d/freebsd-skq
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
b9f5d35a30
freebsd-skq/sys/netinet/if_ether.c

1522 lines
41 KiB
C
Raw Normal View History

/* -> /*- for license, minor formatting changes
2005-01-07 01:45:51 +00:00
/*-
sys: further adoption of SPDX licensing ID tags. Mainly focus on files that use BSD 3-Clause license. The Software Package Data Exchange (SPDX) group provides a specification to make it easier for automated tools to detect and summarize well known opensource licenses. We are gradually adopting the specification, noting that the tags are considered only advisory and do not, in any way, superceed or replace the license texts. Special thanks to Wind River for providing access to "The Duke of Highlander" tool: an older (2014) run over FreeBSD tree was useful as a starting point.
2017-11-20 19:43:44 +00:00
* SPDX-License-Identifier: BSD-3-Clause
*
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
* Copyright (c) 1982, 1986, 1988, 1993
* The Regents of the University of California. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
Renumber copyright clause 4 Renumber cluase 4 to 3, per what everybody else did when BSD granted them permission to remove clause 3. My insistance on keeping the same numbering for legal reasons is too pedantic, so give up on that point. Submitted by: Jan Schaumann <jschauma@stevens.edu> Pull Request: https://github.com/freebsd/freebsd/pull/96
2017-02-28 23:42:47 +00:00
* 3. Neither the name of the University nor the names of its contributors
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
* may be used to endorse or promote products derived from this software
* without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*
* @(#)if_ether.c 8.1 (Berkeley) 6/10/93
*/
/*
* Ethernet address resolution protocol.
* TODO:
* add "inuse/lock" bit (or ref. count) along with valid bit
*/
Add FBSDID to all files in netinet so that people can more easily include file version information in bug reports. Approved by: re (kensmith)
2007-10-07 20:44:24 +00:00
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
Make INET a proper option. This will not make any of object files that LINT create change; there might be differences with INET disabled, but hardly anything compiled before without INET anyway. Now the 'obvious' things will give a proper error if compiled without inet - ipx_ip, ipfw, tcp_debug. The only thing that _should_ work (but can't be made to compile reasonably easily) is sppp :-( This commit move struct arpcom from <netinet/if_ether.h> to <net/if_arp.h>.
1998-01-08 23:42:31 +00:00
#include "opt_inet.h"
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
#include <sys/param.h>
Extract eventfilter declarations to sys/_eventfilter.h This allows replacing "sys/eventfilter.h" includes with "sys/_eventfilter.h" in other header files (e.g., sys/{bus,conf,cpu}.h) and reduces header pollution substantially. EVENTHANDLER_DECLARE and EVENTHANDLER_LIST_DECLAREs were moved out of .c files into appropriate headers (e.g., sys/proc.h, powernv/opal.h). As a side effect of reduced header pollution, many .c files and headers no longer contain needed definitions. The remainder of the patch addresses adding appropriate includes to fix those files. LOCK_DEBUG and LOCK_FILE_LINE_ARG are moved to sys/_lock.h, as required by sys/mutex.h since r326106 (but silently protected by header pollution prior to this change). No functional change (intended). Of course, any out of tree modules that relied on header pollution for sys/eventhandler.h, sys/lock.h, or sys/mutex.h inclusion need to be fixed. __FreeBSD_version has been bumped.
2019-05-20 00:38:23 +00:00
#include <sys/eventhandler.h>
Remove old ballast, clean up a little bit, staticize. Add five sysctl variables that you should probably never tweak. net.arp.t_prune: 300 net.arp.t_keep: 1200 net.arp.t_down: 20 net.arp.maxtries: 5 net.arp.useloopback: 1 net.arp.proxyall: 0 (It's net.arp because arp isn't limited to inet, though our present implementation surely is).
1995-12-09 16:06:54 +00:00
#include <sys/kernel.h>
Convert in_ifaddr_lock and in6_ifaddr_lock to rmlock. Both are used to protect access to IP addresses lists and they can be acquired for reading several times per packet. To reduce lock contention it is better to use rmlock here. Reviewed by: gnn (previous version) Obtained from: Yandex LLC Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D3149
2015-07-29 08:12:05 +00:00
#include <sys/lock.h>
* Change sysctl from using linker_set to construct its tree using SLISTs. This makes it possible to change the sysctl tree at runtime. * Change KLD to find and register any sysctl nodes contained in the loaded file and to unregister them when the file is unloaded. Reviewed by: Archie Cobbs <archie@whistle.com>, Peter Wemm <peter@netplex.com.au> (well they looked at it anyway)
1999-02-16 10:49:55 +00:00
#include <sys/queue.h>
Remove old ballast, clean up a little bit, staticize. Add five sysctl variables that you should probably never tweak. net.arp.t_prune: 300 net.arp.t_keep: 1200 net.arp.t_down: 20 net.arp.maxtries: 5 net.arp.useloopback: 1 net.arp.proxyall: 0 (It's net.arp because arp isn't limited to inet, though our present implementation surely is).
1995-12-09 16:06:54 +00:00
#include <sys/sysctl.h>
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
#include <sys/systm.h>
#include <sys/mbuf.h>
Remove old ballast, clean up a little bit, staticize. Add five sysctl variables that you should probably never tweak. net.arp.t_prune: 300 net.arp.t_keep: 1200 net.arp.t_down: 20 net.arp.maxtries: 5 net.arp.useloopback: 1 net.arp.proxyall: 0 (It's net.arp because arp isn't limited to inet, though our present implementation surely is).
1995-12-09 16:06:54 +00:00
#include <sys/malloc.h>
oops commit the version that compiles
2008-09-14 08:24:45 +00:00
#include <sys/proc.h>
Convert in_ifaddr_lock and in6_ifaddr_lock to rmlock. Both are used to protect access to IP addresses lists and they can be acquired for reading several times per packet. To reduce lock contention it is better to use rmlock here. Reviewed by: gnn (previous version) Obtained from: Yandex LLC Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D3149
2015-07-29 08:12:05 +00:00
#include <sys/rmlock.h>
Removed nested include if <sys/socket.h> from <net/if.h> and <net/if_arp.h> and fixed the things that depended on it. The nested include just allowed unportable programs to compile and made my simple #include checking program report that networking code doesn't need to include <sys/socket.h>.
1996-10-12 19:49:43 +00:00
#include <sys/socket.h>
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
#include <sys/syslog.h>
#include <net/if.h>
The r48589 promised to remove implicit inclusion of if_var.h soon. Prepare to this event, adding if_var.h to files that do need it. Also, include all includes that now are included due to implicit pollution via if_var.h Sponsored by: Netflix Sponsored by: Nginx, Inc.
2013-10-26 17:58:36 +00:00
#include <net/if_var.h>
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
#include <net/if_dl.h>
World, I'd like you to meet the first FreeBSD token Ring driver. This is for various Olicom cards. An IBM driver is following. This patch also adds support to tcpdump to decode packets on tokenring. Congratulations to the proud father.. (below) Submitted by: Larry Lile <lile@stdio.com>
1999-02-20 11:18:00 +00:00
#include <net/if_types.h>
Make networking domains drop-ins, through the magic of GNU ld. (Some day, there may even be LKMs.) Also, change the internal name of `unixdomain' to `localdomain' since AF_LOCAL is now the preferred name of this family. Declare netisr correctly and in the right place.
1995-05-11 00:13:26 +00:00
#include <net/netisr.h>
Make the code act the same in the case of BRIDGE being defined, but not turned on, and the case of it not being defined at all. i.e. Disabling bridging re-enables some of the checks it disables. Submitted by: "Rogier R. Mulhuijzen" <drwilco@drwilco.net>
2001-02-03 17:25:21 +00:00
#include <net/ethernet.h>
After cleaning up rt_tables from vnet.h and cleaning up opt_route.h a lot of files no longer need route.h either. Garbage collect them. While here remove now unneeded vnet.h #includes as well.
2009-06-23 17:03:45 +00:00
#include <net/route.h>
Merge the remainder of kern_vimage.c and vimage.h into vnet.c and vnet.h, we now use jails (rather than vimages) as the abstraction for virtualization management, and what remained was specific to virtual network stacks. Minor cleanups are done in the process, and comments updated to reflect these changes. Reviewed by: bz Approved by: re (vimage blanket)
2009-08-01 19:26:27 +00:00
#include <net/vnet.h>
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
#include <netinet/in.h>
Make in_arpinput(), inp_lookup_mcast_ifp(), icmp_reflect(), ip_dooptions(), icmp6_redirect_input(), in6_lltable_rtcheck(), in6p_lookup_mcast_ifp() and in6_selecthlim() use new routing api. Eliminate now-unused ip_rtaddr(). Fix lookup key fib6_lookup_nh_basic() which was lost diring merge. Make fib6_lookup_nh_basic() and fib6_lookup_nh_extended() always return IPv6 destination address with embedded scope. Currently rw_gateway has it scope embedded, do the same for non-gatewayed destinations. Sponsored by: Yandex LLC
2015-12-09 11:14:27 +00:00
#include <netinet/in_fib.h>
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
#include <netinet/in_var.h>
This main goals of this project are: 1. separating L2 tables (ARP, NDP) from the L3 routing tables 2. removing as much locking dependencies among these layers as possible to allow for some parallelism in the search operations 3. simplify the logic in the routing code, The most notable end result is the obsolescent of the route cloning (RTF_CLONING) concept, which translated into code reduction in both IPv4 ARP and IPv6 NDP related modules, and size reduction in struct rtentry{}. The change in design obsoletes the semantics of RTF_CLONING, RTF_WASCLONE and RTF_LLINFO routing flags. The userland applications such as "arp" and "ndp" have been modified to reflect those changes. The output from "netstat -r" shows only the routing entries. Quite a few developers have contributed to this project in the past: Glebius Smirnoff, Luigi Rizzo, Alessandro Cerri, and Andre Oppermann. And most recently: - Kip Macy revised the locking code completely, thus completing the last piece of the puzzle, Kip has also been conducting active functional testing - Sam Leffler has helped me improving/refactoring the code, and provided valuable reviews - Julian Elischer setup the perforce tree for me and has helped me maintaining that branch before the svn conversion
2008-12-15 06:10:57 +00:00
#include <net/if_llatbl.h>
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
#include <netinet/if_ether.h>
Clean up some #endif comments removing from short sections. Add #endif comments to longer, also refining strange ones. Properly use #ifdef rather than #if defined() where possible. Four #if defined(PCBGROUP) occurances (netinet and netinet6) were ignored to avoid conflicts with eventually upcoming changes for RSS. Reported by: bde (most) Reviewed by: bde MFC after: 3 days
2012-01-22 02:13:19 +00:00
#ifdef INET
Unbreak LINT by moving all carp hooks to net/if.c / netinet/ip_carp.h, with the appropriate ifdefs. Reviewed by: bz Approved by: ken (mentor)
2010-08-11 20:18:19 +00:00
#include <netinet/ip_carp.h>
#endif
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
Complete break-out of sys/sys/mac.h into sys/security/mac/mac_framework.h begun with a repo-copy of mac.h to mac_framework.h. sys/mac.h now contains the userspace and user<->kernel API and definitions, with all in-kernel interfaces moved to mac_framework.h, which is now included across most of the kernel instead. This change is the first step in a larger cleanup and sweep of MAC Framework interfaces in the kernel, and will not be MFC'd. Obtained from: TrustedBSD Project Sponsored by: SPARTA
2006-10-22 11:52:19 +00:00
#include <security/mac/mac_framework.h>
Add const qualifier to the dst parameter of the ifnet if_output method.
2013-04-26 12:50:32 +00:00
#define SIN(s) ((const struct sockaddr_in *)(s))
* Improve logging invalid arp messages * Remove redundant check in ip_arpinput Suggested by: glebius MFC after: 2 weeks
2015-09-15 08:50:44 +00:00
static struct timeval arp_lastlog;
static int arp_curpps;
static int arp_maxpps = 1;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
Remove LLE read lock from IPv4 fast path. LLE structure is mostly unchanged during its lifecycle. To be more specific, there are 2 things relevant for fast path lookup code: 1) link-level address change. Since r286722, these updates are performed under AFDATA WLOCK. 2) Some sort of feedback indicating that this particular entry is used so we re-send arp request to perform reachability verification instead of expiring entry. The only signal that is needed from fast path is something like binary yes/no. The latter is solved by the following changes: 1) introduce special r_skip_req field which is read lockless by fast path, but updated under (new) req_mutex mutex. If this field is non-zero, then fast path will acquire lock and set it back to 0. 2) introduce simple state machine: incomplete->reachable<->verify->deleted. Before that we implicitely had incomplete->reachable->deleted state machine, with V_arpt_keep between "reachable" and "deleted". Verification was performed in runtime 5 seconds before V_arpt_keep expire. This is changed to "change state to verify 5 seconds before V_arpt_keep, set r_skip_req to non-zero value and check it every second". If the value is zero - then send arp verification probe. These changes do not introduce any signifficant control plane overhead: typically lle callout timer would fire 1 time more each V_arpt_keep (1200s) for used lles and up to arp_maxtries (5) for dead lles. As a result, all packets towards "reachable" lle are handled by fast path without acquiring lle read lock. Additional "req_mutex" is needed because callout / arpresolve_slow() or eventhandler might keep LLE lock for signifficant amount of time, which might not be feasible for fast path locking (e.g. having rmlock as ether AFDATA or lltable own lock). Differential Revision: https://reviews.freebsd.org/D3688
2015-12-05 09:50:37 +00:00
/* Simple ARP state machine */
enum arp_llinfo_state {
ARP_LLINFO_INCOMPLETE = 0, /* No LLE data */
ARP_LLINFO_REACHABLE, /* LLE is valid */
ARP_LLINFO_VERIFY, /* LLE is valid, need refresh */
ARP_LLINFO_DELETED, /* LLE is deleted */
};
* Change sysctl from using linker_set to construct its tree using SLISTs. This makes it possible to change the sysctl tree at runtime. * Change KLD to find and register any sysctl nodes contained in the loaded file and to unregister them when the file is unloaded. Reviewed by: Archie Cobbs <archie@whistle.com>, Peter Wemm <peter@netplex.com.au> (well they looked at it anyway)
1999-02-16 10:49:55 +00:00
SYSCTL_DECL(_net_link_ether);
Mark more nodes as CTLFLAG_MPSAFE or CTLFLAG_NEEDGIANT (17 of many) r357614 added CTLFLAG_NEEDGIANT to make it easier to find nodes that are still not MPSAFE (or already are but aren’t properly marked). Use it in preparation for a general review of all nodes. This is non-functional change that adds annotations to SYSCTL_NODE and SYSCTL_PROC nodes using one of the soon-to-be-required flags. Mark all obvious cases as MPSAFE. All entries that haven't been marked as MPSAFE before are by default marked as NEEDGIANT Approved by: kib (mentor, blanket) Commented by: kib, gallatin, melifaro Differential Revision: https://reviews.freebsd.org/D23718
2020-02-26 14:26:36 +00:00
static SYSCTL_NODE(_net_link_ether, PF_INET, inet,
CTLFLAG_RW | CTLFLAG_MPSAFE, 0,
"");
static SYSCTL_NODE(_net_link_ether, PF_ARP, arp,
CTLFLAG_RW | CTLFLAG_MPSAFE, 0,
"");
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
/* timer values */
Use the new VNET_DEFINE_STATIC macro when we are defining static VNET variables. Reviewed by: bz Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D16147
2018-07-24 16:35:52 +00:00
VNET_DEFINE_STATIC(int, arpt_keep) = (20*60); /* once resolved, good for 20
Build on Jeff Roberson's linker-set based dynamic per-CPU allocator (DPCPU), as suggested by Peter Wemm, and implement a new per-virtual network stack memory allocator. Modify vnet to use the allocator instead of monolithic global container structures (vinet, ...). This change solves many binary compatibility problems associated with VIMAGE, and restores ELF symbols for virtualized global variables. Each virtualized global variable exists as a "reference copy", and also once per virtual network stack. Virtualized global variables are tagged at compile-time, placing the in a special linker set, which is loaded into a contiguous region of kernel memory. Virtualized global variables in the base kernel are linked as normal, but those in modules are copied and relocated to a reserved portion of the kernel's vnet region with the help of a the kernel linker. Virtualized global variables exist in per-vnet memory set up when the network stack instance is created, and are initialized statically from the reference copy. Run-time access occurs via an accessor macro, which converts from the current vnet and requested symbol to a per-vnet address. When "options VIMAGE" is not compiled into the kernel, normal global ELF symbols will be used instead and indirection is avoided. This change restores static initialization for network stack global variables, restores support for non-global symbols and types, eliminates the need for many subsystem constructors, eliminates large per-subsystem structures that caused many binary compatibility issues both for monitoring applications (netstat) and kernel modules, removes the per-function INIT_VNET_*() macros throughout the stack, eliminates the need for vnet_symmap ksym(2) munging, and eliminates duplicate definitions of virtualized globals under VIMAGE_GLOBALS. Bump __FreeBSD_version and update UPDATING. Portions submitted by: bz Reviewed by: bz, zec Discussed with: gnn, jamie, jeff, jhb, julian, sam Suggested by: peter Approved by: re (kensmith)
2009-07-14 22:48:30 +00:00
* minutes */
Use the new VNET_DEFINE_STATIC macro when we are defining static VNET variables. Reviewed by: bz Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D16147
2018-07-24 16:35:52 +00:00
VNET_DEFINE_STATIC(int, arp_maxtries) = 5;
VNET_DEFINE_STATIC(int, arp_proxyall) = 0;
VNET_DEFINE_STATIC(int, arpt_down) = 20; /* keep incomplete entries for
Some style(9) and whitespace changes. Together with: Andrey Zonov <andrey zonov.org>
2012-07-31 11:31:12 +00:00
* 20 seconds */
Use the new VNET_DEFINE_STATIC macro when we are defining static VNET variables. Reviewed by: bz Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D16147
2018-07-24 16:35:52 +00:00
VNET_DEFINE_STATIC(int, arpt_rexmit) = 1; /* retransmit arp entries, sec*/
Migrate structs arpstat, icmpstat, mrtstat, pimstat and udpstat to PCPU counters.
2013-07-09 09:50:15 +00:00
VNET_PCPUSTAT_DEFINE(struct arpstat, arpstat); /* ARP statistics, see if_arp.h */
VNET_PCPUSTAT_SYSINIT(arpstat);
#ifdef VIMAGE
VNET_PCPUSTAT_SYSUNINIT(arpstat);
#endif /* VIMAGE */
Add a queue to hold packets while we await an ARP reply. When a fast machine first brings up some non TCP networking program it is quite possible that we will drop packets due to the fact that only one packet can be held per ARP entry. This leads to packets being missed when a program starts or restarts if the ARP data is not currently in the ARP cache. This code adds a new sysctl, net.link.ether.inet.maxhold, which defines a system wide maximum number of packets to be held in each ARP entry. Up to maxhold packets are queued until an ARP reply is received or the ARP times out. The default setting is the old value of 1 which has been part of the BSD networking code since time immemorial. Expose the time we hold an incomplete ARP entry by adding the sysctl net.link.ether.inet.wait, which defaults to 20 seconds, the value used when the new ARP code was added.. Reviewed by: bz, rpaulo MFC after: 3 weeks
2010-11-12 22:03:02 +00:00
Use the new VNET_DEFINE_STATIC macro when we are defining static VNET variables. Reviewed by: bz Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D16147
2018-07-24 16:35:52 +00:00
VNET_DEFINE_STATIC(int, arp_maxhold) = 1;
Build on Jeff Roberson's linker-set based dynamic per-CPU allocator (DPCPU), as suggested by Peter Wemm, and implement a new per-virtual network stack memory allocator. Modify vnet to use the allocator instead of monolithic global container structures (vinet, ...). This change solves many binary compatibility problems associated with VIMAGE, and restores ELF symbols for virtualized global variables. Each virtualized global variable exists as a "reference copy", and also once per virtual network stack. Virtualized global variables are tagged at compile-time, placing the in a special linker set, which is loaded into a contiguous region of kernel memory. Virtualized global variables in the base kernel are linked as normal, but those in modules are copied and relocated to a reserved portion of the kernel's vnet region with the help of a the kernel linker. Virtualized global variables exist in per-vnet memory set up when the network stack instance is created, and are initialized statically from the reference copy. Run-time access occurs via an accessor macro, which converts from the current vnet and requested symbol to a per-vnet address. When "options VIMAGE" is not compiled into the kernel, normal global ELF symbols will be used instead and indirection is avoided. This change restores static initialization for network stack global variables, restores support for non-global symbols and types, eliminates the need for many subsystem constructors, eliminates large per-subsystem structures that caused many binary compatibility issues both for monitoring applications (netstat) and kernel modules, removes the per-function INIT_VNET_*() macros throughout the stack, eliminates the need for vnet_symmap ksym(2) munging, and eliminates duplicate definitions of virtualized globals under VIMAGE_GLOBALS. Bump __FreeBSD_version and update UPDATING. Portions submitted by: bz Reviewed by: bz, zec Discussed with: gnn, jamie, jeff, jhb, julian, sam Suggested by: peter Approved by: re (kensmith)
2009-07-14 22:48:30 +00:00
Remove unused VNET_SET() and related macros; only VNET_GET() is ever actually used. Rename VNET_GET() to VNET() to shorten variable references. Discussed with: bz, julian Reviewed by: bz Approved by: re (kensmith, kib)
2009-07-16 21:13:04 +00:00
#define V_arpt_keep VNET(arpt_keep)
This patch fixes the following issues in the ARP operation: 1. There is a regression issue in the ARP code. The incomplete ARP entry was timing out too quickly (1 second timeout), as such, a new entry is created each time arpresolve() is called. Therefore the maximum attempts made is always 1. Consequently the error code returned to the application is always 0. 2. Set the expiration of each incomplete entry to a 20-second lifetime. 3. Return "incomplete" entries to the application. Reviewed by: kmacy MFC after: 3 days
2009-10-15 06:12:04 +00:00
#define V_arpt_down VNET(arpt_down)
Remove LLE read lock from IPv4 fast path. LLE structure is mostly unchanged during its lifecycle. To be more specific, there are 2 things relevant for fast path lookup code: 1) link-level address change. Since r286722, these updates are performed under AFDATA WLOCK. 2) Some sort of feedback indicating that this particular entry is used so we re-send arp request to perform reachability verification instead of expiring entry. The only signal that is needed from fast path is something like binary yes/no. The latter is solved by the following changes: 1) introduce special r_skip_req field which is read lockless by fast path, but updated under (new) req_mutex mutex. If this field is non-zero, then fast path will acquire lock and set it back to 0. 2) introduce simple state machine: incomplete->reachable<->verify->deleted. Before that we implicitely had incomplete->reachable->deleted state machine, with V_arpt_keep between "reachable" and "deleted". Verification was performed in runtime 5 seconds before V_arpt_keep expire. This is changed to "change state to verify 5 seconds before V_arpt_keep, set r_skip_req to non-zero value and check it every second". If the value is zero - then send arp verification probe. These changes do not introduce any signifficant control plane overhead: typically lle callout timer would fire 1 time more each V_arpt_keep (1200s) for used lles and up to arp_maxtries (5) for dead lles. As a result, all packets towards "reachable" lle are handled by fast path without acquiring lle read lock. Additional "req_mutex" is needed because callout / arpresolve_slow() or eventhandler might keep LLE lock for signifficant amount of time, which might not be feasible for fast path locking (e.g. having rmlock as ether AFDATA or lltable own lock). Differential Revision: https://reviews.freebsd.org/D3688
2015-12-05 09:50:37 +00:00
#define V_arpt_rexmit VNET(arpt_rexmit)
Remove unused VNET_SET() and related macros; only VNET_GET() is ever actually used. Rename VNET_GET() to VNET() to shorten variable references. Discussed with: bz, julian Reviewed by: bz Approved by: re (kensmith, kib)
2009-07-16 21:13:04 +00:00
#define V_arp_maxtries VNET(arp_maxtries)
#define V_arp_proxyall VNET(arp_proxyall)
Add a queue to hold packets while we await an ARP reply. When a fast machine first brings up some non TCP networking program it is quite possible that we will drop packets due to the fact that only one packet can be held per ARP entry. This leads to packets being missed when a program starts or restarts if the ARP data is not currently in the ARP cache. This code adds a new sysctl, net.link.ether.inet.maxhold, which defines a system wide maximum number of packets to be held in each ARP entry. Up to maxhold packets are queued until an ARP reply is received or the ARP times out. The default setting is the old value of 1 which has been part of the BSD networking code since time immemorial. Expose the time we hold an incomplete ARP entry by adding the sysctl net.link.ether.inet.wait, which defaults to 20 seconds, the value used when the new ARP code was added.. Reviewed by: bz, rpaulo MFC after: 3 weeks
2010-11-12 22:03:02 +00:00
#define V_arp_maxhold VNET(arp_maxhold)
Build on Jeff Roberson's linker-set based dynamic per-CPU allocator (DPCPU), as suggested by Peter Wemm, and implement a new per-virtual network stack memory allocator. Modify vnet to use the allocator instead of monolithic global container structures (vinet, ...). This change solves many binary compatibility problems associated with VIMAGE, and restores ELF symbols for virtualized global variables. Each virtualized global variable exists as a "reference copy", and also once per virtual network stack. Virtualized global variables are tagged at compile-time, placing the in a special linker set, which is loaded into a contiguous region of kernel memory. Virtualized global variables in the base kernel are linked as normal, but those in modules are copied and relocated to a reserved portion of the kernel's vnet region with the help of a the kernel linker. Virtualized global variables exist in per-vnet memory set up when the network stack instance is created, and are initialized statically from the reference copy. Run-time access occurs via an accessor macro, which converts from the current vnet and requested symbol to a per-vnet address. When "options VIMAGE" is not compiled into the kernel, normal global ELF symbols will be used instead and indirection is avoided. This change restores static initialization for network stack global variables, restores support for non-global symbols and types, eliminates the need for many subsystem constructors, eliminates large per-subsystem structures that caused many binary compatibility issues both for monitoring applications (netstat) and kernel modules, removes the per-function INIT_VNET_*() macros throughout the stack, eliminates the need for vnet_symmap ksym(2) munging, and eliminates duplicate definitions of virtualized globals under VIMAGE_GLOBALS. Bump __FreeBSD_version and update UPDATING. Portions submitted by: bz Reviewed by: bz, zec Discussed with: gnn, jamie, jeff, jhb, julian, sam Suggested by: peter Approved by: re (kensmith)
2009-07-14 22:48:30 +00:00
Remove SYSCTL_VNET_* macros, and simply put CTLFLAG_VNET where needed. Sponsored by: Nginx, Inc.
2014-11-07 09:39:05 +00:00
SYSCTL_INT(_net_link_ether_inet, OID_AUTO, max_age, CTLFLAG_VNET | CTLFLAG_RW,
Build on Jeff Roberson's linker-set based dynamic per-CPU allocator (DPCPU), as suggested by Peter Wemm, and implement a new per-virtual network stack memory allocator. Modify vnet to use the allocator instead of monolithic global container structures (vinet, ...). This change solves many binary compatibility problems associated with VIMAGE, and restores ELF symbols for virtualized global variables. Each virtualized global variable exists as a "reference copy", and also once per virtual network stack. Virtualized global variables are tagged at compile-time, placing the in a special linker set, which is loaded into a contiguous region of kernel memory. Virtualized global variables in the base kernel are linked as normal, but those in modules are copied and relocated to a reserved portion of the kernel's vnet region with the help of a the kernel linker. Virtualized global variables exist in per-vnet memory set up when the network stack instance is created, and are initialized statically from the reference copy. Run-time access occurs via an accessor macro, which converts from the current vnet and requested symbol to a per-vnet address. When "options VIMAGE" is not compiled into the kernel, normal global ELF symbols will be used instead and indirection is avoided. This change restores static initialization for network stack global variables, restores support for non-global symbols and types, eliminates the need for many subsystem constructors, eliminates large per-subsystem structures that caused many binary compatibility issues both for monitoring applications (netstat) and kernel modules, removes the per-function INIT_VNET_*() macros throughout the stack, eliminates the need for vnet_symmap ksym(2) munging, and eliminates duplicate definitions of virtualized globals under VIMAGE_GLOBALS. Bump __FreeBSD_version and update UPDATING. Portions submitted by: bz Reviewed by: bz, zec Discussed with: gnn, jamie, jeff, jhb, julian, sam Suggested by: peter Approved by: re (kensmith)
2009-07-14 22:48:30 +00:00
&VNET_NAME(arpt_keep), 0,
"ARP entry lifetime in seconds");
Remove SYSCTL_VNET_* macros, and simply put CTLFLAG_VNET where needed. Sponsored by: Nginx, Inc.
2014-11-07 09:39:05 +00:00
SYSCTL_INT(_net_link_ether_inet, OID_AUTO, maxtries, CTLFLAG_VNET | CTLFLAG_RW,
Build on Jeff Roberson's linker-set based dynamic per-CPU allocator (DPCPU), as suggested by Peter Wemm, and implement a new per-virtual network stack memory allocator. Modify vnet to use the allocator instead of monolithic global container structures (vinet, ...). This change solves many binary compatibility problems associated with VIMAGE, and restores ELF symbols for virtualized global variables. Each virtualized global variable exists as a "reference copy", and also once per virtual network stack. Virtualized global variables are tagged at compile-time, placing the in a special linker set, which is loaded into a contiguous region of kernel memory. Virtualized global variables in the base kernel are linked as normal, but those in modules are copied and relocated to a reserved portion of the kernel's vnet region with the help of a the kernel linker. Virtualized global variables exist in per-vnet memory set up when the network stack instance is created, and are initialized statically from the reference copy. Run-time access occurs via an accessor macro, which converts from the current vnet and requested symbol to a per-vnet address. When "options VIMAGE" is not compiled into the kernel, normal global ELF symbols will be used instead and indirection is avoided. This change restores static initialization for network stack global variables, restores support for non-global symbols and types, eliminates the need for many subsystem constructors, eliminates large per-subsystem structures that caused many binary compatibility issues both for monitoring applications (netstat) and kernel modules, removes the per-function INIT_VNET_*() macros throughout the stack, eliminates the need for vnet_symmap ksym(2) munging, and eliminates duplicate definitions of virtualized globals under VIMAGE_GLOBALS. Bump __FreeBSD_version and update UPDATING. Portions submitted by: bz Reviewed by: bz, zec Discussed with: gnn, jamie, jeff, jhb, julian, sam Suggested by: peter Approved by: re (kensmith)
2009-07-14 22:48:30 +00:00
&VNET_NAME(arp_maxtries), 0,
Step 1.5 of importing the network stack virtualization infrastructure from the vimage project, as per plan established at devsummit 08/08: http://wiki.freebsd.org/Image/Notes200808DevSummit Introduce INIT_VNET_*() initializer macros, VNET_FOREACH() iterator macros, and CURVNET_SET() context setting macros, all currently resolving to NOPs. Prepare for virtualization of selected SYSCTL objects by introducing a family of SYSCTL_V_*() macros, currently resolving to their global counterparts, i.e. SYSCTL_V_INT() == SYSCTL_INT(). Move selected #defines from sys/sys/vimage.h to newly introduced header files specific to virtualized subsystems (sys/net/vnet.h, sys/netinet/vinet.h etc.). All the changes are verified to have zero functional impact at this point in time by doing MD5 comparision between pre- and post-change object files(*). (*) netipsec/keysock.c did not validate depending on compile time options. Implemented by: julian, bz, brooks, zec Reviewed by: julian, bz, brooks, kris, rwatson, ... Approved by: julian (mentor) Obtained from: //depot/projects/vimage-commit2/... X-MFC after: never Sponsored by: NLnet Foundation, The FreeBSD Foundation
2008-10-02 15:37:58 +00:00
"ARP resolution attempts before returning error");
Remove SYSCTL_VNET_* macros, and simply put CTLFLAG_VNET where needed. Sponsored by: Nginx, Inc.
2014-11-07 09:39:05 +00:00
SYSCTL_INT(_net_link_ether_inet, OID_AUTO, proxyall, CTLFLAG_VNET | CTLFLAG_RW,
Build on Jeff Roberson's linker-set based dynamic per-CPU allocator (DPCPU), as suggested by Peter Wemm, and implement a new per-virtual network stack memory allocator. Modify vnet to use the allocator instead of monolithic global container structures (vinet, ...). This change solves many binary compatibility problems associated with VIMAGE, and restores ELF symbols for virtualized global variables. Each virtualized global variable exists as a "reference copy", and also once per virtual network stack. Virtualized global variables are tagged at compile-time, placing the in a special linker set, which is loaded into a contiguous region of kernel memory. Virtualized global variables in the base kernel are linked as normal, but those in modules are copied and relocated to a reserved portion of the kernel's vnet region with the help of a the kernel linker. Virtualized global variables exist in per-vnet memory set up when the network stack instance is created, and are initialized statically from the reference copy. Run-time access occurs via an accessor macro, which converts from the current vnet and requested symbol to a per-vnet address. When "options VIMAGE" is not compiled into the kernel, normal global ELF symbols will be used instead and indirection is avoided. This change restores static initialization for network stack global variables, restores support for non-global symbols and types, eliminates the need for many subsystem constructors, eliminates large per-subsystem structures that caused many binary compatibility issues both for monitoring applications (netstat) and kernel modules, removes the per-function INIT_VNET_*() macros throughout the stack, eliminates the need for vnet_symmap ksym(2) munging, and eliminates duplicate definitions of virtualized globals under VIMAGE_GLOBALS. Bump __FreeBSD_version and update UPDATING. Portions submitted by: bz Reviewed by: bz, zec Discussed with: gnn, jamie, jeff, jhb, julian, sam Suggested by: peter Approved by: re (kensmith)
2009-07-14 22:48:30 +00:00
&VNET_NAME(arp_proxyall), 0,
Step 1.5 of importing the network stack virtualization infrastructure from the vimage project, as per plan established at devsummit 08/08: http://wiki.freebsd.org/Image/Notes200808DevSummit Introduce INIT_VNET_*() initializer macros, VNET_FOREACH() iterator macros, and CURVNET_SET() context setting macros, all currently resolving to NOPs. Prepare for virtualization of selected SYSCTL objects by introducing a family of SYSCTL_V_*() macros, currently resolving to their global counterparts, i.e. SYSCTL_V_INT() == SYSCTL_INT(). Move selected #defines from sys/sys/vimage.h to newly introduced header files specific to virtualized subsystems (sys/net/vnet.h, sys/netinet/vinet.h etc.). All the changes are verified to have zero functional impact at this point in time by doing MD5 comparision between pre- and post-change object files(*). (*) netipsec/keysock.c did not validate depending on compile time options. Implemented by: julian, bz, brooks, zec Reviewed by: julian, bz, brooks, kris, rwatson, ... Approved by: julian (mentor) Obtained from: //depot/projects/vimage-commit2/... X-MFC after: never Sponsored by: NLnet Foundation, The FreeBSD Foundation
2008-10-02 15:37:58 +00:00
"Enable proxy ARP for all suitable requests");
Remove SYSCTL_VNET_* macros, and simply put CTLFLAG_VNET where needed. Sponsored by: Nginx, Inc.
2014-11-07 09:39:05 +00:00
SYSCTL_INT(_net_link_ether_inet, OID_AUTO, wait, CTLFLAG_VNET | CTLFLAG_RW,
Add a queue to hold packets while we await an ARP reply. When a fast machine first brings up some non TCP networking program it is quite possible that we will drop packets due to the fact that only one packet can be held per ARP entry. This leads to packets being missed when a program starts or restarts if the ARP data is not currently in the ARP cache. This code adds a new sysctl, net.link.ether.inet.maxhold, which defines a system wide maximum number of packets to be held in each ARP entry. Up to maxhold packets are queued until an ARP reply is received or the ARP times out. The default setting is the old value of 1 which has been part of the BSD networking code since time immemorial. Expose the time we hold an incomplete ARP entry by adding the sysctl net.link.ether.inet.wait, which defaults to 20 seconds, the value used when the new ARP code was added.. Reviewed by: bz, rpaulo MFC after: 3 weeks
2010-11-12 22:03:02 +00:00
&VNET_NAME(arpt_down), 0,
"Incomplete ARP entry lifetime in seconds");
Migrate structs arpstat, icmpstat, mrtstat, pimstat and udpstat to PCPU counters.
2013-07-09 09:50:15 +00:00
SYSCTL_VNET_PCPUSTAT(_net_link_ether_arp, OID_AUTO, stats, struct arpstat,
arpstat, "ARP statistics (struct arpstat, net/if_arp.h)");
Remove SYSCTL_VNET_* macros, and simply put CTLFLAG_VNET where needed. Sponsored by: Nginx, Inc.
2014-11-07 09:39:05 +00:00
SYSCTL_INT(_net_link_ether_inet, OID_AUTO, maxhold, CTLFLAG_VNET | CTLFLAG_RW,
Some style(9) and whitespace changes. Together with: Andrey Zonov <andrey zonov.org>
2012-07-31 11:31:12 +00:00
&VNET_NAME(arp_maxhold), 0,
Add a queue to hold packets while we await an ARP reply. When a fast machine first brings up some non TCP networking program it is quite possible that we will drop packets due to the fact that only one packet can be held per ARP entry. This leads to packets being missed when a program starts or restarts if the ARP data is not currently in the ARP cache. This code adds a new sysctl, net.link.ether.inet.maxhold, which defines a system wide maximum number of packets to be held in each ARP entry. Up to maxhold packets are queued until an ARP reply is received or the ARP times out. The default setting is the old value of 1 which has been part of the BSD networking code since time immemorial. Expose the time we hold an incomplete ARP entry by adding the sysctl net.link.ether.inet.wait, which defaults to 20 seconds, the value used when the new ARP code was added.. Reviewed by: bz, rpaulo MFC after: 3 weeks
2010-11-12 22:03:02 +00:00
"Number of packets to hold per ARP entry");
* Improve logging invalid arp messages * Remove redundant check in ip_arpinput Suggested by: glebius MFC after: 2 weeks
2015-09-15 08:50:44 +00:00
SYSCTL_INT(_net_link_ether_inet, OID_AUTO, max_log_per_second,
CTLFLAG_RW, &arp_maxpps, 0,
"Maximum number of remotely triggered ARP messages that can be "
"logged per second");
Add GARP retransmit capability A single gratuitous ARP (GARP) is always transmitted when an IPv4 address is added to an interface, and that is usually sufficient. However, in some circumstances, such as when a shared address is passed between cluster nodes, this single GARP may occasionally be dropped or lost. This can lead to neighbors on the network link working with a stale ARP cache and sending packets destined for that address to the node that previously owned the address, which may not respond. To avoid this situation, GARP retransmissions can be enabled by setting the net.link.ether.inet.garp_rexmit_count sysctl to a value greater than zero. The setting represents the maximum number of retransmissions. The interval between retransmissions is calculated using an exponential backoff algorithm, doubling each time, so the retransmission intervals are: {1, 2, 4, 8, 16, ...} (seconds). Due to the exponential backoff algorithm used for the interval between GARP retransmissions, the maximum number of retransmissions is limited to 16 for sanity. This limit corresponds to a maximum interval between retransmissions of 2^16 seconds ~= 18 hours. Increasing this limit is possible, but sending out GARPs spaced days apart would be of little use. Submitted by: David A. Bright <david.a.bright@dell.com> MFC after: 1 month Relnotes: yes Sponsored by: Dell EMC Differential Revision: https://reviews.freebsd.org/D7695
2016-10-02 01:42:45 +00:00
/*
* Due to the exponential backoff algorithm used for the interval between GARP
* retransmissions, the maximum number of retransmissions is limited for
* sanity. This limit corresponds to a maximum interval between retransmissions
* of 2^16 seconds ~= 18 hours.
*
* Making this limit more dynamic is more complicated than worthwhile,
* especially since sending out GARPs spaced days apart would be of little
* use. A maximum dynamic limit would look something like:
*
* const int max = fls(INT_MAX / hz) - 1;
*/
#define MAX_GARP_RETRANSMITS 16
static int sysctl_garp_rexmit(SYSCTL_HANDLER_ARGS);
static int garp_rexmit_count = 0; /* GARP retransmission setting. */
SYSCTL_PROC(_net_link_ether_inet, OID_AUTO, garp_rexmit_count,
CTLTYPE_INT|CTLFLAG_RW|CTLFLAG_MPSAFE,
&garp_rexmit_count, 0, sysctl_garp_rexmit, "I",
"Number of times to retransmit GARP packets;"
" 0 to disable, maximum of 16");
Improve ARP logging. r344504 added an extra ARP_LOG() call in case of an if_output() failure. It turns out IPv4 can be noisy. In order to not spam the console by default: (a) add a counter for these events so people can keep better track of how often it happens, and (b) add a sysctl to select the default ARP_LOG log level and set it to INFO avoiding the one (the new) DEBUG level by default. Claim a spare (1st one after 10 years since the stats were added) in order to not break netstat from FreeBSD 12->13 updates in the future. Reviewed by: karels Differential Revision: https://reviews.freebsd.org/D19490
2019-03-09 01:12:59 +00:00
VNET_DEFINE_STATIC(int, arp_log_level) = LOG_INFO; /* Min. log(9) level. */
#define V_arp_log_level VNET(arp_log_level)
SYSCTL_INT(_net_link_ether_arp, OID_AUTO, log_level, CTLFLAG_VNET | CTLFLAG_RW,
&VNET_NAME(arp_log_level), 0,
"Minimum log(9) level for recording rate limited arp log messages. "
"The higher will be log more (emerg=0, info=6 (default), debug=7).");
* Improve logging invalid arp messages * Remove redundant check in ip_arpinput Suggested by: glebius MFC after: 2 weeks
2015-09-15 08:50:44 +00:00
#define ARP_LOG(pri, ...) do { \
Improve ARP logging. r344504 added an extra ARP_LOG() call in case of an if_output() failure. It turns out IPv4 can be noisy. In order to not spam the console by default: (a) add a counter for these events so people can keep better track of how often it happens, and (b) add a sysctl to select the default ARP_LOG log level and set it to INFO avoiding the one (the new) DEBUG level by default. Claim a spare (1st one after 10 years since the stats were added) in order to not break netstat from FreeBSD 12->13 updates in the future. Reviewed by: karels Differential Revision: https://reviews.freebsd.org/D19490
2019-03-09 01:12:59 +00:00
if ((pri) <= V_arp_log_level && \
ppsratecheck(&arp_lastlog, &arp_curpps, arp_maxpps)) \
* Improve logging invalid arp messages * Remove redundant check in ip_arpinput Suggested by: glebius MFC after: 2 weeks
2015-09-15 08:50:44 +00:00
log((pri), "arp: " __VA_ARGS__); \
} while (0)
Remove old ballast, clean up a little bit, staticize. Add five sysctl variables that you should probably never tweak. net.arp.t_prune: 300 net.arp.t_keep: 1200 net.arp.t_down: 20 net.arp.maxtries: 5 net.arp.useloopback: 1 net.arp.proxyall: 0 (It's net.arp because arp isn't limited to inet, though our present implementation surely is).
1995-12-09 16:06:54 +00:00
Update netisr handling; Each SWI now registers its queue, and all queue drain routines are done by swi_net, which allows for better queue control at some future point. Packets may also be directly dispatched to a netisr instead of queued, this may be of interest at some installations, but currently defaults to off. Reviewed by: hsu, silby, jayanth, sam Sponsored by: DARPA, NAI Labs
2003-03-04 23:19:55 +00:00
static void arpintr(struct mbuf *);
Remove __P.
2002-03-19 21:25:46 +00:00
static void arptimer(void *);
Make INET a proper option. This will not make any of object files that LINT create change; there might be differences with INET disabled, but hardly anything compiled before without INET anyway. Now the 'obvious' things will give a proper error if compiled without inet - ipx_ip, ipfw, tcp_debug. The only thing that _should_ work (but can't be made to compile reasonably easily) is sppp :-( This commit move struct arpcom from <netinet/if_ether.h> to <net/if_arp.h>.
1998-01-08 23:42:31 +00:00
#ifdef INET
Remove __P.
2002-03-19 21:25:46 +00:00
static void in_arpinput(struct mbuf *);
Make INET a proper option. This will not make any of object files that LINT create change; there might be differences with INET disabled, but hardly anything compiled before without INET anyway. Now the 'obvious' things will give a proper error if compiled without inet - ipx_ip, ipfw, tcp_debug. The only thing that _should_ work (but can't be made to compile reasonably easily) is sppp :-( This commit move struct arpcom from <netinet/if_ether.h> to <net/if_arp.h>.
1998-01-08 23:42:31 +00:00
#endif
Implement full proxy ARP, gated on option ARP_PROXYALL. This allows a FreeBSD box to do proxy ARP as easily as most commercial routers do, without messing around with (potentially variable) Ethernet addresses. This code is really quite simple; I'm not at all sure why it wasn't implemented in 4.4. It might be worth stealing an interface flag (maybe IFF_LINK1) to use for finer-grained control over which interfaces get proxy treatment. For the moment, it's all or nothing.
1994-10-01 21:50:33 +00:00
Move lle update code from from gigantic ip_arpinput() to separate bunch of functions. The goal is to isolate actual lle updates to permit more fine-grained locking. Do all lle link-level update under AFDATA wlock. Sponsored by: Yandex LLC
2015-08-13 13:38:09 +00:00
static void arp_check_update_lle(struct arphdr *ah, struct in_addr isaddr,
struct ifnet *ifp, int bridged, struct llentry *la);
static void arp_mark_lle_reachable(struct llentry *la);
Use lladdr_event to propagate gratiotus arp. Differential Revision: https://reviews.freebsd.org/D4019
2015-11-09 10:11:14 +00:00
static void arp_iflladdr(void *arg __unused, struct ifnet *ifp);
Move lle update code from from gigantic ip_arpinput() to separate bunch of functions. The goal is to isolate actual lle updates to permit more fine-grained locking. Do all lle link-level update under AFDATA wlock. Sponsored by: Yandex LLC
2015-08-13 13:38:09 +00:00
Use lladdr_event to propagate gratiotus arp. Differential Revision: https://reviews.freebsd.org/D4019
2015-11-09 10:11:14 +00:00
static eventhandler_tag iflladdr_tag;
Move lle update code from from gigantic ip_arpinput() to separate bunch of functions. The goal is to isolate actual lle updates to permit more fine-grained locking. Do all lle link-level update under AFDATA wlock. Sponsored by: Yandex LLC
2015-08-13 13:38:09 +00:00
Reimplement the netisr framework in order to support parallel netisr threads: - Support up to one netisr thread per CPU, each processings its own workstream, or set of per-protocol queues. Threads may be bound to specific CPUs, or allowed to migrate, based on a global policy. In the future it would be desirable to support topology-centric policies, such as "one netisr per package". - Allow each protocol to advertise an ordering policy, which can currently be one of: NETISR_POLICY_SOURCE: packets must maintain ordering with respect to an implicit or explicit source (such as an interface or socket). NETISR_POLICY_FLOW: make use of mbuf flow identifiers to place work, as well as allowing protocols to provide a flow generation function for mbufs without flow identifers (m2flow). Falls back on NETISR_POLICY_SOURCE if now flow ID is available. NETISR_POLICY_CPU: allow protocols to inspect and assign a CPU for each packet handled by netisr (m2cpuid). - Provide utility functions for querying the number of workstreams being used, as well as a mapping function from workstream to CPU ID, which protocols may use in work placement decisions. - Add explicit interfaces to get and set per-protocol queue limits, and get and clear drop counters, which query data or apply changes across all workstreams. - Add a more extensible netisr registration interface, in which protocols declare 'struct netisr_handler' structures for each registered NETISR_ type. These include name, handler function, optional mbuf to flow ID function, optional mbuf to CPU ID function, queue limit, and ordering policy. Padding is present to allow these to be expanded in the future. If no queue limit is declared, then a default is used. - Queue limits are now per-workstream, and raised from the previous IFQ_MAXLEN default of 50 to 256. - All protocols are updated to use the new registration interface, and with the exception of netnatm, default queue limits. Most protocols register as NETISR_POLICY_SOURCE, except IPv4 and IPv6, which use NETISR_POLICY_FLOW, and will therefore take advantage of driver- generated flow IDs if present. - Formalize a non-packet based interface between interface polling and the netisr, rather than having polling pretend to be two protocols. Provide two explicit hooks in the netisr worker for start and end events for runs: netisr_poll() and netisr_pollmore(), as well as a function, netisr_sched_poll(), to allow the polling code to schedule netisr execution. DEVICE_POLLING still embeds single-netisr assumptions in its implementation, so for now if it is compiled into the kernel, a single and un-bound netisr thread is enforced regardless of tunable configuration. In the default configuration, the new netisr implementation maintains the same basic assumptions as the previous implementation: a single, un-bound worker thread processes all deferred work, and direct dispatch is enabled by default wherever possible. Performance measurement shows a marginal performance improvement over the old implementation due to the use of batched dequeue. An rmlock is used to synchronize use and registration/unregistration using the framework; currently, synchronized use is disabled (replicating current netisr policy) due to a measurable 3%-6% hit in ping-pong micro-benchmarking. It will be enabled once further rmlock optimization has taken place. However, in practice, netisrs are rarely registered or unregistered at runtime. A new man page for netisr will follow, but since one doesn't currently exist, it hasn't been updated. This change is not appropriate for MFC, although the polling shutdown handler should be merged to 7-STABLE. Bump __FreeBSD_version. Reviewed by: bz
2009-06-01 10:41:38 +00:00
static const struct netisr_handler arp_nh = {
.nh_name = "arp",
.nh_handler = arpintr,
.nh_proto = NETISR_ARP,
.nh_policy = NETISR_POLICY_SOURCE,
};
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
/*
This main goals of this project are: 1. separating L2 tables (ARP, NDP) from the L3 routing tables 2. removing as much locking dependencies among these layers as possible to allow for some parallelism in the search operations 3. simplify the logic in the routing code, The most notable end result is the obsolescent of the route cloning (RTF_CLONING) concept, which translated into code reduction in both IPv4 ARP and IPv6 NDP related modules, and size reduction in struct rtentry{}. The change in design obsoletes the semantics of RTF_CLONING, RTF_WASCLONE and RTF_LLINFO routing flags. The userland applications such as "arp" and "ndp" have been modified to reflect those changes. The output from "netstat -r" shows only the routing entries. Quite a few developers have contributed to this project in the past: Glebius Smirnoff, Luigi Rizzo, Alessandro Cerri, and Andre Oppermann. And most recently: - Kip Macy revised the locking code completely, thus completing the last piece of the puzzle, Kip has also been conducting active functional testing - Sam Leffler has helped me improving/refactoring the code, and provided valuable reviews - Julian Elischer setup the perforce tree for me and has helped me maintaining that branch before the svn conversion
2008-12-15 06:10:57 +00:00
* Timeout routine. Age arp_tab entries periodically.
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
*/
Make arp_rtrequest() static since nobody needs to referene it any more.
1994-12-22 22:00:30 +00:00
static void
This main goals of this project are: 1. separating L2 tables (ARP, NDP) from the L3 routing tables 2. removing as much locking dependencies among these layers as possible to allow for some parallelism in the search operations 3. simplify the logic in the routing code, The most notable end result is the obsolescent of the route cloning (RTF_CLONING) concept, which translated into code reduction in both IPv4 ARP and IPv6 NDP related modules, and size reduction in struct rtentry{}. The change in design obsoletes the semantics of RTF_CLONING, RTF_WASCLONE and RTF_LLINFO routing flags. The userland applications such as "arp" and "ndp" have been modified to reflect those changes. The output from "netstat -r" shows only the routing entries. Quite a few developers have contributed to this project in the past: Glebius Smirnoff, Luigi Rizzo, Alessandro Cerri, and Andre Oppermann. And most recently: - Kip Macy revised the locking code completely, thus completing the last piece of the puzzle, Kip has also been conducting active functional testing - Sam Leffler has helped me improving/refactoring the code, and provided valuable reviews - Julian Elischer setup the perforce tree for me and has helped me maintaining that branch before the svn conversion
2008-12-15 06:10:57 +00:00
arptimer(void *arg)
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
{
Fix races between in_lltable_prefix_free(), lla_lookup(), llentry_free() and arptimer(): o Use callout_init_rw() for lle timeout, this allows us safely disestablish them. - This allows us to simplify the arptimer() and make it race safe. o Consistently use ifp->if_afdata_lock to lock access to linked lists in the lle hashes. o Introduce new lle flag LLE_LINKED, which marks an entry that is attached to the hash. - Use LLE_LINKED to avoid double unlinking via consequent calls to llentry_free(). - Mark lle with LLE_DELETED via |= operation istead of =, so that other flags won't be lost. o Make LLE_ADDREF(), LLE_REMREF() and LLE_FREE_LOCKED() more consistent and provide more informative KASSERTs. The patch is a collaborative work of all submitters and myself. PR: kern/165863 Submitted by: Andrey Zonov <andrey zonov.org> Submitted by: Ryan Stone <rysto32 gmail.com> Submitted by: Eric van Gyzen <eric_van_gyzen dell.com>
2012-08-02 13:57:49 +00:00
struct llentry *lle = (struct llentry *)arg;
This main goals of this project are: 1. separating L2 tables (ARP, NDP) from the L3 routing tables 2. removing as much locking dependencies among these layers as possible to allow for some parallelism in the search operations 3. simplify the logic in the routing code, The most notable end result is the obsolescent of the route cloning (RTF_CLONING) concept, which translated into code reduction in both IPv4 ARP and IPv6 NDP related modules, and size reduction in struct rtentry{}. The change in design obsoletes the semantics of RTF_CLONING, RTF_WASCLONE and RTF_LLINFO routing flags. The userland applications such as "arp" and "ndp" have been modified to reflect those changes. The output from "netstat -r" shows only the routing entries. Quite a few developers have contributed to this project in the past: Glebius Smirnoff, Luigi Rizzo, Alessandro Cerri, and Andre Oppermann. And most recently: - Kip Macy revised the locking code completely, thus completing the last piece of the puzzle, Kip has also been conducting active functional testing - Sam Leffler has helped me improving/refactoring the code, and provided valuable reviews - Julian Elischer setup the perforce tree for me and has helped me maintaining that branch before the svn conversion
2008-12-15 06:10:57 +00:00
struct ifnet *ifp;
Remove LLE read lock from IPv4 fast path. LLE structure is mostly unchanged during its lifecycle. To be more specific, there are 2 things relevant for fast path lookup code: 1) link-level address change. Since r286722, these updates are performed under AFDATA WLOCK. 2) Some sort of feedback indicating that this particular entry is used so we re-send arp request to perform reachability verification instead of expiring entry. The only signal that is needed from fast path is something like binary yes/no. The latter is solved by the following changes: 1) introduce special r_skip_req field which is read lockless by fast path, but updated under (new) req_mutex mutex. If this field is non-zero, then fast path will acquire lock and set it back to 0. 2) introduce simple state machine: incomplete->reachable<->verify->deleted. Before that we implicitely had incomplete->reachable->deleted state machine, with V_arpt_keep between "reachable" and "deleted". Verification was performed in runtime 5 seconds before V_arpt_keep expire. This is changed to "change state to verify 5 seconds before V_arpt_keep, set r_skip_req to non-zero value and check it every second". If the value is zero - then send arp verification probe. These changes do not introduce any signifficant control plane overhead: typically lle callout timer would fire 1 time more each V_arpt_keep (1200s) for used lles and up to arp_maxtries (5) for dead lles. As a result, all packets towards "reachable" lle are handled by fast path without acquiring lle read lock. Additional "req_mutex" is needed because callout / arpresolve_slow() or eventhandler might keep LLE lock for signifficant amount of time, which might not be feasible for fast path locking (e.g. having rmlock as ether AFDATA or lltable own lock). Differential Revision: https://reviews.freebsd.org/D3688
2015-12-05 09:50:37 +00:00
int r_skip_req;
Fix races between in_lltable_prefix_free(), lla_lookup(), llentry_free() and arptimer(): o Use callout_init_rw() for lle timeout, this allows us safely disestablish them. - This allows us to simplify the arptimer() and make it race safe. o Consistently use ifp->if_afdata_lock to lock access to linked lists in the lle hashes. o Introduce new lle flag LLE_LINKED, which marks an entry that is attached to the hash. - Use LLE_LINKED to avoid double unlinking via consequent calls to llentry_free(). - Mark lle with LLE_DELETED via |= operation istead of =, so that other flags won't be lost. o Make LLE_ADDREF(), LLE_REMREF() and LLE_FREE_LOCKED() more consistent and provide more informative KASSERTs. The patch is a collaborative work of all submitters and myself. PR: kern/165863 Submitted by: Andrey Zonov <andrey zonov.org> Submitted by: Ryan Stone <rysto32 gmail.com> Submitted by: Eric van Gyzen <eric_van_gyzen dell.com>
2012-08-02 13:57:49 +00:00
if (lle->la_flags & LLE_STATIC) {
return;
}
This fixes a bug in the way that the LLE timers for nd6 and arp were being used. They basically would pass in the mutex to the callout_init. Because they used this method to the callout system, it was possible to "stop" the callout. When flushing the table and you stopped the running callout, the callout_stop code would return 1 indicating that it was going to stop the callout (that was about to run on the callout_wheel blocked by the function calling the stop). Now when 1 was returned, it would lower the reference count one extra time for the stopped timer, then a few lines later delete the memory. Of course the callout_wheel was stuck in the lock code and would then crash since it was accessing freed memory. By using callout_init(c, 1) we always get a 0 back and the reference counting bug does not rear its head. We do have to make a few adjustments to the callouts themselves though to make sure it does the proper thing if rescheduled as well as gets the lock. Commented upon by hiren and sbruno See Phabricator D1777 for more details. Commented upon by hiren and sbruno Reviewed by: adrian, jhb and bz Sponsored by: Netflix Inc.
2015-02-09 19:28:11 +00:00
LLE_WLOCK(lle);
Use single 'lle_timer' callout in lltable instead of two different names of the same timer.
2015-08-11 12:38:54 +00:00
if (callout_pending(&lle->lle_timer)) {
This fixes a bug in the way that the LLE timers for nd6 and arp were being used. They basically would pass in the mutex to the callout_init. Because they used this method to the callout system, it was possible to "stop" the callout. When flushing the table and you stopped the running callout, the callout_stop code would return 1 indicating that it was going to stop the callout (that was about to run on the callout_wheel blocked by the function calling the stop). Now when 1 was returned, it would lower the reference count one extra time for the stopped timer, then a few lines later delete the memory. Of course the callout_wheel was stuck in the lock code and would then crash since it was accessing freed memory. By using callout_init(c, 1) we always get a 0 back and the reference counting bug does not rear its head. We do have to make a few adjustments to the callouts themselves though to make sure it does the proper thing if rescheduled as well as gets the lock. Commented upon by hiren and sbruno See Phabricator D1777 for more details. Commented upon by hiren and sbruno Reviewed by: adrian, jhb and bz Sponsored by: Netflix Inc.
2015-02-09 19:28:11 +00:00
/*
White space cleanup -- remove trailing tab's or spaces from any line. Sponsored by: Netflix Inc.
2020-02-12 13:31:36 +00:00
* Here we are a bit odd here in the treatment of
This fixes a bug in the way that the LLE timers for nd6 and arp were being used. They basically would pass in the mutex to the callout_init. Because they used this method to the callout system, it was possible to "stop" the callout. When flushing the table and you stopped the running callout, the callout_stop code would return 1 indicating that it was going to stop the callout (that was about to run on the callout_wheel blocked by the function calling the stop). Now when 1 was returned, it would lower the reference count one extra time for the stopped timer, then a few lines later delete the memory. Of course the callout_wheel was stuck in the lock code and would then crash since it was accessing freed memory. By using callout_init(c, 1) we always get a 0 back and the reference counting bug does not rear its head. We do have to make a few adjustments to the callouts themselves though to make sure it does the proper thing if rescheduled as well as gets the lock. Commented upon by hiren and sbruno See Phabricator D1777 for more details. Commented upon by hiren and sbruno Reviewed by: adrian, jhb and bz Sponsored by: Netflix Inc.
2015-02-09 19:28:11 +00:00
* active/pending. If the pending bit is set, it got
* rescheduled before I ran. The active
* bit we ignore, since if it was stopped
* in ll_tablefree() and was currently running
* it would have return 0 so the code would
* not have deleted it since the callout could
* not be stopped so we want to go through
* with the delete here now. If the callout
* was restarted, the pending bit will be back on and
* we just want to bail since the callout_reset would
* return 1 and our reference would have been removed
* by arpresolve() below.
*/
LLE_WUNLOCK(lle);
return;
}
This main goals of this project are: 1. separating L2 tables (ARP, NDP) from the L3 routing tables 2. removing as much locking dependencies among these layers as possible to allow for some parallelism in the search operations 3. simplify the logic in the routing code, The most notable end result is the obsolescent of the route cloning (RTF_CLONING) concept, which translated into code reduction in both IPv4 ARP and IPv6 NDP related modules, and size reduction in struct rtentry{}. The change in design obsoletes the semantics of RTF_CLONING, RTF_WASCLONE and RTF_LLINFO routing flags. The userland applications such as "arp" and "ndp" have been modified to reflect those changes. The output from "netstat -r" shows only the routing entries. Quite a few developers have contributed to this project in the past: Glebius Smirnoff, Luigi Rizzo, Alessandro Cerri, and Andre Oppermann. And most recently: - Kip Macy revised the locking code completely, thus completing the last piece of the puzzle, Kip has also been conducting active functional testing - Sam Leffler has helped me improving/refactoring the code, and provided valuable reviews - Julian Elischer setup the perforce tree for me and has helped me maintaining that branch before the svn conversion
2008-12-15 06:10:57 +00:00
ifp = lle->lle_tbl->llt_ifp;
Add ARP statistics to the kernel and netstat. New counters now exist for: requests sent replies sent requests received replies received packets received total packets dropped due to no ARP entry entrys timed out Duplicate IPs seen The new statistics are seen in the netstat command when it is given the -s command line switch. MFC after: 2 weeks In collaboration with: bz
2009-09-03 21:10:57 +00:00
CURVNET_SET(ifp->if_vnet);
Fix races between in_lltable_prefix_free(), lla_lookup(), llentry_free() and arptimer(): o Use callout_init_rw() for lle timeout, this allows us safely disestablish them. - This allows us to simplify the arptimer() and make it race safe. o Consistently use ifp->if_afdata_lock to lock access to linked lists in the lle hashes. o Introduce new lle flag LLE_LINKED, which marks an entry that is attached to the hash. - Use LLE_LINKED to avoid double unlinking via consequent calls to llentry_free(). - Mark lle with LLE_DELETED via |= operation istead of =, so that other flags won't be lost. o Make LLE_ADDREF(), LLE_REMREF() and LLE_FREE_LOCKED() more consistent and provide more informative KASSERTs. The patch is a collaborative work of all submitters and myself. PR: kern/165863 Submitted by: Andrey Zonov <andrey zonov.org> Submitted by: Ryan Stone <rysto32 gmail.com> Submitted by: Eric van Gyzen <eric_van_gyzen dell.com>
2012-08-02 13:57:49 +00:00
Remove LLE read lock from IPv4 fast path. LLE structure is mostly unchanged during its lifecycle. To be more specific, there are 2 things relevant for fast path lookup code: 1) link-level address change. Since r286722, these updates are performed under AFDATA WLOCK. 2) Some sort of feedback indicating that this particular entry is used so we re-send arp request to perform reachability verification instead of expiring entry. The only signal that is needed from fast path is something like binary yes/no. The latter is solved by the following changes: 1) introduce special r_skip_req field which is read lockless by fast path, but updated under (new) req_mutex mutex. If this field is non-zero, then fast path will acquire lock and set it back to 0. 2) introduce simple state machine: incomplete->reachable<->verify->deleted. Before that we implicitely had incomplete->reachable->deleted state machine, with V_arpt_keep between "reachable" and "deleted". Verification was performed in runtime 5 seconds before V_arpt_keep expire. This is changed to "change state to verify 5 seconds before V_arpt_keep, set r_skip_req to non-zero value and check it every second". If the value is zero - then send arp verification probe. These changes do not introduce any signifficant control plane overhead: typically lle callout timer would fire 1 time more each V_arpt_keep (1200s) for used lles and up to arp_maxtries (5) for dead lles. As a result, all packets towards "reachable" lle are handled by fast path without acquiring lle read lock. Additional "req_mutex" is needed because callout / arpresolve_slow() or eventhandler might keep LLE lock for signifficant amount of time, which might not be feasible for fast path locking (e.g. having rmlock as ether AFDATA or lltable own lock). Differential Revision: https://reviews.freebsd.org/D3688
2015-12-05 09:50:37 +00:00
switch (lle->ln_state) {
case ARP_LLINFO_REACHABLE:
/*
* Expiration time is approaching.
* Let's try to refresh entry if it is still
* in use.
*
* Set r_skip_req to get feedback from
* fast path. Change state and re-schedule
* ourselves.
*/
LLE_REQ_LOCK(lle);
lle->r_skip_req = 1;
LLE_REQ_UNLOCK(lle);
lle->ln_state = ARP_LLINFO_VERIFY;
callout_schedule(&lle->lle_timer, hz * V_arpt_rexmit);
LLE_WUNLOCK(lle);
CURVNET_RESTORE();
return;
case ARP_LLINFO_VERIFY:
LLE_REQ_LOCK(lle);
r_skip_req = lle->r_skip_req;
LLE_REQ_UNLOCK(lle);
if (r_skip_req == 0 && lle->la_preempt > 0) {
/* Entry was used, issue refresh request */
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
struct epoch_tracker et;
Remove LLE read lock from IPv4 fast path. LLE structure is mostly unchanged during its lifecycle. To be more specific, there are 2 things relevant for fast path lookup code: 1) link-level address change. Since r286722, these updates are performed under AFDATA WLOCK. 2) Some sort of feedback indicating that this particular entry is used so we re-send arp request to perform reachability verification instead of expiring entry. The only signal that is needed from fast path is something like binary yes/no. The latter is solved by the following changes: 1) introduce special r_skip_req field which is read lockless by fast path, but updated under (new) req_mutex mutex. If this field is non-zero, then fast path will acquire lock and set it back to 0. 2) introduce simple state machine: incomplete->reachable<->verify->deleted. Before that we implicitely had incomplete->reachable->deleted state machine, with V_arpt_keep between "reachable" and "deleted". Verification was performed in runtime 5 seconds before V_arpt_keep expire. This is changed to "change state to verify 5 seconds before V_arpt_keep, set r_skip_req to non-zero value and check it every second". If the value is zero - then send arp verification probe. These changes do not introduce any signifficant control plane overhead: typically lle callout timer would fire 1 time more each V_arpt_keep (1200s) for used lles and up to arp_maxtries (5) for dead lles. As a result, all packets towards "reachable" lle are handled by fast path without acquiring lle read lock. Additional "req_mutex" is needed because callout / arpresolve_slow() or eventhandler might keep LLE lock for signifficant amount of time, which might not be feasible for fast path locking (e.g. having rmlock as ether AFDATA or lltable own lock). Differential Revision: https://reviews.freebsd.org/D3688
2015-12-05 09:50:37 +00:00
struct in_addr dst;
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
Remove LLE read lock from IPv4 fast path. LLE structure is mostly unchanged during its lifecycle. To be more specific, there are 2 things relevant for fast path lookup code: 1) link-level address change. Since r286722, these updates are performed under AFDATA WLOCK. 2) Some sort of feedback indicating that this particular entry is used so we re-send arp request to perform reachability verification instead of expiring entry. The only signal that is needed from fast path is something like binary yes/no. The latter is solved by the following changes: 1) introduce special r_skip_req field which is read lockless by fast path, but updated under (new) req_mutex mutex. If this field is non-zero, then fast path will acquire lock and set it back to 0. 2) introduce simple state machine: incomplete->reachable<->verify->deleted. Before that we implicitely had incomplete->reachable->deleted state machine, with V_arpt_keep between "reachable" and "deleted". Verification was performed in runtime 5 seconds before V_arpt_keep expire. This is changed to "change state to verify 5 seconds before V_arpt_keep, set r_skip_req to non-zero value and check it every second". If the value is zero - then send arp verification probe. These changes do not introduce any signifficant control plane overhead: typically lle callout timer would fire 1 time more each V_arpt_keep (1200s) for used lles and up to arp_maxtries (5) for dead lles. As a result, all packets towards "reachable" lle are handled by fast path without acquiring lle read lock. Additional "req_mutex" is needed because callout / arpresolve_slow() or eventhandler might keep LLE lock for signifficant amount of time, which might not be feasible for fast path locking (e.g. having rmlock as ether AFDATA or lltable own lock). Differential Revision: https://reviews.freebsd.org/D3688
2015-12-05 09:50:37 +00:00
dst = lle->r_l3addr.addr4;
lle->la_preempt--;
callout_schedule(&lle->lle_timer, hz * V_arpt_rexmit);
LLE_WUNLOCK(lle);
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
NET_EPOCH_ENTER(et);
Remove LLE read lock from IPv4 fast path. LLE structure is mostly unchanged during its lifecycle. To be more specific, there are 2 things relevant for fast path lookup code: 1) link-level address change. Since r286722, these updates are performed under AFDATA WLOCK. 2) Some sort of feedback indicating that this particular entry is used so we re-send arp request to perform reachability verification instead of expiring entry. The only signal that is needed from fast path is something like binary yes/no. The latter is solved by the following changes: 1) introduce special r_skip_req field which is read lockless by fast path, but updated under (new) req_mutex mutex. If this field is non-zero, then fast path will acquire lock and set it back to 0. 2) introduce simple state machine: incomplete->reachable<->verify->deleted. Before that we implicitely had incomplete->reachable->deleted state machine, with V_arpt_keep between "reachable" and "deleted". Verification was performed in runtime 5 seconds before V_arpt_keep expire. This is changed to "change state to verify 5 seconds before V_arpt_keep, set r_skip_req to non-zero value and check it every second". If the value is zero - then send arp verification probe. These changes do not introduce any signifficant control plane overhead: typically lle callout timer would fire 1 time more each V_arpt_keep (1200s) for used lles and up to arp_maxtries (5) for dead lles. As a result, all packets towards "reachable" lle are handled by fast path without acquiring lle read lock. Additional "req_mutex" is needed because callout / arpresolve_slow() or eventhandler might keep LLE lock for signifficant amount of time, which might not be feasible for fast path locking (e.g. having rmlock as ether AFDATA or lltable own lock). Differential Revision: https://reviews.freebsd.org/D3688
2015-12-05 09:50:37 +00:00
arprequest(ifp, NULL, &dst, NULL);
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
NET_EPOCH_EXIT(et);
Remove LLE read lock from IPv4 fast path. LLE structure is mostly unchanged during its lifecycle. To be more specific, there are 2 things relevant for fast path lookup code: 1) link-level address change. Since r286722, these updates are performed under AFDATA WLOCK. 2) Some sort of feedback indicating that this particular entry is used so we re-send arp request to perform reachability verification instead of expiring entry. The only signal that is needed from fast path is something like binary yes/no. The latter is solved by the following changes: 1) introduce special r_skip_req field which is read lockless by fast path, but updated under (new) req_mutex mutex. If this field is non-zero, then fast path will acquire lock and set it back to 0. 2) introduce simple state machine: incomplete->reachable<->verify->deleted. Before that we implicitely had incomplete->reachable->deleted state machine, with V_arpt_keep between "reachable" and "deleted". Verification was performed in runtime 5 seconds before V_arpt_keep expire. This is changed to "change state to verify 5 seconds before V_arpt_keep, set r_skip_req to non-zero value and check it every second". If the value is zero - then send arp verification probe. These changes do not introduce any signifficant control plane overhead: typically lle callout timer would fire 1 time more each V_arpt_keep (1200s) for used lles and up to arp_maxtries (5) for dead lles. As a result, all packets towards "reachable" lle are handled by fast path without acquiring lle read lock. Additional "req_mutex" is needed because callout / arpresolve_slow() or eventhandler might keep LLE lock for signifficant amount of time, which might not be feasible for fast path locking (e.g. having rmlock as ether AFDATA or lltable own lock). Differential Revision: https://reviews.freebsd.org/D3688
2015-12-05 09:50:37 +00:00
CURVNET_RESTORE();
return;
}
/* Nothing happened. Reschedule if not too late */
if (lle->la_expire > time_uptime) {
callout_schedule(&lle->lle_timer, hz * V_arpt_rexmit);
LLE_WUNLOCK(lle);
CURVNET_RESTORE();
return;
}
break;
case ARP_LLINFO_INCOMPLETE:
case ARP_LLINFO_DELETED:
break;
}
Catch up with r238990. LLE_DELETED does not clobber everything else in la_flags since said revision.
2013-07-03 17:27:32 +00:00
if ((lle->la_flags & LLE_DELETED) == 0) {
Fix races between in_lltable_prefix_free(), lla_lookup(), llentry_free() and arptimer(): o Use callout_init_rw() for lle timeout, this allows us safely disestablish them. - This allows us to simplify the arptimer() and make it race safe. o Consistently use ifp->if_afdata_lock to lock access to linked lists in the lle hashes. o Introduce new lle flag LLE_LINKED, which marks an entry that is attached to the hash. - Use LLE_LINKED to avoid double unlinking via consequent calls to llentry_free(). - Mark lle with LLE_DELETED via |= operation istead of =, so that other flags won't be lost. o Make LLE_ADDREF(), LLE_REMREF() and LLE_FREE_LOCKED() more consistent and provide more informative KASSERTs. The patch is a collaborative work of all submitters and myself. PR: kern/165863 Submitted by: Andrey Zonov <andrey zonov.org> Submitted by: Ryan Stone <rysto32 gmail.com> Submitted by: Eric van Gyzen <eric_van_gyzen dell.com>
2012-08-02 13:57:49 +00:00
int evt;
if (lle->la_flags & LLE_VALID)
evt = LLENTRY_EXPIRED;
else
evt = LLENTRY_TIMEDOUT;
EVENTHANDLER_INVOKE(lle_event, lle, evt);
}
Use single 'lle_timer' callout in lltable instead of two different names of the same timer.
2015-08-11 12:38:54 +00:00
callout_stop(&lle->lle_timer);
Fix races between in_lltable_prefix_free(), lla_lookup(), llentry_free() and arptimer(): o Use callout_init_rw() for lle timeout, this allows us safely disestablish them. - This allows us to simplify the arptimer() and make it race safe. o Consistently use ifp->if_afdata_lock to lock access to linked lists in the lle hashes. o Introduce new lle flag LLE_LINKED, which marks an entry that is attached to the hash. - Use LLE_LINKED to avoid double unlinking via consequent calls to llentry_free(). - Mark lle with LLE_DELETED via |= operation istead of =, so that other flags won't be lost. o Make LLE_ADDREF(), LLE_REMREF() and LLE_FREE_LOCKED() more consistent and provide more informative KASSERTs. The patch is a collaborative work of all submitters and myself. PR: kern/165863 Submitted by: Andrey Zonov <andrey zonov.org> Submitted by: Ryan Stone <rysto32 gmail.com> Submitted by: Eric van Gyzen <eric_van_gyzen dell.com>
2012-08-02 13:57:49 +00:00
/* XXX: LOR avoidance. We still have ref on lle. */
LLE_WUNLOCK(lle);
This main goals of this project are: 1. separating L2 tables (ARP, NDP) from the L3 routing tables 2. removing as much locking dependencies among these layers as possible to allow for some parallelism in the search operations 3. simplify the logic in the routing code, The most notable end result is the obsolescent of the route cloning (RTF_CLONING) concept, which translated into code reduction in both IPv4 ARP and IPv6 NDP related modules, and size reduction in struct rtentry{}. The change in design obsoletes the semantics of RTF_CLONING, RTF_WASCLONE and RTF_LLINFO routing flags. The userland applications such as "arp" and "ndp" have been modified to reflect those changes. The output from "netstat -r" shows only the routing entries. Quite a few developers have contributed to this project in the past: Glebius Smirnoff, Luigi Rizzo, Alessandro Cerri, and Andre Oppermann. And most recently: - Kip Macy revised the locking code completely, thus completing the last piece of the puzzle, Kip has also been conducting active functional testing - Sam Leffler has helped me improving/refactoring the code, and provided valuable reviews - Julian Elischer setup the perforce tree for me and has helped me maintaining that branch before the svn conversion
2008-12-15 06:10:57 +00:00
IF_AFDATA_LOCK(ifp);
LLE_WLOCK(lle);
Fix races between in_lltable_prefix_free(), lla_lookup(), llentry_free() and arptimer(): o Use callout_init_rw() for lle timeout, this allows us safely disestablish them. - This allows us to simplify the arptimer() and make it race safe. o Consistently use ifp->if_afdata_lock to lock access to linked lists in the lle hashes. o Introduce new lle flag LLE_LINKED, which marks an entry that is attached to the hash. - Use LLE_LINKED to avoid double unlinking via consequent calls to llentry_free(). - Mark lle with LLE_DELETED via |= operation istead of =, so that other flags won't be lost. o Make LLE_ADDREF(), LLE_REMREF() and LLE_FREE_LOCKED() more consistent and provide more informative KASSERTs. The patch is a collaborative work of all submitters and myself. PR: kern/165863 Submitted by: Andrey Zonov <andrey zonov.org> Submitted by: Ryan Stone <rysto32 gmail.com> Submitted by: Eric van Gyzen <eric_van_gyzen dell.com>
2012-08-02 13:57:49 +00:00
Fix problem in r238990. The LLE_LINKED flag should be tested prior to entering llentry_free(), and in case if we lose the race, we should simply perform LLE_FREE_LOCKED(). Otherwise, if the race is lost by the thread performing arptimer(), it will remove two references from the lle instead of one. Reported by: Ian FREISLICH <ianf clue.co.za>
2012-12-13 11:11:15 +00:00
/* Guard against race with other llentry_free(). */
if (lle->la_flags & LLE_LINKED) {
LLE_REMREF(lle);
* Require explicitl lle unlink prior to calling llentry_delete(). This one slightly decreases time of holding afdata wlock. * While here, make nd6_free() return void. No one has used its return value since r186119.
2015-09-15 06:48:19 +00:00
lltable_unlink_entry(lle->lle_tbl, lle);
}
This main goals of this project are: 1. separating L2 tables (ARP, NDP) from the L3 routing tables 2. removing as much locking dependencies among these layers as possible to allow for some parallelism in the search operations 3. simplify the logic in the routing code, The most notable end result is the obsolescent of the route cloning (RTF_CLONING) concept, which translated into code reduction in both IPv4 ARP and IPv6 NDP related modules, and size reduction in struct rtentry{}. The change in design obsoletes the semantics of RTF_CLONING, RTF_WASCLONE and RTF_LLINFO routing flags. The userland applications such as "arp" and "ndp" have been modified to reflect those changes. The output from "netstat -r" shows only the routing entries. Quite a few developers have contributed to this project in the past: Glebius Smirnoff, Luigi Rizzo, Alessandro Cerri, and Andre Oppermann. And most recently: - Kip Macy revised the locking code completely, thus completing the last piece of the puzzle, Kip has also been conducting active functional testing - Sam Leffler has helped me improving/refactoring the code, and provided valuable reviews - Julian Elischer setup the perforce tree for me and has helped me maintaining that branch before the svn conversion
2008-12-15 06:10:57 +00:00
IF_AFDATA_UNLOCK(ifp);
Fix problem in r238990. The LLE_LINKED flag should be tested prior to entering llentry_free(), and in case if we lose the race, we should simply perform LLE_FREE_LOCKED(). Otherwise, if the race is lost by the thread performing arptimer(), it will remove two references from the lle instead of one. Reported by: Ian FREISLICH <ianf clue.co.za>
2012-12-13 11:11:15 +00:00
* Require explicitl lle unlink prior to calling llentry_delete(). This one slightly decreases time of holding afdata wlock. * While here, make nd6_free() return void. No one has used its return value since r186119.
2015-09-15 06:48:19 +00:00
size_t pkts_dropped = llentry_free(lle);
ARPSTAT_ADD(dropped, pkts_dropped);
Fix races between in_lltable_prefix_free(), lla_lookup(), llentry_free() and arptimer(): o Use callout_init_rw() for lle timeout, this allows us safely disestablish them. - This allows us to simplify the arptimer() and make it race safe. o Consistently use ifp->if_afdata_lock to lock access to linked lists in the lle hashes. o Introduce new lle flag LLE_LINKED, which marks an entry that is attached to the hash. - Use LLE_LINKED to avoid double unlinking via consequent calls to llentry_free(). - Mark lle with LLE_DELETED via |= operation istead of =, so that other flags won't be lost. o Make LLE_ADDREF(), LLE_REMREF() and LLE_FREE_LOCKED() more consistent and provide more informative KASSERTs. The patch is a collaborative work of all submitters and myself. PR: kern/165863 Submitted by: Andrey Zonov <andrey zonov.org> Submitted by: Ryan Stone <rysto32 gmail.com> Submitted by: Eric van Gyzen <eric_van_gyzen dell.com>
2012-08-02 13:57:49 +00:00
ARPSTAT_INC(timeouts);
Fix problem in r238990. The LLE_LINKED flag should be tested prior to entering llentry_free(), and in case if we lose the race, we should simply perform LLE_FREE_LOCKED(). Otherwise, if the race is lost by the thread performing arptimer(), it will remove two references from the lle instead of one. Reported by: Ian FREISLICH <ianf clue.co.za>
2012-12-13 11:11:15 +00:00
Add ARP statistics to the kernel and netstat. New counters now exist for: requests sent replies sent requests received replies received packets received total packets dropped due to no ARP entry entrys timed out Duplicate IPs seen The new statistics are seen in the netstat command when it is given the -s command line switch. MFC after: 2 weeks In collaboration with: bz
2009-09-03 21:10:57 +00:00
CURVNET_RESTORE();
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
Implement interface link header precomputation API. Add if_requestencap() interface method which is capable of calculating various link headers for given interface. Right now there is support for INET/INET6/ARP llheader calculation (IFENCAP_LL type request). Other types are planned to support more complex calculation (L2 multipath lagg nexthops, tunnel encap nexthops, etc..). Reshape 'struct route' to be able to pass additional data (with is length) to prepend to mbuf. These two changes permits routing code to pass pre-calculated nexthop data (like L2 header for route w/gateway) down to the stack eliminating the need for other lookups. It also brings us closer to more complex scenarios like transparently handling MPLS nexthops and tunnel interfaces. Last, but not least, it removes layering violation introduced by flowtable code (ro_lle) and simplifies handling of existing if_output consumers. ARP/ND changes: Make arp/ndp stack pre-calculate link header upon installing/updating lle record. Interface link address change are handled by re-calculating headers for all lles based on if_lladdr event. After these changes, arpresolve()/nd6_resolve() returns full pre-calculated header for supported interfaces thus simplifying if_output(). Move these lookups to separate ether_resolve_addr() function which ether returs error or fully-prepared link header. Add <arp|nd6_>resolve_addr() compat versions to return link addresses instead of pre-calculated data. BPF changes: Raw bpf writes occupied _two_ cases: AF_UNSPEC and pseudo_AF_HDRCMPLT. Despite the naming, both of there have ther header "complete". The only difference is that interface source mac has to be filled by OS for AF_UNSPEC (controlled via BIOCGHDRCMPLT). This logic has to stay inside BPF and not pollute if_output() routines. Convert BPF to pass prepend data via new 'struct route' mechanism. Note that it does not change non-optimized if_output(): ro_prepend handling is purely optional. Side note: hackish pseudo_AF_HDRCMPLT is supported for ethernet and FDDI. It is not needed for ethernet anymore. The only remaining FDDI user is dev/pdq mostly untouched since 2007. FDDI support was eliminated from OpenBSD in 2013 (sys/net/if_fddisubr.c rev 1.65). Flowtable changes: Flowtable violates layering by saving (and not correctly managing) rtes/lles. Instead of passing lle pointer, pass pointer to pre-calculated header data from that lle. Differential Revision: https://reviews.freebsd.org/D4102
2015-12-31 05:03:27 +00:00
/*
* Stores link-layer header for @ifp in format suitable for if_output()
* into buffer @buf. Resulting header length is stored in @bufsize.
*
* Returns 0 on success.
*/
static int
arp_fillheader(struct ifnet *ifp, struct arphdr *ah, int bcast, u_char *buf,
size_t *bufsize)
{
struct if_encap_req ereq;
int error;
bzero(buf, *bufsize);
bzero(&ereq, sizeof(ereq));
ereq.buf = buf;
ereq.bufsize = *bufsize;
ereq.rtype = IFENCAP_LL;
ereq.family = AF_ARP;
ereq.lladdr = ar_tha(ah);
ereq.hdata = (u_char *)ah;
if (bcast)
ereq.flags = IFENCAP_FLAG_BROADCAST;
error = ifp->if_requestencap(ifp, &ereq);
if (error == 0)
*bufsize = ereq.bufsize;
return (error);
}
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
/*
* Broadcast an ARP request. Caller specifies:
* - arp header source ip address
* - arp header target ip address
* - arp header source ethernet address
*/
Make arp code return (more) errors. arprequest() is a void function and in case of error we simply return without any feedback. In case of any local operation or *if_output() failing no feedback is send up the stack for the packet which triggered the arp request to be sent. arpresolve_full() has three pre-canned possible errors returned (if we have not yet sent enough arp requests or if we tried often enough without success) otherwise "no error" is returned. Make arprequest() an "internal" function arprequest_internal() which does return a possible error to the caller. Preserve arprequest() as a void wrapper function for external consumers. In arpresolve_full() add an extra error checking. Use the arprequest_internal() function and only return an error if non of the three ones (mentioend above) are already set. This will return possible errors all the way up the stack and allows functions and programs to react on the send errors rather than leaving them in the dark. Also they might get more detailed feedback of why packets cannot be sent and they will receive it quicker. Reviewed by: karels, hselasky Differential Revision: https://reviews.freebsd.org/D18904
2019-02-24 22:49:56 +00:00
static int
arprequest_internal(struct ifnet *ifp, const struct in_addr *sip,
Add const qualifier to the dst parameter of the ifnet if_output method.
2013-04-26 12:50:32 +00:00
const struct in_addr *tip, u_char *enaddr)
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
{
De-register.
2004-03-14 00:44:11 +00:00
struct mbuf *m;
struct arphdr *ah;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
struct sockaddr sa;
Make it possible to use alternative source hardware address in the ARP datagram generated by arprequest(). If caller doesn't supply the address, then it is either picked from CARP or hardware address of the interface is taken. While here, make several minor fixes: - Hold IF_ADDR_RLOCK(ifp) while traversing address list. - Remove not true comment. - Access internet address and mask via in_ifaddr fields, rather than ifaddr.
2012-01-08 17:25:15 +00:00
u_char *carpaddr = NULL;
Implement interface link header precomputation API. Add if_requestencap() interface method which is capable of calculating various link headers for given interface. Right now there is support for INET/INET6/ARP llheader calculation (IFENCAP_LL type request). Other types are planned to support more complex calculation (L2 multipath lagg nexthops, tunnel encap nexthops, etc..). Reshape 'struct route' to be able to pass additional data (with is length) to prepend to mbuf. These two changes permits routing code to pass pre-calculated nexthop data (like L2 header for route w/gateway) down to the stack eliminating the need for other lookups. It also brings us closer to more complex scenarios like transparently handling MPLS nexthops and tunnel interfaces. Last, but not least, it removes layering violation introduced by flowtable code (ro_lle) and simplifies handling of existing if_output consumers. ARP/ND changes: Make arp/ndp stack pre-calculate link header upon installing/updating lle record. Interface link address change are handled by re-calculating headers for all lles based on if_lladdr event. After these changes, arpresolve()/nd6_resolve() returns full pre-calculated header for supported interfaces thus simplifying if_output(). Move these lookups to separate ether_resolve_addr() function which ether returs error or fully-prepared link header. Add <arp|nd6_>resolve_addr() compat versions to return link addresses instead of pre-calculated data. BPF changes: Raw bpf writes occupied _two_ cases: AF_UNSPEC and pseudo_AF_HDRCMPLT. Despite the naming, both of there have ther header "complete". The only difference is that interface source mac has to be filled by OS for AF_UNSPEC (controlled via BIOCGHDRCMPLT). This logic has to stay inside BPF and not pollute if_output() routines. Convert BPF to pass prepend data via new 'struct route' mechanism. Note that it does not change non-optimized if_output(): ro_prepend handling is purely optional. Side note: hackish pseudo_AF_HDRCMPLT is supported for ethernet and FDDI. It is not needed for ethernet anymore. The only remaining FDDI user is dev/pdq mostly untouched since 2007. FDDI support was eliminated from OpenBSD in 2013 (sys/net/if_fddisubr.c rev 1.65). Flowtable changes: Flowtable violates layering by saving (and not correctly managing) rtes/lles. Instead of passing lle pointer, pass pointer to pre-calculated header data from that lle. Differential Revision: https://reviews.freebsd.org/D4102
2015-12-31 05:03:27 +00:00
uint8_t linkhdr[LLE_MAX_LINKHDR];
size_t linkhdrsize;
struct route ro;
int error;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
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
NET_EPOCH_ASSERT();
This main goals of this project are: 1. separating L2 tables (ARP, NDP) from the L3 routing tables 2. removing as much locking dependencies among these layers as possible to allow for some parallelism in the search operations 3. simplify the logic in the routing code, The most notable end result is the obsolescent of the route cloning (RTF_CLONING) concept, which translated into code reduction in both IPv4 ARP and IPv6 NDP related modules, and size reduction in struct rtentry{}. The change in design obsoletes the semantics of RTF_CLONING, RTF_WASCLONE and RTF_LLINFO routing flags. The userland applications such as "arp" and "ndp" have been modified to reflect those changes. The output from "netstat -r" shows only the routing entries. Quite a few developers have contributed to this project in the past: Glebius Smirnoff, Luigi Rizzo, Alessandro Cerri, and Andre Oppermann. And most recently: - Kip Macy revised the locking code completely, thus completing the last piece of the puzzle, Kip has also been conducting active functional testing - Sam Leffler has helped me improving/refactoring the code, and provided valuable reviews - Julian Elischer setup the perforce tree for me and has helped me maintaining that branch before the svn conversion
2008-12-15 06:10:57 +00:00
if (sip == NULL) {
/*
* The caller did not supply a source address, try to find
* a compatible one among those assigned to this interface.
*/
struct ifaddr *ifa;
ifnet: Replace if_addr_lock rwlock with epoch + mutex Run on LLNW canaries and tested by pho@ gallatin: Using a 14-core, 28-HTT single socket E5-2697 v3 with a 40GbE MLX5 based ConnectX 4-LX NIC, I see an almost 12% improvement in received packet rate, and a larger improvement in bytes delivered all the way to userspace. When the host receiving 64 streams of netperf -H $DUT -t UDP_STREAM -- -m 1, I see, using nstat -I mce0 1 before the patch: InMpps OMpps InGbs OGbs err TCP Est %CPU syscalls csw irq GBfree 4.98 0.00 4.42 0.00 4235592 33 83.80 4720653 2149771 1235 247.32 4.73 0.00 4.20 0.00 4025260 33 82.99 4724900 2139833 1204 247.32 4.72 0.00 4.20 0.00 4035252 33 82.14 4719162 2132023 1264 247.32 4.71 0.00 4.21 0.00 4073206 33 83.68 4744973 2123317 1347 247.32 4.72 0.00 4.21 0.00 4061118 33 80.82 4713615 2188091 1490 247.32 4.72 0.00 4.21 0.00 4051675 33 85.29 4727399 2109011 1205 247.32 4.73 0.00 4.21 0.00 4039056 33 84.65 4724735 2102603 1053 247.32 After the patch InMpps OMpps InGbs OGbs err TCP Est %CPU syscalls csw irq GBfree 5.43 0.00 4.20 0.00 3313143 33 84.96 5434214 1900162 2656 245.51 5.43 0.00 4.20 0.00 3308527 33 85.24 5439695 1809382 2521 245.51 5.42 0.00 4.19 0.00 3316778 33 87.54 5416028 1805835 2256 245.51 5.42 0.00 4.19 0.00 3317673 33 90.44 5426044 1763056 2332 245.51 5.42 0.00 4.19 0.00 3314839 33 88.11 5435732 1792218 2499 245.52 5.44 0.00 4.19 0.00 3293228 33 91.84 5426301 1668597 2121 245.52 Similarly, netperf reports 230Mb/s before the patch, and 270Mb/s after the patch Reviewed by: gallatin Sponsored by: Limelight Networks Differential Revision: https://reviews.freebsd.org/D15366
2018-05-18 20:13:34 +00:00
CK_STAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) {
Make it possible to use alternative source hardware address in the ARP datagram generated by arprequest(). If caller doesn't supply the address, then it is either picked from CARP or hardware address of the interface is taken. While here, make several minor fixes: - Hold IF_ADDR_RLOCK(ifp) while traversing address list. - Remove not true comment. - Access internet address and mask via in_ifaddr fields, rather than ifaddr.
2012-01-08 17:25:15 +00:00
if (ifa->ifa_addr->sa_family != AF_INET)
This main goals of this project are: 1. separating L2 tables (ARP, NDP) from the L3 routing tables 2. removing as much locking dependencies among these layers as possible to allow for some parallelism in the search operations 3. simplify the logic in the routing code, The most notable end result is the obsolescent of the route cloning (RTF_CLONING) concept, which translated into code reduction in both IPv4 ARP and IPv6 NDP related modules, and size reduction in struct rtentry{}. The change in design obsoletes the semantics of RTF_CLONING, RTF_WASCLONE and RTF_LLINFO routing flags. The userland applications such as "arp" and "ndp" have been modified to reflect those changes. The output from "netstat -r" shows only the routing entries. Quite a few developers have contributed to this project in the past: Glebius Smirnoff, Luigi Rizzo, Alessandro Cerri, and Andre Oppermann. And most recently: - Kip Macy revised the locking code completely, thus completing the last piece of the puzzle, Kip has also been conducting active functional testing - Sam Leffler has helped me improving/refactoring the code, and provided valuable reviews - Julian Elischer setup the perforce tree for me and has helped me maintaining that branch before the svn conversion
2008-12-15 06:10:57 +00:00
continue;
Make it possible to use alternative source hardware address in the ARP datagram generated by arprequest(). If caller doesn't supply the address, then it is either picked from CARP or hardware address of the interface is taken. While here, make several minor fixes: - Hold IF_ADDR_RLOCK(ifp) while traversing address list. - Remove not true comment. - Access internet address and mask via in_ifaddr fields, rather than ifaddr.
2012-01-08 17:25:15 +00:00
if (ifa->ifa_carp) {
if ((*carp_iamatch_p)(ifa, &carpaddr) == 0)
continue;
sip = &IA_SIN(ifa)->sin_addr;
} else {
carpaddr = NULL;
sip = &IA_SIN(ifa)->sin_addr;
}
This main goals of this project are: 1. separating L2 tables (ARP, NDP) from the L3 routing tables 2. removing as much locking dependencies among these layers as possible to allow for some parallelism in the search operations 3. simplify the logic in the routing code, The most notable end result is the obsolescent of the route cloning (RTF_CLONING) concept, which translated into code reduction in both IPv4 ARP and IPv6 NDP related modules, and size reduction in struct rtentry{}. The change in design obsoletes the semantics of RTF_CLONING, RTF_WASCLONE and RTF_LLINFO routing flags. The userland applications such as "arp" and "ndp" have been modified to reflect those changes. The output from "netstat -r" shows only the routing entries. Quite a few developers have contributed to this project in the past: Glebius Smirnoff, Luigi Rizzo, Alessandro Cerri, and Andre Oppermann. And most recently: - Kip Macy revised the locking code completely, thus completing the last piece of the puzzle, Kip has also been conducting active functional testing - Sam Leffler has helped me improving/refactoring the code, and provided valuable reviews - Julian Elischer setup the perforce tree for me and has helped me maintaining that branch before the svn conversion
2008-12-15 06:10:57 +00:00
if (0 == ((sip->s_addr ^ tip->s_addr) &
Make it possible to use alternative source hardware address in the ARP datagram generated by arprequest(). If caller doesn't supply the address, then it is either picked from CARP or hardware address of the interface is taken. While here, make several minor fixes: - Hold IF_ADDR_RLOCK(ifp) while traversing address list. - Remove not true comment. - Access internet address and mask via in_ifaddr fields, rather than ifaddr.
2012-01-08 17:25:15 +00:00
IA_MASKSIN(ifa)->sin_addr.s_addr))
This main goals of this project are: 1. separating L2 tables (ARP, NDP) from the L3 routing tables 2. removing as much locking dependencies among these layers as possible to allow for some parallelism in the search operations 3. simplify the logic in the routing code, The most notable end result is the obsolescent of the route cloning (RTF_CLONING) concept, which translated into code reduction in both IPv4 ARP and IPv6 NDP related modules, and size reduction in struct rtentry{}. The change in design obsoletes the semantics of RTF_CLONING, RTF_WASCLONE and RTF_LLINFO routing flags. The userland applications such as "arp" and "ndp" have been modified to reflect those changes. The output from "netstat -r" shows only the routing entries. Quite a few developers have contributed to this project in the past: Glebius Smirnoff, Luigi Rizzo, Alessandro Cerri, and Andre Oppermann. And most recently: - Kip Macy revised the locking code completely, thus completing the last piece of the puzzle, Kip has also been conducting active functional testing - Sam Leffler has helped me improving/refactoring the code, and provided valuable reviews - Julian Elischer setup the perforce tree for me and has helped me maintaining that branch before the svn conversion
2008-12-15 06:10:57 +00:00
break; /* found it. */
}
Some style(9) and whitespace changes. Together with: Andrey Zonov <andrey zonov.org>
2012-07-31 11:31:12 +00:00
if (sip == NULL) {
This main goals of this project are: 1. separating L2 tables (ARP, NDP) from the L3 routing tables 2. removing as much locking dependencies among these layers as possible to allow for some parallelism in the search operations 3. simplify the logic in the routing code, The most notable end result is the obsolescent of the route cloning (RTF_CLONING) concept, which translated into code reduction in both IPv4 ARP and IPv6 NDP related modules, and size reduction in struct rtentry{}. The change in design obsoletes the semantics of RTF_CLONING, RTF_WASCLONE and RTF_LLINFO routing flags. The userland applications such as "arp" and "ndp" have been modified to reflect those changes. The output from "netstat -r" shows only the routing entries. Quite a few developers have contributed to this project in the past: Glebius Smirnoff, Luigi Rizzo, Alessandro Cerri, and Andre Oppermann. And most recently: - Kip Macy revised the locking code completely, thus completing the last piece of the puzzle, Kip has also been conducting active functional testing - Sam Leffler has helped me improving/refactoring the code, and provided valuable reviews - Julian Elischer setup the perforce tree for me and has helped me maintaining that branch before the svn conversion
2008-12-15 06:10:57 +00:00
printf("%s: cannot find matching address\n", __func__);
Make arp code return (more) errors. arprequest() is a void function and in case of error we simply return without any feedback. In case of any local operation or *if_output() failing no feedback is send up the stack for the packet which triggered the arp request to be sent. arpresolve_full() has three pre-canned possible errors returned (if we have not yet sent enough arp requests or if we tried often enough without success) otherwise "no error" is returned. Make arprequest() an "internal" function arprequest_internal() which does return a possible error to the caller. Preserve arprequest() as a void wrapper function for external consumers. In arpresolve_full() add an extra error checking. Use the arprequest_internal() function and only return an error if non of the three ones (mentioend above) are already set. This will return possible errors all the way up the stack and allows functions and programs to react on the send errors rather than leaving them in the dark. Also they might get more detailed feedback of why packets cannot be sent and they will receive it quicker. Reviewed by: karels, hselasky Differential Revision: https://reviews.freebsd.org/D18904
2019-02-24 22:49:56 +00:00
return (EADDRNOTAVAIL);
This main goals of this project are: 1. separating L2 tables (ARP, NDP) from the L3 routing tables 2. removing as much locking dependencies among these layers as possible to allow for some parallelism in the search operations 3. simplify the logic in the routing code, The most notable end result is the obsolescent of the route cloning (RTF_CLONING) concept, which translated into code reduction in both IPv4 ARP and IPv6 NDP related modules, and size reduction in struct rtentry{}. The change in design obsoletes the semantics of RTF_CLONING, RTF_WASCLONE and RTF_LLINFO routing flags. The userland applications such as "arp" and "ndp" have been modified to reflect those changes. The output from "netstat -r" shows only the routing entries. Quite a few developers have contributed to this project in the past: Glebius Smirnoff, Luigi Rizzo, Alessandro Cerri, and Andre Oppermann. And most recently: - Kip Macy revised the locking code completely, thus completing the last piece of the puzzle, Kip has also been conducting active functional testing - Sam Leffler has helped me improving/refactoring the code, and provided valuable reviews - Julian Elischer setup the perforce tree for me and has helped me maintaining that branch before the svn conversion
2008-12-15 06:10:57 +00:00
}
}
Make it possible to use alternative source hardware address in the ARP datagram generated by arprequest(). If caller doesn't supply the address, then it is either picked from CARP or hardware address of the interface is taken. While here, make several minor fixes: - Hold IF_ADDR_RLOCK(ifp) while traversing address list. - Remove not true comment. - Access internet address and mask via in_ifaddr fields, rather than ifaddr.
2012-01-08 17:25:15 +00:00
if (enaddr == NULL)
enaddr = carpaddr ? carpaddr : (u_char *)IF_LLADDR(ifp);
This main goals of this project are: 1. separating L2 tables (ARP, NDP) from the L3 routing tables 2. removing as much locking dependencies among these layers as possible to allow for some parallelism in the search operations 3. simplify the logic in the routing code, The most notable end result is the obsolescent of the route cloning (RTF_CLONING) concept, which translated into code reduction in both IPv4 ARP and IPv6 NDP related modules, and size reduction in struct rtentry{}. The change in design obsoletes the semantics of RTF_CLONING, RTF_WASCLONE and RTF_LLINFO routing flags. The userland applications such as "arp" and "ndp" have been modified to reflect those changes. The output from "netstat -r" shows only the routing entries. Quite a few developers have contributed to this project in the past: Glebius Smirnoff, Luigi Rizzo, Alessandro Cerri, and Andre Oppermann. And most recently: - Kip Macy revised the locking code completely, thus completing the last piece of the puzzle, Kip has also been conducting active functional testing - Sam Leffler has helped me improving/refactoring the code, and provided valuable reviews - Julian Elischer setup the perforce tree for me and has helped me maintaining that branch before the svn conversion
2008-12-15 06:10:57 +00:00
Mechanically substitute flags from historic mbuf allocator with malloc(9) flags within sys. Exceptions: - sys/contrib not touched - sys/mbuf.h edited manually
2012-12-05 08:04:20 +00:00
if ((m = m_gethdr(M_NOWAIT, MT_DATA)) == NULL)
Make arp code return (more) errors. arprequest() is a void function and in case of error we simply return without any feedback. In case of any local operation or *if_output() failing no feedback is send up the stack for the packet which triggered the arp request to be sent. arpresolve_full() has three pre-canned possible errors returned (if we have not yet sent enough arp requests or if we tried often enough without success) otherwise "no error" is returned. Make arprequest() an "internal" function arprequest_internal() which does return a possible error to the caller. Preserve arprequest() as a void wrapper function for external consumers. In arpresolve_full() add an extra error checking. Use the arprequest_internal() function and only return an error if non of the three ones (mentioend above) are already set. This will return possible errors all the way up the stack and allows functions and programs to react on the send errors rather than leaving them in the dark. Also they might get more detailed feedback of why packets cannot be sent and they will receive it quicker. Reviewed by: karels, hselasky Differential Revision: https://reviews.freebsd.org/D18904
2019-02-24 22:49:56 +00:00
return (ENOMEM);
Use macros instead of referencing struct if_data that resides in ifnet. Sponsored by: Nginx, Inc.
2014-08-31 06:30:50 +00:00
m->m_len = sizeof(*ah) + 2 * sizeof(struct in_addr) +
2 * ifp->if_addrlen;
- Fix indentation lost by 'diff -b'. - Un-wrap short line.
2004-03-21 18:51:26 +00:00
m->m_pkthdr.len = m->m_len;
To ease changes to underlying mbuf structure and the mbuf allocator, reduce the knowledge of mbuf layout, and in particular constants such as M_EXT, MLEN, MHLEN, and so on, in mbuf consumers by unifying various alignment utility functions (M_ALIGN(), MH_ALIGN(), MEXT_ALIGN() in a single M_ALIGN() macro, implemented by a now-inlined m_align() function: - Move m_align() from uipc_mbuf.c to mbuf.h; mark as __inline. - Reimplement M_ALIGN(), MH_ALIGN(), and MEXT_ALIGN() using m_align(). - Update consumers around the tree to simply use M_ALIGN(). This change eliminates a number of cases where mbuf consumers must be aware of whether or not mbufs returned by the allocator use external storage, but also assumptions about the size of the returned mbuf. This will make it easier to introduce changes in how we use external storage, as well as features such as variable-size mbufs. Differential Revision: https://reviews.freebsd.org/D1436 Reviewed by: glebius, trasz, gnn, bz Sponsored by: EMC / Isilon Storage Division
2015-01-05 09:58:32 +00:00
M_ALIGN(m, m->m_len);
- Fix indentation lost by 'diff -b'. - Un-wrap short line.
2004-03-21 18:51:26 +00:00
ah = mtod(m, struct arphdr *);
Remove interface type specific code from arprequest(), and in_arpinput(). The AF_ARP case in the (*if_output)() routine will handle the interface type specific bits. Obtained from: NetBSD
2004-03-21 06:36:05 +00:00
bzero((caddr_t)ah, m->m_len);
#ifdef MAC
Move towards more explicit support for various network protocol stacks in the TrustedBSD MAC Framework: - Add mac_atalk.c and add explicit entry point mac_netatalk_aarp_send() for AARP packet labeling, rather than using a generic link layer entry point. - Add mac_inet6.c and add explicit entry point mac_netinet6_nd6_send() for ND6 packet labeling, rather than using a generic link layer entry point. - Add expliict entry point mac_netinet_arp_send() for ARP packet labeling, and mac_netinet_igmp_send() for IGMP packet labeling, rather than using a generic link layer entry point. - Remove previous genering link layer entry point, mac_mbuf_create_linklayer() as it is no longer used. - Add implementations of new entry points to various policies, largely by replicating the existing link layer entry point for them; remove old link layer entry point implementation. - Make MAC_IFNET_LOCK(), MAC_IFNET_UNLOCK(), and mac_ifnet_mtx global to the MAC Framework rather than static to mac_net.c as it is now needed outside of mac_net.c. Obtained from: TrustedBSD Project
2007-10-28 15:55:23 +00:00
mac_netinet_arp_send(ifp, m);
Remove interface type specific code from arprequest(), and in_arpinput(). The AF_ARP case in the (*if_output)() routine will handle the interface type specific bits. Obtained from: NetBSD
2004-03-21 06:36:05 +00:00
#endif
bring in ARP support for variable length link level addresses Reviewed by: jdp Approved by: jdp Obtained from: NetBSD MFC after: 6 weeks
2001-10-14 20:17:53 +00:00
ah->ar_pro = htons(ETHERTYPE_IP);
ah->ar_hln = ifp->if_addrlen; /* hardware address length */
ah->ar_pln = sizeof(struct in_addr); /* protocol address length */
ah->ar_op = htons(ARPOP_REQUEST);
Add const qualifier to the dst parameter of the ifnet if_output method.
2013-04-26 12:50:32 +00:00
bcopy(enaddr, ar_sha(ah), ah->ar_hln);
bcopy(sip, ar_spa(ah), ah->ar_pln);
bcopy(tip, ar_tpa(ah), ah->ar_pln);
Remove interface type specific code from arprequest(), and in_arpinput(). The AF_ARP case in the (*if_output)() routine will handle the interface type specific bits. Obtained from: NetBSD
2004-03-21 06:36:05 +00:00
sa.sa_family = AF_ARP;
sa.sa_len = 2;
Implement interface link header precomputation API. Add if_requestencap() interface method which is capable of calculating various link headers for given interface. Right now there is support for INET/INET6/ARP llheader calculation (IFENCAP_LL type request). Other types are planned to support more complex calculation (L2 multipath lagg nexthops, tunnel encap nexthops, etc..). Reshape 'struct route' to be able to pass additional data (with is length) to prepend to mbuf. These two changes permits routing code to pass pre-calculated nexthop data (like L2 header for route w/gateway) down to the stack eliminating the need for other lookups. It also brings us closer to more complex scenarios like transparently handling MPLS nexthops and tunnel interfaces. Last, but not least, it removes layering violation introduced by flowtable code (ro_lle) and simplifies handling of existing if_output consumers. ARP/ND changes: Make arp/ndp stack pre-calculate link header upon installing/updating lle record. Interface link address change are handled by re-calculating headers for all lles based on if_lladdr event. After these changes, arpresolve()/nd6_resolve() returns full pre-calculated header for supported interfaces thus simplifying if_output(). Move these lookups to separate ether_resolve_addr() function which ether returs error or fully-prepared link header. Add <arp|nd6_>resolve_addr() compat versions to return link addresses instead of pre-calculated data. BPF changes: Raw bpf writes occupied _two_ cases: AF_UNSPEC and pseudo_AF_HDRCMPLT. Despite the naming, both of there have ther header "complete". The only difference is that interface source mac has to be filled by OS for AF_UNSPEC (controlled via BIOCGHDRCMPLT). This logic has to stay inside BPF and not pollute if_output() routines. Convert BPF to pass prepend data via new 'struct route' mechanism. Note that it does not change non-optimized if_output(): ro_prepend handling is purely optional. Side note: hackish pseudo_AF_HDRCMPLT is supported for ethernet and FDDI. It is not needed for ethernet anymore. The only remaining FDDI user is dev/pdq mostly untouched since 2007. FDDI support was eliminated from OpenBSD in 2013 (sys/net/if_fddisubr.c rev 1.65). Flowtable changes: Flowtable violates layering by saving (and not correctly managing) rtes/lles. Instead of passing lle pointer, pass pointer to pre-calculated header data from that lle. Differential Revision: https://reviews.freebsd.org/D4102
2015-12-31 05:03:27 +00:00
/* Calculate link header for sending frame */
bzero(&ro, sizeof(ro));
linkhdrsize = sizeof(linkhdr);
error = arp_fillheader(ifp, ah, 1, linkhdr, &linkhdrsize);
if (error != 0 && error != EAFNOSUPPORT) {
ARP_LOG(LOG_ERR, "Failed to calculate ARP header on %s: %d\n",
if_name(ifp), error);
Make arp code return (more) errors. arprequest() is a void function and in case of error we simply return without any feedback. In case of any local operation or *if_output() failing no feedback is send up the stack for the packet which triggered the arp request to be sent. arpresolve_full() has three pre-canned possible errors returned (if we have not yet sent enough arp requests or if we tried often enough without success) otherwise "no error" is returned. Make arprequest() an "internal" function arprequest_internal() which does return a possible error to the caller. Preserve arprequest() as a void wrapper function for external consumers. In arpresolve_full() add an extra error checking. Use the arprequest_internal() function and only return an error if non of the three ones (mentioend above) are already set. This will return possible errors all the way up the stack and allows functions and programs to react on the send errors rather than leaving them in the dark. Also they might get more detailed feedback of why packets cannot be sent and they will receive it quicker. Reviewed by: karels, hselasky Differential Revision: https://reviews.freebsd.org/D18904
2019-02-24 22:49:56 +00:00
return (error);
Implement interface link header precomputation API. Add if_requestencap() interface method which is capable of calculating various link headers for given interface. Right now there is support for INET/INET6/ARP llheader calculation (IFENCAP_LL type request). Other types are planned to support more complex calculation (L2 multipath lagg nexthops, tunnel encap nexthops, etc..). Reshape 'struct route' to be able to pass additional data (with is length) to prepend to mbuf. These two changes permits routing code to pass pre-calculated nexthop data (like L2 header for route w/gateway) down to the stack eliminating the need for other lookups. It also brings us closer to more complex scenarios like transparently handling MPLS nexthops and tunnel interfaces. Last, but not least, it removes layering violation introduced by flowtable code (ro_lle) and simplifies handling of existing if_output consumers. ARP/ND changes: Make arp/ndp stack pre-calculate link header upon installing/updating lle record. Interface link address change are handled by re-calculating headers for all lles based on if_lladdr event. After these changes, arpresolve()/nd6_resolve() returns full pre-calculated header for supported interfaces thus simplifying if_output(). Move these lookups to separate ether_resolve_addr() function which ether returs error or fully-prepared link header. Add <arp|nd6_>resolve_addr() compat versions to return link addresses instead of pre-calculated data. BPF changes: Raw bpf writes occupied _two_ cases: AF_UNSPEC and pseudo_AF_HDRCMPLT. Despite the naming, both of there have ther header "complete". The only difference is that interface source mac has to be filled by OS for AF_UNSPEC (controlled via BIOCGHDRCMPLT). This logic has to stay inside BPF and not pollute if_output() routines. Convert BPF to pass prepend data via new 'struct route' mechanism. Note that it does not change non-optimized if_output(): ro_prepend handling is purely optional. Side note: hackish pseudo_AF_HDRCMPLT is supported for ethernet and FDDI. It is not needed for ethernet anymore. The only remaining FDDI user is dev/pdq mostly untouched since 2007. FDDI support was eliminated from OpenBSD in 2013 (sys/net/if_fddisubr.c rev 1.65). Flowtable changes: Flowtable violates layering by saving (and not correctly managing) rtes/lles. Instead of passing lle pointer, pass pointer to pre-calculated header data from that lle. Differential Revision: https://reviews.freebsd.org/D4102
2015-12-31 05:03:27 +00:00
}
ro.ro_prepend = linkhdr;
ro.ro_plen = linkhdrsize;
ro.ro_flags = 0;
Remove interface type specific code from arprequest(), and in_arpinput(). The AF_ARP case in the (*if_output)() routine will handle the interface type specific bits. Obtained from: NetBSD
2004-03-21 06:36:05 +00:00
m->m_flags |= M_BCAST;
Add m_clrprotoflags() to clear protocol specific mbuf flags at up and downwards layer crossings. Consistently use it within IP, IPv6 and ethernet protocols. Discussed with: trociny, glebius
2013-08-19 13:27:32 +00:00
m_clrprotoflags(m); /* Avoid confusing lower layers. */
Make arp code return (more) errors. arprequest() is a void function and in case of error we simply return without any feedback. In case of any local operation or *if_output() failing no feedback is send up the stack for the packet which triggered the arp request to be sent. arpresolve_full() has three pre-canned possible errors returned (if we have not yet sent enough arp requests or if we tried often enough without success) otherwise "no error" is returned. Make arprequest() an "internal" function arprequest_internal() which does return a possible error to the caller. Preserve arprequest() as a void wrapper function for external consumers. In arpresolve_full() add an extra error checking. Use the arprequest_internal() function and only return an error if non of the three ones (mentioend above) are already set. This will return possible errors all the way up the stack and allows functions and programs to react on the send errors rather than leaving them in the dark. Also they might get more detailed feedback of why packets cannot be sent and they will receive it quicker. Reviewed by: karels, hselasky Differential Revision: https://reviews.freebsd.org/D18904
2019-02-24 22:49:56 +00:00
error = (*ifp->if_output)(ifp, m, &sa, &ro);
Add ARP statistics to the kernel and netstat. New counters now exist for: requests sent replies sent requests received replies received packets received total packets dropped due to no ARP entry entrys timed out Duplicate IPs seen The new statistics are seen in the netstat command when it is given the -s command line switch. MFC after: 2 weeks In collaboration with: bz
2009-09-03 21:10:57 +00:00
ARPSTAT_INC(txrequests);
Improve ARP logging. r344504 added an extra ARP_LOG() call in case of an if_output() failure. It turns out IPv4 can be noisy. In order to not spam the console by default: (a) add a counter for these events so people can keep better track of how often it happens, and (b) add a sysctl to select the default ARP_LOG log level and set it to INFO avoiding the one (the new) DEBUG level by default. Claim a spare (1st one after 10 years since the stats were added) in order to not break netstat from FreeBSD 12->13 updates in the future. Reviewed by: karels Differential Revision: https://reviews.freebsd.org/D19490
2019-03-09 01:12:59 +00:00
if (error) {
ARPSTAT_INC(txerrors);
Make arp code return (more) errors. arprequest() is a void function and in case of error we simply return without any feedback. In case of any local operation or *if_output() failing no feedback is send up the stack for the packet which triggered the arp request to be sent. arpresolve_full() has three pre-canned possible errors returned (if we have not yet sent enough arp requests or if we tried often enough without success) otherwise "no error" is returned. Make arprequest() an "internal" function arprequest_internal() which does return a possible error to the caller. Preserve arprequest() as a void wrapper function for external consumers. In arpresolve_full() add an extra error checking. Use the arprequest_internal() function and only return an error if non of the three ones (mentioend above) are already set. This will return possible errors all the way up the stack and allows functions and programs to react on the send errors rather than leaving them in the dark. Also they might get more detailed feedback of why packets cannot be sent and they will receive it quicker. Reviewed by: karels, hselasky Differential Revision: https://reviews.freebsd.org/D18904
2019-02-24 22:49:56 +00:00
ARP_LOG(LOG_DEBUG, "Failed to send ARP packet on %s: %d\n",
if_name(ifp), error);
Improve ARP logging. r344504 added an extra ARP_LOG() call in case of an if_output() failure. It turns out IPv4 can be noisy. In order to not spam the console by default: (a) add a counter for these events so people can keep better track of how often it happens, and (b) add a sysctl to select the default ARP_LOG log level and set it to INFO avoiding the one (the new) DEBUG level by default. Claim a spare (1st one after 10 years since the stats were added) in order to not break netstat from FreeBSD 12->13 updates in the future. Reviewed by: karels Differential Revision: https://reviews.freebsd.org/D19490
2019-03-09 01:12:59 +00:00
}
Make arp code return (more) errors. arprequest() is a void function and in case of error we simply return without any feedback. In case of any local operation or *if_output() failing no feedback is send up the stack for the packet which triggered the arp request to be sent. arpresolve_full() has three pre-canned possible errors returned (if we have not yet sent enough arp requests or if we tried often enough without success) otherwise "no error" is returned. Make arprequest() an "internal" function arprequest_internal() which does return a possible error to the caller. Preserve arprequest() as a void wrapper function for external consumers. In arpresolve_full() add an extra error checking. Use the arprequest_internal() function and only return an error if non of the three ones (mentioend above) are already set. This will return possible errors all the way up the stack and allows functions and programs to react on the send errors rather than leaving them in the dark. Also they might get more detailed feedback of why packets cannot be sent and they will receive it quicker. Reviewed by: karels, hselasky Differential Revision: https://reviews.freebsd.org/D18904
2019-02-24 22:49:56 +00:00
return (error);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
Make arp code return (more) errors. arprequest() is a void function and in case of error we simply return without any feedback. In case of any local operation or *if_output() failing no feedback is send up the stack for the packet which triggered the arp request to be sent. arpresolve_full() has three pre-canned possible errors returned (if we have not yet sent enough arp requests or if we tried often enough without success) otherwise "no error" is returned. Make arprequest() an "internal" function arprequest_internal() which does return a possible error to the caller. Preserve arprequest() as a void wrapper function for external consumers. In arpresolve_full() add an extra error checking. Use the arprequest_internal() function and only return an error if non of the three ones (mentioend above) are already set. This will return possible errors all the way up the stack and allows functions and programs to react on the send errors rather than leaving them in the dark. Also they might get more detailed feedback of why packets cannot be sent and they will receive it quicker. Reviewed by: karels, hselasky Differential Revision: https://reviews.freebsd.org/D18904
2019-02-24 22:49:56 +00:00
void
arprequest(struct ifnet *ifp, const struct in_addr *sip,
const struct in_addr *tip, u_char *enaddr)
{
(void) arprequest_internal(ifp, sip, tip, enaddr);
}
Implement interface link header precomputation API. Add if_requestencap() interface method which is capable of calculating various link headers for given interface. Right now there is support for INET/INET6/ARP llheader calculation (IFENCAP_LL type request). Other types are planned to support more complex calculation (L2 multipath lagg nexthops, tunnel encap nexthops, etc..). Reshape 'struct route' to be able to pass additional data (with is length) to prepend to mbuf. These two changes permits routing code to pass pre-calculated nexthop data (like L2 header for route w/gateway) down to the stack eliminating the need for other lookups. It also brings us closer to more complex scenarios like transparently handling MPLS nexthops and tunnel interfaces. Last, but not least, it removes layering violation introduced by flowtable code (ro_lle) and simplifies handling of existing if_output consumers. ARP/ND changes: Make arp/ndp stack pre-calculate link header upon installing/updating lle record. Interface link address change are handled by re-calculating headers for all lles based on if_lladdr event. After these changes, arpresolve()/nd6_resolve() returns full pre-calculated header for supported interfaces thus simplifying if_output(). Move these lookups to separate ether_resolve_addr() function which ether returs error or fully-prepared link header. Add <arp|nd6_>resolve_addr() compat versions to return link addresses instead of pre-calculated data. BPF changes: Raw bpf writes occupied _two_ cases: AF_UNSPEC and pseudo_AF_HDRCMPLT. Despite the naming, both of there have ther header "complete". The only difference is that interface source mac has to be filled by OS for AF_UNSPEC (controlled via BIOCGHDRCMPLT). This logic has to stay inside BPF and not pollute if_output() routines. Convert BPF to pass prepend data via new 'struct route' mechanism. Note that it does not change non-optimized if_output(): ro_prepend handling is purely optional. Side note: hackish pseudo_AF_HDRCMPLT is supported for ethernet and FDDI. It is not needed for ethernet anymore. The only remaining FDDI user is dev/pdq mostly untouched since 2007. FDDI support was eliminated from OpenBSD in 2013 (sys/net/if_fddisubr.c rev 1.65). Flowtable changes: Flowtable violates layering by saving (and not correctly managing) rtes/lles. Instead of passing lle pointer, pass pointer to pre-calculated header data from that lle. Differential Revision: https://reviews.freebsd.org/D4102
2015-12-31 05:03:27 +00:00
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
/*
Split arpresolve() into fast/slow path. This change isolates the most common case (e.g. successful lookup) from more complicates scenarios. It also (tries to) make code more simple by avoiding retry: cycle. The actual goal is to prepare code to the upcoming change that will allow LL address retrieval without acquiring LLE lock at all. Reviewed by: ae Differential Revision: https://reviews.freebsd.org/D3383
2015-08-16 12:23:58 +00:00
* Resolve an IP address into an ethernet address - heavy version.
* Used internally by arpresolve().
Improve the comment for arpresolve_full() in if_ether.c. No functional changes. MFC after: 6 weeks
2018-11-17 16:13:09 +00:00
* We have already checked that we can't use an existing lle without
* modification so we have to acquire an LLE_EXCLUSIVE lle lock.
This commit does two things: 1. rt_check() cleanup: rt_check() is only necessary for some address families to gain access to the corresponding arp entry, so call it only in/near the *resolve() routines where it is actually used -- at the moment this is arpresolve(), nd6_storelladdr() (the call is embedded here), and atmresolve() (the call is just before atmresolve to reduce the number of changes). This change will make it a lot easier to decouple the arp table from the routing table. There is an extra call to rt_check() in if_iso88025subr.c to determine the routing info length. I have left it alone for the time being. The interface of arpresolve() and nd6_storelladdr() now changes slightly: + the 'rtentry' parameter (really a hint from the upper level layer) is now passed unchanged from *_output(), so it becomes the route to the final destination and not to the gateway. + the routines will return 0 if resolution is possible, non-zero otherwise. + arpresolve() returns EWOULDBLOCK in case the mbuf is being held waiting for an arp reply -- in this case the error code is masked in the caller so the upper layer protocol will not see a failure. 2. arpcom untangling Where possible, use 'struct ifnet' instead of 'struct arpcom' variables, and use the IFP2AC macro to access arpcom fields. This mostly affects the netatalk code. === Detailed changes: === net/if_arcsubr.c rt_check() cleanup, remove a useless variable net/if_atmsubr.c rt_check() cleanup net/if_ethersubr.c rt_check() cleanup, arpcom untangling net/if_fddisubr.c rt_check() cleanup, arpcom untangling net/if_iso88025subr.c rt_check() cleanup netatalk/aarp.c arpcom untangling, remove a block of duplicated code netatalk/at_extern.h arpcom untangling netinet/if_ether.c rt_check() cleanup (change arpresolve) netinet6/nd6.c rt_check() cleanup (change nd6_storelladdr)
2004-04-25 09:24:52 +00:00
*
Improve the comment for arpresolve_full() in if_ether.c. No functional changes. MFC after: 6 weeks
2018-11-17 16:13:09 +00:00
* On success, desten and pflags are filled in and the function returns 0;
This commit does two things: 1. rt_check() cleanup: rt_check() is only necessary for some address families to gain access to the corresponding arp entry, so call it only in/near the *resolve() routines where it is actually used -- at the moment this is arpresolve(), nd6_storelladdr() (the call is embedded here), and atmresolve() (the call is just before atmresolve to reduce the number of changes). This change will make it a lot easier to decouple the arp table from the routing table. There is an extra call to rt_check() in if_iso88025subr.c to determine the routing info length. I have left it alone for the time being. The interface of arpresolve() and nd6_storelladdr() now changes slightly: + the 'rtentry' parameter (really a hint from the upper level layer) is now passed unchanged from *_output(), so it becomes the route to the final destination and not to the gateway. + the routines will return 0 if resolution is possible, non-zero otherwise. + arpresolve() returns EWOULDBLOCK in case the mbuf is being held waiting for an arp reply -- in this case the error code is masked in the caller so the upper layer protocol will not see a failure. 2. arpcom untangling Where possible, use 'struct ifnet' instead of 'struct arpcom' variables, and use the IFP2AC macro to access arpcom fields. This mostly affects the netatalk code. === Detailed changes: === net/if_arcsubr.c rt_check() cleanup, remove a useless variable net/if_atmsubr.c rt_check() cleanup net/if_ethersubr.c rt_check() cleanup, arpcom untangling net/if_fddisubr.c rt_check() cleanup, arpcom untangling net/if_iso88025subr.c rt_check() cleanup netatalk/aarp.c arpcom untangling, remove a block of duplicated code netatalk/at_extern.h arpcom untangling netinet/if_ether.c rt_check() cleanup (change arpresolve) netinet6/nd6.c rt_check() cleanup (change nd6_storelladdr)
2004-04-25 09:24:52 +00:00
* If the packet must be held pending resolution, we return EWOULDBLOCK
* On other errors, we return the corresponding error code.
Don't duplicate the whole of arpresolve to arpresolve 2 for the sake of two compares against 0. The negative effect of cache flushing is probably more than the gain by not doing the two compares (the value is almost certainly in register or at worst, cache). Note that the uses of m_freem() are in error cases and m_freem() handles NULL anyhow. So fast-path really isn't changed much at all.
2007-12-31 23:48:06 +00:00
* Note that m_freem() handles NULL.
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
*/
Split arpresolve() into fast/slow path. This change isolates the most common case (e.g. successful lookup) from more complicates scenarios. It also (tries to) make code more simple by avoiding retry: cycle. The actual goal is to prepare code to the upcoming change that will allow LL address retrieval without acquiring LLE lock at all. Reviewed by: ae Differential Revision: https://reviews.freebsd.org/D3383
2015-08-16 12:23:58 +00:00
static int
Implement interface link header precomputation API. Add if_requestencap() interface method which is capable of calculating various link headers for given interface. Right now there is support for INET/INET6/ARP llheader calculation (IFENCAP_LL type request). Other types are planned to support more complex calculation (L2 multipath lagg nexthops, tunnel encap nexthops, etc..). Reshape 'struct route' to be able to pass additional data (with is length) to prepend to mbuf. These two changes permits routing code to pass pre-calculated nexthop data (like L2 header for route w/gateway) down to the stack eliminating the need for other lookups. It also brings us closer to more complex scenarios like transparently handling MPLS nexthops and tunnel interfaces. Last, but not least, it removes layering violation introduced by flowtable code (ro_lle) and simplifies handling of existing if_output consumers. ARP/ND changes: Make arp/ndp stack pre-calculate link header upon installing/updating lle record. Interface link address change are handled by re-calculating headers for all lles based on if_lladdr event. After these changes, arpresolve()/nd6_resolve() returns full pre-calculated header for supported interfaces thus simplifying if_output(). Move these lookups to separate ether_resolve_addr() function which ether returs error or fully-prepared link header. Add <arp|nd6_>resolve_addr() compat versions to return link addresses instead of pre-calculated data. BPF changes: Raw bpf writes occupied _two_ cases: AF_UNSPEC and pseudo_AF_HDRCMPLT. Despite the naming, both of there have ther header "complete". The only difference is that interface source mac has to be filled by OS for AF_UNSPEC (controlled via BIOCGHDRCMPLT). This logic has to stay inside BPF and not pollute if_output() routines. Convert BPF to pass prepend data via new 'struct route' mechanism. Note that it does not change non-optimized if_output(): ro_prepend handling is purely optional. Side note: hackish pseudo_AF_HDRCMPLT is supported for ethernet and FDDI. It is not needed for ethernet anymore. The only remaining FDDI user is dev/pdq mostly untouched since 2007. FDDI support was eliminated from OpenBSD in 2013 (sys/net/if_fddisubr.c rev 1.65). Flowtable changes: Flowtable violates layering by saving (and not correctly managing) rtes/lles. Instead of passing lle pointer, pass pointer to pre-calculated header data from that lle. Differential Revision: https://reviews.freebsd.org/D4102
2015-12-31 05:03:27 +00:00
arpresolve_full(struct ifnet *ifp, int is_gw, int flags, struct mbuf *m,
This change re-adds L2 caching for TCP and UDP, as originally added in D4306 but removed due to other changes in the system. Restore the llentry pointer to the "struct route", and use it to cache the L2 lookup (ARP or ND6) as appropriate. Submitted by: Mike Karels Differential Revision: https://reviews.freebsd.org/D6262
2016-06-02 17:51:29 +00:00
const struct sockaddr *dst, u_char *desten, uint32_t *pflags,
struct llentry **plle)
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
{
* Split allocation and table linking for lle's. Before that, the logic besides lle_create() was the following: return existing if found, create if not. This behaviour was error-prone since we had to deal with 'sudden' static<>dynamic lle changes. This commit fixes bunch of different issues like: - refcount leak when lle is converted to static. Simple check case: console 1: while true; do for i in `arp -an|awk '$4~/incomp/{print$2}'|tr -d '()'`; do arp -s $i 00:22:44:66:88:00 ; arp -d $i; done; done console 2: ping -f any-dead-host-in-L2 console 3: # watch for memory consumption: vmstat -m | awk '$1~/lltable/{print$2}' - possible problems in arptimer() / nd6_timer() when dropping/reacquiring lock. New logic explicitly handles use-or-create cases in every lla_create user. Basically, most of the changes are purely mechanical. However, we explicitly avoid using existing lle's for interface/static LLE records. * While here, call lle_event handlers on all real table lle change. * Create lltable_free_entry() calling existing per-lltable lle_free_t callback for entry deletion
2015-08-20 12:05:17 +00:00
struct llentry *la = NULL, *la_tmp;
Add a queue to hold packets while we await an ARP reply. When a fast machine first brings up some non TCP networking program it is quite possible that we will drop packets due to the fact that only one packet can be held per ARP entry. This leads to packets being missed when a program starts or restarts if the ARP data is not currently in the ARP cache. This code adds a new sysctl, net.link.ether.inet.maxhold, which defines a system wide maximum number of packets to be held in each ARP entry. Up to maxhold packets are queued until an ARP reply is received or the ARP times out. The default setting is the old value of 1 which has been part of the BSD networking code since time immemorial. Expose the time we hold an incomplete ARP entry by adding the sysctl net.link.ether.inet.wait, which defaults to 20 seconds, the value used when the new ARP code was added.. Reviewed by: bz, rpaulo MFC after: 3 weeks
2010-11-12 22:03:02 +00:00
struct mbuf *curr = NULL;
struct mbuf *next = NULL;
Split arpresolve() into fast/slow path. This change isolates the most common case (e.g. successful lookup) from more complicates scenarios. It also (tries to) make code more simple by avoiding retry: cycle. The actual goal is to prepare code to the upcoming change that will allow LL address retrieval without acquiring LLE lock at all. Reviewed by: ae Differential Revision: https://reviews.freebsd.org/D3383
2015-08-16 12:23:58 +00:00
int error, renew;
Implement interface link header precomputation API. Add if_requestencap() interface method which is capable of calculating various link headers for given interface. Right now there is support for INET/INET6/ARP llheader calculation (IFENCAP_LL type request). Other types are planned to support more complex calculation (L2 multipath lagg nexthops, tunnel encap nexthops, etc..). Reshape 'struct route' to be able to pass additional data (with is length) to prepend to mbuf. These two changes permits routing code to pass pre-calculated nexthop data (like L2 header for route w/gateway) down to the stack eliminating the need for other lookups. It also brings us closer to more complex scenarios like transparently handling MPLS nexthops and tunnel interfaces. Last, but not least, it removes layering violation introduced by flowtable code (ro_lle) and simplifies handling of existing if_output consumers. ARP/ND changes: Make arp/ndp stack pre-calculate link header upon installing/updating lle record. Interface link address change are handled by re-calculating headers for all lles based on if_lladdr event. After these changes, arpresolve()/nd6_resolve() returns full pre-calculated header for supported interfaces thus simplifying if_output(). Move these lookups to separate ether_resolve_addr() function which ether returs error or fully-prepared link header. Add <arp|nd6_>resolve_addr() compat versions to return link addresses instead of pre-calculated data. BPF changes: Raw bpf writes occupied _two_ cases: AF_UNSPEC and pseudo_AF_HDRCMPLT. Despite the naming, both of there have ther header "complete". The only difference is that interface source mac has to be filled by OS for AF_UNSPEC (controlled via BIOCGHDRCMPLT). This logic has to stay inside BPF and not pollute if_output() routines. Convert BPF to pass prepend data via new 'struct route' mechanism. Note that it does not change non-optimized if_output(): ro_prepend handling is purely optional. Side note: hackish pseudo_AF_HDRCMPLT is supported for ethernet and FDDI. It is not needed for ethernet anymore. The only remaining FDDI user is dev/pdq mostly untouched since 2007. FDDI support was eliminated from OpenBSD in 2013 (sys/net/if_fddisubr.c rev 1.65). Flowtable changes: Flowtable violates layering by saving (and not correctly managing) rtes/lles. Instead of passing lle pointer, pass pointer to pre-calculated header data from that lle. Differential Revision: https://reviews.freebsd.org/D4102
2015-12-31 05:03:27 +00:00
char *lladdr;
int ll_len;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
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
NET_EPOCH_ASSERT();
Do not return unlocked/unreferenced lle in arpresolve/nd6_storelladdr - return lle flags IFF needed. Do not pass rte to arpresolve - pass is_gateway flag instead.
2014-11-27 23:06:25 +00:00
if (pflags != NULL)
*pflags = 0;
This change re-adds L2 caching for TCP and UDP, as originally added in D4306 but removed due to other changes in the system. Restore the llentry pointer to the "struct route", and use it to cache the L2 lookup (ARP or ND6) as appropriate. Submitted by: Mike Karels Differential Revision: https://reviews.freebsd.org/D6262
2016-06-02 17:51:29 +00:00
if (plle != NULL)
*plle = NULL;
Do not return unlocked/unreferenced lle in arpresolve/nd6_storelladdr - return lle flags IFF needed. Do not pass rte to arpresolve - pass is_gateway flag instead.
2014-11-27 23:06:25 +00:00
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
if ((flags & LLE_CREATE) == 0)
Split arpresolve() into fast/slow path. This change isolates the most common case (e.g. successful lookup) from more complicates scenarios. It also (tries to) make code more simple by avoiding retry: cycle. The actual goal is to prepare code to the upcoming change that will allow LL address retrieval without acquiring LLE lock at all. Reviewed by: ae Differential Revision: https://reviews.freebsd.org/D3383
2015-08-16 12:23:58 +00:00
la = lla_lookup(LLTABLE(ifp), LLE_EXCLUSIVE, dst);
if (la == NULL && (ifp->if_flags & (IFF_NOARP | IFF_STATICARP)) == 0) {
* Split allocation and table linking for lle's. Before that, the logic besides lle_create() was the following: return existing if found, create if not. This behaviour was error-prone since we had to deal with 'sudden' static<>dynamic lle changes. This commit fixes bunch of different issues like: - refcount leak when lle is converted to static. Simple check case: console 1: while true; do for i in `arp -an|awk '$4~/incomp/{print$2}'|tr -d '()'`; do arp -s $i 00:22:44:66:88:00 ; arp -d $i; done; done console 2: ping -f any-dead-host-in-L2 console 3: # watch for memory consumption: vmstat -m | awk '$1~/lltable/{print$2}' - possible problems in arptimer() / nd6_timer() when dropping/reacquiring lock. New logic explicitly handles use-or-create cases in every lla_create user. Basically, most of the changes are purely mechanical. However, we explicitly avoid using existing lle's for interface/static LLE records. * While here, call lle_event handlers on all real table lle change. * Create lltable_free_entry() calling existing per-lltable lle_free_t callback for entry deletion
2015-08-20 12:05:17 +00:00
la = lltable_alloc_entry(LLTABLE(ifp), 0, dst);
if (la == NULL) {
Use inet_ntoa_r() instead of inet_ntoa() throughout the kernel inet_ntoa() cannot be used safely in a multithreaded environment because it uses a static local buffer. Instead, use inet_ntoa_r() with a buffer on the caller's stack. Suggested by: glebius, emaste Reviewed by: gnn MFC after: 2 weeks Sponsored by: Dell EMC Differential Revision: https://reviews.freebsd.org/D9625
2017-02-16 20:47:41 +00:00
char addrbuf[INET_ADDRSTRLEN];
* Split allocation and table linking for lle's. Before that, the logic besides lle_create() was the following: return existing if found, create if not. This behaviour was error-prone since we had to deal with 'sudden' static<>dynamic lle changes. This commit fixes bunch of different issues like: - refcount leak when lle is converted to static. Simple check case: console 1: while true; do for i in `arp -an|awk '$4~/incomp/{print$2}'|tr -d '()'`; do arp -s $i 00:22:44:66:88:00 ; arp -d $i; done; done console 2: ping -f any-dead-host-in-L2 console 3: # watch for memory consumption: vmstat -m | awk '$1~/lltable/{print$2}' - possible problems in arptimer() / nd6_timer() when dropping/reacquiring lock. New logic explicitly handles use-or-create cases in every lla_create user. Basically, most of the changes are purely mechanical. However, we explicitly avoid using existing lle's for interface/static LLE records. * While here, call lle_event handlers on all real table lle change. * Create lltable_free_entry() calling existing per-lltable lle_free_t callback for entry deletion
2015-08-20 12:05:17 +00:00
log(LOG_DEBUG,
"arpresolve: can't allocate llinfo for %s on %s\n",
Use inet_ntoa_r() instead of inet_ntoa() throughout the kernel inet_ntoa() cannot be used safely in a multithreaded environment because it uses a static local buffer. Instead, use inet_ntoa_r() with a buffer on the caller's stack. Suggested by: glebius, emaste Reviewed by: gnn MFC after: 2 weeks Sponsored by: Dell EMC Differential Revision: https://reviews.freebsd.org/D9625
2017-02-16 20:47:41 +00:00
inet_ntoa_r(SIN(dst)->sin_addr, addrbuf),
if_name(ifp));
* Split allocation and table linking for lle's. Before that, the logic besides lle_create() was the following: return existing if found, create if not. This behaviour was error-prone since we had to deal with 'sudden' static<>dynamic lle changes. This commit fixes bunch of different issues like: - refcount leak when lle is converted to static. Simple check case: console 1: while true; do for i in `arp -an|awk '$4~/incomp/{print$2}'|tr -d '()'`; do arp -s $i 00:22:44:66:88:00 ; arp -d $i; done; done console 2: ping -f any-dead-host-in-L2 console 3: # watch for memory consumption: vmstat -m | awk '$1~/lltable/{print$2}' - possible problems in arptimer() / nd6_timer() when dropping/reacquiring lock. New logic explicitly handles use-or-create cases in every lla_create user. Basically, most of the changes are purely mechanical. However, we explicitly avoid using existing lle's for interface/static LLE records. * While here, call lle_event handlers on all real table lle change. * Create lltable_free_entry() calling existing per-lltable lle_free_t callback for entry deletion
2015-08-20 12:05:17 +00:00
m_freem(m);
return (EINVAL);
}
Some more whitespace cleanup.
2012-08-01 09:00:26 +00:00
IF_AFDATA_WLOCK(ifp);
* Split allocation and table linking for lle's. Before that, the logic besides lle_create() was the following: return existing if found, create if not. This behaviour was error-prone since we had to deal with 'sudden' static<>dynamic lle changes. This commit fixes bunch of different issues like: - refcount leak when lle is converted to static. Simple check case: console 1: while true; do for i in `arp -an|awk '$4~/incomp/{print$2}'|tr -d '()'`; do arp -s $i 00:22:44:66:88:00 ; arp -d $i; done; done console 2: ping -f any-dead-host-in-L2 console 3: # watch for memory consumption: vmstat -m | awk '$1~/lltable/{print$2}' - possible problems in arptimer() / nd6_timer() when dropping/reacquiring lock. New logic explicitly handles use-or-create cases in every lla_create user. Basically, most of the changes are purely mechanical. However, we explicitly avoid using existing lle's for interface/static LLE records. * While here, call lle_event handlers on all real table lle change. * Create lltable_free_entry() calling existing per-lltable lle_free_t callback for entry deletion
2015-08-20 12:05:17 +00:00
LLE_WLOCK(la);
la_tmp = lla_lookup(LLTABLE(ifp), LLE_EXCLUSIVE, dst);
/* Prefer ANY existing lle over newly-created one */
if (la_tmp == NULL)
lltable_link_entry(LLTABLE(ifp), la);
Some more whitespace cleanup.
2012-08-01 09:00:26 +00:00
IF_AFDATA_WUNLOCK(ifp);
* Split allocation and table linking for lle's. Before that, the logic besides lle_create() was the following: return existing if found, create if not. This behaviour was error-prone since we had to deal with 'sudden' static<>dynamic lle changes. This commit fixes bunch of different issues like: - refcount leak when lle is converted to static. Simple check case: console 1: while true; do for i in `arp -an|awk '$4~/incomp/{print$2}'|tr -d '()'`; do arp -s $i 00:22:44:66:88:00 ; arp -d $i; done; done console 2: ping -f any-dead-host-in-L2 console 3: # watch for memory consumption: vmstat -m | awk '$1~/lltable/{print$2}' - possible problems in arptimer() / nd6_timer() when dropping/reacquiring lock. New logic explicitly handles use-or-create cases in every lla_create user. Basically, most of the changes are purely mechanical. However, we explicitly avoid using existing lle's for interface/static LLE records. * While here, call lle_event handlers on all real table lle change. * Create lltable_free_entry() calling existing per-lltable lle_free_t callback for entry deletion
2015-08-20 12:05:17 +00:00
if (la_tmp != NULL) {
lltable_free_entry(LLTABLE(ifp), la);
la = la_tmp;
}
default to doing lla_lookup with shared afdata lock and returning a shared lock on the lle - thus restoring parallel performance to pre-arpv2 level
2008-12-17 00:14:28 +00:00
}
o Fix a race between three threads: output path, incoming ARP packet and route request adding/removing ARP entries. The root of the problem is that struct llinfo_arp was accessed without any locks. To close race we will use locking provided by rtentry, that references this llinfo_arp: - Make arplookup() return a locked rtentry. - In arpresolve() hold the lock provided by rt_check()/arplookup() until the end of function, covering all accesses to the rtentry itself and llinfo_arp it refers to. - In in_arpinput() do not drop lock provided by arplookup() during first part of the function. - Simplify logic in the first part of in_arpinput(), removing one level of indentation. - In the second part of in_arpinput() hold rtentry lock while copying address. o Fix a condition when route entry is destroyed, while another thread is contested on its lock: - When storing a pointer to rtentry in llinfo_arp list, always add a reference to this rtentry, to prevent rtentry being destroyed via RTM_DELETE request. - Remove this reference when removing entry from llinfo_arp list. o Further cleanup of arptimer(): - Inline arptfree() into arptimer(). - Use official queue(3) way to pass LIST. - Hold rtentry lock while reading its structure. - Do not check that sdl_family is AF_LINK, but assert this. Reviewed by: sam Stress test: http://www.holm.cc/stress/log/cons141.html Stress test: http://people.freebsd.org/~pho/stress/log/cons144.html
2005-08-11 08:25:48 +00:00
if (la == NULL) {
This main goals of this project are: 1. separating L2 tables (ARP, NDP) from the L3 routing tables 2. removing as much locking dependencies among these layers as possible to allow for some parallelism in the search operations 3. simplify the logic in the routing code, The most notable end result is the obsolescent of the route cloning (RTF_CLONING) concept, which translated into code reduction in both IPv4 ARP and IPv6 NDP related modules, and size reduction in struct rtentry{}. The change in design obsoletes the semantics of RTF_CLONING, RTF_WASCLONE and RTF_LLINFO routing flags. The userland applications such as "arp" and "ndp" have been modified to reflect those changes. The output from "netstat -r" shows only the routing entries. Quite a few developers have contributed to this project in the past: Glebius Smirnoff, Luigi Rizzo, Alessandro Cerri, and Andre Oppermann. And most recently: - Kip Macy revised the locking code completely, thus completing the last piece of the puzzle, Kip has also been conducting active functional testing - Sam Leffler has helped me improving/refactoring the code, and provided valuable reviews - Julian Elischer setup the perforce tree for me and has helped me maintaining that branch before the svn conversion
2008-12-15 06:10:57 +00:00
m_freem(m);
return (EINVAL);
Some style(9) and whitespace changes. Together with: Andrey Zonov <andrey zonov.org>
2012-07-31 11:31:12 +00:00
}
- Do not hold route entry lock, when calling arprequest(). One such call was introduced by me in 1.139, the other one was present before. - Do all manipulations with rtentry and la before dropping the lock. - Copy interface address from route into local variable before dropping the lock. Supply this copy as argument to arprequest() LORs fixed: http://sources.zabbadoz.net/freebsd/lor/003.html http://sources.zabbadoz.net/freebsd/lor/037.html http://sources.zabbadoz.net/freebsd/lor/061.html http://sources.zabbadoz.net/freebsd/lor/062.html http://sources.zabbadoz.net/freebsd/lor/064.html http://sources.zabbadoz.net/freebsd/lor/068.html http://sources.zabbadoz.net/freebsd/lor/071.html http://sources.zabbadoz.net/freebsd/lor/074.html http://sources.zabbadoz.net/freebsd/lor/077.html http://sources.zabbadoz.net/freebsd/lor/093.html http://sources.zabbadoz.net/freebsd/lor/135.html http://sources.zabbadoz.net/freebsd/lor/140.html http://sources.zabbadoz.net/freebsd/lor/142.html http://sources.zabbadoz.net/freebsd/lor/145.html http://sources.zabbadoz.net/freebsd/lor/152.html http://sources.zabbadoz.net/freebsd/lor/158.html
2005-09-09 10:06:27 +00:00
This main goals of this project are: 1. separating L2 tables (ARP, NDP) from the L3 routing tables 2. removing as much locking dependencies among these layers as possible to allow for some parallelism in the search operations 3. simplify the logic in the routing code, The most notable end result is the obsolescent of the route cloning (RTF_CLONING) concept, which translated into code reduction in both IPv4 ARP and IPv6 NDP related modules, and size reduction in struct rtentry{}. The change in design obsoletes the semantics of RTF_CLONING, RTF_WASCLONE and RTF_LLINFO routing flags. The userland applications such as "arp" and "ndp" have been modified to reflect those changes. The output from "netstat -r" shows only the routing entries. Quite a few developers have contributed to this project in the past: Glebius Smirnoff, Luigi Rizzo, Alessandro Cerri, and Andre Oppermann. And most recently: - Kip Macy revised the locking code completely, thus completing the last piece of the puzzle, Kip has also been conducting active functional testing - Sam Leffler has helped me improving/refactoring the code, and provided valuable reviews - Julian Elischer setup the perforce tree for me and has helped me maintaining that branch before the svn conversion
2008-12-15 06:10:57 +00:00
if ((la->la_flags & LLE_VALID) &&
Use time_uptime instead of non-monotonic time_second to drive ARP timeouts. Suggested by: bde
2010-11-30 15:57:00 +00:00
((la->la_flags & LLE_STATIC) || la->la_expire > time_uptime)) {
Implement interface link header precomputation API. Add if_requestencap() interface method which is capable of calculating various link headers for given interface. Right now there is support for INET/INET6/ARP llheader calculation (IFENCAP_LL type request). Other types are planned to support more complex calculation (L2 multipath lagg nexthops, tunnel encap nexthops, etc..). Reshape 'struct route' to be able to pass additional data (with is length) to prepend to mbuf. These two changes permits routing code to pass pre-calculated nexthop data (like L2 header for route w/gateway) down to the stack eliminating the need for other lookups. It also brings us closer to more complex scenarios like transparently handling MPLS nexthops and tunnel interfaces. Last, but not least, it removes layering violation introduced by flowtable code (ro_lle) and simplifies handling of existing if_output consumers. ARP/ND changes: Make arp/ndp stack pre-calculate link header upon installing/updating lle record. Interface link address change are handled by re-calculating headers for all lles based on if_lladdr event. After these changes, arpresolve()/nd6_resolve() returns full pre-calculated header for supported interfaces thus simplifying if_output(). Move these lookups to separate ether_resolve_addr() function which ether returs error or fully-prepared link header. Add <arp|nd6_>resolve_addr() compat versions to return link addresses instead of pre-calculated data. BPF changes: Raw bpf writes occupied _two_ cases: AF_UNSPEC and pseudo_AF_HDRCMPLT. Despite the naming, both of there have ther header "complete". The only difference is that interface source mac has to be filled by OS for AF_UNSPEC (controlled via BIOCGHDRCMPLT). This logic has to stay inside BPF and not pollute if_output() routines. Convert BPF to pass prepend data via new 'struct route' mechanism. Note that it does not change non-optimized if_output(): ro_prepend handling is purely optional. Side note: hackish pseudo_AF_HDRCMPLT is supported for ethernet and FDDI. It is not needed for ethernet anymore. The only remaining FDDI user is dev/pdq mostly untouched since 2007. FDDI support was eliminated from OpenBSD in 2013 (sys/net/if_fddisubr.c rev 1.65). Flowtable changes: Flowtable violates layering by saving (and not correctly managing) rtes/lles. Instead of passing lle pointer, pass pointer to pre-calculated header data from that lle. Differential Revision: https://reviews.freebsd.org/D4102
2015-12-31 05:03:27 +00:00
if (flags & LLE_ADDRONLY) {
lladdr = la->ll_addr;
ll_len = ifp->if_addrlen;
} else {
lladdr = la->r_linkdata;
ll_len = la->r_hdrlen;
}
bcopy(lladdr, desten, ll_len);
Some more whitespace cleanup.
2012-08-01 09:00:26 +00:00
Fix outgoing TCP/UDP packet drop on arp/ndp entry expiration. Current arp/nd code relies on the feedback from the datapath indicating that the entry is still used. This mechanism is incorporated into the arpresolve()/nd6_resolve() routines. After the inpcb route cache introduction, the packet path for the locally-originated packets changed, passing cached lle pointer to the ether_output() directly. This resulted in the arp/ndp entry expire each time exactly after the configured max_age interval. During the small window between the ARP/NDP request and reply from the router, most of the packets got lost. Fix this behaviour by plugging datapath notification code to the packet path used by route cache. Unify the notification code by using single inlined function with the per-AF callbacks. Reported by: sthaug at nethelp.no Reviewed by: ae MFC after: 2 weeks
2018-03-17 17:05:48 +00:00
/* Notify LLE code that the entry was used by datapath */
llentry_mark_used(la);
Do not return unlocked/unreferenced lle in arpresolve/nd6_storelladdr - return lle flags IFF needed. Do not pass rte to arpresolve - pass is_gateway flag instead.
2014-11-27 23:06:25 +00:00
if (pflags != NULL)
Remove LLE read lock from IPv4 fast path. LLE structure is mostly unchanged during its lifecycle. To be more specific, there are 2 things relevant for fast path lookup code: 1) link-level address change. Since r286722, these updates are performed under AFDATA WLOCK. 2) Some sort of feedback indicating that this particular entry is used so we re-send arp request to perform reachability verification instead of expiring entry. The only signal that is needed from fast path is something like binary yes/no. The latter is solved by the following changes: 1) introduce special r_skip_req field which is read lockless by fast path, but updated under (new) req_mutex mutex. If this field is non-zero, then fast path will acquire lock and set it back to 0. 2) introduce simple state machine: incomplete->reachable<->verify->deleted. Before that we implicitely had incomplete->reachable->deleted state machine, with V_arpt_keep between "reachable" and "deleted". Verification was performed in runtime 5 seconds before V_arpt_keep expire. This is changed to "change state to verify 5 seconds before V_arpt_keep, set r_skip_req to non-zero value and check it every second". If the value is zero - then send arp verification probe. These changes do not introduce any signifficant control plane overhead: typically lle callout timer would fire 1 time more each V_arpt_keep (1200s) for used lles and up to arp_maxtries (5) for dead lles. As a result, all packets towards "reachable" lle are handled by fast path without acquiring lle read lock. Additional "req_mutex" is needed because callout / arpresolve_slow() or eventhandler might keep LLE lock for signifficant amount of time, which might not be feasible for fast path locking (e.g. having rmlock as ether AFDATA or lltable own lock). Differential Revision: https://reviews.freebsd.org/D3688
2015-12-05 09:50:37 +00:00
*pflags = la->la_flags & (LLE_VALID|LLE_IFADDR);
This change re-adds L2 caching for TCP and UDP, as originally added in D4306 but removed due to other changes in the system. Restore the llentry pointer to the "struct route", and use it to cache the L2 lookup (ARP or ND6) as appropriate. Submitted by: Mike Karels Differential Revision: https://reviews.freebsd.org/D6262
2016-06-02 17:51:29 +00:00
if (plle) {
LLE_ADDREF(la);
*plle = la;
}
Split arpresolve() into fast/slow path. This change isolates the most common case (e.g. successful lookup) from more complicates scenarios. It also (tries to) make code more simple by avoiding retry: cycle. The actual goal is to prepare code to the upcoming change that will allow LL address retrieval without acquiring LLE lock at all. Reviewed by: ae Differential Revision: https://reviews.freebsd.org/D3383
2015-08-16 12:23:58 +00:00
LLE_WUNLOCK(la);
If there is a system with a bpf consumer running and a packet is wanted to be transmitted but the arp cache entry expired, which triggers an arp request to be sent, the bpf code might want to sleep but crash the system due to a non sleep lock held from the arp entry not released properly. Release the lock before calling the arp request code to solve the issue as is done on all the other code paths. PR: 200323 Approved by: ae, gnn(mentor) MFC after: 1 week Sponsored by: Netgate Differential Revision: https://reviews.freebsd.org/D2828
2015-06-17 12:23:04 +00:00
return (0);
Some style(9) and whitespace changes. Together with: Andrey Zonov <andrey zonov.org>
2012-07-31 11:31:12 +00:00
}
Use time_uptime instead of non-monotonic time_second to drive ARP timeouts. Suggested by: bde
2010-11-30 15:57:00 +00:00
renew = (la->la_asked == 0 || la->la_expire != time_uptime);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
/*
* There is an arptab entry, but no ethernet address
Add a queue to hold packets while we await an ARP reply. When a fast machine first brings up some non TCP networking program it is quite possible that we will drop packets due to the fact that only one packet can be held per ARP entry. This leads to packets being missed when a program starts or restarts if the ARP data is not currently in the ARP cache. This code adds a new sysctl, net.link.ether.inet.maxhold, which defines a system wide maximum number of packets to be held in each ARP entry. Up to maxhold packets are queued until an ARP reply is received or the ARP times out. The default setting is the old value of 1 which has been part of the BSD networking code since time immemorial. Expose the time we hold an incomplete ARP entry by adding the sysctl net.link.ether.inet.wait, which defaults to 20 seconds, the value used when the new ARP code was added.. Reviewed by: bz, rpaulo MFC after: 3 weeks
2010-11-12 22:03:02 +00:00
* response yet. Add the mbuf to the list, dropping
* the oldest packet if we have exceeded the system
* setting.
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
*/
This main goals of this project are: 1. separating L2 tables (ARP, NDP) from the L3 routing tables 2. removing as much locking dependencies among these layers as possible to allow for some parallelism in the search operations 3. simplify the logic in the routing code, The most notable end result is the obsolescent of the route cloning (RTF_CLONING) concept, which translated into code reduction in both IPv4 ARP and IPv6 NDP related modules, and size reduction in struct rtentry{}. The change in design obsoletes the semantics of RTF_CLONING, RTF_WASCLONE and RTF_LLINFO routing flags. The userland applications such as "arp" and "ndp" have been modified to reflect those changes. The output from "netstat -r" shows only the routing entries. Quite a few developers have contributed to this project in the past: Glebius Smirnoff, Luigi Rizzo, Alessandro Cerri, and Andre Oppermann. And most recently: - Kip Macy revised the locking code completely, thus completing the last piece of the puzzle, Kip has also been conducting active functional testing - Sam Leffler has helped me improving/refactoring the code, and provided valuable reviews - Julian Elischer setup the perforce tree for me and has helped me maintaining that branch before the svn conversion
2008-12-15 06:10:57 +00:00
if (m != NULL) {
Add a queue to hold packets while we await an ARP reply. When a fast machine first brings up some non TCP networking program it is quite possible that we will drop packets due to the fact that only one packet can be held per ARP entry. This leads to packets being missed when a program starts or restarts if the ARP data is not currently in the ARP cache. This code adds a new sysctl, net.link.ether.inet.maxhold, which defines a system wide maximum number of packets to be held in each ARP entry. Up to maxhold packets are queued until an ARP reply is received or the ARP times out. The default setting is the old value of 1 which has been part of the BSD networking code since time immemorial. Expose the time we hold an incomplete ARP entry by adding the sysctl net.link.ether.inet.wait, which defaults to 20 seconds, the value used when the new ARP code was added.. Reviewed by: bz, rpaulo MFC after: 3 weeks
2010-11-12 22:03:02 +00:00
if (la->la_numheld >= V_arp_maxhold) {
if (la->la_hold != NULL) {
next = la->la_hold->m_nextpkt;
m_freem(la->la_hold);
la->la_hold = next;
la->la_numheld--;
ARPSTAT_INC(dropped);
}
Some style(9) and whitespace changes. Together with: Andrey Zonov <andrey zonov.org>
2012-07-31 11:31:12 +00:00
}
Add ARP statistics to the kernel and netstat. New counters now exist for: requests sent replies sent requests received replies received packets received total packets dropped due to no ARP entry entrys timed out Duplicate IPs seen The new statistics are seen in the netstat command when it is given the -s command line switch. MFC after: 2 weeks In collaboration with: bz
2009-09-03 21:10:57 +00:00
if (la->la_hold != NULL) {
Add a queue to hold packets while we await an ARP reply. When a fast machine first brings up some non TCP networking program it is quite possible that we will drop packets due to the fact that only one packet can be held per ARP entry. This leads to packets being missed when a program starts or restarts if the ARP data is not currently in the ARP cache. This code adds a new sysctl, net.link.ether.inet.maxhold, which defines a system wide maximum number of packets to be held in each ARP entry. Up to maxhold packets are queued until an ARP reply is received or the ARP times out. The default setting is the old value of 1 which has been part of the BSD networking code since time immemorial. Expose the time we hold an incomplete ARP entry by adding the sysctl net.link.ether.inet.wait, which defaults to 20 seconds, the value used when the new ARP code was added.. Reviewed by: bz, rpaulo MFC after: 3 weeks
2010-11-12 22:03:02 +00:00
curr = la->la_hold;
while (curr->m_nextpkt != NULL)
curr = curr->m_nextpkt;
curr->m_nextpkt = m;
Some style(9) and whitespace changes. Together with: Andrey Zonov <andrey zonov.org>
2012-07-31 11:31:12 +00:00
} else
Add a queue to hold packets while we await an ARP reply. When a fast machine first brings up some non TCP networking program it is quite possible that we will drop packets due to the fact that only one packet can be held per ARP entry. This leads to packets being missed when a program starts or restarts if the ARP data is not currently in the ARP cache. This code adds a new sysctl, net.link.ether.inet.maxhold, which defines a system wide maximum number of packets to be held in each ARP entry. Up to maxhold packets are queued until an ARP reply is received or the ARP times out. The default setting is the old value of 1 which has been part of the BSD networking code since time immemorial. Expose the time we hold an incomplete ARP entry by adding the sysctl net.link.ether.inet.wait, which defaults to 20 seconds, the value used when the new ARP code was added.. Reviewed by: bz, rpaulo MFC after: 3 weeks
2010-11-12 22:03:02 +00:00
la->la_hold = m;
la->la_numheld++;
Don't panic in arpresolve if we're given a null mbuf. We could insist that the caller just pass in an initialized mbuf even if didn't have any data - but that seems rather contrived.
2007-12-17 04:19:25 +00:00
}
Rework ARP retransmission algorythm so that ARP requests are retransmitted without suppression, while there is demand for such ARP entry. As before, retransmission is rate limited to one packet per second. Details: - Remove net.link.ether.inet.host_down_time - Do not set/clear RTF_REJECT flag on route, to avoid rt_check() returning error. We will generate error ourselves. - Return EWOULDBLOCK on first arp_maxtries failed requests , and return EHOSTDOWN/EHOSTUNREACH on further requests. - Retransmit ARP request always, independently from return code. Ratelimit to 1 pps.
2005-11-08 12:05:57 +00:00
/*
* Return EWOULDBLOCK if we have tried less than arp_maxtries. It
* will be masked by ether_output(). Return EHOSTDOWN/EHOSTUNREACH
* if we have already sent arp_maxtries ARP requests. Retransmit the
* ARP request, but not faster than one request per second.
*/
Commit step 1 of the vimage project, (network stack) virtualization work done by Marko Zec (zec@). This is the first in a series of commits over the course of the next few weeks. Mark all uses of global variables to be virtualized with a V_ prefix. Use macros to map them back to their global names for now, so this is a NOP change only. We hope to have caught at least 85-90% of what is needed so we do not invalidate a lot of outstanding patches again. Obtained from: //depot/projects/vimage-commit2/... Reviewed by: brooks, des, ed, mav, julian, jamie, kris, rwatson, zec, ... (various people I forgot, different versions) md5 (with a bit of help) Sponsored by: NLnet Foundation, The FreeBSD Foundation X-MFC after: never V_Commit_Message_Reviewed_By: more people than the patch
2008-08-17 23:27:27 +00:00
if (la->la_asked < V_arp_maxtries)
Rework ARP retransmission algorythm so that ARP requests are retransmitted without suppression, while there is demand for such ARP entry. As before, retransmission is rate limited to one packet per second. Details: - Remove net.link.ether.inet.host_down_time - Do not set/clear RTF_REJECT flag on route, to avoid rt_check() returning error. We will generate error ourselves. - Return EWOULDBLOCK on first arp_maxtries failed requests , and return EHOSTDOWN/EHOSTUNREACH on further requests. - Retransmit ARP request always, independently from return code. Ratelimit to 1 pps.
2005-11-08 12:05:57 +00:00
error = EWOULDBLOCK; /* First request. */
else
Do not return unlocked/unreferenced lle in arpresolve/nd6_storelladdr - return lle flags IFF needed. Do not pass rte to arpresolve - pass is_gateway flag instead.
2014-11-27 23:06:25 +00:00
error = is_gw != 0 ? EHOSTUNREACH : EHOSTDOWN;
Rework ARP retransmission algorythm so that ARP requests are retransmitted without suppression, while there is demand for such ARP entry. As before, retransmission is rate limited to one packet per second. Details: - Remove net.link.ether.inet.host_down_time - Do not set/clear RTF_REJECT flag on route, to avoid rt_check() returning error. We will generate error ourselves. - Return EWOULDBLOCK on first arp_maxtries failed requests , and return EHOSTDOWN/EHOSTUNREACH on further requests. - Retransmit ARP request always, independently from return code. Ratelimit to 1 pps.
2005-11-08 12:05:57 +00:00
This main goals of this project are: 1. separating L2 tables (ARP, NDP) from the L3 routing tables 2. removing as much locking dependencies among these layers as possible to allow for some parallelism in the search operations 3. simplify the logic in the routing code, The most notable end result is the obsolescent of the route cloning (RTF_CLONING) concept, which translated into code reduction in both IPv4 ARP and IPv6 NDP related modules, and size reduction in struct rtentry{}. The change in design obsoletes the semantics of RTF_CLONING, RTF_WASCLONE and RTF_LLINFO routing flags. The userland applications such as "arp" and "ndp" have been modified to reflect those changes. The output from "netstat -r" shows only the routing entries. Quite a few developers have contributed to this project in the past: Glebius Smirnoff, Luigi Rizzo, Alessandro Cerri, and Andre Oppermann. And most recently: - Kip Macy revised the locking code completely, thus completing the last piece of the puzzle, Kip has also been conducting active functional testing - Sam Leffler has helped me improving/refactoring the code, and provided valuable reviews - Julian Elischer setup the perforce tree for me and has helped me maintaining that branch before the svn conversion
2008-12-15 06:10:57 +00:00
if (renew) {
Make arp code return (more) errors. arprequest() is a void function and in case of error we simply return without any feedback. In case of any local operation or *if_output() failing no feedback is send up the stack for the packet which triggered the arp request to be sent. arpresolve_full() has three pre-canned possible errors returned (if we have not yet sent enough arp requests or if we tried often enough without success) otherwise "no error" is returned. Make arprequest() an "internal" function arprequest_internal() which does return a possible error to the caller. Preserve arprequest() as a void wrapper function for external consumers. In arpresolve_full() add an extra error checking. Use the arprequest_internal() function and only return an error if non of the three ones (mentioend above) are already set. This will return possible errors all the way up the stack and allows functions and programs to react on the send errors rather than leaving them in the dark. Also they might get more detailed feedback of why packets cannot be sent and they will receive it quicker. Reviewed by: karels, hselasky Differential Revision: https://reviews.freebsd.org/D18904
2019-02-24 22:49:56 +00:00
int canceled, e;
Plug reference leaks in the link-layer code ("new-arp") that previously prevented the link-layer entry from being freed. In both in.c and in6.c (though that code path seems to be basically dead) plug a reference leak in case of a pending callout being drained. In if_ether.c consistently add a reference before resetting the callout and in case we canceled a pending one remove the reference for that. In the final case in arptimer, before freeing the expired entry, remove the reference again and explicitly call callout_stop() to clear the active flag. In nd6.c:nd6_free() we are only ever called from the callout function and thus need to remove the reference there as well before calling into llentry_free(). In if_llatbl.c when freeing entire tables make sure that in case we cancel a pending callout to remove the reference as well. Reviewed by: qingli (earlier version) MFC after: 10 days Problem observed, patch tested by: simon on ipv6gw.f.o, Christian Kratzer (ck cksoft.de), Evgenii Davidov (dado korolev-net.ru) PR: kern/144564 Configurations still affected: with options FLOWTABLE
2010-04-11 16:04:08 +00:00
This main goals of this project are: 1. separating L2 tables (ARP, NDP) from the L3 routing tables 2. removing as much locking dependencies among these layers as possible to allow for some parallelism in the search operations 3. simplify the logic in the routing code, The most notable end result is the obsolescent of the route cloning (RTF_CLONING) concept, which translated into code reduction in both IPv4 ARP and IPv6 NDP related modules, and size reduction in struct rtentry{}. The change in design obsoletes the semantics of RTF_CLONING, RTF_WASCLONE and RTF_LLINFO routing flags. The userland applications such as "arp" and "ndp" have been modified to reflect those changes. The output from "netstat -r" shows only the routing entries. Quite a few developers have contributed to this project in the past: Glebius Smirnoff, Luigi Rizzo, Alessandro Cerri, and Andre Oppermann. And most recently: - Kip Macy revised the locking code completely, thus completing the last piece of the puzzle, Kip has also been conducting active functional testing - Sam Leffler has helped me improving/refactoring the code, and provided valuable reviews - Julian Elischer setup the perforce tree for me and has helped me maintaining that branch before the svn conversion
2008-12-15 06:10:57 +00:00
LLE_ADDREF(la);
Use time_uptime instead of non-monotonic time_second to drive ARP timeouts. Suggested by: bde
2010-11-30 15:57:00 +00:00
la->la_expire = time_uptime;
Use single 'lle_timer' callout in lltable instead of two different names of the same timer.
2015-08-11 12:38:54 +00:00
canceled = callout_reset(&la->lle_timer, hz * V_arpt_down,
Plug reference leaks in the link-layer code ("new-arp") that previously prevented the link-layer entry from being freed. In both in.c and in6.c (though that code path seems to be basically dead) plug a reference leak in case of a pending callout being drained. In if_ether.c consistently add a reference before resetting the callout and in case we canceled a pending one remove the reference for that. In the final case in arptimer, before freeing the expired entry, remove the reference again and explicitly call callout_stop() to clear the active flag. In nd6.c:nd6_free() we are only ever called from the callout function and thus need to remove the reference there as well before calling into llentry_free(). In if_llatbl.c when freeing entire tables make sure that in case we cancel a pending callout to remove the reference as well. Reviewed by: qingli (earlier version) MFC after: 10 days Problem observed, patch tested by: simon on ipv6gw.f.o, Christian Kratzer (ck cksoft.de), Evgenii Davidov (dado korolev-net.ru) PR: kern/144564 Configurations still affected: with options FLOWTABLE
2010-04-11 16:04:08 +00:00
arptimer, la);
if (canceled)
LLE_REMREF(la);
o Increment requests counter right before send out an ARP query actually. Otherwise the code could lead to the spurious EHOSTDOWN errors. PR: kern/107807 Submitted by: Dmitrij Tejblum MFC after: 1 month
2007-01-14 18:44:17 +00:00
la->la_asked++;
This main goals of this project are: 1. separating L2 tables (ARP, NDP) from the L3 routing tables 2. removing as much locking dependencies among these layers as possible to allow for some parallelism in the search operations 3. simplify the logic in the routing code, The most notable end result is the obsolescent of the route cloning (RTF_CLONING) concept, which translated into code reduction in both IPv4 ARP and IPv6 NDP related modules, and size reduction in struct rtentry{}. The change in design obsoletes the semantics of RTF_CLONING, RTF_WASCLONE and RTF_LLINFO routing flags. The userland applications such as "arp" and "ndp" have been modified to reflect those changes. The output from "netstat -r" shows only the routing entries. Quite a few developers have contributed to this project in the past: Glebius Smirnoff, Luigi Rizzo, Alessandro Cerri, and Andre Oppermann. And most recently: - Kip Macy revised the locking code completely, thus completing the last piece of the puzzle, Kip has also been conducting active functional testing - Sam Leffler has helped me improving/refactoring the code, and provided valuable reviews - Julian Elischer setup the perforce tree for me and has helped me maintaining that branch before the svn conversion
2008-12-15 06:10:57 +00:00
LLE_WUNLOCK(la);
Make arp code return (more) errors. arprequest() is a void function and in case of error we simply return without any feedback. In case of any local operation or *if_output() failing no feedback is send up the stack for the packet which triggered the arp request to be sent. arpresolve_full() has three pre-canned possible errors returned (if we have not yet sent enough arp requests or if we tried often enough without success) otherwise "no error" is returned. Make arprequest() an "internal" function arprequest_internal() which does return a possible error to the caller. Preserve arprequest() as a void wrapper function for external consumers. In arpresolve_full() add an extra error checking. Use the arprequest_internal() function and only return an error if non of the three ones (mentioend above) are already set. This will return possible errors all the way up the stack and allows functions and programs to react on the send errors rather than leaving them in the dark. Also they might get more detailed feedback of why packets cannot be sent and they will receive it quicker. Reviewed by: karels, hselasky Differential Revision: https://reviews.freebsd.org/D18904
2019-02-24 22:49:56 +00:00
e = arprequest_internal(ifp, NULL, &SIN(dst)->sin_addr, NULL);
/*
* Only overwrite 'error' in case of error; in case of success
* the proper return value was already set above.
*/
if (e != 0)
return (e);
This main goals of this project are: 1. separating L2 tables (ARP, NDP) from the L3 routing tables 2. removing as much locking dependencies among these layers as possible to allow for some parallelism in the search operations 3. simplify the logic in the routing code, The most notable end result is the obsolescent of the route cloning (RTF_CLONING) concept, which translated into code reduction in both IPv4 ARP and IPv6 NDP related modules, and size reduction in struct rtentry{}. The change in design obsoletes the semantics of RTF_CLONING, RTF_WASCLONE and RTF_LLINFO routing flags. The userland applications such as "arp" and "ndp" have been modified to reflect those changes. The output from "netstat -r" shows only the routing entries. Quite a few developers have contributed to this project in the past: Glebius Smirnoff, Luigi Rizzo, Alessandro Cerri, and Andre Oppermann. And most recently: - Kip Macy revised the locking code completely, thus completing the last piece of the puzzle, Kip has also been conducting active functional testing - Sam Leffler has helped me improving/refactoring the code, and provided valuable reviews - Julian Elischer setup the perforce tree for me and has helped me maintaining that branch before the svn conversion
2008-12-15 06:10:57 +00:00
return (error);
}
Split arpresolve() into fast/slow path. This change isolates the most common case (e.g. successful lookup) from more complicates scenarios. It also (tries to) make code more simple by avoiding retry: cycle. The actual goal is to prepare code to the upcoming change that will allow LL address retrieval without acquiring LLE lock at all. Reviewed by: ae Differential Revision: https://reviews.freebsd.org/D3383
2015-08-16 12:23:58 +00:00
LLE_WUNLOCK(la);
Rework ARP retransmission algorythm so that ARP requests are retransmitted without suppression, while there is demand for such ARP entry. As before, retransmission is rate limited to one packet per second. Details: - Remove net.link.ether.inet.host_down_time - Do not set/clear RTF_REJECT flag on route, to avoid rt_check() returning error. We will generate error ourselves. - Return EWOULDBLOCK on first arp_maxtries failed requests , and return EHOSTDOWN/EHOSTUNREACH on further requests. - Retransmit ARP request always, independently from return code. Ratelimit to 1 pps.
2005-11-08 12:05:57 +00:00
return (error);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
Implement interface link header precomputation API. Add if_requestencap() interface method which is capable of calculating various link headers for given interface. Right now there is support for INET/INET6/ARP llheader calculation (IFENCAP_LL type request). Other types are planned to support more complex calculation (L2 multipath lagg nexthops, tunnel encap nexthops, etc..). Reshape 'struct route' to be able to pass additional data (with is length) to prepend to mbuf. These two changes permits routing code to pass pre-calculated nexthop data (like L2 header for route w/gateway) down to the stack eliminating the need for other lookups. It also brings us closer to more complex scenarios like transparently handling MPLS nexthops and tunnel interfaces. Last, but not least, it removes layering violation introduced by flowtable code (ro_lle) and simplifies handling of existing if_output consumers. ARP/ND changes: Make arp/ndp stack pre-calculate link header upon installing/updating lle record. Interface link address change are handled by re-calculating headers for all lles based on if_lladdr event. After these changes, arpresolve()/nd6_resolve() returns full pre-calculated header for supported interfaces thus simplifying if_output(). Move these lookups to separate ether_resolve_addr() function which ether returs error or fully-prepared link header. Add <arp|nd6_>resolve_addr() compat versions to return link addresses instead of pre-calculated data. BPF changes: Raw bpf writes occupied _two_ cases: AF_UNSPEC and pseudo_AF_HDRCMPLT. Despite the naming, both of there have ther header "complete". The only difference is that interface source mac has to be filled by OS for AF_UNSPEC (controlled via BIOCGHDRCMPLT). This logic has to stay inside BPF and not pollute if_output() routines. Convert BPF to pass prepend data via new 'struct route' mechanism. Note that it does not change non-optimized if_output(): ro_prepend handling is purely optional. Side note: hackish pseudo_AF_HDRCMPLT is supported for ethernet and FDDI. It is not needed for ethernet anymore. The only remaining FDDI user is dev/pdq mostly untouched since 2007. FDDI support was eliminated from OpenBSD in 2013 (sys/net/if_fddisubr.c rev 1.65). Flowtable changes: Flowtable violates layering by saving (and not correctly managing) rtes/lles. Instead of passing lle pointer, pass pointer to pre-calculated header data from that lle. Differential Revision: https://reviews.freebsd.org/D4102
2015-12-31 05:03:27 +00:00
/*
* Lookups link header based on an IP address.
Split arpresolve() into fast/slow path. This change isolates the most common case (e.g. successful lookup) from more complicates scenarios. It also (tries to) make code more simple by avoiding retry: cycle. The actual goal is to prepare code to the upcoming change that will allow LL address retrieval without acquiring LLE lock at all. Reviewed by: ae Differential Revision: https://reviews.freebsd.org/D3383
2015-08-16 12:23:58 +00:00
* On input:
* ifp is the interface we use
* is_gw != 0 if @dst represents gateway to some destination
* m is the mbuf. May be NULL if we don't have a packet.
* dst is the next hop,
Implement interface link header precomputation API. Add if_requestencap() interface method which is capable of calculating various link headers for given interface. Right now there is support for INET/INET6/ARP llheader calculation (IFENCAP_LL type request). Other types are planned to support more complex calculation (L2 multipath lagg nexthops, tunnel encap nexthops, etc..). Reshape 'struct route' to be able to pass additional data (with is length) to prepend to mbuf. These two changes permits routing code to pass pre-calculated nexthop data (like L2 header for route w/gateway) down to the stack eliminating the need for other lookups. It also brings us closer to more complex scenarios like transparently handling MPLS nexthops and tunnel interfaces. Last, but not least, it removes layering violation introduced by flowtable code (ro_lle) and simplifies handling of existing if_output consumers. ARP/ND changes: Make arp/ndp stack pre-calculate link header upon installing/updating lle record. Interface link address change are handled by re-calculating headers for all lles based on if_lladdr event. After these changes, arpresolve()/nd6_resolve() returns full pre-calculated header for supported interfaces thus simplifying if_output(). Move these lookups to separate ether_resolve_addr() function which ether returs error or fully-prepared link header. Add <arp|nd6_>resolve_addr() compat versions to return link addresses instead of pre-calculated data. BPF changes: Raw bpf writes occupied _two_ cases: AF_UNSPEC and pseudo_AF_HDRCMPLT. Despite the naming, both of there have ther header "complete". The only difference is that interface source mac has to be filled by OS for AF_UNSPEC (controlled via BIOCGHDRCMPLT). This logic has to stay inside BPF and not pollute if_output() routines. Convert BPF to pass prepend data via new 'struct route' mechanism. Note that it does not change non-optimized if_output(): ro_prepend handling is purely optional. Side note: hackish pseudo_AF_HDRCMPLT is supported for ethernet and FDDI. It is not needed for ethernet anymore. The only remaining FDDI user is dev/pdq mostly untouched since 2007. FDDI support was eliminated from OpenBSD in 2013 (sys/net/if_fddisubr.c rev 1.65). Flowtable changes: Flowtable violates layering by saving (and not correctly managing) rtes/lles. Instead of passing lle pointer, pass pointer to pre-calculated header data from that lle. Differential Revision: https://reviews.freebsd.org/D4102
2015-12-31 05:03:27 +00:00
* desten is the storage to put LL header.
Remove LLE read lock from IPv4 fast path. LLE structure is mostly unchanged during its lifecycle. To be more specific, there are 2 things relevant for fast path lookup code: 1) link-level address change. Since r286722, these updates are performed under AFDATA WLOCK. 2) Some sort of feedback indicating that this particular entry is used so we re-send arp request to perform reachability verification instead of expiring entry. The only signal that is needed from fast path is something like binary yes/no. The latter is solved by the following changes: 1) introduce special r_skip_req field which is read lockless by fast path, but updated under (new) req_mutex mutex. If this field is non-zero, then fast path will acquire lock and set it back to 0. 2) introduce simple state machine: incomplete->reachable<->verify->deleted. Before that we implicitely had incomplete->reachable->deleted state machine, with V_arpt_keep between "reachable" and "deleted". Verification was performed in runtime 5 seconds before V_arpt_keep expire. This is changed to "change state to verify 5 seconds before V_arpt_keep, set r_skip_req to non-zero value and check it every second". If the value is zero - then send arp verification probe. These changes do not introduce any signifficant control plane overhead: typically lle callout timer would fire 1 time more each V_arpt_keep (1200s) for used lles and up to arp_maxtries (5) for dead lles. As a result, all packets towards "reachable" lle are handled by fast path without acquiring lle read lock. Additional "req_mutex" is needed because callout / arpresolve_slow() or eventhandler might keep LLE lock for signifficant amount of time, which might not be feasible for fast path locking (e.g. having rmlock as ether AFDATA or lltable own lock). Differential Revision: https://reviews.freebsd.org/D3688
2015-12-05 09:50:37 +00:00
* flags returns subset of lle flags: LLE_VALID | LLE_IFADDR
Split arpresolve() into fast/slow path. This change isolates the most common case (e.g. successful lookup) from more complicates scenarios. It also (tries to) make code more simple by avoiding retry: cycle. The actual goal is to prepare code to the upcoming change that will allow LL address retrieval without acquiring LLE lock at all. Reviewed by: ae Differential Revision: https://reviews.freebsd.org/D3383
2015-08-16 12:23:58 +00:00
*
Implement interface link header precomputation API. Add if_requestencap() interface method which is capable of calculating various link headers for given interface. Right now there is support for INET/INET6/ARP llheader calculation (IFENCAP_LL type request). Other types are planned to support more complex calculation (L2 multipath lagg nexthops, tunnel encap nexthops, etc..). Reshape 'struct route' to be able to pass additional data (with is length) to prepend to mbuf. These two changes permits routing code to pass pre-calculated nexthop data (like L2 header for route w/gateway) down to the stack eliminating the need for other lookups. It also brings us closer to more complex scenarios like transparently handling MPLS nexthops and tunnel interfaces. Last, but not least, it removes layering violation introduced by flowtable code (ro_lle) and simplifies handling of existing if_output consumers. ARP/ND changes: Make arp/ndp stack pre-calculate link header upon installing/updating lle record. Interface link address change are handled by re-calculating headers for all lles based on if_lladdr event. After these changes, arpresolve()/nd6_resolve() returns full pre-calculated header for supported interfaces thus simplifying if_output(). Move these lookups to separate ether_resolve_addr() function which ether returs error or fully-prepared link header. Add <arp|nd6_>resolve_addr() compat versions to return link addresses instead of pre-calculated data. BPF changes: Raw bpf writes occupied _two_ cases: AF_UNSPEC and pseudo_AF_HDRCMPLT. Despite the naming, both of there have ther header "complete". The only difference is that interface source mac has to be filled by OS for AF_UNSPEC (controlled via BIOCGHDRCMPLT). This logic has to stay inside BPF and not pollute if_output() routines. Convert BPF to pass prepend data via new 'struct route' mechanism. Note that it does not change non-optimized if_output(): ro_prepend handling is purely optional. Side note: hackish pseudo_AF_HDRCMPLT is supported for ethernet and FDDI. It is not needed for ethernet anymore. The only remaining FDDI user is dev/pdq mostly untouched since 2007. FDDI support was eliminated from OpenBSD in 2013 (sys/net/if_fddisubr.c rev 1.65). Flowtable changes: Flowtable violates layering by saving (and not correctly managing) rtes/lles. Instead of passing lle pointer, pass pointer to pre-calculated header data from that lle. Differential Revision: https://reviews.freebsd.org/D4102
2015-12-31 05:03:27 +00:00
* On success, full/partial link header and flags are filled in and
* the function returns 0.
Split arpresolve() into fast/slow path. This change isolates the most common case (e.g. successful lookup) from more complicates scenarios. It also (tries to) make code more simple by avoiding retry: cycle. The actual goal is to prepare code to the upcoming change that will allow LL address retrieval without acquiring LLE lock at all. Reviewed by: ae Differential Revision: https://reviews.freebsd.org/D3383
2015-08-16 12:23:58 +00:00
* If the packet must be held pending resolution, we return EWOULDBLOCK
* On other errors, we return the corresponding error code.
* Note that m_freem() handles NULL.
*/
int
arpresolve(struct ifnet *ifp, int is_gw, struct mbuf *m,
This change re-adds L2 caching for TCP and UDP, as originally added in D4306 but removed due to other changes in the system. Restore the llentry pointer to the "struct route", and use it to cache the L2 lookup (ARP or ND6) as appropriate. Submitted by: Mike Karels Differential Revision: https://reviews.freebsd.org/D6262
2016-06-02 17:51:29 +00:00
const struct sockaddr *dst, u_char *desten, uint32_t *pflags,
struct llentry **plle)
Split arpresolve() into fast/slow path. This change isolates the most common case (e.g. successful lookup) from more complicates scenarios. It also (tries to) make code more simple by avoiding retry: cycle. The actual goal is to prepare code to the upcoming change that will allow LL address retrieval without acquiring LLE lock at all. Reviewed by: ae Differential Revision: https://reviews.freebsd.org/D3383
2015-08-16 12:23:58 +00:00
{
netinet: for pointers replace 0 with NULL. These are mostly cosmetical, no functional change. Found with devel/coccinelle. Reviewed by: ae. tuexen
2016-04-15 15:46:41 +00:00
struct llentry *la = NULL;
Split arpresolve() into fast/slow path. This change isolates the most common case (e.g. successful lookup) from more complicates scenarios. It also (tries to) make code more simple by avoiding retry: cycle. The actual goal is to prepare code to the upcoming change that will allow LL address retrieval without acquiring LLE lock at all. Reviewed by: ae Differential Revision: https://reviews.freebsd.org/D3383
2015-08-16 12:23:58 +00:00
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
NET_EPOCH_ASSERT();
Split arpresolve() into fast/slow path. This change isolates the most common case (e.g. successful lookup) from more complicates scenarios. It also (tries to) make code more simple by avoiding retry: cycle. The actual goal is to prepare code to the upcoming change that will allow LL address retrieval without acquiring LLE lock at all. Reviewed by: ae Differential Revision: https://reviews.freebsd.org/D3383
2015-08-16 12:23:58 +00:00
if (pflags != NULL)
*pflags = 0;
This change re-adds L2 caching for TCP and UDP, as originally added in D4306 but removed due to other changes in the system. Restore the llentry pointer to the "struct route", and use it to cache the L2 lookup (ARP or ND6) as appropriate. Submitted by: Mike Karels Differential Revision: https://reviews.freebsd.org/D6262
2016-06-02 17:51:29 +00:00
if (plle != NULL)
*plle = NULL;
Split arpresolve() into fast/slow path. This change isolates the most common case (e.g. successful lookup) from more complicates scenarios. It also (tries to) make code more simple by avoiding retry: cycle. The actual goal is to prepare code to the upcoming change that will allow LL address retrieval without acquiring LLE lock at all. Reviewed by: ae Differential Revision: https://reviews.freebsd.org/D3383
2015-08-16 12:23:58 +00:00
if (m != NULL) {
if (m->m_flags & M_BCAST) {
/* broadcast */
(void)memcpy(desten,
ifp->if_broadcastaddr, ifp->if_addrlen);
return (0);
}
if (m->m_flags & M_MCAST) {
/* multicast */
ETHER_MAP_IP_MULTICAST(&SIN(dst)->sin_addr, desten);
return (0);
}
}
Fix per-connection L2 caching in fast path r301217 re-added per-connection L2 caching from a previous change, but it omitted caching in the fast path. Add it. Reviewed By: gallatin Approved by: gnn (mentor) Differential Revision: https://reviews.freebsd.org/D7239
2016-07-22 02:11:49 +00:00
la = lla_lookup(LLTABLE(ifp), plle ? LLE_EXCLUSIVE : LLE_UNLOCKED, dst);
Remove LLE read lock from IPv4 fast path. LLE structure is mostly unchanged during its lifecycle. To be more specific, there are 2 things relevant for fast path lookup code: 1) link-level address change. Since r286722, these updates are performed under AFDATA WLOCK. 2) Some sort of feedback indicating that this particular entry is used so we re-send arp request to perform reachability verification instead of expiring entry. The only signal that is needed from fast path is something like binary yes/no. The latter is solved by the following changes: 1) introduce special r_skip_req field which is read lockless by fast path, but updated under (new) req_mutex mutex. If this field is non-zero, then fast path will acquire lock and set it back to 0. 2) introduce simple state machine: incomplete->reachable<->verify->deleted. Before that we implicitely had incomplete->reachable->deleted state machine, with V_arpt_keep between "reachable" and "deleted". Verification was performed in runtime 5 seconds before V_arpt_keep expire. This is changed to "change state to verify 5 seconds before V_arpt_keep, set r_skip_req to non-zero value and check it every second". If the value is zero - then send arp verification probe. These changes do not introduce any signifficant control plane overhead: typically lle callout timer would fire 1 time more each V_arpt_keep (1200s) for used lles and up to arp_maxtries (5) for dead lles. As a result, all packets towards "reachable" lle are handled by fast path without acquiring lle read lock. Additional "req_mutex" is needed because callout / arpresolve_slow() or eventhandler might keep LLE lock for signifficant amount of time, which might not be feasible for fast path locking (e.g. having rmlock as ether AFDATA or lltable own lock). Differential Revision: https://reviews.freebsd.org/D3688
2015-12-05 09:50:37 +00:00
if (la != NULL && (la->r_flags & RLLE_VALID) != 0) {
/* Entry found, let's copy lle info */
Implement interface link header precomputation API. Add if_requestencap() interface method which is capable of calculating various link headers for given interface. Right now there is support for INET/INET6/ARP llheader calculation (IFENCAP_LL type request). Other types are planned to support more complex calculation (L2 multipath lagg nexthops, tunnel encap nexthops, etc..). Reshape 'struct route' to be able to pass additional data (with is length) to prepend to mbuf. These two changes permits routing code to pass pre-calculated nexthop data (like L2 header for route w/gateway) down to the stack eliminating the need for other lookups. It also brings us closer to more complex scenarios like transparently handling MPLS nexthops and tunnel interfaces. Last, but not least, it removes layering violation introduced by flowtable code (ro_lle) and simplifies handling of existing if_output consumers. ARP/ND changes: Make arp/ndp stack pre-calculate link header upon installing/updating lle record. Interface link address change are handled by re-calculating headers for all lles based on if_lladdr event. After these changes, arpresolve()/nd6_resolve() returns full pre-calculated header for supported interfaces thus simplifying if_output(). Move these lookups to separate ether_resolve_addr() function which ether returs error or fully-prepared link header. Add <arp|nd6_>resolve_addr() compat versions to return link addresses instead of pre-calculated data. BPF changes: Raw bpf writes occupied _two_ cases: AF_UNSPEC and pseudo_AF_HDRCMPLT. Despite the naming, both of there have ther header "complete". The only difference is that interface source mac has to be filled by OS for AF_UNSPEC (controlled via BIOCGHDRCMPLT). This logic has to stay inside BPF and not pollute if_output() routines. Convert BPF to pass prepend data via new 'struct route' mechanism. Note that it does not change non-optimized if_output(): ro_prepend handling is purely optional. Side note: hackish pseudo_AF_HDRCMPLT is supported for ethernet and FDDI. It is not needed for ethernet anymore. The only remaining FDDI user is dev/pdq mostly untouched since 2007. FDDI support was eliminated from OpenBSD in 2013 (sys/net/if_fddisubr.c rev 1.65). Flowtable changes: Flowtable violates layering by saving (and not correctly managing) rtes/lles. Instead of passing lle pointer, pass pointer to pre-calculated header data from that lle. Differential Revision: https://reviews.freebsd.org/D4102
2015-12-31 05:03:27 +00:00
bcopy(la->r_linkdata, desten, la->r_hdrlen);
Split arpresolve() into fast/slow path. This change isolates the most common case (e.g. successful lookup) from more complicates scenarios. It also (tries to) make code more simple by avoiding retry: cycle. The actual goal is to prepare code to the upcoming change that will allow LL address retrieval without acquiring LLE lock at all. Reviewed by: ae Differential Revision: https://reviews.freebsd.org/D3383
2015-08-16 12:23:58 +00:00
if (pflags != NULL)
Remove LLE read lock from IPv4 fast path. LLE structure is mostly unchanged during its lifecycle. To be more specific, there are 2 things relevant for fast path lookup code: 1) link-level address change. Since r286722, these updates are performed under AFDATA WLOCK. 2) Some sort of feedback indicating that this particular entry is used so we re-send arp request to perform reachability verification instead of expiring entry. The only signal that is needed from fast path is something like binary yes/no. The latter is solved by the following changes: 1) introduce special r_skip_req field which is read lockless by fast path, but updated under (new) req_mutex mutex. If this field is non-zero, then fast path will acquire lock and set it back to 0. 2) introduce simple state machine: incomplete->reachable<->verify->deleted. Before that we implicitely had incomplete->reachable->deleted state machine, with V_arpt_keep between "reachable" and "deleted". Verification was performed in runtime 5 seconds before V_arpt_keep expire. This is changed to "change state to verify 5 seconds before V_arpt_keep, set r_skip_req to non-zero value and check it every second". If the value is zero - then send arp verification probe. These changes do not introduce any signifficant control plane overhead: typically lle callout timer would fire 1 time more each V_arpt_keep (1200s) for used lles and up to arp_maxtries (5) for dead lles. As a result, all packets towards "reachable" lle are handled by fast path without acquiring lle read lock. Additional "req_mutex" is needed because callout / arpresolve_slow() or eventhandler might keep LLE lock for signifficant amount of time, which might not be feasible for fast path locking (e.g. having rmlock as ether AFDATA or lltable own lock). Differential Revision: https://reviews.freebsd.org/D3688
2015-12-05 09:50:37 +00:00
*pflags = LLE_VALID | (la->r_flags & RLLE_IFADDR);
Fix outgoing TCP/UDP packet drop on arp/ndp entry expiration. Current arp/nd code relies on the feedback from the datapath indicating that the entry is still used. This mechanism is incorporated into the arpresolve()/nd6_resolve() routines. After the inpcb route cache introduction, the packet path for the locally-originated packets changed, passing cached lle pointer to the ether_output() directly. This resulted in the arp/ndp entry expire each time exactly after the configured max_age interval. During the small window between the ARP/NDP request and reply from the router, most of the packets got lost. Fix this behaviour by plugging datapath notification code to the packet path used by route cache. Unify the notification code by using single inlined function with the per-AF callbacks. Reported by: sthaug at nethelp.no Reviewed by: ae MFC after: 2 weeks
2018-03-17 17:05:48 +00:00
/* Notify the LLE handling code that the entry was used. */
llentry_mark_used(la);
Fix per-connection L2 caching in fast path r301217 re-added per-connection L2 caching from a previous change, but it omitted caching in the fast path. Add it. Reviewed By: gallatin Approved by: gnn (mentor) Differential Revision: https://reviews.freebsd.org/D7239
2016-07-22 02:11:49 +00:00
if (plle) {
LLE_ADDREF(la);
*plle = la;
LLE_WUNLOCK(la);
}
Split arpresolve() into fast/slow path. This change isolates the most common case (e.g. successful lookup) from more complicates scenarios. It also (tries to) make code more simple by avoiding retry: cycle. The actual goal is to prepare code to the upcoming change that will allow LL address retrieval without acquiring LLE lock at all. Reviewed by: ae Differential Revision: https://reviews.freebsd.org/D3383
2015-08-16 12:23:58 +00:00
return (0);
}
Fix per-connection L2 caching in fast path r301217 re-added per-connection L2 caching from a previous change, but it omitted caching in the fast path. Add it. Reviewed By: gallatin Approved by: gnn (mentor) Differential Revision: https://reviews.freebsd.org/D7239
2016-07-22 02:11:49 +00:00
if (plle && la)
LLE_WUNLOCK(la);
Split arpresolve() into fast/slow path. This change isolates the most common case (e.g. successful lookup) from more complicates scenarios. It also (tries to) make code more simple by avoiding retry: cycle. The actual goal is to prepare code to the upcoming change that will allow LL address retrieval without acquiring LLE lock at all. Reviewed by: ae Differential Revision: https://reviews.freebsd.org/D3383
2015-08-16 12:23:58 +00:00
Implement interface link header precomputation API. Add if_requestencap() interface method which is capable of calculating various link headers for given interface. Right now there is support for INET/INET6/ARP llheader calculation (IFENCAP_LL type request). Other types are planned to support more complex calculation (L2 multipath lagg nexthops, tunnel encap nexthops, etc..). Reshape 'struct route' to be able to pass additional data (with is length) to prepend to mbuf. These two changes permits routing code to pass pre-calculated nexthop data (like L2 header for route w/gateway) down to the stack eliminating the need for other lookups. It also brings us closer to more complex scenarios like transparently handling MPLS nexthops and tunnel interfaces. Last, but not least, it removes layering violation introduced by flowtable code (ro_lle) and simplifies handling of existing if_output consumers. ARP/ND changes: Make arp/ndp stack pre-calculate link header upon installing/updating lle record. Interface link address change are handled by re-calculating headers for all lles based on if_lladdr event. After these changes, arpresolve()/nd6_resolve() returns full pre-calculated header for supported interfaces thus simplifying if_output(). Move these lookups to separate ether_resolve_addr() function which ether returs error or fully-prepared link header. Add <arp|nd6_>resolve_addr() compat versions to return link addresses instead of pre-calculated data. BPF changes: Raw bpf writes occupied _two_ cases: AF_UNSPEC and pseudo_AF_HDRCMPLT. Despite the naming, both of there have ther header "complete". The only difference is that interface source mac has to be filled by OS for AF_UNSPEC (controlled via BIOCGHDRCMPLT). This logic has to stay inside BPF and not pollute if_output() routines. Convert BPF to pass prepend data via new 'struct route' mechanism. Note that it does not change non-optimized if_output(): ro_prepend handling is purely optional. Side note: hackish pseudo_AF_HDRCMPLT is supported for ethernet and FDDI. It is not needed for ethernet anymore. The only remaining FDDI user is dev/pdq mostly untouched since 2007. FDDI support was eliminated from OpenBSD in 2013 (sys/net/if_fddisubr.c rev 1.65). Flowtable changes: Flowtable violates layering by saving (and not correctly managing) rtes/lles. Instead of passing lle pointer, pass pointer to pre-calculated header data from that lle. Differential Revision: https://reviews.freebsd.org/D4102
2015-12-31 05:03:27 +00:00
return (arpresolve_full(ifp, is_gw, la == NULL ? LLE_CREATE : 0, m, dst,
This change re-adds L2 caching for TCP and UDP, as originally added in D4306 but removed due to other changes in the system. Restore the llentry pointer to the "struct route", and use it to cache the L2 lookup (ARP or ND6) as appropriate. Submitted by: Mike Karels Differential Revision: https://reviews.freebsd.org/D6262
2016-06-02 17:51:29 +00:00
desten, pflags, plle));
Split arpresolve() into fast/slow path. This change isolates the most common case (e.g. successful lookup) from more complicates scenarios. It also (tries to) make code more simple by avoiding retry: cycle. The actual goal is to prepare code to the upcoming change that will allow LL address retrieval without acquiring LLE lock at all. Reviewed by: ae Differential Revision: https://reviews.freebsd.org/D3383
2015-08-16 12:23:58 +00:00
}
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
/*
* Common length and type checks are done here,
* then the protocol-specific routine is called.
*/
Remove old ballast, clean up a little bit, staticize. Add five sysctl variables that you should probably never tweak. net.arp.t_prune: 300 net.arp.t_keep: 1200 net.arp.t_down: 20 net.arp.maxtries: 5 net.arp.useloopback: 1 net.arp.proxyall: 0 (It's net.arp because arp isn't limited to inet, though our present implementation surely is).
1995-12-09 16:06:54 +00:00
static void
Update netisr handling; Each SWI now registers its queue, and all queue drain routines are done by swi_net, which allows for better queue control at some future point. Packets may also be directly dispatched to a netisr instead of queued, this may be of interest at some installations, but currently defaults to off. Reviewed by: hsu, silby, jayanth, sam Sponsored by: DARPA, NAI Labs
2003-03-04 23:19:55 +00:00
arpintr(struct mbuf *m)
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
{
Update netisr handling; Each SWI now registers its queue, and all queue drain routines are done by swi_net, which allows for better queue control at some future point. Packets may also be directly dispatched to a netisr instead of queued, this may be of interest at some installations, but currently defaults to off. Reviewed by: hsu, silby, jayanth, sam Sponsored by: DARPA, NAI Labs
2003-03-04 23:19:55 +00:00
struct arphdr *ar;
* Improve logging invalid arp messages * Remove redundant check in ip_arpinput Suggested by: glebius MFC after: 2 weeks
2015-09-15 08:50:44 +00:00
struct ifnet *ifp;
* Improve error checking for arp messages. * Clean stale headers from if_ether.c. Reported by: rozhuk.im at gmail.com Reviewed by: ae MFC after: 2 weeks
2015-09-14 10:28:47 +00:00
char *layer;
int hlen;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
* Improve logging invalid arp messages * Remove redundant check in ip_arpinput Suggested by: glebius MFC after: 2 weeks
2015-09-15 08:50:44 +00:00
ifp = m->m_pkthdr.rcvif;
Update netisr handling; Each SWI now registers its queue, and all queue drain routines are done by swi_net, which allows for better queue control at some future point. Packets may also be directly dispatched to a netisr instead of queued, this may be of interest at some installations, but currently defaults to off. Reviewed by: hsu, silby, jayanth, sam Sponsored by: DARPA, NAI Labs
2003-03-04 23:19:55 +00:00
if (m->m_len < sizeof(struct arphdr) &&
((m = m_pullup(m, sizeof(struct arphdr))) == NULL)) {
* Improve logging invalid arp messages * Remove redundant check in ip_arpinput Suggested by: glebius MFC after: 2 weeks
2015-09-15 08:50:44 +00:00
ARP_LOG(LOG_NOTICE, "packet with short header received on %s\n",
if_name(ifp));
Update netisr handling; Each SWI now registers its queue, and all queue drain routines are done by swi_net, which allows for better queue control at some future point. Packets may also be directly dispatched to a netisr instead of queued, this may be of interest at some installations, but currently defaults to off. Reviewed by: hsu, silby, jayanth, sam Sponsored by: DARPA, NAI Labs
2003-03-04 23:19:55 +00:00
return;
}
ar = mtod(m, struct arphdr *);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
* Improve error checking for arp messages. * Clean stale headers from if_ether.c. Reported by: rozhuk.im at gmail.com Reviewed by: ae MFC after: 2 weeks
2015-09-14 10:28:47 +00:00
/* Check if length is sufficient */
Fix regression from r287779, that bite me. If we call m_pullup() unconditionally, we end up with an mbuf chain of two mbufs, which later in in_arpreply() is rewritten from ARP request to ARP reply and is sent out. Looks like igb(4) (at least mine, and at least at my network) fails on such mbuf chain, so ARP reply doesn't go out wire. Thus, make the m_pullup() call conditional, as it is everywhere. Of course, the bug in igb(?) should be investigated, but better first fix the head. And unconditional m_pullup() was suboptimal, anyway.
2015-10-07 13:10:26 +00:00
if (m->m_len < arphdr_len(ar)) {
m = m_pullup(m, arphdr_len(ar));
if (m == NULL) {
ARP_LOG(LOG_NOTICE, "short packet received on %s\n",
if_name(ifp));
return;
}
ar = mtod(m, struct arphdr *);
* Improve error checking for arp messages. * Clean stale headers from if_ether.c. Reported by: rozhuk.im at gmail.com Reviewed by: ae MFC after: 2 weeks
2015-09-14 10:28:47 +00:00
}
hlen = 0;
layer = "";
switch (ntohs(ar->ar_hrd)) {
case ARPHRD_ETHER:
hlen = ETHER_ADDR_LEN; /* RFC 826 */
layer = "ethernet";
break;
case ARPHRD_INFINIBAND:
White space cleanup -- remove trailing tab's or spaces from any line. Sponsored by: Netflix Inc.
2020-02-12 13:31:36 +00:00
hlen = 20; /* RFC 4391, INFINIBAND_ALEN */
* Improve error checking for arp messages. * Clean stale headers from if_ether.c. Reported by: rozhuk.im at gmail.com Reviewed by: ae MFC after: 2 weeks
2015-09-14 10:28:47 +00:00
layer = "infiniband";
break;
case ARPHRD_IEEE1394:
hlen = 0; /* SHALL be 16 */ /* RFC 2734 */
layer = "firewire";
/*
Fix typo (s/harware/hardware/)
2015-12-25 14:51:36 +00:00
* Restrict too long hardware addresses.
* Improve error checking for arp messages. * Clean stale headers from if_ether.c. Reported by: rozhuk.im at gmail.com Reviewed by: ae MFC after: 2 weeks
2015-09-14 10:28:47 +00:00
* Currently we are capable of handling 20-byte
* addresses ( sizeof(lle->ll_addr) )
*/
if (ar->ar_hln >= 20)
hlen = 16;
break;
default:
* Improve logging invalid arp messages * Remove redundant check in ip_arpinput Suggested by: glebius MFC after: 2 weeks
2015-09-15 08:50:44 +00:00
ARP_LOG(LOG_NOTICE,
Fix typo (s/harware/hardware/)
2015-12-25 14:51:36 +00:00
"packet with unknown hardware format 0x%02d received on "
"%s\n", ntohs(ar->ar_hrd), if_name(ifp));
Update netisr handling; Each SWI now registers its queue, and all queue drain routines are done by swi_net, which allows for better queue control at some future point. Packets may also be directly dispatched to a netisr instead of queued, this may be of interest at some installations, but currently defaults to off. Reviewed by: hsu, silby, jayanth, sam Sponsored by: DARPA, NAI Labs
2003-03-04 23:19:55 +00:00
m_freem(m);
return;
}
* Improve error checking for arp messages. * Clean stale headers from if_ether.c. Reported by: rozhuk.im at gmail.com Reviewed by: ae MFC after: 2 weeks
2015-09-14 10:28:47 +00:00
if (hlen != 0 && hlen != ar->ar_hln) {
* Improve logging invalid arp messages * Remove redundant check in ip_arpinput Suggested by: glebius MFC after: 2 weeks
2015-09-15 08:50:44 +00:00
ARP_LOG(LOG_NOTICE,
"packet with invalid %s address length %d received on %s\n",
layer, ar->ar_hln, if_name(ifp));
* Improve error checking for arp messages. * Clean stale headers from if_ether.c. Reported by: rozhuk.im at gmail.com Reviewed by: ae MFC after: 2 weeks
2015-09-14 10:28:47 +00:00
m_freem(m);
return;
Update netisr handling; Each SWI now registers its queue, and all queue drain routines are done by swi_net, which allows for better queue control at some future point. Packets may also be directly dispatched to a netisr instead of queued, this may be of interest at some installations, but currently defaults to off. Reviewed by: hsu, silby, jayanth, sam Sponsored by: DARPA, NAI Labs
2003-03-04 23:19:55 +00:00
}
The function arpintr() incorrectly checks m->m_len to detect incomplete ARP packets. This can incorrectly reject complete frames since the frame could be stored in more than one mbuf. The following patches fix the length comparisson, and add several diagnostic log messages to the interrupt handler for out-of-the-norm ARP packets. This should make ARP problems easier to detect, diagnose and fix. Submitted by: C. Stephen Gunn <csg@waterspout.com> Approved by: jkh Reviewed by: rwatson
2000-03-11 00:24:29 +00:00
Add ARP statistics to the kernel and netstat. New counters now exist for: requests sent replies sent requests received replies received packets received total packets dropped due to no ARP entry entrys timed out Duplicate IPs seen The new statistics are seen in the netstat command when it is given the -s command line switch. MFC after: 2 weeks In collaboration with: bz
2009-09-03 21:10:57 +00:00
ARPSTAT_INC(received);
Update netisr handling; Each SWI now registers its queue, and all queue drain routines are done by swi_net, which allows for better queue control at some future point. Packets may also be directly dispatched to a netisr instead of queued, this may be of interest at some installations, but currently defaults to off. Reviewed by: hsu, silby, jayanth, sam Sponsored by: DARPA, NAI Labs
2003-03-04 23:19:55 +00:00
switch (ntohs(ar->ar_pro)) {
Make INET a proper option. This will not make any of object files that LINT create change; there might be differences with INET disabled, but hardly anything compiled before without INET anyway. Now the 'obvious' things will give a proper error if compiled without inet - ipx_ip, ipfw, tcp_debug. The only thing that _should_ work (but can't be made to compile reasonably easily) is sppp :-( This commit move struct arpcom from <netinet/if_ether.h> to <net/if_arp.h>.
1998-01-08 23:42:31 +00:00
#ifdef INET
Update netisr handling; Each SWI now registers its queue, and all queue drain routines are done by swi_net, which allows for better queue control at some future point. Packets may also be directly dispatched to a netisr instead of queued, this may be of interest at some installations, but currently defaults to off. Reviewed by: hsu, silby, jayanth, sam Sponsored by: DARPA, NAI Labs
2003-03-04 23:19:55 +00:00
case ETHERTYPE_IP:
in_arpinput(m);
return;
Make INET a proper option. This will not make any of object files that LINT create change; there might be differences with INET disabled, but hardly anything compiled before without INET anyway. Now the 'obvious' things will give a proper error if compiled without inet - ipx_ip, ipfw, tcp_debug. The only thing that _should_ work (but can't be made to compile reasonably easily) is sppp :-( This commit move struct arpcom from <netinet/if_ether.h> to <net/if_arp.h>.
1998-01-08 23:42:31 +00:00
#endif
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
Update netisr handling; Each SWI now registers its queue, and all queue drain routines are done by swi_net, which allows for better queue control at some future point. Packets may also be directly dispatched to a netisr instead of queued, this may be of interest at some installations, but currently defaults to off. Reviewed by: hsu, silby, jayanth, sam Sponsored by: DARPA, NAI Labs
2003-03-04 23:19:55 +00:00
m_freem(m);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
Make INET a proper option. This will not make any of object files that LINT create change; there might be differences with INET disabled, but hardly anything compiled before without INET anyway. Now the 'obvious' things will give a proper error if compiled without inet - ipx_ip, ipfw, tcp_debug. The only thing that _should_ work (but can't be made to compile reasonably easily) is sppp :-( This commit move struct arpcom from <netinet/if_ether.h> to <net/if_arp.h>.
1998-01-08 23:42:31 +00:00
#ifdef INET
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
/*
* ARP for Internet protocols on 10 Mb/s Ethernet.
* Algorithm is that given in RFC 826.
* In addition, a sanity check is performed on the sender
* protocol address, to catch impersonators.
* We no longer handle negotiations for use of trailer protocol:
* Formerly, ARP replied for protocol type ETHERTYPE_TRAIL sent
* along with IP replies if we wanted trailers sent to us,
* and also sent them in response to IP replies.
* This allowed either end to announce the desire to receive
* trailer packets.
* We no longer reply to requests for ETHERTYPE_TRAIL protocol either,
* but formerly didn't normally send requests.
*/
provide a sysctl 'net.link.ether.inet.log_arp_wrong_iface' to allow one to supress logging when ARP replies arrive on the wrong interface: "/kernel: arp: 1.2.3.4 is on dc0 but got reply from 00:00:c5:79:d0:0c on dc1" the default is to log just to give notice about possibly incorrectly configured networks.
2001-01-06 00:45:08 +00:00
static int log_arp_wrong_iface = 1;
Allow disabling of "arp moved" messages. Submitted by: Stephen Hurd <deuce@lordlegacy.org>
2001-09-03 21:53:15 +00:00
static int log_arp_movements = 1;
Add a knob to suppress logging of attempts to modify permanent ARP entries. Submitted by: Andrew Alcheyev <buddy telenet.ru>
2005-12-18 19:11:56 +00:00
static int log_arp_permanent_modify = 1;
Provide a sysctl switch that allows to install ARP entries with multicast bit set. FreeBSD refuses to install such entries since 9.0, and this broke installations running Microsoft NLB, which are violating standards. Tested by: Tarasov Oleg <oleg_tarasov sg-tea.com>
2012-09-03 14:29:28 +00:00
static int allow_multicast = 0;
provide a sysctl 'net.link.ether.inet.log_arp_wrong_iface' to allow one to supress logging when ARP replies arrive on the wrong interface: "/kernel: arp: 1.2.3.4 is on dc0 but got reply from 00:00:c5:79:d0:0c on dc1" the default is to log just to give notice about possibly incorrectly configured networks.
2001-01-06 00:45:08 +00:00
SYSCTL_INT(_net_link_ether_inet, OID_AUTO, log_arp_wrong_iface, CTLFLAG_RW,
&log_arp_wrong_iface, 0,
"log arp packets arriving on the wrong interface");
Allow disabling of "arp moved" messages. Submitted by: Stephen Hurd <deuce@lordlegacy.org>
2001-09-03 21:53:15 +00:00
SYSCTL_INT(_net_link_ether_inet, OID_AUTO, log_arp_movements, CTLFLAG_RW,
Some style(9) and whitespace changes. Together with: Andrey Zonov <andrey zonov.org>
2012-07-31 11:31:12 +00:00
&log_arp_movements, 0,
"log arp replies from MACs different than the one in the cache");
Add a knob to suppress logging of attempts to modify permanent ARP entries. Submitted by: Andrew Alcheyev <buddy telenet.ru>
2005-12-18 19:11:56 +00:00
SYSCTL_INT(_net_link_ether_inet, OID_AUTO, log_arp_permanent_modify, CTLFLAG_RW,
Some style(9) and whitespace changes. Together with: Andrey Zonov <andrey zonov.org>
2012-07-31 11:31:12 +00:00
&log_arp_permanent_modify, 0,
"log arp replies from MACs different than the one in the permanent arp entry");
Provide a sysctl switch that allows to install ARP entries with multicast bit set. FreeBSD refuses to install such entries since 9.0, and this broke installations running Microsoft NLB, which are violating standards. Tested by: Tarasov Oleg <oleg_tarasov sg-tea.com>
2012-09-03 14:29:28 +00:00
SYSCTL_INT(_net_link_ether_inet, OID_AUTO, allow_multicast, CTLFLAG_RW,
&allow_multicast, 0, "accept multicast addresses");
provide a sysctl 'net.link.ether.inet.log_arp_wrong_iface' to allow one to supress logging when ARP replies arrive on the wrong interface: "/kernel: arp: 1.2.3.4 is on dc0 but got reply from 00:00:c5:79:d0:0c on dc1" the default is to log just to give notice about possibly incorrectly configured networks.
2001-01-06 00:45:08 +00:00
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
static void
Move universally to ANSI C function declarations, with relatively consistent style(9)-ish layout.
2007-05-10 15:58:48 +00:00
in_arpinput(struct mbuf *m)
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
{
Convert in_ifaddr_lock and in6_ifaddr_lock to rmlock. Both are used to protect access to IP addresses lists and they can be acquired for reading several times per packet. To reduce lock contention it is better to use rmlock here. Reviewed by: gnn (previous version) Obtained from: Yandex LLC Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D3149
2015-07-29 08:12:05 +00:00
struct rm_priotracker in_ifa_tracker;
De-register.
2004-03-14 00:44:11 +00:00
struct arphdr *ah;
struct ifnet *ifp = m->m_pkthdr.rcvif;
* Split allocation and table linking for lle's. Before that, the logic besides lle_create() was the following: return existing if found, create if not. This behaviour was error-prone since we had to deal with 'sudden' static<>dynamic lle changes. This commit fixes bunch of different issues like: - refcount leak when lle is converted to static. Simple check case: console 1: while true; do for i in `arp -an|awk '$4~/incomp/{print$2}'|tr -d '()'`; do arp -s $i 00:22:44:66:88:00 ; arp -d $i; done; done console 2: ping -f any-dead-host-in-L2 console 3: # watch for memory consumption: vmstat -m | awk '$1~/lltable/{print$2}' - possible problems in arptimer() / nd6_timer() when dropping/reacquiring lock. New logic explicitly handles use-or-create cases in every lla_create user. Basically, most of the changes are purely mechanical. However, we explicitly avoid using existing lle's for interface/static LLE records. * While here, call lle_event handlers on all real table lle change. * Create lltable_free_entry() calling existing per-lltable lle_free_t callback for entry deletion
2015-08-20 12:05:17 +00:00
struct llentry *la = NULL, *la_tmp;
Add a hash table that contains the list of internet addresses, and use this in place of the in_ifaddr list when appropriate. This improves performance on hosts which have a large number of IP aliases.
2001-09-29 04:34:11 +00:00
struct ifaddr *ifa;
struct in_ifaddr *ia;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
struct sockaddr sa;
struct in_addr isaddr, itaddr, myaddr;
Add CARP (Common Address Redundancy Protocol), which allows multiple hosts to share an IP address, providing high availability and load balancing. Original work on CARP done by Michael Shalayeff, with many additions by Marco Pfatschbacher and Ryan McBride. FreeBSD port done solely by Max Laier. Patch by: mlaier Obtained from: OpenBSD (mickey, mcbride)
2005-02-22 13:04:05 +00:00
u_int8_t *enaddr = NULL;
Move lle update code from from gigantic ip_arpinput() to separate bunch of functions. The goal is to isolate actual lle updates to permit more fine-grained locking. Do all lle link-level update under AFDATA wlock. Sponsored by: Yandex LLC
2015-08-13 13:38:09 +00:00
int op;
Fix ARP in bridging scenarios where the bridge shares its MAC address with one of its members (see my r180140). Pointy hat to: philip Submitted by: Eygene Ryabinkin <rea-fbsd@codelabs.ru> MFC after: 3 days
2008-08-18 09:06:11 +00:00
int bridged = 0, is_bridge = 0;
Move lle update code from from gigantic ip_arpinput() to separate bunch of functions. The goal is to isolate actual lle updates to permit more fine-grained locking. Do all lle link-level update under AFDATA wlock. Sponsored by: Yandex LLC
2015-08-13 13:38:09 +00:00
int carped;
add interface for allowing consumers to register for ARP updates, redirects, and path MTU changes Reviewed by: silby
2007-12-12 20:53:25 +00:00
struct sockaddr_in sin;
* Split allocation and table linking for lle's. Before that, the logic besides lle_create() was the following: return existing if found, create if not. This behaviour was error-prone since we had to deal with 'sudden' static<>dynamic lle changes. This commit fixes bunch of different issues like: - refcount leak when lle is converted to static. Simple check case: console 1: while true; do for i in `arp -an|awk '$4~/incomp/{print$2}'|tr -d '()'`; do arp -s $i 00:22:44:66:88:00 ; arp -d $i; done; done console 2: ping -f any-dead-host-in-L2 console 3: # watch for memory consumption: vmstat -m | awk '$1~/lltable/{print$2}' - possible problems in arptimer() / nd6_timer() when dropping/reacquiring lock. New logic explicitly handles use-or-create cases in every lla_create user. Basically, most of the changes are purely mechanical. However, we explicitly avoid using existing lle's for interface/static LLE records. * While here, call lle_event handlers on all real table lle change. * Create lltable_free_entry() calling existing per-lltable lle_free_t callback for entry deletion
2015-08-20 12:05:17 +00:00
struct sockaddr *dst;
Make in_arpinput(), inp_lookup_mcast_ifp(), icmp_reflect(), ip_dooptions(), icmp6_redirect_input(), in6_lltable_rtcheck(), in6p_lookup_mcast_ifp() and in6_selecthlim() use new routing api. Eliminate now-unused ip_rtaddr(). Fix lookup key fib6_lookup_nh_basic() which was lost diring merge. Make fib6_lookup_nh_basic() and fib6_lookup_nh_extended() always return IPv6 destination address with embedded scope. Currently rw_gateway has it scope embedded, do the same for non-gatewayed destinations. Sponsored by: Yandex LLC
2015-12-09 11:14:27 +00:00
struct nhop4_basic nh4;
Implement interface link header precomputation API. Add if_requestencap() interface method which is capable of calculating various link headers for given interface. Right now there is support for INET/INET6/ARP llheader calculation (IFENCAP_LL type request). Other types are planned to support more complex calculation (L2 multipath lagg nexthops, tunnel encap nexthops, etc..). Reshape 'struct route' to be able to pass additional data (with is length) to prepend to mbuf. These two changes permits routing code to pass pre-calculated nexthop data (like L2 header for route w/gateway) down to the stack eliminating the need for other lookups. It also brings us closer to more complex scenarios like transparently handling MPLS nexthops and tunnel interfaces. Last, but not least, it removes layering violation introduced by flowtable code (ro_lle) and simplifies handling of existing if_output consumers. ARP/ND changes: Make arp/ndp stack pre-calculate link header upon installing/updating lle record. Interface link address change are handled by re-calculating headers for all lles based on if_lladdr event. After these changes, arpresolve()/nd6_resolve() returns full pre-calculated header for supported interfaces thus simplifying if_output(). Move these lookups to separate ether_resolve_addr() function which ether returs error or fully-prepared link header. Add <arp|nd6_>resolve_addr() compat versions to return link addresses instead of pre-calculated data. BPF changes: Raw bpf writes occupied _two_ cases: AF_UNSPEC and pseudo_AF_HDRCMPLT. Despite the naming, both of there have ther header "complete". The only difference is that interface source mac has to be filled by OS for AF_UNSPEC (controlled via BIOCGHDRCMPLT). This logic has to stay inside BPF and not pollute if_output() routines. Convert BPF to pass prepend data via new 'struct route' mechanism. Note that it does not change non-optimized if_output(): ro_prepend handling is purely optional. Side note: hackish pseudo_AF_HDRCMPLT is supported for ethernet and FDDI. It is not needed for ethernet anymore. The only remaining FDDI user is dev/pdq mostly untouched since 2007. FDDI support was eliminated from OpenBSD in 2013 (sys/net/if_fddisubr.c rev 1.65). Flowtable changes: Flowtable violates layering by saving (and not correctly managing) rtes/lles. Instead of passing lle pointer, pass pointer to pre-calculated header data from that lle. Differential Revision: https://reviews.freebsd.org/D4102
2015-12-31 05:03:27 +00:00
uint8_t linkhdr[LLE_MAX_LINKHDR];
struct route ro;
size_t linkhdrsize;
int lladdr_off;
int error;
Use inet_ntoa_r() instead of inet_ntoa() throughout the kernel inet_ntoa() cannot be used safely in a multithreaded environment because it uses a static local buffer. Instead, use inet_ntoa_r() with a buffer on the caller's stack. Suggested by: glebius, emaste Reviewed by: gnn MFC after: 2 weeks Sponsored by: Dell EMC Differential Revision: https://reviews.freebsd.org/D9625
2017-02-16 20:47:41 +00:00
char addrbuf[INET_ADDRSTRLEN];
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
NET_EPOCH_ASSERT();
Make in_arpinput(), inp_lookup_mcast_ifp(), icmp_reflect(), ip_dooptions(), icmp6_redirect_input(), in6_lltable_rtcheck(), in6p_lookup_mcast_ifp() and in6_selecthlim() use new routing api. Eliminate now-unused ip_rtaddr(). Fix lookup key fib6_lookup_nh_basic() which was lost diring merge. Make fib6_lookup_nh_basic() and fib6_lookup_nh_extended() always return IPv6 destination address with embedded scope. Currently rw_gateway has it scope embedded, do the same for non-gatewayed destinations. Sponsored by: Yandex LLC
2015-12-09 11:14:27 +00:00
add interface for allowing consumers to register for ARP updates, redirects, and path MTU changes Reviewed by: silby
2007-12-12 20:53:25 +00:00
sin.sin_len = sizeof(struct sockaddr_in);
sin.sin_family = AF_INET;
widen the routing event interface (arp update, redirect, and eventually pmtu change) into separate functions revert previous commit's changes to arpresolve and add a new interface arpresolve2 which does arp resolution without an mbuf
2007-12-17 07:40:34 +00:00
sin.sin_addr.s_addr = 0;
Step 1.5 of importing the network stack virtualization infrastructure from the vimage project, as per plan established at devsummit 08/08: http://wiki.freebsd.org/Image/Notes200808DevSummit Introduce INIT_VNET_*() initializer macros, VNET_FOREACH() iterator macros, and CURVNET_SET() context setting macros, all currently resolving to NOPs. Prepare for virtualization of selected SYSCTL objects by introducing a family of SYSCTL_V_*() macros, currently resolving to their global counterparts, i.e. SYSCTL_V_INT() == SYSCTL_INT(). Move selected #defines from sys/sys/vimage.h to newly introduced header files specific to virtualized subsystems (sys/net/vnet.h, sys/netinet/vinet.h etc.). All the changes are verified to have zero functional impact at this point in time by doing MD5 comparision between pre- and post-change object files(*). (*) netipsec/keysock.c did not validate depending on compile time options. Implemented by: julian, bz, brooks, zec Reviewed by: julian, bz, brooks, kris, rwatson, ... Approved by: julian (mentor) Obtained from: //depot/projects/vimage-commit2/... X-MFC after: never Sponsored by: NLnet Foundation, The FreeBSD Foundation
2008-10-02 15:37:58 +00:00
Back out of r1.148, it causes two arp replies to be sent with different mac addresses. One for the bridged interface with the IP address assigned but then another with the mac for the bridge itself.
2006-01-29 23:21:01 +00:00
if (ifp->if_bridge)
Add hooks into the networking layer to support if_bridge. This changes struct ifnet so a buildworld is necessary. Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 03:13:13 +00:00
bridged = 1;
Fix ARP in bridging scenarios where the bridge shares its MAC address with one of its members (see my r180140). Pointy hat to: philip Submitted by: Eygene Ryabinkin <rea-fbsd@codelabs.ru> MFC after: 3 days
2008-08-18 09:06:11 +00:00
if (ifp->if_type == IFT_BRIDGE)
is_bridge = 1;
Add hooks into the networking layer to support if_bridge. This changes struct ifnet so a buildworld is necessary. Approved by: mlaier (mentor) Obtained from: NetBSD
2005-06-05 03:13:13 +00:00
* Improve logging invalid arp messages * Remove redundant check in ip_arpinput Suggested by: glebius MFC after: 2 weeks
2015-09-15 08:50:44 +00:00
/*
* We already have checked that mbuf contains enough contiguous data
* to hold entire arp message according to the arp header.
*/
bring in ARP support for variable length link level addresses Reviewed by: jdp Approved by: jdp Obtained from: NetBSD MFC after: 6 weeks
2001-10-14 20:17:53 +00:00
ah = mtod(m, struct arphdr *);
* Improve logging invalid arp messages * Remove redundant check in ip_arpinput Suggested by: glebius MFC after: 2 weeks
2015-09-15 08:50:44 +00:00
Some style(9) and whitespace changes. Together with: Andrey Zonov <andrey zonov.org>
2012-07-31 11:31:12 +00:00
/*
Fix several bugs in the ARP code related to improperly formatted packets. *) Reject requests with a protocol length not equal to 4. This is IPv4 and there is no reason to accept anything else. *) Reject packets that have a multicast source hardware address. *) Drop requests where the hardware address length is not equal to the hardware address length of the interface. Pointed out by: Rozhuk Ivan MFC after: 1 week
2011-01-12 19:11:17 +00:00
* ARP is only for IPv4 so we can reject packets with
* a protocol length not equal to an IPv4 address.
*/
if (ah->ar_pln != sizeof(struct in_addr)) {
Rate limit the number of remotely triggered ARP log messages to 1 log message per second.
2013-05-11 10:51:32 +00:00
ARP_LOG(LOG_NOTICE, "requested protocol length != %zu\n",
Fix several bugs in the ARP code related to improperly formatted packets. *) Reject requests with a protocol length not equal to 4. This is IPv4 and there is no reason to accept anything else. *) Reject packets that have a multicast source hardware address. *) Drop requests where the hardware address length is not equal to the hardware address length of the interface. Pointed out by: Rozhuk Ivan MFC after: 1 week
2011-01-12 19:11:17 +00:00
sizeof(struct in_addr));
Fix couple of mbuf leaks in incoming ARP processing.
2013-04-25 17:38:04 +00:00
goto drop;
Fix several bugs in the ARP code related to improperly formatted packets. *) Reject requests with a protocol length not equal to 4. This is IPv4 and there is no reason to accept anything else. *) Reject packets that have a multicast source hardware address. *) Drop requests where the hardware address length is not equal to the hardware address length of the interface. Pointed out by: Rozhuk Ivan MFC after: 1 week
2011-01-12 19:11:17 +00:00
}
Provide a sysctl switch that allows to install ARP entries with multicast bit set. FreeBSD refuses to install such entries since 9.0, and this broke installations running Microsoft NLB, which are violating standards. Tested by: Tarasov Oleg <oleg_tarasov sg-tea.com>
2012-09-03 14:29:28 +00:00
if (allow_multicast == 0 && ETHER_IS_MULTICAST(ar_sha(ah))) {
Rate limit the number of remotely triggered ARP log messages to 1 log message per second.
2013-05-11 10:51:32 +00:00
ARP_LOG(LOG_NOTICE, "%*D is multicast\n",
- Reduce severity for all ARP events, that can be triggered from remote machine to LOG_NOTICE. Exception left to "using my IP address". - Fix multicast ARP warning: add newline and also log the bad MAC address. Tested by: Alexander Wittig <wittigal msu.edu>
2011-11-21 12:07:18 +00:00
ifp->if_addrlen, (u_char *)ar_sha(ah), ":");
Fix couple of mbuf leaks in incoming ARP processing.
2013-04-25 17:38:04 +00:00
goto drop;
Fix several bugs in the ARP code related to improperly formatted packets. *) Reject requests with a protocol length not equal to 4. This is IPv4 and there is no reason to accept anything else. *) Reject packets that have a multicast source hardware address. *) Drop requests where the hardware address length is not equal to the hardware address length of the interface. Pointed out by: Rozhuk Ivan MFC after: 1 week
2011-01-12 19:11:17 +00:00
}
bring in ARP support for variable length link level addresses Reviewed by: jdp Approved by: jdp Obtained from: NetBSD MFC after: 6 weeks
2001-10-14 20:17:53 +00:00
op = ntohs(ah->ar_op);
(void)memcpy(&isaddr, ar_spa(ah), sizeof (isaddr));
(void)memcpy(&itaddr, ar_tpa(ah), sizeof (itaddr));
Check flag do_bridge always, even if kernel was compiled without BRIDGE support. This makes dynamic bridge.ko working. Reviewed by: sam Approved by: julian (mentor) MFC after: 1 week
2004-09-09 12:34:07 +00:00
Add ARP statistics to the kernel and netstat. New counters now exist for: requests sent replies sent requests received replies received packets received total packets dropped due to no ARP entry entrys timed out Duplicate IPs seen The new statistics are seen in the netstat command when it is given the -s command line switch. MFC after: 2 weeks In collaboration with: bz
2009-09-03 21:10:57 +00:00
if (op == ARPOP_REPLY)
ARPSTAT_INC(rxreplies);
Add a hash table that contains the list of internet addresses, and use this in place of the in_ifaddr list when appropriate. This improves performance on hosts which have a large number of IP aliases.
2001-09-29 04:34:11 +00:00
/*
* For a bridge, we want to check the address irrespective
* of the receive interface. (This will change slightly
* when we have clusters of interfaces).
*/
Convert in_ifaddr_lock and in6_ifaddr_lock to rmlock. Both are used to protect access to IP addresses lists and they can be acquired for reading several times per packet. To reduce lock contention it is better to use rmlock here. Reviewed by: gnn (previous version) Obtained from: Yandex LLC Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D3149
2015-07-29 08:12:05 +00:00
IN_IFADDR_RLOCK(&in_ifa_tracker);
Make ARP do not complain about wrong interface if correct interface is a carp one and address matched it. Reviewed by: brooks
2005-03-09 10:00:01 +00:00
LIST_FOREACH(ia, INADDR_HASH(itaddr.s_addr), ia_hash) {
When matching an incoming ARP against a bridge, ensure both interfaces belong to the same bridge. Submitted by: Alexander Zagrebin
2011-01-25 17:15:23 +00:00
if (((bridged && ia->ia_ifp->if_bridge == ifp->if_bridge) ||
This main goals of this project are: 1. separating L2 tables (ARP, NDP) from the L3 routing tables 2. removing as much locking dependencies among these layers as possible to allow for some parallelism in the search operations 3. simplify the logic in the routing code, The most notable end result is the obsolescent of the route cloning (RTF_CLONING) concept, which translated into code reduction in both IPv4 ARP and IPv6 NDP related modules, and size reduction in struct rtentry{}. The change in design obsoletes the semantics of RTF_CLONING, RTF_WASCLONE and RTF_LLINFO routing flags. The userland applications such as "arp" and "ndp" have been modified to reflect those changes. The output from "netstat -r" shows only the routing entries. Quite a few developers have contributed to this project in the past: Glebius Smirnoff, Luigi Rizzo, Alessandro Cerri, and Andre Oppermann. And most recently: - Kip Macy revised the locking code completely, thus completing the last piece of the puzzle, Kip has also been conducting active functional testing - Sam Leffler has helped me improving/refactoring the code, and provided valuable reviews - Julian Elischer setup the perforce tree for me and has helped me maintaining that branch before the svn conversion
2008-12-15 06:10:57 +00:00
ia->ia_ifp == ifp) &&
A major overhaul of the CARP implementation. The ip_carp.c was started from scratch, copying needed functionality from the old implemenation on demand, with a thorough review of all code. The main change is that interface layer has been removed from the CARP. Now redundant addresses are configured exactly on the interfaces, they run on. The CARP configuration itself is, as before, configured and read via SIOCSVH/SIOCGVH ioctls. A new prefix created with SIOCAIFADDR or SIOCAIFADDR_IN6 may now be configured to a particular virtual host id, which makes the prefix redundant. ifconfig(8) semantics has been changed too: now one doesn't need to clone carpXX interface, he/she should directly configure a vhid on a Ethernet interface. To supply vhid data from the kernel to an application the getifaddrs(8) function had been changed to pass ifam_data with each address. [1] The new implementation definitely closes all PRs related to carp(4) being an interface, and may close several others. It also allows to run a single redundant IP per interface. Big thanks to Bjoern Zeeb for his help with inet6 part of patch, for idea on using ifam_data and for several rounds of reviewing! PR: kern/117000, kern/126945, kern/126714, kern/120130, kern/117448 Reviewed by: bz Submitted by: bz [1]
2011-12-16 12:16:56 +00:00
itaddr.s_addr == ia->ia_addr.sin_addr.s_addr &&
(ia->ia_ifa.ifa_carp == NULL ||
(*carp_iamatch_p)(&ia->ia_ifa, &enaddr))) {
In ARP input, more consistently acquire and release ifaddr references. MFC after: 6 weeks
2009-06-24 10:33:35 +00:00
ifa_ref(&ia->ia_ifa);
Convert in_ifaddr_lock and in6_ifaddr_lock to rmlock. Both are used to protect access to IP addresses lists and they can be acquired for reading several times per packet. To reduce lock contention it is better to use rmlock here. Reviewed by: gnn (previous version) Obtained from: Yandex LLC Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D3149
2015-07-29 08:12:05 +00:00
IN_IFADDR_RUNLOCK(&in_ifa_tracker);
Add a hash table that contains the list of internet addresses, and use this in place of the in_ifaddr list when appropriate. This improves performance on hosts which have a large number of IP aliases.
2001-09-29 04:34:11 +00:00
goto match;
Make ARP do not complain about wrong interface if correct interface is a carp one and address matched it. Reviewed by: brooks
2005-03-09 10:00:01 +00:00
}
}
Add a hash table that contains the list of internet addresses, and use this in place of the in_ifaddr list when appropriate. This improves performance on hosts which have a large number of IP aliases.
2001-09-29 04:34:11 +00:00
LIST_FOREACH(ia, INADDR_HASH(isaddr.s_addr), ia_hash)
When matching an incoming ARP against a bridge, ensure both interfaces belong to the same bridge. Submitted by: Alexander Zagrebin
2011-01-25 17:15:23 +00:00
if (((bridged && ia->ia_ifp->if_bridge == ifp->if_bridge) ||
This main goals of this project are: 1. separating L2 tables (ARP, NDP) from the L3 routing tables 2. removing as much locking dependencies among these layers as possible to allow for some parallelism in the search operations 3. simplify the logic in the routing code, The most notable end result is the obsolescent of the route cloning (RTF_CLONING) concept, which translated into code reduction in both IPv4 ARP and IPv6 NDP related modules, and size reduction in struct rtentry{}. The change in design obsoletes the semantics of RTF_CLONING, RTF_WASCLONE and RTF_LLINFO routing flags. The userland applications such as "arp" and "ndp" have been modified to reflect those changes. The output from "netstat -r" shows only the routing entries. Quite a few developers have contributed to this project in the past: Glebius Smirnoff, Luigi Rizzo, Alessandro Cerri, and Andre Oppermann. And most recently: - Kip Macy revised the locking code completely, thus completing the last piece of the puzzle, Kip has also been conducting active functional testing - Sam Leffler has helped me improving/refactoring the code, and provided valuable reviews - Julian Elischer setup the perforce tree for me and has helped me maintaining that branch before the svn conversion
2008-12-15 06:10:57 +00:00
ia->ia_ifp == ifp) &&
In ARP input, more consistently acquire and release ifaddr references. MFC after: 6 weeks
2009-06-24 10:33:35 +00:00
isaddr.s_addr == ia->ia_addr.sin_addr.s_addr) {
ifa_ref(&ia->ia_ifa);
Convert in_ifaddr_lock and in6_ifaddr_lock to rmlock. Both are used to protect access to IP addresses lists and they can be acquired for reading several times per packet. To reduce lock contention it is better to use rmlock here. Reviewed by: gnn (previous version) Obtained from: Yandex LLC Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D3149
2015-07-29 08:12:05 +00:00
IN_IFADDR_RUNLOCK(&in_ifa_tracker);
Add a hash table that contains the list of internet addresses, and use this in place of the in_ifaddr list when appropriate. This improves performance on hosts which have a large number of IP aliases.
2001-09-29 04:34:11 +00:00
goto match;
In ARP input, more consistently acquire and release ifaddr references. MFC after: 6 weeks
2009-06-24 10:33:35 +00:00
}
Fix ARP in bridging scenarios where the bridge shares its MAC address with one of its members (see my r180140). Pointy hat to: philip Submitted by: Eygene Ryabinkin <rea-fbsd@codelabs.ru> MFC after: 3 days
2008-08-18 09:06:11 +00:00
#define BDG_MEMBER_MATCHES_ARP(addr, ifp, ia) \
(ia->ia_ifp->if_bridge == ifp->if_softc && \
!bcmp(IF_LLADDR(ia->ia_ifp), IF_LLADDR(ifp), ifp->if_addrlen) && \
addr == ia->ia_addr.sin_addr.s_addr)
/*
* Check the case when bridge shares its MAC address with
* some of its children, so packets are claimed by bridge
* itself (bridge_input() does it first), but they are really
* meant to be destined to the bridge member.
*/
if (is_bridge) {
LIST_FOREACH(ia, INADDR_HASH(itaddr.s_addr), ia_hash) {
if (BDG_MEMBER_MATCHES_ARP(itaddr.s_addr, ifp, ia)) {
In ARP input, more consistently acquire and release ifaddr references. MFC after: 6 weeks
2009-06-24 10:33:35 +00:00
ifa_ref(&ia->ia_ifa);
Fix ARP in bridging scenarios where the bridge shares its MAC address with one of its members (see my r180140). Pointy hat to: philip Submitted by: Eygene Ryabinkin <rea-fbsd@codelabs.ru> MFC after: 3 days
2008-08-18 09:06:11 +00:00
ifp = ia->ia_ifp;
Convert in_ifaddr_lock and in6_ifaddr_lock to rmlock. Both are used to protect access to IP addresses lists and they can be acquired for reading several times per packet. To reduce lock contention it is better to use rmlock here. Reviewed by: gnn (previous version) Obtained from: Yandex LLC Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D3149
2015-07-29 08:12:05 +00:00
IN_IFADDR_RUNLOCK(&in_ifa_tracker);
Fix ARP in bridging scenarios where the bridge shares its MAC address with one of its members (see my r180140). Pointy hat to: philip Submitted by: Eygene Ryabinkin <rea-fbsd@codelabs.ru> MFC after: 3 days
2008-08-18 09:06:11 +00:00
goto match;
}
}
}
#undef BDG_MEMBER_MATCHES_ARP
Convert in_ifaddr_lock and in6_ifaddr_lock to rmlock. Both are used to protect access to IP addresses lists and they can be acquired for reading several times per packet. To reduce lock contention it is better to use rmlock here. Reviewed by: gnn (previous version) Obtained from: Yandex LLC Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D3149
2015-07-29 08:12:05 +00:00
IN_IFADDR_RUNLOCK(&in_ifa_tracker);
Fix ARP in bridging scenarios where the bridge shares its MAC address with one of its members (see my r180140). Pointy hat to: philip Submitted by: Eygene Ryabinkin <rea-fbsd@codelabs.ru> MFC after: 3 days
2008-08-18 09:06:11 +00:00
Add a hash table that contains the list of internet addresses, and use this in place of the in_ifaddr list when appropriate. This improves performance on hosts which have a large number of IP aliases.
2001-09-29 04:34:11 +00:00
/*
Only examine inet addresses of the interface. This was broken in r1.83, with the result that the system would reply to an ARP request of 0.0.0.0
2001-10-20 05:14:06 +00:00
* No match, use the first inet address on the receive interface
Add a hash table that contains the list of internet addresses, and use this in place of the in_ifaddr list when appropriate. This improves performance on hosts which have a large number of IP aliases.
2001-09-29 04:34:11 +00:00
* as a dummy address for the rest of the function.
*/
ifnet: Replace if_addr_lock rwlock with epoch + mutex Run on LLNW canaries and tested by pho@ gallatin: Using a 14-core, 28-HTT single socket E5-2697 v3 with a 40GbE MLX5 based ConnectX 4-LX NIC, I see an almost 12% improvement in received packet rate, and a larger improvement in bytes delivered all the way to userspace. When the host receiving 64 streams of netperf -H $DUT -t UDP_STREAM -- -m 1, I see, using nstat -I mce0 1 before the patch: InMpps OMpps InGbs OGbs err TCP Est %CPU syscalls csw irq GBfree 4.98 0.00 4.42 0.00 4235592 33 83.80 4720653 2149771 1235 247.32 4.73 0.00 4.20 0.00 4025260 33 82.99 4724900 2139833 1204 247.32 4.72 0.00 4.20 0.00 4035252 33 82.14 4719162 2132023 1264 247.32 4.71 0.00 4.21 0.00 4073206 33 83.68 4744973 2123317 1347 247.32 4.72 0.00 4.21 0.00 4061118 33 80.82 4713615 2188091 1490 247.32 4.72 0.00 4.21 0.00 4051675 33 85.29 4727399 2109011 1205 247.32 4.73 0.00 4.21 0.00 4039056 33 84.65 4724735 2102603 1053 247.32 After the patch InMpps OMpps InGbs OGbs err TCP Est %CPU syscalls csw irq GBfree 5.43 0.00 4.20 0.00 3313143 33 84.96 5434214 1900162 2656 245.51 5.43 0.00 4.20 0.00 3308527 33 85.24 5439695 1809382 2521 245.51 5.42 0.00 4.19 0.00 3316778 33 87.54 5416028 1805835 2256 245.51 5.42 0.00 4.19 0.00 3317673 33 90.44 5426044 1763056 2332 245.51 5.42 0.00 4.19 0.00 3314839 33 88.11 5435732 1792218 2499 245.52 5.44 0.00 4.19 0.00 3293228 33 91.84 5426301 1668597 2121 245.52 Similarly, netperf reports 230Mb/s before the patch, and 270Mb/s after the patch Reviewed by: gallatin Sponsored by: Limelight Networks Differential Revision: https://reviews.freebsd.org/D15366
2018-05-18 20:13:34 +00:00
CK_STAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link)
Don't fallback to a CARP address in BACKUP state.
2011-12-29 15:59:14 +00:00
if (ifa->ifa_addr->sa_family == AF_INET &&
(ifa->ifa_carp == NULL ||
(*carp_iamatch_p)(ifa, &enaddr))) {
If we are bridging, fall back to using any inet address in the system, irrespective of receive interface, as a last resort. Submitted by: ru
2001-10-25 06:14:21 +00:00
ia = ifatoia(ifa);
In ARP input, more consistently acquire and release ifaddr references. MFC after: 6 weeks
2009-06-24 10:33:35 +00:00
ifa_ref(ifa);
If we are bridging, fall back to using any inet address in the system, irrespective of receive interface, as a last resort. Submitted by: ru
2001-10-25 06:14:21 +00:00
goto match;
}
In ARP input, more consistently acquire and release ifaddr references. MFC after: 6 weeks
2009-06-24 10:33:35 +00:00
If we are bridging, fall back to using any inet address in the system, irrespective of receive interface, as a last resort. Submitted by: ru
2001-10-25 06:14:21 +00:00
/*
* If bridging, fall back to using any inet address.
*/
Convert in_ifaddr_lock and in6_ifaddr_lock to rmlock. Both are used to protect access to IP addresses lists and they can be acquired for reading several times per packet. To reduce lock contention it is better to use rmlock here. Reviewed by: gnn (previous version) Obtained from: Yandex LLC Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D3149
2015-07-29 08:12:05 +00:00
IN_IFADDR_RLOCK(&in_ifa_tracker);
ifnet: Replace if_addr_lock rwlock with epoch + mutex Run on LLNW canaries and tested by pho@ gallatin: Using a 14-core, 28-HTT single socket E5-2697 v3 with a 40GbE MLX5 based ConnectX 4-LX NIC, I see an almost 12% improvement in received packet rate, and a larger improvement in bytes delivered all the way to userspace. When the host receiving 64 streams of netperf -H $DUT -t UDP_STREAM -- -m 1, I see, using nstat -I mce0 1 before the patch: InMpps OMpps InGbs OGbs err TCP Est %CPU syscalls csw irq GBfree 4.98 0.00 4.42 0.00 4235592 33 83.80 4720653 2149771 1235 247.32 4.73 0.00 4.20 0.00 4025260 33 82.99 4724900 2139833 1204 247.32 4.72 0.00 4.20 0.00 4035252 33 82.14 4719162 2132023 1264 247.32 4.71 0.00 4.21 0.00 4073206 33 83.68 4744973 2123317 1347 247.32 4.72 0.00 4.21 0.00 4061118 33 80.82 4713615 2188091 1490 247.32 4.72 0.00 4.21 0.00 4051675 33 85.29 4727399 2109011 1205 247.32 4.73 0.00 4.21 0.00 4039056 33 84.65 4724735 2102603 1053 247.32 After the patch InMpps OMpps InGbs OGbs err TCP Est %CPU syscalls csw irq GBfree 5.43 0.00 4.20 0.00 3313143 33 84.96 5434214 1900162 2656 245.51 5.43 0.00 4.20 0.00 3308527 33 85.24 5439695 1809382 2521 245.51 5.42 0.00 4.19 0.00 3316778 33 87.54 5416028 1805835 2256 245.51 5.42 0.00 4.19 0.00 3317673 33 90.44 5426044 1763056 2332 245.51 5.42 0.00 4.19 0.00 3314839 33 88.11 5435732 1792218 2499 245.52 5.44 0.00 4.19 0.00 3293228 33 91.84 5426301 1668597 2121 245.52 Similarly, netperf reports 230Mb/s before the patch, and 270Mb/s after the patch Reviewed by: gallatin Sponsored by: Limelight Networks Differential Revision: https://reviews.freebsd.org/D15366
2018-05-18 20:13:34 +00:00
if (!bridged || (ia = CK_STAILQ_FIRST(&V_in_ifaddrhead)) == NULL) {
Convert in_ifaddr_lock and in6_ifaddr_lock to rmlock. Both are used to protect access to IP addresses lists and they can be acquired for reading several times per packet. To reduce lock contention it is better to use rmlock here. Reviewed by: gnn (previous version) Obtained from: Yandex LLC Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D3149
2015-07-29 08:12:05 +00:00
IN_IFADDR_RUNLOCK(&in_ifa_tracker);
Another small set of changes to reduce diffs with the new arp code.
2004-04-25 15:00:17 +00:00
goto drop;
Add a new global rwlock, in_ifaddr_lock, which will synchronize use of the in_ifaddrhead and INADDR_HASH address lists. Previously, these lists were used unsynchronized as they were effectively never changed in steady state, but we've seen increasing reports of writer-writer races on very busy VPN servers as core count has gone up (and similar configurations where address lists change frequently and concurrently). For the time being, use rwlocks rather than rmlocks in order to take advantage of their better lock debugging support. As a result, we don't enable ip_input()'s read-locking of INADDR_HASH until an rmlock conversion is complete and a performance analysis has been done. This means that one class of reader-writer races still exists. MFC after: 6 weeks Reviewed by: bz
2009-06-25 11:52:33 +00:00
}
In ARP input, more consistently acquire and release ifaddr references. MFC after: 6 weeks
2009-06-24 10:33:35 +00:00
ifa_ref(&ia->ia_ifa);
Convert in_ifaddr_lock and in6_ifaddr_lock to rmlock. Both are used to protect access to IP addresses lists and they can be acquired for reading several times per packet. To reduce lock contention it is better to use rmlock here. Reviewed by: gnn (previous version) Obtained from: Yandex LLC Sponsored by: Yandex LLC Differential Revision: https://reviews.freebsd.org/D3149
2015-07-29 08:12:05 +00:00
IN_IFADDR_RUNLOCK(&in_ifa_tracker);
Add a hash table that contains the list of internet addresses, and use this in place of the in_ifaddr list when appropriate. This improves performance on hosts which have a large number of IP aliases.
2001-09-29 04:34:11 +00:00
match:
Add CARP (Common Address Redundancy Protocol), which allows multiple hosts to share an IP address, providing high availability and load balancing. Original work on CARP done by Michael Shalayeff, with many additions by Marco Pfatschbacher and Ryan McBride. FreeBSD port done solely by Max Laier. Patch by: mlaier Obtained from: OpenBSD (mickey, mcbride)
2005-02-22 13:04:05 +00:00
if (!enaddr)
enaddr = (u_int8_t *)IF_LLADDR(ifp);
A major overhaul of the CARP implementation. The ip_carp.c was started from scratch, copying needed functionality from the old implemenation on demand, with a thorough review of all code. The main change is that interface layer has been removed from the CARP. Now redundant addresses are configured exactly on the interfaces, they run on. The CARP configuration itself is, as before, configured and read via SIOCSVH/SIOCGVH ioctls. A new prefix created with SIOCAIFADDR or SIOCAIFADDR_IN6 may now be configured to a particular virtual host id, which makes the prefix redundant. ifconfig(8) semantics has been changed too: now one doesn't need to clone carpXX interface, he/she should directly configure a vhid on a Ethernet interface. To supply vhid data from the kernel to an application the getifaddrs(8) function had been changed to pass ifam_data with each address. [1] The new implementation definitely closes all PRs related to carp(4) being an interface, and may close several others. It also allows to run a single redundant IP per interface. Big thanks to Bjoern Zeeb for his help with inet6 part of patch, for idea on using ifam_data and for several rounds of reviewing! PR: kern/117000, kern/126945, kern/126714, kern/120130, kern/117448 Reviewed by: bz Submitted by: bz [1]
2011-12-16 12:16:56 +00:00
carped = (ia->ia_ifa.ifa_carp != NULL);
Add a hash table that contains the list of internet addresses, and use this in place of the in_ifaddr list when appropriate. This improves performance on hosts which have a large number of IP aliases.
2001-09-29 04:34:11 +00:00
myaddr = ia->ia_addr.sin_addr;
In ARP input, more consistently acquire and release ifaddr references. MFC after: 6 weeks
2009-06-24 10:33:35 +00:00
ifa_free(&ia->ia_ifa);
Add CARP (Common Address Redundancy Protocol), which allows multiple hosts to share an IP address, providing high availability and load balancing. Original work on CARP done by Michael Shalayeff, with many additions by Marco Pfatschbacher and Ryan McBride. FreeBSD port done solely by Max Laier. Patch by: mlaier Obtained from: OpenBSD (mickey, mcbride)
2005-02-22 13:04:05 +00:00
if (!bcmp(ar_sha(ah), enaddr, ifp->if_addrlen))
Another small set of changes to reduce diffs with the new arp code.
2004-04-25 15:00:17 +00:00
goto drop; /* it's from me, ignore it. */
bring in ARP support for variable length link level addresses Reviewed by: jdp Approved by: jdp Obtained from: NetBSD MFC after: 6 weeks
2001-10-14 20:17:53 +00:00
if (!bcmp(ar_sha(ah), ifp->if_broadcastaddr, ifp->if_addrlen)) {
Rate limit the number of remotely triggered ARP log messages to 1 log message per second.
2013-05-11 10:51:32 +00:00
ARP_LOG(LOG_NOTICE, "link address is broadcast for IP address "
Use inet_ntoa_r() instead of inet_ntoa() throughout the kernel inet_ntoa() cannot be used safely in a multithreaded environment because it uses a static local buffer. Instead, use inet_ntoa_r() with a buffer on the caller's stack. Suggested by: glebius, emaste Reviewed by: gnn MFC after: 2 weeks Sponsored by: Dell EMC Differential Revision: https://reviews.freebsd.org/D9625
2017-02-16 20:47:41 +00:00
"%s!\n", inet_ntoa_r(isaddr, addrbuf));
Another small set of changes to reduce diffs with the new arp code.
2004-04-25 15:00:17 +00:00
goto drop;
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
Move lle update code from from gigantic ip_arpinput() to separate bunch of functions. The goal is to isolate actual lle updates to permit more fine-grained locking. Do all lle link-level update under AFDATA wlock. Sponsored by: Yandex LLC
2015-08-13 13:38:09 +00:00
if (ifp->if_addrlen != ah->ar_hln) {
ARP_LOG(LOG_WARNING, "from %*D: addr len: new %d, "
"i/f %d (ignored)\n", ifp->if_addrlen,
(u_char *) ar_sha(ah), ":", ah->ar_hln,
ifp->if_addrlen);
goto drop;
}
Modify the thrilling "%D is using my IP address %s!" message so that it isn't printed if the IP address in question is '0.0.0.0', which is used by nodes performing DHCP lookup, and so constitute a false positive as a report of misconfiguration.
2004-10-12 17:10:40 +00:00
/*
* Warn if another host is using the same IP address, but only if the
* IP address isn't 0.0.0.0, which is used for DHCP only, in which
* case we suppress the warning to avoid false positive complaints of
* potential misconfiguration.
*/
A major overhaul of the CARP implementation. The ip_carp.c was started from scratch, copying needed functionality from the old implemenation on demand, with a thorough review of all code. The main change is that interface layer has been removed from the CARP. Now redundant addresses are configured exactly on the interfaces, they run on. The CARP configuration itself is, as before, configured and read via SIOCSVH/SIOCGVH ioctls. A new prefix created with SIOCAIFADDR or SIOCAIFADDR_IN6 may now be configured to a particular virtual host id, which makes the prefix redundant. ifconfig(8) semantics has been changed too: now one doesn't need to clone carpXX interface, he/she should directly configure a vhid on a Ethernet interface. To supply vhid data from the kernel to an application the getifaddrs(8) function had been changed to pass ifam_data with each address. [1] The new implementation definitely closes all PRs related to carp(4) being an interface, and may close several others. It also allows to run a single redundant IP per interface. Big thanks to Bjoern Zeeb for his help with inet6 part of patch, for idea on using ifam_data and for several rounds of reviewing! PR: kern/117000, kern/126945, kern/126714, kern/120130, kern/117448 Reviewed by: bz Submitted by: bz [1]
2011-12-16 12:16:56 +00:00
if (!bridged && !carped && isaddr.s_addr == myaddr.s_addr &&
myaddr.s_addr != 0) {
Rate limit the number of remotely triggered ARP log messages to 1 log message per second.
2013-05-11 10:51:32 +00:00
ARP_LOG(LOG_ERR, "%*D is using my IP address %s on %s!\n",
bring in ARP support for variable length link level addresses Reviewed by: jdp Approved by: jdp Obtained from: NetBSD MFC after: 6 weeks
2001-10-14 20:17:53 +00:00
ifp->if_addrlen, (u_char *)ar_sha(ah), ":",
Use inet_ntoa_r() instead of inet_ntoa() throughout the kernel inet_ntoa() cannot be used safely in a multithreaded environment because it uses a static local buffer. Instead, use inet_ntoa_r() with a buffer on the caller's stack. Suggested by: glebius, emaste Reviewed by: gnn MFC after: 2 weeks Sponsored by: Dell EMC Differential Revision: https://reviews.freebsd.org/D9625
2017-02-16 20:47:41 +00:00
inet_ntoa_r(isaddr, addrbuf), ifp->if_xname);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
itaddr = myaddr;
Add ARP statistics to the kernel and netstat. New counters now exist for: requests sent replies sent requests received replies received packets received total packets dropped due to no ARP entry entrys timed out Duplicate IPs seen The new statistics are seen in the netstat command when it is given the -s command line switch. MFC after: 2 weeks In collaboration with: bz
2009-09-03 21:10:57 +00:00
ARPSTAT_INC(dupips);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
goto reply;
}
By popular demand, added the "static ARP" per-interface option.
2003-10-01 08:32:37 +00:00
if (ifp->if_flags & IFF_STATICARP)
goto reply;
o Fix a race between three threads: output path, incoming ARP packet and route request adding/removing ARP entries. The root of the problem is that struct llinfo_arp was accessed without any locks. To close race we will use locking provided by rtentry, that references this llinfo_arp: - Make arplookup() return a locked rtentry. - In arpresolve() hold the lock provided by rt_check()/arplookup() until the end of function, covering all accesses to the rtentry itself and llinfo_arp it refers to. - In in_arpinput() do not drop lock provided by arplookup() during first part of the function. - Simplify logic in the first part of in_arpinput(), removing one level of indentation. - In the second part of in_arpinput() hold rtentry lock while copying address. o Fix a condition when route entry is destroyed, while another thread is contested on its lock: - When storing a pointer to rtentry in llinfo_arp list, always add a reference to this rtentry, to prevent rtentry being destroyed via RTM_DELETE request. - Remove this reference when removing entry from llinfo_arp list. o Further cleanup of arptimer(): - Inline arptfree() into arptimer(). - Use official queue(3) way to pass LIST. - Hold rtentry lock while reading its structure. - Do not check that sdl_family is AF_LINK, but assert this. Reviewed by: sam Stress test: http://www.holm.cc/stress/log/cons141.html Stress test: http://people.freebsd.org/~pho/stress/log/cons144.html
2005-08-11 08:25:48 +00:00
This main goals of this project are: 1. separating L2 tables (ARP, NDP) from the L3 routing tables 2. removing as much locking dependencies among these layers as possible to allow for some parallelism in the search operations 3. simplify the logic in the routing code, The most notable end result is the obsolescent of the route cloning (RTF_CLONING) concept, which translated into code reduction in both IPv4 ARP and IPv6 NDP related modules, and size reduction in struct rtentry{}. The change in design obsoletes the semantics of RTF_CLONING, RTF_WASCLONE and RTF_LLINFO routing flags. The userland applications such as "arp" and "ndp" have been modified to reflect those changes. The output from "netstat -r" shows only the routing entries. Quite a few developers have contributed to this project in the past: Glebius Smirnoff, Luigi Rizzo, Alessandro Cerri, and Andre Oppermann. And most recently: - Kip Macy revised the locking code completely, thus completing the last piece of the puzzle, Kip has also been conducting active functional testing - Sam Leffler has helped me improving/refactoring the code, and provided valuable reviews - Julian Elischer setup the perforce tree for me and has helped me maintaining that branch before the svn conversion
2008-12-15 06:10:57 +00:00
bzero(&sin, sizeof(sin));
sin.sin_len = sizeof(struct sockaddr_in);
sin.sin_family = AF_INET;
sin.sin_addr = isaddr;
* Split allocation and table linking for lle's. Before that, the logic besides lle_create() was the following: return existing if found, create if not. This behaviour was error-prone since we had to deal with 'sudden' static<>dynamic lle changes. This commit fixes bunch of different issues like: - refcount leak when lle is converted to static. Simple check case: console 1: while true; do for i in `arp -an|awk '$4~/incomp/{print$2}'|tr -d '()'`; do arp -s $i 00:22:44:66:88:00 ; arp -d $i; done; done console 2: ping -f any-dead-host-in-L2 console 3: # watch for memory consumption: vmstat -m | awk '$1~/lltable/{print$2}' - possible problems in arptimer() / nd6_timer() when dropping/reacquiring lock. New logic explicitly handles use-or-create cases in every lla_create user. Basically, most of the changes are purely mechanical. However, we explicitly avoid using existing lle's for interface/static LLE records. * While here, call lle_event handlers on all real table lle change. * Create lltable_free_entry() calling existing per-lltable lle_free_t callback for entry deletion
2015-08-20 12:05:17 +00:00
dst = (struct sockaddr *)&sin;
la = lla_lookup(LLTABLE(ifp), LLE_EXCLUSIVE, dst);
Move lle update code from from gigantic ip_arpinput() to separate bunch of functions. The goal is to isolate actual lle updates to permit more fine-grained locking. Do all lle link-level update under AFDATA wlock. Sponsored by: Yandex LLC
2015-08-13 13:38:09 +00:00
if (la != NULL)
arp_check_update_lle(ah, isaddr, ifp, bridged, la);
else if (itaddr.s_addr == myaddr.s_addr) {
Some style(9) and whitespace changes. Together with: Andrey Zonov <andrey zonov.org>
2012-07-31 11:31:12 +00:00
/*
Fix ARP reply handling changed in r286955. If source of ARP request didn't pass the routing check (e.g. not in directly connected network), be polite and still answer the request instead of dropping frame. Reported by: quadro at irc@rusnet
2015-12-16 09:16:06 +00:00
* Request/reply to our address, but no lle exists yet.
Implement interface link header precomputation API. Add if_requestencap() interface method which is capable of calculating various link headers for given interface. Right now there is support for INET/INET6/ARP llheader calculation (IFENCAP_LL type request). Other types are planned to support more complex calculation (L2 multipath lagg nexthops, tunnel encap nexthops, etc..). Reshape 'struct route' to be able to pass additional data (with is length) to prepend to mbuf. These two changes permits routing code to pass pre-calculated nexthop data (like L2 header for route w/gateway) down to the stack eliminating the need for other lookups. It also brings us closer to more complex scenarios like transparently handling MPLS nexthops and tunnel interfaces. Last, but not least, it removes layering violation introduced by flowtable code (ro_lle) and simplifies handling of existing if_output consumers. ARP/ND changes: Make arp/ndp stack pre-calculate link header upon installing/updating lle record. Interface link address change are handled by re-calculating headers for all lles based on if_lladdr event. After these changes, arpresolve()/nd6_resolve() returns full pre-calculated header for supported interfaces thus simplifying if_output(). Move these lookups to separate ether_resolve_addr() function which ether returs error or fully-prepared link header. Add <arp|nd6_>resolve_addr() compat versions to return link addresses instead of pre-calculated data. BPF changes: Raw bpf writes occupied _two_ cases: AF_UNSPEC and pseudo_AF_HDRCMPLT. Despite the naming, both of there have ther header "complete". The only difference is that interface source mac has to be filled by OS for AF_UNSPEC (controlled via BIOCGHDRCMPLT). This logic has to stay inside BPF and not pollute if_output() routines. Convert BPF to pass prepend data via new 'struct route' mechanism. Note that it does not change non-optimized if_output(): ro_prepend handling is purely optional. Side note: hackish pseudo_AF_HDRCMPLT is supported for ethernet and FDDI. It is not needed for ethernet anymore. The only remaining FDDI user is dev/pdq mostly untouched since 2007. FDDI support was eliminated from OpenBSD in 2013 (sys/net/if_fddisubr.c rev 1.65). Flowtable changes: Flowtable violates layering by saving (and not correctly managing) rtes/lles. Instead of passing lle pointer, pass pointer to pre-calculated header data from that lle. Differential Revision: https://reviews.freebsd.org/D4102
2015-12-31 05:03:27 +00:00
* Calculate full link prepend to use in lle.
Add a queue to hold packets while we await an ARP reply. When a fast machine first brings up some non TCP networking program it is quite possible that we will drop packets due to the fact that only one packet can be held per ARP entry. This leads to packets being missed when a program starts or restarts if the ARP data is not currently in the ARP cache. This code adds a new sysctl, net.link.ether.inet.maxhold, which defines a system wide maximum number of packets to be held in each ARP entry. Up to maxhold packets are queued until an ARP reply is received or the ARP times out. The default setting is the old value of 1 which has been part of the BSD networking code since time immemorial. Expose the time we hold an incomplete ARP entry by adding the sysctl net.link.ether.inet.wait, which defaults to 20 seconds, the value used when the new ARP code was added.. Reviewed by: bz, rpaulo MFC after: 3 weeks
2010-11-12 22:03:02 +00:00
*/
Implement interface link header precomputation API. Add if_requestencap() interface method which is capable of calculating various link headers for given interface. Right now there is support for INET/INET6/ARP llheader calculation (IFENCAP_LL type request). Other types are planned to support more complex calculation (L2 multipath lagg nexthops, tunnel encap nexthops, etc..). Reshape 'struct route' to be able to pass additional data (with is length) to prepend to mbuf. These two changes permits routing code to pass pre-calculated nexthop data (like L2 header for route w/gateway) down to the stack eliminating the need for other lookups. It also brings us closer to more complex scenarios like transparently handling MPLS nexthops and tunnel interfaces. Last, but not least, it removes layering violation introduced by flowtable code (ro_lle) and simplifies handling of existing if_output consumers. ARP/ND changes: Make arp/ndp stack pre-calculate link header upon installing/updating lle record. Interface link address change are handled by re-calculating headers for all lles based on if_lladdr event. After these changes, arpresolve()/nd6_resolve() returns full pre-calculated header for supported interfaces thus simplifying if_output(). Move these lookups to separate ether_resolve_addr() function which ether returs error or fully-prepared link header. Add <arp|nd6_>resolve_addr() compat versions to return link addresses instead of pre-calculated data. BPF changes: Raw bpf writes occupied _two_ cases: AF_UNSPEC and pseudo_AF_HDRCMPLT. Despite the naming, both of there have ther header "complete". The only difference is that interface source mac has to be filled by OS for AF_UNSPEC (controlled via BIOCGHDRCMPLT). This logic has to stay inside BPF and not pollute if_output() routines. Convert BPF to pass prepend data via new 'struct route' mechanism. Note that it does not change non-optimized if_output(): ro_prepend handling is purely optional. Side note: hackish pseudo_AF_HDRCMPLT is supported for ethernet and FDDI. It is not needed for ethernet anymore. The only remaining FDDI user is dev/pdq mostly untouched since 2007. FDDI support was eliminated from OpenBSD in 2013 (sys/net/if_fddisubr.c rev 1.65). Flowtable changes: Flowtable violates layering by saving (and not correctly managing) rtes/lles. Instead of passing lle pointer, pass pointer to pre-calculated header data from that lle. Differential Revision: https://reviews.freebsd.org/D4102
2015-12-31 05:03:27 +00:00
linkhdrsize = sizeof(linkhdr);
if (lltable_calc_llheader(ifp, AF_INET, ar_sha(ah), linkhdr,
&linkhdrsize, &lladdr_off) != 0)
goto reply;
/* Allocate new entry */
* Split allocation and table linking for lle's. Before that, the logic besides lle_create() was the following: return existing if found, create if not. This behaviour was error-prone since we had to deal with 'sudden' static<>dynamic lle changes. This commit fixes bunch of different issues like: - refcount leak when lle is converted to static. Simple check case: console 1: while true; do for i in `arp -an|awk '$4~/incomp/{print$2}'|tr -d '()'`; do arp -s $i 00:22:44:66:88:00 ; arp -d $i; done; done console 2: ping -f any-dead-host-in-L2 console 3: # watch for memory consumption: vmstat -m | awk '$1~/lltable/{print$2}' - possible problems in arptimer() / nd6_timer() when dropping/reacquiring lock. New logic explicitly handles use-or-create cases in every lla_create user. Basically, most of the changes are purely mechanical. However, we explicitly avoid using existing lle's for interface/static LLE records. * While here, call lle_event handlers on all real table lle change. * Create lltable_free_entry() calling existing per-lltable lle_free_t callback for entry deletion
2015-08-20 12:05:17 +00:00
la = lltable_alloc_entry(LLTABLE(ifp), 0, dst);
Fix ARP reply handling changed in r286955. If source of ARP request didn't pass the routing check (e.g. not in directly connected network), be polite and still answer the request instead of dropping frame. Reported by: quadro at irc@rusnet
2015-12-16 09:16:06 +00:00
if (la == NULL) {
/*
* lle creation may fail if source address belongs
* to non-directly connected subnet. However, we
* will try to answer the request instead of dropping
* frame.
*/
goto reply;
}
Implement interface link header precomputation API. Add if_requestencap() interface method which is capable of calculating various link headers for given interface. Right now there is support for INET/INET6/ARP llheader calculation (IFENCAP_LL type request). Other types are planned to support more complex calculation (L2 multipath lagg nexthops, tunnel encap nexthops, etc..). Reshape 'struct route' to be able to pass additional data (with is length) to prepend to mbuf. These two changes permits routing code to pass pre-calculated nexthop data (like L2 header for route w/gateway) down to the stack eliminating the need for other lookups. It also brings us closer to more complex scenarios like transparently handling MPLS nexthops and tunnel interfaces. Last, but not least, it removes layering violation introduced by flowtable code (ro_lle) and simplifies handling of existing if_output consumers. ARP/ND changes: Make arp/ndp stack pre-calculate link header upon installing/updating lle record. Interface link address change are handled by re-calculating headers for all lles based on if_lladdr event. After these changes, arpresolve()/nd6_resolve() returns full pre-calculated header for supported interfaces thus simplifying if_output(). Move these lookups to separate ether_resolve_addr() function which ether returs error or fully-prepared link header. Add <arp|nd6_>resolve_addr() compat versions to return link addresses instead of pre-calculated data. BPF changes: Raw bpf writes occupied _two_ cases: AF_UNSPEC and pseudo_AF_HDRCMPLT. Despite the naming, both of there have ther header "complete". The only difference is that interface source mac has to be filled by OS for AF_UNSPEC (controlled via BIOCGHDRCMPLT). This logic has to stay inside BPF and not pollute if_output() routines. Convert BPF to pass prepend data via new 'struct route' mechanism. Note that it does not change non-optimized if_output(): ro_prepend handling is purely optional. Side note: hackish pseudo_AF_HDRCMPLT is supported for ethernet and FDDI. It is not needed for ethernet anymore. The only remaining FDDI user is dev/pdq mostly untouched since 2007. FDDI support was eliminated from OpenBSD in 2013 (sys/net/if_fddisubr.c rev 1.65). Flowtable changes: Flowtable violates layering by saving (and not correctly managing) rtes/lles. Instead of passing lle pointer, pass pointer to pre-calculated header data from that lle. Differential Revision: https://reviews.freebsd.org/D4102
2015-12-31 05:03:27 +00:00
lltable_set_entry_addr(ifp, la, linkhdr, linkhdrsize,
lladdr_off);
* Split allocation and table linking for lle's. Before that, the logic besides lle_create() was the following: return existing if found, create if not. This behaviour was error-prone since we had to deal with 'sudden' static<>dynamic lle changes. This commit fixes bunch of different issues like: - refcount leak when lle is converted to static. Simple check case: console 1: while true; do for i in `arp -an|awk '$4~/incomp/{print$2}'|tr -d '()'`; do arp -s $i 00:22:44:66:88:00 ; arp -d $i; done; done console 2: ping -f any-dead-host-in-L2 console 3: # watch for memory consumption: vmstat -m | awk '$1~/lltable/{print$2}' - possible problems in arptimer() / nd6_timer() when dropping/reacquiring lock. New logic explicitly handles use-or-create cases in every lla_create user. Basically, most of the changes are purely mechanical. However, we explicitly avoid using existing lle's for interface/static LLE records. * While here, call lle_event handlers on all real table lle change. * Create lltable_free_entry() calling existing per-lltable lle_free_t callback for entry deletion
2015-08-20 12:05:17 +00:00
Move lle update code from from gigantic ip_arpinput() to separate bunch of functions. The goal is to isolate actual lle updates to permit more fine-grained locking. Do all lle link-level update under AFDATA wlock. Sponsored by: Yandex LLC
2015-08-13 13:38:09 +00:00
IF_AFDATA_WLOCK(ifp);
* Split allocation and table linking for lle's. Before that, the logic besides lle_create() was the following: return existing if found, create if not. This behaviour was error-prone since we had to deal with 'sudden' static<>dynamic lle changes. This commit fixes bunch of different issues like: - refcount leak when lle is converted to static. Simple check case: console 1: while true; do for i in `arp -an|awk '$4~/incomp/{print$2}'|tr -d '()'`; do arp -s $i 00:22:44:66:88:00 ; arp -d $i; done; done console 2: ping -f any-dead-host-in-L2 console 3: # watch for memory consumption: vmstat -m | awk '$1~/lltable/{print$2}' - possible problems in arptimer() / nd6_timer() when dropping/reacquiring lock. New logic explicitly handles use-or-create cases in every lla_create user. Basically, most of the changes are purely mechanical. However, we explicitly avoid using existing lle's for interface/static LLE records. * While here, call lle_event handlers on all real table lle change. * Create lltable_free_entry() calling existing per-lltable lle_free_t callback for entry deletion
2015-08-20 12:05:17 +00:00
LLE_WLOCK(la);
la_tmp = lla_lookup(LLTABLE(ifp), LLE_EXCLUSIVE, dst);
/*
* Check if lle still does not exists.
* If it does, that means that we either
* 1) have configured it explicitly, via
* 1a) 'arp -s' static entry or
* 1b) interface address static record
* or
* 2) it was the result of sending first packet to-host
* or
* 3) it was another arp reply packet we handled in
* different thread.
*
* In all cases except 3) we definitely need to prefer
* existing lle. For the sake of simplicity, prefer any
* existing lle over newly-create one.
*/
if (la_tmp == NULL)
lltable_link_entry(LLTABLE(ifp), la);
Move lle update code from from gigantic ip_arpinput() to separate bunch of functions. The goal is to isolate actual lle updates to permit more fine-grained locking. Do all lle link-level update under AFDATA wlock. Sponsored by: Yandex LLC
2015-08-13 13:38:09 +00:00
IF_AFDATA_WUNLOCK(ifp);
* Split allocation and table linking for lle's. Before that, the logic besides lle_create() was the following: return existing if found, create if not. This behaviour was error-prone since we had to deal with 'sudden' static<>dynamic lle changes. This commit fixes bunch of different issues like: - refcount leak when lle is converted to static. Simple check case: console 1: while true; do for i in `arp -an|awk '$4~/incomp/{print$2}'|tr -d '()'`; do arp -s $i 00:22:44:66:88:00 ; arp -d $i; done; done console 2: ping -f any-dead-host-in-L2 console 3: # watch for memory consumption: vmstat -m | awk '$1~/lltable/{print$2}' - possible problems in arptimer() / nd6_timer() when dropping/reacquiring lock. New logic explicitly handles use-or-create cases in every lla_create user. Basically, most of the changes are purely mechanical. However, we explicitly avoid using existing lle's for interface/static LLE records. * While here, call lle_event handlers on all real table lle change. * Create lltable_free_entry() calling existing per-lltable lle_free_t callback for entry deletion
2015-08-20 12:05:17 +00:00
if (la_tmp == NULL) {
Check value return from lle_create() for NULL. This bug sneaked unnoticed in r286722. Reported by: adrian
2015-08-19 21:08:42 +00:00
arp_mark_lle_reachable(la);
LLE_WUNLOCK(la);
* Split allocation and table linking for lle's. Before that, the logic besides lle_create() was the following: return existing if found, create if not. This behaviour was error-prone since we had to deal with 'sudden' static<>dynamic lle changes. This commit fixes bunch of different issues like: - refcount leak when lle is converted to static. Simple check case: console 1: while true; do for i in `arp -an|awk '$4~/incomp/{print$2}'|tr -d '()'`; do arp -s $i 00:22:44:66:88:00 ; arp -d $i; done; done console 2: ping -f any-dead-host-in-L2 console 3: # watch for memory consumption: vmstat -m | awk '$1~/lltable/{print$2}' - possible problems in arptimer() / nd6_timer() when dropping/reacquiring lock. New logic explicitly handles use-or-create cases in every lla_create user. Basically, most of the changes are purely mechanical. However, we explicitly avoid using existing lle's for interface/static LLE records. * While here, call lle_event handlers on all real table lle change. * Create lltable_free_entry() calling existing per-lltable lle_free_t callback for entry deletion
2015-08-20 12:05:17 +00:00
} else {
/* Free newly-create entry and handle packet */
lltable_free_entry(LLTABLE(ifp), la);
la = la_tmp;
la_tmp = NULL;
arp_check_update_lle(ah, isaddr, ifp, bridged, la);
/* arp_check_update_lle() returns @la unlocked */
Check value return from lle_create() for NULL. This bug sneaked unnoticed in r286722. Reported by: adrian
2015-08-19 21:08:42 +00:00
}
* Split allocation and table linking for lle's. Before that, the logic besides lle_create() was the following: return existing if found, create if not. This behaviour was error-prone since we had to deal with 'sudden' static<>dynamic lle changes. This commit fixes bunch of different issues like: - refcount leak when lle is converted to static. Simple check case: console 1: while true; do for i in `arp -an|awk '$4~/incomp/{print$2}'|tr -d '()'`; do arp -s $i 00:22:44:66:88:00 ; arp -d $i; done; done console 2: ping -f any-dead-host-in-L2 console 3: # watch for memory consumption: vmstat -m | awk '$1~/lltable/{print$2}' - possible problems in arptimer() / nd6_timer() when dropping/reacquiring lock. New logic explicitly handles use-or-create cases in every lla_create user. Basically, most of the changes are purely mechanical. However, we explicitly avoid using existing lle's for interface/static LLE records. * While here, call lle_event handlers on all real table lle change. * Create lltable_free_entry() calling existing per-lltable lle_free_t callback for entry deletion
2015-08-20 12:05:17 +00:00
la = NULL;
Remove a these days incorrect comment left from before new-arp. MFC after: 1 week
2011-06-18 13:54:36 +00:00
}
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
reply:
Another small set of changes to reduce diffs with the new arp code.
2004-04-25 15:00:17 +00:00
if (op != ARPOP_REQUEST)
goto drop;
Add ARP statistics to the kernel and netstat. New counters now exist for: requests sent replies sent requests received replies received packets received total packets dropped due to no ARP entry entrys timed out Duplicate IPs seen The new statistics are seen in the netstat command when it is given the -s command line switch. MFC after: 2 weeks In collaboration with: bz
2009-09-03 21:10:57 +00:00
ARPSTAT_INC(rxrequests);
This main goals of this project are: 1. separating L2 tables (ARP, NDP) from the L3 routing tables 2. removing as much locking dependencies among these layers as possible to allow for some parallelism in the search operations 3. simplify the logic in the routing code, The most notable end result is the obsolescent of the route cloning (RTF_CLONING) concept, which translated into code reduction in both IPv4 ARP and IPv6 NDP related modules, and size reduction in struct rtentry{}. The change in design obsoletes the semantics of RTF_CLONING, RTF_WASCLONE and RTF_LLINFO routing flags. The userland applications such as "arp" and "ndp" have been modified to reflect those changes. The output from "netstat -r" shows only the routing entries. Quite a few developers have contributed to this project in the past: Glebius Smirnoff, Luigi Rizzo, Alessandro Cerri, and Andre Oppermann. And most recently: - Kip Macy revised the locking code completely, thus completing the last piece of the puzzle, Kip has also been conducting active functional testing - Sam Leffler has helped me improving/refactoring the code, and provided valuable reviews - Julian Elischer setup the perforce tree for me and has helped me maintaining that branch before the svn conversion
2008-12-15 06:10:57 +00:00
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
if (itaddr.s_addr == myaddr.s_addr) {
Add code to allow the system to handle multiple routing tables. This particular implementation is designed to be fully backwards compatible and to be MFC-able to 7.x (and 6.x) Currently the only protocol that can make use of the multiple tables is IPv4 Similar functionality exists in OpenBSD and Linux. From my notes: ----- One thing where FreeBSD has been falling behind, and which by chance I have some time to work on is "policy based routing", which allows different packet streams to be routed by more than just the destination address. Constraints: ------------ I want to make some form of this available in the 6.x tree (and by extension 7.x) , but FreeBSD in general needs it so I might as well do it in -current and back port the portions I need. One of the ways that this can be done is to have the ability to instantiate multiple kernel routing tables (which I will now refer to as "Forwarding Information Bases" or "FIBs" for political correctness reasons). Which FIB a particular packet uses to make the next hop decision can be decided by a number of mechanisms. The policies these mechanisms implement are the "Policies" referred to in "Policy based routing". One of the constraints I have if I try to back port this work to 6.x is that it must be implemented as a EXTENSION to the existing ABIs in 6.x so that third party applications do not need to be recompiled in timespan of the branch. This first version will not have some of the bells and whistles that will come with later versions. It will, for example, be limited to 16 tables in the first commit. Implementation method, Compatible version. (part 1) ------------------------------- For this reason I have implemented a "sufficient subset" of a multiple routing table solution in Perforce, and back-ported it to 6.x. (also in Perforce though not always caught up with what I have done in -current/P4). The subset allows a number of FIBs to be defined at compile time (8 is sufficient for my purposes in 6.x) and implements the changes needed to allow IPV4 to use them. I have not done the changes for ipv6 simply because I do not need it, and I do not have enough knowledge of ipv6 (e.g. neighbor discovery) needed to do it. Other protocol families are left untouched and should there be users with proprietary protocol families, they should continue to work and be oblivious to the existence of the extra FIBs. To understand how this is done, one must know that the current FIB code starts everything off with a single dimensional array of pointers to FIB head structures (One per protocol family), each of which in turn points to the trie of routes available to that family. The basic change in the ABI compatible version of the change is to extent that array to be a 2 dimensional array, so that instead of protocol family X looking at rt_tables[X] for the table it needs, it looks at rt_tables[Y][X] when for all protocol families except ipv4 Y is always 0. Code that is unaware of the change always just sees the first row of the table, which of course looks just like the one dimensional array that existed before. The entry points rtrequest(), rtalloc(), rtalloc1(), rtalloc_ign() are all maintained, but refer only to the first row of the array, so that existing callers in proprietary protocols can continue to do the "right thing". Some new entry points are added, for the exclusive use of ipv4 code called in_rtrequest(), in_rtalloc(), in_rtalloc1() and in_rtalloc_ign(), which have an extra argument which refers the code to the correct row. In addition, there are some new entry points (currently called rtalloc_fib() and friends) that check the Address family being looked up and call either rtalloc() (and friends) if the protocol is not IPv4 forcing the action to row 0 or to the appropriate row if it IS IPv4 (and that info is available). These are for calling from code that is not specific to any particular protocol. The way these are implemented would change in the non ABI preserving code to be added later. One feature of the first version of the code is that for ipv4, the interface routes show up automatically on all the FIBs, so that no matter what FIB you select you always have the basic direct attached hosts available to you. (rtinit() does this automatically). You CAN delete an interface route from one FIB should you want to but by default it's there. ARP information is also available in each FIB. It's assumed that the same machine would have the same MAC address, regardless of which FIB you are using to get to it. This brings us as to how the correct FIB is selected for an outgoing IPV4 packet. Firstly, all packets have a FIB associated with them. if nothing has been done to change it, it will be FIB 0. The FIB is changed in the following ways. Packets fall into one of a number of classes. 1/ locally generated packets, coming from a socket/PCB. Such packets select a FIB from a number associated with the socket/PCB. This in turn is inherited from the process, but can be changed by a socket option. The process in turn inherits it on fork. I have written a utility call setfib that acts a bit like nice.. setfib -3 ping target.example.com # will use fib 3 for ping. It is an obvious extension to make it a property of a jail but I have not done so. It can be achieved by combining the setfib and jail commands. 2/ packets received on an interface for forwarding. By default these packets would use table 0, (or possibly a number settable in a sysctl(not yet)). but prior to routing the firewall can inspect them (see below). (possibly in the future you may be able to associate a FIB with packets received on an interface.. An ifconfig arg, but not yet.) 3/ packets inspected by a packet classifier, which can arbitrarily associate a fib with it on a packet by packet basis. A fib assigned to a packet by a packet classifier (such as ipfw) would over-ride a fib associated by a more default source. (such as cases 1 or 2). 4/ a tcp listen socket associated with a fib will generate accept sockets that are associated with that same fib. 5/ Packets generated in response to some other packet (e.g. reset or icmp packets). These should use the FIB associated with the packet being reponded to. 6/ Packets generated during encapsulation. gif, tun and other tunnel interfaces will encapsulate using the FIB that was in effect withthe proces that set up the tunnel. thus setfib 1 ifconfig gif0 [tunnel instructions] will set the fib for the tunnel to use to be fib 1. Routing messages would be associated with their process, and thus select one FIB or another. messages from the kernel would be associated with the fib they refer to and would only be received by a routing socket associated with that fib. (not yet implemented) In addition Netstat has been edited to be able to cope with the fact that the array is now 2 dimensional. (It looks in system memory using libkvm (!)). Old versions of netstat see only the first FIB. In addition two sysctls are added to give: a) the number of FIBs compiled in (active) b) the default FIB of the calling process. Early testing experience: ------------------------- Basically our (IronPort's) appliance does this functionality already using ipfw fwd but that method has some drawbacks. For example, It can't fully simulate a routing table because it can't influence the socket's choice of local address when a connect() is done. Testing during the generating of these changes has been remarkably smooth so far. Multiple tables have co-existed with no notable side effects, and packets have been routes accordingly. ipfw has grown 2 new keywords: setfib N ip from anay to any count ip from any to any fib N In pf there seems to be a requirement to be able to give symbolic names to the fibs but I do not have that capacity. I am not sure if it is required. SCTP has interestingly enough built in support for this, called VRFs in Cisco parlance. it will be interesting to see how that handles it when it suddenly actually does something. Where to next: -------------------- After committing the ABI compatible version and MFCing it, I'd like to proceed in a forward direction in -current. this will result in some roto-tilling in the routing code. Firstly: the current code's idea of having a separate tree per protocol family, all of the same format, and pointed to by the 1 dimensional array is a bit silly. Especially when one considers that there is code that makes assumptions about every protocol having the same internal structures there. Some protocols don't WANT that sort of structure. (for example the whole idea of a netmask is foreign to appletalk). This needs to be made opaque to the external code. My suggested first change is to add routing method pointers to the 'domain' structure, along with information pointing the data. instead of having an array of pointers to uniform structures, there would be an array pointing to the 'domain' structures for each protocol address domain (protocol family), and the methods this reached would be called. The methods would have an argument that gives FIB number, but the protocol would be free to ignore it. When the ABI can be changed it raises the possibilty of the addition of a fib entry into the "struct route". Currently, the structure contains the sockaddr of the desination, and the resulting fib entry. To make this work fully, one could add a fib number so that given an address and a fib, one can find the third element, the fib entry. Interaction with the ARP layer/ LL layer would need to be revisited as well. Qing Li has been working on this already. This work was sponsored by Ironport Systems/Cisco Reviewed by: several including rwatson, bz and mlair (parts each) Obtained from: Ironport systems/Cisco
2008-05-09 23:03:00 +00:00
/* Shortcut.. the receiving interface is the target. */
bring in ARP support for variable length link level addresses Reviewed by: jdp Approved by: jdp Obtained from: NetBSD MFC after: 6 weeks
2001-10-14 20:17:53 +00:00
(void)memcpy(ar_tha(ah), ar_sha(ah), ah->ar_hln);
Add CARP (Common Address Redundancy Protocol), which allows multiple hosts to share an IP address, providing high availability and load balancing. Original work on CARP done by Michael Shalayeff, with many additions by Marco Pfatschbacher and Ryan McBride. FreeBSD port done solely by Max Laier. Patch by: mlaier Obtained from: OpenBSD (mickey, mcbride)
2005-02-22 13:04:05 +00:00
(void)memcpy(ar_sha(ah), enaddr, ah->ar_hln);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
} else {
Use interface fib for proxyarp checks. Before the change, proxyarp checks for src and dst addresses were performed using default fib, breaking multi-fib scenario. PR: 245181 Submitted by: Scott Aitken (original version) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D24244
2020-04-02 20:06:37 +00:00
/*
* Destination address is not ours. Check if
* proxyarp entry exists or proxyarp is turned on globally.
*/
struct llentry *lle;
The proxy-arp code was broken and responds to ARP requests for addresses that are not proxied locally.
2008-12-19 11:07:34 +00:00
sin.sin_addr = itaddr;
This patch enables the node to respond to ARP requests for configured proxy ARP entries. Reviewed by: bz MFC after: immediately
2009-09-15 18:39:27 +00:00
lle = lla_lookup(LLTABLE(ifp), 0, (struct sockaddr *)&sin);
The proxy-arp code was broken and responds to ARP requests for addresses that are not proxied locally.
2008-12-19 11:07:34 +00:00
This patch enables the node to respond to ARP requests for configured proxy ARP entries. Reviewed by: bz MFC after: immediately
2009-09-15 18:39:27 +00:00
if ((lle != NULL) && (lle->la_flags & LLE_PUB)) {
The proxy-arp code was broken and responds to ARP requests for addresses that are not proxied locally.
2008-12-19 11:07:34 +00:00
(void)memcpy(ar_tha(ah), ar_sha(ah), ah->ar_hln);
Implement interface link header precomputation API. Add if_requestencap() interface method which is capable of calculating various link headers for given interface. Right now there is support for INET/INET6/ARP llheader calculation (IFENCAP_LL type request). Other types are planned to support more complex calculation (L2 multipath lagg nexthops, tunnel encap nexthops, etc..). Reshape 'struct route' to be able to pass additional data (with is length) to prepend to mbuf. These two changes permits routing code to pass pre-calculated nexthop data (like L2 header for route w/gateway) down to the stack eliminating the need for other lookups. It also brings us closer to more complex scenarios like transparently handling MPLS nexthops and tunnel interfaces. Last, but not least, it removes layering violation introduced by flowtable code (ro_lle) and simplifies handling of existing if_output consumers. ARP/ND changes: Make arp/ndp stack pre-calculate link header upon installing/updating lle record. Interface link address change are handled by re-calculating headers for all lles based on if_lladdr event. After these changes, arpresolve()/nd6_resolve() returns full pre-calculated header for supported interfaces thus simplifying if_output(). Move these lookups to separate ether_resolve_addr() function which ether returs error or fully-prepared link header. Add <arp|nd6_>resolve_addr() compat versions to return link addresses instead of pre-calculated data. BPF changes: Raw bpf writes occupied _two_ cases: AF_UNSPEC and pseudo_AF_HDRCMPLT. Despite the naming, both of there have ther header "complete". The only difference is that interface source mac has to be filled by OS for AF_UNSPEC (controlled via BIOCGHDRCMPLT). This logic has to stay inside BPF and not pollute if_output() routines. Convert BPF to pass prepend data via new 'struct route' mechanism. Note that it does not change non-optimized if_output(): ro_prepend handling is purely optional. Side note: hackish pseudo_AF_HDRCMPLT is supported for ethernet and FDDI. It is not needed for ethernet anymore. The only remaining FDDI user is dev/pdq mostly untouched since 2007. FDDI support was eliminated from OpenBSD in 2013 (sys/net/if_fddisubr.c rev 1.65). Flowtable changes: Flowtable violates layering by saving (and not correctly managing) rtes/lles. Instead of passing lle pointer, pass pointer to pre-calculated header data from that lle. Differential Revision: https://reviews.freebsd.org/D4102
2015-12-31 05:03:27 +00:00
(void)memcpy(ar_sha(ah), lle->ll_addr, ah->ar_hln);
The proxy-arp code was broken and responds to ARP requests for addresses that are not proxied locally.
2008-12-19 11:07:34 +00:00
LLE_RUNLOCK(lle);
This patch enables the node to respond to ARP requests for configured proxy ARP entries. Reviewed by: bz MFC after: immediately
2009-09-15 18:39:27 +00:00
} else {
if (lle != NULL)
LLE_RUNLOCK(lle);
if (!V_arp_proxyall)
goto drop;
Some more whitespace cleanup.
2012-08-01 09:00:26 +00:00
Use interface fib for proxyarp checks. Before the change, proxyarp checks for src and dst addresses were performed using default fib, breaking multi-fib scenario. PR: 245181 Submitted by: Scott Aitken (original version) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D24244
2020-04-02 20:06:37 +00:00
if (fib4_lookup_nh_basic(ifp->if_fib, itaddr, 0, 0,
&nh4) != 0)
This patch enables the node to respond to ARP requests for configured proxy ARP entries. Reviewed by: bz MFC after: immediately
2009-09-15 18:39:27 +00:00
goto drop;
/*
* Don't send proxies for nodes on the same interface
* as this one came out of, or we'll get into a fight
* over who claims what Ether address.
*/
Make in_arpinput(), inp_lookup_mcast_ifp(), icmp_reflect(), ip_dooptions(), icmp6_redirect_input(), in6_lltable_rtcheck(), in6p_lookup_mcast_ifp() and in6_selecthlim() use new routing api. Eliminate now-unused ip_rtaddr(). Fix lookup key fib6_lookup_nh_basic() which was lost diring merge. Make fib6_lookup_nh_basic() and fib6_lookup_nh_extended() always return IPv6 destination address with embedded scope. Currently rw_gateway has it scope embedded, do the same for non-gatewayed destinations. Sponsored by: Yandex LLC
2015-12-09 11:14:27 +00:00
if (nh4.nh_ifp == ifp)
This patch enables the node to respond to ARP requests for configured proxy ARP entries. Reviewed by: bz MFC after: immediately
2009-09-15 18:39:27 +00:00
goto drop;
The proxy-arp code was broken and responds to ARP requests for addresses that are not proxied locally.
2008-12-19 11:07:34 +00:00
This patch enables the node to respond to ARP requests for configured proxy ARP entries. Reviewed by: bz MFC after: immediately
2009-09-15 18:39:27 +00:00
(void)memcpy(ar_tha(ah), ar_sha(ah), ah->ar_hln);
(void)memcpy(ar_sha(ah), enaddr, ah->ar_hln);
/*
* Also check that the node which sent the ARP packet
Fix typos - remove duplicate "the". PR: bin/154928 Submitted by: Eitan Adler <lists at eitanadler.com> MFC after: 3 days
2011-02-21 09:01:34 +00:00
* is on the interface we expect it to be on. This
This patch enables the node to respond to ARP requests for configured proxy ARP entries. Reviewed by: bz MFC after: immediately
2009-09-15 18:39:27 +00:00
* avoids ARP chaos if an interface is connected to the
* wrong network.
*/
Some more whitespace cleanup.
2012-08-01 09:00:26 +00:00
Use interface fib for proxyarp checks. Before the change, proxyarp checks for src and dst addresses were performed using default fib, breaking multi-fib scenario. PR: 245181 Submitted by: Scott Aitken (original version) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D24244
2020-04-02 20:06:37 +00:00
if (fib4_lookup_nh_basic(ifp->if_fib, isaddr, 0, 0,
&nh4) != 0)
This patch enables the node to respond to ARP requests for configured proxy ARP entries. Reviewed by: bz MFC after: immediately
2009-09-15 18:39:27 +00:00
goto drop;
Make in_arpinput(), inp_lookup_mcast_ifp(), icmp_reflect(), ip_dooptions(), icmp6_redirect_input(), in6_lltable_rtcheck(), in6p_lookup_mcast_ifp() and in6_selecthlim() use new routing api. Eliminate now-unused ip_rtaddr(). Fix lookup key fib6_lookup_nh_basic() which was lost diring merge. Make fib6_lookup_nh_basic() and fib6_lookup_nh_extended() always return IPv6 destination address with embedded scope. Currently rw_gateway has it scope embedded, do the same for non-gatewayed destinations. Sponsored by: Yandex LLC
2015-12-09 11:14:27 +00:00
if (nh4.nh_ifp != ifp) {
Rate limit the number of remotely triggered ARP log messages to 1 log message per second.
2013-05-11 10:51:32 +00:00
ARP_LOG(LOG_INFO, "proxy: ignoring request"
Make in_arpinput(), inp_lookup_mcast_ifp(), icmp_reflect(), ip_dooptions(), icmp6_redirect_input(), in6_lltable_rtcheck(), in6p_lookup_mcast_ifp() and in6_selecthlim() use new routing api. Eliminate now-unused ip_rtaddr(). Fix lookup key fib6_lookup_nh_basic() which was lost diring merge. Make fib6_lookup_nh_basic() and fib6_lookup_nh_extended() always return IPv6 destination address with embedded scope. Currently rw_gateway has it scope embedded, do the same for non-gatewayed destinations. Sponsored by: Yandex LLC
2015-12-09 11:14:27 +00:00
" from %s via %s\n",
Use inet_ntoa_r() instead of inet_ntoa() throughout the kernel inet_ntoa() cannot be used safely in a multithreaded environment because it uses a static local buffer. Instead, use inet_ntoa_r() with a buffer on the caller's stack. Suggested by: glebius, emaste Reviewed by: gnn MFC after: 2 weeks Sponsored by: Dell EMC Differential Revision: https://reviews.freebsd.org/D9625
2017-02-16 20:47:41 +00:00
inet_ntoa_r(isaddr, addrbuf),
ifp->if_xname);
This patch enables the node to respond to ARP requests for configured proxy ARP entries. Reviewed by: bz MFC after: immediately
2009-09-15 18:39:27 +00:00
goto drop;
}
Extra sanity check when arp proxyall is enabled. Don't send an arp reply if the requesting machine isn't on the interface we believe it should be. Prevents arp wars when you plug cables in the wrong way around. PR: 9848 Submitted by: Ian Dowse <iedowse@maths.tcd.ie> Not objected to by: wollman
2000-07-13 19:31:01 +00:00
Clean up ARP error messages: format IP addresses, explain arplookup() failures in English.
1994-11-02 00:58:29 +00:00
#ifdef DEBUG_PROXY
Use inet_ntoa_r() instead of inet_ntoa() throughout the kernel inet_ntoa() cannot be used safely in a multithreaded environment because it uses a static local buffer. Instead, use inet_ntoa_r() with a buffer on the caller's stack. Suggested by: glebius, emaste Reviewed by: gnn MFC after: 2 weeks Sponsored by: Dell EMC Differential Revision: https://reviews.freebsd.org/D9625
2017-02-16 20:47:41 +00:00
printf("arp: proxying for %s\n",
inet_ntoa_r(itaddr, addrbuf));
Implement full proxy ARP, gated on option ARP_PROXYALL. This allows a FreeBSD box to do proxy ARP as easily as most commercial routers do, without messing around with (potentially variable) Ethernet addresses. This code is really quite simple; I'm not at all sure why it wasn't implemented in 4.4. It might be worth stealing an interface flag (maybe IFF_LINK1) to use for finer-grained control over which interfaces get proxy treatment. For the moment, it's all or nothing.
1994-10-01 21:50:33 +00:00
#endif
This patch enables the node to respond to ARP requests for configured proxy ARP entries. Reviewed by: bz MFC after: immediately
2009-09-15 18:39:27 +00:00
}
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
Comply with RFC 3927, by forcing ARP replies which contain a source address within the link-local IPv4 prefix 169.254.0.0/16, to be broadcast at link layer. Reviewed by: fenner MFC after: 2 weeks
2007-02-02 20:31:44 +00:00
if (itaddr.s_addr == myaddr.s_addr &&
IN_LINKLOCAL(ntohl(itaddr.s_addr))) {
/* RFC 3927 link-local IPv4; always reply by broadcast. */
#ifdef DEBUG_LINKLOCAL
printf("arp: sending reply for link-local addr %s\n",
Use inet_ntoa_r() instead of inet_ntoa() throughout the kernel inet_ntoa() cannot be used safely in a multithreaded environment because it uses a static local buffer. Instead, use inet_ntoa_r() with a buffer on the caller's stack. Suggested by: glebius, emaste Reviewed by: gnn MFC after: 2 weeks Sponsored by: Dell EMC Differential Revision: https://reviews.freebsd.org/D9625
2017-02-16 20:47:41 +00:00
inet_ntoa_r(itaddr, addrbuf));
Comply with RFC 3927, by forcing ARP replies which contain a source address within the link-local IPv4 prefix 169.254.0.0/16, to be broadcast at link layer. Reviewed by: fenner MFC after: 2 weeks
2007-02-02 20:31:44 +00:00
#endif
m->m_flags |= M_BCAST;
m->m_flags &= ~M_MCAST;
} else {
/* default behaviour; never reply by broadcast. */
m->m_flags &= ~(M_BCAST|M_MCAST);
}
bring in ARP support for variable length link level addresses Reviewed by: jdp Approved by: jdp Obtained from: NetBSD MFC after: 6 weeks
2001-10-14 20:17:53 +00:00
(void)memcpy(ar_tpa(ah), ar_spa(ah), ah->ar_pln);
(void)memcpy(ar_spa(ah), &itaddr, ah->ar_pln);
ah->ar_op = htons(ARPOP_REPLY);
ah->ar_pro = htons(ETHERTYPE_IP); /* let's be sure! */
Some style(9) and whitespace changes. Together with: Andrey Zonov <andrey zonov.org>
2012-07-31 11:31:12 +00:00
m->m_len = sizeof(*ah) + (2 * ah->ar_pln) + (2 * ah->ar_hln);
m->m_pkthdr.len = m->m_len;
ARP code reuses mbuf from ARP request to make a reply, but it does not reset rcvif to NULL. Since rcvif is not NULL, ipfw(4) supposes that ARP replies were received on specified interface. Reset rcvif to NULL for ARP replies to fix this issue. PR: kern/131817 Reviewed by: glebius MFC after: 1 month
2011-07-04 05:47:48 +00:00
m->m_pkthdr.rcvif = NULL;
Remove interface type specific code from arprequest(), and in_arpinput(). The AF_ARP case in the (*if_output)() routine will handle the interface type specific bits. Obtained from: NetBSD
2004-03-21 06:36:05 +00:00
sa.sa_family = AF_ARP;
sa.sa_len = 2;
Implement interface link header precomputation API. Add if_requestencap() interface method which is capable of calculating various link headers for given interface. Right now there is support for INET/INET6/ARP llheader calculation (IFENCAP_LL type request). Other types are planned to support more complex calculation (L2 multipath lagg nexthops, tunnel encap nexthops, etc..). Reshape 'struct route' to be able to pass additional data (with is length) to prepend to mbuf. These two changes permits routing code to pass pre-calculated nexthop data (like L2 header for route w/gateway) down to the stack eliminating the need for other lookups. It also brings us closer to more complex scenarios like transparently handling MPLS nexthops and tunnel interfaces. Last, but not least, it removes layering violation introduced by flowtable code (ro_lle) and simplifies handling of existing if_output consumers. ARP/ND changes: Make arp/ndp stack pre-calculate link header upon installing/updating lle record. Interface link address change are handled by re-calculating headers for all lles based on if_lladdr event. After these changes, arpresolve()/nd6_resolve() returns full pre-calculated header for supported interfaces thus simplifying if_output(). Move these lookups to separate ether_resolve_addr() function which ether returs error or fully-prepared link header. Add <arp|nd6_>resolve_addr() compat versions to return link addresses instead of pre-calculated data. BPF changes: Raw bpf writes occupied _two_ cases: AF_UNSPEC and pseudo_AF_HDRCMPLT. Despite the naming, both of there have ther header "complete". The only difference is that interface source mac has to be filled by OS for AF_UNSPEC (controlled via BIOCGHDRCMPLT). This logic has to stay inside BPF and not pollute if_output() routines. Convert BPF to pass prepend data via new 'struct route' mechanism. Note that it does not change non-optimized if_output(): ro_prepend handling is purely optional. Side note: hackish pseudo_AF_HDRCMPLT is supported for ethernet and FDDI. It is not needed for ethernet anymore. The only remaining FDDI user is dev/pdq mostly untouched since 2007. FDDI support was eliminated from OpenBSD in 2013 (sys/net/if_fddisubr.c rev 1.65). Flowtable changes: Flowtable violates layering by saving (and not correctly managing) rtes/lles. Instead of passing lle pointer, pass pointer to pre-calculated header data from that lle. Differential Revision: https://reviews.freebsd.org/D4102
2015-12-31 05:03:27 +00:00
/* Calculate link header for sending frame */
bzero(&ro, sizeof(ro));
linkhdrsize = sizeof(linkhdr);
error = arp_fillheader(ifp, ah, 0, linkhdr, &linkhdrsize);
/*
* arp_fillheader() may fail due to lack of support inside encap request
* routing. This is not necessary an error, AF_ARP can/should be handled
* by if_output().
*/
if (error != 0 && error != EAFNOSUPPORT) {
ARP_LOG(LOG_ERR, "Failed to calculate ARP header on %s: %d\n",
if_name(ifp), error);
return;
}
ro.ro_prepend = linkhdr;
ro.ro_plen = linkhdrsize;
ro.ro_flags = 0;
Add m_clrprotoflags() to clear protocol specific mbuf flags at up and downwards layer crossings. Consistently use it within IP, IPv6 and ethernet protocols. Discussed with: trociny, glebius
2013-08-19 13:27:32 +00:00
m_clrprotoflags(m); /* Avoid confusing lower layers. */
Implement interface link header precomputation API. Add if_requestencap() interface method which is capable of calculating various link headers for given interface. Right now there is support for INET/INET6/ARP llheader calculation (IFENCAP_LL type request). Other types are planned to support more complex calculation (L2 multipath lagg nexthops, tunnel encap nexthops, etc..). Reshape 'struct route' to be able to pass additional data (with is length) to prepend to mbuf. These two changes permits routing code to pass pre-calculated nexthop data (like L2 header for route w/gateway) down to the stack eliminating the need for other lookups. It also brings us closer to more complex scenarios like transparently handling MPLS nexthops and tunnel interfaces. Last, but not least, it removes layering violation introduced by flowtable code (ro_lle) and simplifies handling of existing if_output consumers. ARP/ND changes: Make arp/ndp stack pre-calculate link header upon installing/updating lle record. Interface link address change are handled by re-calculating headers for all lles based on if_lladdr event. After these changes, arpresolve()/nd6_resolve() returns full pre-calculated header for supported interfaces thus simplifying if_output(). Move these lookups to separate ether_resolve_addr() function which ether returs error or fully-prepared link header. Add <arp|nd6_>resolve_addr() compat versions to return link addresses instead of pre-calculated data. BPF changes: Raw bpf writes occupied _two_ cases: AF_UNSPEC and pseudo_AF_HDRCMPLT. Despite the naming, both of there have ther header "complete". The only difference is that interface source mac has to be filled by OS for AF_UNSPEC (controlled via BIOCGHDRCMPLT). This logic has to stay inside BPF and not pollute if_output() routines. Convert BPF to pass prepend data via new 'struct route' mechanism. Note that it does not change non-optimized if_output(): ro_prepend handling is purely optional. Side note: hackish pseudo_AF_HDRCMPLT is supported for ethernet and FDDI. It is not needed for ethernet anymore. The only remaining FDDI user is dev/pdq mostly untouched since 2007. FDDI support was eliminated from OpenBSD in 2013 (sys/net/if_fddisubr.c rev 1.65). Flowtable changes: Flowtable violates layering by saving (and not correctly managing) rtes/lles. Instead of passing lle pointer, pass pointer to pre-calculated header data from that lle. Differential Revision: https://reviews.freebsd.org/D4102
2015-12-31 05:03:27 +00:00
(*ifp->if_output)(ifp, m, &sa, &ro);
Add ARP statistics to the kernel and netstat. New counters now exist for: requests sent replies sent requests received replies received packets received total packets dropped due to no ARP entry entrys timed out Duplicate IPs seen The new statistics are seen in the netstat command when it is given the -s command line switch. MFC after: 2 weeks In collaboration with: bz
2009-09-03 21:10:57 +00:00
ARPSTAT_INC(txreplies);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
return;
Another small set of changes to reduce diffs with the new arp code.
2004-04-25 15:00:17 +00:00
drop:
m_freem(m);
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
}
Make INET a proper option. This will not make any of object files that LINT create change; there might be differences with INET disabled, but hardly anything compiled before without INET anyway. Now the 'obvious' things will give a proper error if compiled without inet - ipx_ip, ipfw, tcp_debug. The only thing that _should_ work (but can't be made to compile reasonably easily) is sppp :-( This commit move struct arpcom from <netinet/if_ether.h> to <net/if_arp.h>.
1998-01-08 23:42:31 +00:00
#endif
BSD 4.4 Lite Kernel Sources
1994-05-24 10:09:53 +00:00
Move lle update code from from gigantic ip_arpinput() to separate bunch of functions. The goal is to isolate actual lle updates to permit more fine-grained locking. Do all lle link-level update under AFDATA wlock. Sponsored by: Yandex LLC
2015-08-13 13:38:09 +00:00
/*
* Checks received arp data against existing @la.
* Updates lle state/performs notification if necessary.
*/
static void
arp_check_update_lle(struct arphdr *ah, struct in_addr isaddr, struct ifnet *ifp,
int bridged, struct llentry *la)
{
struct sockaddr sa;
struct mbuf *m_hold, *m_hold_next;
Implement interface link header precomputation API. Add if_requestencap() interface method which is capable of calculating various link headers for given interface. Right now there is support for INET/INET6/ARP llheader calculation (IFENCAP_LL type request). Other types are planned to support more complex calculation (L2 multipath lagg nexthops, tunnel encap nexthops, etc..). Reshape 'struct route' to be able to pass additional data (with is length) to prepend to mbuf. These two changes permits routing code to pass pre-calculated nexthop data (like L2 header for route w/gateway) down to the stack eliminating the need for other lookups. It also brings us closer to more complex scenarios like transparently handling MPLS nexthops and tunnel interfaces. Last, but not least, it removes layering violation introduced by flowtable code (ro_lle) and simplifies handling of existing if_output consumers. ARP/ND changes: Make arp/ndp stack pre-calculate link header upon installing/updating lle record. Interface link address change are handled by re-calculating headers for all lles based on if_lladdr event. After these changes, arpresolve()/nd6_resolve() returns full pre-calculated header for supported interfaces thus simplifying if_output(). Move these lookups to separate ether_resolve_addr() function which ether returs error or fully-prepared link header. Add <arp|nd6_>resolve_addr() compat versions to return link addresses instead of pre-calculated data. BPF changes: Raw bpf writes occupied _two_ cases: AF_UNSPEC and pseudo_AF_HDRCMPLT. Despite the naming, both of there have ther header "complete". The only difference is that interface source mac has to be filled by OS for AF_UNSPEC (controlled via BIOCGHDRCMPLT). This logic has to stay inside BPF and not pollute if_output() routines. Convert BPF to pass prepend data via new 'struct route' mechanism. Note that it does not change non-optimized if_output(): ro_prepend handling is purely optional. Side note: hackish pseudo_AF_HDRCMPLT is supported for ethernet and FDDI. It is not needed for ethernet anymore. The only remaining FDDI user is dev/pdq mostly untouched since 2007. FDDI support was eliminated from OpenBSD in 2013 (sys/net/if_fddisubr.c rev 1.65). Flowtable changes: Flowtable violates layering by saving (and not correctly managing) rtes/lles. Instead of passing lle pointer, pass pointer to pre-calculated header data from that lle. Differential Revision: https://reviews.freebsd.org/D4102
2015-12-31 05:03:27 +00:00
uint8_t linkhdr[LLE_MAX_LINKHDR];
size_t linkhdrsize;
int lladdr_off;
Use inet_ntoa_r() instead of inet_ntoa() throughout the kernel inet_ntoa() cannot be used safely in a multithreaded environment because it uses a static local buffer. Instead, use inet_ntoa_r() with a buffer on the caller's stack. Suggested by: glebius, emaste Reviewed by: gnn MFC after: 2 weeks Sponsored by: Dell EMC Differential Revision: https://reviews.freebsd.org/D9625
2017-02-16 20:47:41 +00:00
char addrbuf[INET_ADDRSTRLEN];
Move lle update code from from gigantic ip_arpinput() to separate bunch of functions. The goal is to isolate actual lle updates to permit more fine-grained locking. Do all lle link-level update under AFDATA wlock. Sponsored by: Yandex LLC
2015-08-13 13:38:09 +00:00
LLE_WLOCK_ASSERT(la);
/* the following is not an error when doing bridging */
if (!bridged && la->lle_tbl->llt_ifp != ifp) {
if (log_arp_wrong_iface)
ARP_LOG(LOG_WARNING, "%s is on %s "
"but got reply from %*D on %s\n",
Use inet_ntoa_r() instead of inet_ntoa() throughout the kernel inet_ntoa() cannot be used safely in a multithreaded environment because it uses a static local buffer. Instead, use inet_ntoa_r() with a buffer on the caller's stack. Suggested by: glebius, emaste Reviewed by: gnn MFC after: 2 weeks Sponsored by: Dell EMC Differential Revision: https://reviews.freebsd.org/D9625
2017-02-16 20:47:41 +00:00
inet_ntoa_r(isaddr, addrbuf),
Move lle update code from from gigantic ip_arpinput() to separate bunch of functions. The goal is to isolate actual lle updates to permit more fine-grained locking. Do all lle link-level update under AFDATA wlock. Sponsored by: Yandex LLC
2015-08-13 13:38:09 +00:00
la->lle_tbl->llt_ifp->if_xname,
ifp->if_addrlen, (u_char *)ar_sha(ah), ":",
ifp->if_xname);
LLE_WUNLOCK(la);
return;
}
if ((la->la_flags & LLE_VALID) &&
Implement interface link header precomputation API. Add if_requestencap() interface method which is capable of calculating various link headers for given interface. Right now there is support for INET/INET6/ARP llheader calculation (IFENCAP_LL type request). Other types are planned to support more complex calculation (L2 multipath lagg nexthops, tunnel encap nexthops, etc..). Reshape 'struct route' to be able to pass additional data (with is length) to prepend to mbuf. These two changes permits routing code to pass pre-calculated nexthop data (like L2 header for route w/gateway) down to the stack eliminating the need for other lookups. It also brings us closer to more complex scenarios like transparently handling MPLS nexthops and tunnel interfaces. Last, but not least, it removes layering violation introduced by flowtable code (ro_lle) and simplifies handling of existing if_output consumers. ARP/ND changes: Make arp/ndp stack pre-calculate link header upon installing/updating lle record. Interface link address change are handled by re-calculating headers for all lles based on if_lladdr event. After these changes, arpresolve()/nd6_resolve() returns full pre-calculated header for supported interfaces thus simplifying if_output(). Move these lookups to separate ether_resolve_addr() function which ether returs error or fully-prepared link header. Add <arp|nd6_>resolve_addr() compat versions to return link addresses instead of pre-calculated data. BPF changes: Raw bpf writes occupied _two_ cases: AF_UNSPEC and pseudo_AF_HDRCMPLT. Despite the naming, both of there have ther header "complete". The only difference is that interface source mac has to be filled by OS for AF_UNSPEC (controlled via BIOCGHDRCMPLT). This logic has to stay inside BPF and not pollute if_output() routines. Convert BPF to pass prepend data via new 'struct route' mechanism. Note that it does not change non-optimized if_output(): ro_prepend handling is purely optional. Side note: hackish pseudo_AF_HDRCMPLT is supported for ethernet and FDDI. It is not needed for ethernet anymore. The only remaining FDDI user is dev/pdq mostly untouched since 2007. FDDI support was eliminated from OpenBSD in 2013 (sys/net/if_fddisubr.c rev 1.65). Flowtable changes: Flowtable violates layering by saving (and not correctly managing) rtes/lles. Instead of passing lle pointer, pass pointer to pre-calculated header data from that lle. Differential Revision: https://reviews.freebsd.org/D4102
2015-12-31 05:03:27 +00:00
bcmp(ar_sha(ah), la->ll_addr, ifp->if_addrlen)) {
Move lle update code from from gigantic ip_arpinput() to separate bunch of functions. The goal is to isolate actual lle updates to permit more fine-grained locking. Do all lle link-level update under AFDATA wlock. Sponsored by: Yandex LLC
2015-08-13 13:38:09 +00:00
if (la->la_flags & LLE_STATIC) {
LLE_WUNLOCK(la);
if (log_arp_permanent_modify)
ARP_LOG(LOG_ERR,
"%*D attempts to modify "
"permanent entry for %s on %s\n",
ifp->if_addrlen,
(u_char *)ar_sha(ah), ":",
Use inet_ntoa_r() instead of inet_ntoa() throughout the kernel inet_ntoa() cannot be used safely in a multithreaded environment because it uses a static local buffer. Instead, use inet_ntoa_r() with a buffer on the caller's stack. Suggested by: glebius, emaste Reviewed by: gnn MFC after: 2 weeks Sponsored by: Dell EMC Differential Revision: https://reviews.freebsd.org/D9625
2017-02-16 20:47:41 +00:00
inet_ntoa_r(isaddr, addrbuf),
ifp->if_xname);
Move lle update code from from gigantic ip_arpinput() to separate bunch of functions. The goal is to isolate actual lle updates to permit more fine-grained locking. Do all lle link-level update under AFDATA wlock. Sponsored by: Yandex LLC
2015-08-13 13:38:09 +00:00
return;
}
if (log_arp_movements) {
ARP_LOG(LOG_INFO, "%s moved from %*D "
"to %*D on %s\n",
Use inet_ntoa_r() instead of inet_ntoa() throughout the kernel inet_ntoa() cannot be used safely in a multithreaded environment because it uses a static local buffer. Instead, use inet_ntoa_r() with a buffer on the caller's stack. Suggested by: glebius, emaste Reviewed by: gnn MFC after: 2 weeks Sponsored by: Dell EMC Differential Revision: https://reviews.freebsd.org/D9625
2017-02-16 20:47:41 +00:00
inet_ntoa_r(isaddr, addrbuf),
Move lle update code from from gigantic ip_arpinput() to separate bunch of functions. The goal is to isolate actual lle updates to permit more fine-grained locking. Do all lle link-level update under AFDATA wlock. Sponsored by: Yandex LLC
2015-08-13 13:38:09 +00:00
ifp->if_addrlen,
Fix the L2 address printed in the "arp: %s moved from %*D" message. In the r292978 struct llentry was changed and the ll_addr field become the pointer. PR: 217667 MFC after: 1 week
2017-03-11 04:57:52 +00:00
(u_char *)la->ll_addr, ":",
Move lle update code from from gigantic ip_arpinput() to separate bunch of functions. The goal is to isolate actual lle updates to permit more fine-grained locking. Do all lle link-level update under AFDATA wlock. Sponsored by: Yandex LLC
2015-08-13 13:38:09 +00:00
ifp->if_addrlen, (u_char *)ar_sha(ah), ":",
ifp->if_xname);
}
}
Implement interface link header precomputation API. Add if_requestencap() interface method which is capable of calculating various link headers for given interface. Right now there is support for INET/INET6/ARP llheader calculation (IFENCAP_LL type request). Other types are planned to support more complex calculation (L2 multipath lagg nexthops, tunnel encap nexthops, etc..). Reshape 'struct route' to be able to pass additional data (with is length) to prepend to mbuf. These two changes permits routing code to pass pre-calculated nexthop data (like L2 header for route w/gateway) down to the stack eliminating the need for other lookups. It also brings us closer to more complex scenarios like transparently handling MPLS nexthops and tunnel interfaces. Last, but not least, it removes layering violation introduced by flowtable code (ro_lle) and simplifies handling of existing if_output consumers. ARP/ND changes: Make arp/ndp stack pre-calculate link header upon installing/updating lle record. Interface link address change are handled by re-calculating headers for all lles based on if_lladdr event. After these changes, arpresolve()/nd6_resolve() returns full pre-calculated header for supported interfaces thus simplifying if_output(). Move these lookups to separate ether_resolve_addr() function which ether returs error or fully-prepared link header. Add <arp|nd6_>resolve_addr() compat versions to return link addresses instead of pre-calculated data. BPF changes: Raw bpf writes occupied _two_ cases: AF_UNSPEC and pseudo_AF_HDRCMPLT. Despite the naming, both of there have ther header "complete". The only difference is that interface source mac has to be filled by OS for AF_UNSPEC (controlled via BIOCGHDRCMPLT). This logic has to stay inside BPF and not pollute if_output() routines. Convert BPF to pass prepend data via new 'struct route' mechanism. Note that it does not change non-optimized if_output(): ro_prepend handling is purely optional. Side note: hackish pseudo_AF_HDRCMPLT is supported for ethernet and FDDI. It is not needed for ethernet anymore. The only remaining FDDI user is dev/pdq mostly untouched since 2007. FDDI support was eliminated from OpenBSD in 2013 (sys/net/if_fddisubr.c rev 1.65). Flowtable changes: Flowtable violates layering by saving (and not correctly managing) rtes/lles. Instead of passing lle pointer, pass pointer to pre-calculated header data from that lle. Differential Revision: https://reviews.freebsd.org/D4102
2015-12-31 05:03:27 +00:00
/* Calculate full link prepend to use in lle */
linkhdrsize = sizeof(linkhdr);
if (lltable_calc_llheader(ifp, AF_INET, ar_sha(ah), linkhdr,
&linkhdrsize, &lladdr_off) != 0)
return;
Move lle update code from from gigantic ip_arpinput() to separate bunch of functions. The goal is to isolate actual lle updates to permit more fine-grained locking. Do all lle link-level update under AFDATA wlock. Sponsored by: Yandex LLC
2015-08-13 13:38:09 +00:00
/* Check if something has changed */
Implement interface link header precomputation API. Add if_requestencap() interface method which is capable of calculating various link headers for given interface. Right now there is support for INET/INET6/ARP llheader calculation (IFENCAP_LL type request). Other types are planned to support more complex calculation (L2 multipath lagg nexthops, tunnel encap nexthops, etc..). Reshape 'struct route' to be able to pass additional data (with is length) to prepend to mbuf. These two changes permits routing code to pass pre-calculated nexthop data (like L2 header for route w/gateway) down to the stack eliminating the need for other lookups. It also brings us closer to more complex scenarios like transparently handling MPLS nexthops and tunnel interfaces. Last, but not least, it removes layering violation introduced by flowtable code (ro_lle) and simplifies handling of existing if_output consumers. ARP/ND changes: Make arp/ndp stack pre-calculate link header upon installing/updating lle record. Interface link address change are handled by re-calculating headers for all lles based on if_lladdr event. After these changes, arpresolve()/nd6_resolve() returns full pre-calculated header for supported interfaces thus simplifying if_output(). Move these lookups to separate ether_resolve_addr() function which ether returs error or fully-prepared link header. Add <arp|nd6_>resolve_addr() compat versions to return link addresses instead of pre-calculated data. BPF changes: Raw bpf writes occupied _two_ cases: AF_UNSPEC and pseudo_AF_HDRCMPLT. Despite the naming, both of there have ther header "complete". The only difference is that interface source mac has to be filled by OS for AF_UNSPEC (controlled via BIOCGHDRCMPLT). This logic has to stay inside BPF and not pollute if_output() routines. Convert BPF to pass prepend data via new 'struct route' mechanism. Note that it does not change non-optimized if_output(): ro_prepend handling is purely optional. Side note: hackish pseudo_AF_HDRCMPLT is supported for ethernet and FDDI. It is not needed for ethernet anymore. The only remaining FDDI user is dev/pdq mostly untouched since 2007. FDDI support was eliminated from OpenBSD in 2013 (sys/net/if_fddisubr.c rev 1.65). Flowtable changes: Flowtable violates layering by saving (and not correctly managing) rtes/lles. Instead of passing lle pointer, pass pointer to pre-calculated header data from that lle. Differential Revision: https://reviews.freebsd.org/D4102
2015-12-31 05:03:27 +00:00
if (memcmp(la->r_linkdata, linkhdr, linkhdrsize) != 0 ||
Move lle update code from from gigantic ip_arpinput() to separate bunch of functions. The goal is to isolate actual lle updates to permit more fine-grained locking. Do all lle link-level update under AFDATA wlock. Sponsored by: Yandex LLC
2015-08-13 13:38:09 +00:00
(la->la_flags & LLE_VALID) == 0) {
Implement interface link header precomputation API. Add if_requestencap() interface method which is capable of calculating various link headers for given interface. Right now there is support for INET/INET6/ARP llheader calculation (IFENCAP_LL type request). Other types are planned to support more complex calculation (L2 multipath lagg nexthops, tunnel encap nexthops, etc..). Reshape 'struct route' to be able to pass additional data (with is length) to prepend to mbuf. These two changes permits routing code to pass pre-calculated nexthop data (like L2 header for route w/gateway) down to the stack eliminating the need for other lookups. It also brings us closer to more complex scenarios like transparently handling MPLS nexthops and tunnel interfaces. Last, but not least, it removes layering violation introduced by flowtable code (ro_lle) and simplifies handling of existing if_output consumers. ARP/ND changes: Make arp/ndp stack pre-calculate link header upon installing/updating lle record. Interface link address change are handled by re-calculating headers for all lles based on if_lladdr event. After these changes, arpresolve()/nd6_resolve() returns full pre-calculated header for supported interfaces thus simplifying if_output(). Move these lookups to separate ether_resolve_addr() function which ether returs error or fully-prepared link header. Add <arp|nd6_>resolve_addr() compat versions to return link addresses instead of pre-calculated data. BPF changes: Raw bpf writes occupied _two_ cases: AF_UNSPEC and pseudo_AF_HDRCMPLT. Despite the naming, both of there have ther header "complete". The only difference is that interface source mac has to be filled by OS for AF_UNSPEC (controlled via BIOCGHDRCMPLT). This logic has to stay inside BPF and not pollute if_output() routines. Convert BPF to pass prepend data via new 'struct route' mechanism. Note that it does not change non-optimized if_output(): ro_prepend handling is purely optional. Side note: hackish pseudo_AF_HDRCMPLT is supported for ethernet and FDDI. It is not needed for ethernet anymore. The only remaining FDDI user is dev/pdq mostly untouched since 2007. FDDI support was eliminated from OpenBSD in 2013 (sys/net/if_fddisubr.c rev 1.65). Flowtable changes: Flowtable violates layering by saving (and not correctly managing) rtes/lles. Instead of passing lle pointer, pass pointer to pre-calculated header data from that lle. Differential Revision: https://reviews.freebsd.org/D4102
2015-12-31 05:03:27 +00:00
/* Try to perform LLE update */
if (lltable_try_set_entry_addr(ifp, la, linkhdr, linkhdrsize,
lladdr_off) == 0)
Move lle update code from from gigantic ip_arpinput() to separate bunch of functions. The goal is to isolate actual lle updates to permit more fine-grained locking. Do all lle link-level update under AFDATA wlock. Sponsored by: Yandex LLC
2015-08-13 13:38:09 +00:00
return;
Remove LLE read lock from IPv4 fast path. LLE structure is mostly unchanged during its lifecycle. To be more specific, there are 2 things relevant for fast path lookup code: 1) link-level address change. Since r286722, these updates are performed under AFDATA WLOCK. 2) Some sort of feedback indicating that this particular entry is used so we re-send arp request to perform reachability verification instead of expiring entry. The only signal that is needed from fast path is something like binary yes/no. The latter is solved by the following changes: 1) introduce special r_skip_req field which is read lockless by fast path, but updated under (new) req_mutex mutex. If this field is non-zero, then fast path will acquire lock and set it back to 0. 2) introduce simple state machine: incomplete->reachable<->verify->deleted. Before that we implicitely had incomplete->reachable->deleted state machine, with V_arpt_keep between "reachable" and "deleted". Verification was performed in runtime 5 seconds before V_arpt_keep expire. This is changed to "change state to verify 5 seconds before V_arpt_keep, set r_skip_req to non-zero value and check it every second". If the value is zero - then send arp verification probe. These changes do not introduce any signifficant control plane overhead: typically lle callout timer would fire 1 time more each V_arpt_keep (1200s) for used lles and up to arp_maxtries (5) for dead lles. As a result, all packets towards "reachable" lle are handled by fast path without acquiring lle read lock. Additional "req_mutex" is needed because callout / arpresolve_slow() or eventhandler might keep LLE lock for signifficant amount of time, which might not be feasible for fast path locking (e.g. having rmlock as ether AFDATA or lltable own lock). Differential Revision: https://reviews.freebsd.org/D3688
2015-12-05 09:50:37 +00:00
/* Clear fast path feedback request if set */
la->r_skip_req = 0;
Move lle update code from from gigantic ip_arpinput() to separate bunch of functions. The goal is to isolate actual lle updates to permit more fine-grained locking. Do all lle link-level update under AFDATA wlock. Sponsored by: Yandex LLC
2015-08-13 13:38:09 +00:00
}
arp_mark_lle_reachable(la);
/*
* The packets are all freed within the call to the output
* routine.
*
* NB: The lock MUST be released before the call to the
* output routine.
*/
if (la->la_hold != NULL) {
m_hold = la->la_hold;
la->la_hold = NULL;
la->la_numheld = 0;
lltable_fill_sa_entry(la, &sa);
LLE_WUNLOCK(la);
for (; m_hold != NULL; m_hold = m_hold_next) {
m_hold_next = m_hold->m_nextpkt;
m_hold->m_nextpkt = NULL;
/* Avoid confusing lower layers. */
m_clrprotoflags(m_hold);
(*ifp->if_output)(ifp, m_hold, &sa, NULL);
}
} else
LLE_WUNLOCK(la);
}
static void
arp_mark_lle_reachable(struct llentry *la)
{
Remove LLE read lock from IPv4 fast path. LLE structure is mostly unchanged during its lifecycle. To be more specific, there are 2 things relevant for fast path lookup code: 1) link-level address change. Since r286722, these updates are performed under AFDATA WLOCK. 2) Some sort of feedback indicating that this particular entry is used so we re-send arp request to perform reachability verification instead of expiring entry. The only signal that is needed from fast path is something like binary yes/no. The latter is solved by the following changes: 1) introduce special r_skip_req field which is read lockless by fast path, but updated under (new) req_mutex mutex. If this field is non-zero, then fast path will acquire lock and set it back to 0. 2) introduce simple state machine: incomplete->reachable<->verify->deleted. Before that we implicitely had incomplete->reachable->deleted state machine, with V_arpt_keep between "reachable" and "deleted". Verification was performed in runtime 5 seconds before V_arpt_keep expire. This is changed to "change state to verify 5 seconds before V_arpt_keep, set r_skip_req to non-zero value and check it every second". If the value is zero - then send arp verification probe. These changes do not introduce any signifficant control plane overhead: typically lle callout timer would fire 1 time more each V_arpt_keep (1200s) for used lles and up to arp_maxtries (5) for dead lles. As a result, all packets towards "reachable" lle are handled by fast path without acquiring lle read lock. Additional "req_mutex" is needed because callout / arpresolve_slow() or eventhandler might keep LLE lock for signifficant amount of time, which might not be feasible for fast path locking (e.g. having rmlock as ether AFDATA or lltable own lock). Differential Revision: https://reviews.freebsd.org/D3688
2015-12-05 09:50:37 +00:00
int canceled, wtime;
Move lle update code from from gigantic ip_arpinput() to separate bunch of functions. The goal is to isolate actual lle updates to permit more fine-grained locking. Do all lle link-level update under AFDATA wlock. Sponsored by: Yandex LLC
2015-08-13 13:38:09 +00:00
LLE_WLOCK_ASSERT(la);
Remove LLE read lock from IPv4 fast path. LLE structure is mostly unchanged during its lifecycle. To be more specific, there are 2 things relevant for fast path lookup code: 1) link-level address change. Since r286722, these updates are performed under AFDATA WLOCK. 2) Some sort of feedback indicating that this particular entry is used so we re-send arp request to perform reachability verification instead of expiring entry. The only signal that is needed from fast path is something like binary yes/no. The latter is solved by the following changes: 1) introduce special r_skip_req field which is read lockless by fast path, but updated under (new) req_mutex mutex. If this field is non-zero, then fast path will acquire lock and set it back to 0. 2) introduce simple state machine: incomplete->reachable<->verify->deleted. Before that we implicitely had incomplete->reachable->deleted state machine, with V_arpt_keep between "reachable" and "deleted". Verification was performed in runtime 5 seconds before V_arpt_keep expire. This is changed to "change state to verify 5 seconds before V_arpt_keep, set r_skip_req to non-zero value and check it every second". If the value is zero - then send arp verification probe. These changes do not introduce any signifficant control plane overhead: typically lle callout timer would fire 1 time more each V_arpt_keep (1200s) for used lles and up to arp_maxtries (5) for dead lles. As a result, all packets towards "reachable" lle are handled by fast path without acquiring lle read lock. Additional "req_mutex" is needed because callout / arpresolve_slow() or eventhandler might keep LLE lock for signifficant amount of time, which might not be feasible for fast path locking (e.g. having rmlock as ether AFDATA or lltable own lock). Differential Revision: https://reviews.freebsd.org/D3688
2015-12-05 09:50:37 +00:00
la->ln_state = ARP_LLINFO_REACHABLE;
Move lle update code from from gigantic ip_arpinput() to separate bunch of functions. The goal is to isolate actual lle updates to permit more fine-grained locking. Do all lle link-level update under AFDATA wlock. Sponsored by: Yandex LLC
2015-08-13 13:38:09 +00:00
EVENTHANDLER_INVOKE(lle_event, la, LLENTRY_RESOLVED);
if (!(la->la_flags & LLE_STATIC)) {
LLE_ADDREF(la);
la->la_expire = time_uptime + V_arpt_keep;
Remove LLE read lock from IPv4 fast path. LLE structure is mostly unchanged during its lifecycle. To be more specific, there are 2 things relevant for fast path lookup code: 1) link-level address change. Since r286722, these updates are performed under AFDATA WLOCK. 2) Some sort of feedback indicating that this particular entry is used so we re-send arp request to perform reachability verification instead of expiring entry. The only signal that is needed from fast path is something like binary yes/no. The latter is solved by the following changes: 1) introduce special r_skip_req field which is read lockless by fast path, but updated under (new) req_mutex mutex. If this field is non-zero, then fast path will acquire lock and set it back to 0. 2) introduce simple state machine: incomplete->reachable<->verify->deleted. Before that we implicitely had incomplete->reachable->deleted state machine, with V_arpt_keep between "reachable" and "deleted". Verification was performed in runtime 5 seconds before V_arpt_keep expire. This is changed to "change state to verify 5 seconds before V_arpt_keep, set r_skip_req to non-zero value and check it every second". If the value is zero - then send arp verification probe. These changes do not introduce any signifficant control plane overhead: typically lle callout timer would fire 1 time more each V_arpt_keep (1200s) for used lles and up to arp_maxtries (5) for dead lles. As a result, all packets towards "reachable" lle are handled by fast path without acquiring lle read lock. Additional "req_mutex" is needed because callout / arpresolve_slow() or eventhandler might keep LLE lock for signifficant amount of time, which might not be feasible for fast path locking (e.g. having rmlock as ether AFDATA or lltable own lock). Differential Revision: https://reviews.freebsd.org/D3688
2015-12-05 09:50:37 +00:00
wtime = V_arpt_keep - V_arp_maxtries * V_arpt_rexmit;
if (wtime < 0)
wtime = V_arpt_keep;
Move lle update code from from gigantic ip_arpinput() to separate bunch of functions. The goal is to isolate actual lle updates to permit more fine-grained locking. Do all lle link-level update under AFDATA wlock. Sponsored by: Yandex LLC
2015-08-13 13:38:09 +00:00
canceled = callout_reset(&la->lle_timer,
Remove LLE read lock from IPv4 fast path. LLE structure is mostly unchanged during its lifecycle. To be more specific, there are 2 things relevant for fast path lookup code: 1) link-level address change. Since r286722, these updates are performed under AFDATA WLOCK. 2) Some sort of feedback indicating that this particular entry is used so we re-send arp request to perform reachability verification instead of expiring entry. The only signal that is needed from fast path is something like binary yes/no. The latter is solved by the following changes: 1) introduce special r_skip_req field which is read lockless by fast path, but updated under (new) req_mutex mutex. If this field is non-zero, then fast path will acquire lock and set it back to 0. 2) introduce simple state machine: incomplete->reachable<->verify->deleted. Before that we implicitely had incomplete->reachable->deleted state machine, with V_arpt_keep between "reachable" and "deleted". Verification was performed in runtime 5 seconds before V_arpt_keep expire. This is changed to "change state to verify 5 seconds before V_arpt_keep, set r_skip_req to non-zero value and check it every second". If the value is zero - then send arp verification probe. These changes do not introduce any signifficant control plane overhead: typically lle callout timer would fire 1 time more each V_arpt_keep (1200s) for used lles and up to arp_maxtries (5) for dead lles. As a result, all packets towards "reachable" lle are handled by fast path without acquiring lle read lock. Additional "req_mutex" is needed because callout / arpresolve_slow() or eventhandler might keep LLE lock for signifficant amount of time, which might not be feasible for fast path locking (e.g. having rmlock as ether AFDATA or lltable own lock). Differential Revision: https://reviews.freebsd.org/D3688
2015-12-05 09:50:37 +00:00
hz * wtime, arptimer, la);
Move lle update code from from gigantic ip_arpinput() to separate bunch of functions. The goal is to isolate actual lle updates to permit more fine-grained locking. Do all lle link-level update under AFDATA wlock. Sponsored by: Yandex LLC
2015-08-13 13:38:09 +00:00
if (canceled)
LLE_REMREF(la);
}
la->la_asked = 0;
la->la_preempt = V_arp_maxtries;
}
Decompose arp_ifinit() into arp_add_ifa_lle() and arp_announce_ifaddr(). Rename arp_ifinit2() into arp_announce_ifaddr(). Eliminate zeroing ifa_rtrequest: it was used for calling arp_rtrequest() which was responsible for handling route cloning requests. It became obsolete since r186119 (L2/L3 split).
2015-11-09 10:35:33 +00:00
/*
Fix comment typo.
2017-08-09 10:46:34 +00:00
* Add permanent link-layer record for given interface address.
Decompose arp_ifinit() into arp_add_ifa_lle() and arp_announce_ifaddr(). Rename arp_ifinit2() into arp_announce_ifaddr(). Eliminate zeroing ifa_rtrequest: it was used for calling arp_rtrequest() which was responsible for handling route cloning requests. It became obsolete since r186119 (L2/L3 split).
2015-11-09 10:35:33 +00:00
*/
static __noinline void
arp_add_ifa_lle(struct ifnet *ifp, const struct sockaddr *dst)
Move ARP interface initialization into if_ether.c:arp_ifinit().
1994-12-22 21:56:22 +00:00
{
* Split allocation and table linking for lle's. Before that, the logic besides lle_create() was the following: return existing if found, create if not. This behaviour was error-prone since we had to deal with 'sudden' static<>dynamic lle changes. This commit fixes bunch of different issues like: - refcount leak when lle is converted to static. Simple check case: console 1: while true; do for i in `arp -an|awk '$4~/incomp/{print$2}'|tr -d '()'`; do arp -s $i 00:22:44:66:88:00 ; arp -d $i; done; done console 2: ping -f any-dead-host-in-L2 console 3: # watch for memory consumption: vmstat -m | awk '$1~/lltable/{print$2}' - possible problems in arptimer() / nd6_timer() when dropping/reacquiring lock. New logic explicitly handles use-or-create cases in every lla_create user. Basically, most of the changes are purely mechanical. However, we explicitly avoid using existing lle's for interface/static LLE records. * While here, call lle_event handlers on all real table lle change. * Create lltable_free_entry() calling existing per-lltable lle_free_t callback for entry deletion
2015-08-20 12:05:17 +00:00
struct llentry *lle, *lle_tmp;
/*
* Interface address LLE record is considered static
* because kernel code relies on LLE_STATIC flag to check
* if these entries can be rewriten by arp updates.
*/
lle = lltable_alloc_entry(LLTABLE(ifp), LLE_IFADDR | LLE_STATIC, dst);
if (lle == NULL) {
log(LOG_INFO, "arp_ifinit: cannot create arp "
"entry for interface address\n");
return;
}
IF_AFDATA_WLOCK(ifp);
LLE_WLOCK(lle);
/* Unlink any entry if exists */
lle_tmp = lla_lookup(LLTABLE(ifp), LLE_EXCLUSIVE, dst);
if (lle_tmp != NULL)
lltable_unlink_entry(LLTABLE(ifp), lle_tmp);
lltable_link_entry(LLTABLE(ifp), lle);
IF_AFDATA_WUNLOCK(ifp);
if (lle_tmp != NULL)
EVENTHANDLER_INVOKE(lle_event, lle_tmp, LLENTRY_EXPIRED);
EVENTHANDLER_INVOKE(lle_event, lle, LLENTRY_RESOLVED);
LLE_WUNLOCK(lle);
if (lle_tmp != NULL)
lltable_free_entry(LLTABLE(ifp), lle_tmp);
Move ARP interface initialization into if_ether.c:arp_ifinit().
1994-12-22 21:56:22 +00:00
}
Lock down the network interface queues. The queue mutex must be obtained before adding/removing packets from the queue. Also, the if_obytes and if_omcasts fields should only be manipulated under protection of the mutex. IF_ENQUEUE, IF_PREPEND, and IF_DEQUEUE perform all necessary locking on the queue. An IF_LOCK macro is provided, as well as the old (mutex-less) versions of the macros in the form _IF_ENQUEUE, _IF_QFULL, for code which needs them, but their use is discouraged. Two new macros are introduced: IF_DRAIN() to drain a queue, and IF_HANDOFF, which takes care of locking/enqueue, and also statistics updating/start if necessary.
2000-11-25 07:35:38 +00:00
Add GARP retransmit capability A single gratuitous ARP (GARP) is always transmitted when an IPv4 address is added to an interface, and that is usually sufficient. However, in some circumstances, such as when a shared address is passed between cluster nodes, this single GARP may occasionally be dropped or lost. This can lead to neighbors on the network link working with a stale ARP cache and sending packets destined for that address to the node that previously owned the address, which may not respond. To avoid this situation, GARP retransmissions can be enabled by setting the net.link.ether.inet.garp_rexmit_count sysctl to a value greater than zero. The setting represents the maximum number of retransmissions. The interval between retransmissions is calculated using an exponential backoff algorithm, doubling each time, so the retransmission intervals are: {1, 2, 4, 8, 16, ...} (seconds). Due to the exponential backoff algorithm used for the interval between GARP retransmissions, the maximum number of retransmissions is limited to 16 for sanity. This limit corresponds to a maximum interval between retransmissions of 2^16 seconds ~= 18 hours. Increasing this limit is possible, but sending out GARPs spaced days apart would be of little use. Submitted by: David A. Bright <david.a.bright@dell.com> MFC after: 1 month Relnotes: yes Sponsored by: Dell EMC Differential Revision: https://reviews.freebsd.org/D7695
2016-10-02 01:42:45 +00:00
/*
* Handle the garp_rexmit_count. Like sysctl_handle_int(), but limits the range
* of valid values.
*/
static int
sysctl_garp_rexmit(SYSCTL_HANDLER_ARGS)
{
int error;
int rexmit_count = *(int *)arg1;
error = sysctl_handle_int(oidp, &rexmit_count, 0, req);
/* Enforce limits on any new value that may have been set. */
if (!error && req->newptr) {
/* A new value was set. */
if (rexmit_count < 0) {
rexmit_count = 0;
} else if (rexmit_count > MAX_GARP_RETRANSMITS) {
rexmit_count = MAX_GARP_RETRANSMITS;
}
*(int *)arg1 = rexmit_count;
}
return (error);
}
/*
* Retransmit a Gratuitous ARP (GARP) and, if necessary, schedule a callout to
* retransmit it again. A pending callout owns a reference to the ifa.
*/
static void
garp_rexmit(void *arg)
{
struct in_ifaddr *ia = arg;
if (callout_pending(&ia->ia_garp_timer) ||
!callout_active(&ia->ia_garp_timer)) {
IF_ADDR_WUNLOCK(ia->ia_ifa.ifa_ifp);
ifa_free(&ia->ia_ifa);
return;
}
garp: Fix vnet related panic for gratuitous arp Gratuitous ARP packets are sent from a timer, which means we don't have a vnet context set. As a result we panic trying to send the packet. Set the vnet context based on the interface associated with the interface address. To reproduce: sysctl net.link.ether.inet.garp_rexmit_count=2 ifconfig vtnet1 10.0.0.1/24 up PR: 235699 Reviewed by: vangyzen@ MFC after: 1 week
2019-02-12 21:22:57 +00:00
CURVNET_SET(ia->ia_ifa.ifa_ifp->if_vnet);
Add GARP retransmit capability A single gratuitous ARP (GARP) is always transmitted when an IPv4 address is added to an interface, and that is usually sufficient. However, in some circumstances, such as when a shared address is passed between cluster nodes, this single GARP may occasionally be dropped or lost. This can lead to neighbors on the network link working with a stale ARP cache and sending packets destined for that address to the node that previously owned the address, which may not respond. To avoid this situation, GARP retransmissions can be enabled by setting the net.link.ether.inet.garp_rexmit_count sysctl to a value greater than zero. The setting represents the maximum number of retransmissions. The interval between retransmissions is calculated using an exponential backoff algorithm, doubling each time, so the retransmission intervals are: {1, 2, 4, 8, 16, ...} (seconds). Due to the exponential backoff algorithm used for the interval between GARP retransmissions, the maximum number of retransmissions is limited to 16 for sanity. This limit corresponds to a maximum interval between retransmissions of 2^16 seconds ~= 18 hours. Increasing this limit is possible, but sending out GARPs spaced days apart would be of little use. Submitted by: David A. Bright <david.a.bright@dell.com> MFC after: 1 month Relnotes: yes Sponsored by: Dell EMC Differential Revision: https://reviews.freebsd.org/D7695
2016-10-02 01:42:45 +00:00
/*
* Drop lock while the ARP request is generated.
*/
IF_ADDR_WUNLOCK(ia->ia_ifa.ifa_ifp);
arprequest(ia->ia_ifa.ifa_ifp, &IA_SIN(ia)->sin_addr,
&IA_SIN(ia)->sin_addr, IF_LLADDR(ia->ia_ifa.ifa_ifp));
/*
* Increment the count of retransmissions. If the count has reached the
* maximum value, stop sending the GARP packets. Otherwise, schedule
* the callout to retransmit another GARP packet.
*/
++ia->ia_garp_count;
if (ia->ia_garp_count >= garp_rexmit_count) {
ifa_free(&ia->ia_ifa);
} else {
int rescheduled;
IF_ADDR_WLOCK(ia->ia_ifa.ifa_ifp);
rescheduled = callout_reset(&ia->ia_garp_timer,
(1 << ia->ia_garp_count) * hz,
garp_rexmit, ia);
IF_ADDR_WUNLOCK(ia->ia_ifa.ifa_ifp);
if (rescheduled) {
ifa_free(&ia->ia_ifa);
}
}
garp: Fix vnet related panic for gratuitous arp Gratuitous ARP packets are sent from a timer, which means we don't have a vnet context set. As a result we panic trying to send the packet. Set the vnet context based on the interface associated with the interface address. To reproduce: sysctl net.link.ether.inet.garp_rexmit_count=2 ifconfig vtnet1 10.0.0.1/24 up PR: 235699 Reviewed by: vangyzen@ MFC after: 1 week
2019-02-12 21:22:57 +00:00
CURVNET_RESTORE();
Add GARP retransmit capability A single gratuitous ARP (GARP) is always transmitted when an IPv4 address is added to an interface, and that is usually sufficient. However, in some circumstances, such as when a shared address is passed between cluster nodes, this single GARP may occasionally be dropped or lost. This can lead to neighbors on the network link working with a stale ARP cache and sending packets destined for that address to the node that previously owned the address, which may not respond. To avoid this situation, GARP retransmissions can be enabled by setting the net.link.ether.inet.garp_rexmit_count sysctl to a value greater than zero. The setting represents the maximum number of retransmissions. The interval between retransmissions is calculated using an exponential backoff algorithm, doubling each time, so the retransmission intervals are: {1, 2, 4, 8, 16, ...} (seconds). Due to the exponential backoff algorithm used for the interval between GARP retransmissions, the maximum number of retransmissions is limited to 16 for sanity. This limit corresponds to a maximum interval between retransmissions of 2^16 seconds ~= 18 hours. Increasing this limit is possible, but sending out GARPs spaced days apart would be of little use. Submitted by: David A. Bright <david.a.bright@dell.com> MFC after: 1 month Relnotes: yes Sponsored by: Dell EMC Differential Revision: https://reviews.freebsd.org/D7695
2016-10-02 01:42:45 +00:00
}
/*
* Start the GARP retransmit timer.
*
* A single GARP is always transmitted when an IPv4 address is added
* to an interface and that is usually sufficient. However, in some
* circumstances, such as when a shared address is passed between
* cluster nodes, this single GARP may occasionally be dropped or
* lost. This can lead to neighbors on the network link working with a
* stale ARP cache and sending packets destined for that address to
* the node that previously owned the address, which may not respond.
*
* To avoid this situation, GARP retransmits can be enabled by setting
* the net.link.ether.inet.garp_rexmit_count sysctl to a value greater
* than zero. The setting represents the maximum number of
* retransmissions. The interval between retransmissions is calculated
* using an exponential backoff algorithm, doubling each time, so the
* retransmission intervals are: {1, 2, 4, 8, 16, ...} (seconds).
*/
static void
garp_timer_start(struct ifaddr *ifa)
{
struct in_ifaddr *ia = (struct in_ifaddr *) ifa;
IF_ADDR_WLOCK(ia->ia_ifa.ifa_ifp);
ia->ia_garp_count = 0;
if (callout_reset(&ia->ia_garp_timer, (1 << ia->ia_garp_count) * hz,
garp_rexmit, ia) == 0) {
ifa_ref(ifa);
}
IF_ADDR_WUNLOCK(ia->ia_ifa.ifa_ifp);
}
Add CARP (Common Address Redundancy Protocol), which allows multiple hosts to share an IP address, providing high availability and load balancing. Original work on CARP done by Michael Shalayeff, with many additions by Marco Pfatschbacher and Ryan McBride. FreeBSD port done solely by Max Laier. Patch by: mlaier Obtained from: OpenBSD (mickey, mcbride)
2005-02-22 13:04:05 +00:00
void
Decompose arp_ifinit() into arp_add_ifa_lle() and arp_announce_ifaddr(). Rename arp_ifinit2() into arp_announce_ifaddr(). Eliminate zeroing ifa_rtrequest: it was used for calling arp_rtrequest() which was responsible for handling route cloning requests. It became obsolete since r186119 (L2/L3 split).
2015-11-09 10:35:33 +00:00
arp_ifinit(struct ifnet *ifp, struct ifaddr *ifa)
Add CARP (Common Address Redundancy Protocol), which allows multiple hosts to share an IP address, providing high availability and load balancing. Original work on CARP done by Michael Shalayeff, with many additions by Marco Pfatschbacher and Ryan McBride. FreeBSD port done solely by Max Laier. Patch by: mlaier Obtained from: OpenBSD (mickey, mcbride)
2005-02-22 13:04:05 +00:00
{
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
struct epoch_tracker et;
Decompose arp_ifinit() into arp_add_ifa_lle() and arp_announce_ifaddr(). Rename arp_ifinit2() into arp_announce_ifaddr(). Eliminate zeroing ifa_rtrequest: it was used for calling arp_rtrequest() which was responsible for handling route cloning requests. It became obsolete since r186119 (L2/L3 split).
2015-11-09 10:35:33 +00:00
const struct sockaddr_in *dst_in;
const struct sockaddr *dst;
if (ifa->ifa_carp != NULL)
return;
dst = ifa->ifa_addr;
dst_in = (const struct sockaddr_in *)dst;
if (ntohl(dst_in->sin_addr.s_addr) == INADDR_ANY)
return;
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
NET_EPOCH_ENTER(et);
Revert r292275 & r292379 glebius has concerns about these changes so reverting those can be discussed and addressed. Sponsored by: Multiplay
2015-12-17 14:41:30 +00:00
arp_announce_ifaddr(ifp, dst_in->sin_addr, IF_LLADDR(ifp));
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
NET_EPOCH_EXIT(et);
Add GARP retransmit capability A single gratuitous ARP (GARP) is always transmitted when an IPv4 address is added to an interface, and that is usually sufficient. However, in some circumstances, such as when a shared address is passed between cluster nodes, this single GARP may occasionally be dropped or lost. This can lead to neighbors on the network link working with a stale ARP cache and sending packets destined for that address to the node that previously owned the address, which may not respond. To avoid this situation, GARP retransmissions can be enabled by setting the net.link.ether.inet.garp_rexmit_count sysctl to a value greater than zero. The setting represents the maximum number of retransmissions. The interval between retransmissions is calculated using an exponential backoff algorithm, doubling each time, so the retransmission intervals are: {1, 2, 4, 8, 16, ...} (seconds). Due to the exponential backoff algorithm used for the interval between GARP retransmissions, the maximum number of retransmissions is limited to 16 for sanity. This limit corresponds to a maximum interval between retransmissions of 2^16 seconds ~= 18 hours. Increasing this limit is possible, but sending out GARPs spaced days apart would be of little use. Submitted by: David A. Bright <david.a.bright@dell.com> MFC after: 1 month Relnotes: yes Sponsored by: Dell EMC Differential Revision: https://reviews.freebsd.org/D7695
2016-10-02 01:42:45 +00:00
if (garp_rexmit_count > 0) {
garp_timer_start(ifa);
}
Decompose arp_ifinit() into arp_add_ifa_lle() and arp_announce_ifaddr(). Rename arp_ifinit2() into arp_announce_ifaddr(). Eliminate zeroing ifa_rtrequest: it was used for calling arp_rtrequest() which was responsible for handling route cloning requests. It became obsolete since r186119 (L2/L3 split).
2015-11-09 10:35:33 +00:00
arp_add_ifa_lle(ifp, dst);
}
Revert r292275 & r292379 glebius has concerns about these changes so reverting those can be discussed and addressed. Sponsored by: Multiplay
2015-12-17 14:41:30 +00:00
void
arp_announce_ifaddr(struct ifnet *ifp, struct in_addr addr, u_char *enaddr)
Decompose arp_ifinit() into arp_add_ifa_lle() and arp_announce_ifaddr(). Rename arp_ifinit2() into arp_announce_ifaddr(). Eliminate zeroing ifa_rtrequest: it was used for calling arp_rtrequest() which was responsible for handling route cloning requests. It became obsolete since r186119 (L2/L3 split).
2015-11-09 10:35:33 +00:00
{
Revert r292275 & r292379 glebius has concerns about these changes so reverting those can be discussed and addressed. Sponsored by: Multiplay
2015-12-17 14:41:30 +00:00
if (ntohl(addr.s_addr) != INADDR_ANY)
arprequest(ifp, &addr, &addr, enaddr);
Add CARP (Common Address Redundancy Protocol), which allows multiple hosts to share an IP address, providing high availability and load balancing. Original work on CARP done by Michael Shalayeff, with many additions by Marco Pfatschbacher and Ryan McBride. FreeBSD port done solely by Max Laier. Patch by: mlaier Obtained from: OpenBSD (mickey, mcbride)
2005-02-22 13:04:05 +00:00
}
Use lladdr_event to propagate gratiotus arp. Differential Revision: https://reviews.freebsd.org/D4019
2015-11-09 10:11:14 +00:00
/*
Revert r292275 & r292379 glebius has concerns about these changes so reverting those can be discussed and addressed. Sponsored by: Multiplay
2015-12-17 14:41:30 +00:00
* Sends gratuitous ARPs for each ifaddr to notify other
* nodes about the address change.
Use lladdr_event to propagate gratiotus arp. Differential Revision: https://reviews.freebsd.org/D4019
2015-11-09 10:11:14 +00:00
*/
Revert r292275 & r292379 glebius has concerns about these changes so reverting those can be discussed and addressed. Sponsored by: Multiplay
2015-12-17 14:41:30 +00:00
static __noinline void
arp_handle_ifllchange(struct ifnet *ifp)
Use lladdr_event to propagate gratiotus arp. Differential Revision: https://reviews.freebsd.org/D4019
2015-11-09 10:11:14 +00:00
{
struct ifaddr *ifa;
Fix lagg failover due to missing notifications When using lagg failover mode neither Gratuitous ARP (IPv4) or Unsolicited Neighbour Advertisements (IPv6) are sent to notify other nodes that the address may have moved. This results is slow failover, dropped packets and network outages for the lagg interface when the primary link goes down. We now use the new if_link_state_change_cond with the force param set to allow lagg to force through link state changes and hence fire a ifnet_link_event which are now monitored by rip and nd6. Upon receiving these events each protocol trigger the relevant notifications: * inet4 => Gratuitous ARP * inet6 => Unsolicited Neighbour Announce This also fixes the carp IPv6 NA's that stopped working after r251584 which added the ipv6_route__llma route. The new behavour can be controlled using the sysctls: * net.link.ether.inet.arp_on_link * net.inet6.icmp6.nd6_on_link Also removed unused param from lagg_port_state and added descriptions for the sysctls while here. PR: 156226 MFC after: 1 month Sponsored by: Multiplay Differential Revision: https://reviews.freebsd.org/D4111
2015-12-15 16:02:11 +00:00
ifnet: Replace if_addr_lock rwlock with epoch + mutex Run on LLNW canaries and tested by pho@ gallatin: Using a 14-core, 28-HTT single socket E5-2697 v3 with a 40GbE MLX5 based ConnectX 4-LX NIC, I see an almost 12% improvement in received packet rate, and a larger improvement in bytes delivered all the way to userspace. When the host receiving 64 streams of netperf -H $DUT -t UDP_STREAM -- -m 1, I see, using nstat -I mce0 1 before the patch: InMpps OMpps InGbs OGbs err TCP Est %CPU syscalls csw irq GBfree 4.98 0.00 4.42 0.00 4235592 33 83.80 4720653 2149771 1235 247.32 4.73 0.00 4.20 0.00 4025260 33 82.99 4724900 2139833 1204 247.32 4.72 0.00 4.20 0.00 4035252 33 82.14 4719162 2132023 1264 247.32 4.71 0.00 4.21 0.00 4073206 33 83.68 4744973 2123317 1347 247.32 4.72 0.00 4.21 0.00 4061118 33 80.82 4713615 2188091 1490 247.32 4.72 0.00 4.21 0.00 4051675 33 85.29 4727399 2109011 1205 247.32 4.73 0.00 4.21 0.00 4039056 33 84.65 4724735 2102603 1053 247.32 After the patch InMpps OMpps InGbs OGbs err TCP Est %CPU syscalls csw irq GBfree 5.43 0.00 4.20 0.00 3313143 33 84.96 5434214 1900162 2656 245.51 5.43 0.00 4.20 0.00 3308527 33 85.24 5439695 1809382 2521 245.51 5.42 0.00 4.19 0.00 3316778 33 87.54 5416028 1805835 2256 245.51 5.42 0.00 4.19 0.00 3317673 33 90.44 5426044 1763056 2332 245.51 5.42 0.00 4.19 0.00 3314839 33 88.11 5435732 1792218 2499 245.52 5.44 0.00 4.19 0.00 3293228 33 91.84 5426301 1668597 2121 245.52 Similarly, netperf reports 230Mb/s before the patch, and 270Mb/s after the patch Reviewed by: gallatin Sponsored by: Limelight Networks Differential Revision: https://reviews.freebsd.org/D15366
2018-05-18 20:13:34 +00:00
CK_STAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) {
Revert r292275 & r292379 glebius has concerns about these changes so reverting those can be discussed and addressed. Sponsored by: Multiplay
2015-12-17 14:41:30 +00:00
if (ifa->ifa_addr->sa_family == AF_INET)
arp_ifinit(ifp, ifa);
Use lladdr_event to propagate gratiotus arp. Differential Revision: https://reviews.freebsd.org/D4019
2015-11-09 10:11:14 +00:00
}
}
/*
Revert r292275 & r292379 glebius has concerns about these changes so reverting those can be discussed and addressed. Sponsored by: Multiplay
2015-12-17 14:41:30 +00:00
* A handler for interface link layer address change event.
Use lladdr_event to propagate gratiotus arp. Differential Revision: https://reviews.freebsd.org/D4019
2015-11-09 10:11:14 +00:00
*/
Implement interface link header precomputation API. Add if_requestencap() interface method which is capable of calculating various link headers for given interface. Right now there is support for INET/INET6/ARP llheader calculation (IFENCAP_LL type request). Other types are planned to support more complex calculation (L2 multipath lagg nexthops, tunnel encap nexthops, etc..). Reshape 'struct route' to be able to pass additional data (with is length) to prepend to mbuf. These two changes permits routing code to pass pre-calculated nexthop data (like L2 header for route w/gateway) down to the stack eliminating the need for other lookups. It also brings us closer to more complex scenarios like transparently handling MPLS nexthops and tunnel interfaces. Last, but not least, it removes layering violation introduced by flowtable code (ro_lle) and simplifies handling of existing if_output consumers. ARP/ND changes: Make arp/ndp stack pre-calculate link header upon installing/updating lle record. Interface link address change are handled by re-calculating headers for all lles based on if_lladdr event. After these changes, arpresolve()/nd6_resolve() returns full pre-calculated header for supported interfaces thus simplifying if_output(). Move these lookups to separate ether_resolve_addr() function which ether returs error or fully-prepared link header. Add <arp|nd6_>resolve_addr() compat versions to return link addresses instead of pre-calculated data. BPF changes: Raw bpf writes occupied _two_ cases: AF_UNSPEC and pseudo_AF_HDRCMPLT. Despite the naming, both of there have ther header "complete". The only difference is that interface source mac has to be filled by OS for AF_UNSPEC (controlled via BIOCGHDRCMPLT). This logic has to stay inside BPF and not pollute if_output() routines. Convert BPF to pass prepend data via new 'struct route' mechanism. Note that it does not change non-optimized if_output(): ro_prepend handling is purely optional. Side note: hackish pseudo_AF_HDRCMPLT is supported for ethernet and FDDI. It is not needed for ethernet anymore. The only remaining FDDI user is dev/pdq mostly untouched since 2007. FDDI support was eliminated from OpenBSD in 2013 (sys/net/if_fddisubr.c rev 1.65). Flowtable changes: Flowtable violates layering by saving (and not correctly managing) rtes/lles. Instead of passing lle pointer, pass pointer to pre-calculated header data from that lle. Differential Revision: https://reviews.freebsd.org/D4102
2015-12-31 05:03:27 +00:00
static void
Use lladdr_event to propagate gratiotus arp. Differential Revision: https://reviews.freebsd.org/D4019
2015-11-09 10:11:14 +00:00
arp_iflladdr(void *arg __unused, struct ifnet *ifp)
{
Implement interface link header precomputation API. Add if_requestencap() interface method which is capable of calculating various link headers for given interface. Right now there is support for INET/INET6/ARP llheader calculation (IFENCAP_LL type request). Other types are planned to support more complex calculation (L2 multipath lagg nexthops, tunnel encap nexthops, etc..). Reshape 'struct route' to be able to pass additional data (with is length) to prepend to mbuf. These two changes permits routing code to pass pre-calculated nexthop data (like L2 header for route w/gateway) down to the stack eliminating the need for other lookups. It also brings us closer to more complex scenarios like transparently handling MPLS nexthops and tunnel interfaces. Last, but not least, it removes layering violation introduced by flowtable code (ro_lle) and simplifies handling of existing if_output consumers. ARP/ND changes: Make arp/ndp stack pre-calculate link header upon installing/updating lle record. Interface link address change are handled by re-calculating headers for all lles based on if_lladdr event. After these changes, arpresolve()/nd6_resolve() returns full pre-calculated header for supported interfaces thus simplifying if_output(). Move these lookups to separate ether_resolve_addr() function which ether returs error or fully-prepared link header. Add <arp|nd6_>resolve_addr() compat versions to return link addresses instead of pre-calculated data. BPF changes: Raw bpf writes occupied _two_ cases: AF_UNSPEC and pseudo_AF_HDRCMPLT. Despite the naming, both of there have ther header "complete". The only difference is that interface source mac has to be filled by OS for AF_UNSPEC (controlled via BIOCGHDRCMPLT). This logic has to stay inside BPF and not pollute if_output() routines. Convert BPF to pass prepend data via new 'struct route' mechanism. Note that it does not change non-optimized if_output(): ro_prepend handling is purely optional. Side note: hackish pseudo_AF_HDRCMPLT is supported for ethernet and FDDI. It is not needed for ethernet anymore. The only remaining FDDI user is dev/pdq mostly untouched since 2007. FDDI support was eliminated from OpenBSD in 2013 (sys/net/if_fddisubr.c rev 1.65). Flowtable changes: Flowtable violates layering by saving (and not correctly managing) rtes/lles. Instead of passing lle pointer, pass pointer to pre-calculated header data from that lle. Differential Revision: https://reviews.freebsd.org/D4102
2015-12-31 05:03:27 +00:00
lltable_update_ifaddr(LLTABLE(ifp));
Revert r292275 & r292379 glebius has concerns about these changes so reverting those can be discussed and addressed. Sponsored by: Multiplay
2015-12-17 14:41:30 +00:00
if ((ifp->if_flags & IFF_UP) != 0)
arp_handle_ifllchange(ifp);
Use lladdr_event to propagate gratiotus arp. Differential Revision: https://reviews.freebsd.org/D4019
2015-11-09 10:11:14 +00:00
}
First pass at separating per-vnet initializer functions from existing functions for initializing global state. At this stage, the new per-vnet initializer functions are directly called from the existing global initialization code, which should in most cases result in compiler inlining those new functions, hence yielding a near-zero functional change. Modify the existing initializer functions which are invoked via protosw, like ip_init() et. al., to allow them to be invoked multiple times, i.e. per each vnet. Global state, if any, is initialized only if such functions are called within the context of vnet0, which will be determined via the IS_DEFAULT_VNET(curvnet) check (currently always true). While here, V_irtualize a few remaining global UMA zones used by net/netinet/netipsec networking code. While it is not yet clear to me or anybody else whether this is the right thing to do, at this stage this makes the code more readable, and makes it easier to track uncollected UMA-zone-backed objects on vnet removal. In the long run, it's quite possible that some form of shared use of UMA zone pools among multiple vnets should be considered. Bump __FreeBSD_version due to changes in layout of structs vnet_ipfw, vnet_inet and vnet_net. Approved by: julian (mentor)
2009-04-06 22:29:41 +00:00
static void
Introduce a per-VNET flag to enable/disable netisr prcessing on that VNET. Add accessor functions to toggle the state per VNET. The base system (vnet0) will always enable itself with the normal registration. We will share the registered protocol handlers in all VNETs minimising duplication and management. Upon disabling netisr processing for a VNET drain the netisr queue from packets for that VNET. Update netisr consumers to (de)register on a per-VNET start/teardown using VNET_SYS(UN)INIT functionality. The change should be transparent for non-VIMAGE kernels. Reviewed by: gnn (, hiren) Obtained from: projects/vnet MFC after: 2 weeks Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D6691
2016-06-03 13:57:10 +00:00
vnet_arp_init(void)
First pass at separating per-vnet initializer functions from existing functions for initializing global state. At this stage, the new per-vnet initializer functions are directly called from the existing global initialization code, which should in most cases result in compiler inlining those new functions, hence yielding a near-zero functional change. Modify the existing initializer functions which are invoked via protosw, like ip_init() et. al., to allow them to be invoked multiple times, i.e. per each vnet. Global state, if any, is initialized only if such functions are called within the context of vnet0, which will be determined via the IS_DEFAULT_VNET(curvnet) check (currently always true). While here, V_irtualize a few remaining global UMA zones used by net/netinet/netipsec networking code. While it is not yet clear to me or anybody else whether this is the right thing to do, at this stage this makes the code more readable, and makes it easier to track uncollected UMA-zone-backed objects on vnet removal. In the long run, it's quite possible that some form of shared use of UMA zone pools among multiple vnets should be considered. Bump __FreeBSD_version due to changes in layout of structs vnet_ipfw, vnet_inet and vnet_net. Approved by: julian (mentor)
2009-04-06 22:29:41 +00:00
{
Introduce a per-VNET flag to enable/disable netisr prcessing on that VNET. Add accessor functions to toggle the state per VNET. The base system (vnet0) will always enable itself with the normal registration. We will share the registered protocol handlers in all VNETs minimising duplication and management. Upon disabling netisr processing for a VNET drain the netisr queue from packets for that VNET. Update netisr consumers to (de)register on a per-VNET start/teardown using VNET_SYS(UN)INIT functionality. The change should be transparent for non-VIMAGE kernels. Reviewed by: gnn (, hiren) Obtained from: projects/vnet MFC after: 2 weeks Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D6691
2016-06-03 13:57:10 +00:00
if (IS_DEFAULT_VNET(curvnet)) {
netisr_register(&arp_nh);
Use lladdr_event to propagate gratiotus arp. Differential Revision: https://reviews.freebsd.org/D4019
2015-11-09 10:11:14 +00:00
iflladdr_tag = EVENTHANDLER_REGISTER(iflladdr_event,
arp_iflladdr, NULL, EVENTHANDLER_PRI_ANY);
Introduce a per-VNET flag to enable/disable netisr prcessing on that VNET. Add accessor functions to toggle the state per VNET. The base system (vnet0) will always enable itself with the normal registration. We will share the registered protocol handlers in all VNETs minimising duplication and management. Upon disabling netisr processing for a VNET drain the netisr queue from packets for that VNET. Update netisr consumers to (de)register on a per-VNET start/teardown using VNET_SYS(UN)INIT functionality. The change should be transparent for non-VIMAGE kernels. Reviewed by: gnn (, hiren) Obtained from: projects/vnet MFC after: 2 weeks Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D6691
2016-06-03 13:57:10 +00:00
}
#ifdef VIMAGE
else
netisr_register_vnet(&arp_nh);
#endif
Lock down the network interface queues. The queue mutex must be obtained before adding/removing packets from the queue. Also, the if_obytes and if_omcasts fields should only be manipulated under protection of the mutex. IF_ENQUEUE, IF_PREPEND, and IF_DEQUEUE perform all necessary locking on the queue. An IF_LOCK macro is provided, as well as the old (mutex-less) versions of the macros in the form _IF_ENQUEUE, _IF_QFULL, for code which needs them, but their use is discouraged. Two new macros are introduced: IF_DRAIN() to drain a queue, and IF_HANDOFF, which takes care of locking/enqueue, and also statistics updating/start if necessary.
2000-11-25 07:35:38 +00:00
}
Introduce a per-VNET flag to enable/disable netisr prcessing on that VNET. Add accessor functions to toggle the state per VNET. The base system (vnet0) will always enable itself with the normal registration. We will share the registered protocol handlers in all VNETs minimising duplication and management. Upon disabling netisr processing for a VNET drain the netisr queue from packets for that VNET. Update netisr consumers to (de)register on a per-VNET start/teardown using VNET_SYS(UN)INIT functionality. The change should be transparent for non-VIMAGE kernels. Reviewed by: gnn (, hiren) Obtained from: projects/vnet MFC after: 2 weeks Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D6691
2016-06-03 13:57:10 +00:00
VNET_SYSINIT(vnet_arp_init, SI_SUB_PROTO_DOMAIN, SI_ORDER_SECOND,
vnet_arp_init, 0);
#ifdef VIMAGE
/*
* We have to unregister ARP along with IP otherwise we risk doing INADDR_HASH
* lookups after destroying the hash. Ideally this would go on SI_ORDER_3.5.
*/
static void
vnet_arp_destroy(__unused void *arg)
{
netisr_unregister_vnet(&arp_nh);
}
VNET_SYSUNINIT(vnet_arp_uninit, SI_SUB_PROTO_DOMAIN, SI_ORDER_THIRD,
vnet_arp_destroy, NULL);
#endif
Reference in New Issue Copy Permalink