE_RT_LOCK, to protect bridge_rtnode lists. That lock is taken in the data path (while in NET_EPOCH), so it cannot be a sleepable lock. While here document the locking strategy. MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D26418
2020-10-06 19:19:56 +00:00
BRIDGE_RT_UNLOCK(sc);
Add the vlan tag to the bridge route table. This allows a vlan trunk to be bridged, previously legitimate traffic was not passed as the bridge could not tell that it was on a different Ethernet segment. All non-tagged traffic is treated as vlan1 as per IEEE 802.1Q-2003
2007-06-13 18:58:04 +00:00
return (found ? 0 : ENOENT);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
}
/*
* bridge_rtdelete:
*
* Delete routes to a speicifc member interface.
*/
Remove the dependency of bridgestp.h on if_bridgevar.h by moving a couple of private structures to if_bridge.c.
2006-07-27 21:01:48 +00:00
static void
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge_rtdelete(struct bridge_softc *sc, struct ifnet *ifp, int full)
{
struct bridge_rtnode *brt, *nbrt;
bridge: call member interface ioctl() without NET_EPOCH We're not allowed to hold NET_EPOCH while sleeping, so when we call ioctl() handlers for member interfaces we cannot be in NET_EPOCH. We still need some protection of our CK_LISTs, so hold BRIDGE_LOCK instead. That requires changing BRIDGE_LOCK into a sleepable lock, and separating the BRIDGE_RT_LOCK, to protect bridge_rtnode lists. That lock is taken in the data path (while in NET_EPOCH), so it cannot be a sleepable lock. While here document the locking strategy. MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D26418
2020-10-06 19:19:56 +00:00
BRIDGE_RT_LOCK_ASSERT(sc);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge: Change lists to CK_LIST as a peparation for epochification Prepare the ground for a rework of the bridge locking approach. We will use an epoch-based approach in the datapath and making it safe to iterate over the interface, span and rtnode lists without holding the BRIDGE_LOCK. Replace the relevant lists by their ConcurrencyKit equivalents. No functional change in this commit. Reviewed by: emaste, ae, philip (previous version) Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24249
2020-04-05 17:15:20 +00:00
CK_LIST_FOREACH_SAFE(brt, &sc->sc_rtlist, brt_list, nbrt) {
Minor whitespace cleanup.
2005-12-17 10:03:48 +00:00
if (brt->brt_ifp == ifp && (full ||
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
(brt->brt_flags & IFBAF_TYPEMASK) == IFBAF_DYNAMIC))
bridge_rtnode_destroy(sc, brt);
}
}
/*
* bridge_rtable_init:
*
* Initialize the route table for this bridge.
*/
Remove the M_NOWAIT from bridge_rtable_init as it isn't needed. The function return value is not even checked and could lead to a panic on a null sc_rthash. MFC after: 2 weeks
2012-10-04 07:40:55 +00:00
static void
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge_rtable_init(struct bridge_softc *sc)
{
int i;
sc->sc_rthash = malloc(sizeof(*sc->sc_rthash) * BRIDGE_RTHASH_SIZE,
Remove the M_NOWAIT from bridge_rtable_init as it isn't needed. The function return value is not even checked and could lead to a panic on a null sc_rthash. MFC after: 2 weeks
2012-10-04 07:40:55 +00:00
M_DEVBUF, M_WAITOK);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
for (i = 0; i < BRIDGE_RTHASH_SIZE; i++)
bridge: Change lists to CK_LIST as a peparation for epochification Prepare the ground for a rework of the bridge locking approach. We will use an epoch-based approach in the datapath and making it safe to iterate over the interface, span and rtnode lists without holding the BRIDGE_LOCK. Replace the relevant lists by their ConcurrencyKit equivalents. No functional change in this commit. Reviewed by: emaste, ae, philip (previous version) Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24249
2020-04-05 17:15:20 +00:00
CK_LIST_INIT(&sc->sc_rthash[i]);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
sc->sc_rthash_key = arc4random();
bridge: Change lists to CK_LIST as a peparation for epochification Prepare the ground for a rework of the bridge locking approach. We will use an epoch-based approach in the datapath and making it safe to iterate over the interface, span and rtnode lists without holding the BRIDGE_LOCK. Replace the relevant lists by their ConcurrencyKit equivalents. No functional change in this commit. Reviewed by: emaste, ae, philip (previous version) Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24249
2020-04-05 17:15:20 +00:00
CK_LIST_INIT(&sc->sc_rtlist);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
}
/*
* bridge_rtable_fini:
*
* Deconstruct the route table for this bridge.
*/
Change most of the bridge and stp funtions to static. This has highlighted that the following funtions are not used, wrap in '#ifdef noused' for the moment. bstp_enable_change_detection bstp_disable_change_detection bstp_set_bridge_priority bstp_set_port_priority bstp_set_path_cost
2005-10-14 10:38:12 +00:00
static void
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge_rtable_fini(struct bridge_softc *sc)
{
Add an option to limit the number of source MACs that can be behind a bridge interface. Once the limit is reached packets with unknown source addresses are dropped until an existing host cache entry expires or is removed. Useful to use with the STICKY cache option. Sponsored by: miniSuperHappyDevHouse NZ
2007-11-04 08:32:27 +00:00
KASSERT(sc->sc_brtcnt == 0,
("%s: %d bridge routes referenced", __func__, sc->sc_brtcnt));
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
free(sc->sc_rthash, M_DEVBUF);
}
/*
* The following hash function is adapted from "Hash Functions" by Bob Jenkins
* ("Algorithm Alley", Dr. Dobbs Journal, September 1997).
*/
#define mix(a, b, c) \
do { \
a -= b; a -= c; a ^= (c >> 13); \
b -= c; b -= a; b ^= (a << 8); \
c -= a; c -= b; c ^= (b >> 13); \
a -= b; a -= c; a ^= (c >> 12); \
b -= c; b -= a; b ^= (a << 16); \
c -= a; c -= b; c ^= (b >> 5); \
a -= b; a -= c; a ^= (c >> 3); \
b -= c; b -= a; b ^= (a << 10); \
c -= a; c -= b; c ^= (b >> 15); \
} while (/*CONSTCOND*/0)
static __inline uint32_t
bridge_rthash(struct bridge_softc *sc, const uint8_t *addr)
{
uint32_t a = 0x9e3779b9, b = 0x9e3779b9, c = sc->sc_rthash_key;
b += addr[5] << 8;
b += addr[4];
a += addr[3] << 24;
a += addr[2] << 16;
a += addr[1] << 8;
a += addr[0];
mix(a, b, c);
return (c & BRIDGE_RTHASH_MASK);
}
#undef mix
Fix two bugs with the bridge - code expects memcmp() to return a signed value, our memcmp() returns 0 if args are equal and > 0 if not. - It's possible to hijack interface for static entry. If bridge recieves packet from interface marked as learning it will replace the bridge_rtnode entry for the source address even if such entry marked as static. Submitted by: Gleb Kurtsov <k-gleb yandex.ru> MFC after: 3 days
2006-01-31 21:21:28 +00:00
static int
bridge_rtnode_addr_cmp(const uint8_t *a, const uint8_t *b)
{
int i, d;
for (i = 0, d = 0; i < ETHER_ADDR_LEN && d == 0; i++) {
d = ((int)a[i]) - ((int)b[i]);
}
return (d);
}
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
/*
* bridge_rtnode_lookup:
*
Add the vlan tag to the bridge route table. This allows a vlan trunk to be bridged, previously legitimate traffic was not passed as the bridge could not tell that it was on a different Ethernet segment. All non-tagged traffic is treated as vlan1 as per IEEE 802.1Q-2003
2007-06-13 18:58:04 +00:00
* Look up a bridge route node for the specified destination. Compare the
* vlan id or if zero then just return the first match.
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
*/
Make four more functions static that were missed in the last commit.
2005-10-14 20:57:02 +00:00
static struct bridge_rtnode *
Add the vlan tag to the bridge route table. This allows a vlan trunk to be bridged, previously legitimate traffic was not passed as the bridge could not tell that it was on a different Ethernet segment. All non-tagged traffic is treated as vlan1 as per IEEE 802.1Q-2003
2007-06-13 18:58:04 +00:00
bridge_rtnode_lookup(struct bridge_softc *sc, const uint8_t *addr, uint16_t vlan)
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
{
struct bridge_rtnode *brt;
uint32_t hash;
int dir;
bridge: call member interface ioctl() without NET_EPOCH We're not allowed to hold NET_EPOCH while sleeping, so when we call ioctl() handlers for member interfaces we cannot be in NET_EPOCH. We still need some protection of our CK_LISTs, so hold BRIDGE_LOCK instead. That requires changing BRIDGE_LOCK into a sleepable lock, and separating the BRIDGE_RT_LOCK, to protect bridge_rtnode lists. That lock is taken in the data path (while in NET_EPOCH), so it cannot be a sleepable lock. While here document the locking strategy. MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D26418
2020-10-06 19:19:56 +00:00
BRIDGE_RT_LOCK_OR_NET_EPOCH_ASSERT(sc);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
hash = bridge_rthash(sc, addr);
bridge: Change lists to CK_LIST as a peparation for epochification Prepare the ground for a rework of the bridge locking approach. We will use an epoch-based approach in the datapath and making it safe to iterate over the interface, span and rtnode lists without holding the BRIDGE_LOCK. Replace the relevant lists by their ConcurrencyKit equivalents. No functional change in this commit. Reviewed by: emaste, ae, philip (previous version) Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24249
2020-04-05 17:15:20 +00:00
CK_LIST_FOREACH(brt, &sc->sc_rthash[hash], brt_hash) {
Fix two bugs with the bridge - code expects memcmp() to return a signed value, our memcmp() returns 0 if args are equal and > 0 if not. - It's possible to hijack interface for static entry. If bridge recieves packet from interface marked as learning it will replace the bridge_rtnode entry for the source address even if such entry marked as static. Submitted by: Gleb Kurtsov <k-gleb yandex.ru> MFC after: 3 days
2006-01-31 21:21:28 +00:00
dir = bridge_rtnode_addr_cmp(addr, brt->brt_addr);
Add the vlan tag to the bridge route table. This allows a vlan trunk to be bridged, previously legitimate traffic was not passed as the bridge could not tell that it was on a different Ethernet segment. All non-tagged traffic is treated as vlan1 as per IEEE 802.1Q-2003
2007-06-13 18:58:04 +00:00
if (dir == 0 && (brt->brt_vlan == vlan || vlan == 0))
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
return (brt);
if (dir > 0)
return (NULL);
}
return (NULL);
}
/*
* bridge_rtnode_insert:
*
* Insert the specified bridge node into the route table. We
* assume the entry is not already in the table.
*/
Change most of the bridge and stp funtions to static. This has highlighted that the following funtions are not used, wrap in '#ifdef noused' for the moment. bstp_enable_change_detection bstp_disable_change_detection bstp_set_bridge_priority bstp_set_port_priority bstp_set_path_cost
2005-10-14 10:38:12 +00:00
static int
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge_rtnode_insert(struct bridge_softc *sc, struct bridge_rtnode *brt)
{
struct bridge_rtnode *lbrt;
uint32_t hash;
int dir;
bridge: call member interface ioctl() without NET_EPOCH We're not allowed to hold NET_EPOCH while sleeping, so when we call ioctl() handlers for member interfaces we cannot be in NET_EPOCH. We still need some protection of our CK_LISTs, so hold BRIDGE_LOCK instead. That requires changing BRIDGE_LOCK into a sleepable lock, and separating the BRIDGE_RT_LOCK, to protect bridge_rtnode lists. That lock is taken in the data path (while in NET_EPOCH), so it cannot be a sleepable lock. While here document the locking strategy. MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D26418
2020-10-06 19:19:56 +00:00
BRIDGE_RT_LOCK_ASSERT(sc);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
hash = bridge_rthash(sc, brt->brt_addr);
bridge: Change lists to CK_LIST as a peparation for epochification Prepare the ground for a rework of the bridge locking approach. We will use an epoch-based approach in the datapath and making it safe to iterate over the interface, span and rtnode lists without holding the BRIDGE_LOCK. Replace the relevant lists by their ConcurrencyKit equivalents. No functional change in this commit. Reviewed by: emaste, ae, philip (previous version) Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24249
2020-04-05 17:15:20 +00:00
lbrt = CK_LIST_FIRST(&sc->sc_rthash[hash]);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
if (lbrt == NULL) {
bridge: Change lists to CK_LIST as a peparation for epochification Prepare the ground for a rework of the bridge locking approach. We will use an epoch-based approach in the datapath and making it safe to iterate over the interface, span and rtnode lists without holding the BRIDGE_LOCK. Replace the relevant lists by their ConcurrencyKit equivalents. No functional change in this commit. Reviewed by: emaste, ae, philip (previous version) Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24249
2020-04-05 17:15:20 +00:00
CK_LIST_INSERT_HEAD(&sc->sc_rthash[hash], brt, brt_hash);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
goto out;
}
do {
Fix two bugs with the bridge - code expects memcmp() to return a signed value, our memcmp() returns 0 if args are equal and > 0 if not. - It's possible to hijack interface for static entry. If bridge recieves packet from interface marked as learning it will replace the bridge_rtnode entry for the source address even if such entry marked as static. Submitted by: Gleb Kurtsov <k-gleb yandex.ru> MFC after: 3 days
2006-01-31 21:21:28 +00:00
dir = bridge_rtnode_addr_cmp(brt->brt_addr, lbrt->brt_addr);
Add the vlan tag to the bridge route table. This allows a vlan trunk to be bridged, previously legitimate traffic was not passed as the bridge could not tell that it was on a different Ethernet segment. All non-tagged traffic is treated as vlan1 as per IEEE 802.1Q-2003
2007-06-13 18:58:04 +00:00
if (dir == 0 && brt->brt_vlan == lbrt->brt_vlan)
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
return (EEXIST);
if (dir > 0) {
bridge: Change lists to CK_LIST as a peparation for epochification Prepare the ground for a rework of the bridge locking approach. We will use an epoch-based approach in the datapath and making it safe to iterate over the interface, span and rtnode lists without holding the BRIDGE_LOCK. Replace the relevant lists by their ConcurrencyKit equivalents. No functional change in this commit. Reviewed by: emaste, ae, philip (previous version) Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24249
2020-04-05 17:15:20 +00:00
CK_LIST_INSERT_BEFORE(lbrt, brt, brt_hash);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
goto out;
}
bridge: Change lists to CK_LIST as a peparation for epochification Prepare the ground for a rework of the bridge locking approach. We will use an epoch-based approach in the datapath and making it safe to iterate over the interface, span and rtnode lists without holding the BRIDGE_LOCK. Replace the relevant lists by their ConcurrencyKit equivalents. No functional change in this commit. Reviewed by: emaste, ae, philip (previous version) Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24249
2020-04-05 17:15:20 +00:00
if (CK_LIST_NEXT(lbrt, brt_hash) == NULL) {
CK_LIST_INSERT_AFTER(lbrt, brt, brt_hash);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
goto out;
}
bridge: Change lists to CK_LIST as a peparation for epochification Prepare the ground for a rework of the bridge locking approach. We will use an epoch-based approach in the datapath and making it safe to iterate over the interface, span and rtnode lists without holding the BRIDGE_LOCK. Replace the relevant lists by their ConcurrencyKit equivalents. No functional change in this commit. Reviewed by: emaste, ae, philip (previous version) Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24249
2020-04-05 17:15:20 +00:00
lbrt = CK_LIST_NEXT(lbrt, brt_hash);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
} while (lbrt != NULL);
#ifdef DIAGNOSTIC
panic("bridge_rtnode_insert: impossible");
#endif
Minor whitespace cleanup.
2005-12-17 10:03:48 +00:00
out:
bridge: Change lists to CK_LIST as a peparation for epochification Prepare the ground for a rework of the bridge locking approach. We will use an epoch-based approach in the datapath and making it safe to iterate over the interface, span and rtnode lists without holding the BRIDGE_LOCK. Replace the relevant lists by their ConcurrencyKit equivalents. No functional change in this commit. Reviewed by: emaste, ae, philip (previous version) Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24249
2020-04-05 17:15:20 +00:00
CK_LIST_INSERT_HEAD(&sc->sc_rtlist, brt, brt_list);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
sc->sc_brtcnt++;
return (0);
}
bridge: epoch-ification Run the bridge datapath under epoch, rather than under the BRIDGE_LOCK(). We still take the BRIDGE_LOCK() whenever we insert or delete items in the relevant lists, but we use epoch callbacks to free items so that it's safe to iterate the lists without the BRIDGE_LOCK. Tests on mercat5/6 shows this increases bridge throughput significantly, from 3.7Mpps to 18.6Mpps. Reviewed by: emaste, philip, melifaro MFC after: 2 months Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24250
2020-04-26 16:22:35 +00:00
static void
bridge_rtnode_destroy_cb(struct epoch_context *ctx)
{
struct bridge_rtnode *brt;
brt = __containerof(ctx, struct bridge_rtnode, brt_epoch_ctx);
CURVNET_SET(brt->brt_vnet);
uma_zfree(V_bridge_rtnode_zone, brt);
CURVNET_RESTORE();
}
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
/*
* bridge_rtnode_destroy:
*
* Destroy a bridge rtnode.
*/
Change most of the bridge and stp funtions to static. This has highlighted that the following funtions are not used, wrap in '#ifdef noused' for the moment. bstp_enable_change_detection bstp_disable_change_detection bstp_set_bridge_priority bstp_set_port_priority bstp_set_path_cost
2005-10-14 10:38:12 +00:00
static void
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge_rtnode_destroy(struct bridge_softc *sc, struct bridge_rtnode *brt)
{
bridge: call member interface ioctl() without NET_EPOCH We're not allowed to hold NET_EPOCH while sleeping, so when we call ioctl() handlers for member interfaces we cannot be in NET_EPOCH. We still need some protection of our CK_LISTs, so hold BRIDGE_LOCK instead. That requires changing BRIDGE_LOCK into a sleepable lock, and separating the BRIDGE_RT_LOCK, to protect bridge_rtnode lists. That lock is taken in the data path (while in NET_EPOCH), so it cannot be a sleepable lock. While here document the locking strategy. MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D26418
2020-10-06 19:19:56 +00:00
BRIDGE_RT_LOCK_ASSERT(sc);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge: Change lists to CK_LIST as a peparation for epochification Prepare the ground for a rework of the bridge locking approach. We will use an epoch-based approach in the datapath and making it safe to iterate over the interface, span and rtnode lists without holding the BRIDGE_LOCK. Replace the relevant lists by their ConcurrencyKit equivalents. No functional change in this commit. Reviewed by: emaste, ae, philip (previous version) Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24249
2020-04-05 17:15:20 +00:00
CK_LIST_REMOVE(brt, brt_hash);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bridge: Change lists to CK_LIST as a peparation for epochification Prepare the ground for a rework of the bridge locking approach. We will use an epoch-based approach in the datapath and making it safe to iterate over the interface, span and rtnode lists without holding the BRIDGE_LOCK. Replace the relevant lists by their ConcurrencyKit equivalents. No functional change in this commit. Reviewed by: emaste, ae, philip (previous version) Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24249
2020-04-05 17:15:20 +00:00
CK_LIST_REMOVE(brt, brt_list);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
sc->sc_brtcnt--;
Add an option to limit the number of source MACs that can be behind a bridge interface. Once the limit is reached packets with unknown source addresses are dropped until an existing host cache entry expires or is removed. Useful to use with the STICKY cache option. Sponsored by: miniSuperHappyDevHouse NZ
2007-11-04 08:32:27 +00:00
brt->brt_dst->bif_addrcnt--;
bridge: epoch-ification Run the bridge datapath under epoch, rather than under the BRIDGE_LOCK(). We still take the BRIDGE_LOCK() whenever we insert or delete items in the relevant lists, but we use epoch callbacks to free items so that it's safe to iterate the lists without the BRIDGE_LOCK. Tests on mercat5/6 shows this increases bridge throughput significantly, from 3.7Mpps to 18.6Mpps. Reviewed by: emaste, philip, melifaro MFC after: 2 months Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24250
2020-04-26 16:22:35 +00:00
NET_EPOCH_CALL(bridge_rtnode_destroy_cb, &brt->brt_epoch_ctx);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
}
Bring in support for the Rapid Spanning Tree Protocol (802.1w). RSTP provides faster spanning tree convergence, the protocol will exchange information with neighboring switches to quickly transition to forwarding without creating loops. The code will default to RSTP mode but will downgrade any port connected to a legacy STP network so is fully backward compatible. Reviewed by: syrinx Tested by: syrinx
2006-11-01 09:07:47 +00:00
/*
* bridge_rtable_expire:
*
* Set the expiry time for all routes on an interface.
*/
static void
bridge_rtable_expire(struct ifnet *ifp, int age)
{
struct bridge_softc *sc = ifp->if_bridge;
struct bridge_rtnode *brt;
bridge: Fix STP-related panic After r345180 we need to have the appropriate vnet context set to delete an rtnode in bridge_rtnode_destroy(). That's usually the case, but not when it's called by the STP code (through bstp_notify_rtage()). We have to set the vnet context in bridge_rtable_expire() just as we do in the other STP callback bridge_state_change(). Reviewed by: kevans
2019-03-15 15:52:36 +00:00
CURVNET_SET(ifp->if_vnet);
bridge: call member interface ioctl() without NET_EPOCH We're not allowed to hold NET_EPOCH while sleeping, so when we call ioctl() handlers for member interfaces we cannot be in NET_EPOCH. We still need some protection of our CK_LISTs, so hold BRIDGE_LOCK instead. That requires changing BRIDGE_LOCK into a sleepable lock, and separating the BRIDGE_RT_LOCK, to protect bridge_rtnode lists. That lock is taken in the data path (while in NET_EPOCH), so it cannot be a sleepable lock. While here document the locking strategy. MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D26418
2020-10-06 19:19:56 +00:00
BRIDGE_RT_LOCK(sc);
Bring in support for the Rapid Spanning Tree Protocol (802.1w). RSTP provides faster spanning tree convergence, the protocol will exchange information with neighboring switches to quickly transition to forwarding without creating loops. The code will default to RSTP mode but will downgrade any port connected to a legacy STP network so is fully backward compatible. Reviewed by: syrinx Tested by: syrinx
2006-11-01 09:07:47 +00:00
/*
* If the age is zero then flush, otherwise set all the expiry times to
* age for the interface
*/
if (age == 0)
bridge_rtdelete(sc, ifp, IFBF_FLUSHDYN);
else {
bridge: Change lists to CK_LIST as a peparation for epochification Prepare the ground for a rework of the bridge locking approach. We will use an epoch-based approach in the datapath and making it safe to iterate over the interface, span and rtnode lists without holding the BRIDGE_LOCK. Replace the relevant lists by their ConcurrencyKit equivalents. No functional change in this commit. Reviewed by: emaste, ae, philip (previous version) Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24249
2020-04-05 17:15:20 +00:00
CK_LIST_FOREACH(brt, &sc->sc_rtlist, brt_list) {
Bring in support for the Rapid Spanning Tree Protocol (802.1w). RSTP provides faster spanning tree convergence, the protocol will exchange information with neighboring switches to quickly transition to forwarding without creating loops. The code will default to RSTP mode but will downgrade any port connected to a legacy STP network so is fully backward compatible. Reviewed by: syrinx Tested by: syrinx
2006-11-01 09:07:47 +00:00
/* Cap the expiry time to 'age' */
if (brt->brt_ifp == ifp &&
brt->brt_expire > time_uptime + age &&
(brt->brt_flags & IFBAF_TYPEMASK) == IFBAF_DYNAMIC)
brt->brt_expire = time_uptime + age;
}
}
bridge: call member interface ioctl() without NET_EPOCH We're not allowed to hold NET_EPOCH while sleeping, so when we call ioctl() handlers for member interfaces we cannot be in NET_EPOCH. We still need some protection of our CK_LISTs, so hold BRIDGE_LOCK instead. That requires changing BRIDGE_LOCK into a sleepable lock, and separating the BRIDGE_RT_LOCK, to protect bridge_rtnode lists. That lock is taken in the data path (while in NET_EPOCH), so it cannot be a sleepable lock. While here document the locking strategy. MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D26418
2020-10-06 19:19:56 +00:00
BRIDGE_RT_UNLOCK(sc);
bridge: Fix STP-related panic After r345180 we need to have the appropriate vnet context set to delete an rtnode in bridge_rtnode_destroy(). That's usually the case, but not when it's called by the STP code (through bstp_notify_rtage()). We have to set the vnet context in bridge_rtable_expire() just as we do in the other STP callback bridge_state_change(). Reviewed by: kevans
2019-03-15 15:52:36 +00:00
CURVNET_RESTORE();
Bring in support for the Rapid Spanning Tree Protocol (802.1w). RSTP provides faster spanning tree convergence, the protocol will exchange information with neighboring switches to quickly transition to forwarding without creating loops. The code will default to RSTP mode but will downgrade any port connected to a legacy STP network so is fully backward compatible. Reviewed by: syrinx Tested by: syrinx
2006-11-01 09:07:47 +00:00
}
- Use the new bridgestp callback to once again flush our bridge routes when an interface is disabled. - Log port changes to syslog, defaulting to off
2006-08-02 03:54:28 +00:00
/*
* bridge_state_change:
*
* Callback from the bridgestp code when a port changes states.
*/
static void
bridge_state_change(struct ifnet *ifp, int state)
{
struct bridge_softc *sc = ifp->if_bridge;
static const char *stpstates[] = {
"disabled",
"listening",
"learning",
"forwarding",
"blocking",
Bring in support for the Rapid Spanning Tree Protocol (802.1w). RSTP provides faster spanning tree convergence, the protocol will exchange information with neighboring switches to quickly transition to forwarding without creating loops. The code will default to RSTP mode but will downgrade any port connected to a legacy STP network so is fully backward compatible. Reviewed by: syrinx Tested by: syrinx
2006-11-01 09:07:47 +00:00
"discarding"
- Use the new bridgestp callback to once again flush our bridge routes when an interface is disabled. - Log port changes to syslog, defaulting to off
2006-08-02 03:54:28 +00:00
};
Fix a panic when VIMAGE is enabled. Spotted by: Nikos Vassiliadis
2015-05-12 03:35:45 +00:00
CURVNET_SET(ifp->if_vnet);
Virtualize if_bridge(4) cloner.
2014-10-05 19:43:37 +00:00
if (V_log_stp)
- Use the new bridgestp callback to once again flush our bridge routes when an interface is disabled. - Log port changes to syslog, defaulting to off
2006-08-02 03:54:28 +00:00
log(LOG_NOTICE, "%s: state changed to %s on %s\n",
sc->sc_ifp->if_xname, stpstates[state], ifp->if_xname);
Fix a panic when VIMAGE is enabled. Spotted by: Nikos Vassiliadis
2015-05-12 03:35:45 +00:00
CURVNET_RESTORE();
- Use the new bridgestp callback to once again flush our bridge routes when an interface is disabled. - Log port changes to syslog, defaulting to off
2006-08-02 03:54:28 +00:00
}
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
/*
* Send bridge packets through pfil if they are one of the types pfil can deal
* with, or if they are ARP or REVARP. (pfil will pass ARP and REVARP without
- Previously when broadcasting to N number of interfaces we would run pfil hooks for each outgoing interface but also run pfil hooks _N times_ on the bridge interface. This is changed so pfil hooks are run once for the bridge interface (bridge0) and then only on the outgoing interfaces in the broadcast loop. - Simplify bridge_enqueue() by moving bridge_pfil() to the callers. - Check (inet6_pfil_hook.ph_busy_count >= 0), it may be possible to have a packet filter hooked for only ipv6 but we were only checking if ipv4 hooks were busy. - Minor optimisation for null mbuf check after bridge_pfil(), move it into the if-block as it couldnt possibly be null outside. Prodded by: mlaier Approved by: re (scottl), mlaier (mentor)
2005-07-06 01:24:45 +00:00
* question.) If *bifp or *ifp are NULL then packet filtering is skipped for
* that interface.
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
*/
Minor whitespace cleanup.
2005-12-17 10:03:48 +00:00
static int
bridge_pfil(struct mbuf **mp, struct ifnet *bifp, struct ifnet *ifp, int dir)
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
{
Recalculate IP checksum after running pfil hooks. Reviewed by: thompsa Tested by: Adam McDougall <mcdouga9@egr.msu.edu>
2006-05-15 11:49:01 +00:00
int snap, error, i, hlen;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
struct ether_header *eh1, eh2;
struct ip *ip;
Sync if_bridge to NetBSD r1.31 Rename conflicting variables when handling SNAP Ethernet frames. Obtained from: NetBSD Approved by: mlaier (mentor) Approved by: re (blanket)
2005-06-29 19:23:32 +00:00
struct llc llc1;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
u_int16_t ether_type;
New pfil(9) KPI together with newborn pfil API and control utility. The KPI have been reviewed and cleansed of features that were planned back 20 years ago and never implemented. The pfil(9) internals have been made opaque to protocols with only returned types and function declarations exposed. The KPI is made more strict, but at the same time more extensible, as kernel uses same command structures that userland ioctl uses. In nutshell [KA]PI is about declaring filtering points, declaring filters and linking and unlinking them together. New [KA]PI makes it possible to reconfigure pfil(9) configuration: change order of hooks, rehook filter from one filtering point to a different one, disconnect a hook on output leaving it on input only, prepend/append a filter to existing list of filters. Now it possible for a single packet filter to provide multiple rulesets that may be linked to different points. Think of per-interface ACLs in Cisco or Juniper. None of existing packet filters yet support that, however limited usage is already possible, e.g. default ruleset can be moved to single interface, as soon as interface would pride their filtering points. Another future feature is possiblity to create pfil heads, that provide not an mbuf pointer but just a memory pointer with length. That would allow filtering at very early stages of a packet lifecycle, e.g. when packet has just been received by a NIC and no mbuf was yet allocated. Differential Revision: https://reviews.freebsd.org/D18951
2019-01-31 23:01:03 +00:00
pfil_return_t rv;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
snap = 0;
error = -1; /* Default error if not error == 0 */
Bring in IPFW layer2 filtering from bridge.c, this allows Ethernet filtering using the layer2, mac and mac-type keywords. This is one of the last features that bridge.c has over if_bridge and gets us very close to a full functional replacement. Approved by: mlaier (mentor)
2005-06-07 21:20:18 +00:00
Remove a KASSERT intended to help the developer, the condition is no longer valid since the span code was added. PR: kern/113170 MFC after: 1 week
2007-05-30 19:39:55 +00:00
#if 0
Assert that the mbuf is not shared to ensure problems like the last commit are not reintroduced.
2006-03-26 20:52:47 +00:00
/* we may return with the IP fields swapped, ensure its not shared */
KASSERT(M_WRITABLE(*mp), ("%s: modifying a shared mbuf", __func__));
Remove a KASSERT intended to help the developer, the condition is no longer valid since the span code was added. PR: kern/113170 MFC after: 1 week
2007-05-30 19:39:55 +00:00
#endif
Assert that the mbuf is not shared to ensure problems like the last commit are not reintroduced.
2006-03-26 20:52:47 +00:00
Virtualize if_bridge(4) cloner.
2014-10-05 19:43:37 +00:00
if (V_pfil_bridge == 0 && V_pfil_member == 0 && V_pfil_ipfw == 0)
Fix style(9) nits, whitespace and parentheses.
2006-05-16 22:50:41 +00:00
return (0); /* filtering is disabled */
When pfil(9) is enabled the bridge only considers ETHERTYPE_ARP, ETHERTYPE_IP and ETHERTYPE_IPV6 frames. Change this to be a sysctl knob so that is able to still bridge non-IP packets if desired. Also return early if all pfil_* sysctls are turned off, the user obviously does not want to filter on the bridge.
2005-12-29 09:39:15 +00:00
Bring in IPFW layer2 filtering from bridge.c, this allows Ethernet filtering using the layer2, mac and mac-type keywords. This is one of the last features that bridge.c has over if_bridge and gets us very close to a full functional replacement. Approved by: mlaier (mentor)
2005-06-07 21:20:18 +00:00
i = min((*mp)->m_pkthdr.len, max_protohdr);
if ((*mp)->m_len < i) {
*mp = m_pullup(*mp, i);
if (*mp == NULL) {
printf("%s: m_pullup failed\n", __func__);
Fix style(9) nits, whitespace and parentheses.
2006-05-16 22:50:41 +00:00
return (-1);
Bring in IPFW layer2 filtering from bridge.c, this allows Ethernet filtering using the layer2, mac and mac-type keywords. This is one of the last features that bridge.c has over if_bridge and gets us very close to a full functional replacement. Approved by: mlaier (mentor)
2005-06-07 21:20:18 +00:00
}
}
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
eh1 = mtod(*mp, struct ether_header *);
ether_type = ntohs(eh1->ether_type);
/*
* Check for SNAP/LLC.
*/
if (ether_type < ETHERMTU) {
Sync if_bridge to NetBSD r1.31 Rename conflicting variables when handling SNAP Ethernet frames. Obtained from: NetBSD Approved by: mlaier (mentor) Approved by: re (blanket)
2005-06-29 19:23:32 +00:00
struct llc *llc2 = (struct llc *)(eh1 + 1);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
if ((*mp)->m_len >= ETHER_HDR_LEN + 8 &&
Sync if_bridge to NetBSD r1.31 Rename conflicting variables when handling SNAP Ethernet frames. Obtained from: NetBSD Approved by: mlaier (mentor) Approved by: re (blanket)
2005-06-29 19:23:32 +00:00
llc2->llc_dsap == LLC_SNAP_LSAP &&
llc2->llc_ssap == LLC_SNAP_LSAP &&
llc2->llc_control == LLC_UI) {
ether_type = htons(llc2->llc_un.type_snap.ether_type);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
snap = 1;
}
}
Revert r1.80 as the ethernet header was inadvertently stripped from ARP packets. Reimplement this correctly and use a sysctl that defaults to off so the user doesnt get any suprises if ipfw blocks the ARP packet. MFC after: 3 days
2006-09-22 21:57:52 +00:00
/*
* If we're trying to filter bridge traffic, don't look at anything
* other than IP and ARP traffic. If the filter doesn't understand
* IPv6, don't allow IPv6 through the bridge either. This is lame
* since if we really wanted, say, an AppleTalk filter, we are hosed,
* but of course we don't have an AppleTalk filter to begin with.
* (Note that since pfil doesn't understand ARP it will pass *ALL*
* ARP traffic.)
*/
switch (ether_type) {
case ETHERTYPE_ARP:
case ETHERTYPE_REVARP:
Virtualize if_bridge(4) cloner.
2014-10-05 19:43:37 +00:00
if (V_pfil_ipfw_arp == 0)
Revert r1.80 as the ethernet header was inadvertently stripped from ARP packets. Reimplement this correctly and use a sysctl that defaults to off so the user doesnt get any suprises if ipfw blocks the ARP packet. MFC after: 3 days
2006-09-22 21:57:52 +00:00
return (0); /* Automatically pass */
break;
case ETHERTYPE_IP:
#ifdef INET6
case ETHERTYPE_IPV6:
#endif /* INET6 */
break;
default:
/*
* Check to see if the user wants to pass non-ip
* packets, these will not be checked by pfil(9) and
* passed unconditionally so the default is to drop.
*/
Virtualize if_bridge(4) cloner.
2014-10-05 19:43:37 +00:00
if (V_pfil_onlyip)
Revert r1.80 as the ethernet header was inadvertently stripped from ARP packets. Reimplement this correctly and use a sysctl that defaults to off so the user doesnt get any suprises if ipfw blocks the ARP packet. MFC after: 3 days
2006-09-22 21:57:52 +00:00
goto bad;
}
Introduce new link-layer PFIL hook V_link_pfil_hook. Merge ether_ipfw_chk() and part of bridge_pfil() into unified ipfw_check_frame() function called by PFIL. This change was suggested by rwatson? @ DevSummit. Remove ipfw headers from ether/bridge code since they are unneeded now. Note this thange introduce some (temporary) performance penalty since PFIL read lock has to be acquired for every link-level packet. MFC after: 3 weeks
2012-09-04 19:43:26 +00:00
/* Run the packet through pfil before stripping link headers */
New pfil(9) KPI together with newborn pfil API and control utility. The KPI have been reviewed and cleansed of features that were planned back 20 years ago and never implemented. The pfil(9) internals have been made opaque to protocols with only returned types and function declarations exposed. The KPI is made more strict, but at the same time more extensible, as kernel uses same command structures that userland ioctl uses. In nutshell [KA]PI is about declaring filtering points, declaring filters and linking and unlinking them together. New [KA]PI makes it possible to reconfigure pfil(9) configuration: change order of hooks, rehook filter from one filtering point to a different one, disconnect a hook on output leaving it on input only, prepend/append a filter to existing list of filters. Now it possible for a single packet filter to provide multiple rulesets that may be linked to different points. Think of per-interface ACLs in Cisco or Juniper. None of existing packet filters yet support that, however limited usage is already possible, e.g. default ruleset can be moved to single interface, as soon as interface would pride their filtering points. Another future feature is possiblity to create pfil heads, that provide not an mbuf pointer but just a memory pointer with length. That would allow filtering at very early stages of a packet lifecycle, e.g. when packet has just been received by a NIC and no mbuf was yet allocated. Differential Revision: https://reviews.freebsd.org/D18951
2019-01-31 23:01:03 +00:00
if (PFIL_HOOKED_OUT(V_link_pfil_head) && V_pfil_ipfw != 0 &&
dir == PFIL_OUT && ifp != NULL) {
switch (pfil_run_hooks(V_link_pfil_head, mp, ifp, dir, NULL)) {
case PFIL_DROPPED:
In r343631 error code for a packet blocked by a firewall was changed from EACCES to EPERM. This change was not intentional, so fix that. Return EACCESS if a firewall forbids sending. Noticed by: ae
2020-01-01 17:31:43 +00:00
return (EACCES);
New pfil(9) KPI together with newborn pfil API and control utility. The KPI have been reviewed and cleansed of features that were planned back 20 years ago and never implemented. The pfil(9) internals have been made opaque to protocols with only returned types and function declarations exposed. The KPI is made more strict, but at the same time more extensible, as kernel uses same command structures that userland ioctl uses. In nutshell [KA]PI is about declaring filtering points, declaring filters and linking and unlinking them together. New [KA]PI makes it possible to reconfigure pfil(9) configuration: change order of hooks, rehook filter from one filtering point to a different one, disconnect a hook on output leaving it on input only, prepend/append a filter to existing list of filters. Now it possible for a single packet filter to provide multiple rulesets that may be linked to different points. Think of per-interface ACLs in Cisco or Juniper. None of existing packet filters yet support that, however limited usage is already possible, e.g. default ruleset can be moved to single interface, as soon as interface would pride their filtering points. Another future feature is possiblity to create pfil heads, that provide not an mbuf pointer but just a memory pointer with length. That would allow filtering at very early stages of a packet lifecycle, e.g. when packet has just been received by a NIC and no mbuf was yet allocated. Differential Revision: https://reviews.freebsd.org/D18951
2019-01-31 23:01:03 +00:00
case PFIL_CONSUMED:
return (0);
}
Introduce new link-layer PFIL hook V_link_pfil_hook. Merge ether_ipfw_chk() and part of bridge_pfil() into unified ipfw_check_frame() function called by PFIL. This change was suggested by rwatson? @ DevSummit. Remove ipfw headers from ether/bridge code since they are unneeded now. Note this thange introduce some (temporary) performance penalty since PFIL read lock has to be acquired for every link-level packet. MFC after: 3 weeks
2012-09-04 19:43:26 +00:00
}
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
/* Strip off the Ethernet header and keep a copy. */
m_copydata(*mp, 0, ETHER_HDR_LEN, (caddr_t) &eh2);
m_adj(*mp, ETHER_HDR_LEN);
/* Strip off snap header, if present */
if (snap) {
Sync if_bridge to NetBSD r1.31 Rename conflicting variables when handling SNAP Ethernet frames. Obtained from: NetBSD Approved by: mlaier (mentor) Approved by: re (blanket)
2005-06-29 19:23:32 +00:00
m_copydata(*mp, 0, sizeof(struct llc), (caddr_t) &llc1);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
m_adj(*mp, sizeof(struct llc));
}
Check the alignment of the IP header before passing the packet up to the packet filter. This would cause a panic on architectures that require strict alignment such as sparc64 (tier1) and ia64/ppc (tier2). This adds two new macros that check the alignment, these are compile time dependent on __NO_STRICT_ALIGNMENT which is set for i386 and amd64 where alignment isn't need so the cost is avoided. IP_HDR_ALIGNED_P() IP6_HDR_ALIGNED_P() Move bridge_ip_checkbasic()/bridge_ip6_checkbasic() up so that the alignment is checked for ipfw and dummynet too. PR: ia64/81284 Obtained from: NetBSD Approved by: re (dwhite), mlaier (mentor)
2005-07-02 23:13:31 +00:00
/*
* Check the IP header for alignment and errors
*/
if (dir == PFIL_IN) {
switch (ether_type) {
case ETHERTYPE_IP:
error = bridge_ip_checkbasic(mp);
break;
Fix style(9) nits, whitespace and parentheses.
2006-05-16 22:50:41 +00:00
#ifdef INET6
Check the alignment of the IP header before passing the packet up to the packet filter. This would cause a panic on architectures that require strict alignment such as sparc64 (tier1) and ia64/ppc (tier2). This adds two new macros that check the alignment, these are compile time dependent on __NO_STRICT_ALIGNMENT which is set for i386 and amd64 where alignment isn't need so the cost is avoided. IP_HDR_ALIGNED_P() IP6_HDR_ALIGNED_P() Move bridge_ip_checkbasic()/bridge_ip6_checkbasic() up so that the alignment is checked for ipfw and dummynet too. PR: ia64/81284 Obtained from: NetBSD Approved by: re (dwhite), mlaier (mentor)
2005-07-02 23:13:31 +00:00
case ETHERTYPE_IPV6:
error = bridge_ip6_checkbasic(mp);
break;
Fix style(9) nits, whitespace and parentheses.
2006-05-16 22:50:41 +00:00
#endif /* INET6 */
Check the alignment of the IP header before passing the packet up to the packet filter. This would cause a panic on architectures that require strict alignment such as sparc64 (tier1) and ia64/ppc (tier2). This adds two new macros that check the alignment, these are compile time dependent on __NO_STRICT_ALIGNMENT which is set for i386 and amd64 where alignment isn't need so the cost is avoided. IP_HDR_ALIGNED_P() IP6_HDR_ALIGNED_P() Move bridge_ip_checkbasic()/bridge_ip6_checkbasic() up so that the alignment is checked for ipfw and dummynet too. PR: ia64/81284 Obtained from: NetBSD Approved by: re (dwhite), mlaier (mentor)
2005-07-02 23:13:31 +00:00
default:
error = 0;
}
if (error)
goto bad;
}
error = 0;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
/*
Revert r1.80 as the ethernet header was inadvertently stripped from ARP packets. Reimplement this correctly and use a sysctl that defaults to off so the user doesnt get any suprises if ipfw blocks the ARP packet. MFC after: 3 days
2006-09-22 21:57:52 +00:00
* Run the packet through pfil
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
*/
New pfil(9) KPI together with newborn pfil API and control utility. The KPI have been reviewed and cleansed of features that were planned back 20 years ago and never implemented. The pfil(9) internals have been made opaque to protocols with only returned types and function declarations exposed. The KPI is made more strict, but at the same time more extensible, as kernel uses same command structures that userland ioctl uses. In nutshell [KA]PI is about declaring filtering points, declaring filters and linking and unlinking them together. New [KA]PI makes it possible to reconfigure pfil(9) configuration: change order of hooks, rehook filter from one filtering point to a different one, disconnect a hook on output leaving it on input only, prepend/append a filter to existing list of filters. Now it possible for a single packet filter to provide multiple rulesets that may be linked to different points. Think of per-interface ACLs in Cisco or Juniper. None of existing packet filters yet support that, however limited usage is already possible, e.g. default ruleset can be moved to single interface, as soon as interface would pride their filtering points. Another future feature is possiblity to create pfil heads, that provide not an mbuf pointer but just a memory pointer with length. That would allow filtering at very early stages of a packet lifecycle, e.g. when packet has just been received by a NIC and no mbuf was yet allocated. Differential Revision: https://reviews.freebsd.org/D18951
2019-01-31 23:01:03 +00:00
rv = PFIL_PASS;
Fix style(9) nits, whitespace and parentheses.
2006-05-16 22:50:41 +00:00
switch (ether_type) {
case ETHERTYPE_IP:
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
/*
* Run pfil on the member interface and the bridge, both can
* be skipped by clearing pfil_member or pfil_bridge.
*
* Keep the order:
* in_if -> bridge_if -> out_if
*/
New pfil(9) KPI together with newborn pfil API and control utility. The KPI have been reviewed and cleansed of features that were planned back 20 years ago and never implemented. The pfil(9) internals have been made opaque to protocols with only returned types and function declarations exposed. The KPI is made more strict, but at the same time more extensible, as kernel uses same command structures that userland ioctl uses. In nutshell [KA]PI is about declaring filtering points, declaring filters and linking and unlinking them together. New [KA]PI makes it possible to reconfigure pfil(9) configuration: change order of hooks, rehook filter from one filtering point to a different one, disconnect a hook on output leaving it on input only, prepend/append a filter to existing list of filters. Now it possible for a single packet filter to provide multiple rulesets that may be linked to different points. Think of per-interface ACLs in Cisco or Juniper. None of existing packet filters yet support that, however limited usage is already possible, e.g. default ruleset can be moved to single interface, as soon as interface would pride their filtering points. Another future feature is possiblity to create pfil heads, that provide not an mbuf pointer but just a memory pointer with length. That would allow filtering at very early stages of a packet lifecycle, e.g. when packet has just been received by a NIC and no mbuf was yet allocated. Differential Revision: https://reviews.freebsd.org/D18951
2019-01-31 23:01:03 +00:00
if (V_pfil_bridge && dir == PFIL_OUT && bifp != NULL && (rv =
pfil_run_hooks(V_inet_pfil_head, mp, bifp, dir, NULL)) !=
PFIL_PASS)
Add dummynet(4) support to if_bridge, this code is largely based on bridge.c. This is the final piece to match bridge.c in functionality, we can now be a drop-in replacement. Approved by: mlaier (mentor)
2005-06-10 01:25:22 +00:00
break;
New pfil(9) KPI together with newborn pfil API and control utility. The KPI have been reviewed and cleansed of features that were planned back 20 years ago and never implemented. The pfil(9) internals have been made opaque to protocols with only returned types and function declarations exposed. The KPI is made more strict, but at the same time more extensible, as kernel uses same command structures that userland ioctl uses. In nutshell [KA]PI is about declaring filtering points, declaring filters and linking and unlinking them together. New [KA]PI makes it possible to reconfigure pfil(9) configuration: change order of hooks, rehook filter from one filtering point to a different one, disconnect a hook on output leaving it on input only, prepend/append a filter to existing list of filters. Now it possible for a single packet filter to provide multiple rulesets that may be linked to different points. Think of per-interface ACLs in Cisco or Juniper. None of existing packet filters yet support that, however limited usage is already possible, e.g. default ruleset can be moved to single interface, as soon as interface would pride their filtering points. Another future feature is possiblity to create pfil heads, that provide not an mbuf pointer but just a memory pointer with length. That would allow filtering at very early stages of a packet lifecycle, e.g. when packet has just been received by a NIC and no mbuf was yet allocated. Differential Revision: https://reviews.freebsd.org/D18951
2019-01-31 23:01:03 +00:00
if (V_pfil_member && ifp != NULL && (rv =
pfil_run_hooks(V_inet_pfil_head, mp, ifp, dir, NULL)) !=
PFIL_PASS)
Add dummynet(4) support to if_bridge, this code is largely based on bridge.c. This is the final piece to match bridge.c in functionality, we can now be a drop-in replacement. Approved by: mlaier (mentor)
2005-06-10 01:25:22 +00:00
break;
New pfil(9) KPI together with newborn pfil API and control utility. The KPI have been reviewed and cleansed of features that were planned back 20 years ago and never implemented. The pfil(9) internals have been made opaque to protocols with only returned types and function declarations exposed. The KPI is made more strict, but at the same time more extensible, as kernel uses same command structures that userland ioctl uses. In nutshell [KA]PI is about declaring filtering points, declaring filters and linking and unlinking them together. New [KA]PI makes it possible to reconfigure pfil(9) configuration: change order of hooks, rehook filter from one filtering point to a different one, disconnect a hook on output leaving it on input only, prepend/append a filter to existing list of filters. Now it possible for a single packet filter to provide multiple rulesets that may be linked to different points. Think of per-interface ACLs in Cisco or Juniper. None of existing packet filters yet support that, however limited usage is already possible, e.g. default ruleset can be moved to single interface, as soon as interface would pride their filtering points. Another future feature is possiblity to create pfil heads, that provide not an mbuf pointer but just a memory pointer with length. That would allow filtering at very early stages of a packet lifecycle, e.g. when packet has just been received by a NIC and no mbuf was yet allocated. Differential Revision: https://reviews.freebsd.org/D18951
2019-01-31 23:01:03 +00:00
if (V_pfil_bridge && dir == PFIL_IN && bifp != NULL && (rv =
pfil_run_hooks(V_inet_pfil_head, mp, bifp, dir, NULL)) !=
PFIL_PASS)
Add support for fragmenting ipv4 packets. The packet filter may reassemble the ip fragments and return a packet that is larger than the MTU of the sending interface. There is no check for DF or icmp replies as we can only get a large packet to fragment by reassembling a previous fragment, and this only happens after a call to pfil(9). Obtained from: OpenBSD (mostly) Glanced at by: mlaier MFC after: 1 month
2006-04-29 05:37:25 +00:00
break;
/* check if we need to fragment the packet */
bridge: Fix fragment handling and memory leak Fragmented UDP and ICMP packets were corrupted if a firewall with reassembling feature (like pf'scrub) is enabled on the bridge. This patch fixes corrupted packet problem and the panic (triggered easly with low RAM) as explain in PR 185633. bridge_pfil and bridge_fragment relationship: bridge_pfil() receive (IN direction) packets and sent it to the firewall The firewall can be configured for reassembling fragmented packet (like pf'scrubing) in one mbuf chain when bridge_pfil() need to send this reassembled packet to the outgoing interface, it needs to re-fragment it by using bridge_fragment() bridge_fragment() had to split this mbuf (using ip_fragment) first then had to M_PREPEND each packet in the mbuf chain for adding Ethernet header. But M_PREPEND can sometime create a new mbuf on the begining of the mbuf chain, then the "main" pointer of this mbuf chain should be updated and this case is tottaly forgotten. The original bridge_fragment code (Revision 158140, 2006 April 29) came from OpenBSD, and the call to bridge_enqueue was embedded. But on FreeBSD, bridge_enqueue() is done after bridge_fragment(), then the original OpenBSD code can't work as-it of FreeBSD. PR: 185633 Submitted by: Olivier Cochard-Labbé Differential Revision: https://reviews.freebsd.org/D7780
2016-09-24 07:09:43 +00:00
/* bridge_fragment generates a mbuf chain of packets */
/* that already include eth headers */
Virtualize if_bridge(4) cloner.
2014-10-05 19:43:37 +00:00
if (V_pfil_member && ifp != NULL && dir == PFIL_OUT) {
Add support for fragmenting ipv4 packets. The packet filter may reassemble the ip fragments and return a packet that is larger than the MTU of the sending interface. There is no check for DF or icmp replies as we can only get a large packet to fragment by reassembling a previous fragment, and this only happens after a call to pfil(9). Obtained from: OpenBSD (mostly) Glanced at by: mlaier MFC after: 1 month
2006-04-29 05:37:25 +00:00
i = (*mp)->m_pkthdr.len;
if (i > ifp->if_mtu) {
bridge: Fix fragment handling and memory leak Fragmented UDP and ICMP packets were corrupted if a firewall with reassembling feature (like pf'scrub) is enabled on the bridge. This patch fixes corrupted packet problem and the panic (triggered easly with low RAM) as explain in PR 185633. bridge_pfil and bridge_fragment relationship: bridge_pfil() receive (IN direction) packets and sent it to the firewall The firewall can be configured for reassembling fragmented packet (like pf'scrubing) in one mbuf chain when bridge_pfil() need to send this reassembled packet to the outgoing interface, it needs to re-fragment it by using bridge_fragment() bridge_fragment() had to split this mbuf (using ip_fragment) first then had to M_PREPEND each packet in the mbuf chain for adding Ethernet header. But M_PREPEND can sometime create a new mbuf on the begining of the mbuf chain, then the "main" pointer of this mbuf chain should be updated and this case is tottaly forgotten. The original bridge_fragment code (Revision 158140, 2006 April 29) came from OpenBSD, and the call to bridge_enqueue was embedded. But on FreeBSD, bridge_enqueue() is done after bridge_fragment(), then the original OpenBSD code can't work as-it of FreeBSD. PR: 185633 Submitted by: Olivier Cochard-Labbé Differential Revision: https://reviews.freebsd.org/D7780
2016-09-24 07:09:43 +00:00
error = bridge_fragment(ifp, mp, &eh2, snap,
Add support for fragmenting ipv4 packets. The packet filter may reassemble the ip fragments and return a packet that is larger than the MTU of the sending interface. There is no check for DF or icmp replies as we can only get a large packet to fragment by reassembling a previous fragment, and this only happens after a call to pfil(9). Obtained from: OpenBSD (mostly) Glanced at by: mlaier MFC after: 1 month
2006-04-29 05:37:25 +00:00
&llc1);
return (error);
}
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
}
A step in resolving mess with byte ordering for AF_INET. After this change: - All packets in NETISR_IP queue are in net byte order. - ip_input() is entered in net byte order and converts packet to host byte order right _after_ processing pfil(9) hooks. - ip_output() is entered in host byte order and converts packet to net byte order right _before_ processing pfil(9) hooks. - ip_fragment() accepts and emits packet in net byte order. - ip_forward(), ip_mloopback() use host byte order (untouched actually). - ip_fastforward() no longer modifies packet at all (except ip_ttl). - Swapping of byte order there and back removed from the following modules: pf(4), ipfw(4), enc(4), if_bridge(4). - Swapping of byte order added to ipfilter(4), based on __FreeBSD_version - __FreeBSD_version bumped. - pfil(9) manual page updated. Reviewed by: ray, luigi, eri, melifaro Tested by: glebius (LE), ray (BE)
2012-10-06 10:02:11 +00:00
/* Recalculate the ip checksum. */
Add support for fragmenting ipv4 packets. The packet filter may reassemble the ip fragments and return a packet that is larger than the MTU of the sending interface. There is no check for DF or icmp replies as we can only get a large packet to fragment by reassembling a previous fragment, and this only happens after a call to pfil(9). Obtained from: OpenBSD (mostly) Glanced at by: mlaier MFC after: 1 month
2006-04-29 05:37:25 +00:00
ip = mtod(*mp, struct ip *);
Recalculate IP checksum after running pfil hooks. Reviewed by: thompsa Tested by: Adam McDougall <mcdouga9@egr.msu.edu>
2006-05-15 11:49:01 +00:00
hlen = ip->ip_hl << 2;
Fix style(9) nits, whitespace and parentheses.
2006-05-16 22:50:41 +00:00
if (hlen < sizeof(struct ip))
goto bad;
if (hlen > (*mp)->m_len) {
sys/net* : for pointers replace 0 with NULL. Mostly cosmetical, no functional change. Found with devel/coccinelle.
2016-04-15 17:30:33 +00:00
if ((*mp = m_pullup(*mp, hlen)) == NULL)
Fix style(9) nits, whitespace and parentheses.
2006-05-16 22:50:41 +00:00
goto bad;
ip = mtod(*mp, struct ip *);
if (ip == NULL)
goto bad;
}
ip->ip_sum = 0;
if (hlen == sizeof(struct ip))
ip->ip_sum = in_cksum_hdr(ip);
else
ip->ip_sum = in_cksum(*mp, hlen);
Add support for fragmenting ipv4 packets. The packet filter may reassemble the ip fragments and return a packet that is larger than the MTU of the sending interface. There is no check for DF or icmp replies as we can only get a large packet to fragment by reassembling a previous fragment, and this only happens after a call to pfil(9). Obtained from: OpenBSD (mostly) Glanced at by: mlaier MFC after: 1 month
2006-04-29 05:37:25 +00:00
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
break;
Fix style(9) nits, whitespace and parentheses.
2006-05-16 22:50:41 +00:00
#ifdef INET6
case ETHERTYPE_IPV6:
New pfil(9) KPI together with newborn pfil API and control utility. The KPI have been reviewed and cleansed of features that were planned back 20 years ago and never implemented. The pfil(9) internals have been made opaque to protocols with only returned types and function declarations exposed. The KPI is made more strict, but at the same time more extensible, as kernel uses same command structures that userland ioctl uses. In nutshell [KA]PI is about declaring filtering points, declaring filters and linking and unlinking them together. New [KA]PI makes it possible to reconfigure pfil(9) configuration: change order of hooks, rehook filter from one filtering point to a different one, disconnect a hook on output leaving it on input only, prepend/append a filter to existing list of filters. Now it possible for a single packet filter to provide multiple rulesets that may be linked to different points. Think of per-interface ACLs in Cisco or Juniper. None of existing packet filters yet support that, however limited usage is already possible, e.g. default ruleset can be moved to single interface, as soon as interface would pride their filtering points. Another future feature is possiblity to create pfil heads, that provide not an mbuf pointer but just a memory pointer with length. That would allow filtering at very early stages of a packet lifecycle, e.g. when packet has just been received by a NIC and no mbuf was yet allocated. Differential Revision: https://reviews.freebsd.org/D18951
2019-01-31 23:01:03 +00:00
if (V_pfil_bridge && dir == PFIL_OUT && bifp != NULL && (rv =
pfil_run_hooks(V_inet6_pfil_head, mp, bifp, dir, NULL)) !=
PFIL_PASS)
Add dummynet(4) support to if_bridge, this code is largely based on bridge.c. This is the final piece to match bridge.c in functionality, we can now be a drop-in replacement. Approved by: mlaier (mentor)
2005-06-10 01:25:22 +00:00
break;
New pfil(9) KPI together with newborn pfil API and control utility. The KPI have been reviewed and cleansed of features that were planned back 20 years ago and never implemented. The pfil(9) internals have been made opaque to protocols with only returned types and function declarations exposed. The KPI is made more strict, but at the same time more extensible, as kernel uses same command structures that userland ioctl uses. In nutshell [KA]PI is about declaring filtering points, declaring filters and linking and unlinking them together. New [KA]PI makes it possible to reconfigure pfil(9) configuration: change order of hooks, rehook filter from one filtering point to a different one, disconnect a hook on output leaving it on input only, prepend/append a filter to existing list of filters. Now it possible for a single packet filter to provide multiple rulesets that may be linked to different points. Think of per-interface ACLs in Cisco or Juniper. None of existing packet filters yet support that, however limited usage is already possible, e.g. default ruleset can be moved to single interface, as soon as interface would pride their filtering points. Another future feature is possiblity to create pfil heads, that provide not an mbuf pointer but just a memory pointer with length. That would allow filtering at very early stages of a packet lifecycle, e.g. when packet has just been received by a NIC and no mbuf was yet allocated. Differential Revision: https://reviews.freebsd.org/D18951
2019-01-31 23:01:03 +00:00
if (V_pfil_member && ifp != NULL && (rv =
pfil_run_hooks(V_inet6_pfil_head, mp, ifp, dir, NULL)) !=
PFIL_PASS)
Add dummynet(4) support to if_bridge, this code is largely based on bridge.c. This is the final piece to match bridge.c in functionality, we can now be a drop-in replacement. Approved by: mlaier (mentor)
2005-06-10 01:25:22 +00:00
break;
New pfil(9) KPI together with newborn pfil API and control utility. The KPI have been reviewed and cleansed of features that were planned back 20 years ago and never implemented. The pfil(9) internals have been made opaque to protocols with only returned types and function declarations exposed. The KPI is made more strict, but at the same time more extensible, as kernel uses same command structures that userland ioctl uses. In nutshell [KA]PI is about declaring filtering points, declaring filters and linking and unlinking them together. New [KA]PI makes it possible to reconfigure pfil(9) configuration: change order of hooks, rehook filter from one filtering point to a different one, disconnect a hook on output leaving it on input only, prepend/append a filter to existing list of filters. Now it possible for a single packet filter to provide multiple rulesets that may be linked to different points. Think of per-interface ACLs in Cisco or Juniper. None of existing packet filters yet support that, however limited usage is already possible, e.g. default ruleset can be moved to single interface, as soon as interface would pride their filtering points. Another future feature is possiblity to create pfil heads, that provide not an mbuf pointer but just a memory pointer with length. That would allow filtering at very early stages of a packet lifecycle, e.g. when packet has just been received by a NIC and no mbuf was yet allocated. Differential Revision: https://reviews.freebsd.org/D18951
2019-01-31 23:01:03 +00:00
if (V_pfil_bridge && dir == PFIL_IN && bifp != NULL && (rv =
pfil_run_hooks(V_inet6_pfil_head, mp, bifp, dir, NULL)) !=
PFIL_PASS)
break;
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
break;
Fix style(9) nits, whitespace and parentheses.
2006-05-16 22:50:41 +00:00
#endif
New pfil(9) KPI together with newborn pfil API and control utility. The KPI have been reviewed and cleansed of features that were planned back 20 years ago and never implemented. The pfil(9) internals have been made opaque to protocols with only returned types and function declarations exposed. The KPI is made more strict, but at the same time more extensible, as kernel uses same command structures that userland ioctl uses. In nutshell [KA]PI is about declaring filtering points, declaring filters and linking and unlinking them together. New [KA]PI makes it possible to reconfigure pfil(9) configuration: change order of hooks, rehook filter from one filtering point to a different one, disconnect a hook on output leaving it on input only, prepend/append a filter to existing list of filters. Now it possible for a single packet filter to provide multiple rulesets that may be linked to different points. Think of per-interface ACLs in Cisco or Juniper. None of existing packet filters yet support that, however limited usage is already possible, e.g. default ruleset can be moved to single interface, as soon as interface would pride their filtering points. Another future feature is possiblity to create pfil heads, that provide not an mbuf pointer but just a memory pointer with length. That would allow filtering at very early stages of a packet lifecycle, e.g. when packet has just been received by a NIC and no mbuf was yet allocated. Differential Revision: https://reviews.freebsd.org/D18951
2019-01-31 23:01:03 +00:00
}
switch (rv) {
case PFIL_CONSUMED:
return (0);
case PFIL_DROPPED:
In r343631 error code for a packet blocked by a firewall was changed from EACCES to EPERM. This change was not intentional, so fix that. Return EACCESS if a firewall forbids sending. Noticed by: ae
2020-01-01 17:31:43 +00:00
return (EACCES);
Fix style(9) nits, whitespace and parentheses.
2006-05-16 22:50:41 +00:00
default:
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
break;
}
error = -1;
/*
* Finally, put everything back the way it was and return
*/
if (snap) {
Mechanically substitute flags from historic mbuf allocator with malloc(9) flags within sys. Exceptions: - sys/contrib not touched - sys/mbuf.h edited manually
2012-12-05 08:04:20 +00:00
M_PREPEND(*mp, sizeof(struct llc), M_NOWAIT);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
if (*mp == NULL)
Fix style(9) nits, whitespace and parentheses.
2006-05-16 22:50:41 +00:00
return (error);
Sync if_bridge to NetBSD r1.31 Rename conflicting variables when handling SNAP Ethernet frames. Obtained from: NetBSD Approved by: mlaier (mentor) Approved by: re (blanket)
2005-06-29 19:23:32 +00:00
bcopy(&llc1, mtod(*mp, caddr_t), sizeof(struct llc));
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
}
Mechanically substitute flags from historic mbuf allocator with malloc(9) flags within sys. Exceptions: - sys/contrib not touched - sys/mbuf.h edited manually
2012-12-05 08:04:20 +00:00
M_PREPEND(*mp, ETHER_HDR_LEN, M_NOWAIT);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
if (*mp == NULL)
Fix style(9) nits, whitespace and parentheses.
2006-05-16 22:50:41 +00:00
return (error);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
bcopy(&eh2, mtod(*mp, caddr_t), ETHER_HDR_LEN);
Fix style(9) nits, whitespace and parentheses.
2006-05-16 22:50:41 +00:00
return (0);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
Minor whitespace cleanup.
2005-12-17 10:03:48 +00:00
bad:
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
m_freem(*mp);
*mp = NULL;
Fix style(9) nits, whitespace and parentheses.
2006-05-16 22:50:41 +00:00
return (error);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
}
/*
* Perform basic checks on header size since
* pfil assumes ip_input has already processed
* it for it. Cut-and-pasted from ip_input.c.
* Given how simple the IPv6 version is,
* does the IPv4 version really need to be
* this complicated?
*
* XXX Should we update ipstat here, or not?
* XXX Right now we update ipstat but not
* XXX csum_counter.
*/
static int
bridge_ip_checkbasic(struct mbuf **mp)
{
struct mbuf *m = *mp;
struct ip *ip;
int len, hlen;
u_short sum;
if (*mp == NULL)
Fix style(9) nits, whitespace and parentheses.
2006-05-16 22:50:41 +00:00
return (-1);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
Check the alignment of the IP header before passing the packet up to the packet filter. This would cause a panic on architectures that require strict alignment such as sparc64 (tier1) and ia64/ppc (tier2). This adds two new macros that check the alignment, these are compile time dependent on __NO_STRICT_ALIGNMENT which is set for i386 and amd64 where alignment isn't need so the cost is avoided. IP_HDR_ALIGNED_P() IP6_HDR_ALIGNED_P() Move bridge_ip_checkbasic()/bridge_ip6_checkbasic() up so that the alignment is checked for ipfw and dummynet too. PR: ia64/81284 Obtained from: NetBSD Approved by: re (dwhite), mlaier (mentor)
2005-07-02 23:13:31 +00:00
if (IP_HDR_ALIGNED_P(mtod(m, caddr_t)) == 0) {
if ((m = m_copyup(m, sizeof(struct ip),
(max_linkhdr + 3) & ~3)) == NULL) {
/* XXXJRT new stat, please */
Many network stack subsystems use a single global data structure to hold all pertinent statatistics for the subsystem. These structures are sometimes "borrowed" by kernel modules that require a place to store statistics for similar events. Add KPI accessor functions for statistics structures referenced by kernel modules so that they no longer encode certain specifics of how the data structures are named and stored. This change is intended to make it easier to move to per-CPU network stats following 8.0-RELEASE. The following modules are affected by this change: if_bridge if_cxgb if_gif ip_mroute ipdivert pf In practice, most of these statistics consumers should, in fact, maintain their own statistics data structures rather than borrowing structures from the base network stack. However, that change is too agressive for this point in the release cycle. Reviewed by: bz Approved by: re (kib)
2009-08-02 19:43:32 +00:00
KMOD_IPSTAT_INC(ips_toosmall);
Check the alignment of the IP header before passing the packet up to the packet filter. This would cause a panic on architectures that require strict alignment such as sparc64 (tier1) and ia64/ppc (tier2). This adds two new macros that check the alignment, these are compile time dependent on __NO_STRICT_ALIGNMENT which is set for i386 and amd64 where alignment isn't need so the cost is avoided. IP_HDR_ALIGNED_P() IP6_HDR_ALIGNED_P() Move bridge_ip_checkbasic()/bridge_ip6_checkbasic() up so that the alignment is checked for ipfw and dummynet too. PR: ia64/81284 Obtained from: NetBSD Approved by: re (dwhite), mlaier (mentor)
2005-07-02 23:13:31 +00:00
goto bad;
}
} else if (__predict_false(m->m_len < sizeof (struct ip))) {
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
if ((m = m_pullup(m, sizeof (struct ip))) == NULL) {
Many network stack subsystems use a single global data structure to hold all pertinent statatistics for the subsystem. These structures are sometimes "borrowed" by kernel modules that require a place to store statistics for similar events. Add KPI accessor functions for statistics structures referenced by kernel modules so that they no longer encode certain specifics of how the data structures are named and stored. This change is intended to make it easier to move to per-CPU network stats following 8.0-RELEASE. The following modules are affected by this change: if_bridge if_cxgb if_gif ip_mroute ipdivert pf In practice, most of these statistics consumers should, in fact, maintain their own statistics data structures rather than borrowing structures from the base network stack. However, that change is too agressive for this point in the release cycle. Reviewed by: bz Approved by: re (kib)
2009-08-02 19:43:32 +00:00
KMOD_IPSTAT_INC(ips_toosmall);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
goto bad;
}
}
ip = mtod(m, struct ip *);
if (ip == NULL) goto bad;
if (ip->ip_v != IPVERSION) {
Many network stack subsystems use a single global data structure to hold all pertinent statatistics for the subsystem. These structures are sometimes "borrowed" by kernel modules that require a place to store statistics for similar events. Add KPI accessor functions for statistics structures referenced by kernel modules so that they no longer encode certain specifics of how the data structures are named and stored. This change is intended to make it easier to move to per-CPU network stats following 8.0-RELEASE. The following modules are affected by this change: if_bridge if_cxgb if_gif ip_mroute ipdivert pf In practice, most of these statistics consumers should, in fact, maintain their own statistics data structures rather than borrowing structures from the base network stack. However, that change is too agressive for this point in the release cycle. Reviewed by: bz Approved by: re (kib)
2009-08-02 19:43:32 +00:00
KMOD_IPSTAT_INC(ips_badvers);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
goto bad;
}
hlen = ip->ip_hl << 2;
if (hlen < sizeof(struct ip)) { /* minimum header length */
Many network stack subsystems use a single global data structure to hold all pertinent statatistics for the subsystem. These structures are sometimes "borrowed" by kernel modules that require a place to store statistics for similar events. Add KPI accessor functions for statistics structures referenced by kernel modules so that they no longer encode certain specifics of how the data structures are named and stored. This change is intended to make it easier to move to per-CPU network stats following 8.0-RELEASE. The following modules are affected by this change: if_bridge if_cxgb if_gif ip_mroute ipdivert pf In practice, most of these statistics consumers should, in fact, maintain their own statistics data structures rather than borrowing structures from the base network stack. However, that change is too agressive for this point in the release cycle. Reviewed by: bz Approved by: re (kib)
2009-08-02 19:43:32 +00:00
KMOD_IPSTAT_INC(ips_badhlen);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
goto bad;
}
if (hlen > m->m_len) {
sys/net* : for pointers replace 0 with NULL. Mostly cosmetical, no functional change. Found with devel/coccinelle.
2016-04-15 17:30:33 +00:00
if ((m = m_pullup(m, hlen)) == NULL) {
Many network stack subsystems use a single global data structure to hold all pertinent statatistics for the subsystem. These structures are sometimes "borrowed" by kernel modules that require a place to store statistics for similar events. Add KPI accessor functions for statistics structures referenced by kernel modules so that they no longer encode certain specifics of how the data structures are named and stored. This change is intended to make it easier to move to per-CPU network stats following 8.0-RELEASE. The following modules are affected by this change: if_bridge if_cxgb if_gif ip_mroute ipdivert pf In practice, most of these statistics consumers should, in fact, maintain their own statistics data structures rather than borrowing structures from the base network stack. However, that change is too agressive for this point in the release cycle. Reviewed by: bz Approved by: re (kib)
2009-08-02 19:43:32 +00:00
KMOD_IPSTAT_INC(ips_badhlen);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
goto bad;
}
ip = mtod(m, struct ip *);
if (ip == NULL) goto bad;
}
if (m->m_pkthdr.csum_flags & CSUM_IP_CHECKED) {
sum = !(m->m_pkthdr.csum_flags & CSUM_IP_VALID);
} else {
if (hlen == sizeof(struct ip)) {
sum = in_cksum_hdr(ip);
} else {
sum = in_cksum(m, hlen);
}
}
if (sum) {
Many network stack subsystems use a single global data structure to hold all pertinent statatistics for the subsystem. These structures are sometimes "borrowed" by kernel modules that require a place to store statistics for similar events. Add KPI accessor functions for statistics structures referenced by kernel modules so that they no longer encode certain specifics of how the data structures are named and stored. This change is intended to make it easier to move to per-CPU network stats following 8.0-RELEASE. The following modules are affected by this change: if_bridge if_cxgb if_gif ip_mroute ipdivert pf In practice, most of these statistics consumers should, in fact, maintain their own statistics data structures rather than borrowing structures from the base network stack. However, that change is too agressive for this point in the release cycle. Reviewed by: bz Approved by: re (kib)
2009-08-02 19:43:32 +00:00
KMOD_IPSTAT_INC(ips_badsum);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
goto bad;
}
/* Retrieve the packet length. */
len = ntohs(ip->ip_len);
/*
* Check for additional length bogosity
*/
if (len < hlen) {
Many network stack subsystems use a single global data structure to hold all pertinent statatistics for the subsystem. These structures are sometimes "borrowed" by kernel modules that require a place to store statistics for similar events. Add KPI accessor functions for statistics structures referenced by kernel modules so that they no longer encode certain specifics of how the data structures are named and stored. This change is intended to make it easier to move to per-CPU network stats following 8.0-RELEASE. The following modules are affected by this change: if_bridge if_cxgb if_gif ip_mroute ipdivert pf In practice, most of these statistics consumers should, in fact, maintain their own statistics data structures rather than borrowing structures from the base network stack. However, that change is too agressive for this point in the release cycle. Reviewed by: bz Approved by: re (kib)
2009-08-02 19:43:32 +00:00
KMOD_IPSTAT_INC(ips_badlen);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
goto bad;
}
/*
* Check that the amount of data in the buffers
* is as at least much as the IP header would have us expect.
* Drop packet if shorter than we expect.
*/
if (m->m_pkthdr.len < len) {
Many network stack subsystems use a single global data structure to hold all pertinent statatistics for the subsystem. These structures are sometimes "borrowed" by kernel modules that require a place to store statistics for similar events. Add KPI accessor functions for statistics structures referenced by kernel modules so that they no longer encode certain specifics of how the data structures are named and stored. This change is intended to make it easier to move to per-CPU network stats following 8.0-RELEASE. The following modules are affected by this change: if_bridge if_cxgb if_gif ip_mroute ipdivert pf In practice, most of these statistics consumers should, in fact, maintain their own statistics data structures rather than borrowing structures from the base network stack. However, that change is too agressive for this point in the release cycle. Reviewed by: bz Approved by: re (kib)
2009-08-02 19:43:32 +00:00
KMOD_IPSTAT_INC(ips_tooshort);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
goto bad;
}
/* Checks out, proceed */
*mp = m;
Fix style(9) nits, whitespace and parentheses.
2006-05-16 22:50:41 +00:00
return (0);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
Minor whitespace cleanup.
2005-12-17 10:03:48 +00:00
bad:
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
*mp = m;
Fix style(9) nits, whitespace and parentheses.
2006-05-16 22:50:41 +00:00
return (-1);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
}
Fix style(9) nits, whitespace and parentheses.
2006-05-16 22:50:41 +00:00
#ifdef INET6
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
/*
* Same as above, but for IPv6.
* Cut-and-pasted from ip6_input.c.
* XXX Should we update ip6stat, or not?
*/
static int
bridge_ip6_checkbasic(struct mbuf **mp)
{
struct mbuf *m = *mp;
struct ip6_hdr *ip6;
/*
* If the IPv6 header is not aligned, slurp it up into a new
* mbuf with space for link headers, in the event we forward
* it. Otherwise, if it is aligned, make sure the entire base
* IPv6 header is in the first mbuf of the chain.
Check the alignment of the IP header before passing the packet up to the packet filter. This would cause a panic on architectures that require strict alignment such as sparc64 (tier1) and ia64/ppc (tier2). This adds two new macros that check the alignment, these are compile time dependent on __NO_STRICT_ALIGNMENT which is set for i386 and amd64 where alignment isn't need so the cost is avoided. IP_HDR_ALIGNED_P() IP6_HDR_ALIGNED_P() Move bridge_ip_checkbasic()/bridge_ip6_checkbasic() up so that the alignment is checked for ipfw and dummynet too. PR: ia64/81284 Obtained from: NetBSD Approved by: re (dwhite), mlaier (mentor)
2005-07-02 23:13:31 +00:00
*/
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
if (IP6_HDR_ALIGNED_P(mtod(m, caddr_t)) == 0) {
struct ifnet *inifp = m->m_pkthdr.rcvif;
if ((m = m_copyup(m, sizeof(struct ip6_hdr),
(max_linkhdr + 3) & ~3)) == NULL) {
Check the alignment of the IP header before passing the packet up to the packet filter. This would cause a panic on architectures that require strict alignment such as sparc64 (tier1) and ia64/ppc (tier2). This adds two new macros that check the alignment, these are compile time dependent on __NO_STRICT_ALIGNMENT which is set for i386 and amd64 where alignment isn't need so the cost is avoided. IP_HDR_ALIGNED_P() IP6_HDR_ALIGNED_P() Move bridge_ip_checkbasic()/bridge_ip6_checkbasic() up so that the alignment is checked for ipfw and dummynet too. PR: ia64/81284 Obtained from: NetBSD Approved by: re (dwhite), mlaier (mentor)
2005-07-02 23:13:31 +00:00
/* XXXJRT new stat, please */
Use IP6STAT_INC/IP6STAT_DEC macros to update ip6 stats. MFC after: 1 week
2013-04-09 07:11:22 +00:00
IP6STAT_INC(ip6s_toosmall);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
in6_ifstat_inc(inifp, ifs6_in_hdrerr);
goto bad;
}
Check the alignment of the IP header before passing the packet up to the packet filter. This would cause a panic on architectures that require strict alignment such as sparc64 (tier1) and ia64/ppc (tier2). This adds two new macros that check the alignment, these are compile time dependent on __NO_STRICT_ALIGNMENT which is set for i386 and amd64 where alignment isn't need so the cost is avoided. IP_HDR_ALIGNED_P() IP6_HDR_ALIGNED_P() Move bridge_ip_checkbasic()/bridge_ip6_checkbasic() up so that the alignment is checked for ipfw and dummynet too. PR: ia64/81284 Obtained from: NetBSD Approved by: re (dwhite), mlaier (mentor)
2005-07-02 23:13:31 +00:00
} else if (__predict_false(m->m_len < sizeof(struct ip6_hdr))) {
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
struct ifnet *inifp = m->m_pkthdr.rcvif;
if ((m = m_pullup(m, sizeof(struct ip6_hdr))) == NULL) {
Use IP6STAT_INC/IP6STAT_DEC macros to update ip6 stats. MFC after: 1 week
2013-04-09 07:11:22 +00:00
IP6STAT_INC(ip6s_toosmall);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
in6_ifstat_inc(inifp, ifs6_in_hdrerr);
goto bad;
}
}
ip6 = mtod(m, struct ip6_hdr *);
if ((ip6->ip6_vfc & IPV6_VERSION_MASK) != IPV6_VERSION) {
Use IP6STAT_INC/IP6STAT_DEC macros to update ip6 stats. MFC after: 1 week
2013-04-09 07:11:22 +00:00
IP6STAT_INC(ip6s_badvers);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_hdrerr);
goto bad;
}
/* Checks out, proceed */
*mp = m;
Fix style(9) nits, whitespace and parentheses.
2006-05-16 22:50:41 +00:00
return (0);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
Minor whitespace cleanup.
2005-12-17 10:03:48 +00:00
bad:
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
*mp = m;
Fix style(9) nits, whitespace and parentheses.
2006-05-16 22:50:41 +00:00
return (-1);
Add if_bridge, which provides more advanced Ethernet bridging and 802.1d spanning tree support. Based on Jason Wright's bridge driver from OpenBSD, and modified by Jason R. Thorpe in NetBSD. Reviewed by: mlaier, bms, green Silence from: -net Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 02:59:26 +00:00
}
Fix style(9) nits, whitespace and parentheses.
2006-05-16 22:50:41 +00:00
#endif /* INET6 */
Add support for fragmenting ipv4 packets. The packet filter may reassemble the ip fragments and return a packet that is larger than the MTU of the sending interface. There is no check for DF or icmp replies as we can only get a large packet to fragment by reassembling a previous fragment, and this only happens after a call to pfil(9). Obtained from: OpenBSD (mostly) Glanced at by: mlaier MFC after: 1 month
2006-04-29 05:37:25 +00:00
/*
* bridge_fragment:
*
bridge: Fix fragment handling and memory leak Fragmented UDP and ICMP packets were corrupted if a firewall with reassembling feature (like pf'scrub) is enabled on the bridge. This patch fixes corrupted packet problem and the panic (triggered easly with low RAM) as explain in PR 185633. bridge_pfil and bridge_fragment relationship: bridge_pfil() receive (IN direction) packets and sent it to the firewall The firewall can be configured for reassembling fragmented packet (like pf'scrubing) in one mbuf chain when bridge_pfil() need to send this reassembled packet to the outgoing interface, it needs to re-fragment it by using bridge_fragment() bridge_fragment() had to split this mbuf (using ip_fragment) first then had to M_PREPEND each packet in the mbuf chain for adding Ethernet header. But M_PREPEND can sometime create a new mbuf on the begining of the mbuf chain, then the "main" pointer of this mbuf chain should be updated and this case is tottaly forgotten. The original bridge_fragment code (Revision 158140, 2006 April 29) came from OpenBSD, and the call to bridge_enqueue was embedded. But on FreeBSD, bridge_enqueue() is done after bridge_fragment(), then the original OpenBSD code can't work as-it of FreeBSD. PR: 185633 Submitted by: Olivier Cochard-Labbé Differential Revision: https://reviews.freebsd.org/D7780
2016-09-24 07:09:43 +00:00
* Fragment mbuf chain in multiple packets and prepend ethernet header.
Add support for fragmenting ipv4 packets. The packet filter may reassemble the ip fragments and return a packet that is larger than the MTU of the sending interface. There is no check for DF or icmp replies as we can only get a large packet to fragment by reassembling a previous fragment, and this only happens after a call to pfil(9). Obtained from: OpenBSD (mostly) Glanced at by: mlaier MFC after: 1 month
2006-04-29 05:37:25 +00:00
*/
static int
bridge: Fix fragment handling and memory leak Fragmented UDP and ICMP packets were corrupted if a firewall with reassembling feature (like pf'scrub) is enabled on the bridge. This patch fixes corrupted packet problem and the panic (triggered easly with low RAM) as explain in PR 185633. bridge_pfil and bridge_fragment relationship: bridge_pfil() receive (IN direction) packets and sent it to the firewall The firewall can be configured for reassembling fragmented packet (like pf'scrubing) in one mbuf chain when bridge_pfil() need to send this reassembled packet to the outgoing interface, it needs to re-fragment it by using bridge_fragment() bridge_fragment() had to split this mbuf (using ip_fragment) first then had to M_PREPEND each packet in the mbuf chain for adding Ethernet header. But M_PREPEND can sometime create a new mbuf on the begining of the mbuf chain, then the "main" pointer of this mbuf chain should be updated and this case is tottaly forgotten. The original bridge_fragment code (Revision 158140, 2006 April 29) came from OpenBSD, and the call to bridge_enqueue was embedded. But on FreeBSD, bridge_enqueue() is done after bridge_fragment(), then the original OpenBSD code can't work as-it of FreeBSD. PR: 185633 Submitted by: Olivier Cochard-Labbé Differential Revision: https://reviews.freebsd.org/D7780
2016-09-24 07:09:43 +00:00
bridge_fragment(struct ifnet *ifp, struct mbuf **mp, struct ether_header *eh,
Add support for fragmenting ipv4 packets. The packet filter may reassemble the ip fragments and return a packet that is larger than the MTU of the sending interface. There is no check for DF or icmp replies as we can only get a large packet to fragment by reassembling a previous fragment, and this only happens after a call to pfil(9). Obtained from: OpenBSD (mostly) Glanced at by: mlaier MFC after: 1 month
2006-04-29 05:37:25 +00:00
int snap, struct llc *llc)
{
bridge: Fix fragment handling and memory leak Fragmented UDP and ICMP packets were corrupted if a firewall with reassembling feature (like pf'scrub) is enabled on the bridge. This patch fixes corrupted packet problem and the panic (triggered easly with low RAM) as explain in PR 185633. bridge_pfil and bridge_fragment relationship: bridge_pfil() receive (IN direction) packets and sent it to the firewall The firewall can be configured for reassembling fragmented packet (like pf'scrubing) in one mbuf chain when bridge_pfil() need to send this reassembled packet to the outgoing interface, it needs to re-fragment it by using bridge_fragment() bridge_fragment() had to split this mbuf (using ip_fragment) first then had to M_PREPEND each packet in the mbuf chain for adding Ethernet header. But M_PREPEND can sometime create a new mbuf on the begining of the mbuf chain, then the "main" pointer of this mbuf chain should be updated and this case is tottaly forgotten. The original bridge_fragment code (Revision 158140, 2006 April 29) came from OpenBSD, and the call to bridge_enqueue was embedded. But on FreeBSD, bridge_enqueue() is done after bridge_fragment(), then the original OpenBSD code can't work as-it of FreeBSD. PR: 185633 Submitted by: Olivier Cochard-Labbé Differential Revision: https://reviews.freebsd.org/D7780
2016-09-24 07:09:43 +00:00
struct mbuf *m = *mp, *nextpkt = NULL, *mprev = NULL, *mcur = NULL;
Add support for fragmenting ipv4 packets. The packet filter may reassemble the ip fragments and return a packet that is larger than the MTU of the sending interface. There is no check for DF or icmp replies as we can only get a large packet to fragment by reassembling a previous fragment, and this only happens after a call to pfil(9). Obtained from: OpenBSD (mostly) Glanced at by: mlaier MFC after: 1 month
2006-04-29 05:37:25 +00:00
struct ip *ip;
int error = -1;
if (m->m_len < sizeof(struct ip) &&
(m = m_pullup(m, sizeof(struct ip))) == NULL)
bridge: Fix fragment handling and memory leak Fragmented UDP and ICMP packets were corrupted if a firewall with reassembling feature (like pf'scrub) is enabled on the bridge. This patch fixes corrupted packet problem and the panic (triggered easly with low RAM) as explain in PR 185633. bridge_pfil and bridge_fragment relationship: bridge_pfil() receive (IN direction) packets and sent it to the firewall The firewall can be configured for reassembling fragmented packet (like pf'scrubing) in one mbuf chain when bridge_pfil() need to send this reassembled packet to the outgoing interface, it needs to re-fragment it by using bridge_fragment() bridge_fragment() had to split this mbuf (using ip_fragment) first then had to M_PREPEND each packet in the mbuf chain for adding Ethernet header. But M_PREPEND can sometime create a new mbuf on the begining of the mbuf chain, then the "main" pointer of this mbuf chain should be updated and this case is tottaly forgotten. The original bridge_fragment code (Revision 158140, 2006 April 29) came from OpenBSD, and the call to bridge_enqueue was embedded. But on FreeBSD, bridge_enqueue() is done after bridge_fragment(), then the original OpenBSD code can't work as-it of FreeBSD. PR: 185633 Submitted by: Olivier Cochard-Labbé Differential Revision: https://reviews.freebsd.org/D7780
2016-09-24 07:09:43 +00:00
goto dropit;
Add support for fragmenting ipv4 packets. The packet filter may reassemble the ip fragments and return a packet that is larger than the MTU of the sending interface. There is no check for DF or icmp replies as we can only get a large packet to fragment by reassembling a previous fragment, and this only happens after a call to pfil(9). Obtained from: OpenBSD (mostly) Glanced at by: mlaier MFC after: 1 month
2006-04-29 05:37:25 +00:00
ip = mtod(m, struct ip *);
o Remove last argument to ip_fragment(), and obtain all needed information on checksums directly from mbuf flags. This simplifies code. o Clear CSUM_IP from the mbuf in ip_fragment() if we did checksums in hardware. Some driver may not announce CSUM_IP in theur if_hwassist, although try to do checksums if CSUM_IP set on mbuf. Example is em(4). o While here, consistently use CSUM_IP instead of its alias CSUM_DELAY_IP. After this change CSUM_DELAY_IP vanishes from the stack. Submitted by: Sebastian Kuzminsky <seb lineratesystems.com>
2012-10-26 21:06:33 +00:00
m->m_pkthdr.csum_flags |= CSUM_IP;
error = ip_fragment(ip, &m, ifp->if_mtu, ifp->if_hwassist);
Add support for fragmenting ipv4 packets. The packet filter may reassemble the ip fragments and return a packet that is larger than the MTU of the sending interface. There is no check for DF or icmp replies as we can only get a large packet to fragment by reassembling a previous fragment, and this only happens after a call to pfil(9). Obtained from: OpenBSD (mostly) Glanced at by: mlaier MFC after: 1 month
2006-04-29 05:37:25 +00:00
if (error)
bridge: Fix fragment handling and memory leak Fragmented UDP and ICMP packets were corrupted if a firewall with reassembling feature (like pf'scrub) is enabled on the bridge. This patch fixes corrupted packet problem and the panic (triggered easly with low RAM) as explain in PR 185633. bridge_pfil and bridge_fragment relationship: bridge_pfil() receive (IN direction) packets and sent it to the firewall The firewall can be configured for reassembling fragmented packet (like pf'scrubing) in one mbuf chain when bridge_pfil() need to send this reassembled packet to the outgoing interface, it needs to re-fragment it by using bridge_fragment() bridge_fragment() had to split this mbuf (using ip_fragment) first then had to M_PREPEND each packet in the mbuf chain for adding Ethernet header. But M_PREPEND can sometime create a new mbuf on the begining of the mbuf chain, then the "main" pointer of this mbuf chain should be updated and this case is tottaly forgotten. The original bridge_fragment code (Revision 158140, 2006 April 29) came from OpenBSD, and the call to bridge_enqueue was embedded. But on FreeBSD, bridge_enqueue() is done after bridge_fragment(), then the original OpenBSD code can't work as-it of FreeBSD. PR: 185633 Submitted by: Olivier Cochard-Labbé Differential Revision: https://reviews.freebsd.org/D7780
2016-09-24 07:09:43 +00:00
goto dropit;
Add support for fragmenting ipv4 packets. The packet filter may reassemble the ip fragments and return a packet that is larger than the MTU of the sending interface. There is no check for DF or icmp replies as we can only get a large packet to fragment by reassembling a previous fragment, and this only happens after a call to pfil(9). Obtained from: OpenBSD (mostly) Glanced at by: mlaier MFC after: 1 month
2006-04-29 05:37:25 +00:00
bridge: Fix fragment handling and memory leak Fragmented UDP and ICMP packets were corrupted if a firewall with reassembling feature (like pf'scrub) is enabled on the bridge. This patch fixes corrupted packet problem and the panic (triggered easly with low RAM) as explain in PR 185633. bridge_pfil and bridge_fragment relationship: bridge_pfil() receive (IN direction) packets and sent it to the firewall The firewall can be configured for reassembling fragmented packet (like pf'scrubing) in one mbuf chain when bridge_pfil() need to send this reassembled packet to the outgoing interface, it needs to re-fragment it by using bridge_fragment() bridge_fragment() had to split this mbuf (using ip_fragment) first then had to M_PREPEND each packet in the mbuf chain for adding Ethernet header. But M_PREPEND can sometime create a new mbuf on the begining of the mbuf chain, then the "main" pointer of this mbuf chain should be updated and this case is tottaly forgotten. The original bridge_fragment code (Revision 158140, 2006 April 29) came from OpenBSD, and the call to bridge_enqueue was embedded. But on FreeBSD, bridge_enqueue() is done after bridge_fragment(), then the original OpenBSD code can't work as-it of FreeBSD. PR: 185633 Submitted by: Olivier Cochard-Labbé Differential Revision: https://reviews.freebsd.org/D7780
2016-09-24 07:09:43 +00:00
/*
* Walk the chain and re-add the Ethernet header for
* each mbuf packet.
*/
for (mcur = m; mcur; mcur = mcur->m_nextpkt) {
nextpkt = mcur->m_nextpkt;
mcur->m_nextpkt = NULL;
if (snap) {
M_PREPEND(mcur, sizeof(struct llc), M_NOWAIT);
if (mcur == NULL) {
Add support for fragmenting ipv4 packets. The packet filter may reassemble the ip fragments and return a packet that is larger than the MTU of the sending interface. There is no check for DF or icmp replies as we can only get a large packet to fragment by reassembling a previous fragment, and this only happens after a call to pfil(9). Obtained from: OpenBSD (mostly) Glanced at by: mlaier MFC after: 1 month
2006-04-29 05:37:25 +00:00
error = ENOBUFS;
bridge: Fix fragment handling and memory leak Fragmented UDP and ICMP packets were corrupted if a firewall with reassembling feature (like pf'scrub) is enabled on the bridge. This patch fixes corrupted packet problem and the panic (triggered easly with low RAM) as explain in PR 185633. bridge_pfil and bridge_fragment relationship: bridge_pfil() receive (IN direction) packets and sent it to the firewall The firewall can be configured for reassembling fragmented packet (like pf'scrubing) in one mbuf chain when bridge_pfil() need to send this reassembled packet to the outgoing interface, it needs to re-fragment it by using bridge_fragment() bridge_fragment() had to split this mbuf (using ip_fragment) first then had to M_PREPEND each packet in the mbuf chain for adding Ethernet header. But M_PREPEND can sometime create a new mbuf on the begining of the mbuf chain, then the "main" pointer of this mbuf chain should be updated and this case is tottaly forgotten. The original bridge_fragment code (Revision 158140, 2006 April 29) came from OpenBSD, and the call to bridge_enqueue was embedded. But on FreeBSD, bridge_enqueue() is done after bridge_fragment(), then the original OpenBSD code can't work as-it of FreeBSD. PR: 185633 Submitted by: Olivier Cochard-Labbé Differential Revision: https://reviews.freebsd.org/D7780
2016-09-24 07:09:43 +00:00
if (mprev != NULL)
mprev->m_nextpkt = nextpkt;
goto dropit;
Add support for fragmenting ipv4 packets. The packet filter may reassemble the ip fragments and return a packet that is larger than the MTU of the sending interface. There is no check for DF or icmp replies as we can only get a large packet to fragment by reassembling a previous fragment, and this only happens after a call to pfil(9). Obtained from: OpenBSD (mostly) Glanced at by: mlaier MFC after: 1 month
2006-04-29 05:37:25 +00:00
}
bridge: Fix fragment handling and memory leak Fragmented UDP and ICMP packets were corrupted if a firewall with reassembling feature (like pf'scrub) is enabled on the bridge. This patch fixes corrupted packet problem and the panic (triggered easly with low RAM) as explain in PR 185633. bridge_pfil and bridge_fragment relationship: bridge_pfil() receive (IN direction) packets and sent it to the firewall The firewall can be configured for reassembling fragmented packet (like pf'scrubing) in one mbuf chain when bridge_pfil() need to send this reassembled packet to the outgoing interface, it needs to re-fragment it by using bridge_fragment() bridge_fragment() had to split this mbuf (using ip_fragment) first then had to M_PREPEND each packet in the mbuf chain for adding Ethernet header. But M_PREPEND can sometime create a new mbuf on the begining of the mbuf chain, then the "main" pointer of this mbuf chain should be updated and this case is tottaly forgotten. The original bridge_fragment code (Revision 158140, 2006 April 29) came from OpenBSD, and the call to bridge_enqueue was embedded. But on FreeBSD, bridge_enqueue() is done after bridge_fragment(), then the original OpenBSD code can't work as-it of FreeBSD. PR: 185633 Submitted by: Olivier Cochard-Labbé Differential Revision: https://reviews.freebsd.org/D7780
2016-09-24 07:09:43 +00:00
bcopy(llc, mtod(mcur, caddr_t),sizeof(struct llc));
}
M_PREPEND(mcur, ETHER_HDR_LEN, M_NOWAIT);
if (mcur == NULL) {
error = ENOBUFS;
if (mprev != NULL)
mprev->m_nextpkt = nextpkt;
goto dropit;
}
bcopy(eh, mtod(mcur, caddr_t), ETHER_HDR_LEN);
Add support for fragmenting ipv4 packets. The packet filter may reassemble the ip fragments and return a packet that is larger than the MTU of the sending interface. There is no check for DF or icmp replies as we can only get a large packet to fragment by reassembling a previous fragment, and this only happens after a call to pfil(9). Obtained from: OpenBSD (mostly) Glanced at by: mlaier MFC after: 1 month
2006-04-29 05:37:25 +00:00
bridge: Fix fragment handling and memory leak Fragmented UDP and ICMP packets were corrupted if a firewall with reassembling feature (like pf'scrub) is enabled on the bridge. This patch fixes corrupted packet problem and the panic (triggered easly with low RAM) as explain in PR 185633. bridge_pfil and bridge_fragment relationship: bridge_pfil() receive (IN direction) packets and sent it to the firewall The firewall can be configured for reassembling fragmented packet (like pf'scrubing) in one mbuf chain when bridge_pfil() need to send this reassembled packet to the outgoing interface, it needs to re-fragment it by using bridge_fragment() bridge_fragment() had to split this mbuf (using ip_fragment) first then had to M_PREPEND each packet in the mbuf chain for adding Ethernet header. But M_PREPEND can sometime create a new mbuf on the begining of the mbuf chain, then the "main" pointer of this mbuf chain should be updated and this case is tottaly forgotten. The original bridge_fragment code (Revision 158140, 2006 April 29) came from OpenBSD, and the call to bridge_enqueue was embedded. But on FreeBSD, bridge_enqueue() is done after bridge_fragment(), then the original OpenBSD code can't work as-it of FreeBSD. PR: 185633 Submitted by: Olivier Cochard-Labbé Differential Revision: https://reviews.freebsd.org/D7780
2016-09-24 07:09:43 +00:00
/*
* The previous two M_PREPEND could have inserted one or two
* mbufs in front so we have to update the previous packet's
* m_nextpkt.
*/
mcur->m_nextpkt = nextpkt;
if (mprev != NULL)
mprev->m_nextpkt = mcur;
else {
/* The first mbuf in the original chain needs to be
* updated. */
*mp = mcur;
}
mprev = mcur;
}
Add support for fragmenting ipv4 packets. The packet filter may reassemble the ip fragments and return a packet that is larger than the MTU of the sending interface. There is no check for DF or icmp replies as we can only get a large packet to fragment by reassembling a previous fragment, and this only happens after a call to pfil(9). Obtained from: OpenBSD (mostly) Glanced at by: mlaier MFC after: 1 month
2006-04-29 05:37:25 +00:00
bridge: Fix fragment handling and memory leak Fragmented UDP and ICMP packets were corrupted if a firewall with reassembling feature (like pf'scrub) is enabled on the bridge. This patch fixes corrupted packet problem and the panic (triggered easly with low RAM) as explain in PR 185633. bridge_pfil and bridge_fragment relationship: bridge_pfil() receive (IN direction) packets and sent it to the firewall The firewall can be configured for reassembling fragmented packet (like pf'scrubing) in one mbuf chain when bridge_pfil() need to send this reassembled packet to the outgoing interface, it needs to re-fragment it by using bridge_fragment() bridge_fragment() had to split this mbuf (using ip_fragment) first then had to M_PREPEND each packet in the mbuf chain for adding Ethernet header. But M_PREPEND can sometime create a new mbuf on the begining of the mbuf chain, then the "main" pointer of this mbuf chain should be updated and this case is tottaly forgotten. The original bridge_fragment code (Revision 158140, 2006 April 29) came from OpenBSD, and the call to bridge_enqueue was embedded. But on FreeBSD, bridge_enqueue() is done after bridge_fragment(), then the original OpenBSD code can't work as-it of FreeBSD. PR: 185633 Submitted by: Olivier Cochard-Labbé Differential Revision: https://reviews.freebsd.org/D7780
2016-09-24 07:09:43 +00:00
KMOD_IPSTAT_INC(ips_fragmented);
Add support for fragmenting ipv4 packets. The packet filter may reassemble the ip fragments and return a packet that is larger than the MTU of the sending interface. There is no check for DF or icmp replies as we can only get a large packet to fragment by reassembling a previous fragment, and this only happens after a call to pfil(9). Obtained from: OpenBSD (mostly) Glanced at by: mlaier MFC after: 1 month
2006-04-29 05:37:25 +00:00
return (error);
bridge: Fix fragment handling and memory leak Fragmented UDP and ICMP packets were corrupted if a firewall with reassembling feature (like pf'scrub) is enabled on the bridge. This patch fixes corrupted packet problem and the panic (triggered easly with low RAM) as explain in PR 185633. bridge_pfil and bridge_fragment relationship: bridge_pfil() receive (IN direction) packets and sent it to the firewall The firewall can be configured for reassembling fragmented packet (like pf'scrubing) in one mbuf chain when bridge_pfil() need to send this reassembled packet to the outgoing interface, it needs to re-fragment it by using bridge_fragment() bridge_fragment() had to split this mbuf (using ip_fragment) first then had to M_PREPEND each packet in the mbuf chain for adding Ethernet header. But M_PREPEND can sometime create a new mbuf on the begining of the mbuf chain, then the "main" pointer of this mbuf chain should be updated and this case is tottaly forgotten. The original bridge_fragment code (Revision 158140, 2006 April 29) came from OpenBSD, and the call to bridge_enqueue was embedded. But on FreeBSD, bridge_enqueue() is done after bridge_fragment(), then the original OpenBSD code can't work as-it of FreeBSD. PR: 185633 Submitted by: Olivier Cochard-Labbé Differential Revision: https://reviews.freebsd.org/D7780
2016-09-24 07:09:43 +00:00
dropit:
for (mcur = *mp; mcur; mcur = m) { /* droping the full packet chain */
m = mcur->m_nextpkt;
m_freem(mcur);
}
Add support for fragmenting ipv4 packets. The packet filter may reassemble the ip fragments and return a packet that is larger than the MTU of the sending interface. There is no check for DF or icmp replies as we can only get a large packet to fragment by reassembling a previous fragment, and this only happens after a call to pfil(9). Obtained from: OpenBSD (mostly) Glanced at by: mlaier MFC after: 1 month
2006-04-29 05:37:25 +00:00
return (error);
}
Add linkstate to bridge(4), set the link to up when at least one underlying interface is up, otherwise the link is down. This, among other things, allows carp to work on a bridge. Prodded by: glebius Tested by: Alexander Lunev
2012-04-20 09:55:50 +00:00
static void
bridge_linkstate(struct ifnet *ifp)
{
struct bridge_softc *sc = ifp->if_bridge;
Fix a panic I introduced in r234487, the bridge softc pointer is set to null early in the detach so rearrange things not to explode. Reported by: David Roffiaen, Gustau Perez Querol Tested by: David Roffiaen MFC after: 3 days
2012-06-11 20:12:13 +00:00
struct bridge_iflist *bif;
bridge: epoch-ification Run the bridge datapath under epoch, rather than under the BRIDGE_LOCK(). We still take the BRIDGE_LOCK() whenever we insert or delete items in the relevant lists, but we use epoch callbacks to free items so that it's safe to iterate the lists without the BRIDGE_LOCK. Tests on mercat5/6 shows this increases bridge throughput significantly, from 3.7Mpps to 18.6Mpps. Reviewed by: emaste, philip, melifaro MFC after: 2 months Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24250
2020-04-26 16:22:35 +00:00
struct epoch_tracker et;
NET_EPOCH_ENTER(et);
Add linkstate to bridge(4), set the link to up when at least one underlying interface is up, otherwise the link is down. This, among other things, allows carp to work on a bridge. Prodded by: glebius Tested by: Alexander Lunev
2012-04-20 09:55:50 +00:00
bif = bridge_lookup_member_if(sc, ifp);
if (bif == NULL) {
bridge: epoch-ification Run the bridge datapath under epoch, rather than under the BRIDGE_LOCK(). We still take the BRIDGE_LOCK() whenever we insert or delete items in the relevant lists, but we use epoch callbacks to free items so that it's safe to iterate the lists without the BRIDGE_LOCK. Tests on mercat5/6 shows this increases bridge throughput significantly, from 3.7Mpps to 18.6Mpps. Reviewed by: emaste, philip, melifaro MFC after: 2 months Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24250
2020-04-26 16:22:35 +00:00
NET_EPOCH_EXIT(et);
Add linkstate to bridge(4), set the link to up when at least one underlying interface is up, otherwise the link is down. This, among other things, allows carp to work on a bridge. Prodded by: glebius Tested by: Alexander Lunev
2012-04-20 09:55:50 +00:00
return;
}
Fix a panic I introduced in r234487, the bridge softc pointer is set to null early in the detach so rearrange things not to explode. Reported by: David Roffiaen, Gustau Perez Querol Tested by: David Roffiaen MFC after: 3 days
2012-06-11 20:12:13 +00:00
bridge_linkcheck(sc);
bstp_linkstate(&bif->bif_stp);
bridge: epoch-ification Run the bridge datapath under epoch, rather than under the BRIDGE_LOCK(). We still take the BRIDGE_LOCK() whenever we insert or delete items in the relevant lists, but we use epoch callbacks to free items so that it's safe to iterate the lists without the BRIDGE_LOCK. Tests on mercat5/6 shows this increases bridge throughput significantly, from 3.7Mpps to 18.6Mpps. Reviewed by: emaste, philip, melifaro MFC after: 2 months Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24250
2020-04-26 16:22:35 +00:00
NET_EPOCH_EXIT(et);
Fix a panic I introduced in r234487, the bridge softc pointer is set to null early in the detach so rearrange things not to explode. Reported by: David Roffiaen, Gustau Perez Querol Tested by: David Roffiaen MFC after: 3 days
2012-06-11 20:12:13 +00:00
}
static void
bridge_linkcheck(struct bridge_softc *sc)
{
struct bridge_iflist *bif;
int new_link, hasls;
bridge: call member interface ioctl() without NET_EPOCH We're not allowed to hold NET_EPOCH while sleeping, so when we call ioctl() handlers for member interfaces we cannot be in NET_EPOCH. We still need some protection of our CK_LISTs, so hold BRIDGE_LOCK instead. That requires changing BRIDGE_LOCK into a sleepable lock, and separating the BRIDGE_RT_LOCK, to protect bridge_rtnode lists. That lock is taken in the data path (while in NET_EPOCH), so it cannot be a sleepable lock. While here document the locking strategy. MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D26418
2020-10-06 19:19:56 +00:00
BRIDGE_LOCK_OR_NET_EPOCH_ASSERT(sc);
bridge: epoch-ification Run the bridge datapath under epoch, rather than under the BRIDGE_LOCK(). We still take the BRIDGE_LOCK() whenever we insert or delete items in the relevant lists, but we use epoch callbacks to free items so that it's safe to iterate the lists without the BRIDGE_LOCK. Tests on mercat5/6 shows this increases bridge throughput significantly, from 3.7Mpps to 18.6Mpps. Reviewed by: emaste, philip, melifaro MFC after: 2 months Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24250
2020-04-26 16:22:35 +00:00
Add linkstate to bridge(4), set the link to up when at least one underlying interface is up, otherwise the link is down. This, among other things, allows carp to work on a bridge. Prodded by: glebius Tested by: Alexander Lunev
2012-04-20 09:55:50 +00:00
new_link = LINK_STATE_DOWN;
hasls = 0;
/* Our link is considered up if at least one of our ports is active */
bridge: Change lists to CK_LIST as a peparation for epochification Prepare the ground for a rework of the bridge locking approach. We will use an epoch-based approach in the datapath and making it safe to iterate over the interface, span and rtnode lists without holding the BRIDGE_LOCK. Replace the relevant lists by their ConcurrencyKit equivalents. No functional change in this commit. Reviewed by: emaste, ae, philip (previous version) Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24249
2020-04-05 17:15:20 +00:00
CK_LIST_FOREACH(bif, &sc->sc_iflist, bif_next) {
Fix a panic I introduced in r234487, the bridge softc pointer is set to null early in the detach so rearrange things not to explode. Reported by: David Roffiaen, Gustau Perez Querol Tested by: David Roffiaen MFC after: 3 days
2012-06-11 20:12:13 +00:00
if (bif->bif_ifp->if_capabilities & IFCAP_LINKSTATE)
Add linkstate to bridge(4), set the link to up when at least one underlying interface is up, otherwise the link is down. This, among other things, allows carp to work on a bridge. Prodded by: glebius Tested by: Alexander Lunev
2012-04-20 09:55:50 +00:00
hasls++;
Fix a panic I introduced in r234487, the bridge softc pointer is set to null early in the detach so rearrange things not to explode. Reported by: David Roffiaen, Gustau Perez Querol Tested by: David Roffiaen MFC after: 3 days
2012-06-11 20:12:13 +00:00
if (bif->bif_ifp->if_link_state == LINK_STATE_UP) {
Add linkstate to bridge(4), set the link to up when at least one underlying interface is up, otherwise the link is down. This, among other things, allows carp to work on a bridge. Prodded by: glebius Tested by: Alexander Lunev
2012-04-20 09:55:50 +00:00
new_link = LINK_STATE_UP;
break;
}
}
bridge: Change lists to CK_LIST as a peparation for epochification Prepare the ground for a rework of the bridge locking approach. We will use an epoch-based approach in the datapath and making it safe to iterate over the interface, span and rtnode lists without holding the BRIDGE_LOCK. Replace the relevant lists by their ConcurrencyKit equivalents. No functional change in this commit. Reviewed by: emaste, ae, philip (previous version) Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D24249
2020-04-05 17:15:20 +00:00
if (!CK_LIST_EMPTY(&sc->sc_iflist) && !hasls) {
Add linkstate to bridge(4), set the link to up when at least one underlying interface is up, otherwise the link is down. This, among other things, allows carp to work on a bridge. Prodded by: glebius Tested by: Alexander Lunev
2012-04-20 09:55:50 +00:00
/* If no interfaces support link-state then we default to up */
new_link = LINK_STATE_UP;
}
if_link_state_change(sc->sc_ifp, new_link);
}
Reference in New Issue Copy Permalink