2005-01-07 01:45:51 +00:00
|
|
|
/*-
|
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.
|
|
|
|
* 4. Neither the name of the University nor the names of its contributors
|
|
|
|
* may be used to endorse or promote products derived from this software
|
|
|
|
* without specific prior written permission.
|
|
|
|
*
|
|
|
|
* THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
|
|
|
|
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
|
|
|
|
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
|
|
|
|
* ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
|
|
|
|
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
|
|
|
|
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
|
|
|
|
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
|
|
|
|
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
|
|
|
|
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
|
|
|
|
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
|
|
|
|
* SUCH DAMAGE.
|
|
|
|
*
|
|
|
|
* @(#)ip_input.c 8.2 (Berkeley) 1/4/94
|
|
|
|
*/
|
|
|
|
|
2007-10-07 20:44:24 +00:00
|
|
|
#include <sys/cdefs.h>
|
|
|
|
__FBSDID("$FreeBSD$");
|
|
|
|
|
2004-07-08 22:35:36 +00:00
|
|
|
#include "opt_bootp.h"
|
1996-06-12 19:34:33 +00:00
|
|
|
#include "opt_ipfw.h"
|
1999-08-29 12:18:39 +00:00
|
|
|
#include "opt_ipstealth.h"
|
1999-12-22 19:13:38 +00:00
|
|
|
#include "opt_ipsec.h"
|
2009-02-27 14:12:05 +00:00
|
|
|
#include "opt_route.h"
|
2005-02-22 13:04:05 +00:00
|
|
|
#include "opt_carp.h"
|
1996-06-12 19:34:33 +00:00
|
|
|
|
1994-05-24 10:09:53 +00:00
|
|
|
#include <sys/param.h>
|
|
|
|
#include <sys/systm.h>
|
2005-01-02 01:50:57 +00:00
|
|
|
#include <sys/callout.h>
|
1994-05-24 10:09:53 +00:00
|
|
|
#include <sys/mbuf.h>
|
1998-12-14 18:09:13 +00:00
|
|
|
#include <sys/malloc.h>
|
1994-05-24 10:09:53 +00:00
|
|
|
#include <sys/domain.h>
|
|
|
|
#include <sys/protosw.h>
|
|
|
|
#include <sys/socket.h>
|
|
|
|
#include <sys/time.h>
|
|
|
|
#include <sys/kernel.h>
|
Conditionally compile out V_ globals while instantiating the appropriate
container structures, depending on VIMAGE_GLOBALS compile time option.
Make VIMAGE_GLOBALS a new compile-time option, which by default will not
be defined, resulting in instatiations of global variables selected for
V_irtualization (enclosed in #ifdef VIMAGE_GLOBALS blocks) to be
effectively compiled out. Instantiate new global container structures
to hold V_irtualized variables: vnet_net_0, vnet_inet_0, vnet_inet6_0,
vnet_ipsec_0, vnet_netgraph_0, and vnet_gif_0.
Update the VSYM() macro so that depending on VIMAGE_GLOBALS the V_
macros resolve either to the original globals, or to fields inside
container structures, i.e. effectively
#ifdef VIMAGE_GLOBALS
#define V_rt_tables rt_tables
#else
#define V_rt_tables vnet_net_0._rt_tables
#endif
Update SYSCTL_V_*() macros to operate either on globals or on fields
inside container structs.
Extend the internal kldsym() lookups with the ability to resolve
selected fields inside the virtualization container structs. This
applies only to the fields which are explicitly registered for kldsym()
visibility via VNET_MOD_DECLARE() and vnet_mod_register(), currently
this is done only in sys/net/if.c.
Fix a few broken instances of MODULE_GLOBAL() macro use in SCTP code,
and modify the MODULE_GLOBAL() macro to resolve to V_ macros, which in
turn result in proper code being generated depending on VIMAGE_GLOBALS.
De-virtualize local static variables in sys/contrib/pf/net/pf_subr.c
which were prematurely V_irtualized by automated V_ prepending scripts
during earlier merging steps. PF virtualization will be done
separately, most probably after next PF import.
Convert a few variable initializations at instantiation to
initialization in init functions, most notably in ipfw. Also convert
TUNABLE_INT() initializers for V_ variables to TUNABLE_FETCH_INT() in
initializer functions.
Discussed at: devsummit Strassburg
Reviewed by: bz, julian
Approved by: julian (mentor)
Obtained from: //depot/projects/vimage-commit2/...
X-MFC after: never
Sponsored by: NLnet Foundation, The FreeBSD Foundation
2008-12-10 23:12:39 +00:00
|
|
|
#include <sys/lock.h>
|
|
|
|
#include <sys/rwlock.h>
|
1995-03-16 18:22:28 +00:00
|
|
|
#include <sys/syslog.h>
|
1995-03-16 18:17:34 +00:00
|
|
|
#include <sys/sysctl.h>
|
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
|
|
|
#include <sys/vimage.h>
|
1994-05-24 10:09:53 +00:00
|
|
|
|
2000-07-31 23:41:47 +00:00
|
|
|
#include <net/pfil.h>
|
1994-05-24 10:09:53 +00:00
|
|
|
#include <net/if.h>
|
2001-09-25 18:40:52 +00:00
|
|
|
#include <net/if_types.h>
|
1997-11-05 02:51:32 +00:00
|
|
|
#include <net/if_var.h>
|
1996-11-11 04:56:32 +00:00
|
|
|
#include <net/if_dl.h>
|
1994-05-24 10:09:53 +00:00
|
|
|
#include <net/route.h>
|
1995-05-11 00:13:26 +00:00
|
|
|
#include <net/netisr.h>
|
2008-12-02 21:37:28 +00:00
|
|
|
#include <net/vnet.h>
|
2009-04-19 04:44:05 +00:00
|
|
|
#include <net/flowtable.h>
|
1994-05-24 10:09:53 +00:00
|
|
|
|
|
|
|
#include <netinet/in.h>
|
|
|
|
#include <netinet/in_systm.h>
|
1995-03-16 18:17:34 +00:00
|
|
|
#include <netinet/in_var.h>
|
1994-05-24 10:09:53 +00:00
|
|
|
#include <netinet/ip.h>
|
|
|
|
#include <netinet/in_pcb.h>
|
|
|
|
#include <netinet/ip_var.h>
|
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
|
|
|
#include <netinet/ip_fw.h>
|
1994-05-24 10:09:53 +00:00
|
|
|
#include <netinet/ip_icmp.h>
|
2005-11-18 20:12:40 +00:00
|
|
|
#include <netinet/ip_options.h>
|
1996-10-07 19:21:46 +00:00
|
|
|
#include <machine/in_cksum.h>
|
2005-02-22 13:04:05 +00:00
|
|
|
#ifdef DEV_CARP
|
|
|
|
#include <netinet/ip_carp.h>
|
|
|
|
#endif
|
2007-07-03 12:13:45 +00:00
|
|
|
#ifdef IPSEC
|
2006-02-01 13:55:03 +00:00
|
|
|
#include <netinet/ip_ipsec.h>
|
2007-07-03 12:13:45 +00:00
|
|
|
#endif /* IPSEC */
|
1994-05-24 10:09:53 +00:00
|
|
|
|
1994-09-06 22:42:31 +00:00
|
|
|
#include <sys/socketvar.h>
|
1996-05-08 04:29:08 +00:00
|
|
|
|
2006-10-22 11:52:19 +00:00
|
|
|
#include <security/mac/mac_framework.h>
|
|
|
|
|
2008-09-26 18:30:11 +00:00
|
|
|
#ifdef CTASSERT
|
|
|
|
CTASSERT(sizeof(struct ip) == 20);
|
|
|
|
#endif
|
|
|
|
|
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
|
|
|
static VNET_DEFINE(int, ipsendredirects) = 1; /* XXX */
|
|
|
|
static VNET_DEFINE(int, ip_checkinterface);
|
|
|
|
static VNET_DEFINE(int, ip_keepfaith);
|
|
|
|
static VNET_DEFINE(int, ip_sendsourcequench);
|
Conditionally compile out V_ globals while instantiating the appropriate
container structures, depending on VIMAGE_GLOBALS compile time option.
Make VIMAGE_GLOBALS a new compile-time option, which by default will not
be defined, resulting in instatiations of global variables selected for
V_irtualization (enclosed in #ifdef VIMAGE_GLOBALS blocks) to be
effectively compiled out. Instantiate new global container structures
to hold V_irtualized variables: vnet_net_0, vnet_inet_0, vnet_inet6_0,
vnet_ipsec_0, vnet_netgraph_0, and vnet_gif_0.
Update the VSYM() macro so that depending on VIMAGE_GLOBALS the V_
macros resolve either to the original globals, or to fields inside
container structures, i.e. effectively
#ifdef VIMAGE_GLOBALS
#define V_rt_tables rt_tables
#else
#define V_rt_tables vnet_net_0._rt_tables
#endif
Update SYSCTL_V_*() macros to operate either on globals or on fields
inside container structs.
Extend the internal kldsym() lookups with the ability to resolve
selected fields inside the virtualization container structs. This
applies only to the fields which are explicitly registered for kldsym()
visibility via VNET_MOD_DECLARE() and vnet_mod_register(), currently
this is done only in sys/net/if.c.
Fix a few broken instances of MODULE_GLOBAL() macro use in SCTP code,
and modify the MODULE_GLOBAL() macro to resolve to V_ macros, which in
turn result in proper code being generated depending on VIMAGE_GLOBALS.
De-virtualize local static variables in sys/contrib/pf/net/pf_subr.c
which were prematurely V_irtualized by automated V_ prepending scripts
during earlier merging steps. PF virtualization will be done
separately, most probably after next PF import.
Convert a few variable initializations at instantiation to
initialization in init functions, most notably in ipfw. Also convert
TUNABLE_INT() initializers for V_ variables to TUNABLE_FETCH_INT() in
initializer functions.
Discussed at: devsummit Strassburg
Reviewed by: bz, julian
Approved by: julian (mentor)
Obtained from: //depot/projects/vimage-commit2/...
X-MFC after: never
Sponsored by: NLnet Foundation, The FreeBSD Foundation
2008-12-10 23:12:39 +00:00
|
|
|
|
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
|
|
|
#define V_ipsendredirects VNET_GET(ipsendredirects)
|
|
|
|
#define V_ip_checkinterface VNET_GET(ip_checkinterface)
|
|
|
|
#define V_ip_keepfaith VNET_GET(ip_keepfaith)
|
|
|
|
#define V_ip_sendsourcequench VNET_GET(ip_sendsourcequench)
|
|
|
|
|
|
|
|
VNET_DEFINE(int, ip_defttl) = IPDEFTTL;
|
|
|
|
VNET_DEFINE(int, ip_do_randomid);
|
|
|
|
VNET_DEFINE(int, ipforwarding);
|
|
|
|
|
|
|
|
VNET_DEFINE(struct in_ifaddrhead, in_ifaddrhead); /* first inet address */
|
|
|
|
VNET_DEFINE(struct in_ifaddrhashhead *, in_ifaddrhashtbl); /* inet addr hash table */
|
|
|
|
VNET_DEFINE(u_long, in_ifaddrhmask); /* mask for hash table */
|
|
|
|
VNET_DEFINE(struct ipstat, ipstat);
|
|
|
|
|
|
|
|
static VNET_DEFINE(int, ip_rsvp_on);
|
|
|
|
VNET_DEFINE(struct socket *, ip_rsvpd);
|
|
|
|
VNET_DEFINE(int, rsvp_on);
|
|
|
|
|
|
|
|
#define V_ip_rsvp_on VNET_GET(ip_rsvp_on)
|
|
|
|
|
|
|
|
static VNET_DEFINE(TAILQ_HEAD(ipqhead, ipq), ipq[IPREASS_NHASH]);
|
|
|
|
static VNET_DEFINE(int, maxnipq); /* Administrative limit on # reass queues. */
|
|
|
|
static VNET_DEFINE(int, maxfragsperpacket);
|
|
|
|
static VNET_DEFINE(int, nipq); /* Total # of reass queues */
|
|
|
|
|
|
|
|
#define V_ipq VNET_GET(ipq)
|
|
|
|
#define V_maxnipq VNET_GET(maxnipq)
|
|
|
|
#define V_maxfragsperpacket VNET_GET(maxfragsperpacket)
|
|
|
|
#define V_nipq VNET_GET(nipq)
|
|
|
|
|
|
|
|
VNET_DEFINE(int, ipstealth);
|
2009-06-25 14:44:00 +00:00
|
|
|
|
2009-06-25 11:52:33 +00:00
|
|
|
struct rwlock in_ifaddr_lock;
|
2009-06-25 14:44:00 +00:00
|
|
|
RW_SYSINIT(in_ifaddr_lock, &in_ifaddr_lock, "in_ifaddr_lock");
|
1994-09-06 22:42:31 +00:00
|
|
|
|
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
|
|
|
SYSCTL_VNET_INT(_net_inet_ip, IPCTL_FORWARDING, forwarding, CTLFLAG_RW,
|
|
|
|
&VNET_NAME(ipforwarding), 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 IP forwarding between interfaces");
|
1995-11-14 20:34:56 +00:00
|
|
|
|
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
|
|
|
SYSCTL_VNET_INT(_net_inet_ip, IPCTL_SENDREDIRECTS, redirect, CTLFLAG_RW,
|
|
|
|
&VNET_NAME(ipsendredirects), 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 sending IP redirects");
|
1995-11-14 20:34:56 +00:00
|
|
|
|
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
|
|
|
SYSCTL_VNET_INT(_net_inet_ip, IPCTL_DEFTTL, ttl, CTLFLAG_RW,
|
|
|
|
&VNET_NAME(ip_defttl), 0,
|
|
|
|
"Maximum TTL on IP packets");
|
1995-11-14 20:34:56 +00:00
|
|
|
|
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
|
|
|
SYSCTL_VNET_INT(_net_inet_ip, IPCTL_KEEPFAITH, keepfaith, CTLFLAG_RW,
|
|
|
|
&VNET_NAME(ip_keepfaith), 0,
|
2007-03-19 19:00:51 +00:00
|
|
|
"Enable packet capture for FAITH IPv4->IPv6 translater daemon");
|
1999-12-22 19:13:38 +00:00
|
|
|
|
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
|
|
|
SYSCTL_VNET_INT(_net_inet_ip, OID_AUTO, sendsourcequench, CTLFLAG_RW,
|
|
|
|
&VNET_NAME(ip_sendsourcequench), 0,
|
2007-03-19 19:00:51 +00:00
|
|
|
"Enable the transmission of source quench packets");
|
2002-11-19 17:06:06 +00:00
|
|
|
|
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
|
|
|
SYSCTL_VNET_INT(_net_inet_ip, OID_AUTO, random_id, CTLFLAG_RW,
|
|
|
|
&VNET_NAME(ip_do_randomid), 0,
|
|
|
|
"Assign random ip_id values");
|
2004-08-14 15:32:40 +00:00
|
|
|
|
2001-03-04 01:39:19 +00:00
|
|
|
/*
|
|
|
|
* XXX - Setting ip_checkinterface mostly implements the receive side of
|
|
|
|
* the Strong ES model described in RFC 1122, but since the routing table
|
2001-03-05 22:40:27 +00:00
|
|
|
* and transmit implementation do not implement the Strong ES model,
|
2001-03-04 01:39:19 +00:00
|
|
|
* setting this to 1 results in an odd hybrid.
|
2001-03-05 08:45:05 +00:00
|
|
|
*
|
2001-03-05 22:40:27 +00:00
|
|
|
* XXX - ip_checkinterface currently must be disabled if you use ipnat
|
|
|
|
* to translate the destination address to another local interface.
|
2001-03-05 08:45:05 +00:00
|
|
|
*
|
|
|
|
* XXX - ip_checkinterface must be disabled if you add IP aliases
|
|
|
|
* to the loopback interface instead of the interface where the
|
|
|
|
* packets for those addresses are received.
|
2001-03-04 01:39:19 +00:00
|
|
|
*/
|
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
|
|
|
SYSCTL_VNET_INT(_net_inet_ip, OID_AUTO, check_interface, CTLFLAG_RW,
|
|
|
|
&VNET_NAME(ip_checkinterface), 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
|
|
|
"Verify packet arrives on correct interface");
|
2001-03-02 20:54:03 +00:00
|
|
|
|
2004-08-27 15:16:24 +00:00
|
|
|
struct pfil_head inet_pfil_hook; /* Packet filter hooks */
|
1994-05-24 10:09:53 +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 struct netisr_handler ip_nh = {
|
|
|
|
.nh_name = "ip",
|
|
|
|
.nh_handler = ip_input,
|
|
|
|
.nh_proto = NETISR_IP,
|
|
|
|
.nh_policy = NETISR_POLICY_FLOW,
|
|
|
|
};
|
2001-09-29 04:34:11 +00:00
|
|
|
|
1994-05-24 10:09:53 +00:00
|
|
|
extern struct domain inetdomain;
|
2001-09-03 20:03:55 +00:00
|
|
|
extern struct protosw inetsw[];
|
1994-05-24 10:09:53 +00:00
|
|
|
u_char ip_protox[IPPROTO_MAX];
|
2001-09-29 04:34:11 +00:00
|
|
|
|
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
|
|
|
SYSCTL_VNET_STRUCT(_net_inet_ip, IPCTL_STATS, stats, CTLFLAG_RW,
|
|
|
|
&VNET_NAME(ipstat), ipstat,
|
|
|
|
"IP statistics (struct ipstat, netinet/ip_var.h)");
|
1994-05-24 10:09:53 +00:00
|
|
|
|
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
|
|
|
static VNET_DEFINE(uma_zone_t, ipq_zone);
|
|
|
|
#define V_ipq_zone VNET_GET(ipq_zone)
|
1997-09-15 23:07:01 +00:00
|
|
|
|
2006-01-15 17:05:48 +00:00
|
|
|
static struct mtx ipqlock;
|
2003-09-05 00:10:33 +00:00
|
|
|
|
|
|
|
#define IPQ_LOCK() mtx_lock(&ipqlock)
|
|
|
|
#define IPQ_UNLOCK() mtx_unlock(&ipqlock)
|
2003-10-14 18:45:50 +00:00
|
|
|
#define IPQ_LOCK_INIT() mtx_init(&ipqlock, "ipqlock", NULL, MTX_DEF)
|
|
|
|
#define IPQ_LOCK_ASSERT() mtx_assert(&ipqlock, MA_OWNED)
|
1994-08-18 22:36:09 +00:00
|
|
|
|
2006-01-15 18:58:21 +00:00
|
|
|
static void maxnipq_update(void);
|
2006-04-21 09:25:40 +00:00
|
|
|
static void ipq_zone_change(void *);
|
2006-01-15 18:58:21 +00:00
|
|
|
|
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
|
|
|
SYSCTL_VNET_INT(_net_inet_ip, OID_AUTO, fragpackets, CTLFLAG_RD,
|
|
|
|
&VNET_NAME(nipq), 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
|
|
|
"Current number of IPv4 fragment reassembly queue entries");
|
2006-01-15 18:58:21 +00:00
|
|
|
|
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
|
|
|
SYSCTL_VNET_INT(_net_inet_ip, OID_AUTO, maxfragsperpacket, CTLFLAG_RW,
|
|
|
|
&VNET_NAME(maxfragsperpacket), 0,
|
2007-03-19 19:00:51 +00:00
|
|
|
"Maximum number of IPv4 fragments allowed per packet");
|
2006-01-15 18:58:21 +00:00
|
|
|
|
|
|
|
struct callout ipport_tick_callout;
|
|
|
|
|
1995-11-14 20:34:56 +00:00
|
|
|
#ifdef IPCTL_DEFMTU
|
|
|
|
SYSCTL_INT(_net_inet_ip, IPCTL_DEFMTU, mtu, CTLFLAG_RW,
|
1999-05-03 23:57:32 +00:00
|
|
|
&ip_mtu, 0, "Default MTU");
|
1995-11-14 20:34:56 +00:00
|
|
|
#endif
|
|
|
|
|
1999-02-22 18:19:57 +00:00
|
|
|
#ifdef IPSTEALTH
|
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
|
|
|
SYSCTL_VNET_INT(_net_inet_ip, OID_AUTO, stealth, CTLFLAG_RW,
|
|
|
|
&VNET_NAME(ipstealth), 0,
|
|
|
|
"IP stealth mode, no TTL decrementation on forwarding");
|
1999-02-22 18:19:57 +00:00
|
|
|
#endif
|
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
|
|
|
|
2009-06-12 20:46:36 +00:00
|
|
|
#ifdef FLOWTABLE
|
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
|
|
|
static VNET_DEFINE(int, ip_output_flowtable_size) = 2048;
|
|
|
|
VNET_DEFINE(struct flowtable *, ip_ft);
|
|
|
|
#define V_ip_output_flowtable_size VNET_GET(ip_output_flowtable_size)
|
|
|
|
|
|
|
|
SYSCTL_VNET_INT(_net_inet_ip, OID_AUTO, output_flowtable_size, CTLFLAG_RDTUN,
|
|
|
|
&VNET_NAME(ip_output_flowtable_size), 2048,
|
2009-04-19 04:44:05 +00:00
|
|
|
"number of entries in the per-cpu output flow caches");
|
2009-06-12 20:46:36 +00:00
|
|
|
#endif
|
|
|
|
|
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_DEFINE(int, fw_one_pass) = 1;
|
2006-10-21 00:16:31 +00:00
|
|
|
|
2002-03-19 21:25:46 +00:00
|
|
|
static void ip_freef(struct ipqhead *, struct ipq *);
|
1999-12-06 00:43:07 +00:00
|
|
|
|
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
|
|
|
#ifdef VIMAGE
|
|
|
|
/* XXX only has to stay for .vmi_dependson elsewhere. */
|
Introduce vnet module registration / initialization framework with
dependency tracking and ordering enforcement.
With this change, per-vnet initialization functions introduced with
r190787 are no longer directly called from traditional initialization
functions (which cc in most cases inlined to pre-r190787 code), but are
instead registered via the vnet framework first, and are invoked only
after all prerequisite modules have been initialized. In the long run,
this framework should allow us to both initialize and dismantle
multiple vnet instances in a correct order.
The problem this change aims to solve is how to replay the
initialization sequence of various network stack components, which
have been traditionally triggered via different mechanisms (SYSINIT,
protosw). Note that this initialization sequence was and still can be
subtly different depending on whether certain pieces of code have been
statically compiled into the kernel, loaded as modules by boot
loader, or kldloaded at run time.
The approach is simple - we record the initialization sequence
established by the traditional mechanisms whenever vnet_mod_register()
is called for a particular vnet module. The vnet_mod_register_multi()
variant allows a single initializer function to be registered multiple
times but with different arguments - currently this is only used in
kern/uipc_domain.c by net_add_domain() with different struct domain *
as arguments, which allows for protosw-registered initialization
routines to be invoked in a correct order by the new vnet
initialization framework.
For the purpose of identifying vnet modules, each vnet module has to
have a unique ID, which is statically assigned in sys/vimage.h.
Dynamic assignment of vnet module IDs is not supported yet.
A vnet module may specify a single prerequisite module at registration
time by filling in the vmi_dependson field of its vnet_modinfo struct
with the ID of the module it depends on. Unless specified otherwise,
all vnet modules depend on VNET_MOD_NET (container for ifnet list head,
rt_tables etc.), which thus has to and will always be initialized
first. The framework will panic if it detects any unresolved
dependencies before completing system initialization. Detection of
unresolved dependencies for vnet modules registered after boot
(kldloaded modules) is not provided.
Note that the fact that each module can specify only a single
prerequisite may become problematic in the long run. In particular,
INET6 depends on INET being already instantiated, due to TCP / UDP
structures residing in INET container. IPSEC also depends on INET,
which will in turn additionally complicate making INET6-only kernel
configs a reality.
The entire registration framework can be compiled out by turning on the
VIMAGE_GLOBALS kernel config option.
Reviewed by: bz
Approved by: julian (mentor)
2009-04-11 05:58:58 +00:00
|
|
|
static void vnet_inet_register(void);
|
|
|
|
|
|
|
|
static const vnet_modinfo_t vnet_inet_modinfo = {
|
|
|
|
.vmi_id = VNET_MOD_INET,
|
|
|
|
.vmi_name = "inet",
|
|
|
|
};
|
|
|
|
|
|
|
|
static void vnet_inet_register()
|
|
|
|
{
|
|
|
|
|
|
|
|
vnet_mod_register(&vnet_inet_modinfo);
|
|
|
|
}
|
|
|
|
|
|
|
|
SYSINIT(inet, SI_SUB_PROTO_BEGIN, SI_ORDER_FIRST, vnet_inet_register, 0);
|
|
|
|
#endif
|
|
|
|
|
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 int
|
|
|
|
sysctl_netinet_intr_queue_maxlen(SYSCTL_HANDLER_ARGS)
|
|
|
|
{
|
|
|
|
int error, qlimit;
|
|
|
|
|
|
|
|
netisr_getqlimit(&ip_nh, &qlimit);
|
|
|
|
error = sysctl_handle_int(oidp, &qlimit, 0, req);
|
|
|
|
if (error || !req->newptr)
|
|
|
|
return (error);
|
|
|
|
if (qlimit < 1)
|
|
|
|
return (EINVAL);
|
|
|
|
return (netisr_setqlimit(&ip_nh, qlimit));
|
|
|
|
}
|
|
|
|
SYSCTL_PROC(_net_inet_ip, IPCTL_INTRQMAXLEN, intr_queue_maxlen,
|
|
|
|
CTLTYPE_INT|CTLFLAG_RW, 0, 0, sysctl_netinet_intr_queue_maxlen, "I",
|
|
|
|
"Maximum size of the IP input queue");
|
|
|
|
|
|
|
|
static int
|
|
|
|
sysctl_netinet_intr_queue_drops(SYSCTL_HANDLER_ARGS)
|
|
|
|
{
|
|
|
|
u_int64_t qdrops_long;
|
|
|
|
int error, qdrops;
|
|
|
|
|
|
|
|
netisr_getqdrops(&ip_nh, &qdrops_long);
|
|
|
|
qdrops = qdrops_long;
|
|
|
|
error = sysctl_handle_int(oidp, &qdrops, 0, req);
|
|
|
|
if (error || !req->newptr)
|
|
|
|
return (error);
|
|
|
|
if (qdrops != 0)
|
|
|
|
return (EINVAL);
|
|
|
|
netisr_clearqdrops(&ip_nh);
|
|
|
|
return (0);
|
|
|
|
}
|
|
|
|
|
|
|
|
SYSCTL_PROC(_net_inet_ip, IPCTL_INTRQDROPS, intr_queue_drops,
|
|
|
|
CTLTYPE_INT|CTLFLAG_RD, 0, 0, sysctl_netinet_intr_queue_drops, "I",
|
|
|
|
"Number of packets dropped from the IP input queue");
|
|
|
|
|
1994-05-24 10:09:53 +00:00
|
|
|
/*
|
|
|
|
* IP initialization: fill in IP protocol switch table.
|
|
|
|
* All protocols not implemented in kernel go to raw IP protocol handler.
|
|
|
|
*/
|
|
|
|
void
|
2007-05-10 15:58:48 +00:00
|
|
|
ip_init(void)
|
1994-05-24 10:09:53 +00:00
|
|
|
{
|
2007-05-10 15:58:48 +00:00
|
|
|
struct protosw *pr;
|
|
|
|
int i;
|
1994-05-24 10:09:53 +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
|
|
|
V_ip_id = time_second & 0xffff;
|
2008-11-19 09:39:34 +00:00
|
|
|
|
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
|
|
|
TAILQ_INIT(&V_in_ifaddrhead);
|
|
|
|
V_in_ifaddrhashtbl = hashinit(INADDR_NHASH, M_IFADDR, &V_in_ifaddrhmask);
|
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
|
|
|
|
|
|
|
/* Initialize IP reassembly queue. */
|
|
|
|
for (i = 0; i < IPREASS_NHASH; i++)
|
|
|
|
TAILQ_INIT(&V_ipq[i]);
|
|
|
|
V_maxnipq = nmbclusters / 32;
|
|
|
|
V_maxfragsperpacket = 16;
|
|
|
|
V_ipq_zone = uma_zcreate("ipq", sizeof(struct ipq), NULL, NULL, NULL,
|
|
|
|
NULL, UMA_ALIGN_PTR, 0);
|
|
|
|
maxnipq_update();
|
|
|
|
|
2009-06-22 21:19:24 +00:00
|
|
|
#ifdef FLOWTABLE
|
|
|
|
TUNABLE_INT_FETCH("net.inet.ip.output_flowtable_size",
|
|
|
|
&V_ip_output_flowtable_size);
|
|
|
|
V_ip_ft = flowtable_alloc(V_ip_output_flowtable_size, FL_PCPU);
|
|
|
|
#endif
|
|
|
|
|
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
|
|
|
/* Skip initialization of globals for non-default instances. */
|
|
|
|
if (!IS_DEFAULT_VNET(curvnet))
|
|
|
|
return;
|
|
|
|
|
2001-09-03 20:03:55 +00:00
|
|
|
pr = pffindproto(PF_INET, IPPROTO_RAW, SOCK_RAW);
|
2005-01-30 19:29:47 +00:00
|
|
|
if (pr == NULL)
|
2004-09-16 18:33:39 +00:00
|
|
|
panic("ip_init: PF_INET not found");
|
|
|
|
|
|
|
|
/* Initialize the entire ip_protox[] array to IPPROTO_RAW. */
|
1994-05-24 10:09:53 +00:00
|
|
|
for (i = 0; i < IPPROTO_MAX; i++)
|
|
|
|
ip_protox[i] = pr - inetsw;
|
2004-09-16 18:33:39 +00:00
|
|
|
/*
|
|
|
|
* Cycle through IP protocols and put them into the appropriate place
|
|
|
|
* in ip_protox[].
|
|
|
|
*/
|
2001-09-03 20:03:55 +00:00
|
|
|
for (pr = inetdomain.dom_protosw;
|
|
|
|
pr < inetdomain.dom_protoswNPROTOSW; pr++)
|
1994-05-24 10:09:53 +00:00
|
|
|
if (pr->pr_domain->dom_family == PF_INET &&
|
2004-09-16 18:33:39 +00:00
|
|
|
pr->pr_protocol && pr->pr_protocol != IPPROTO_RAW) {
|
|
|
|
/* Be careful to only index valid IP protocols. */
|
2005-02-23 00:38:12 +00:00
|
|
|
if (pr->pr_protocol < IPPROTO_MAX)
|
2004-09-16 18:33:39 +00:00
|
|
|
ip_protox[pr->pr_protocol] = pr - inetsw;
|
|
|
|
}
|
1997-09-15 23:07:01 +00:00
|
|
|
|
2004-08-27 15:16:24 +00:00
|
|
|
/* Initialize packet filter hooks. */
|
2003-09-23 17:54:04 +00:00
|
|
|
inet_pfil_hook.ph_type = PFIL_TYPE_AF;
|
|
|
|
inet_pfil_hook.ph_af = AF_INET;
|
|
|
|
if ((i = pfil_head_register(&inet_pfil_hook)) != 0)
|
|
|
|
printf("%s: WARNING: unable to register pfil hook, "
|
|
|
|
"error %d\n", __func__, i);
|
|
|
|
|
2005-01-02 01:50:57 +00:00
|
|
|
/* Start ipport_tick. */
|
|
|
|
callout_init(&ipport_tick_callout, CALLOUT_MPSAFE);
|
Change the curvnet variable from a global const struct vnet *,
previously always pointing to the default vnet context, to a
dynamically changing thread-local one. The currvnet context
should be set on entry to networking code via CURVNET_SET() macros,
and reverted to previous state via CURVNET_RESTORE(). Recursions
on curvnet are permitted, though strongly discuouraged.
This change should have no functional impact on nooptions VIMAGE
kernel builds, where CURVNET_* macros expand to whitespace.
The curthread->td_vnet (aka curvnet) variable's purpose is to be an
indicator of the vnet context in which the current network-related
operation takes place, in case we cannot deduce the current vnet
context from any other source, such as by looking at mbuf's
m->m_pkthdr.rcvif->if_vnet, sockets's so->so_vnet etc. Moreover, so
far curvnet has turned out to be an invaluable consistency checking
aid: it helps to catch cases when sockets, ifnets or any other
vnet-aware structures may have leaked from one vnet to another.
The exact placement of the CURVNET_SET() / CURVNET_RESTORE() macros
was a result of an empirical iterative process, whith an aim to
reduce recursions on CURVNET_SET() to a minimum, while still reducing
the scope of CURVNET_SET() to networking only operations - the
alternative would be calling CURVNET_SET() on each system call entry.
In general, curvnet has to be set in three typicall cases: when
processing socket-related requests from userspace or from within the
kernel; when processing inbound traffic flowing from device drivers
to upper layers of the networking stack, and when executing
timer-driven networking functions.
This change also introduces a DDB subcommand to show the list of all
vnet instances.
Approved by: julian (mentor)
2009-05-05 10:56:12 +00:00
|
|
|
callout_reset(&ipport_tick_callout, 1, ipport_tick, NULL);
|
2005-01-02 01:50:57 +00:00
|
|
|
EVENTHANDLER_REGISTER(shutdown_pre_sync, ip_fini, NULL,
|
|
|
|
SHUTDOWN_PRI_DEFAULT);
|
2006-04-21 09:25:40 +00:00
|
|
|
EVENTHANDLER_REGISTER(nmbclusters_change, ipq_zone_change,
|
|
|
|
NULL, EVENTHANDLER_PRI_ANY);
|
2005-01-02 01:50:57 +00:00
|
|
|
|
2004-09-16 18:33:39 +00:00
|
|
|
/* Initialize various other remaining things. */
|
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
|
|
|
IPQ_LOCK_INIT();
|
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
|
|
|
netisr_register(&ip_nh);
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
|
|
|
|
2007-05-10 15:58:48 +00:00
|
|
|
void
|
|
|
|
ip_fini(void *xtp)
|
2005-01-02 01:50:57 +00:00
|
|
|
{
|
2007-05-10 15:58:48 +00:00
|
|
|
|
2005-01-02 01:50:57 +00:00
|
|
|
callout_stop(&ipport_tick_callout);
|
|
|
|
}
|
|
|
|
|
1994-05-24 10:09:53 +00:00
|
|
|
/*
|
|
|
|
* Ip input routine. Checksum and byte swap header. If fragmented
|
|
|
|
* try to reassemble. Process options. Pass to next level.
|
|
|
|
*/
|
1996-02-05 20:36:02 +00:00
|
|
|
void
|
|
|
|
ip_input(struct mbuf *m)
|
1994-05-24 10:09:53 +00:00
|
|
|
{
|
2003-11-14 21:02:22 +00:00
|
|
|
struct ip *ip = NULL;
|
2000-10-19 23:15:54 +00:00
|
|
|
struct in_ifaddr *ia = NULL;
|
2001-09-29 04:34:11 +00:00
|
|
|
struct ifaddr *ifa;
|
2009-04-20 14:35:42 +00:00
|
|
|
struct ifnet *ifp;
|
2004-08-17 22:05:54 +00:00
|
|
|
int checkif, hlen = 0;
|
1997-02-06 11:14:22 +00:00
|
|
|
u_short sum;
|
2003-11-14 21:48:57 +00:00
|
|
|
int dchg = 0; /* dest changed after fw */
|
2003-10-16 16:25:25 +00:00
|
|
|
struct in_addr odst; /* original dst address */
|
1999-12-06 00:43:07 +00:00
|
|
|
|
2007-05-16 20:46:58 +00:00
|
|
|
M_ASSERTPKTHDR(m);
|
2006-02-01 13:55:03 +00:00
|
|
|
|
2004-02-25 19:55:29 +00:00
|
|
|
if (m->m_flags & M_FASTFWD_OURS) {
|
2004-08-17 22:05:54 +00:00
|
|
|
/*
|
2004-09-15 20:17:03 +00:00
|
|
|
* Firewall or NAT changed destination to local.
|
|
|
|
* We expect ip_len and ip_off to be in host byte order.
|
2004-08-17 22:05:54 +00:00
|
|
|
*/
|
2004-09-15 20:17:03 +00:00
|
|
|
m->m_flags &= ~M_FASTFWD_OURS;
|
|
|
|
/* Set up some basics that will be used later. */
|
2004-08-17 22:05:54 +00:00
|
|
|
ip = mtod(m, struct ip *);
|
|
|
|
hlen = ip->ip_hl << 2;
|
2007-05-16 20:46:58 +00:00
|
|
|
goto ours;
|
|
|
|
}
|
2004-02-18 00:04:52 +00:00
|
|
|
|
2009-04-11 23:35:20 +00:00
|
|
|
IPSTAT_INC(ips_total);
|
1996-10-07 19:21:46 +00:00
|
|
|
|
|
|
|
if (m->m_pkthdr.len < sizeof(struct ip))
|
|
|
|
goto tooshort;
|
|
|
|
|
1994-05-24 10:09:53 +00:00
|
|
|
if (m->m_len < sizeof (struct ip) &&
|
2004-08-11 10:46:15 +00:00
|
|
|
(m = m_pullup(m, sizeof (struct ip))) == NULL) {
|
2009-04-11 23:35:20 +00:00
|
|
|
IPSTAT_INC(ips_toosmall);
|
1996-02-05 20:36:02 +00:00
|
|
|
return;
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
|
|
|
ip = mtod(m, struct ip *);
|
1996-10-07 19:21:46 +00:00
|
|
|
|
2002-10-20 22:52:07 +00:00
|
|
|
if (ip->ip_v != IPVERSION) {
|
2009-04-11 23:35:20 +00:00
|
|
|
IPSTAT_INC(ips_badvers);
|
1994-05-24 10:09:53 +00:00
|
|
|
goto bad;
|
|
|
|
}
|
1996-10-07 19:21:46 +00:00
|
|
|
|
2002-10-20 22:52:07 +00:00
|
|
|
hlen = ip->ip_hl << 2;
|
1994-05-24 10:09:53 +00:00
|
|
|
if (hlen < sizeof(struct ip)) { /* minimum header length */
|
2009-04-11 23:35:20 +00:00
|
|
|
IPSTAT_INC(ips_badhlen);
|
1994-05-24 10:09:53 +00:00
|
|
|
goto bad;
|
|
|
|
}
|
|
|
|
if (hlen > m->m_len) {
|
2004-08-11 10:46:15 +00:00
|
|
|
if ((m = m_pullup(m, hlen)) == NULL) {
|
2009-04-11 23:35:20 +00:00
|
|
|
IPSTAT_INC(ips_badhlen);
|
1996-02-05 20:36:02 +00:00
|
|
|
return;
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
|
|
|
ip = mtod(m, struct ip *);
|
|
|
|
}
|
2001-06-11 12:39:29 +00:00
|
|
|
|
|
|
|
/* 127/8 must not appear on wire - RFC1122 */
|
2009-04-20 14:35:42 +00:00
|
|
|
ifp = m->m_pkthdr.rcvif;
|
2001-06-11 12:39:29 +00:00
|
|
|
if ((ntohl(ip->ip_dst.s_addr) >> IN_CLASSA_NSHIFT) == IN_LOOPBACKNET ||
|
|
|
|
(ntohl(ip->ip_src.s_addr) >> IN_CLASSA_NSHIFT) == IN_LOOPBACKNET) {
|
2009-04-20 14:35:42 +00:00
|
|
|
if ((ifp->if_flags & IFF_LOOPBACK) == 0) {
|
2009-04-11 23:35:20 +00:00
|
|
|
IPSTAT_INC(ips_badaddr);
|
2001-06-11 12:39:29 +00:00
|
|
|
goto bad;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2000-03-27 19:14:27 +00:00
|
|
|
if (m->m_pkthdr.csum_flags & CSUM_IP_CHECKED) {
|
|
|
|
sum = !(m->m_pkthdr.csum_flags & CSUM_IP_VALID);
|
1996-10-07 19:21:46 +00:00
|
|
|
} else {
|
2000-03-27 19:14:27 +00:00
|
|
|
if (hlen == sizeof(struct ip)) {
|
|
|
|
sum = in_cksum_hdr(ip);
|
|
|
|
} else {
|
|
|
|
sum = in_cksum(m, hlen);
|
|
|
|
}
|
1996-10-07 19:21:46 +00:00
|
|
|
}
|
1997-02-06 11:14:22 +00:00
|
|
|
if (sum) {
|
2009-04-11 23:35:20 +00:00
|
|
|
IPSTAT_INC(ips_badsum);
|
1994-05-24 10:09:53 +00:00
|
|
|
goto bad;
|
|
|
|
}
|
|
|
|
|
2004-06-13 17:29:10 +00:00
|
|
|
#ifdef ALTQ
|
|
|
|
if (altq_input != NULL && (*altq_input)(m, AF_INET) == 0)
|
|
|
|
/* packet is dropped by traffic conditioner */
|
|
|
|
return;
|
|
|
|
#endif
|
|
|
|
|
1994-05-24 10:09:53 +00:00
|
|
|
/*
|
|
|
|
* Convert fields to host representation.
|
|
|
|
*/
|
2002-02-18 20:35:27 +00:00
|
|
|
ip->ip_len = ntohs(ip->ip_len);
|
1994-05-24 10:09:53 +00:00
|
|
|
if (ip->ip_len < hlen) {
|
2009-04-11 23:35:20 +00:00
|
|
|
IPSTAT_INC(ips_badlen);
|
1994-05-24 10:09:53 +00:00
|
|
|
goto bad;
|
|
|
|
}
|
2002-02-18 20:35:27 +00:00
|
|
|
ip->ip_off = ntohs(ip->ip_off);
|
1994-05-24 10:09:53 +00:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Check that the amount of data in the buffers
|
|
|
|
* is as at least much as the IP header would have us expect.
|
|
|
|
* Trim mbufs if longer than we expect.
|
|
|
|
* Drop packet if shorter than we expect.
|
|
|
|
*/
|
|
|
|
if (m->m_pkthdr.len < ip->ip_len) {
|
1996-10-07 19:21:46 +00:00
|
|
|
tooshort:
|
2009-04-11 23:35:20 +00:00
|
|
|
IPSTAT_INC(ips_tooshort);
|
1994-05-24 10:09:53 +00:00
|
|
|
goto bad;
|
|
|
|
}
|
|
|
|
if (m->m_pkthdr.len > ip->ip_len) {
|
|
|
|
if (m->m_len == m->m_pkthdr.len) {
|
|
|
|
m->m_len = ip->ip_len;
|
|
|
|
m->m_pkthdr.len = ip->ip_len;
|
|
|
|
} else
|
|
|
|
m_adj(m, ip->ip_len - m->m_pkthdr.len);
|
|
|
|
}
|
2007-07-03 12:13:45 +00:00
|
|
|
#ifdef IPSEC
|
2003-07-22 18:58:34 +00:00
|
|
|
/*
|
|
|
|
* Bypass packet filtering for packets from a tunnel (gif).
|
|
|
|
*/
|
2007-08-05 16:16:15 +00:00
|
|
|
if (ip_ipsec_filtertunnel(m))
|
2004-08-27 15:16:24 +00:00
|
|
|
goto passin;
|
2007-07-03 12:13:45 +00:00
|
|
|
#endif /* IPSEC */
|
2001-03-05 08:45:05 +00:00
|
|
|
|
1997-04-03 10:47:12 +00:00
|
|
|
/*
|
2003-09-23 17:54:04 +00:00
|
|
|
* Run through list of hooks for input packets.
|
2003-10-16 16:25:25 +00:00
|
|
|
*
|
|
|
|
* NB: Beware of the destination address changing (e.g.
|
|
|
|
* by NAT rewriting). When this happens, tell
|
|
|
|
* ip_forward to do the right thing.
|
1997-04-03 10:47:12 +00:00
|
|
|
*/
|
2004-08-27 15:16:24 +00:00
|
|
|
|
|
|
|
/* Jump over all PFIL processing if hooks are not active. */
|
2006-02-02 03:13:16 +00:00
|
|
|
if (!PFIL_HOOKED(&inet_pfil_hook))
|
2004-08-27 15:16:24 +00:00
|
|
|
goto passin;
|
|
|
|
|
2003-10-16 16:25:25 +00:00
|
|
|
odst = ip->ip_dst;
|
2009-04-20 14:35:42 +00:00
|
|
|
if (pfil_run_hooks(&inet_pfil_hook, &m, ifp, PFIL_IN, NULL) != 0)
|
2003-09-23 17:54:04 +00:00
|
|
|
return;
|
|
|
|
if (m == NULL) /* consumed by filter */
|
|
|
|
return;
|
2004-08-17 22:05:54 +00:00
|
|
|
|
2003-09-23 17:54:04 +00:00
|
|
|
ip = mtod(m, struct ip *);
|
2003-11-14 21:48:57 +00:00
|
|
|
dchg = (odst.s_addr != ip->ip_dst.s_addr);
|
2009-04-20 14:35:42 +00:00
|
|
|
ifp = m->m_pkthdr.rcvif;
|
1995-01-12 13:06:32 +00:00
|
|
|
|
2004-08-17 22:05:54 +00:00
|
|
|
#ifdef IPFIREWALL_FORWARD
|
|
|
|
if (m->m_flags & M_FASTFWD_OURS) {
|
|
|
|
m->m_flags &= ~M_FASTFWD_OURS;
|
|
|
|
goto ours;
|
|
|
|
}
|
2005-02-22 17:40:40 +00:00
|
|
|
if ((dchg = (m_tag_find(m, PACKET_TAG_IPFORWARD, NULL) != NULL)) != 0) {
|
|
|
|
/*
|
|
|
|
* Directly ship on the packet. This allows to forward packets
|
|
|
|
* that were destined for us to some other directly connected
|
|
|
|
* host.
|
|
|
|
*/
|
|
|
|
ip_forward(m, dchg);
|
|
|
|
return;
|
|
|
|
}
|
2004-08-17 22:05:54 +00:00
|
|
|
#endif /* IPFIREWALL_FORWARD */
|
Remove (almost all) global variables that were used to hold
packet forwarding state ("annotations") during ip processing.
The code is considerably cleaner now.
The variables removed by this change are:
ip_divert_cookie used by divert sockets
ip_fw_fwd_addr used for transparent ip redirection
last_pkt used by dynamic pipes in dummynet
Removal of the first two has been done by carrying the annotations
into volatile structs prepended to the mbuf chains, and adding
appropriate code to add/remove annotations in the routines which
make use of them, i.e. ip_input(), ip_output(), tcp_input(),
bdg_forward(), ether_demux(), ether_output_frame(), div_output().
On passing, remove a bug in divert handling of fragmented packet.
Now it is the fragment at offset 0 which sets the divert status of
the whole packet, whereas formerly it was the last incoming fragment
to decide.
Removal of last_pkt required a change in the interface of ip_fw_chk()
and dummynet_io(). On passing, use the same mechanism for dummynet
annotations and for divert/forward annotations.
option IPFIREWALL_FORWARD is effectively useless, the code to
implement it is very small and is now in by default to avoid the
obfuscation of conditionally compiled code.
NOTES:
* there is at least one global variable left, sro_fwd, in ip_output().
I am not sure if/how this can be removed.
* I have deliberately avoided gratuitous style changes in this commit
to avoid cluttering the diffs. Minor stule cleanup will likely be
necessary
* this commit only focused on the IP layer. I am sure there is a
number of global variables used in the TCP and maybe UDP stack.
* despite the number of files touched, there are absolutely no API's
or data structures changed by this commit (except the interfaces of
ip_fw_chk() and dummynet_io(), which are internal anyways), so
an MFC is quite safe and unintrusive (and desirable, given the
improved readability of the code).
MFC after: 10 days
2002-06-22 11:51:02 +00:00
|
|
|
|
2004-08-27 15:16:24 +00:00
|
|
|
passin:
|
1994-05-24 10:09:53 +00:00
|
|
|
/*
|
|
|
|
* Process options and, if not destined for us,
|
|
|
|
* ship it on. ip_dooptions returns 1 when an
|
|
|
|
* error was detected (causing an icmp message
|
|
|
|
* to be sent and the original packet to be freed).
|
|
|
|
*/
|
2004-08-17 22:05:54 +00:00
|
|
|
if (hlen > sizeof (struct ip) && ip_dooptions(m, 0))
|
1996-02-05 20:36:02 +00:00
|
|
|
return;
|
1994-05-24 10:09:53 +00:00
|
|
|
|
1994-09-06 22:42:31 +00:00
|
|
|
/* greedy RSVP, snatches any PATH packet of the RSVP protocol and no
|
1995-06-13 17:51:16 +00:00
|
|
|
* matter if it is destined to another node, or whether it is
|
1994-09-06 22:42:31 +00:00
|
|
|
* a multicast one, RSVP wants it! and prevents it from being forwarded
|
|
|
|
* anywhere else. Also checks if the rsvp daemon is running before
|
|
|
|
* grabbing the packet.
|
|
|
|
*/
|
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 (V_rsvp_on && ip->ip_p==IPPROTO_RSVP)
|
1994-09-06 22:42:31 +00:00
|
|
|
goto ours;
|
|
|
|
|
1994-05-24 10:09:53 +00:00
|
|
|
/*
|
|
|
|
* Check our list of addresses, to see if the packet is for us.
|
1999-02-09 16:55:46 +00:00
|
|
|
* If we don't have any addresses, assume any unicast packet
|
|
|
|
* we receive might be for us (and let the upper layers deal
|
|
|
|
* with it).
|
1994-05-24 10:09:53 +00:00
|
|
|
*/
|
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 (TAILQ_EMPTY(&V_in_ifaddrhead) &&
|
1999-02-09 16:55:46 +00:00
|
|
|
(m->m_flags & (M_MCAST|M_BCAST)) == 0)
|
|
|
|
goto ours;
|
|
|
|
|
2001-03-04 01:39:19 +00:00
|
|
|
/*
|
|
|
|
* Enable a consistency check between the destination address
|
|
|
|
* and the arrival interface for a unicast packet (the RFC 1122
|
|
|
|
* strong ES model) if IP forwarding is disabled and the packet
|
2001-03-04 03:22:36 +00:00
|
|
|
* is not locally generated and the packet is not subject to
|
|
|
|
* 'ipfw fwd'.
|
2001-03-05 08:45:05 +00:00
|
|
|
*
|
2001-11-04 22:56:25 +00:00
|
|
|
* XXX - Checking also should be disabled if the destination
|
2001-03-05 08:45:05 +00:00
|
|
|
* address is ipnat'ed to a different interface.
|
|
|
|
*
|
2001-03-05 22:40:27 +00:00
|
|
|
* XXX - Checking is incompatible with IP aliases added
|
2001-03-05 08:45:05 +00:00
|
|
|
* to the loopback interface instead of the interface where
|
|
|
|
* the packets are received.
|
2005-02-22 13:04:05 +00:00
|
|
|
*
|
|
|
|
* XXX - This is the case for carp vhost IPs as well so we
|
|
|
|
* insert a workaround. If the packet got here, we already
|
|
|
|
* checked with carp_iamatch() and carp_forus().
|
2001-03-04 01:39:19 +00:00
|
|
|
*/
|
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
|
|
|
checkif = V_ip_checkinterface && (V_ipforwarding == 0) &&
|
2009-04-20 14:35:42 +00:00
|
|
|
ifp != NULL && ((ifp->if_flags & IFF_LOOPBACK) == 0) &&
|
2005-02-22 13:04:05 +00:00
|
|
|
#ifdef DEV_CARP
|
2009-04-20 14:35:42 +00:00
|
|
|
!ifp->if_carp &&
|
2005-02-22 13:04:05 +00:00
|
|
|
#endif
|
2004-08-17 22:05:54 +00:00
|
|
|
(dchg == 0);
|
2001-03-04 01:39:19 +00:00
|
|
|
|
2001-09-29 04:34:11 +00:00
|
|
|
/*
|
|
|
|
* Check for exact addresses in the hash bucket.
|
|
|
|
*/
|
2009-06-25 11:52:33 +00:00
|
|
|
/* IN_IFADDR_RLOCK(); */
|
2004-08-17 22:05:54 +00:00
|
|
|
LIST_FOREACH(ia, INADDR_HASH(ip->ip_dst.s_addr), ia_hash) {
|
1998-07-06 03:20:19 +00:00
|
|
|
/*
|
2001-03-04 01:39:19 +00:00
|
|
|
* If the address matches, verify that the packet
|
|
|
|
* arrived via the correct interface if checking is
|
|
|
|
* enabled.
|
1998-07-06 03:20:19 +00:00
|
|
|
*/
|
2004-08-17 22:05:54 +00:00
|
|
|
if (IA_SIN(ia)->sin_addr.s_addr == ip->ip_dst.s_addr &&
|
2009-06-23 20:19:09 +00:00
|
|
|
(!checkif || ia->ia_ifp == ifp)) {
|
|
|
|
ifa_ref(&ia->ia_ifa);
|
2009-06-25 11:52:33 +00:00
|
|
|
/* IN_IFADDR_RUNLOCK(); */
|
1999-03-12 01:15:57 +00:00
|
|
|
goto ours;
|
2009-06-23 20:19:09 +00:00
|
|
|
}
|
2001-09-29 04:34:11 +00:00
|
|
|
}
|
2009-06-25 11:52:33 +00:00
|
|
|
/* IN_IFADDR_RUNLOCK(); */
|
|
|
|
|
2001-09-29 04:34:11 +00:00
|
|
|
/*
|
|
|
|
* Check for broadcast addresses.
|
|
|
|
*
|
|
|
|
* Only accept broadcast packets that arrive via the matching
|
|
|
|
* interface. Reception of forwarded directed broadcasts would
|
|
|
|
* be handled via ip_forward() and ether_output() with the loopback
|
|
|
|
* into the stack for SIMPLEX interfaces handled by ether_output().
|
|
|
|
*/
|
2009-04-20 14:35:42 +00:00
|
|
|
if (ifp != NULL && ifp->if_flags & IFF_BROADCAST) {
|
|
|
|
IF_ADDR_LOCK(ifp);
|
|
|
|
TAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) {
|
2001-09-29 04:34:11 +00:00
|
|
|
if (ifa->ifa_addr->sa_family != AF_INET)
|
|
|
|
continue;
|
|
|
|
ia = ifatoia(ifa);
|
1994-05-24 10:09:53 +00:00
|
|
|
if (satosin(&ia->ia_broadaddr)->sin_addr.s_addr ==
|
2009-04-20 14:35:42 +00:00
|
|
|
ip->ip_dst.s_addr) {
|
2009-06-23 20:19:09 +00:00
|
|
|
ifa_ref(ifa);
|
2009-04-20 14:35:42 +00:00
|
|
|
IF_ADDR_UNLOCK(ifp);
|
1994-05-24 10:09:53 +00:00
|
|
|
goto ours;
|
2009-04-20 14:35:42 +00:00
|
|
|
}
|
|
|
|
if (ia->ia_netbroadcast.s_addr == ip->ip_dst.s_addr) {
|
2009-06-23 20:19:09 +00:00
|
|
|
ifa_ref(ifa);
|
2009-04-20 14:35:42 +00:00
|
|
|
IF_ADDR_UNLOCK(ifp);
|
1994-05-24 10:09:53 +00:00
|
|
|
goto ours;
|
2009-04-20 14:35:42 +00:00
|
|
|
}
|
2004-07-08 22:35:36 +00:00
|
|
|
#ifdef BOOTP_COMPAT
|
2009-04-20 14:35:42 +00:00
|
|
|
if (IA_SIN(ia)->sin_addr.s_addr == INADDR_ANY) {
|
2009-06-23 20:19:09 +00:00
|
|
|
ifa_ref(ifa);
|
2009-04-20 14:35:42 +00:00
|
|
|
IF_ADDR_UNLOCK(ifp);
|
2004-07-08 22:35:36 +00:00
|
|
|
goto ours;
|
2009-04-20 14:35:42 +00:00
|
|
|
}
|
2004-07-08 22:35:36 +00:00
|
|
|
#endif
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
2009-04-20 14:35:42 +00:00
|
|
|
IF_ADDR_UNLOCK(ifp);
|
2009-06-24 14:29:40 +00:00
|
|
|
ia = NULL;
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
2007-02-03 06:45:51 +00:00
|
|
|
/* RFC 3927 2.7: Do not forward datagrams for 169.254.0.0/16. */
|
|
|
|
if (IN_LINKLOCAL(ntohl(ip->ip_dst.s_addr))) {
|
2009-04-11 23:35:20 +00:00
|
|
|
IPSTAT_INC(ips_cantforward);
|
2007-02-03 06:45:51 +00:00
|
|
|
m_freem(m);
|
|
|
|
return;
|
|
|
|
}
|
1994-05-24 10:09:53 +00:00
|
|
|
if (IN_MULTICAST(ntohl(ip->ip_dst.s_addr))) {
|
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 (V_ip_mrouter) {
|
1994-05-24 10:09:53 +00:00
|
|
|
/*
|
|
|
|
* If we are acting as a multicast router, all
|
|
|
|
* incoming multicast packets are passed to the
|
|
|
|
* kernel-level multicast forwarding function.
|
|
|
|
* The packet is returned (relatively) intact; if
|
|
|
|
* ip_mforward() returns a non-zero value, the packet
|
|
|
|
* must be discarded, else it may be accepted below.
|
|
|
|
*/
|
2009-04-20 14:35:42 +00:00
|
|
|
if (ip_mforward && ip_mforward(ip, ifp, m, 0) != 0) {
|
2009-04-11 23:35:20 +00:00
|
|
|
IPSTAT_INC(ips_cantforward);
|
1994-05-24 10:09:53 +00:00
|
|
|
m_freem(m);
|
1996-02-05 20:36:02 +00:00
|
|
|
return;
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
2002-05-12 00:22:38 +00:00
|
|
|
* The process-level routing daemon needs to receive
|
1994-05-24 10:09:53 +00:00
|
|
|
* all multicast IGMP packets, whether or not this
|
|
|
|
* host belongs to their destination groups.
|
|
|
|
*/
|
|
|
|
if (ip->ip_p == IPPROTO_IGMP)
|
|
|
|
goto ours;
|
2009-04-11 23:35:20 +00:00
|
|
|
IPSTAT_INC(ips_forward);
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
|
|
|
/*
|
2009-03-09 17:53:05 +00:00
|
|
|
* Assume the packet is for us, to avoid prematurely taking
|
|
|
|
* a lock on the in_multi hash. Protocols must perform
|
|
|
|
* their own filtering and update statistics accordingly.
|
1994-05-24 10:09:53 +00:00
|
|
|
*/
|
|
|
|
goto ours;
|
|
|
|
}
|
|
|
|
if (ip->ip_dst.s_addr == (u_long)INADDR_BROADCAST)
|
|
|
|
goto ours;
|
|
|
|
if (ip->ip_dst.s_addr == INADDR_ANY)
|
|
|
|
goto ours;
|
|
|
|
|
1999-12-22 19:13:38 +00:00
|
|
|
/*
|
|
|
|
* FAITH(Firewall Aided Internet Translator)
|
|
|
|
*/
|
2009-04-20 14:35:42 +00:00
|
|
|
if (ifp && ifp->if_type == IFT_FAITH) {
|
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 (V_ip_keepfaith) {
|
1999-12-22 19:13:38 +00:00
|
|
|
if (ip->ip_p == IPPROTO_TCP || ip->ip_p == IPPROTO_ICMP)
|
|
|
|
goto ours;
|
|
|
|
}
|
|
|
|
m_freem(m);
|
|
|
|
return;
|
|
|
|
}
|
2001-09-25 18:40:52 +00:00
|
|
|
|
1994-05-24 10:09:53 +00:00
|
|
|
/*
|
|
|
|
* Not for us; forward if possible and desirable.
|
|
|
|
*/
|
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 (V_ipforwarding == 0) {
|
2009-04-11 23:35:20 +00:00
|
|
|
IPSTAT_INC(ips_cantforward);
|
2001-02-21 16:59:47 +00:00
|
|
|
m_freem(m);
|
2002-02-26 02:11:13 +00:00
|
|
|
} else {
|
2007-07-03 12:13:45 +00:00
|
|
|
#ifdef IPSEC
|
2006-02-01 13:55:03 +00:00
|
|
|
if (ip_ipsec_fwd(m))
|
2002-02-26 02:11:13 +00:00
|
|
|
goto bad;
|
2007-07-03 12:13:45 +00:00
|
|
|
#endif /* IPSEC */
|
2004-08-17 22:05:54 +00:00
|
|
|
ip_forward(m, dchg);
|
2002-02-26 02:11:13 +00:00
|
|
|
}
|
1996-02-05 20:36:02 +00:00
|
|
|
return;
|
1994-05-24 10:09:53 +00:00
|
|
|
|
|
|
|
ours:
|
2001-12-29 09:24:18 +00:00
|
|
|
#ifdef IPSTEALTH
|
|
|
|
/*
|
|
|
|
* IPSTEALTH: Process non-routing options only
|
|
|
|
* if the packet is destined for us.
|
|
|
|
*/
|
2009-06-24 14:29:40 +00:00
|
|
|
if (V_ipstealth && hlen > sizeof (struct ip) && ip_dooptions(m, 1)) {
|
|
|
|
if (ia != NULL)
|
|
|
|
ifa_free(&ia->ia_ifa);
|
2001-12-29 09:24:18 +00:00
|
|
|
return;
|
2009-06-24 14:29:40 +00:00
|
|
|
}
|
2001-12-29 09:24:18 +00:00
|
|
|
#endif /* IPSTEALTH */
|
|
|
|
|
2000-10-19 23:15:54 +00:00
|
|
|
/* Count the packet in the ip address stats */
|
|
|
|
if (ia != NULL) {
|
|
|
|
ia->ia_ifa.if_ipackets++;
|
|
|
|
ia->ia_ifa.if_ibytes += m->m_pkthdr.len;
|
2009-06-23 20:19:09 +00:00
|
|
|
ifa_free(&ia->ia_ifa);
|
2000-10-19 23:15:54 +00:00
|
|
|
}
|
1994-10-28 15:09:49 +00:00
|
|
|
|
1994-05-24 10:09:53 +00:00
|
|
|
/*
|
2004-08-03 12:31:38 +00:00
|
|
|
* Attempt reassembly; if it succeeds, proceed.
|
|
|
|
* ip_reass() will return a different mbuf.
|
1994-05-24 10:09:53 +00:00
|
|
|
*/
|
2000-10-26 13:14:48 +00:00
|
|
|
if (ip->ip_off & (IP_MF | IP_OFFMASK)) {
|
2004-08-03 12:31:38 +00:00
|
|
|
m = ip_reass(m);
|
|
|
|
if (m == NULL)
|
2000-10-26 13:14:48 +00:00
|
|
|
return;
|
|
|
|
ip = mtod(m, struct ip *);
|
|
|
|
/* Get the header length of the reassembled packet */
|
2002-10-20 22:52:07 +00:00
|
|
|
hlen = ip->ip_hl << 2;
|
2004-08-03 12:31:38 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Further protocols expect the packet length to be w/o the
|
|
|
|
* IP header.
|
|
|
|
*/
|
|
|
|
ip->ip_len -= hlen;
|
1994-05-24 10:09:53 +00:00
|
|
|
|
2007-07-03 12:13:45 +00:00
|
|
|
#ifdef IPSEC
|
2001-06-11 12:39:29 +00:00
|
|
|
/*
|
|
|
|
* enforce IPsec policy checking if we are seeing last header.
|
|
|
|
* note that we do not visit this with protocols with pcb layer
|
|
|
|
* code - like udp/tcp/raw ip.
|
|
|
|
*/
|
2006-02-01 13:55:03 +00:00
|
|
|
if (ip_ipsec_input(m))
|
2001-06-11 12:39:29 +00:00
|
|
|
goto bad;
|
2007-07-03 12:13:45 +00:00
|
|
|
#endif /* IPSEC */
|
2001-06-11 12:39:29 +00:00
|
|
|
|
1994-05-24 10:09:53 +00:00
|
|
|
/*
|
|
|
|
* Switch out to protocol's input routine.
|
|
|
|
*/
|
2009-04-11 23:35:20 +00:00
|
|
|
IPSTAT_INC(ips_delivered);
|
2004-08-17 22:05:54 +00:00
|
|
|
|
2004-02-25 19:55:29 +00:00
|
|
|
(*inetsw[ip_protox[ip->ip_p]].pr_input)(m, hlen);
|
1996-02-05 20:36:02 +00:00
|
|
|
return;
|
1994-05-24 10:09:53 +00:00
|
|
|
bad:
|
|
|
|
m_freem(m);
|
1996-02-05 20:36:02 +00:00
|
|
|
}
|
|
|
|
|
2006-01-15 18:58:21 +00:00
|
|
|
/*
|
|
|
|
* After maxnipq has been updated, propagate the change to UMA. The UMA zone
|
|
|
|
* max has slightly different semantics than the sysctl, for historical
|
|
|
|
* reasons.
|
|
|
|
*/
|
|
|
|
static void
|
|
|
|
maxnipq_update(void)
|
|
|
|
{
|
|
|
|
|
|
|
|
/*
|
|
|
|
* -1 for unlimited allocation.
|
|
|
|
*/
|
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 (V_maxnipq < 0)
|
|
|
|
uma_zone_set_max(V_ipq_zone, 0);
|
2006-01-15 18:58:21 +00:00
|
|
|
/*
|
|
|
|
* Positive number for specific bound.
|
|
|
|
*/
|
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 (V_maxnipq > 0)
|
|
|
|
uma_zone_set_max(V_ipq_zone, V_maxnipq);
|
2006-01-15 18:58:21 +00:00
|
|
|
/*
|
|
|
|
* Zero specifies no further fragment queue allocation -- set the
|
|
|
|
* bound very low, but rely on implementation elsewhere to actually
|
|
|
|
* prevent allocation and reclaim current queues.
|
|
|
|
*/
|
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 (V_maxnipq == 0)
|
|
|
|
uma_zone_set_max(V_ipq_zone, 1);
|
2006-01-15 18:58:21 +00:00
|
|
|
}
|
|
|
|
|
2006-04-21 09:25:40 +00:00
|
|
|
static void
|
|
|
|
ipq_zone_change(void *tag)
|
|
|
|
{
|
|
|
|
|
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 (V_maxnipq > 0 && V_maxnipq < (nmbclusters / 32)) {
|
|
|
|
V_maxnipq = nmbclusters / 32;
|
2006-04-21 09:25:40 +00:00
|
|
|
maxnipq_update();
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2006-01-15 18:58:21 +00:00
|
|
|
static int
|
|
|
|
sysctl_maxnipq(SYSCTL_HANDLER_ARGS)
|
|
|
|
{
|
|
|
|
int error, i;
|
|
|
|
|
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
|
|
|
i = V_maxnipq;
|
2006-01-15 18:58:21 +00:00
|
|
|
error = sysctl_handle_int(oidp, &i, 0, req);
|
|
|
|
if (error || !req->newptr)
|
|
|
|
return (error);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* XXXRW: Might be a good idea to sanity check the argument and place
|
|
|
|
* an extreme upper bound.
|
|
|
|
*/
|
|
|
|
if (i < -1)
|
|
|
|
return (EINVAL);
|
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
|
|
|
V_maxnipq = i;
|
2006-01-15 18:58:21 +00:00
|
|
|
maxnipq_update();
|
|
|
|
return (0);
|
|
|
|
}
|
|
|
|
|
|
|
|
SYSCTL_PROC(_net_inet_ip, OID_AUTO, maxfragpackets, CTLTYPE_INT|CTLFLAG_RW,
|
|
|
|
NULL, 0, sysctl_maxnipq, "I",
|
|
|
|
"Maximum number of IPv4 fragment reassembly queue entries");
|
|
|
|
|
1994-05-24 10:09:53 +00:00
|
|
|
/*
|
1999-12-06 00:43:07 +00:00
|
|
|
* Take incoming datagram fragment and try to reassemble it into
|
2004-08-03 12:31:38 +00:00
|
|
|
* whole datagram. If the argument is the first fragment or one
|
|
|
|
* in between the function will return NULL and store the mbuf
|
|
|
|
* in the fragment chain. If the argument is the last fragment
|
|
|
|
* the packet will be reassembled and the pointer to the new
|
|
|
|
* mbuf returned for further processing. Only m_tags attached
|
|
|
|
* to the first packet/fragment are preserved.
|
|
|
|
* The IP header is *NOT* adjusted out of iplen.
|
1994-05-24 10:09:53 +00:00
|
|
|
*/
|
2004-08-03 12:31:38 +00:00
|
|
|
struct mbuf *
|
|
|
|
ip_reass(struct mbuf *m)
|
1994-05-24 10:09:53 +00:00
|
|
|
{
|
2004-08-03 12:31:38 +00:00
|
|
|
struct ip *ip;
|
|
|
|
struct mbuf *p, *q, *nq, *t;
|
|
|
|
struct ipq *fp = NULL;
|
|
|
|
struct ipqhead *head;
|
|
|
|
int i, hlen, next;
|
2003-10-29 15:07:04 +00:00
|
|
|
u_int8_t ecn, ecn0;
|
2004-08-03 12:31:38 +00:00
|
|
|
u_short hash;
|
1994-05-24 10:09:53 +00:00
|
|
|
|
2005-04-08 10:25:13 +00:00
|
|
|
/* If maxnipq or maxfragsperpacket are 0, never accept fragments. */
|
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 (V_maxnipq == 0 || V_maxfragsperpacket == 0) {
|
2009-04-11 23:35:20 +00:00
|
|
|
IPSTAT_INC(ips_fragments);
|
|
|
|
IPSTAT_INC(ips_fragdropped);
|
2004-08-12 08:37:42 +00:00
|
|
|
m_freem(m);
|
|
|
|
return (NULL);
|
2004-08-03 12:31:38 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
ip = mtod(m, struct ip *);
|
|
|
|
hlen = ip->ip_hl << 2;
|
|
|
|
|
|
|
|
hash = IPREASS_HASH(ip->ip_src.s_addr, ip->ip_id);
|
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
|
|
|
head = &V_ipq[hash];
|
2004-08-03 12:31:38 +00:00
|
|
|
IPQ_LOCK();
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Look for queue of fragments
|
|
|
|
* of this datagram.
|
|
|
|
*/
|
|
|
|
TAILQ_FOREACH(fp, head, ipq_list)
|
|
|
|
if (ip->ip_id == fp->ipq_id &&
|
|
|
|
ip->ip_src.s_addr == fp->ipq_src.s_addr &&
|
|
|
|
ip->ip_dst.s_addr == fp->ipq_dst.s_addr &&
|
|
|
|
#ifdef MAC
|
2007-10-24 19:04:04 +00:00
|
|
|
mac_ipq_match(m, fp) &&
|
2004-08-03 12:31:38 +00:00
|
|
|
#endif
|
|
|
|
ip->ip_p == fp->ipq_p)
|
|
|
|
goto found;
|
|
|
|
|
|
|
|
fp = NULL;
|
|
|
|
|
|
|
|
/*
|
2006-01-15 18:58:21 +00:00
|
|
|
* Attempt to trim the number of allocated fragment queues if it
|
|
|
|
* exceeds the administrative limit.
|
2004-08-03 12:31:38 +00:00
|
|
|
*/
|
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 ((V_nipq > V_maxnipq) && (V_maxnipq > 0)) {
|
2004-08-03 12:31:38 +00:00
|
|
|
/*
|
|
|
|
* drop something from the tail of the current queue
|
|
|
|
* before proceeding further
|
|
|
|
*/
|
|
|
|
struct ipq *q = TAILQ_LAST(head, ipqhead);
|
|
|
|
if (q == NULL) { /* gak */
|
|
|
|
for (i = 0; i < IPREASS_NHASH; i++) {
|
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
|
|
|
struct ipq *r = TAILQ_LAST(&V_ipq[i], ipqhead);
|
2004-08-03 12:31:38 +00:00
|
|
|
if (r) {
|
2009-04-11 23:35:20 +00:00
|
|
|
IPSTAT_ADD(ips_fragtimeout,
|
|
|
|
r->ipq_nfrags);
|
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
|
|
|
ip_freef(&V_ipq[i], r);
|
2004-08-03 12:31:38 +00:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
} else {
|
2009-04-11 23:35:20 +00:00
|
|
|
IPSTAT_ADD(ips_fragtimeout, q->ipq_nfrags);
|
2004-08-03 12:31:38 +00:00
|
|
|
ip_freef(head, q);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
found:
|
|
|
|
/*
|
|
|
|
* Adjust ip_len to not reflect header,
|
|
|
|
* convert offset of this to bytes.
|
|
|
|
*/
|
|
|
|
ip->ip_len -= hlen;
|
|
|
|
if (ip->ip_off & IP_MF) {
|
|
|
|
/*
|
|
|
|
* Make sure that fragments have a data length
|
|
|
|
* that's a non-zero multiple of 8 bytes.
|
|
|
|
*/
|
|
|
|
if (ip->ip_len == 0 || (ip->ip_len & 0x7) != 0) {
|
2009-04-11 23:35:20 +00:00
|
|
|
IPSTAT_INC(ips_toosmall); /* XXX */
|
2004-08-03 12:31:38 +00:00
|
|
|
goto dropfrag;
|
|
|
|
}
|
|
|
|
m->m_flags |= M_FRAG;
|
|
|
|
} else
|
|
|
|
m->m_flags &= ~M_FRAG;
|
|
|
|
ip->ip_off <<= 3;
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Attempt reassembly; if it succeeds, proceed.
|
|
|
|
* ip_reass() will return a different mbuf.
|
|
|
|
*/
|
2009-04-11 23:35:20 +00:00
|
|
|
IPSTAT_INC(ips_fragments);
|
2004-08-03 12:31:38 +00:00
|
|
|
m->m_pkthdr.header = ip;
|
2003-09-05 00:10:33 +00:00
|
|
|
|
2004-08-03 12:31:38 +00:00
|
|
|
/* Previous ip_reass() started here. */
|
1994-05-24 10:09:53 +00:00
|
|
|
/*
|
|
|
|
* Presence of header sizes in mbufs
|
|
|
|
* would confuse code below.
|
|
|
|
*/
|
|
|
|
m->m_data += hlen;
|
|
|
|
m->m_len -= hlen;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* If first fragment to arrive, create a reassembly queue.
|
|
|
|
*/
|
2003-06-06 19:32:48 +00:00
|
|
|
if (fp == NULL) {
|
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
|
|
|
fp = uma_zalloc(V_ipq_zone, M_NOWAIT);
|
2006-01-15 18:58:21 +00:00
|
|
|
if (fp == NULL)
|
1994-05-24 10:09:53 +00:00
|
|
|
goto dropfrag;
|
2002-07-31 17:17:51 +00:00
|
|
|
#ifdef MAC
|
2007-10-24 19:04:04 +00:00
|
|
|
if (mac_ipq_init(fp, M_NOWAIT) != 0) {
|
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
|
|
|
uma_zfree(V_ipq_zone, fp);
|
2006-05-05 06:24:34 +00:00
|
|
|
fp = NULL;
|
2003-03-26 15:12:03 +00:00
|
|
|
goto dropfrag;
|
|
|
|
}
|
2007-10-24 19:04:04 +00:00
|
|
|
mac_ipq_create(m, fp);
|
2002-07-31 17:17:51 +00:00
|
|
|
#endif
|
2001-03-16 20:00:53 +00:00
|
|
|
TAILQ_INSERT_HEAD(head, fp, ipq_list);
|
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
|
|
|
V_nipq++;
|
2003-02-22 06:41:47 +00:00
|
|
|
fp->ipq_nfrags = 1;
|
1994-05-24 10:09:53 +00:00
|
|
|
fp->ipq_ttl = IPFRAGTTL;
|
|
|
|
fp->ipq_p = ip->ip_p;
|
|
|
|
fp->ipq_id = ip->ip_id;
|
1998-08-24 07:47:39 +00:00
|
|
|
fp->ipq_src = ip->ip_src;
|
|
|
|
fp->ipq_dst = ip->ip_dst;
|
1998-12-21 22:40:54 +00:00
|
|
|
fp->ipq_frags = m;
|
|
|
|
m->m_nextpkt = NULL;
|
2005-04-08 10:25:13 +00:00
|
|
|
goto done;
|
2002-07-31 17:17:51 +00:00
|
|
|
} else {
|
2003-02-22 06:41:47 +00:00
|
|
|
fp->ipq_nfrags++;
|
2002-07-31 17:17:51 +00:00
|
|
|
#ifdef MAC
|
2007-10-24 19:04:04 +00:00
|
|
|
mac_ipq_update(m, fp);
|
2002-07-31 17:17:51 +00:00
|
|
|
#endif
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
|
|
|
|
1998-08-24 07:47:39 +00:00
|
|
|
#define GETIP(m) ((struct ip*)((m)->m_pkthdr.header))
|
|
|
|
|
2003-10-29 15:07:04 +00:00
|
|
|
/*
|
|
|
|
* Handle ECN by comparing this segment with the first one;
|
|
|
|
* if CE is set, do not lose CE.
|
|
|
|
* drop if CE and not-ECT are mixed for the same packet.
|
|
|
|
*/
|
|
|
|
ecn = ip->ip_tos & IPTOS_ECN_MASK;
|
|
|
|
ecn0 = GETIP(fp->ipq_frags)->ip_tos & IPTOS_ECN_MASK;
|
|
|
|
if (ecn == IPTOS_ECN_CE) {
|
|
|
|
if (ecn0 == IPTOS_ECN_NOTECT)
|
|
|
|
goto dropfrag;
|
|
|
|
if (ecn0 != IPTOS_ECN_CE)
|
|
|
|
GETIP(fp->ipq_frags)->ip_tos |= IPTOS_ECN_CE;
|
|
|
|
}
|
|
|
|
if (ecn == IPTOS_ECN_NOTECT && ecn0 != IPTOS_ECN_NOTECT)
|
|
|
|
goto dropfrag;
|
|
|
|
|
1994-05-24 10:09:53 +00:00
|
|
|
/*
|
|
|
|
* Find a segment which begins after this one does.
|
|
|
|
*/
|
1998-08-24 07:47:39 +00:00
|
|
|
for (p = NULL, q = fp->ipq_frags; q; p = q, q = q->m_nextpkt)
|
|
|
|
if (GETIP(q)->ip_off > ip->ip_off)
|
1994-05-24 10:09:53 +00:00
|
|
|
break;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* If there is a preceding segment, it may provide some of
|
|
|
|
* our data already. If so, drop the data from the incoming
|
1998-12-21 22:40:54 +00:00
|
|
|
* segment. If it provides all of our data, drop us, otherwise
|
|
|
|
* stick new segment in the proper place.
|
2000-03-27 19:14:27 +00:00
|
|
|
*
|
|
|
|
* If some of the data is dropped from the the preceding
|
|
|
|
* segment, then it's checksum is invalidated.
|
1994-05-24 10:09:53 +00:00
|
|
|
*/
|
1998-08-24 07:47:39 +00:00
|
|
|
if (p) {
|
|
|
|
i = GETIP(p)->ip_off + GETIP(p)->ip_len - ip->ip_off;
|
1994-05-24 10:09:53 +00:00
|
|
|
if (i > 0) {
|
|
|
|
if (i >= ip->ip_len)
|
|
|
|
goto dropfrag;
|
1999-12-22 19:13:38 +00:00
|
|
|
m_adj(m, i);
|
2000-03-27 19:14:27 +00:00
|
|
|
m->m_pkthdr.csum_flags = 0;
|
1994-05-24 10:09:53 +00:00
|
|
|
ip->ip_off += i;
|
|
|
|
ip->ip_len -= i;
|
|
|
|
}
|
1998-12-21 22:40:54 +00:00
|
|
|
m->m_nextpkt = p->m_nextpkt;
|
|
|
|
p->m_nextpkt = m;
|
|
|
|
} else {
|
|
|
|
m->m_nextpkt = fp->ipq_frags;
|
|
|
|
fp->ipq_frags = m;
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* While we overlap succeeding segments trim them or,
|
|
|
|
* if they are completely covered, dequeue them.
|
|
|
|
*/
|
1998-08-24 07:47:39 +00:00
|
|
|
for (; q != NULL && ip->ip_off + ip->ip_len > GETIP(q)->ip_off;
|
1998-12-21 22:40:54 +00:00
|
|
|
q = nq) {
|
2003-02-25 11:53:11 +00:00
|
|
|
i = (ip->ip_off + ip->ip_len) - GETIP(q)->ip_off;
|
1998-08-24 07:47:39 +00:00
|
|
|
if (i < GETIP(q)->ip_len) {
|
|
|
|
GETIP(q)->ip_len -= i;
|
|
|
|
GETIP(q)->ip_off += i;
|
|
|
|
m_adj(q, i);
|
2000-03-27 19:14:27 +00:00
|
|
|
q->m_pkthdr.csum_flags = 0;
|
1994-05-24 10:09:53 +00:00
|
|
|
break;
|
|
|
|
}
|
1998-08-24 07:47:39 +00:00
|
|
|
nq = q->m_nextpkt;
|
1998-12-21 22:40:54 +00:00
|
|
|
m->m_nextpkt = nq;
|
2009-04-11 23:35:20 +00:00
|
|
|
IPSTAT_INC(ips_fragdropped);
|
2003-02-22 06:41:47 +00:00
|
|
|
fp->ipq_nfrags--;
|
1998-08-24 07:47:39 +00:00
|
|
|
m_freem(q);
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
2003-02-22 06:41:47 +00:00
|
|
|
* Check for complete reassembly and perform frag per packet
|
|
|
|
* limiting.
|
|
|
|
*
|
|
|
|
* Frag limiting is performed here so that the nth frag has
|
|
|
|
* a chance to complete the packet before we drop the packet.
|
|
|
|
* As a result, n+1 frags are actually allowed per packet, but
|
|
|
|
* only n will ever be stored. (n = maxfragsperpacket.)
|
|
|
|
*
|
1994-05-24 10:09:53 +00:00
|
|
|
*/
|
|
|
|
next = 0;
|
1998-08-24 07:47:39 +00:00
|
|
|
for (p = NULL, q = fp->ipq_frags; q; p = q, q = q->m_nextpkt) {
|
2003-02-22 06:41:47 +00:00
|
|
|
if (GETIP(q)->ip_off != next) {
|
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 (fp->ipq_nfrags > V_maxfragsperpacket) {
|
2009-04-11 23:35:20 +00:00
|
|
|
IPSTAT_ADD(ips_fragdropped, fp->ipq_nfrags);
|
2003-02-22 06:41:47 +00:00
|
|
|
ip_freef(head, fp);
|
2003-02-25 11:49:01 +00:00
|
|
|
}
|
2004-08-03 12:31:38 +00:00
|
|
|
goto done;
|
2003-02-22 06:41:47 +00:00
|
|
|
}
|
1998-08-24 07:47:39 +00:00
|
|
|
next += GETIP(q)->ip_len;
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
1998-08-24 07:47:39 +00:00
|
|
|
/* Make sure the last packet didn't have the IP_MF flag */
|
2003-02-22 06:41:47 +00:00
|
|
|
if (p->m_flags & M_FRAG) {
|
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 (fp->ipq_nfrags > V_maxfragsperpacket) {
|
2009-04-11 23:35:20 +00:00
|
|
|
IPSTAT_ADD(ips_fragdropped, fp->ipq_nfrags);
|
2003-02-22 06:41:47 +00:00
|
|
|
ip_freef(head, fp);
|
2003-02-25 11:49:01 +00:00
|
|
|
}
|
2004-08-03 12:31:38 +00:00
|
|
|
goto done;
|
2003-02-22 06:41:47 +00:00
|
|
|
}
|
1994-05-24 10:09:53 +00:00
|
|
|
|
|
|
|
/*
|
1996-10-25 17:57:53 +00:00
|
|
|
* Reassembly is complete. Make sure the packet is a sane size.
|
|
|
|
*/
|
1998-08-24 07:47:39 +00:00
|
|
|
q = fp->ipq_frags;
|
|
|
|
ip = GETIP(q);
|
2002-10-20 22:52:07 +00:00
|
|
|
if (next + (ip->ip_hl << 2) > IP_MAXPACKET) {
|
2009-04-11 23:35:20 +00:00
|
|
|
IPSTAT_INC(ips_toolong);
|
|
|
|
IPSTAT_ADD(ips_fragdropped, fp->ipq_nfrags);
|
2001-03-16 20:00:53 +00:00
|
|
|
ip_freef(head, fp);
|
2004-08-03 12:31:38 +00:00
|
|
|
goto done;
|
1996-10-25 17:57:53 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Concatenate fragments.
|
1994-05-24 10:09:53 +00:00
|
|
|
*/
|
1998-08-24 07:47:39 +00:00
|
|
|
m = q;
|
1994-05-24 10:09:53 +00:00
|
|
|
t = m->m_next;
|
2005-01-30 19:29:47 +00:00
|
|
|
m->m_next = NULL;
|
1994-05-24 10:09:53 +00:00
|
|
|
m_cat(m, t);
|
1998-08-24 07:47:39 +00:00
|
|
|
nq = q->m_nextpkt;
|
2005-01-30 19:29:47 +00:00
|
|
|
q->m_nextpkt = NULL;
|
1998-08-24 07:47:39 +00:00
|
|
|
for (q = nq; q != NULL; q = nq) {
|
|
|
|
nq = q->m_nextpkt;
|
1998-09-10 08:56:40 +00:00
|
|
|
q->m_nextpkt = NULL;
|
2000-03-27 19:14:27 +00:00
|
|
|
m->m_pkthdr.csum_flags &= q->m_pkthdr.csum_flags;
|
|
|
|
m->m_pkthdr.csum_data += q->m_pkthdr.csum_data;
|
2000-09-14 21:06:48 +00:00
|
|
|
m_cat(m, q);
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
2006-02-07 11:48:10 +00:00
|
|
|
/*
|
|
|
|
* In order to do checksumming faster we do 'end-around carry' here
|
|
|
|
* (and not in for{} loop), though it implies we are not going to
|
|
|
|
* reassemble more than 64k fragments.
|
|
|
|
*/
|
|
|
|
m->m_pkthdr.csum_data =
|
|
|
|
(m->m_pkthdr.csum_data & 0xffff) + (m->m_pkthdr.csum_data >> 16);
|
2002-07-31 17:17:51 +00:00
|
|
|
#ifdef MAC
|
2007-10-24 19:04:04 +00:00
|
|
|
mac_ipq_reassemble(fp, m);
|
|
|
|
mac_ipq_destroy(fp);
|
2002-07-31 17:17:51 +00:00
|
|
|
#endif
|
1994-05-24 10:09:53 +00:00
|
|
|
|
|
|
|
/*
|
2004-08-03 12:31:38 +00:00
|
|
|
* Create header for new ip packet by modifying header of first
|
|
|
|
* packet; dequeue and discard fragment reassembly header.
|
1994-05-24 10:09:53 +00:00
|
|
|
* Make header visible.
|
|
|
|
*/
|
2004-08-03 12:31:38 +00:00
|
|
|
ip->ip_len = (ip->ip_hl << 2) + next;
|
1998-08-24 07:47:39 +00:00
|
|
|
ip->ip_src = fp->ipq_src;
|
|
|
|
ip->ip_dst = fp->ipq_dst;
|
2001-03-16 20:00:53 +00:00
|
|
|
TAILQ_REMOVE(head, fp, ipq_list);
|
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
|
|
|
V_nipq--;
|
|
|
|
uma_zfree(V_ipq_zone, fp);
|
2002-10-20 22:52:07 +00:00
|
|
|
m->m_len += (ip->ip_hl << 2);
|
|
|
|
m->m_data -= (ip->ip_hl << 2);
|
1994-05-24 10:09:53 +00:00
|
|
|
/* some debugging cruft by sklower, below, will go away soon */
|
2002-09-18 19:43:01 +00:00
|
|
|
if (m->m_flags & M_PKTHDR) /* XXX this should be done elsewhere */
|
|
|
|
m_fixhdr(m);
|
2009-04-11 23:35:20 +00:00
|
|
|
IPSTAT_INC(ips_reassembled);
|
2004-08-03 12:31:38 +00:00
|
|
|
IPQ_UNLOCK();
|
1999-12-22 19:13:38 +00:00
|
|
|
return (m);
|
1994-05-24 10:09:53 +00:00
|
|
|
|
|
|
|
dropfrag:
|
2009-04-11 23:35:20 +00:00
|
|
|
IPSTAT_INC(ips_fragdropped);
|
2003-06-06 19:32:48 +00:00
|
|
|
if (fp != NULL)
|
2003-02-22 06:41:47 +00:00
|
|
|
fp->ipq_nfrags--;
|
1994-05-24 10:09:53 +00:00
|
|
|
m_freem(m);
|
2004-08-03 12:31:38 +00:00
|
|
|
done:
|
|
|
|
IPQ_UNLOCK();
|
|
|
|
return (NULL);
|
1998-08-24 07:47:39 +00:00
|
|
|
|
|
|
|
#undef GETIP
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Free a fragment reassembly header and all
|
|
|
|
* associated datagrams.
|
|
|
|
*/
|
1995-11-14 20:34:56 +00:00
|
|
|
static void
|
2007-05-10 15:58:48 +00:00
|
|
|
ip_freef(struct ipqhead *fhp, struct ipq *fp)
|
1994-05-24 10:09:53 +00:00
|
|
|
{
|
2007-05-10 15:58:48 +00:00
|
|
|
struct mbuf *q;
|
1994-05-24 10:09:53 +00:00
|
|
|
|
2003-09-05 00:10:33 +00:00
|
|
|
IPQ_LOCK_ASSERT();
|
|
|
|
|
1998-08-24 07:47:39 +00:00
|
|
|
while (fp->ipq_frags) {
|
|
|
|
q = fp->ipq_frags;
|
|
|
|
fp->ipq_frags = q->m_nextpkt;
|
|
|
|
m_freem(q);
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
2001-03-16 20:00:53 +00:00
|
|
|
TAILQ_REMOVE(fhp, fp, ipq_list);
|
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
|
|
|
uma_zfree(V_ipq_zone, fp);
|
|
|
|
V_nipq--;
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* IP timer processing;
|
|
|
|
* if a timer expires on a reassembly
|
|
|
|
* queue, discard it.
|
|
|
|
*/
|
|
|
|
void
|
2007-05-10 15:58:48 +00:00
|
|
|
ip_slowtimo(void)
|
1994-05-24 10:09:53 +00:00
|
|
|
{
|
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
|
|
|
VNET_ITERATOR_DECL(vnet_iter);
|
2007-05-10 15:58:48 +00:00
|
|
|
struct ipq *fp;
|
1997-09-15 23:07:01 +00:00
|
|
|
int i;
|
|
|
|
|
2003-09-05 00:10:33 +00:00
|
|
|
IPQ_LOCK();
|
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
|
|
|
VNET_LIST_RLOCK();
|
|
|
|
VNET_FOREACH(vnet_iter) {
|
|
|
|
CURVNET_SET(vnet_iter);
|
|
|
|
for (i = 0; i < IPREASS_NHASH; i++) {
|
|
|
|
for(fp = TAILQ_FIRST(&V_ipq[i]); fp;) {
|
|
|
|
struct ipq *fpp;
|
|
|
|
|
|
|
|
fpp = fp;
|
|
|
|
fp = TAILQ_NEXT(fp, ipq_list);
|
|
|
|
if(--fpp->ipq_ttl == 0) {
|
2009-04-11 23:35:20 +00:00
|
|
|
IPSTAT_ADD(ips_fragtimeout,
|
|
|
|
fpp->ipq_nfrags);
|
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
|
|
|
ip_freef(&V_ipq[i], fpp);
|
|
|
|
}
|
1997-09-15 23:07:01 +00:00
|
|
|
}
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
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
|
|
|
/*
|
|
|
|
* If we are over the maximum number of fragments
|
|
|
|
* (due to the limit being lowered), drain off
|
|
|
|
* enough to get down to the new limit.
|
|
|
|
*/
|
|
|
|
if (V_maxnipq >= 0 && V_nipq > V_maxnipq) {
|
|
|
|
for (i = 0; i < IPREASS_NHASH; i++) {
|
|
|
|
while (V_nipq > V_maxnipq &&
|
|
|
|
!TAILQ_EMPTY(&V_ipq[i])) {
|
2009-04-11 23:35:20 +00:00
|
|
|
IPSTAT_ADD(ips_fragdropped,
|
|
|
|
TAILQ_FIRST(&V_ipq[i])->ipq_nfrags);
|
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
|
|
|
ip_freef(&V_ipq[i],
|
|
|
|
TAILQ_FIRST(&V_ipq[i]));
|
|
|
|
}
|
2001-06-03 23:33:23 +00:00
|
|
|
}
|
|
|
|
}
|
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
|
|
|
CURVNET_RESTORE();
|
2001-06-03 23:33:23 +00:00
|
|
|
}
|
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
|
|
|
VNET_LIST_RUNLOCK();
|
2003-09-05 00:10:33 +00:00
|
|
|
IPQ_UNLOCK();
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Drain off all datagram fragments.
|
|
|
|
*/
|
|
|
|
void
|
2007-05-10 15:58:48 +00:00
|
|
|
ip_drain(void)
|
1994-05-24 10:09:53 +00:00
|
|
|
{
|
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
|
|
|
VNET_ITERATOR_DECL(vnet_iter);
|
1997-09-15 23:07:01 +00:00
|
|
|
int i;
|
1995-12-19 20:46:15 +00:00
|
|
|
|
2003-09-05 00:10:33 +00:00
|
|
|
IPQ_LOCK();
|
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
|
|
|
VNET_LIST_RLOCK();
|
|
|
|
VNET_FOREACH(vnet_iter) {
|
|
|
|
CURVNET_SET(vnet_iter);
|
|
|
|
for (i = 0; i < IPREASS_NHASH; i++) {
|
|
|
|
while(!TAILQ_EMPTY(&V_ipq[i])) {
|
2009-04-11 23:35:20 +00:00
|
|
|
IPSTAT_ADD(ips_fragdropped,
|
|
|
|
TAILQ_FIRST(&V_ipq[i])->ipq_nfrags);
|
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
|
|
|
ip_freef(&V_ipq[i], TAILQ_FIRST(&V_ipq[i]));
|
|
|
|
}
|
1997-09-15 23:07:01 +00:00
|
|
|
}
|
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
|
|
|
CURVNET_RESTORE();
|
1997-09-15 23:07:01 +00:00
|
|
|
}
|
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
|
|
|
VNET_LIST_RUNLOCK();
|
2003-09-05 00:10:33 +00:00
|
|
|
IPQ_UNLOCK();
|
1995-12-19 20:46:15 +00:00
|
|
|
in_rtqdrain();
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
|
|
|
|
2004-10-19 15:45:57 +00:00
|
|
|
/*
|
|
|
|
* The protocol to be inserted into ip_protox[] must be already registered
|
|
|
|
* in inetsw[], either statically or through pf_proto_register().
|
|
|
|
*/
|
|
|
|
int
|
|
|
|
ipproto_register(u_char ipproto)
|
|
|
|
{
|
|
|
|
struct protosw *pr;
|
|
|
|
|
|
|
|
/* Sanity checks. */
|
|
|
|
if (ipproto == 0)
|
|
|
|
return (EPROTONOSUPPORT);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* The protocol slot must not be occupied by another protocol
|
|
|
|
* already. An index pointing to IPPROTO_RAW is unused.
|
|
|
|
*/
|
|
|
|
pr = pffindproto(PF_INET, IPPROTO_RAW, SOCK_RAW);
|
|
|
|
if (pr == NULL)
|
|
|
|
return (EPFNOSUPPORT);
|
|
|
|
if (ip_protox[ipproto] != pr - inetsw) /* IPPROTO_RAW */
|
|
|
|
return (EEXIST);
|
|
|
|
|
|
|
|
/* Find the protocol position in inetsw[] and set the index. */
|
|
|
|
for (pr = inetdomain.dom_protosw;
|
|
|
|
pr < inetdomain.dom_protoswNPROTOSW; pr++) {
|
|
|
|
if (pr->pr_domain->dom_family == PF_INET &&
|
|
|
|
pr->pr_protocol && pr->pr_protocol == ipproto) {
|
|
|
|
/* Be careful to only index valid IP protocols. */
|
2005-02-23 00:38:12 +00:00
|
|
|
if (pr->pr_protocol < IPPROTO_MAX) {
|
2004-10-19 15:45:57 +00:00
|
|
|
ip_protox[pr->pr_protocol] = pr - inetsw;
|
|
|
|
return (0);
|
|
|
|
} else
|
|
|
|
return (EINVAL);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
return (EPROTONOSUPPORT);
|
|
|
|
}
|
|
|
|
|
|
|
|
int
|
|
|
|
ipproto_unregister(u_char ipproto)
|
|
|
|
{
|
|
|
|
struct protosw *pr;
|
|
|
|
|
|
|
|
/* Sanity checks. */
|
|
|
|
if (ipproto == 0)
|
|
|
|
return (EPROTONOSUPPORT);
|
|
|
|
|
|
|
|
/* Check if the protocol was indeed registered. */
|
|
|
|
pr = pffindproto(PF_INET, IPPROTO_RAW, SOCK_RAW);
|
|
|
|
if (pr == NULL)
|
|
|
|
return (EPFNOSUPPORT);
|
|
|
|
if (ip_protox[ipproto] == pr - inetsw) /* IPPROTO_RAW */
|
|
|
|
return (ENOENT);
|
|
|
|
|
|
|
|
/* Reset the protocol slot to IPPROTO_RAW. */
|
|
|
|
ip_protox[ipproto] = pr - inetsw;
|
|
|
|
return (0);
|
|
|
|
}
|
|
|
|
|
1994-05-24 10:09:53 +00:00
|
|
|
/*
|
2009-06-23 20:19:09 +00:00
|
|
|
* Given address of next destination (final or next hop), return (referenced)
|
|
|
|
* internet address info of interface to be used to get there.
|
1994-05-24 10:09:53 +00:00
|
|
|
*/
|
2001-11-30 10:40:28 +00:00
|
|
|
struct in_ifaddr *
|
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
|
|
|
ip_rtaddr(struct in_addr dst, u_int fibnum)
|
1994-05-24 10:09:53 +00:00
|
|
|
{
|
2003-11-20 20:07:39 +00:00
|
|
|
struct route sro;
|
2003-11-14 21:48:57 +00:00
|
|
|
struct sockaddr_in *sin;
|
2009-06-24 14:29:40 +00:00
|
|
|
struct in_ifaddr *ia;
|
2003-11-14 21:48:57 +00:00
|
|
|
|
2003-11-26 20:31:13 +00:00
|
|
|
bzero(&sro, sizeof(sro));
|
2003-11-20 20:07:39 +00:00
|
|
|
sin = (struct sockaddr_in *)&sro.ro_dst;
|
2003-11-14 21:48:57 +00:00
|
|
|
sin->sin_family = AF_INET;
|
|
|
|
sin->sin_len = sizeof(*sin);
|
|
|
|
sin->sin_addr = dst;
|
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
|
|
|
in_rtalloc_ign(&sro, 0, fibnum);
|
2003-11-14 21:48:57 +00:00
|
|
|
|
2003-11-20 20:07:39 +00:00
|
|
|
if (sro.ro_rt == NULL)
|
2005-01-30 19:29:47 +00:00
|
|
|
return (NULL);
|
2003-11-14 21:48:57 +00:00
|
|
|
|
2009-06-24 14:29:40 +00:00
|
|
|
ia = ifatoia(sro.ro_rt->rt_ifa);
|
|
|
|
ifa_ref(&ia->ia_ifa);
|
2003-11-20 20:07:39 +00:00
|
|
|
RTFREE(sro.ro_rt);
|
2009-06-24 14:29:40 +00:00
|
|
|
return (ia);
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
u_char inetctlerrmap[PRC_NCMDS] = {
|
|
|
|
0, 0, 0, 0,
|
|
|
|
0, EMSGSIZE, EHOSTDOWN, EHOSTUNREACH,
|
|
|
|
EHOSTUNREACH, EHOSTUNREACH, ECONNREFUSED, ECONNREFUSED,
|
|
|
|
EMSGSIZE, EHOSTUNREACH, 0, 0,
|
2003-06-17 06:21:08 +00:00
|
|
|
0, 0, EHOSTUNREACH, 0,
|
2001-08-27 22:10:07 +00:00
|
|
|
ENOPROTOOPT, ECONNREFUSED
|
1994-05-24 10:09:53 +00:00
|
|
|
};
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Forward a packet. If some error occurs return the sender
|
|
|
|
* an icmp packet. Note we can't always generate a meaningful
|
|
|
|
* icmp message because icmp doesn't have a large enough repertoire
|
|
|
|
* of codes and types.
|
|
|
|
*
|
|
|
|
* If not forwarding, just drop the packet. This could be confusing
|
|
|
|
* if ipforwarding was zero but some routing protocol was advancing
|
|
|
|
* us as a gateway to somewhere. However, we must let the routing
|
|
|
|
* protocol deal with that.
|
|
|
|
*
|
|
|
|
* The srcrt parameter indicates whether the packet is being forwarded
|
|
|
|
* via a source route.
|
|
|
|
*/
|
2004-08-17 22:05:54 +00:00
|
|
|
void
|
|
|
|
ip_forward(struct mbuf *m, int srcrt)
|
1994-05-24 10:09:53 +00:00
|
|
|
{
|
Remove (almost all) global variables that were used to hold
packet forwarding state ("annotations") during ip processing.
The code is considerably cleaner now.
The variables removed by this change are:
ip_divert_cookie used by divert sockets
ip_fw_fwd_addr used for transparent ip redirection
last_pkt used by dynamic pipes in dummynet
Removal of the first two has been done by carrying the annotations
into volatile structs prepended to the mbuf chains, and adding
appropriate code to add/remove annotations in the routines which
make use of them, i.e. ip_input(), ip_output(), tcp_input(),
bdg_forward(), ether_demux(), ether_output_frame(), div_output().
On passing, remove a bug in divert handling of fragmented packet.
Now it is the fragment at offset 0 which sets the divert status of
the whole packet, whereas formerly it was the last incoming fragment
to decide.
Removal of last_pkt required a change in the interface of ip_fw_chk()
and dummynet_io(). On passing, use the same mechanism for dummynet
annotations and for divert/forward annotations.
option IPFIREWALL_FORWARD is effectively useless, the code to
implement it is very small and is now in by default to avoid the
obfuscation of conditionally compiled code.
NOTES:
* there is at least one global variable left, sro_fwd, in ip_output().
I am not sure if/how this can be removed.
* I have deliberately avoided gratuitous style changes in this commit
to avoid cluttering the diffs. Minor stule cleanup will likely be
necessary
* this commit only focused on the IP layer. I am sure there is a
number of global variables used in the TCP and maybe UDP stack.
* despite the number of files touched, there are absolutely no API's
or data structures changed by this commit (except the interfaces of
ip_fw_chk() and dummynet_io(), which are internal anyways), so
an MFC is quite safe and unintrusive (and desirable, given the
improved readability of the code).
MFC after: 10 days
2002-06-22 11:51:02 +00:00
|
|
|
struct ip *ip = mtod(m, struct ip *);
|
2009-05-27 12:44:36 +00:00
|
|
|
struct in_ifaddr *ia;
|
1994-05-24 10:09:53 +00:00
|
|
|
struct mbuf *mcopy;
|
2004-08-17 22:05:54 +00:00
|
|
|
struct in_addr dest;
|
2008-04-09 05:17:18 +00:00
|
|
|
struct route ro;
|
2005-05-04 13:09:19 +00:00
|
|
|
int error, type = 0, code = 0, mtu = 0;
|
2001-12-28 21:21:57 +00:00
|
|
|
|
2004-08-17 22:05:54 +00:00
|
|
|
if (m->m_flags & (M_BCAST|M_MCAST) || in_canforward(ip->ip_dst) == 0) {
|
2009-04-11 23:35:20 +00:00
|
|
|
IPSTAT_INC(ips_cantforward);
|
1994-05-24 10:09:53 +00:00
|
|
|
m_freem(m);
|
|
|
|
return;
|
|
|
|
}
|
1999-02-22 18:19:57 +00:00
|
|
|
#ifdef IPSTEALTH
|
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 (!V_ipstealth) {
|
1999-02-22 18:19:57 +00:00
|
|
|
#endif
|
|
|
|
if (ip->ip_ttl <= IPTTLDEC) {
|
|
|
|
icmp_error(m, ICMP_TIMXCEED, ICMP_TIMXCEED_INTRANS,
|
2003-11-14 21:48:57 +00:00
|
|
|
0, 0);
|
1999-02-22 18:19:57 +00:00
|
|
|
return;
|
|
|
|
}
|
|
|
|
#ifdef IPSTEALTH
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
1999-02-22 18:19:57 +00:00
|
|
|
#endif
|
1994-05-24 10:09:53 +00:00
|
|
|
|
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
|
|
|
ia = ip_rtaddr(ip->ip_dst, M_GETFIB(m));
|
2009-05-27 12:44:36 +00:00
|
|
|
#ifndef IPSEC
|
|
|
|
/*
|
|
|
|
* 'ia' may be NULL if there is no route for this destination.
|
|
|
|
* In case of IPsec, Don't discard it just yet, but pass it to
|
|
|
|
* ip_output in case of outgoing IPsec policy.
|
|
|
|
*/
|
2007-12-02 13:00:47 +00:00
|
|
|
if (!srcrt && ia == NULL) {
|
2003-11-14 21:48:57 +00:00
|
|
|
icmp_error(m, ICMP_UNREACH, ICMP_UNREACH_HOST, 0, 0);
|
Make sure the cached forwarding route (ipforward_rt) is still up before
using it. Not checking this may have caused the wrong IP address to be
used when processing certain IP options (see example below). This also
caused the wrong route to be passed to ip_output() when forwarding, but
fortunately ip_output() is smart enough to detect this.
This example demonstrates the wrong behavior of the Record Route option
observed with this bug. Host ``freebsd'' is acting as the gateway for
the ``sysv''.
1. On the gateway, we add the route to the destination. The new route
will use the primary address of the loopback interface, 127.0.0.1:
: freebsd# route add 10.0.0.66 -iface lo0 -reject
: add host 10.0.0.66: gateway lo0
2. From the client, we ping the destination. We see the correct replies.
Please note that this also causes the relevant route on the ``freebsd''
gateway to be cached in ipforward_rt variable:
: sysv# ping -snv 10.0.0.66
: PING 10.0.0.66: 56 data bytes
: ICMP Host Unreachable from gateway 192.168.0.115
: ICMP Host Unreachable from gateway 192.168.0.115
: ICMP Host Unreachable from gateway 192.168.0.115
:
: ----10.0.0.66 PING Statistics----
: 3 packets transmitted, 0 packets received, 100% packet loss
3. On the gateway, we delete the route to the destination, thus making
the destination reachable through the `default' route:
: freebsd# route delete 10.0.0.66
: delete host 10.0.0.66
4. From the client, we ping destination again, now with the RR option
turned on. The surprise here is the 127.0.0.1 in the first reply.
This is caused by the bug in ip_rtaddr() not checking the cached
route is still up befor use. The debug code also shows that the
wrong (down) route is further passed to ip_output(). The latter
detects that the route is down, and replaces the bogus route with
the valid one, so we see the correct replies (192.168.0.115) on
further probes:
: sysv# ping -snRv 10.0.0.66
: PING 10.0.0.66: 56 data bytes
: 64 bytes from 10.0.0.66: icmp_seq=0. time=10. ms
: IP options: <record route> 127.0.0.1, 10.0.0.65, 10.0.0.66,
: 192.168.0.65, 192.168.0.115, 192.168.0.120,
: 0.0.0.0(Current), 0.0.0.0, 0.0.0.0
: 64 bytes from 10.0.0.66: icmp_seq=1. time=0. ms
: IP options: <record route> 192.168.0.115, 10.0.0.65, 10.0.0.66,
: 192.168.0.65, 192.168.0.115, 192.168.0.120,
: 0.0.0.0(Current), 0.0.0.0, 0.0.0.0
: 64 bytes from 10.0.0.66: icmp_seq=2. time=0. ms
: IP options: <record route> 192.168.0.115, 10.0.0.65, 10.0.0.66,
: 192.168.0.65, 192.168.0.115, 192.168.0.120,
: 0.0.0.0(Current), 0.0.0.0, 0.0.0.0
:
: ----10.0.0.66 PING Statistics----
: 3 packets transmitted, 3 packets received, 0% packet loss
: round-trip (ms) min/avg/max = 0/3/10
2001-03-18 13:04:07 +00:00
|
|
|
return;
|
2003-11-14 21:48:57 +00:00
|
|
|
}
|
2009-05-27 12:44:36 +00:00
|
|
|
#endif
|
1994-05-24 10:09:53 +00:00
|
|
|
|
|
|
|
/*
|
It was possible for ip_forward() to supply to icmp_error()
an IP header with ip_len in network byte order. For certain
values of ip_len, this could cause icmp_error() to write
beyond the end of an mbuf, causing mbuf free-list corruption.
This problem was observed during generation of ICMP redirects.
We now make quite sure that the copy of the IP header kept
for icmp_error() is stored in a non-shared mbuf header so
that it will not be modified by ip_output().
Also:
- Calculate the correct number of bytes that need to be
retained for icmp_error(), instead of assuming that 64
is enough (it's not).
- In icmp_error(), use m_copydata instead of bcopy() to
copy from the supplied mbuf chain, in case the first 8
bytes of IP payload are not stored directly after the IP
header.
- Sanity-check ip_len in icmp_error(), and panic if it is
less than sizeof(struct ip). Incoming packets with bad
ip_len values are discarded in ip_input(), so this should
only be triggered by bugs in the code, not by bad packets.
This patch results from code and suggestions from Ruslan, Bosko,
Jonathan Lemon and Matt Dillon, with important testing by Mike
Tancsa, who could reproduce this problem at will.
Reported by: Mike Tancsa <mike@sentex.net>
Reviewed by: ru, bmilekic, jlemon, dillon
2001-03-08 19:03:26 +00:00
|
|
|
* Save the IP header and at most 8 bytes of the payload,
|
|
|
|
* in case we need to generate an ICMP message to the src.
|
|
|
|
*
|
2002-06-23 20:48:26 +00:00
|
|
|
* XXX this can be optimized a lot by saving the data in a local
|
|
|
|
* buffer on the stack (72 bytes at most), and only allocating the
|
|
|
|
* mbuf if really necessary. The vast majority of the packets
|
|
|
|
* are forwarded without having to send an ICMP back (either
|
|
|
|
* because unnecessary, or because rate limited), so we are
|
|
|
|
* really we are wasting a lot of work here.
|
|
|
|
*
|
It was possible for ip_forward() to supply to icmp_error()
an IP header with ip_len in network byte order. For certain
values of ip_len, this could cause icmp_error() to write
beyond the end of an mbuf, causing mbuf free-list corruption.
This problem was observed during generation of ICMP redirects.
We now make quite sure that the copy of the IP header kept
for icmp_error() is stored in a non-shared mbuf header so
that it will not be modified by ip_output().
Also:
- Calculate the correct number of bytes that need to be
retained for icmp_error(), instead of assuming that 64
is enough (it's not).
- In icmp_error(), use m_copydata instead of bcopy() to
copy from the supplied mbuf chain, in case the first 8
bytes of IP payload are not stored directly after the IP
header.
- Sanity-check ip_len in icmp_error(), and panic if it is
less than sizeof(struct ip). Incoming packets with bad
ip_len values are discarded in ip_input(), so this should
only be triggered by bugs in the code, not by bad packets.
This patch results from code and suggestions from Ruslan, Bosko,
Jonathan Lemon and Matt Dillon, with important testing by Mike
Tancsa, who could reproduce this problem at will.
Reported by: Mike Tancsa <mike@sentex.net>
Reviewed by: ru, bmilekic, jlemon, dillon
2001-03-08 19:03:26 +00:00
|
|
|
* We don't use m_copy() because it might return a reference
|
|
|
|
* to a shared cluster. Both this function and ip_output()
|
|
|
|
* assume exclusive access to the IP header in `m', so any
|
|
|
|
* data in a cluster may change before we reach icmp_error().
|
1994-05-24 10:09:53 +00:00
|
|
|
*/
|
2005-11-18 14:44:48 +00:00
|
|
|
MGETHDR(mcopy, M_DONTWAIT, m->m_type);
|
2003-02-19 05:47:46 +00:00
|
|
|
if (mcopy != NULL && !m_dup_pkthdr(mcopy, m, M_DONTWAIT)) {
|
2002-12-30 20:22:40 +00:00
|
|
|
/*
|
|
|
|
* It's probably ok if the pkthdr dup fails (because
|
|
|
|
* the deep copy of the tag chain failed), but for now
|
|
|
|
* be conservative and just discard the copy since
|
|
|
|
* code below may some day want the tags.
|
|
|
|
*/
|
|
|
|
m_free(mcopy);
|
|
|
|
mcopy = NULL;
|
|
|
|
}
|
It was possible for ip_forward() to supply to icmp_error()
an IP header with ip_len in network byte order. For certain
values of ip_len, this could cause icmp_error() to write
beyond the end of an mbuf, causing mbuf free-list corruption.
This problem was observed during generation of ICMP redirects.
We now make quite sure that the copy of the IP header kept
for icmp_error() is stored in a non-shared mbuf header so
that it will not be modified by ip_output().
Also:
- Calculate the correct number of bytes that need to be
retained for icmp_error(), instead of assuming that 64
is enough (it's not).
- In icmp_error(), use m_copydata instead of bcopy() to
copy from the supplied mbuf chain, in case the first 8
bytes of IP payload are not stored directly after the IP
header.
- Sanity-check ip_len in icmp_error(), and panic if it is
less than sizeof(struct ip). Incoming packets with bad
ip_len values are discarded in ip_input(), so this should
only be triggered by bugs in the code, not by bad packets.
This patch results from code and suggestions from Ruslan, Bosko,
Jonathan Lemon and Matt Dillon, with important testing by Mike
Tancsa, who could reproduce this problem at will.
Reported by: Mike Tancsa <mike@sentex.net>
Reviewed by: ru, bmilekic, jlemon, dillon
2001-03-08 19:03:26 +00:00
|
|
|
if (mcopy != NULL) {
|
2005-11-18 14:44:48 +00:00
|
|
|
mcopy->m_len = min(ip->ip_len, M_TRAILINGSPACE(mcopy));
|
2004-06-16 08:28:54 +00:00
|
|
|
mcopy->m_pkthdr.len = mcopy->m_len;
|
It was possible for ip_forward() to supply to icmp_error()
an IP header with ip_len in network byte order. For certain
values of ip_len, this could cause icmp_error() to write
beyond the end of an mbuf, causing mbuf free-list corruption.
This problem was observed during generation of ICMP redirects.
We now make quite sure that the copy of the IP header kept
for icmp_error() is stored in a non-shared mbuf header so
that it will not be modified by ip_output().
Also:
- Calculate the correct number of bytes that need to be
retained for icmp_error(), instead of assuming that 64
is enough (it's not).
- In icmp_error(), use m_copydata instead of bcopy() to
copy from the supplied mbuf chain, in case the first 8
bytes of IP payload are not stored directly after the IP
header.
- Sanity-check ip_len in icmp_error(), and panic if it is
less than sizeof(struct ip). Incoming packets with bad
ip_len values are discarded in ip_input(), so this should
only be triggered by bugs in the code, not by bad packets.
This patch results from code and suggestions from Ruslan, Bosko,
Jonathan Lemon and Matt Dillon, with important testing by Mike
Tancsa, who could reproduce this problem at will.
Reported by: Mike Tancsa <mike@sentex.net>
Reviewed by: ru, bmilekic, jlemon, dillon
2001-03-08 19:03:26 +00:00
|
|
|
m_copydata(m, 0, mcopy->m_len, mtod(mcopy, caddr_t));
|
|
|
|
}
|
Fixed broken ICMP error generation, unified conversion of IP header
fields between host and network byte order. The details:
o icmp_error() now does not add IP header length. This fixes the problem
when icmp_error() is called from ip_forward(). In this case the ip_len
of the original IP datagram returned with ICMP error was wrong.
o icmp_error() expects all three fields, ip_len, ip_id and ip_off in host
byte order, so DTRT and convert these fields back to network byte order
before sending a message. This fixes the problem described in PR 16240
and PR 20877 (ip_id field was returned in host byte order).
o ip_ttl decrement operation in ip_forward() was moved down to make sure
that it does not corrupt the copy of original IP datagram passed later
to icmp_error().
o A copy of original IP datagram in ip_forward() was made a read-write,
independent copy. This fixes the problem I first reported to Garrett
Wollman and Bill Fenner and later put in audit trail of PR 16240:
ip_output() (not always) converts fields of original datagram to network
byte order, but because copy (mcopy) and its original (m) most likely
share the same mbuf cluster, ip_output()'s manipulations on original
also corrupted the copy.
o ip_output() now expects all three fields, ip_len, ip_off and (what is
significant) ip_id in host byte order. It was a headache for years that
ip_id was handled differently. The only compatibility issue here is the
raw IP socket interface with IP_HDRINCL socket option set and a non-zero
ip_id field, but ip.4 manual page was unclear on whether in this case
ip_id field should be in host or network byte order.
2000-09-01 12:33:03 +00:00
|
|
|
|
|
|
|
#ifdef IPSTEALTH
|
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 (!V_ipstealth) {
|
Fixed broken ICMP error generation, unified conversion of IP header
fields between host and network byte order. The details:
o icmp_error() now does not add IP header length. This fixes the problem
when icmp_error() is called from ip_forward(). In this case the ip_len
of the original IP datagram returned with ICMP error was wrong.
o icmp_error() expects all three fields, ip_len, ip_id and ip_off in host
byte order, so DTRT and convert these fields back to network byte order
before sending a message. This fixes the problem described in PR 16240
and PR 20877 (ip_id field was returned in host byte order).
o ip_ttl decrement operation in ip_forward() was moved down to make sure
that it does not corrupt the copy of original IP datagram passed later
to icmp_error().
o A copy of original IP datagram in ip_forward() was made a read-write,
independent copy. This fixes the problem I first reported to Garrett
Wollman and Bill Fenner and later put in audit trail of PR 16240:
ip_output() (not always) converts fields of original datagram to network
byte order, but because copy (mcopy) and its original (m) most likely
share the same mbuf cluster, ip_output()'s manipulations on original
also corrupted the copy.
o ip_output() now expects all three fields, ip_len, ip_off and (what is
significant) ip_id in host byte order. It was a headache for years that
ip_id was handled differently. The only compatibility issue here is the
raw IP socket interface with IP_HDRINCL socket option set and a non-zero
ip_id field, but ip.4 manual page was unclear on whether in this case
ip_id field should be in host or network byte order.
2000-09-01 12:33:03 +00:00
|
|
|
#endif
|
|
|
|
ip->ip_ttl -= IPTTLDEC;
|
|
|
|
#ifdef IPSTEALTH
|
|
|
|
}
|
|
|
|
#endif
|
1994-05-24 10:09:53 +00:00
|
|
|
|
|
|
|
/*
|
|
|
|
* If forwarding packet using same interface that it came in on,
|
|
|
|
* perhaps should send a redirect to sender to shortcut a hop.
|
|
|
|
* Only send redirect if source is sending directly to us,
|
|
|
|
* and if packet was not source routed (or has any options).
|
|
|
|
* Also, don't send redirect if forwarding using a default route
|
|
|
|
* or a route modified by a redirect.
|
|
|
|
*/
|
2004-08-17 22:05:54 +00:00
|
|
|
dest.s_addr = 0;
|
2009-05-27 12:44:36 +00:00
|
|
|
if (!srcrt && V_ipsendredirects &&
|
|
|
|
ia != NULL && ia->ia_ifp == m->m_pkthdr.rcvif) {
|
2003-11-14 21:48:57 +00:00
|
|
|
struct sockaddr_in *sin;
|
|
|
|
struct rtentry *rt;
|
|
|
|
|
2003-11-26 20:31:13 +00:00
|
|
|
bzero(&ro, sizeof(ro));
|
2003-11-14 21:48:57 +00:00
|
|
|
sin = (struct sockaddr_in *)&ro.ro_dst;
|
|
|
|
sin->sin_family = AF_INET;
|
|
|
|
sin->sin_len = sizeof(*sin);
|
2004-08-17 22:05:54 +00:00
|
|
|
sin->sin_addr = ip->ip_dst;
|
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
|
|
|
in_rtalloc_ign(&ro, 0, M_GETFIB(m));
|
2003-11-14 21:48:57 +00:00
|
|
|
|
|
|
|
rt = ro.ro_rt;
|
|
|
|
|
|
|
|
if (rt && (rt->rt_flags & (RTF_DYNAMIC|RTF_MODIFIED)) == 0 &&
|
2004-08-17 22:05:54 +00:00
|
|
|
satosin(rt_key(rt))->sin_addr.s_addr != 0) {
|
1994-05-24 10:09:53 +00:00
|
|
|
#define RTA(rt) ((struct in_ifaddr *)(rt->rt_ifa))
|
2003-11-14 21:48:57 +00:00
|
|
|
u_long src = ntohl(ip->ip_src.s_addr);
|
|
|
|
|
|
|
|
if (RTA(rt) &&
|
|
|
|
(src & RTA(rt)->ia_subnetmask) == RTA(rt)->ia_subnet) {
|
|
|
|
if (rt->rt_flags & RTF_GATEWAY)
|
2004-08-17 22:05:54 +00:00
|
|
|
dest.s_addr = satosin(rt->rt_gateway)->sin_addr.s_addr;
|
2003-11-14 21:48:57 +00:00
|
|
|
else
|
2004-08-17 22:05:54 +00:00
|
|
|
dest.s_addr = ip->ip_dst.s_addr;
|
2003-11-14 21:48:57 +00:00
|
|
|
/* Router requirements says to only send host redirects */
|
|
|
|
type = ICMP_REDIRECT;
|
|
|
|
code = ICMP_REDIRECT_HOST;
|
|
|
|
}
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
2003-11-14 21:48:57 +00:00
|
|
|
if (rt)
|
|
|
|
RTFREE(rt);
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
|
|
|
|
2008-04-09 05:17:18 +00:00
|
|
|
/*
|
|
|
|
* Try to cache the route MTU from ip_output so we can consider it for
|
|
|
|
* the ICMP_UNREACH_NEEDFRAG "Next-Hop MTU" field described in RFC1191.
|
|
|
|
*/
|
|
|
|
bzero(&ro, sizeof(ro));
|
|
|
|
|
|
|
|
error = ip_output(m, NULL, &ro, IP_FORWARDING, NULL, NULL);
|
|
|
|
|
|
|
|
if (error == EMSGSIZE && ro.ro_rt)
|
|
|
|
mtu = ro.ro_rt->rt_rmx.rmx_mtu;
|
|
|
|
if (ro.ro_rt)
|
|
|
|
RTFREE(ro.ro_rt);
|
|
|
|
|
1994-05-24 10:09:53 +00:00
|
|
|
if (error)
|
2009-04-11 23:35:20 +00:00
|
|
|
IPSTAT_INC(ips_cantforward);
|
1994-05-24 10:09:53 +00:00
|
|
|
else {
|
2009-04-11 23:35:20 +00:00
|
|
|
IPSTAT_INC(ips_forward);
|
1994-05-24 10:09:53 +00:00
|
|
|
if (type)
|
2009-04-11 23:35:20 +00:00
|
|
|
IPSTAT_INC(ips_redirectsent);
|
1994-05-24 10:09:53 +00:00
|
|
|
else {
|
2003-11-14 21:02:22 +00:00
|
|
|
if (mcopy)
|
1994-05-24 10:09:53 +00:00
|
|
|
m_freem(mcopy);
|
2009-06-23 20:19:09 +00:00
|
|
|
if (ia != NULL)
|
|
|
|
ifa_free(&ia->ia_ifa);
|
1994-05-24 10:09:53 +00:00
|
|
|
return;
|
|
|
|
}
|
|
|
|
}
|
2009-06-23 20:19:09 +00:00
|
|
|
if (mcopy == NULL) {
|
|
|
|
if (ia != NULL)
|
|
|
|
ifa_free(&ia->ia_ifa);
|
1994-05-24 10:09:53 +00:00
|
|
|
return;
|
2009-06-23 20:19:09 +00:00
|
|
|
}
|
1994-05-24 10:09:53 +00:00
|
|
|
|
|
|
|
switch (error) {
|
|
|
|
|
|
|
|
case 0: /* forwarded, but need redirect */
|
|
|
|
/* type, code set above */
|
|
|
|
break;
|
|
|
|
|
2009-05-27 12:44:36 +00:00
|
|
|
case ENETUNREACH:
|
1994-05-24 10:09:53 +00:00
|
|
|
case EHOSTUNREACH:
|
|
|
|
case ENETDOWN:
|
|
|
|
case EHOSTDOWN:
|
|
|
|
default:
|
|
|
|
type = ICMP_UNREACH;
|
|
|
|
code = ICMP_UNREACH_HOST;
|
|
|
|
break;
|
|
|
|
|
|
|
|
case EMSGSIZE:
|
|
|
|
type = ICMP_UNREACH;
|
|
|
|
code = ICMP_UNREACH_NEEDFRAG;
|
2002-10-16 02:25:05 +00:00
|
|
|
|
2007-07-03 12:13:45 +00:00
|
|
|
#ifdef IPSEC
|
2008-04-09 05:17:18 +00:00
|
|
|
/*
|
|
|
|
* If IPsec is configured for this path,
|
|
|
|
* override any possibly mtu value set by ip_output.
|
|
|
|
*/
|
|
|
|
mtu = ip_ipsec_mtu(m, mtu);
|
2007-07-03 12:13:45 +00:00
|
|
|
#endif /* IPSEC */
|
2004-08-17 22:05:54 +00:00
|
|
|
/*
|
2008-04-09 05:17:18 +00:00
|
|
|
* If the MTU was set before make sure we are below the
|
|
|
|
* interface MTU.
|
2006-01-24 17:57:19 +00:00
|
|
|
* If the MTU wasn't set before use the interface mtu or
|
|
|
|
* fall back to the next smaller mtu step compared to the
|
|
|
|
* current packet size.
|
2004-08-17 22:05:54 +00:00
|
|
|
*/
|
2008-04-09 05:17:18 +00:00
|
|
|
if (mtu != 0) {
|
|
|
|
if (ia != NULL)
|
|
|
|
mtu = min(mtu, ia->ia_ifp->if_mtu);
|
|
|
|
} else {
|
2006-01-24 17:57:19 +00:00
|
|
|
if (ia != NULL)
|
|
|
|
mtu = ia->ia_ifp->if_mtu;
|
|
|
|
else
|
|
|
|
mtu = ip_next_mtu(ip->ip_len, 0);
|
|
|
|
}
|
2009-04-11 23:35:20 +00:00
|
|
|
IPSTAT_INC(ips_cantfrag);
|
1994-05-24 10:09:53 +00:00
|
|
|
break;
|
|
|
|
|
|
|
|
case ENOBUFS:
|
2002-11-19 17:06:06 +00:00
|
|
|
/*
|
|
|
|
* A router should not generate ICMP_SOURCEQUENCH as
|
|
|
|
* required in RFC1812 Requirements for IP Version 4 Routers.
|
|
|
|
* Source quench could be a big problem under DoS attacks,
|
|
|
|
* or if the underlying interface is rate-limited.
|
|
|
|
* Those who need source quench packets may re-enable them
|
|
|
|
* via the net.inet.ip.sendsourcequench sysctl.
|
|
|
|
*/
|
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 (V_ip_sendsourcequench == 0) {
|
2002-11-19 17:06:06 +00:00
|
|
|
m_freem(mcopy);
|
2009-06-23 20:19:09 +00:00
|
|
|
if (ia != NULL)
|
|
|
|
ifa_free(&ia->ia_ifa);
|
2002-11-19 17:06:06 +00:00
|
|
|
return;
|
|
|
|
} else {
|
|
|
|
type = ICMP_SOURCEQUENCH;
|
|
|
|
code = 0;
|
|
|
|
}
|
1994-05-24 10:09:53 +00:00
|
|
|
break;
|
2000-05-15 18:41:01 +00:00
|
|
|
|
|
|
|
case EACCES: /* ipfw denied packet */
|
|
|
|
m_freem(mcopy);
|
2009-06-23 20:19:09 +00:00
|
|
|
if (ia != NULL)
|
|
|
|
ifa_free(&ia->ia_ifa);
|
2000-05-15 18:41:01 +00:00
|
|
|
return;
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
2009-06-23 20:19:09 +00:00
|
|
|
if (ia != NULL)
|
|
|
|
ifa_free(&ia->ia_ifa);
|
2005-05-04 13:09:19 +00:00
|
|
|
icmp_error(mcopy, type, code, dest.s_addr, mtu);
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
|
|
|
|
1996-11-11 04:56:32 +00:00
|
|
|
void
|
2007-05-10 15:58:48 +00:00
|
|
|
ip_savecontrol(struct inpcb *inp, struct mbuf **mp, struct ip *ip,
|
|
|
|
struct mbuf *m)
|
1996-11-11 04:56:32 +00:00
|
|
|
{
|
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
|
|
|
|
2004-01-31 10:40:25 +00:00
|
|
|
if (inp->inp_socket->so_options & (SO_BINTIME | SO_TIMESTAMP)) {
|
|
|
|
struct bintime bt;
|
|
|
|
|
|
|
|
bintime(&bt);
|
|
|
|
if (inp->inp_socket->so_options & SO_BINTIME) {
|
|
|
|
*mp = sbcreatecontrol((caddr_t) &bt, sizeof(bt),
|
|
|
|
SCM_BINTIME, SOL_SOCKET);
|
|
|
|
if (*mp)
|
|
|
|
mp = &(*mp)->m_next;
|
|
|
|
}
|
|
|
|
if (inp->inp_socket->so_options & SO_TIMESTAMP) {
|
|
|
|
struct timeval tv;
|
|
|
|
|
|
|
|
bintime2timeval(&bt, &tv);
|
|
|
|
*mp = sbcreatecontrol((caddr_t) &tv, sizeof(tv),
|
|
|
|
SCM_TIMESTAMP, SOL_SOCKET);
|
|
|
|
if (*mp)
|
|
|
|
mp = &(*mp)->m_next;
|
|
|
|
}
|
2002-05-31 11:52:35 +00:00
|
|
|
}
|
1996-11-11 04:56:32 +00:00
|
|
|
if (inp->inp_flags & INP_RECVDSTADDR) {
|
|
|
|
*mp = sbcreatecontrol((caddr_t) &ip->ip_dst,
|
|
|
|
sizeof(struct in_addr), IP_RECVDSTADDR, IPPROTO_IP);
|
|
|
|
if (*mp)
|
|
|
|
mp = &(*mp)->m_next;
|
|
|
|
}
|
2003-04-29 21:36:18 +00:00
|
|
|
if (inp->inp_flags & INP_RECVTTL) {
|
|
|
|
*mp = sbcreatecontrol((caddr_t) &ip->ip_ttl,
|
|
|
|
sizeof(u_char), IP_RECVTTL, IPPROTO_IP);
|
|
|
|
if (*mp)
|
|
|
|
mp = &(*mp)->m_next;
|
|
|
|
}
|
1996-11-11 04:56:32 +00:00
|
|
|
#ifdef notyet
|
|
|
|
/* XXX
|
|
|
|
* Moving these out of udp_input() made them even more broken
|
|
|
|
* than they already were.
|
|
|
|
*/
|
|
|
|
/* options were tossed already */
|
|
|
|
if (inp->inp_flags & INP_RECVOPTS) {
|
|
|
|
*mp = sbcreatecontrol((caddr_t) opts_deleted_above,
|
|
|
|
sizeof(struct in_addr), IP_RECVOPTS, IPPROTO_IP);
|
|
|
|
if (*mp)
|
|
|
|
mp = &(*mp)->m_next;
|
|
|
|
}
|
|
|
|
/* ip_srcroute doesn't do what we want here, need to fix */
|
|
|
|
if (inp->inp_flags & INP_RECVRETOPTS) {
|
2004-09-15 20:13:26 +00:00
|
|
|
*mp = sbcreatecontrol((caddr_t) ip_srcroute(m),
|
1996-11-11 04:56:32 +00:00
|
|
|
sizeof(struct in_addr), IP_RECVRETOPTS, IPPROTO_IP);
|
|
|
|
if (*mp)
|
|
|
|
mp = &(*mp)->m_next;
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
if (inp->inp_flags & INP_RECVIF) {
|
1997-11-05 02:51:32 +00:00
|
|
|
struct ifnet *ifp;
|
|
|
|
struct sdlbuf {
|
|
|
|
struct sockaddr_dl sdl;
|
|
|
|
u_char pad[32];
|
|
|
|
} sdlbuf;
|
|
|
|
struct sockaddr_dl *sdp;
|
|
|
|
struct sockaddr_dl *sdl2 = &sdlbuf.sdl;
|
|
|
|
|
|
|
|
if (((ifp = m->m_pkthdr.rcvif))
|
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
|
|
|
&& ( ifp->if_index && (ifp->if_index <= V_if_index))) {
|
2005-11-11 16:04:59 +00:00
|
|
|
sdp = (struct sockaddr_dl *)ifp->if_addr->ifa_addr;
|
1997-11-05 02:51:32 +00:00
|
|
|
/*
|
|
|
|
* Change our mind and don't try copy.
|
|
|
|
*/
|
|
|
|
if ((sdp->sdl_family != AF_LINK)
|
|
|
|
|| (sdp->sdl_len > sizeof(sdlbuf))) {
|
|
|
|
goto makedummy;
|
|
|
|
}
|
|
|
|
bcopy(sdp, sdl2, sdp->sdl_len);
|
|
|
|
} else {
|
|
|
|
makedummy:
|
|
|
|
sdl2->sdl_len
|
|
|
|
= offsetof(struct sockaddr_dl, sdl_data[0]);
|
|
|
|
sdl2->sdl_family = AF_LINK;
|
|
|
|
sdl2->sdl_index = 0;
|
|
|
|
sdl2->sdl_nlen = sdl2->sdl_alen = sdl2->sdl_slen = 0;
|
|
|
|
}
|
|
|
|
*mp = sbcreatecontrol((caddr_t) sdl2, sdl2->sdl_len,
|
1996-11-11 04:56:32 +00:00
|
|
|
IP_RECVIF, IPPROTO_IP);
|
|
|
|
if (*mp)
|
|
|
|
mp = &(*mp)->m_next;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2002-06-23 20:48:26 +00:00
|
|
|
/*
|
2007-03-25 21:49:50 +00:00
|
|
|
* XXXRW: Multicast routing code in ip_mroute.c is generally MPSAFE, but the
|
|
|
|
* ip_rsvp and ip_rsvp_on variables need to be interlocked with rsvp_on
|
|
|
|
* locking. This code remains in ip_input.c as ip_mroute.c is optionally
|
|
|
|
* compiled.
|
2002-06-23 20:48:26 +00:00
|
|
|
*/
|
1994-09-06 22:42:31 +00:00
|
|
|
int
|
|
|
|
ip_rsvp_init(struct socket *so)
|
|
|
|
{
|
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
|
|
|
|
1994-09-06 22:42:31 +00:00
|
|
|
if (so->so_type != SOCK_RAW ||
|
|
|
|
so->so_proto->pr_protocol != IPPROTO_RSVP)
|
Massive cleanup of the ip_mroute code.
No functional changes, but:
+ the mrouting module now should behave the same as the compiled-in
version (it did not before, some of the rsvp code was not loaded
properly);
+ netinet/ip_mroute.c is now truly optional;
+ removed some redundant/unused code;
+ changed many instances of '0' to NULL and INADDR_ANY as appropriate;
+ removed several static variables to make the code more SMP-friendly;
+ fixed some minor bugs in the mrouting code (mostly, incorrect return
values from functions).
This commit is also a prerequisite to the addition of support for PIM,
which i would like to put in before DP2 (it does not change any of
the existing APIs, anyways).
Note, in the process we found out that some device drivers fail to
properly handle changes in IFF_ALLMULTI, leading to interesting
behaviour when a multicast router is started. This bug is not
corrected by this commit, and will be fixed with a separate commit.
Detailed changes:
--------------------
netinet/ip_mroute.c all the above.
conf/files make ip_mroute.c optional
net/route.c fix mrt_ioctl hook
netinet/ip_input.c fix ip_mforward hook, move rsvp_input() here
together with other rsvp code, and a couple
of indentation fixes.
netinet/ip_output.c fix ip_mforward and ip_mcast_src hooks
netinet/ip_var.h rsvp function hooks
netinet/raw_ip.c hooks for mrouting and rsvp functions, plus
interface cleanup.
netinet/ip_mroute.h remove an unused and optional field from a struct
Most of the code is from Pavlin Radoslavov and the XORP project
Reviewed by: sam
MFC after: 1 week
2002-11-15 22:53:53 +00:00
|
|
|
return EOPNOTSUPP;
|
1994-09-06 22:42:31 +00:00
|
|
|
|
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 (V_ip_rsvpd != NULL)
|
Massive cleanup of the ip_mroute code.
No functional changes, but:
+ the mrouting module now should behave the same as the compiled-in
version (it did not before, some of the rsvp code was not loaded
properly);
+ netinet/ip_mroute.c is now truly optional;
+ removed some redundant/unused code;
+ changed many instances of '0' to NULL and INADDR_ANY as appropriate;
+ removed several static variables to make the code more SMP-friendly;
+ fixed some minor bugs in the mrouting code (mostly, incorrect return
values from functions).
This commit is also a prerequisite to the addition of support for PIM,
which i would like to put in before DP2 (it does not change any of
the existing APIs, anyways).
Note, in the process we found out that some device drivers fail to
properly handle changes in IFF_ALLMULTI, leading to interesting
behaviour when a multicast router is started. This bug is not
corrected by this commit, and will be fixed with a separate commit.
Detailed changes:
--------------------
netinet/ip_mroute.c all the above.
conf/files make ip_mroute.c optional
net/route.c fix mrt_ioctl hook
netinet/ip_input.c fix ip_mforward hook, move rsvp_input() here
together with other rsvp code, and a couple
of indentation fixes.
netinet/ip_output.c fix ip_mforward and ip_mcast_src hooks
netinet/ip_var.h rsvp function hooks
netinet/raw_ip.c hooks for mrouting and rsvp functions, plus
interface cleanup.
netinet/ip_mroute.h remove an unused and optional field from a struct
Most of the code is from Pavlin Radoslavov and the XORP project
Reviewed by: sam
MFC after: 1 week
2002-11-15 22:53:53 +00:00
|
|
|
return EADDRINUSE;
|
1994-09-06 22:42:31 +00:00
|
|
|
|
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
|
|
|
V_ip_rsvpd = so;
|
1995-06-13 17:51:16 +00:00
|
|
|
/*
|
|
|
|
* This may seem silly, but we need to be sure we don't over-increment
|
|
|
|
* the RSVP counter, in case something slips up.
|
|
|
|
*/
|
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 (!V_ip_rsvp_on) {
|
|
|
|
V_ip_rsvp_on = 1;
|
|
|
|
V_rsvp_on++;
|
1995-06-13 17:51:16 +00:00
|
|
|
}
|
1994-09-06 22:42:31 +00:00
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
int
|
|
|
|
ip_rsvp_done(void)
|
|
|
|
{
|
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
|
|
|
|
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
|
|
|
V_ip_rsvpd = NULL;
|
1995-06-13 17:51:16 +00:00
|
|
|
/*
|
|
|
|
* This may seem silly, but we need to be sure we don't over-decrement
|
|
|
|
* the RSVP counter, in case something slips up.
|
|
|
|
*/
|
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 (V_ip_rsvp_on) {
|
|
|
|
V_ip_rsvp_on = 0;
|
|
|
|
V_rsvp_on--;
|
1995-06-13 17:51:16 +00:00
|
|
|
}
|
1994-09-06 22:42:31 +00:00
|
|
|
return 0;
|
|
|
|
}
|
Massive cleanup of the ip_mroute code.
No functional changes, but:
+ the mrouting module now should behave the same as the compiled-in
version (it did not before, some of the rsvp code was not loaded
properly);
+ netinet/ip_mroute.c is now truly optional;
+ removed some redundant/unused code;
+ changed many instances of '0' to NULL and INADDR_ANY as appropriate;
+ removed several static variables to make the code more SMP-friendly;
+ fixed some minor bugs in the mrouting code (mostly, incorrect return
values from functions).
This commit is also a prerequisite to the addition of support for PIM,
which i would like to put in before DP2 (it does not change any of
the existing APIs, anyways).
Note, in the process we found out that some device drivers fail to
properly handle changes in IFF_ALLMULTI, leading to interesting
behaviour when a multicast router is started. This bug is not
corrected by this commit, and will be fixed with a separate commit.
Detailed changes:
--------------------
netinet/ip_mroute.c all the above.
conf/files make ip_mroute.c optional
net/route.c fix mrt_ioctl hook
netinet/ip_input.c fix ip_mforward hook, move rsvp_input() here
together with other rsvp code, and a couple
of indentation fixes.
netinet/ip_output.c fix ip_mforward and ip_mcast_src hooks
netinet/ip_var.h rsvp function hooks
netinet/raw_ip.c hooks for mrouting and rsvp functions, plus
interface cleanup.
netinet/ip_mroute.h remove an unused and optional field from a struct
Most of the code is from Pavlin Radoslavov and the XORP project
Reviewed by: sam
MFC after: 1 week
2002-11-15 22:53:53 +00:00
|
|
|
|
|
|
|
void
|
|
|
|
rsvp_input(struct mbuf *m, int off) /* XXX must fixup manually */
|
|
|
|
{
|
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
|
|
|
|
Massive cleanup of the ip_mroute code.
No functional changes, but:
+ the mrouting module now should behave the same as the compiled-in
version (it did not before, some of the rsvp code was not loaded
properly);
+ netinet/ip_mroute.c is now truly optional;
+ removed some redundant/unused code;
+ changed many instances of '0' to NULL and INADDR_ANY as appropriate;
+ removed several static variables to make the code more SMP-friendly;
+ fixed some minor bugs in the mrouting code (mostly, incorrect return
values from functions).
This commit is also a prerequisite to the addition of support for PIM,
which i would like to put in before DP2 (it does not change any of
the existing APIs, anyways).
Note, in the process we found out that some device drivers fail to
properly handle changes in IFF_ALLMULTI, leading to interesting
behaviour when a multicast router is started. This bug is not
corrected by this commit, and will be fixed with a separate commit.
Detailed changes:
--------------------
netinet/ip_mroute.c all the above.
conf/files make ip_mroute.c optional
net/route.c fix mrt_ioctl hook
netinet/ip_input.c fix ip_mforward hook, move rsvp_input() here
together with other rsvp code, and a couple
of indentation fixes.
netinet/ip_output.c fix ip_mforward and ip_mcast_src hooks
netinet/ip_var.h rsvp function hooks
netinet/raw_ip.c hooks for mrouting and rsvp functions, plus
interface cleanup.
netinet/ip_mroute.h remove an unused and optional field from a struct
Most of the code is from Pavlin Radoslavov and the XORP project
Reviewed by: sam
MFC after: 1 week
2002-11-15 22:53:53 +00:00
|
|
|
if (rsvp_input_p) { /* call the real one if loaded */
|
|
|
|
rsvp_input_p(m, off);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Can still get packets with rsvp_on = 0 if there is a local member
|
|
|
|
* of the group to which the RSVP packet is addressed. But in this
|
|
|
|
* case we want to throw the packet away.
|
|
|
|
*/
|
|
|
|
|
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 (!V_rsvp_on) {
|
Massive cleanup of the ip_mroute code.
No functional changes, but:
+ the mrouting module now should behave the same as the compiled-in
version (it did not before, some of the rsvp code was not loaded
properly);
+ netinet/ip_mroute.c is now truly optional;
+ removed some redundant/unused code;
+ changed many instances of '0' to NULL and INADDR_ANY as appropriate;
+ removed several static variables to make the code more SMP-friendly;
+ fixed some minor bugs in the mrouting code (mostly, incorrect return
values from functions).
This commit is also a prerequisite to the addition of support for PIM,
which i would like to put in before DP2 (it does not change any of
the existing APIs, anyways).
Note, in the process we found out that some device drivers fail to
properly handle changes in IFF_ALLMULTI, leading to interesting
behaviour when a multicast router is started. This bug is not
corrected by this commit, and will be fixed with a separate commit.
Detailed changes:
--------------------
netinet/ip_mroute.c all the above.
conf/files make ip_mroute.c optional
net/route.c fix mrt_ioctl hook
netinet/ip_input.c fix ip_mforward hook, move rsvp_input() here
together with other rsvp code, and a couple
of indentation fixes.
netinet/ip_output.c fix ip_mforward and ip_mcast_src hooks
netinet/ip_var.h rsvp function hooks
netinet/raw_ip.c hooks for mrouting and rsvp functions, plus
interface cleanup.
netinet/ip_mroute.h remove an unused and optional field from a struct
Most of the code is from Pavlin Radoslavov and the XORP project
Reviewed by: sam
MFC after: 1 week
2002-11-15 22:53:53 +00:00
|
|
|
m_freem(m);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
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 (V_ip_rsvpd != NULL) {
|
Massive cleanup of the ip_mroute code.
No functional changes, but:
+ the mrouting module now should behave the same as the compiled-in
version (it did not before, some of the rsvp code was not loaded
properly);
+ netinet/ip_mroute.c is now truly optional;
+ removed some redundant/unused code;
+ changed many instances of '0' to NULL and INADDR_ANY as appropriate;
+ removed several static variables to make the code more SMP-friendly;
+ fixed some minor bugs in the mrouting code (mostly, incorrect return
values from functions).
This commit is also a prerequisite to the addition of support for PIM,
which i would like to put in before DP2 (it does not change any of
the existing APIs, anyways).
Note, in the process we found out that some device drivers fail to
properly handle changes in IFF_ALLMULTI, leading to interesting
behaviour when a multicast router is started. This bug is not
corrected by this commit, and will be fixed with a separate commit.
Detailed changes:
--------------------
netinet/ip_mroute.c all the above.
conf/files make ip_mroute.c optional
net/route.c fix mrt_ioctl hook
netinet/ip_input.c fix ip_mforward hook, move rsvp_input() here
together with other rsvp code, and a couple
of indentation fixes.
netinet/ip_output.c fix ip_mforward and ip_mcast_src hooks
netinet/ip_var.h rsvp function hooks
netinet/raw_ip.c hooks for mrouting and rsvp functions, plus
interface cleanup.
netinet/ip_mroute.h remove an unused and optional field from a struct
Most of the code is from Pavlin Radoslavov and the XORP project
Reviewed by: sam
MFC after: 1 week
2002-11-15 22:53:53 +00:00
|
|
|
rip_input(m, off);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
/* Drop the packet */
|
|
|
|
m_freem(m);
|
|
|
|
}
|