freebsd-skq/sys/netinet/ip_encap.c

416 lines
11 KiB
C
Raw Normal View History

/* $KAME: ip_encap.c,v 1.41 2001/03/15 08:35:08 itojun Exp $ */
/*-
* SPDX-License-Identifier: BSD-3-Clause
*
* Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project.
Rework IP encapsulation handling code. Currently it has several disadvantages: - it uses single mutex to protect internal structures. It is used by data- and control- path, thus there are no parallelism at all. - it uses single list to keep encap handlers for both INET and INET6 families. - struct encaptab keeps unneeded information (src, dst, masks, protosw), that isn't used by code in the source tree. - matches are prioritized and when many tunneling interfaces are registered, encapcheck handler of each interface is invoked for each packet. The search takes O(n) for n interfaces. All this work is done with exclusive lock held. What this patch includes: - the datapath is converted to be lockless using epoch(9) KPI. - struct encaptab now linked using CK_LIST. - all unused fields removed from struct encaptab. Several new fields addedr: min_length is the minimum packet length, that encapsulation handler expects to see; exact_match is maximum number of bits, that can return an encapsulation handler, when it wants to consume a packet. - IPv6 and IPv4 handlers are stored in separate lists; - added new "encap_lookup_t" method, that will be used later. It is targeted to speedup lookup of needed interface, when gif(4)/gre(4) have many interfaces. - the need to use protosw structure is eliminated. The only pr_input method was used from this structure, so I don't see the need to keep using it. - encap_input_t method changed to avoid using mbuf tags to store softc pointer. Now it is passed directly trough encap_input_t method. encap_getarg() funtions is removed. - all sockaddr structures and code that uses them removed. We don't have any code in the tree that uses them. All consumers use encap_attach_func() method, that relies on invoking of encapcheck() to determine the needed handler. - introduced struct encap_config, it contains parameters of encap handler that is going to be registered by encap_attach() function. - encap handlers are stored in lists ordered by exact_match value, thus handlers that need more bits to match will be checked first, and if encapcheck method returns exact_match value, the search will be stopped. - all current consumers changed to use new KPI. Reviewed by: mmacy Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D15617
2018-06-05 20:51:01 +00:00
* Copyright (c) 2018 Andrey V. Elsukov <ae@FreeBSD.org>
* 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.
* 3. Neither the name of the project nor the names of its contributors
* may be used to endorse or promote products derived from this software
* without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*/
/*
* My grandfather said that there's a devil inside tunnelling technology...
*
* We have surprisingly many protocols that want packets with IP protocol
* #4 or #41. Here's a list of protocols that want protocol #41:
* RFC1933 configured tunnel
* RFC1933 automatic tunnel
* RFC2401 IPsec tunnel
* RFC2473 IPv6 generic packet tunnelling
* RFC2529 6over4 tunnel
* mobile-ip6 (uses RFC2473)
* RFC3056 6to4 tunnel
* isatap tunnel
* Here's a list of protocol that want protocol #4:
* RFC1853 IPv4-in-IPv4 tunnelling
* RFC2003 IPv4 encapsulation within IPv4
* RFC2344 reverse tunnelling for mobile-ip4
* RFC2401 IPsec tunnel
* Well, what can I say. They impose different en/decapsulation mechanism
* from each other, so they need separate protocol handler. The only one
* we can easily determine by protocol # is IPsec, which always has
* AH/ESP/IPComp header right after outer IP header.
*
* So, clearly good old protosw does not work for protocol #4 and #41.
* The code will let you match protocol via src/dst address pair.
*/
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
#include "opt_inet.h"
#include "opt_inet6.h"
#include <sys/param.h>
#include <sys/systm.h>
#include <sys/eventhandler.h>
Rework IP encapsulation handling code. Currently it has several disadvantages: - it uses single mutex to protect internal structures. It is used by data- and control- path, thus there are no parallelism at all. - it uses single list to keep encap handlers for both INET and INET6 families. - struct encaptab keeps unneeded information (src, dst, masks, protosw), that isn't used by code in the source tree. - matches are prioritized and when many tunneling interfaces are registered, encapcheck handler of each interface is invoked for each packet. The search takes O(n) for n interfaces. All this work is done with exclusive lock held. What this patch includes: - the datapath is converted to be lockless using epoch(9) KPI. - struct encaptab now linked using CK_LIST. - all unused fields removed from struct encaptab. Several new fields addedr: min_length is the minimum packet length, that encapsulation handler expects to see; exact_match is maximum number of bits, that can return an encapsulation handler, when it wants to consume a packet. - IPv6 and IPv4 handlers are stored in separate lists; - added new "encap_lookup_t" method, that will be used later. It is targeted to speedup lookup of needed interface, when gif(4)/gre(4) have many interfaces. - the need to use protosw structure is eliminated. The only pr_input method was used from this structure, so I don't see the need to keep using it. - encap_input_t method changed to avoid using mbuf tags to store softc pointer. Now it is passed directly trough encap_input_t method. encap_getarg() funtions is removed. - all sockaddr structures and code that uses them removed. We don't have any code in the tree that uses them. All consumers use encap_attach_func() method, that relies on invoking of encapcheck() to determine the needed handler. - introduced struct encap_config, it contains parameters of encap handler that is going to be registered by encap_attach() function. - encap handlers are stored in lists ordered by exact_match value, thus handlers that need more bits to match will be checked first, and if encapcheck method returns exact_match value, the search will be stopped. - all current consumers changed to use new KPI. Reviewed by: mmacy Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D15617
2018-06-05 20:51:01 +00:00
#include <sys/kernel.h>
#include <sys/lock.h>
Rework IP encapsulation handling code. Currently it has several disadvantages: - it uses single mutex to protect internal structures. It is used by data- and control- path, thus there are no parallelism at all. - it uses single list to keep encap handlers for both INET and INET6 families. - struct encaptab keeps unneeded information (src, dst, masks, protosw), that isn't used by code in the source tree. - matches are prioritized and when many tunneling interfaces are registered, encapcheck handler of each interface is invoked for each packet. The search takes O(n) for n interfaces. All this work is done with exclusive lock held. What this patch includes: - the datapath is converted to be lockless using epoch(9) KPI. - struct encaptab now linked using CK_LIST. - all unused fields removed from struct encaptab. Several new fields addedr: min_length is the minimum packet length, that encapsulation handler expects to see; exact_match is maximum number of bits, that can return an encapsulation handler, when it wants to consume a packet. - IPv6 and IPv4 handlers are stored in separate lists; - added new "encap_lookup_t" method, that will be used later. It is targeted to speedup lookup of needed interface, when gif(4)/gre(4) have many interfaces. - the need to use protosw structure is eliminated. The only pr_input method was used from this structure, so I don't see the need to keep using it. - encap_input_t method changed to avoid using mbuf tags to store softc pointer. Now it is passed directly trough encap_input_t method. encap_getarg() funtions is removed. - all sockaddr structures and code that uses them removed. We don't have any code in the tree that uses them. All consumers use encap_attach_func() method, that relies on invoking of encapcheck() to determine the needed handler. - introduced struct encap_config, it contains parameters of encap handler that is going to be registered by encap_attach() function. - encap handlers are stored in lists ordered by exact_match value, thus handlers that need more bits to match will be checked first, and if encapcheck method returns exact_match value, the search will be stopped. - all current consumers changed to use new KPI. Reviewed by: mmacy Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D15617
2018-06-05 20:51:01 +00:00
#include <sys/malloc.h>
#include <sys/mutex.h>
#include <sys/mbuf.h>
#include <sys/errno.h>
Rework IP encapsulation handling code. Currently it has several disadvantages: - it uses single mutex to protect internal structures. It is used by data- and control- path, thus there are no parallelism at all. - it uses single list to keep encap handlers for both INET and INET6 families. - struct encaptab keeps unneeded information (src, dst, masks, protosw), that isn't used by code in the source tree. - matches are prioritized and when many tunneling interfaces are registered, encapcheck handler of each interface is invoked for each packet. The search takes O(n) for n interfaces. All this work is done with exclusive lock held. What this patch includes: - the datapath is converted to be lockless using epoch(9) KPI. - struct encaptab now linked using CK_LIST. - all unused fields removed from struct encaptab. Several new fields addedr: min_length is the minimum packet length, that encapsulation handler expects to see; exact_match is maximum number of bits, that can return an encapsulation handler, when it wants to consume a packet. - IPv6 and IPv4 handlers are stored in separate lists; - added new "encap_lookup_t" method, that will be used later. It is targeted to speedup lookup of needed interface, when gif(4)/gre(4) have many interfaces. - the need to use protosw structure is eliminated. The only pr_input method was used from this structure, so I don't see the need to keep using it. - encap_input_t method changed to avoid using mbuf tags to store softc pointer. Now it is passed directly trough encap_input_t method. encap_getarg() funtions is removed. - all sockaddr structures and code that uses them removed. We don't have any code in the tree that uses them. All consumers use encap_attach_func() method, that relies on invoking of encapcheck() to determine the needed handler. - introduced struct encap_config, it contains parameters of encap handler that is going to be registered by encap_attach() function. - encap handlers are stored in lists ordered by exact_match value, thus handlers that need more bits to match will be checked first, and if encapcheck method returns exact_match value, the search will be stopped. - all current consumers changed to use new KPI. Reviewed by: mmacy Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D15617
2018-06-05 20:51:01 +00:00
#include <sys/socket.h>
#include <net/if.h>
Rework IP encapsulation handling code. Currently it has several disadvantages: - it uses single mutex to protect internal structures. It is used by data- and control- path, thus there are no parallelism at all. - it uses single list to keep encap handlers for both INET and INET6 families. - struct encaptab keeps unneeded information (src, dst, masks, protosw), that isn't used by code in the source tree. - matches are prioritized and when many tunneling interfaces are registered, encapcheck handler of each interface is invoked for each packet. The search takes O(n) for n interfaces. All this work is done with exclusive lock held. What this patch includes: - the datapath is converted to be lockless using epoch(9) KPI. - struct encaptab now linked using CK_LIST. - all unused fields removed from struct encaptab. Several new fields addedr: min_length is the minimum packet length, that encapsulation handler expects to see; exact_match is maximum number of bits, that can return an encapsulation handler, when it wants to consume a packet. - IPv6 and IPv4 handlers are stored in separate lists; - added new "encap_lookup_t" method, that will be used later. It is targeted to speedup lookup of needed interface, when gif(4)/gre(4) have many interfaces. - the need to use protosw structure is eliminated. The only pr_input method was used from this structure, so I don't see the need to keep using it. - encap_input_t method changed to avoid using mbuf tags to store softc pointer. Now it is passed directly trough encap_input_t method. encap_getarg() funtions is removed. - all sockaddr structures and code that uses them removed. We don't have any code in the tree that uses them. All consumers use encap_attach_func() method, that relies on invoking of encapcheck() to determine the needed handler. - introduced struct encap_config, it contains parameters of encap handler that is going to be registered by encap_attach() function. - encap handlers are stored in lists ordered by exact_match value, thus handlers that need more bits to match will be checked first, and if encapcheck method returns exact_match value, the search will be stopped. - all current consumers changed to use new KPI. Reviewed by: mmacy Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D15617
2018-06-05 20:51:01 +00:00
#include <net/if_var.h>
#include <netinet/in.h>
#include <netinet/ip_var.h>
#include <netinet/ip_encap.h>
#ifdef INET6
#include <netinet6/ip6_var.h>
#endif
Rework IP encapsulation handling code. Currently it has several disadvantages: - it uses single mutex to protect internal structures. It is used by data- and control- path, thus there are no parallelism at all. - it uses single list to keep encap handlers for both INET and INET6 families. - struct encaptab keeps unneeded information (src, dst, masks, protosw), that isn't used by code in the source tree. - matches are prioritized and when many tunneling interfaces are registered, encapcheck handler of each interface is invoked for each packet. The search takes O(n) for n interfaces. All this work is done with exclusive lock held. What this patch includes: - the datapath is converted to be lockless using epoch(9) KPI. - struct encaptab now linked using CK_LIST. - all unused fields removed from struct encaptab. Several new fields addedr: min_length is the minimum packet length, that encapsulation handler expects to see; exact_match is maximum number of bits, that can return an encapsulation handler, when it wants to consume a packet. - IPv6 and IPv4 handlers are stored in separate lists; - added new "encap_lookup_t" method, that will be used later. It is targeted to speedup lookup of needed interface, when gif(4)/gre(4) have many interfaces. - the need to use protosw structure is eliminated. The only pr_input method was used from this structure, so I don't see the need to keep using it. - encap_input_t method changed to avoid using mbuf tags to store softc pointer. Now it is passed directly trough encap_input_t method. encap_getarg() funtions is removed. - all sockaddr structures and code that uses them removed. We don't have any code in the tree that uses them. All consumers use encap_attach_func() method, that relies on invoking of encapcheck() to determine the needed handler. - introduced struct encap_config, it contains parameters of encap handler that is going to be registered by encap_attach() function. - encap handlers are stored in lists ordered by exact_match value, thus handlers that need more bits to match will be checked first, and if encapcheck method returns exact_match value, the search will be stopped. - all current consumers changed to use new KPI. Reviewed by: mmacy Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D15617
2018-06-05 20:51:01 +00:00
static MALLOC_DEFINE(M_NETADDR, "encap_export_host",
"Export host address structure");
Rework IP encapsulation handling code. Currently it has several disadvantages: - it uses single mutex to protect internal structures. It is used by data- and control- path, thus there are no parallelism at all. - it uses single list to keep encap handlers for both INET and INET6 families. - struct encaptab keeps unneeded information (src, dst, masks, protosw), that isn't used by code in the source tree. - matches are prioritized and when many tunneling interfaces are registered, encapcheck handler of each interface is invoked for each packet. The search takes O(n) for n interfaces. All this work is done with exclusive lock held. What this patch includes: - the datapath is converted to be lockless using epoch(9) KPI. - struct encaptab now linked using CK_LIST. - all unused fields removed from struct encaptab. Several new fields addedr: min_length is the minimum packet length, that encapsulation handler expects to see; exact_match is maximum number of bits, that can return an encapsulation handler, when it wants to consume a packet. - IPv6 and IPv4 handlers are stored in separate lists; - added new "encap_lookup_t" method, that will be used later. It is targeted to speedup lookup of needed interface, when gif(4)/gre(4) have many interfaces. - the need to use protosw structure is eliminated. The only pr_input method was used from this structure, so I don't see the need to keep using it. - encap_input_t method changed to avoid using mbuf tags to store softc pointer. Now it is passed directly trough encap_input_t method. encap_getarg() funtions is removed. - all sockaddr structures and code that uses them removed. We don't have any code in the tree that uses them. All consumers use encap_attach_func() method, that relies on invoking of encapcheck() to determine the needed handler. - introduced struct encap_config, it contains parameters of encap handler that is going to be registered by encap_attach() function. - encap handlers are stored in lists ordered by exact_match value, thus handlers that need more bits to match will be checked first, and if encapcheck method returns exact_match value, the search will be stopped. - all current consumers changed to use new KPI. Reviewed by: mmacy Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D15617
2018-06-05 20:51:01 +00:00
struct encaptab {
CK_LIST_ENTRY(encaptab) chain;
int proto;
int min_length;
int exact_match;
void *arg;
Rework IP encapsulation handling code. Currently it has several disadvantages: - it uses single mutex to protect internal structures. It is used by data- and control- path, thus there are no parallelism at all. - it uses single list to keep encap handlers for both INET and INET6 families. - struct encaptab keeps unneeded information (src, dst, masks, protosw), that isn't used by code in the source tree. - matches are prioritized and when many tunneling interfaces are registered, encapcheck handler of each interface is invoked for each packet. The search takes O(n) for n interfaces. All this work is done with exclusive lock held. What this patch includes: - the datapath is converted to be lockless using epoch(9) KPI. - struct encaptab now linked using CK_LIST. - all unused fields removed from struct encaptab. Several new fields addedr: min_length is the minimum packet length, that encapsulation handler expects to see; exact_match is maximum number of bits, that can return an encapsulation handler, when it wants to consume a packet. - IPv6 and IPv4 handlers are stored in separate lists; - added new "encap_lookup_t" method, that will be used later. It is targeted to speedup lookup of needed interface, when gif(4)/gre(4) have many interfaces. - the need to use protosw structure is eliminated. The only pr_input method was used from this structure, so I don't see the need to keep using it. - encap_input_t method changed to avoid using mbuf tags to store softc pointer. Now it is passed directly trough encap_input_t method. encap_getarg() funtions is removed. - all sockaddr structures and code that uses them removed. We don't have any code in the tree that uses them. All consumers use encap_attach_func() method, that relies on invoking of encapcheck() to determine the needed handler. - introduced struct encap_config, it contains parameters of encap handler that is going to be registered by encap_attach() function. - encap handlers are stored in lists ordered by exact_match value, thus handlers that need more bits to match will be checked first, and if encapcheck method returns exact_match value, the search will be stopped. - all current consumers changed to use new KPI. Reviewed by: mmacy Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D15617
2018-06-05 20:51:01 +00:00
encap_lookup_t lookup;
encap_check_t check;
encap_input_t input;
};
struct srcaddrtab {
CK_LIST_ENTRY(srcaddrtab) chain;
encap_srcaddr_t srcaddr;
void *arg;
};
Rework IP encapsulation handling code. Currently it has several disadvantages: - it uses single mutex to protect internal structures. It is used by data- and control- path, thus there are no parallelism at all. - it uses single list to keep encap handlers for both INET and INET6 families. - struct encaptab keeps unneeded information (src, dst, masks, protosw), that isn't used by code in the source tree. - matches are prioritized and when many tunneling interfaces are registered, encapcheck handler of each interface is invoked for each packet. The search takes O(n) for n interfaces. All this work is done with exclusive lock held. What this patch includes: - the datapath is converted to be lockless using epoch(9) KPI. - struct encaptab now linked using CK_LIST. - all unused fields removed from struct encaptab. Several new fields addedr: min_length is the minimum packet length, that encapsulation handler expects to see; exact_match is maximum number of bits, that can return an encapsulation handler, when it wants to consume a packet. - IPv6 and IPv4 handlers are stored in separate lists; - added new "encap_lookup_t" method, that will be used later. It is targeted to speedup lookup of needed interface, when gif(4)/gre(4) have many interfaces. - the need to use protosw structure is eliminated. The only pr_input method was used from this structure, so I don't see the need to keep using it. - encap_input_t method changed to avoid using mbuf tags to store softc pointer. Now it is passed directly trough encap_input_t method. encap_getarg() funtions is removed. - all sockaddr structures and code that uses them removed. We don't have any code in the tree that uses them. All consumers use encap_attach_func() method, that relies on invoking of encapcheck() to determine the needed handler. - introduced struct encap_config, it contains parameters of encap handler that is going to be registered by encap_attach() function. - encap handlers are stored in lists ordered by exact_match value, thus handlers that need more bits to match will be checked first, and if encapcheck method returns exact_match value, the search will be stopped. - all current consumers changed to use new KPI. Reviewed by: mmacy Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D15617
2018-06-05 20:51:01 +00:00
CK_LIST_HEAD(encaptab_head, encaptab);
CK_LIST_HEAD(srcaddrtab_head, srcaddrtab);
Rework IP encapsulation handling code. Currently it has several disadvantages: - it uses single mutex to protect internal structures. It is used by data- and control- path, thus there are no parallelism at all. - it uses single list to keep encap handlers for both INET and INET6 families. - struct encaptab keeps unneeded information (src, dst, masks, protosw), that isn't used by code in the source tree. - matches are prioritized and when many tunneling interfaces are registered, encapcheck handler of each interface is invoked for each packet. The search takes O(n) for n interfaces. All this work is done with exclusive lock held. What this patch includes: - the datapath is converted to be lockless using epoch(9) KPI. - struct encaptab now linked using CK_LIST. - all unused fields removed from struct encaptab. Several new fields addedr: min_length is the minimum packet length, that encapsulation handler expects to see; exact_match is maximum number of bits, that can return an encapsulation handler, when it wants to consume a packet. - IPv6 and IPv4 handlers are stored in separate lists; - added new "encap_lookup_t" method, that will be used later. It is targeted to speedup lookup of needed interface, when gif(4)/gre(4) have many interfaces. - the need to use protosw structure is eliminated. The only pr_input method was used from this structure, so I don't see the need to keep using it. - encap_input_t method changed to avoid using mbuf tags to store softc pointer. Now it is passed directly trough encap_input_t method. encap_getarg() funtions is removed. - all sockaddr structures and code that uses them removed. We don't have any code in the tree that uses them. All consumers use encap_attach_func() method, that relies on invoking of encapcheck() to determine the needed handler. - introduced struct encap_config, it contains parameters of encap handler that is going to be registered by encap_attach() function. - encap handlers are stored in lists ordered by exact_match value, thus handlers that need more bits to match will be checked first, and if encapcheck method returns exact_match value, the search will be stopped. - all current consumers changed to use new KPI. Reviewed by: mmacy Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D15617
2018-06-05 20:51:01 +00:00
#ifdef INET
static struct encaptab_head ipv4_encaptab = CK_LIST_HEAD_INITIALIZER();
static struct srcaddrtab_head ipv4_srcaddrtab = CK_LIST_HEAD_INITIALIZER();
Rework IP encapsulation handling code. Currently it has several disadvantages: - it uses single mutex to protect internal structures. It is used by data- and control- path, thus there are no parallelism at all. - it uses single list to keep encap handlers for both INET and INET6 families. - struct encaptab keeps unneeded information (src, dst, masks, protosw), that isn't used by code in the source tree. - matches are prioritized and when many tunneling interfaces are registered, encapcheck handler of each interface is invoked for each packet. The search takes O(n) for n interfaces. All this work is done with exclusive lock held. What this patch includes: - the datapath is converted to be lockless using epoch(9) KPI. - struct encaptab now linked using CK_LIST. - all unused fields removed from struct encaptab. Several new fields addedr: min_length is the minimum packet length, that encapsulation handler expects to see; exact_match is maximum number of bits, that can return an encapsulation handler, when it wants to consume a packet. - IPv6 and IPv4 handlers are stored in separate lists; - added new "encap_lookup_t" method, that will be used later. It is targeted to speedup lookup of needed interface, when gif(4)/gre(4) have many interfaces. - the need to use protosw structure is eliminated. The only pr_input method was used from this structure, so I don't see the need to keep using it. - encap_input_t method changed to avoid using mbuf tags to store softc pointer. Now it is passed directly trough encap_input_t method. encap_getarg() funtions is removed. - all sockaddr structures and code that uses them removed. We don't have any code in the tree that uses them. All consumers use encap_attach_func() method, that relies on invoking of encapcheck() to determine the needed handler. - introduced struct encap_config, it contains parameters of encap handler that is going to be registered by encap_attach() function. - encap handlers are stored in lists ordered by exact_match value, thus handlers that need more bits to match will be checked first, and if encapcheck method returns exact_match value, the search will be stopped. - all current consumers changed to use new KPI. Reviewed by: mmacy Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D15617
2018-06-05 20:51:01 +00:00
#endif
#ifdef INET6
static struct encaptab_head ipv6_encaptab = CK_LIST_HEAD_INITIALIZER();
static struct srcaddrtab_head ipv6_srcaddrtab = CK_LIST_HEAD_INITIALIZER();
Rework IP encapsulation handling code. Currently it has several disadvantages: - it uses single mutex to protect internal structures. It is used by data- and control- path, thus there are no parallelism at all. - it uses single list to keep encap handlers for both INET and INET6 families. - struct encaptab keeps unneeded information (src, dst, masks, protosw), that isn't used by code in the source tree. - matches are prioritized and when many tunneling interfaces are registered, encapcheck handler of each interface is invoked for each packet. The search takes O(n) for n interfaces. All this work is done with exclusive lock held. What this patch includes: - the datapath is converted to be lockless using epoch(9) KPI. - struct encaptab now linked using CK_LIST. - all unused fields removed from struct encaptab. Several new fields addedr: min_length is the minimum packet length, that encapsulation handler expects to see; exact_match is maximum number of bits, that can return an encapsulation handler, when it wants to consume a packet. - IPv6 and IPv4 handlers are stored in separate lists; - added new "encap_lookup_t" method, that will be used later. It is targeted to speedup lookup of needed interface, when gif(4)/gre(4) have many interfaces. - the need to use protosw structure is eliminated. The only pr_input method was used from this structure, so I don't see the need to keep using it. - encap_input_t method changed to avoid using mbuf tags to store softc pointer. Now it is passed directly trough encap_input_t method. encap_getarg() funtions is removed. - all sockaddr structures and code that uses them removed. We don't have any code in the tree that uses them. All consumers use encap_attach_func() method, that relies on invoking of encapcheck() to determine the needed handler. - introduced struct encap_config, it contains parameters of encap handler that is going to be registered by encap_attach() function. - encap handlers are stored in lists ordered by exact_match value, thus handlers that need more bits to match will be checked first, and if encapcheck method returns exact_match value, the search will be stopped. - all current consumers changed to use new KPI. Reviewed by: mmacy Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D15617
2018-06-05 20:51:01 +00:00
#endif
static struct mtx encapmtx, srcaddrmtx;
MTX_SYSINIT(encapmtx, &encapmtx, "encapmtx", MTX_DEF);
MTX_SYSINIT(srcaddrmtx, &srcaddrmtx, "srcaddrmtx", MTX_DEF);
Rework IP encapsulation handling code. Currently it has several disadvantages: - it uses single mutex to protect internal structures. It is used by data- and control- path, thus there are no parallelism at all. - it uses single list to keep encap handlers for both INET and INET6 families. - struct encaptab keeps unneeded information (src, dst, masks, protosw), that isn't used by code in the source tree. - matches are prioritized and when many tunneling interfaces are registered, encapcheck handler of each interface is invoked for each packet. The search takes O(n) for n interfaces. All this work is done with exclusive lock held. What this patch includes: - the datapath is converted to be lockless using epoch(9) KPI. - struct encaptab now linked using CK_LIST. - all unused fields removed from struct encaptab. Several new fields addedr: min_length is the minimum packet length, that encapsulation handler expects to see; exact_match is maximum number of bits, that can return an encapsulation handler, when it wants to consume a packet. - IPv6 and IPv4 handlers are stored in separate lists; - added new "encap_lookup_t" method, that will be used later. It is targeted to speedup lookup of needed interface, when gif(4)/gre(4) have many interfaces. - the need to use protosw structure is eliminated. The only pr_input method was used from this structure, so I don't see the need to keep using it. - encap_input_t method changed to avoid using mbuf tags to store softc pointer. Now it is passed directly trough encap_input_t method. encap_getarg() funtions is removed. - all sockaddr structures and code that uses them removed. We don't have any code in the tree that uses them. All consumers use encap_attach_func() method, that relies on invoking of encapcheck() to determine the needed handler. - introduced struct encap_config, it contains parameters of encap handler that is going to be registered by encap_attach() function. - encap handlers are stored in lists ordered by exact_match value, thus handlers that need more bits to match will be checked first, and if encapcheck method returns exact_match value, the search will be stopped. - all current consumers changed to use new KPI. Reviewed by: mmacy Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D15617
2018-06-05 20:51:01 +00:00
#define ENCAP_WLOCK() mtx_lock(&encapmtx)
#define ENCAP_WUNLOCK() mtx_unlock(&encapmtx)
#define ENCAP_RLOCK_TRACKER struct epoch_tracker encap_et
Widen NET_EPOCH coverage. When epoch(9) was introduced to network stack, it was basically dropped in place of existing locking, which was mutexes and rwlocks. For the sake of performance mutex covered areas were as small as possible, so became epoch covered areas. However, epoch doesn't introduce any contention, it just delays memory reclaim. So, there is no point to minimise epoch covered areas in sense of performance. Meanwhile entering/exiting epoch also has non-zero CPU usage, so doing this less often is a win. Not the least is also code maintainability. In the new paradigm we can assume that at any stage of processing a packet, we are inside network epoch. This makes coding both input and output path way easier. On output path we already enter epoch quite early - in the ip_output(), in the ip6_output(). This patch does the same for the input path. All ISR processing, network related callouts, other ways of packet injection to the network stack shall be performed in net_epoch. Any leaf function that walks network configuration now asserts epoch. Tricky part is configuration code paths - ioctls, sysctls. They also call into leaf functions, so some need to be changed. This patch would introduce more epoch recursions (see EPOCH_TRACE) than we had before. They will be cleaned up separately, as several of them aren't trivial. Note, that unlike a lock recursion the epoch recursion is safe and just wastes a bit of resources. Reviewed by: gallatin, hselasky, cy, adrian, kristof Differential Revision: https://reviews.freebsd.org/D19111
2019-10-07 22:40:05 +00:00
#define ENCAP_RLOCK() NET_EPOCH_ENTER(encap_et)
#define ENCAP_RUNLOCK() NET_EPOCH_EXIT(encap_et)
#define ENCAP_WAIT() NET_EPOCH_WAIT()
Rework IP encapsulation handling code. Currently it has several disadvantages: - it uses single mutex to protect internal structures. It is used by data- and control- path, thus there are no parallelism at all. - it uses single list to keep encap handlers for both INET and INET6 families. - struct encaptab keeps unneeded information (src, dst, masks, protosw), that isn't used by code in the source tree. - matches are prioritized and when many tunneling interfaces are registered, encapcheck handler of each interface is invoked for each packet. The search takes O(n) for n interfaces. All this work is done with exclusive lock held. What this patch includes: - the datapath is converted to be lockless using epoch(9) KPI. - struct encaptab now linked using CK_LIST. - all unused fields removed from struct encaptab. Several new fields addedr: min_length is the minimum packet length, that encapsulation handler expects to see; exact_match is maximum number of bits, that can return an encapsulation handler, when it wants to consume a packet. - IPv6 and IPv4 handlers are stored in separate lists; - added new "encap_lookup_t" method, that will be used later. It is targeted to speedup lookup of needed interface, when gif(4)/gre(4) have many interfaces. - the need to use protosw structure is eliminated. The only pr_input method was used from this structure, so I don't see the need to keep using it. - encap_input_t method changed to avoid using mbuf tags to store softc pointer. Now it is passed directly trough encap_input_t method. encap_getarg() funtions is removed. - all sockaddr structures and code that uses them removed. We don't have any code in the tree that uses them. All consumers use encap_attach_func() method, that relies on invoking of encapcheck() to determine the needed handler. - introduced struct encap_config, it contains parameters of encap handler that is going to be registered by encap_attach() function. - encap handlers are stored in lists ordered by exact_match value, thus handlers that need more bits to match will be checked first, and if encapcheck method returns exact_match value, the search will be stopped. - all current consumers changed to use new KPI. Reviewed by: mmacy Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D15617
2018-06-05 20:51:01 +00:00
#define SRCADDR_WLOCK() mtx_lock(&srcaddrmtx)
#define SRCADDR_WUNLOCK() mtx_unlock(&srcaddrmtx)
#define SRCADDR_RLOCK_TRACKER struct epoch_tracker srcaddr_et
#define SRCADDR_RLOCK() \
epoch_enter_preempt(net_epoch_preempt, &srcaddr_et)
#define SRCADDR_RUNLOCK() \
epoch_exit_preempt(net_epoch_preempt, &srcaddr_et)
#define SRCADDR_WAIT() epoch_wait_preempt(net_epoch_preempt)
/*
* ifaddr_event_ext handler.
*
* Tunnelling interfaces may request the kernel to notify when
* some interface addresses appears or disappears. Usually tunnelling
* interface must use an address configured on the local machine as
* ingress address to be able receive datagramms and do not send
* spoofed packets.
*/
static void
srcaddr_change_event(void *arg __unused, struct ifnet *ifp,
struct ifaddr *ifa, int event)
{
SRCADDR_RLOCK_TRACKER;
struct srcaddrtab_head *head;
struct srcaddrtab *p;
/* Support for old ifaddr_event. */
EVENTHANDLER_INVOKE(ifaddr_event, ifp);
switch (ifa->ifa_addr->sa_family) {
#ifdef INET
case AF_INET:
head = &ipv4_srcaddrtab;
break;
#endif
#ifdef INET6
case AF_INET6:
head = &ipv6_srcaddrtab;
break;
#endif
default:
/* ignore event */
return;
}
SRCADDR_RLOCK();
CK_LIST_FOREACH(p, head, chain) {
(*p->srcaddr)(p->arg, ifa->ifa_addr, event);
}
SRCADDR_RUNLOCK();
}
EVENTHANDLER_DEFINE(ifaddr_event_ext, srcaddr_change_event, NULL, 0);
static struct srcaddrtab *
encap_register_srcaddr(struct srcaddrtab_head *head, encap_srcaddr_t func,
void *arg, int mflags)
{
struct srcaddrtab *p, *tmp;
if (func == NULL)
return (NULL);
p = malloc(sizeof(*p), M_NETADDR, mflags);
if (p == NULL)
return (NULL);
p->srcaddr = func;
p->arg = arg;
SRCADDR_WLOCK();
CK_LIST_FOREACH(tmp, head, chain) {
if (func == tmp->srcaddr && arg == tmp->arg)
break;
}
if (tmp == NULL)
CK_LIST_INSERT_HEAD(head, p, chain);
SRCADDR_WUNLOCK();
if (tmp != NULL) {
free(p, M_NETADDR);
p = tmp;
}
return (p);
}
static int
encap_unregister_srcaddr(struct srcaddrtab_head *head,
const struct srcaddrtab *cookie)
{
struct srcaddrtab *p;
SRCADDR_WLOCK();
CK_LIST_FOREACH(p, head, chain) {
if (p == cookie) {
CK_LIST_REMOVE(p, chain);
SRCADDR_WUNLOCK();
SRCADDR_WAIT();
free(p, M_NETADDR);
return (0);
}
}
SRCADDR_WUNLOCK();
return (EINVAL);
}
Rework IP encapsulation handling code. Currently it has several disadvantages: - it uses single mutex to protect internal structures. It is used by data- and control- path, thus there are no parallelism at all. - it uses single list to keep encap handlers for both INET and INET6 families. - struct encaptab keeps unneeded information (src, dst, masks, protosw), that isn't used by code in the source tree. - matches are prioritized and when many tunneling interfaces are registered, encapcheck handler of each interface is invoked for each packet. The search takes O(n) for n interfaces. All this work is done with exclusive lock held. What this patch includes: - the datapath is converted to be lockless using epoch(9) KPI. - struct encaptab now linked using CK_LIST. - all unused fields removed from struct encaptab. Several new fields addedr: min_length is the minimum packet length, that encapsulation handler expects to see; exact_match is maximum number of bits, that can return an encapsulation handler, when it wants to consume a packet. - IPv6 and IPv4 handlers are stored in separate lists; - added new "encap_lookup_t" method, that will be used later. It is targeted to speedup lookup of needed interface, when gif(4)/gre(4) have many interfaces. - the need to use protosw structure is eliminated. The only pr_input method was used from this structure, so I don't see the need to keep using it. - encap_input_t method changed to avoid using mbuf tags to store softc pointer. Now it is passed directly trough encap_input_t method. encap_getarg() funtions is removed. - all sockaddr structures and code that uses them removed. We don't have any code in the tree that uses them. All consumers use encap_attach_func() method, that relies on invoking of encapcheck() to determine the needed handler. - introduced struct encap_config, it contains parameters of encap handler that is going to be registered by encap_attach() function. - encap handlers are stored in lists ordered by exact_match value, thus handlers that need more bits to match will be checked first, and if encapcheck method returns exact_match value, the search will be stopped. - all current consumers changed to use new KPI. Reviewed by: mmacy Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D15617
2018-06-05 20:51:01 +00:00
static struct encaptab *
encap_attach(struct encaptab_head *head, const struct encap_config *cfg,
void *arg, int mflags)
{
Rework IP encapsulation handling code. Currently it has several disadvantages: - it uses single mutex to protect internal structures. It is used by data- and control- path, thus there are no parallelism at all. - it uses single list to keep encap handlers for both INET and INET6 families. - struct encaptab keeps unneeded information (src, dst, masks, protosw), that isn't used by code in the source tree. - matches are prioritized and when many tunneling interfaces are registered, encapcheck handler of each interface is invoked for each packet. The search takes O(n) for n interfaces. All this work is done with exclusive lock held. What this patch includes: - the datapath is converted to be lockless using epoch(9) KPI. - struct encaptab now linked using CK_LIST. - all unused fields removed from struct encaptab. Several new fields addedr: min_length is the minimum packet length, that encapsulation handler expects to see; exact_match is maximum number of bits, that can return an encapsulation handler, when it wants to consume a packet. - IPv6 and IPv4 handlers are stored in separate lists; - added new "encap_lookup_t" method, that will be used later. It is targeted to speedup lookup of needed interface, when gif(4)/gre(4) have many interfaces. - the need to use protosw structure is eliminated. The only pr_input method was used from this structure, so I don't see the need to keep using it. - encap_input_t method changed to avoid using mbuf tags to store softc pointer. Now it is passed directly trough encap_input_t method. encap_getarg() funtions is removed. - all sockaddr structures and code that uses them removed. We don't have any code in the tree that uses them. All consumers use encap_attach_func() method, that relies on invoking of encapcheck() to determine the needed handler. - introduced struct encap_config, it contains parameters of encap handler that is going to be registered by encap_attach() function. - encap handlers are stored in lists ordered by exact_match value, thus handlers that need more bits to match will be checked first, and if encapcheck method returns exact_match value, the search will be stopped. - all current consumers changed to use new KPI. Reviewed by: mmacy Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D15617
2018-06-05 20:51:01 +00:00
struct encaptab *ep, *tmp;
Rework IP encapsulation handling code. Currently it has several disadvantages: - it uses single mutex to protect internal structures. It is used by data- and control- path, thus there are no parallelism at all. - it uses single list to keep encap handlers for both INET and INET6 families. - struct encaptab keeps unneeded information (src, dst, masks, protosw), that isn't used by code in the source tree. - matches are prioritized and when many tunneling interfaces are registered, encapcheck handler of each interface is invoked for each packet. The search takes O(n) for n interfaces. All this work is done with exclusive lock held. What this patch includes: - the datapath is converted to be lockless using epoch(9) KPI. - struct encaptab now linked using CK_LIST. - all unused fields removed from struct encaptab. Several new fields addedr: min_length is the minimum packet length, that encapsulation handler expects to see; exact_match is maximum number of bits, that can return an encapsulation handler, when it wants to consume a packet. - IPv6 and IPv4 handlers are stored in separate lists; - added new "encap_lookup_t" method, that will be used later. It is targeted to speedup lookup of needed interface, when gif(4)/gre(4) have many interfaces. - the need to use protosw structure is eliminated. The only pr_input method was used from this structure, so I don't see the need to keep using it. - encap_input_t method changed to avoid using mbuf tags to store softc pointer. Now it is passed directly trough encap_input_t method. encap_getarg() funtions is removed. - all sockaddr structures and code that uses them removed. We don't have any code in the tree that uses them. All consumers use encap_attach_func() method, that relies on invoking of encapcheck() to determine the needed handler. - introduced struct encap_config, it contains parameters of encap handler that is going to be registered by encap_attach() function. - encap handlers are stored in lists ordered by exact_match value, thus handlers that need more bits to match will be checked first, and if encapcheck method returns exact_match value, the search will be stopped. - all current consumers changed to use new KPI. Reviewed by: mmacy Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D15617
2018-06-05 20:51:01 +00:00
if (cfg == NULL || cfg->input == NULL ||
(cfg->check == NULL && cfg->lookup == NULL) ||
(cfg->lookup != NULL && cfg->exact_match != ENCAP_DRV_LOOKUP) ||
(cfg->exact_match == ENCAP_DRV_LOOKUP && cfg->lookup == NULL))
return (NULL);
Rework IP encapsulation handling code. Currently it has several disadvantages: - it uses single mutex to protect internal structures. It is used by data- and control- path, thus there are no parallelism at all. - it uses single list to keep encap handlers for both INET and INET6 families. - struct encaptab keeps unneeded information (src, dst, masks, protosw), that isn't used by code in the source tree. - matches are prioritized and when many tunneling interfaces are registered, encapcheck handler of each interface is invoked for each packet. The search takes O(n) for n interfaces. All this work is done with exclusive lock held. What this patch includes: - the datapath is converted to be lockless using epoch(9) KPI. - struct encaptab now linked using CK_LIST. - all unused fields removed from struct encaptab. Several new fields addedr: min_length is the minimum packet length, that encapsulation handler expects to see; exact_match is maximum number of bits, that can return an encapsulation handler, when it wants to consume a packet. - IPv6 and IPv4 handlers are stored in separate lists; - added new "encap_lookup_t" method, that will be used later. It is targeted to speedup lookup of needed interface, when gif(4)/gre(4) have many interfaces. - the need to use protosw structure is eliminated. The only pr_input method was used from this structure, so I don't see the need to keep using it. - encap_input_t method changed to avoid using mbuf tags to store softc pointer. Now it is passed directly trough encap_input_t method. encap_getarg() funtions is removed. - all sockaddr structures and code that uses them removed. We don't have any code in the tree that uses them. All consumers use encap_attach_func() method, that relies on invoking of encapcheck() to determine the needed handler. - introduced struct encap_config, it contains parameters of encap handler that is going to be registered by encap_attach() function. - encap handlers are stored in lists ordered by exact_match value, thus handlers that need more bits to match will be checked first, and if encapcheck method returns exact_match value, the search will be stopped. - all current consumers changed to use new KPI. Reviewed by: mmacy Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D15617
2018-06-05 20:51:01 +00:00
ep = malloc(sizeof(*ep), M_NETADDR, mflags);
if (ep == NULL)
return (NULL);
ep->proto = cfg->proto;
ep->min_length = cfg->min_length;
ep->exact_match = cfg->exact_match;
ep->arg = arg;
ep->lookup = cfg->exact_match == ENCAP_DRV_LOOKUP ? cfg->lookup: NULL;
ep->check = cfg->exact_match != ENCAP_DRV_LOOKUP ? cfg->check: NULL;
ep->input = cfg->input;
ENCAP_WLOCK();
CK_LIST_FOREACH(tmp, head, chain) {
if (tmp->exact_match <= ep->exact_match)
break;
}
Rework IP encapsulation handling code. Currently it has several disadvantages: - it uses single mutex to protect internal structures. It is used by data- and control- path, thus there are no parallelism at all. - it uses single list to keep encap handlers for both INET and INET6 families. - struct encaptab keeps unneeded information (src, dst, masks, protosw), that isn't used by code in the source tree. - matches are prioritized and when many tunneling interfaces are registered, encapcheck handler of each interface is invoked for each packet. The search takes O(n) for n interfaces. All this work is done with exclusive lock held. What this patch includes: - the datapath is converted to be lockless using epoch(9) KPI. - struct encaptab now linked using CK_LIST. - all unused fields removed from struct encaptab. Several new fields addedr: min_length is the minimum packet length, that encapsulation handler expects to see; exact_match is maximum number of bits, that can return an encapsulation handler, when it wants to consume a packet. - IPv6 and IPv4 handlers are stored in separate lists; - added new "encap_lookup_t" method, that will be used later. It is targeted to speedup lookup of needed interface, when gif(4)/gre(4) have many interfaces. - the need to use protosw structure is eliminated. The only pr_input method was used from this structure, so I don't see the need to keep using it. - encap_input_t method changed to avoid using mbuf tags to store softc pointer. Now it is passed directly trough encap_input_t method. encap_getarg() funtions is removed. - all sockaddr structures and code that uses them removed. We don't have any code in the tree that uses them. All consumers use encap_attach_func() method, that relies on invoking of encapcheck() to determine the needed handler. - introduced struct encap_config, it contains parameters of encap handler that is going to be registered by encap_attach() function. - encap handlers are stored in lists ordered by exact_match value, thus handlers that need more bits to match will be checked first, and if encapcheck method returns exact_match value, the search will be stopped. - all current consumers changed to use new KPI. Reviewed by: mmacy Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D15617
2018-06-05 20:51:01 +00:00
if (tmp == NULL)
CK_LIST_INSERT_HEAD(head, ep, chain);
else
CK_LIST_INSERT_BEFORE(tmp, ep, chain);
ENCAP_WUNLOCK();
return (ep);
}
static int
encap_detach(struct encaptab_head *head, const struct encaptab *cookie)
{
struct encaptab *ep;
Rework IP encapsulation handling code. Currently it has several disadvantages: - it uses single mutex to protect internal structures. It is used by data- and control- path, thus there are no parallelism at all. - it uses single list to keep encap handlers for both INET and INET6 families. - struct encaptab keeps unneeded information (src, dst, masks, protosw), that isn't used by code in the source tree. - matches are prioritized and when many tunneling interfaces are registered, encapcheck handler of each interface is invoked for each packet. The search takes O(n) for n interfaces. All this work is done with exclusive lock held. What this patch includes: - the datapath is converted to be lockless using epoch(9) KPI. - struct encaptab now linked using CK_LIST. - all unused fields removed from struct encaptab. Several new fields addedr: min_length is the minimum packet length, that encapsulation handler expects to see; exact_match is maximum number of bits, that can return an encapsulation handler, when it wants to consume a packet. - IPv6 and IPv4 handlers are stored in separate lists; - added new "encap_lookup_t" method, that will be used later. It is targeted to speedup lookup of needed interface, when gif(4)/gre(4) have many interfaces. - the need to use protosw structure is eliminated. The only pr_input method was used from this structure, so I don't see the need to keep using it. - encap_input_t method changed to avoid using mbuf tags to store softc pointer. Now it is passed directly trough encap_input_t method. encap_getarg() funtions is removed. - all sockaddr structures and code that uses them removed. We don't have any code in the tree that uses them. All consumers use encap_attach_func() method, that relies on invoking of encapcheck() to determine the needed handler. - introduced struct encap_config, it contains parameters of encap handler that is going to be registered by encap_attach() function. - encap handlers are stored in lists ordered by exact_match value, thus handlers that need more bits to match will be checked first, and if encapcheck method returns exact_match value, the search will be stopped. - all current consumers changed to use new KPI. Reviewed by: mmacy Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D15617
2018-06-05 20:51:01 +00:00
ENCAP_WLOCK();
CK_LIST_FOREACH(ep, head, chain) {
if (ep == cookie) {
CK_LIST_REMOVE(ep, chain);
ENCAP_WUNLOCK();
ENCAP_WAIT();
free(ep, M_NETADDR);
return (0);
}
}
ENCAP_WUNLOCK();
return (EINVAL);
}
Rework IP encapsulation handling code. Currently it has several disadvantages: - it uses single mutex to protect internal structures. It is used by data- and control- path, thus there are no parallelism at all. - it uses single list to keep encap handlers for both INET and INET6 families. - struct encaptab keeps unneeded information (src, dst, masks, protosw), that isn't used by code in the source tree. - matches are prioritized and when many tunneling interfaces are registered, encapcheck handler of each interface is invoked for each packet. The search takes O(n) for n interfaces. All this work is done with exclusive lock held. What this patch includes: - the datapath is converted to be lockless using epoch(9) KPI. - struct encaptab now linked using CK_LIST. - all unused fields removed from struct encaptab. Several new fields addedr: min_length is the minimum packet length, that encapsulation handler expects to see; exact_match is maximum number of bits, that can return an encapsulation handler, when it wants to consume a packet. - IPv6 and IPv4 handlers are stored in separate lists; - added new "encap_lookup_t" method, that will be used later. It is targeted to speedup lookup of needed interface, when gif(4)/gre(4) have many interfaces. - the need to use protosw structure is eliminated. The only pr_input method was used from this structure, so I don't see the need to keep using it. - encap_input_t method changed to avoid using mbuf tags to store softc pointer. Now it is passed directly trough encap_input_t method. encap_getarg() funtions is removed. - all sockaddr structures and code that uses them removed. We don't have any code in the tree that uses them. All consumers use encap_attach_func() method, that relies on invoking of encapcheck() to determine the needed handler. - introduced struct encap_config, it contains parameters of encap handler that is going to be registered by encap_attach() function. - encap handlers are stored in lists ordered by exact_match value, thus handlers that need more bits to match will be checked first, and if encapcheck method returns exact_match value, the search will be stopped. - all current consumers changed to use new KPI. Reviewed by: mmacy Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D15617
2018-06-05 20:51:01 +00:00
static int
encap_input(struct encaptab_head *head, struct mbuf *m, int off, int proto)
{
ENCAP_RLOCK_TRACKER;
struct encaptab *ep, *match;
void *arg;
Rework IP encapsulation handling code. Currently it has several disadvantages: - it uses single mutex to protect internal structures. It is used by data- and control- path, thus there are no parallelism at all. - it uses single list to keep encap handlers for both INET and INET6 families. - struct encaptab keeps unneeded information (src, dst, masks, protosw), that isn't used by code in the source tree. - matches are prioritized and when many tunneling interfaces are registered, encapcheck handler of each interface is invoked for each packet. The search takes O(n) for n interfaces. All this work is done with exclusive lock held. What this patch includes: - the datapath is converted to be lockless using epoch(9) KPI. - struct encaptab now linked using CK_LIST. - all unused fields removed from struct encaptab. Several new fields addedr: min_length is the minimum packet length, that encapsulation handler expects to see; exact_match is maximum number of bits, that can return an encapsulation handler, when it wants to consume a packet. - IPv6 and IPv4 handlers are stored in separate lists; - added new "encap_lookup_t" method, that will be used later. It is targeted to speedup lookup of needed interface, when gif(4)/gre(4) have many interfaces. - the need to use protosw structure is eliminated. The only pr_input method was used from this structure, so I don't see the need to keep using it. - encap_input_t method changed to avoid using mbuf tags to store softc pointer. Now it is passed directly trough encap_input_t method. encap_getarg() funtions is removed. - all sockaddr structures and code that uses them removed. We don't have any code in the tree that uses them. All consumers use encap_attach_func() method, that relies on invoking of encapcheck() to determine the needed handler. - introduced struct encap_config, it contains parameters of encap handler that is going to be registered by encap_attach() function. - encap handlers are stored in lists ordered by exact_match value, thus handlers that need more bits to match will be checked first, and if encapcheck method returns exact_match value, the search will be stopped. - all current consumers changed to use new KPI. Reviewed by: mmacy Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D15617
2018-06-05 20:51:01 +00:00
int matchprio, ret;
match = NULL;
matchprio = 0;
Rework IP encapsulation handling code. Currently it has several disadvantages: - it uses single mutex to protect internal structures. It is used by data- and control- path, thus there are no parallelism at all. - it uses single list to keep encap handlers for both INET and INET6 families. - struct encaptab keeps unneeded information (src, dst, masks, protosw), that isn't used by code in the source tree. - matches are prioritized and when many tunneling interfaces are registered, encapcheck handler of each interface is invoked for each packet. The search takes O(n) for n interfaces. All this work is done with exclusive lock held. What this patch includes: - the datapath is converted to be lockless using epoch(9) KPI. - struct encaptab now linked using CK_LIST. - all unused fields removed from struct encaptab. Several new fields addedr: min_length is the minimum packet length, that encapsulation handler expects to see; exact_match is maximum number of bits, that can return an encapsulation handler, when it wants to consume a packet. - IPv6 and IPv4 handlers are stored in separate lists; - added new "encap_lookup_t" method, that will be used later. It is targeted to speedup lookup of needed interface, when gif(4)/gre(4) have many interfaces. - the need to use protosw structure is eliminated. The only pr_input method was used from this structure, so I don't see the need to keep using it. - encap_input_t method changed to avoid using mbuf tags to store softc pointer. Now it is passed directly trough encap_input_t method. encap_getarg() funtions is removed. - all sockaddr structures and code that uses them removed. We don't have any code in the tree that uses them. All consumers use encap_attach_func() method, that relies on invoking of encapcheck() to determine the needed handler. - introduced struct encap_config, it contains parameters of encap handler that is going to be registered by encap_attach() function. - encap handlers are stored in lists ordered by exact_match value, thus handlers that need more bits to match will be checked first, and if encapcheck method returns exact_match value, the search will be stopped. - all current consumers changed to use new KPI. Reviewed by: mmacy Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D15617
2018-06-05 20:51:01 +00:00
ENCAP_RLOCK();
CK_LIST_FOREACH(ep, head, chain) {
if (ep->proto >= 0 && ep->proto != proto)
continue;
Rework IP encapsulation handling code. Currently it has several disadvantages: - it uses single mutex to protect internal structures. It is used by data- and control- path, thus there are no parallelism at all. - it uses single list to keep encap handlers for both INET and INET6 families. - struct encaptab keeps unneeded information (src, dst, masks, protosw), that isn't used by code in the source tree. - matches are prioritized and when many tunneling interfaces are registered, encapcheck handler of each interface is invoked for each packet. The search takes O(n) for n interfaces. All this work is done with exclusive lock held. What this patch includes: - the datapath is converted to be lockless using epoch(9) KPI. - struct encaptab now linked using CK_LIST. - all unused fields removed from struct encaptab. Several new fields addedr: min_length is the minimum packet length, that encapsulation handler expects to see; exact_match is maximum number of bits, that can return an encapsulation handler, when it wants to consume a packet. - IPv6 and IPv4 handlers are stored in separate lists; - added new "encap_lookup_t" method, that will be used later. It is targeted to speedup lookup of needed interface, when gif(4)/gre(4) have many interfaces. - the need to use protosw structure is eliminated. The only pr_input method was used from this structure, so I don't see the need to keep using it. - encap_input_t method changed to avoid using mbuf tags to store softc pointer. Now it is passed directly trough encap_input_t method. encap_getarg() funtions is removed. - all sockaddr structures and code that uses them removed. We don't have any code in the tree that uses them. All consumers use encap_attach_func() method, that relies on invoking of encapcheck() to determine the needed handler. - introduced struct encap_config, it contains parameters of encap handler that is going to be registered by encap_attach() function. - encap handlers are stored in lists ordered by exact_match value, thus handlers that need more bits to match will be checked first, and if encapcheck method returns exact_match value, the search will be stopped. - all current consumers changed to use new KPI. Reviewed by: mmacy Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D15617
2018-06-05 20:51:01 +00:00
if (ep->min_length > m->m_pkthdr.len)
continue;
Rework IP encapsulation handling code. Currently it has several disadvantages: - it uses single mutex to protect internal structures. It is used by data- and control- path, thus there are no parallelism at all. - it uses single list to keep encap handlers for both INET and INET6 families. - struct encaptab keeps unneeded information (src, dst, masks, protosw), that isn't used by code in the source tree. - matches are prioritized and when many tunneling interfaces are registered, encapcheck handler of each interface is invoked for each packet. The search takes O(n) for n interfaces. All this work is done with exclusive lock held. What this patch includes: - the datapath is converted to be lockless using epoch(9) KPI. - struct encaptab now linked using CK_LIST. - all unused fields removed from struct encaptab. Several new fields addedr: min_length is the minimum packet length, that encapsulation handler expects to see; exact_match is maximum number of bits, that can return an encapsulation handler, when it wants to consume a packet. - IPv6 and IPv4 handlers are stored in separate lists; - added new "encap_lookup_t" method, that will be used later. It is targeted to speedup lookup of needed interface, when gif(4)/gre(4) have many interfaces. - the need to use protosw structure is eliminated. The only pr_input method was used from this structure, so I don't see the need to keep using it. - encap_input_t method changed to avoid using mbuf tags to store softc pointer. Now it is passed directly trough encap_input_t method. encap_getarg() funtions is removed. - all sockaddr structures and code that uses them removed. We don't have any code in the tree that uses them. All consumers use encap_attach_func() method, that relies on invoking of encapcheck() to determine the needed handler. - introduced struct encap_config, it contains parameters of encap handler that is going to be registered by encap_attach() function. - encap handlers are stored in lists ordered by exact_match value, thus handlers that need more bits to match will be checked first, and if encapcheck method returns exact_match value, the search will be stopped. - all current consumers changed to use new KPI. Reviewed by: mmacy Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D15617
2018-06-05 20:51:01 +00:00
if (ep->exact_match == ENCAP_DRV_LOOKUP)
ret = (*ep->lookup)(m, off, proto, &arg);
else
ret = (*ep->check)(m, off, proto, ep->arg);
if (ret <= 0)
continue;
if (ret > matchprio) {
match = ep;
Rework IP encapsulation handling code. Currently it has several disadvantages: - it uses single mutex to protect internal structures. It is used by data- and control- path, thus there are no parallelism at all. - it uses single list to keep encap handlers for both INET and INET6 families. - struct encaptab keeps unneeded information (src, dst, masks, protosw), that isn't used by code in the source tree. - matches are prioritized and when many tunneling interfaces are registered, encapcheck handler of each interface is invoked for each packet. The search takes O(n) for n interfaces. All this work is done with exclusive lock held. What this patch includes: - the datapath is converted to be lockless using epoch(9) KPI. - struct encaptab now linked using CK_LIST. - all unused fields removed from struct encaptab. Several new fields addedr: min_length is the minimum packet length, that encapsulation handler expects to see; exact_match is maximum number of bits, that can return an encapsulation handler, when it wants to consume a packet. - IPv6 and IPv4 handlers are stored in separate lists; - added new "encap_lookup_t" method, that will be used later. It is targeted to speedup lookup of needed interface, when gif(4)/gre(4) have many interfaces. - the need to use protosw structure is eliminated. The only pr_input method was used from this structure, so I don't see the need to keep using it. - encap_input_t method changed to avoid using mbuf tags to store softc pointer. Now it is passed directly trough encap_input_t method. encap_getarg() funtions is removed. - all sockaddr structures and code that uses them removed. We don't have any code in the tree that uses them. All consumers use encap_attach_func() method, that relies on invoking of encapcheck() to determine the needed handler. - introduced struct encap_config, it contains parameters of encap handler that is going to be registered by encap_attach() function. - encap handlers are stored in lists ordered by exact_match value, thus handlers that need more bits to match will be checked first, and if encapcheck method returns exact_match value, the search will be stopped. - all current consumers changed to use new KPI. Reviewed by: mmacy Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D15617
2018-06-05 20:51:01 +00:00
if (ep->exact_match != ENCAP_DRV_LOOKUP)
arg = ep->arg;
/*
* No need to continue the search, we got the
* exact match.
*/
if (ret >= ep->exact_match)
break;
matchprio = ret;
}
}
if (match != NULL) {
Rework IP encapsulation handling code. Currently it has several disadvantages: - it uses single mutex to protect internal structures. It is used by data- and control- path, thus there are no parallelism at all. - it uses single list to keep encap handlers for both INET and INET6 families. - struct encaptab keeps unneeded information (src, dst, masks, protosw), that isn't used by code in the source tree. - matches are prioritized and when many tunneling interfaces are registered, encapcheck handler of each interface is invoked for each packet. The search takes O(n) for n interfaces. All this work is done with exclusive lock held. What this patch includes: - the datapath is converted to be lockless using epoch(9) KPI. - struct encaptab now linked using CK_LIST. - all unused fields removed from struct encaptab. Several new fields addedr: min_length is the minimum packet length, that encapsulation handler expects to see; exact_match is maximum number of bits, that can return an encapsulation handler, when it wants to consume a packet. - IPv6 and IPv4 handlers are stored in separate lists; - added new "encap_lookup_t" method, that will be used later. It is targeted to speedup lookup of needed interface, when gif(4)/gre(4) have many interfaces. - the need to use protosw structure is eliminated. The only pr_input method was used from this structure, so I don't see the need to keep using it. - encap_input_t method changed to avoid using mbuf tags to store softc pointer. Now it is passed directly trough encap_input_t method. encap_getarg() funtions is removed. - all sockaddr structures and code that uses them removed. We don't have any code in the tree that uses them. All consumers use encap_attach_func() method, that relies on invoking of encapcheck() to determine the needed handler. - introduced struct encap_config, it contains parameters of encap handler that is going to be registered by encap_attach() function. - encap handlers are stored in lists ordered by exact_match value, thus handlers that need more bits to match will be checked first, and if encapcheck method returns exact_match value, the search will be stopped. - all current consumers changed to use new KPI. Reviewed by: mmacy Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D15617
2018-06-05 20:51:01 +00:00
/* found a match, "match" has the best one */
ret = (*match->input)(m, off, proto, arg);
ENCAP_RUNLOCK();
MPASS(ret == IPPROTO_DONE);
return (IPPROTO_DONE);
}
Rework IP encapsulation handling code. Currently it has several disadvantages: - it uses single mutex to protect internal structures. It is used by data- and control- path, thus there are no parallelism at all. - it uses single list to keep encap handlers for both INET and INET6 families. - struct encaptab keeps unneeded information (src, dst, masks, protosw), that isn't used by code in the source tree. - matches are prioritized and when many tunneling interfaces are registered, encapcheck handler of each interface is invoked for each packet. The search takes O(n) for n interfaces. All this work is done with exclusive lock held. What this patch includes: - the datapath is converted to be lockless using epoch(9) KPI. - struct encaptab now linked using CK_LIST. - all unused fields removed from struct encaptab. Several new fields addedr: min_length is the minimum packet length, that encapsulation handler expects to see; exact_match is maximum number of bits, that can return an encapsulation handler, when it wants to consume a packet. - IPv6 and IPv4 handlers are stored in separate lists; - added new "encap_lookup_t" method, that will be used later. It is targeted to speedup lookup of needed interface, when gif(4)/gre(4) have many interfaces. - the need to use protosw structure is eliminated. The only pr_input method was used from this structure, so I don't see the need to keep using it. - encap_input_t method changed to avoid using mbuf tags to store softc pointer. Now it is passed directly trough encap_input_t method. encap_getarg() funtions is removed. - all sockaddr structures and code that uses them removed. We don't have any code in the tree that uses them. All consumers use encap_attach_func() method, that relies on invoking of encapcheck() to determine the needed handler. - introduced struct encap_config, it contains parameters of encap handler that is going to be registered by encap_attach() function. - encap handlers are stored in lists ordered by exact_match value, thus handlers that need more bits to match will be checked first, and if encapcheck method returns exact_match value, the search will be stopped. - all current consumers changed to use new KPI. Reviewed by: mmacy Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D15617
2018-06-05 20:51:01 +00:00
ENCAP_RUNLOCK();
return (0);
}
Rework IP encapsulation handling code. Currently it has several disadvantages: - it uses single mutex to protect internal structures. It is used by data- and control- path, thus there are no parallelism at all. - it uses single list to keep encap handlers for both INET and INET6 families. - struct encaptab keeps unneeded information (src, dst, masks, protosw), that isn't used by code in the source tree. - matches are prioritized and when many tunneling interfaces are registered, encapcheck handler of each interface is invoked for each packet. The search takes O(n) for n interfaces. All this work is done with exclusive lock held. What this patch includes: - the datapath is converted to be lockless using epoch(9) KPI. - struct encaptab now linked using CK_LIST. - all unused fields removed from struct encaptab. Several new fields addedr: min_length is the minimum packet length, that encapsulation handler expects to see; exact_match is maximum number of bits, that can return an encapsulation handler, when it wants to consume a packet. - IPv6 and IPv4 handlers are stored in separate lists; - added new "encap_lookup_t" method, that will be used later. It is targeted to speedup lookup of needed interface, when gif(4)/gre(4) have many interfaces. - the need to use protosw structure is eliminated. The only pr_input method was used from this structure, so I don't see the need to keep using it. - encap_input_t method changed to avoid using mbuf tags to store softc pointer. Now it is passed directly trough encap_input_t method. encap_getarg() funtions is removed. - all sockaddr structures and code that uses them removed. We don't have any code in the tree that uses them. All consumers use encap_attach_func() method, that relies on invoking of encapcheck() to determine the needed handler. - introduced struct encap_config, it contains parameters of encap handler that is going to be registered by encap_attach() function. - encap handlers are stored in lists ordered by exact_match value, thus handlers that need more bits to match will be checked first, and if encapcheck method returns exact_match value, the search will be stopped. - all current consumers changed to use new KPI. Reviewed by: mmacy Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D15617
2018-06-05 20:51:01 +00:00
#ifdef INET
const struct srcaddrtab *
ip_encap_register_srcaddr(encap_srcaddr_t func, void *arg, int mflags)
{
return (encap_register_srcaddr(&ipv4_srcaddrtab, func, arg, mflags));
}
int
ip_encap_unregister_srcaddr(const struct srcaddrtab *cookie)
{
return (encap_unregister_srcaddr(&ipv4_srcaddrtab, cookie));
}
const struct encaptab *
Rework IP encapsulation handling code. Currently it has several disadvantages: - it uses single mutex to protect internal structures. It is used by data- and control- path, thus there are no parallelism at all. - it uses single list to keep encap handlers for both INET and INET6 families. - struct encaptab keeps unneeded information (src, dst, masks, protosw), that isn't used by code in the source tree. - matches are prioritized and when many tunneling interfaces are registered, encapcheck handler of each interface is invoked for each packet. The search takes O(n) for n interfaces. All this work is done with exclusive lock held. What this patch includes: - the datapath is converted to be lockless using epoch(9) KPI. - struct encaptab now linked using CK_LIST. - all unused fields removed from struct encaptab. Several new fields addedr: min_length is the minimum packet length, that encapsulation handler expects to see; exact_match is maximum number of bits, that can return an encapsulation handler, when it wants to consume a packet. - IPv6 and IPv4 handlers are stored in separate lists; - added new "encap_lookup_t" method, that will be used later. It is targeted to speedup lookup of needed interface, when gif(4)/gre(4) have many interfaces. - the need to use protosw structure is eliminated. The only pr_input method was used from this structure, so I don't see the need to keep using it. - encap_input_t method changed to avoid using mbuf tags to store softc pointer. Now it is passed directly trough encap_input_t method. encap_getarg() funtions is removed. - all sockaddr structures and code that uses them removed. We don't have any code in the tree that uses them. All consumers use encap_attach_func() method, that relies on invoking of encapcheck() to determine the needed handler. - introduced struct encap_config, it contains parameters of encap handler that is going to be registered by encap_attach() function. - encap handlers are stored in lists ordered by exact_match value, thus handlers that need more bits to match will be checked first, and if encapcheck method returns exact_match value, the search will be stopped. - all current consumers changed to use new KPI. Reviewed by: mmacy Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D15617
2018-06-05 20:51:01 +00:00
ip_encap_attach(const struct encap_config *cfg, void *arg, int mflags)
{
Rework IP encapsulation handling code. Currently it has several disadvantages: - it uses single mutex to protect internal structures. It is used by data- and control- path, thus there are no parallelism at all. - it uses single list to keep encap handlers for both INET and INET6 families. - struct encaptab keeps unneeded information (src, dst, masks, protosw), that isn't used by code in the source tree. - matches are prioritized and when many tunneling interfaces are registered, encapcheck handler of each interface is invoked for each packet. The search takes O(n) for n interfaces. All this work is done with exclusive lock held. What this patch includes: - the datapath is converted to be lockless using epoch(9) KPI. - struct encaptab now linked using CK_LIST. - all unused fields removed from struct encaptab. Several new fields addedr: min_length is the minimum packet length, that encapsulation handler expects to see; exact_match is maximum number of bits, that can return an encapsulation handler, when it wants to consume a packet. - IPv6 and IPv4 handlers are stored in separate lists; - added new "encap_lookup_t" method, that will be used later. It is targeted to speedup lookup of needed interface, when gif(4)/gre(4) have many interfaces. - the need to use protosw structure is eliminated. The only pr_input method was used from this structure, so I don't see the need to keep using it. - encap_input_t method changed to avoid using mbuf tags to store softc pointer. Now it is passed directly trough encap_input_t method. encap_getarg() funtions is removed. - all sockaddr structures and code that uses them removed. We don't have any code in the tree that uses them. All consumers use encap_attach_func() method, that relies on invoking of encapcheck() to determine the needed handler. - introduced struct encap_config, it contains parameters of encap handler that is going to be registered by encap_attach() function. - encap handlers are stored in lists ordered by exact_match value, thus handlers that need more bits to match will be checked first, and if encapcheck method returns exact_match value, the search will be stopped. - all current consumers changed to use new KPI. Reviewed by: mmacy Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D15617
2018-06-05 20:51:01 +00:00
return (encap_attach(&ipv4_encaptab, cfg, arg, mflags));
}
Rework IP encapsulation handling code. Currently it has several disadvantages: - it uses single mutex to protect internal structures. It is used by data- and control- path, thus there are no parallelism at all. - it uses single list to keep encap handlers for both INET and INET6 families. - struct encaptab keeps unneeded information (src, dst, masks, protosw), that isn't used by code in the source tree. - matches are prioritized and when many tunneling interfaces are registered, encapcheck handler of each interface is invoked for each packet. The search takes O(n) for n interfaces. All this work is done with exclusive lock held. What this patch includes: - the datapath is converted to be lockless using epoch(9) KPI. - struct encaptab now linked using CK_LIST. - all unused fields removed from struct encaptab. Several new fields addedr: min_length is the minimum packet length, that encapsulation handler expects to see; exact_match is maximum number of bits, that can return an encapsulation handler, when it wants to consume a packet. - IPv6 and IPv4 handlers are stored in separate lists; - added new "encap_lookup_t" method, that will be used later. It is targeted to speedup lookup of needed interface, when gif(4)/gre(4) have many interfaces. - the need to use protosw structure is eliminated. The only pr_input method was used from this structure, so I don't see the need to keep using it. - encap_input_t method changed to avoid using mbuf tags to store softc pointer. Now it is passed directly trough encap_input_t method. encap_getarg() funtions is removed. - all sockaddr structures and code that uses them removed. We don't have any code in the tree that uses them. All consumers use encap_attach_func() method, that relies on invoking of encapcheck() to determine the needed handler. - introduced struct encap_config, it contains parameters of encap handler that is going to be registered by encap_attach() function. - encap handlers are stored in lists ordered by exact_match value, thus handlers that need more bits to match will be checked first, and if encapcheck method returns exact_match value, the search will be stopped. - all current consumers changed to use new KPI. Reviewed by: mmacy Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D15617
2018-06-05 20:51:01 +00:00
int
ip_encap_detach(const struct encaptab *cookie)
{
Rework IP encapsulation handling code. Currently it has several disadvantages: - it uses single mutex to protect internal structures. It is used by data- and control- path, thus there are no parallelism at all. - it uses single list to keep encap handlers for both INET and INET6 families. - struct encaptab keeps unneeded information (src, dst, masks, protosw), that isn't used by code in the source tree. - matches are prioritized and when many tunneling interfaces are registered, encapcheck handler of each interface is invoked for each packet. The search takes O(n) for n interfaces. All this work is done with exclusive lock held. What this patch includes: - the datapath is converted to be lockless using epoch(9) KPI. - struct encaptab now linked using CK_LIST. - all unused fields removed from struct encaptab. Several new fields addedr: min_length is the minimum packet length, that encapsulation handler expects to see; exact_match is maximum number of bits, that can return an encapsulation handler, when it wants to consume a packet. - IPv6 and IPv4 handlers are stored in separate lists; - added new "encap_lookup_t" method, that will be used later. It is targeted to speedup lookup of needed interface, when gif(4)/gre(4) have many interfaces. - the need to use protosw structure is eliminated. The only pr_input method was used from this structure, so I don't see the need to keep using it. - encap_input_t method changed to avoid using mbuf tags to store softc pointer. Now it is passed directly trough encap_input_t method. encap_getarg() funtions is removed. - all sockaddr structures and code that uses them removed. We don't have any code in the tree that uses them. All consumers use encap_attach_func() method, that relies on invoking of encapcheck() to determine the needed handler. - introduced struct encap_config, it contains parameters of encap handler that is going to be registered by encap_attach() function. - encap handlers are stored in lists ordered by exact_match value, thus handlers that need more bits to match will be checked first, and if encapcheck method returns exact_match value, the search will be stopped. - all current consumers changed to use new KPI. Reviewed by: mmacy Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D15617
2018-06-05 20:51:01 +00:00
return (encap_detach(&ipv4_encaptab, cookie));
}
int
Rework IP encapsulation handling code. Currently it has several disadvantages: - it uses single mutex to protect internal structures. It is used by data- and control- path, thus there are no parallelism at all. - it uses single list to keep encap handlers for both INET and INET6 families. - struct encaptab keeps unneeded information (src, dst, masks, protosw), that isn't used by code in the source tree. - matches are prioritized and when many tunneling interfaces are registered, encapcheck handler of each interface is invoked for each packet. The search takes O(n) for n interfaces. All this work is done with exclusive lock held. What this patch includes: - the datapath is converted to be lockless using epoch(9) KPI. - struct encaptab now linked using CK_LIST. - all unused fields removed from struct encaptab. Several new fields addedr: min_length is the minimum packet length, that encapsulation handler expects to see; exact_match is maximum number of bits, that can return an encapsulation handler, when it wants to consume a packet. - IPv6 and IPv4 handlers are stored in separate lists; - added new "encap_lookup_t" method, that will be used later. It is targeted to speedup lookup of needed interface, when gif(4)/gre(4) have many interfaces. - the need to use protosw structure is eliminated. The only pr_input method was used from this structure, so I don't see the need to keep using it. - encap_input_t method changed to avoid using mbuf tags to store softc pointer. Now it is passed directly trough encap_input_t method. encap_getarg() funtions is removed. - all sockaddr structures and code that uses them removed. We don't have any code in the tree that uses them. All consumers use encap_attach_func() method, that relies on invoking of encapcheck() to determine the needed handler. - introduced struct encap_config, it contains parameters of encap handler that is going to be registered by encap_attach() function. - encap handlers are stored in lists ordered by exact_match value, thus handlers that need more bits to match will be checked first, and if encapcheck method returns exact_match value, the search will be stopped. - all current consumers changed to use new KPI. Reviewed by: mmacy Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D15617
2018-06-05 20:51:01 +00:00
encap4_input(struct mbuf **mp, int *offp, int proto)
{
Rework IP encapsulation handling code. Currently it has several disadvantages: - it uses single mutex to protect internal structures. It is used by data- and control- path, thus there are no parallelism at all. - it uses single list to keep encap handlers for both INET and INET6 families. - struct encaptab keeps unneeded information (src, dst, masks, protosw), that isn't used by code in the source tree. - matches are prioritized and when many tunneling interfaces are registered, encapcheck handler of each interface is invoked for each packet. The search takes O(n) for n interfaces. All this work is done with exclusive lock held. What this patch includes: - the datapath is converted to be lockless using epoch(9) KPI. - struct encaptab now linked using CK_LIST. - all unused fields removed from struct encaptab. Several new fields addedr: min_length is the minimum packet length, that encapsulation handler expects to see; exact_match is maximum number of bits, that can return an encapsulation handler, when it wants to consume a packet. - IPv6 and IPv4 handlers are stored in separate lists; - added new "encap_lookup_t" method, that will be used later. It is targeted to speedup lookup of needed interface, when gif(4)/gre(4) have many interfaces. - the need to use protosw structure is eliminated. The only pr_input method was used from this structure, so I don't see the need to keep using it. - encap_input_t method changed to avoid using mbuf tags to store softc pointer. Now it is passed directly trough encap_input_t method. encap_getarg() funtions is removed. - all sockaddr structures and code that uses them removed. We don't have any code in the tree that uses them. All consumers use encap_attach_func() method, that relies on invoking of encapcheck() to determine the needed handler. - introduced struct encap_config, it contains parameters of encap handler that is going to be registered by encap_attach() function. - encap handlers are stored in lists ordered by exact_match value, thus handlers that need more bits to match will be checked first, and if encapcheck method returns exact_match value, the search will be stopped. - all current consumers changed to use new KPI. Reviewed by: mmacy Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D15617
2018-06-05 20:51:01 +00:00
if (encap_input(&ipv4_encaptab, *mp, *offp, proto) != IPPROTO_DONE)
return (rip_input(mp, offp, proto));
return (IPPROTO_DONE);
}
Rework IP encapsulation handling code. Currently it has several disadvantages: - it uses single mutex to protect internal structures. It is used by data- and control- path, thus there are no parallelism at all. - it uses single list to keep encap handlers for both INET and INET6 families. - struct encaptab keeps unneeded information (src, dst, masks, protosw), that isn't used by code in the source tree. - matches are prioritized and when many tunneling interfaces are registered, encapcheck handler of each interface is invoked for each packet. The search takes O(n) for n interfaces. All this work is done with exclusive lock held. What this patch includes: - the datapath is converted to be lockless using epoch(9) KPI. - struct encaptab now linked using CK_LIST. - all unused fields removed from struct encaptab. Several new fields addedr: min_length is the minimum packet length, that encapsulation handler expects to see; exact_match is maximum number of bits, that can return an encapsulation handler, when it wants to consume a packet. - IPv6 and IPv4 handlers are stored in separate lists; - added new "encap_lookup_t" method, that will be used later. It is targeted to speedup lookup of needed interface, when gif(4)/gre(4) have many interfaces. - the need to use protosw structure is eliminated. The only pr_input method was used from this structure, so I don't see the need to keep using it. - encap_input_t method changed to avoid using mbuf tags to store softc pointer. Now it is passed directly trough encap_input_t method. encap_getarg() funtions is removed. - all sockaddr structures and code that uses them removed. We don't have any code in the tree that uses them. All consumers use encap_attach_func() method, that relies on invoking of encapcheck() to determine the needed handler. - introduced struct encap_config, it contains parameters of encap handler that is going to be registered by encap_attach() function. - encap handlers are stored in lists ordered by exact_match value, thus handlers that need more bits to match will be checked first, and if encapcheck method returns exact_match value, the search will be stopped. - all current consumers changed to use new KPI. Reviewed by: mmacy Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D15617
2018-06-05 20:51:01 +00:00
#endif /* INET */
Rework IP encapsulation handling code. Currently it has several disadvantages: - it uses single mutex to protect internal structures. It is used by data- and control- path, thus there are no parallelism at all. - it uses single list to keep encap handlers for both INET and INET6 families. - struct encaptab keeps unneeded information (src, dst, masks, protosw), that isn't used by code in the source tree. - matches are prioritized and when many tunneling interfaces are registered, encapcheck handler of each interface is invoked for each packet. The search takes O(n) for n interfaces. All this work is done with exclusive lock held. What this patch includes: - the datapath is converted to be lockless using epoch(9) KPI. - struct encaptab now linked using CK_LIST. - all unused fields removed from struct encaptab. Several new fields addedr: min_length is the minimum packet length, that encapsulation handler expects to see; exact_match is maximum number of bits, that can return an encapsulation handler, when it wants to consume a packet. - IPv6 and IPv4 handlers are stored in separate lists; - added new "encap_lookup_t" method, that will be used later. It is targeted to speedup lookup of needed interface, when gif(4)/gre(4) have many interfaces. - the need to use protosw structure is eliminated. The only pr_input method was used from this structure, so I don't see the need to keep using it. - encap_input_t method changed to avoid using mbuf tags to store softc pointer. Now it is passed directly trough encap_input_t method. encap_getarg() funtions is removed. - all sockaddr structures and code that uses them removed. We don't have any code in the tree that uses them. All consumers use encap_attach_func() method, that relies on invoking of encapcheck() to determine the needed handler. - introduced struct encap_config, it contains parameters of encap handler that is going to be registered by encap_attach() function. - encap handlers are stored in lists ordered by exact_match value, thus handlers that need more bits to match will be checked first, and if encapcheck method returns exact_match value, the search will be stopped. - all current consumers changed to use new KPI. Reviewed by: mmacy Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D15617
2018-06-05 20:51:01 +00:00
#ifdef INET6
const struct srcaddrtab *
ip6_encap_register_srcaddr(encap_srcaddr_t func, void *arg, int mflags)
{
return (encap_register_srcaddr(&ipv6_srcaddrtab, func, arg, mflags));
}
int
ip6_encap_unregister_srcaddr(const struct srcaddrtab *cookie)
{
return (encap_unregister_srcaddr(&ipv6_srcaddrtab, cookie));
}
Rework IP encapsulation handling code. Currently it has several disadvantages: - it uses single mutex to protect internal structures. It is used by data- and control- path, thus there are no parallelism at all. - it uses single list to keep encap handlers for both INET and INET6 families. - struct encaptab keeps unneeded information (src, dst, masks, protosw), that isn't used by code in the source tree. - matches are prioritized and when many tunneling interfaces are registered, encapcheck handler of each interface is invoked for each packet. The search takes O(n) for n interfaces. All this work is done with exclusive lock held. What this patch includes: - the datapath is converted to be lockless using epoch(9) KPI. - struct encaptab now linked using CK_LIST. - all unused fields removed from struct encaptab. Several new fields addedr: min_length is the minimum packet length, that encapsulation handler expects to see; exact_match is maximum number of bits, that can return an encapsulation handler, when it wants to consume a packet. - IPv6 and IPv4 handlers are stored in separate lists; - added new "encap_lookup_t" method, that will be used later. It is targeted to speedup lookup of needed interface, when gif(4)/gre(4) have many interfaces. - the need to use protosw structure is eliminated. The only pr_input method was used from this structure, so I don't see the need to keep using it. - encap_input_t method changed to avoid using mbuf tags to store softc pointer. Now it is passed directly trough encap_input_t method. encap_getarg() funtions is removed. - all sockaddr structures and code that uses them removed. We don't have any code in the tree that uses them. All consumers use encap_attach_func() method, that relies on invoking of encapcheck() to determine the needed handler. - introduced struct encap_config, it contains parameters of encap handler that is going to be registered by encap_attach() function. - encap handlers are stored in lists ordered by exact_match value, thus handlers that need more bits to match will be checked first, and if encapcheck method returns exact_match value, the search will be stopped. - all current consumers changed to use new KPI. Reviewed by: mmacy Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D15617
2018-06-05 20:51:01 +00:00
const struct encaptab *
ip6_encap_attach(const struct encap_config *cfg, void *arg, int mflags)
{
Rework IP encapsulation handling code. Currently it has several disadvantages: - it uses single mutex to protect internal structures. It is used by data- and control- path, thus there are no parallelism at all. - it uses single list to keep encap handlers for both INET and INET6 families. - struct encaptab keeps unneeded information (src, dst, masks, protosw), that isn't used by code in the source tree. - matches are prioritized and when many tunneling interfaces are registered, encapcheck handler of each interface is invoked for each packet. The search takes O(n) for n interfaces. All this work is done with exclusive lock held. What this patch includes: - the datapath is converted to be lockless using epoch(9) KPI. - struct encaptab now linked using CK_LIST. - all unused fields removed from struct encaptab. Several new fields addedr: min_length is the minimum packet length, that encapsulation handler expects to see; exact_match is maximum number of bits, that can return an encapsulation handler, when it wants to consume a packet. - IPv6 and IPv4 handlers are stored in separate lists; - added new "encap_lookup_t" method, that will be used later. It is targeted to speedup lookup of needed interface, when gif(4)/gre(4) have many interfaces. - the need to use protosw structure is eliminated. The only pr_input method was used from this structure, so I don't see the need to keep using it. - encap_input_t method changed to avoid using mbuf tags to store softc pointer. Now it is passed directly trough encap_input_t method. encap_getarg() funtions is removed. - all sockaddr structures and code that uses them removed. We don't have any code in the tree that uses them. All consumers use encap_attach_func() method, that relies on invoking of encapcheck() to determine the needed handler. - introduced struct encap_config, it contains parameters of encap handler that is going to be registered by encap_attach() function. - encap handlers are stored in lists ordered by exact_match value, thus handlers that need more bits to match will be checked first, and if encapcheck method returns exact_match value, the search will be stopped. - all current consumers changed to use new KPI. Reviewed by: mmacy Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D15617
2018-06-05 20:51:01 +00:00
return (encap_attach(&ipv6_encaptab, cfg, arg, mflags));
}
Rework IP encapsulation handling code. Currently it has several disadvantages: - it uses single mutex to protect internal structures. It is used by data- and control- path, thus there are no parallelism at all. - it uses single list to keep encap handlers for both INET and INET6 families. - struct encaptab keeps unneeded information (src, dst, masks, protosw), that isn't used by code in the source tree. - matches are prioritized and when many tunneling interfaces are registered, encapcheck handler of each interface is invoked for each packet. The search takes O(n) for n interfaces. All this work is done with exclusive lock held. What this patch includes: - the datapath is converted to be lockless using epoch(9) KPI. - struct encaptab now linked using CK_LIST. - all unused fields removed from struct encaptab. Several new fields addedr: min_length is the minimum packet length, that encapsulation handler expects to see; exact_match is maximum number of bits, that can return an encapsulation handler, when it wants to consume a packet. - IPv6 and IPv4 handlers are stored in separate lists; - added new "encap_lookup_t" method, that will be used later. It is targeted to speedup lookup of needed interface, when gif(4)/gre(4) have many interfaces. - the need to use protosw structure is eliminated. The only pr_input method was used from this structure, so I don't see the need to keep using it. - encap_input_t method changed to avoid using mbuf tags to store softc pointer. Now it is passed directly trough encap_input_t method. encap_getarg() funtions is removed. - all sockaddr structures and code that uses them removed. We don't have any code in the tree that uses them. All consumers use encap_attach_func() method, that relies on invoking of encapcheck() to determine the needed handler. - introduced struct encap_config, it contains parameters of encap handler that is going to be registered by encap_attach() function. - encap handlers are stored in lists ordered by exact_match value, thus handlers that need more bits to match will be checked first, and if encapcheck method returns exact_match value, the search will be stopped. - all current consumers changed to use new KPI. Reviewed by: mmacy Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D15617
2018-06-05 20:51:01 +00:00
int
ip6_encap_detach(const struct encaptab *cookie)
{
Rework IP encapsulation handling code. Currently it has several disadvantages: - it uses single mutex to protect internal structures. It is used by data- and control- path, thus there are no parallelism at all. - it uses single list to keep encap handlers for both INET and INET6 families. - struct encaptab keeps unneeded information (src, dst, masks, protosw), that isn't used by code in the source tree. - matches are prioritized and when many tunneling interfaces are registered, encapcheck handler of each interface is invoked for each packet. The search takes O(n) for n interfaces. All this work is done with exclusive lock held. What this patch includes: - the datapath is converted to be lockless using epoch(9) KPI. - struct encaptab now linked using CK_LIST. - all unused fields removed from struct encaptab. Several new fields addedr: min_length is the minimum packet length, that encapsulation handler expects to see; exact_match is maximum number of bits, that can return an encapsulation handler, when it wants to consume a packet. - IPv6 and IPv4 handlers are stored in separate lists; - added new "encap_lookup_t" method, that will be used later. It is targeted to speedup lookup of needed interface, when gif(4)/gre(4) have many interfaces. - the need to use protosw structure is eliminated. The only pr_input method was used from this structure, so I don't see the need to keep using it. - encap_input_t method changed to avoid using mbuf tags to store softc pointer. Now it is passed directly trough encap_input_t method. encap_getarg() funtions is removed. - all sockaddr structures and code that uses them removed. We don't have any code in the tree that uses them. All consumers use encap_attach_func() method, that relies on invoking of encapcheck() to determine the needed handler. - introduced struct encap_config, it contains parameters of encap handler that is going to be registered by encap_attach() function. - encap handlers are stored in lists ordered by exact_match value, thus handlers that need more bits to match will be checked first, and if encapcheck method returns exact_match value, the search will be stopped. - all current consumers changed to use new KPI. Reviewed by: mmacy Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D15617
2018-06-05 20:51:01 +00:00
return (encap_detach(&ipv6_encaptab, cookie));
}
Rework IP encapsulation handling code. Currently it has several disadvantages: - it uses single mutex to protect internal structures. It is used by data- and control- path, thus there are no parallelism at all. - it uses single list to keep encap handlers for both INET and INET6 families. - struct encaptab keeps unneeded information (src, dst, masks, protosw), that isn't used by code in the source tree. - matches are prioritized and when many tunneling interfaces are registered, encapcheck handler of each interface is invoked for each packet. The search takes O(n) for n interfaces. All this work is done with exclusive lock held. What this patch includes: - the datapath is converted to be lockless using epoch(9) KPI. - struct encaptab now linked using CK_LIST. - all unused fields removed from struct encaptab. Several new fields addedr: min_length is the minimum packet length, that encapsulation handler expects to see; exact_match is maximum number of bits, that can return an encapsulation handler, when it wants to consume a packet. - IPv6 and IPv4 handlers are stored in separate lists; - added new "encap_lookup_t" method, that will be used later. It is targeted to speedup lookup of needed interface, when gif(4)/gre(4) have many interfaces. - the need to use protosw structure is eliminated. The only pr_input method was used from this structure, so I don't see the need to keep using it. - encap_input_t method changed to avoid using mbuf tags to store softc pointer. Now it is passed directly trough encap_input_t method. encap_getarg() funtions is removed. - all sockaddr structures and code that uses them removed. We don't have any code in the tree that uses them. All consumers use encap_attach_func() method, that relies on invoking of encapcheck() to determine the needed handler. - introduced struct encap_config, it contains parameters of encap handler that is going to be registered by encap_attach() function. - encap handlers are stored in lists ordered by exact_match value, thus handlers that need more bits to match will be checked first, and if encapcheck method returns exact_match value, the search will be stopped. - all current consumers changed to use new KPI. Reviewed by: mmacy Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D15617
2018-06-05 20:51:01 +00:00
int
encap6_input(struct mbuf **mp, int *offp, int proto)
{
Rework IP encapsulation handling code. Currently it has several disadvantages: - it uses single mutex to protect internal structures. It is used by data- and control- path, thus there are no parallelism at all. - it uses single list to keep encap handlers for both INET and INET6 families. - struct encaptab keeps unneeded information (src, dst, masks, protosw), that isn't used by code in the source tree. - matches are prioritized and when many tunneling interfaces are registered, encapcheck handler of each interface is invoked for each packet. The search takes O(n) for n interfaces. All this work is done with exclusive lock held. What this patch includes: - the datapath is converted to be lockless using epoch(9) KPI. - struct encaptab now linked using CK_LIST. - all unused fields removed from struct encaptab. Several new fields addedr: min_length is the minimum packet length, that encapsulation handler expects to see; exact_match is maximum number of bits, that can return an encapsulation handler, when it wants to consume a packet. - IPv6 and IPv4 handlers are stored in separate lists; - added new "encap_lookup_t" method, that will be used later. It is targeted to speedup lookup of needed interface, when gif(4)/gre(4) have many interfaces. - the need to use protosw structure is eliminated. The only pr_input method was used from this structure, so I don't see the need to keep using it. - encap_input_t method changed to avoid using mbuf tags to store softc pointer. Now it is passed directly trough encap_input_t method. encap_getarg() funtions is removed. - all sockaddr structures and code that uses them removed. We don't have any code in the tree that uses them. All consumers use encap_attach_func() method, that relies on invoking of encapcheck() to determine the needed handler. - introduced struct encap_config, it contains parameters of encap handler that is going to be registered by encap_attach() function. - encap handlers are stored in lists ordered by exact_match value, thus handlers that need more bits to match will be checked first, and if encapcheck method returns exact_match value, the search will be stopped. - all current consumers changed to use new KPI. Reviewed by: mmacy Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D15617
2018-06-05 20:51:01 +00:00
if (encap_input(&ipv6_encaptab, *mp, *offp, proto) != IPPROTO_DONE)
return (rip6_input(mp, offp, proto));
return (IPPROTO_DONE);
}
Rework IP encapsulation handling code. Currently it has several disadvantages: - it uses single mutex to protect internal structures. It is used by data- and control- path, thus there are no parallelism at all. - it uses single list to keep encap handlers for both INET and INET6 families. - struct encaptab keeps unneeded information (src, dst, masks, protosw), that isn't used by code in the source tree. - matches are prioritized and when many tunneling interfaces are registered, encapcheck handler of each interface is invoked for each packet. The search takes O(n) for n interfaces. All this work is done with exclusive lock held. What this patch includes: - the datapath is converted to be lockless using epoch(9) KPI. - struct encaptab now linked using CK_LIST. - all unused fields removed from struct encaptab. Several new fields addedr: min_length is the minimum packet length, that encapsulation handler expects to see; exact_match is maximum number of bits, that can return an encapsulation handler, when it wants to consume a packet. - IPv6 and IPv4 handlers are stored in separate lists; - added new "encap_lookup_t" method, that will be used later. It is targeted to speedup lookup of needed interface, when gif(4)/gre(4) have many interfaces. - the need to use protosw structure is eliminated. The only pr_input method was used from this structure, so I don't see the need to keep using it. - encap_input_t method changed to avoid using mbuf tags to store softc pointer. Now it is passed directly trough encap_input_t method. encap_getarg() funtions is removed. - all sockaddr structures and code that uses them removed. We don't have any code in the tree that uses them. All consumers use encap_attach_func() method, that relies on invoking of encapcheck() to determine the needed handler. - introduced struct encap_config, it contains parameters of encap handler that is going to be registered by encap_attach() function. - encap handlers are stored in lists ordered by exact_match value, thus handlers that need more bits to match will be checked first, and if encapcheck method returns exact_match value, the search will be stopped. - all current consumers changed to use new KPI. Reviewed by: mmacy Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D15617
2018-06-05 20:51:01 +00:00
#endif /* INET6 */