2002-10-16 02:10:08 +00:00
|
|
|
/* $FreeBSD$ */
|
|
|
|
/* $OpenBSD: ip_esp.c,v 1.69 2001/06/26 06:18:59 angelos Exp $ */
|
2005-01-07 01:45:51 +00:00
|
|
|
/*-
|
2002-10-16 02:10:08 +00:00
|
|
|
* The authors of this code are John Ioannidis (ji@tla.org),
|
|
|
|
* Angelos D. Keromytis (kermit@csd.uch.gr) and
|
|
|
|
* Niels Provos (provos@physnet.uni-hamburg.de).
|
|
|
|
*
|
|
|
|
* The original version of this code was written by John Ioannidis
|
|
|
|
* for BSD/OS in Athens, Greece, in November 1995.
|
|
|
|
*
|
|
|
|
* Ported to OpenBSD and NetBSD, with additional transforms, in December 1996,
|
|
|
|
* by Angelos D. Keromytis.
|
|
|
|
*
|
|
|
|
* Additional transforms and features in 1997 and 1998 by Angelos D. Keromytis
|
|
|
|
* and Niels Provos.
|
|
|
|
*
|
|
|
|
* Additional features in 1999 by Angelos D. Keromytis.
|
|
|
|
*
|
|
|
|
* Copyright (C) 1995, 1996, 1997, 1998, 1999 by John Ioannidis,
|
|
|
|
* Angelos D. Keromytis and Niels Provos.
|
|
|
|
* Copyright (c) 2001 Angelos D. Keromytis.
|
|
|
|
*
|
|
|
|
* Permission to use, copy, and modify this software with or without fee
|
|
|
|
* is hereby granted, provided that this entire notice is included in
|
|
|
|
* all copies of any software which is or includes a copy or
|
|
|
|
* modification of this software.
|
|
|
|
* You may use this code under the GNU public license if you so wish. Please
|
|
|
|
* contribute changes back to the authors under this freer than GPL license
|
|
|
|
* so that we may further the use of strong encryption without limitations to
|
|
|
|
* all.
|
|
|
|
*
|
|
|
|
* THIS SOFTWARE IS BEING PROVIDED "AS IS", WITHOUT ANY EXPRESS OR
|
|
|
|
* IMPLIED WARRANTY. IN PARTICULAR, NONE OF THE AUTHORS MAKES ANY
|
|
|
|
* REPRESENTATION OR WARRANTY OF ANY KIND CONCERNING THE
|
|
|
|
* MERCHANTABILITY OF THIS SOFTWARE OR ITS FITNESS FOR ANY PARTICULAR
|
|
|
|
* PURPOSE.
|
|
|
|
*/
|
|
|
|
#include "opt_inet.h"
|
|
|
|
#include "opt_inet6.h"
|
|
|
|
|
|
|
|
#include <sys/param.h>
|
|
|
|
#include <sys/systm.h>
|
|
|
|
#include <sys/mbuf.h>
|
|
|
|
#include <sys/socket.h>
|
|
|
|
#include <sys/syslog.h>
|
|
|
|
#include <sys/kernel.h>
|
|
|
|
#include <sys/random.h>
|
|
|
|
#include <sys/sysctl.h>
|
|
|
|
|
|
|
|
#include <net/if.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 <net/vnet.h>
|
2002-10-16 02:10:08 +00:00
|
|
|
|
|
|
|
#include <netinet/in.h>
|
|
|
|
#include <netinet/in_systm.h>
|
|
|
|
#include <netinet/ip.h>
|
|
|
|
#include <netinet/ip_ecn.h>
|
|
|
|
#include <netinet/ip6.h>
|
|
|
|
|
|
|
|
#include <net/route.h>
|
|
|
|
#include <netipsec/ipsec.h>
|
|
|
|
#include <netipsec/ah.h>
|
|
|
|
#include <netipsec/ah_var.h>
|
|
|
|
#include <netipsec/esp.h>
|
|
|
|
#include <netipsec/esp_var.h>
|
|
|
|
#include <netipsec/xform.h>
|
|
|
|
|
|
|
|
#ifdef INET6
|
|
|
|
#include <netinet6/ip6_var.h>
|
|
|
|
#include <netipsec/ipsec6.h>
|
|
|
|
#include <netinet6/ip6_ecn.h>
|
|
|
|
#endif
|
|
|
|
|
|
|
|
#include <netipsec/key.h>
|
|
|
|
#include <netipsec/key_debug.h>
|
|
|
|
|
|
|
|
#include <opencrypto/cryptodev.h>
|
|
|
|
#include <opencrypto/xform.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
|
|
|
VNET_DEFINE(int, esp_enable) = 1;
|
|
|
|
VNET_DEFINE(struct espstat, espstat);
|
2002-10-16 02:10:08 +00:00
|
|
|
|
|
|
|
SYSCTL_DECL(_net_inet_esp);
|
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_esp, OID_AUTO,
|
|
|
|
esp_enable, CTLFLAG_RW, &VNET_NAME(esp_enable), 0, "");
|
|
|
|
SYSCTL_VNET_STRUCT(_net_inet_esp, IPSECCTL_STATS,
|
|
|
|
stats, CTLFLAG_RD, &VNET_NAME(espstat), espstat, "");
|
|
|
|
|
2010-11-22 19:32:54 +00:00
|
|
|
static VNET_DEFINE(int, esp_max_ivlen); /* max iv length over all algorithms */
|
2009-07-16 21:13:04 +00:00
|
|
|
#define V_esp_max_ivlen VNET(esp_max_ivlen)
|
2002-10-16 02:10:08 +00:00
|
|
|
|
|
|
|
static int esp_input_cb(struct cryptop *op);
|
|
|
|
static int esp_output_cb(struct cryptop *crp);
|
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
|
|
|
|
2002-10-16 02:10:08 +00:00
|
|
|
/*
|
|
|
|
* NB: this is public for use by the PF_KEY support.
|
|
|
|
* NB: if you add support here; be sure to add code to esp_attach below!
|
|
|
|
*/
|
|
|
|
struct enc_xform *
|
|
|
|
esp_algorithm_lookup(int alg)
|
|
|
|
{
|
|
|
|
if (alg >= ESP_ALG_MAX)
|
|
|
|
return NULL;
|
|
|
|
switch (alg) {
|
|
|
|
case SADB_EALG_DESCBC:
|
|
|
|
return &enc_xform_des;
|
|
|
|
case SADB_EALG_3DESCBC:
|
|
|
|
return &enc_xform_3des;
|
|
|
|
case SADB_X_EALG_AES:
|
|
|
|
return &enc_xform_rijndael128;
|
|
|
|
case SADB_X_EALG_BLOWFISHCBC:
|
|
|
|
return &enc_xform_blf;
|
|
|
|
case SADB_X_EALG_CAST128CBC:
|
|
|
|
return &enc_xform_cast5;
|
|
|
|
case SADB_X_EALG_SKIPJACK:
|
|
|
|
return &enc_xform_skipjack;
|
|
|
|
case SADB_EALG_NULL:
|
|
|
|
return &enc_xform_null;
|
2007-05-09 19:37:02 +00:00
|
|
|
case SADB_X_EALG_CAMELLIACBC:
|
|
|
|
return &enc_xform_camellia;
|
2002-10-16 02:10:08 +00:00
|
|
|
}
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
|
|
|
size_t
|
|
|
|
esp_hdrsiz(struct secasvar *sav)
|
|
|
|
{
|
|
|
|
size_t size;
|
|
|
|
|
|
|
|
if (sav != NULL) {
|
|
|
|
/*XXX not right for null algorithm--does it matter??*/
|
2003-09-29 22:57:43 +00:00
|
|
|
IPSEC_ASSERT(sav->tdb_encalgxform != NULL,
|
|
|
|
("SA with null xform"));
|
2002-10-16 02:10:08 +00:00
|
|
|
if (sav->flags & SADB_X_EXT_OLD)
|
|
|
|
size = sizeof (struct esp);
|
|
|
|
else
|
|
|
|
size = sizeof (struct newesp);
|
|
|
|
size += sav->tdb_encalgxform->blocksize + 9;
|
|
|
|
/*XXX need alg check???*/
|
|
|
|
if (sav->tdb_authalgxform != NULL && sav->replay)
|
|
|
|
size += ah_hdrsiz(sav);
|
|
|
|
} else {
|
|
|
|
/*
|
|
|
|
* base header size
|
|
|
|
* + max iv length for CBC mode
|
|
|
|
* + max pad length
|
|
|
|
* + sizeof (pad length field)
|
|
|
|
* + sizeof (next header field)
|
|
|
|
* + max icv supported.
|
|
|
|
*/
|
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
|
|
|
size = sizeof (struct newesp) + V_esp_max_ivlen + 9 + 16;
|
2002-10-16 02:10:08 +00:00
|
|
|
}
|
|
|
|
return size;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* esp_init() is called when an SPI is being set up.
|
|
|
|
*/
|
|
|
|
static int
|
|
|
|
esp_init(struct secasvar *sav, struct xformsw *xsp)
|
|
|
|
{
|
|
|
|
struct enc_xform *txform;
|
|
|
|
struct cryptoini cria, crie;
|
|
|
|
int keylen;
|
|
|
|
int error;
|
|
|
|
|
|
|
|
txform = esp_algorithm_lookup(sav->alg_enc);
|
|
|
|
if (txform == NULL) {
|
2003-09-29 22:57:43 +00:00
|
|
|
DPRINTF(("%s: unsupported encryption algorithm %d\n",
|
|
|
|
__func__, sav->alg_enc));
|
2002-10-16 02:10:08 +00:00
|
|
|
return EINVAL;
|
|
|
|
}
|
|
|
|
if (sav->key_enc == NULL) {
|
2003-09-29 22:57:43 +00:00
|
|
|
DPRINTF(("%s: no encoding key for %s algorithm\n",
|
|
|
|
__func__, txform->name));
|
2002-10-16 02:10:08 +00:00
|
|
|
return EINVAL;
|
|
|
|
}
|
|
|
|
if ((sav->flags&(SADB_X_EXT_OLD|SADB_X_EXT_IV4B)) == SADB_X_EXT_IV4B) {
|
2003-09-29 22:57:43 +00:00
|
|
|
DPRINTF(("%s: 4-byte IV not supported with protocol\n",
|
|
|
|
__func__));
|
2002-10-16 02:10:08 +00:00
|
|
|
return EINVAL;
|
|
|
|
}
|
|
|
|
keylen = _KEYLEN(sav->key_enc);
|
|
|
|
if (txform->minkey > keylen || keylen > txform->maxkey) {
|
2003-09-29 22:57:43 +00:00
|
|
|
DPRINTF(("%s: invalid key length %u, must be in the range "
|
|
|
|
"[%u..%u] for algorithm %s\n", __func__,
|
2002-10-16 02:10:08 +00:00
|
|
|
keylen, txform->minkey, txform->maxkey,
|
|
|
|
txform->name));
|
|
|
|
return EINVAL;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* NB: The null xform needs a non-zero blocksize to keep the
|
|
|
|
* crypto code happy but if we use it to set ivlen then
|
|
|
|
* the ESP header will be processed incorrectly. The
|
|
|
|
* compromise is to force it to zero here.
|
|
|
|
*/
|
|
|
|
sav->ivlen = (txform == &enc_xform_null ? 0 : txform->blocksize);
|
2003-02-19 05:47:46 +00:00
|
|
|
sav->iv = (caddr_t) malloc(sav->ivlen, M_XDATA, M_WAITOK);
|
2002-10-16 02:10:08 +00:00
|
|
|
if (sav->iv == NULL) {
|
2003-09-29 22:57:43 +00:00
|
|
|
DPRINTF(("%s: no memory for IV\n", __func__));
|
2002-10-16 02:10:08 +00:00
|
|
|
return EINVAL;
|
|
|
|
}
|
|
|
|
key_randomfill(sav->iv, sav->ivlen); /*XXX*/
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Setup AH-related state.
|
|
|
|
*/
|
|
|
|
if (sav->alg_auth != 0) {
|
|
|
|
error = ah_init0(sav, xsp, &cria);
|
|
|
|
if (error)
|
|
|
|
return error;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* NB: override anything set in ah_init0 */
|
|
|
|
sav->tdb_xform = xsp;
|
|
|
|
sav->tdb_encalgxform = txform;
|
|
|
|
|
|
|
|
/* Initialize crypto session. */
|
|
|
|
bzero(&crie, sizeof (crie));
|
|
|
|
crie.cri_alg = sav->tdb_encalgxform->type;
|
|
|
|
crie.cri_klen = _KEYBITS(sav->key_enc);
|
2006-03-25 13:38:52 +00:00
|
|
|
crie.cri_key = sav->key_enc->key_data;
|
2002-10-16 02:10:08 +00:00
|
|
|
/* XXX Rounds ? */
|
|
|
|
|
|
|
|
if (sav->tdb_authalgxform && sav->tdb_encalgxform) {
|
|
|
|
/* init both auth & enc */
|
|
|
|
crie.cri_next = &cria;
|
|
|
|
error = crypto_newsession(&sav->tdb_cryptoid,
|
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
|
|
|
&crie, V_crypto_support);
|
2002-10-16 02:10:08 +00:00
|
|
|
} else if (sav->tdb_encalgxform) {
|
|
|
|
error = crypto_newsession(&sav->tdb_cryptoid,
|
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
|
|
|
&crie, V_crypto_support);
|
2002-10-16 02:10:08 +00:00
|
|
|
} else if (sav->tdb_authalgxform) {
|
|
|
|
error = crypto_newsession(&sav->tdb_cryptoid,
|
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
|
|
|
&cria, V_crypto_support);
|
2002-10-16 02:10:08 +00:00
|
|
|
} else {
|
|
|
|
/* XXX cannot happen? */
|
2003-09-29 22:57:43 +00:00
|
|
|
DPRINTF(("%s: no encoding OR authentication xform!\n",
|
|
|
|
__func__));
|
2002-10-16 02:10:08 +00:00
|
|
|
error = EINVAL;
|
|
|
|
}
|
|
|
|
return error;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Paranoia.
|
|
|
|
*/
|
|
|
|
static int
|
|
|
|
esp_zeroize(struct secasvar *sav)
|
|
|
|
{
|
|
|
|
/* NB: ah_zerorize free's the crypto session state */
|
|
|
|
int error = ah_zeroize(sav);
|
|
|
|
|
|
|
|
if (sav->key_enc)
|
2006-03-25 13:38:52 +00:00
|
|
|
bzero(sav->key_enc->key_data, _KEYLEN(sav->key_enc));
|
2003-09-29 22:57:43 +00:00
|
|
|
if (sav->iv) {
|
|
|
|
free(sav->iv, M_XDATA);
|
|
|
|
sav->iv = NULL;
|
|
|
|
}
|
2002-10-16 02:10:08 +00:00
|
|
|
sav->tdb_encalgxform = NULL;
|
|
|
|
sav->tdb_xform = NULL;
|
|
|
|
return error;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* ESP input processing, called (eventually) through the protocol switch.
|
|
|
|
*/
|
|
|
|
static int
|
|
|
|
esp_input(struct mbuf *m, struct secasvar *sav, int skip, int protoff)
|
|
|
|
{
|
|
|
|
struct auth_hash *esph;
|
|
|
|
struct enc_xform *espx;
|
|
|
|
struct tdb_ident *tdbi;
|
|
|
|
struct tdb_crypto *tc;
|
|
|
|
int plen, alen, hlen;
|
|
|
|
struct m_tag *mtag;
|
|
|
|
struct newesp *esp;
|
|
|
|
|
|
|
|
struct cryptodesc *crde;
|
|
|
|
struct cryptop *crp;
|
|
|
|
|
2003-09-29 22:57:43 +00:00
|
|
|
IPSEC_ASSERT(sav != NULL, ("null SA"));
|
|
|
|
IPSEC_ASSERT(sav->tdb_encalgxform != NULL, ("null encoding xform"));
|
2009-10-01 15:33:53 +00:00
|
|
|
|
|
|
|
/* Valid IP Packet length ? */
|
|
|
|
if ( (skip&3) || (m->m_pkthdr.len&3) ){
|
|
|
|
DPRINTF(("%s: misaligned packet, skip %u pkt len %u",
|
|
|
|
__func__, skip, m->m_pkthdr.len));
|
|
|
|
V_espstat.esps_badilen++;
|
|
|
|
m_freem(m);
|
|
|
|
return EINVAL;
|
|
|
|
}
|
2002-10-16 02:10:08 +00:00
|
|
|
|
|
|
|
/* XXX don't pullup, just copy header */
|
|
|
|
IP6_EXTHDR_GET(esp, struct newesp *, m, skip, sizeof (struct newesp));
|
|
|
|
|
|
|
|
esph = sav->tdb_authalgxform;
|
|
|
|
espx = sav->tdb_encalgxform;
|
|
|
|
|
|
|
|
/* Determine the ESP header length */
|
|
|
|
if (sav->flags & SADB_X_EXT_OLD)
|
|
|
|
hlen = sizeof (struct esp) + sav->ivlen;
|
|
|
|
else
|
|
|
|
hlen = sizeof (struct newesp) + sav->ivlen;
|
|
|
|
/* Authenticator hash size */
|
2011-02-18 09:40:13 +00:00
|
|
|
if (esph != NULL) {
|
|
|
|
switch (esph->type) {
|
|
|
|
case CRYPTO_SHA2_256_HMAC:
|
|
|
|
case CRYPTO_SHA2_384_HMAC:
|
|
|
|
case CRYPTO_SHA2_512_HMAC:
|
|
|
|
alen = esph->hashsize/2;
|
|
|
|
break;
|
|
|
|
default:
|
|
|
|
alen = AH_HMAC_HASHLEN;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}else
|
|
|
|
alen = 0;
|
2002-10-16 02:10:08 +00:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Verify payload length is multiple of encryption algorithm
|
|
|
|
* block size.
|
|
|
|
*
|
|
|
|
* NB: This works for the null algorithm because the blocksize
|
|
|
|
* is 4 and all packets must be 4-byte aligned regardless
|
|
|
|
* of the algorithm.
|
|
|
|
*/
|
|
|
|
plen = m->m_pkthdr.len - (skip + hlen + alen);
|
|
|
|
if ((plen & (espx->blocksize - 1)) || (plen <= 0)) {
|
2003-09-29 22:57:43 +00:00
|
|
|
DPRINTF(("%s: payload of %d octets not a multiple of %d octets,"
|
|
|
|
" SA %s/%08lx\n", __func__,
|
2002-10-16 02:10:08 +00:00
|
|
|
plen, espx->blocksize,
|
|
|
|
ipsec_address(&sav->sah->saidx.dst),
|
|
|
|
(u_long) ntohl(sav->spi)));
|
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_espstat.esps_badilen++;
|
2002-10-16 02:10:08 +00:00
|
|
|
m_freem(m);
|
|
|
|
return EINVAL;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Check sequence number.
|
|
|
|
*/
|
|
|
|
if (esph && sav->replay && !ipsec_chkreplay(ntohl(esp->esp_seq), sav)) {
|
2003-09-29 22:57:43 +00:00
|
|
|
DPRINTF(("%s: packet replay check for %s\n", __func__,
|
2002-10-16 02:10:08 +00:00
|
|
|
ipsec_logsastr(sav))); /*XXX*/
|
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_espstat.esps_replay++;
|
2002-10-16 02:10:08 +00:00
|
|
|
m_freem(m);
|
|
|
|
return ENOBUFS; /*XXX*/
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Update the counters */
|
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_espstat.esps_ibytes += m->m_pkthdr.len - (skip + hlen + alen);
|
2002-10-16 02:10:08 +00:00
|
|
|
|
|
|
|
/* Find out if we've already done crypto */
|
|
|
|
for (mtag = m_tag_find(m, PACKET_TAG_IPSEC_IN_CRYPTO_DONE, NULL);
|
|
|
|
mtag != NULL;
|
|
|
|
mtag = m_tag_find(m, PACKET_TAG_IPSEC_IN_CRYPTO_DONE, mtag)) {
|
|
|
|
tdbi = (struct tdb_ident *) (mtag + 1);
|
|
|
|
if (tdbi->proto == sav->sah->saidx.proto &&
|
|
|
|
tdbi->spi == sav->spi &&
|
|
|
|
!bcmp(&tdbi->dst, &sav->sah->saidx.dst,
|
|
|
|
sizeof(union sockaddr_union)))
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Get crypto descriptors */
|
|
|
|
crp = crypto_getreq(esph && espx ? 2 : 1);
|
|
|
|
if (crp == NULL) {
|
2003-09-29 22:57:43 +00:00
|
|
|
DPRINTF(("%s: failed to acquire crypto descriptors\n",
|
|
|
|
__func__));
|
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_espstat.esps_crypto++;
|
2002-10-16 02:10:08 +00:00
|
|
|
m_freem(m);
|
|
|
|
return ENOBUFS;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Get IPsec-specific opaque pointer */
|
|
|
|
if (esph == NULL || mtag != NULL)
|
|
|
|
tc = (struct tdb_crypto *) malloc(sizeof(struct tdb_crypto),
|
|
|
|
M_XDATA, M_NOWAIT|M_ZERO);
|
|
|
|
else
|
|
|
|
tc = (struct tdb_crypto *) malloc(sizeof(struct tdb_crypto) + alen,
|
|
|
|
M_XDATA, M_NOWAIT|M_ZERO);
|
|
|
|
if (tc == NULL) {
|
|
|
|
crypto_freereq(crp);
|
2003-09-29 22:57:43 +00:00
|
|
|
DPRINTF(("%s: failed to allocate tdb_crypto\n", __func__));
|
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_espstat.esps_crypto++;
|
2002-10-16 02:10:08 +00:00
|
|
|
m_freem(m);
|
|
|
|
return ENOBUFS;
|
|
|
|
}
|
|
|
|
|
|
|
|
tc->tc_ptr = (caddr_t) mtag;
|
|
|
|
|
|
|
|
if (esph) {
|
|
|
|
struct cryptodesc *crda = crp->crp_desc;
|
|
|
|
|
2003-09-29 22:57:43 +00:00
|
|
|
IPSEC_ASSERT(crda != NULL, ("null ah crypto descriptor"));
|
2002-10-16 02:10:08 +00:00
|
|
|
|
|
|
|
/* Authentication descriptor */
|
|
|
|
crda->crd_skip = skip;
|
|
|
|
crda->crd_len = m->m_pkthdr.len - (skip + alen);
|
|
|
|
crda->crd_inject = m->m_pkthdr.len - alen;
|
|
|
|
|
|
|
|
crda->crd_alg = esph->type;
|
2006-03-25 13:38:52 +00:00
|
|
|
crda->crd_key = sav->key_auth->key_data;
|
2002-10-16 02:10:08 +00:00
|
|
|
crda->crd_klen = _KEYBITS(sav->key_auth);
|
|
|
|
|
|
|
|
/* Copy the authenticator */
|
|
|
|
if (mtag == NULL)
|
|
|
|
m_copydata(m, m->m_pkthdr.len - alen, alen,
|
|
|
|
(caddr_t) (tc + 1));
|
|
|
|
|
|
|
|
/* Chain authentication request */
|
|
|
|
crde = crda->crd_next;
|
|
|
|
} else {
|
|
|
|
crde = crp->crp_desc;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Crypto operation descriptor */
|
|
|
|
crp->crp_ilen = m->m_pkthdr.len; /* Total input length */
|
2003-06-30 05:09:32 +00:00
|
|
|
crp->crp_flags = CRYPTO_F_IMBUF | CRYPTO_F_CBIFSYNC;
|
2002-10-16 02:10:08 +00:00
|
|
|
crp->crp_buf = (caddr_t) m;
|
|
|
|
crp->crp_callback = esp_input_cb;
|
|
|
|
crp->crp_sid = sav->tdb_cryptoid;
|
|
|
|
crp->crp_opaque = (caddr_t) tc;
|
|
|
|
|
|
|
|
/* These are passed as-is to the callback */
|
|
|
|
tc->tc_spi = sav->spi;
|
|
|
|
tc->tc_dst = sav->sah->saidx.dst;
|
|
|
|
tc->tc_proto = sav->sah->saidx.proto;
|
|
|
|
tc->tc_protoff = protoff;
|
|
|
|
tc->tc_skip = skip;
|
2011-03-31 15:23:32 +00:00
|
|
|
KEY_ADDREFSA(sav);
|
|
|
|
tc->tc_sav = sav;
|
2002-10-16 02:10:08 +00:00
|
|
|
|
|
|
|
/* Decryption descriptor */
|
|
|
|
if (espx) {
|
2003-09-29 22:57:43 +00:00
|
|
|
IPSEC_ASSERT(crde != NULL, ("null esp crypto descriptor"));
|
2002-10-16 02:10:08 +00:00
|
|
|
crde->crd_skip = skip + hlen;
|
|
|
|
crde->crd_len = m->m_pkthdr.len - (skip + hlen + alen);
|
|
|
|
crde->crd_inject = skip + hlen - sav->ivlen;
|
|
|
|
|
|
|
|
crde->crd_alg = espx->type;
|
2006-03-25 13:38:52 +00:00
|
|
|
crde->crd_key = sav->key_enc->key_data;
|
2002-10-16 02:10:08 +00:00
|
|
|
crde->crd_klen = _KEYBITS(sav->key_enc);
|
|
|
|
/* XXX Rounds ? */
|
|
|
|
}
|
|
|
|
|
|
|
|
if (mtag == NULL)
|
|
|
|
return crypto_dispatch(crp);
|
|
|
|
else
|
|
|
|
return esp_input_cb(crp);
|
|
|
|
}
|
|
|
|
|
|
|
|
#ifdef INET6
|
|
|
|
#define IPSEC_COMMON_INPUT_CB(m, sav, skip, protoff, mtag) do { \
|
|
|
|
if (saidx->dst.sa.sa_family == AF_INET6) { \
|
|
|
|
error = ipsec6_common_input_cb(m, sav, skip, protoff, mtag); \
|
|
|
|
} else { \
|
|
|
|
error = ipsec4_common_input_cb(m, sav, skip, protoff, mtag); \
|
|
|
|
} \
|
|
|
|
} while (0)
|
|
|
|
#else
|
|
|
|
#define IPSEC_COMMON_INPUT_CB(m, sav, skip, protoff, mtag) \
|
|
|
|
(error = ipsec4_common_input_cb(m, sav, skip, protoff, mtag))
|
|
|
|
#endif
|
|
|
|
|
|
|
|
/*
|
|
|
|
* ESP input callback from the crypto driver.
|
|
|
|
*/
|
|
|
|
static int
|
|
|
|
esp_input_cb(struct cryptop *crp)
|
|
|
|
{
|
2011-02-18 09:40:13 +00:00
|
|
|
u_int8_t lastthree[3], aalg[AH_HMAC_MAXHASHLEN];
|
|
|
|
int hlen, skip, protoff, error, alen;
|
2002-10-16 02:10:08 +00:00
|
|
|
struct mbuf *m;
|
|
|
|
struct cryptodesc *crd;
|
|
|
|
struct auth_hash *esph;
|
|
|
|
struct enc_xform *espx;
|
|
|
|
struct tdb_crypto *tc;
|
|
|
|
struct m_tag *mtag;
|
|
|
|
struct secasvar *sav;
|
|
|
|
struct secasindex *saidx;
|
|
|
|
caddr_t ptr;
|
|
|
|
|
|
|
|
crd = crp->crp_desc;
|
2003-09-29 22:57:43 +00:00
|
|
|
IPSEC_ASSERT(crd != NULL, ("null crypto descriptor!"));
|
2002-10-16 02:10:08 +00:00
|
|
|
|
|
|
|
tc = (struct tdb_crypto *) crp->crp_opaque;
|
2003-09-29 22:57:43 +00:00
|
|
|
IPSEC_ASSERT(tc != NULL, ("null opaque crypto data area!"));
|
2002-10-16 02:10:08 +00:00
|
|
|
skip = tc->tc_skip;
|
|
|
|
protoff = tc->tc_protoff;
|
|
|
|
mtag = (struct m_tag *) tc->tc_ptr;
|
|
|
|
m = (struct mbuf *) crp->crp_buf;
|
|
|
|
|
2011-03-31 15:23:32 +00:00
|
|
|
sav = tc->tc_sav;
|
|
|
|
IPSEC_ASSERT(sav != NULL, ("null SA!"));
|
2002-10-16 02:10:08 +00:00
|
|
|
|
|
|
|
saidx = &sav->sah->saidx;
|
2003-09-29 22:57:43 +00:00
|
|
|
IPSEC_ASSERT(saidx->dst.sa.sa_family == AF_INET ||
|
2002-10-16 02:10:08 +00:00
|
|
|
saidx->dst.sa.sa_family == AF_INET6,
|
2003-09-29 22:57:43 +00:00
|
|
|
("unexpected protocol family %u", saidx->dst.sa.sa_family));
|
2002-10-16 02:10:08 +00:00
|
|
|
|
|
|
|
esph = sav->tdb_authalgxform;
|
|
|
|
espx = sav->tdb_encalgxform;
|
|
|
|
|
|
|
|
/* Check for crypto errors */
|
|
|
|
if (crp->crp_etype) {
|
|
|
|
/* Reset the session ID */
|
|
|
|
if (sav->tdb_cryptoid != 0)
|
|
|
|
sav->tdb_cryptoid = crp->crp_sid;
|
|
|
|
|
|
|
|
if (crp->crp_etype == EAGAIN) {
|
2006-03-23 23:26:34 +00:00
|
|
|
error = crypto_dispatch(crp);
|
|
|
|
return error;
|
2002-10-16 02:10:08 +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_espstat.esps_noxform++;
|
2003-09-29 22:57:43 +00:00
|
|
|
DPRINTF(("%s: crypto error %d\n", __func__, crp->crp_etype));
|
2002-10-16 02:10:08 +00:00
|
|
|
error = crp->crp_etype;
|
|
|
|
goto bad;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Shouldn't happen... */
|
|
|
|
if (m == 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
|
|
|
V_espstat.esps_crypto++;
|
2003-09-29 22:57:43 +00:00
|
|
|
DPRINTF(("%s: bogus returned buffer from crypto\n", __func__));
|
2002-10-16 02:10:08 +00:00
|
|
|
error = EINVAL;
|
|
|
|
goto bad;
|
|
|
|
}
|
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_espstat.esps_hist[sav->alg_enc]++;
|
2002-10-16 02:10:08 +00:00
|
|
|
|
|
|
|
/* If authentication was performed, check now. */
|
|
|
|
if (esph != NULL) {
|
2011-02-18 09:40:13 +00:00
|
|
|
switch (esph->type) {
|
|
|
|
case CRYPTO_SHA2_256_HMAC:
|
|
|
|
case CRYPTO_SHA2_384_HMAC:
|
|
|
|
case CRYPTO_SHA2_512_HMAC:
|
|
|
|
alen = esph->hashsize/2;
|
|
|
|
break;
|
|
|
|
default:
|
|
|
|
alen = AH_HMAC_HASHLEN;
|
|
|
|
break;
|
|
|
|
}
|
2002-10-16 02:10:08 +00:00
|
|
|
/*
|
|
|
|
* If we have a tag, it means an IPsec-aware NIC did
|
|
|
|
* the verification for us. Otherwise we need to
|
|
|
|
* check the authentication calculation.
|
|
|
|
*/
|
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_ahstat.ahs_hist[sav->alg_auth]++;
|
2002-10-16 02:10:08 +00:00
|
|
|
if (mtag == NULL) {
|
|
|
|
/* Copy the authenticator from the packet */
|
2011-02-18 09:40:13 +00:00
|
|
|
m_copydata(m, m->m_pkthdr.len - alen,
|
|
|
|
alen, aalg);
|
2002-10-16 02:10:08 +00:00
|
|
|
|
|
|
|
ptr = (caddr_t) (tc + 1);
|
|
|
|
|
|
|
|
/* Verify authenticator */
|
2011-02-18 09:40:13 +00:00
|
|
|
if (bcmp(ptr, aalg, alen) != 0) {
|
2003-09-29 22:57:43 +00:00
|
|
|
DPRINTF(("%s: "
|
2002-10-16 02:10:08 +00:00
|
|
|
"authentication hash mismatch for packet in SA %s/%08lx\n",
|
2003-09-29 22:57:43 +00:00
|
|
|
__func__,
|
2002-10-16 02:10:08 +00:00
|
|
|
ipsec_address(&saidx->dst),
|
|
|
|
(u_long) ntohl(sav->spi)));
|
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_espstat.esps_badauth++;
|
2002-10-16 02:10:08 +00:00
|
|
|
error = EACCES;
|
|
|
|
goto bad;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Remove trailing authenticator */
|
2011-02-18 09:40:13 +00:00
|
|
|
m_adj(m, -alen);
|
2002-10-16 02:10:08 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
/* Release the crypto descriptors */
|
|
|
|
free(tc, M_XDATA), tc = NULL;
|
|
|
|
crypto_freereq(crp), crp = NULL;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Packet is now decrypted.
|
|
|
|
*/
|
|
|
|
m->m_flags |= M_DECRYPTED;
|
|
|
|
|
2006-03-22 16:00:42 +00:00
|
|
|
/*
|
|
|
|
* Update replay sequence number, if appropriate.
|
|
|
|
*/
|
|
|
|
if (sav->replay) {
|
|
|
|
u_int32_t seq;
|
|
|
|
|
|
|
|
m_copydata(m, skip + offsetof(struct newesp, esp_seq),
|
|
|
|
sizeof (seq), (caddr_t) &seq);
|
|
|
|
if (ipsec_updatereplay(ntohl(seq), sav)) {
|
|
|
|
DPRINTF(("%s: packet replay check for %s\n", __func__,
|
|
|
|
ipsec_logsastr(sav)));
|
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_espstat.esps_replay++;
|
2006-03-22 16:00:42 +00:00
|
|
|
error = ENOBUFS;
|
|
|
|
goto bad;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2002-10-16 02:10:08 +00:00
|
|
|
/* Determine the ESP header length */
|
|
|
|
if (sav->flags & SADB_X_EXT_OLD)
|
|
|
|
hlen = sizeof (struct esp) + sav->ivlen;
|
|
|
|
else
|
|
|
|
hlen = sizeof (struct newesp) + sav->ivlen;
|
|
|
|
|
|
|
|
/* Remove the ESP header and IV from the mbuf. */
|
|
|
|
error = m_striphdr(m, skip, hlen);
|
|
|
|
if (error) {
|
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_espstat.esps_hdrops++;
|
2003-09-29 22:57:43 +00:00
|
|
|
DPRINTF(("%s: bad mbuf chain, SA %s/%08lx\n", __func__,
|
2002-10-16 02:10:08 +00:00
|
|
|
ipsec_address(&sav->sah->saidx.dst),
|
|
|
|
(u_long) ntohl(sav->spi)));
|
|
|
|
goto bad;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Save the last three bytes of decrypted data */
|
|
|
|
m_copydata(m, m->m_pkthdr.len - 3, 3, lastthree);
|
|
|
|
|
|
|
|
/* Verify pad length */
|
|
|
|
if (lastthree[1] + 2 > m->m_pkthdr.len - skip) {
|
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_espstat.esps_badilen++;
|
2003-09-29 22:57:43 +00:00
|
|
|
DPRINTF(("%s: invalid padding length %d for %u byte packet "
|
|
|
|
"in SA %s/%08lx\n", __func__,
|
2002-10-16 02:10:08 +00:00
|
|
|
lastthree[1], m->m_pkthdr.len - skip,
|
|
|
|
ipsec_address(&sav->sah->saidx.dst),
|
|
|
|
(u_long) ntohl(sav->spi)));
|
|
|
|
error = EINVAL;
|
|
|
|
goto bad;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Verify correct decryption by checking the last padding bytes */
|
|
|
|
if ((sav->flags & SADB_X_EXT_PMASK) != SADB_X_EXT_PRAND) {
|
|
|
|
if (lastthree[1] != lastthree[0] && lastthree[1] != 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
|
|
|
V_espstat.esps_badenc++;
|
2003-09-29 22:57:43 +00:00
|
|
|
DPRINTF(("%s: decryption failed for packet in "
|
|
|
|
"SA %s/%08lx\n", __func__,
|
2002-10-16 02:10:08 +00:00
|
|
|
ipsec_address(&sav->sah->saidx.dst),
|
|
|
|
(u_long) ntohl(sav->spi)));
|
|
|
|
error = EINVAL;
|
|
|
|
goto bad;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Trim the mbuf chain to remove trailing authenticator and padding */
|
|
|
|
m_adj(m, -(lastthree[1] + 2));
|
|
|
|
|
|
|
|
/* Restore the Next Protocol field */
|
|
|
|
m_copyback(m, protoff, sizeof (u_int8_t), lastthree + 2);
|
|
|
|
|
|
|
|
IPSEC_COMMON_INPUT_CB(m, sav, skip, protoff, mtag);
|
|
|
|
|
|
|
|
KEY_FREESAV(&sav);
|
|
|
|
return error;
|
|
|
|
bad:
|
|
|
|
if (sav)
|
|
|
|
KEY_FREESAV(&sav);
|
|
|
|
if (m != NULL)
|
|
|
|
m_freem(m);
|
|
|
|
if (tc != NULL)
|
|
|
|
free(tc, M_XDATA);
|
|
|
|
if (crp != NULL)
|
|
|
|
crypto_freereq(crp);
|
|
|
|
return error;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* ESP output routine, called by ipsec[46]_process_packet().
|
|
|
|
*/
|
|
|
|
static int
|
|
|
|
esp_output(
|
|
|
|
struct mbuf *m,
|
|
|
|
struct ipsecrequest *isr,
|
|
|
|
struct mbuf **mp,
|
|
|
|
int skip,
|
|
|
|
int protoff
|
|
|
|
)
|
|
|
|
{
|
|
|
|
struct enc_xform *espx;
|
|
|
|
struct auth_hash *esph;
|
|
|
|
int hlen, rlen, plen, padding, blks, alen, i, roff;
|
|
|
|
struct mbuf *mo = (struct mbuf *) NULL;
|
|
|
|
struct tdb_crypto *tc;
|
|
|
|
struct secasvar *sav;
|
|
|
|
struct secasindex *saidx;
|
|
|
|
unsigned char *pad;
|
|
|
|
u_int8_t prot;
|
|
|
|
int error, maxpacketsize;
|
|
|
|
|
|
|
|
struct cryptodesc *crde = NULL, *crda = NULL;
|
|
|
|
struct cryptop *crp;
|
|
|
|
|
|
|
|
sav = isr->sav;
|
2003-09-29 22:57:43 +00:00
|
|
|
IPSEC_ASSERT(sav != NULL, ("null SA"));
|
2002-10-16 02:10:08 +00:00
|
|
|
esph = sav->tdb_authalgxform;
|
|
|
|
espx = sav->tdb_encalgxform;
|
2003-09-29 22:57:43 +00:00
|
|
|
IPSEC_ASSERT(espx != NULL, ("null encoding xform"));
|
2002-10-16 02:10:08 +00:00
|
|
|
|
|
|
|
if (sav->flags & SADB_X_EXT_OLD)
|
|
|
|
hlen = sizeof (struct esp) + sav->ivlen;
|
|
|
|
else
|
|
|
|
hlen = sizeof (struct newesp) + sav->ivlen;
|
|
|
|
|
|
|
|
rlen = m->m_pkthdr.len - skip; /* Raw payload length. */
|
|
|
|
/*
|
|
|
|
* NB: The null encoding transform has a blocksize of 4
|
|
|
|
* so that headers are properly aligned.
|
|
|
|
*/
|
|
|
|
blks = espx->blocksize; /* IV blocksize */
|
|
|
|
|
|
|
|
/* XXX clamp padding length a la KAME??? */
|
|
|
|
padding = ((blks - ((rlen + 2) % blks)) % blks) + 2;
|
|
|
|
plen = rlen + padding; /* Padded payload length. */
|
|
|
|
|
|
|
|
if (esph)
|
2011-02-18 09:40:13 +00:00
|
|
|
switch (esph->type) {
|
|
|
|
case CRYPTO_SHA2_256_HMAC:
|
|
|
|
case CRYPTO_SHA2_384_HMAC:
|
|
|
|
case CRYPTO_SHA2_512_HMAC:
|
|
|
|
alen = esph->hashsize/2;
|
|
|
|
break;
|
|
|
|
default:
|
2002-10-16 02:10:08 +00:00
|
|
|
alen = AH_HMAC_HASHLEN;
|
2011-02-18 09:40:13 +00:00
|
|
|
break;
|
|
|
|
}
|
2002-10-16 02:10:08 +00:00
|
|
|
else
|
|
|
|
alen = 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
|
|
|
V_espstat.esps_output++;
|
2002-10-16 02:10:08 +00:00
|
|
|
|
|
|
|
saidx = &sav->sah->saidx;
|
|
|
|
/* Check for maximum packet size violations. */
|
|
|
|
switch (saidx->dst.sa.sa_family) {
|
|
|
|
#ifdef INET
|
|
|
|
case AF_INET:
|
|
|
|
maxpacketsize = IP_MAXPACKET;
|
|
|
|
break;
|
|
|
|
#endif /* INET */
|
|
|
|
#ifdef INET6
|
|
|
|
case AF_INET6:
|
|
|
|
maxpacketsize = IPV6_MAXPACKET;
|
|
|
|
break;
|
|
|
|
#endif /* INET6 */
|
|
|
|
default:
|
2003-09-29 22:57:43 +00:00
|
|
|
DPRINTF(("%s: unknown/unsupported protocol "
|
|
|
|
"family %d, SA %s/%08lx\n", __func__,
|
2002-10-16 02:10:08 +00:00
|
|
|
saidx->dst.sa.sa_family, ipsec_address(&saidx->dst),
|
|
|
|
(u_long) ntohl(sav->spi)));
|
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_espstat.esps_nopf++;
|
2002-10-16 02:10:08 +00:00
|
|
|
error = EPFNOSUPPORT;
|
|
|
|
goto bad;
|
|
|
|
}
|
|
|
|
if (skip + hlen + rlen + padding + alen > maxpacketsize) {
|
2003-09-29 22:57:43 +00:00
|
|
|
DPRINTF(("%s: packet in SA %s/%08lx got too big "
|
|
|
|
"(len %u, max len %u)\n", __func__,
|
2002-10-16 02:10:08 +00:00
|
|
|
ipsec_address(&saidx->dst), (u_long) ntohl(sav->spi),
|
|
|
|
skip + hlen + rlen + padding + alen, maxpacketsize));
|
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_espstat.esps_toobig++;
|
2002-10-16 02:10:08 +00:00
|
|
|
error = EMSGSIZE;
|
|
|
|
goto bad;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Update the counters. */
|
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_espstat.esps_obytes += m->m_pkthdr.len - skip;
|
2002-10-16 02:10:08 +00:00
|
|
|
|
2006-03-15 21:11:11 +00:00
|
|
|
m = m_unshare(m, M_NOWAIT);
|
2002-10-16 02:10:08 +00:00
|
|
|
if (m == NULL) {
|
2003-09-29 22:57:43 +00:00
|
|
|
DPRINTF(("%s: cannot clone mbuf chain, SA %s/%08lx\n", __func__,
|
2002-10-16 02:10:08 +00:00
|
|
|
ipsec_address(&saidx->dst), (u_long) ntohl(sav->spi)));
|
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_espstat.esps_hdrops++;
|
2002-10-16 02:10:08 +00:00
|
|
|
error = ENOBUFS;
|
|
|
|
goto bad;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Inject ESP header. */
|
|
|
|
mo = m_makespace(m, skip, hlen, &roff);
|
|
|
|
if (mo == NULL) {
|
2003-09-29 22:57:43 +00:00
|
|
|
DPRINTF(("%s: %u byte ESP hdr inject failed for SA %s/%08lx\n",
|
|
|
|
__func__, hlen, ipsec_address(&saidx->dst),
|
2002-10-16 02:10:08 +00:00
|
|
|
(u_long) ntohl(sav->spi)));
|
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_espstat.esps_hdrops++; /* XXX diffs from openbsd */
|
2002-10-16 02:10:08 +00:00
|
|
|
error = ENOBUFS;
|
|
|
|
goto bad;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Initialize ESP header. */
|
|
|
|
bcopy((caddr_t) &sav->spi, mtod(mo, caddr_t) + roff, sizeof(u_int32_t));
|
|
|
|
if (sav->replay) {
|
2006-04-09 19:11:45 +00:00
|
|
|
u_int32_t replay;
|
|
|
|
|
2006-04-10 15:04:36 +00:00
|
|
|
#ifdef REGRESSION
|
2006-04-09 19:11:45 +00:00
|
|
|
/* Emulate replay attack when ipsec_replay is TRUE. */
|
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_ipsec_replay)
|
2006-04-10 15:04:36 +00:00
|
|
|
#endif
|
2006-04-09 19:11:45 +00:00
|
|
|
sav->replay->count++;
|
|
|
|
replay = htonl(sav->replay->count);
|
2002-10-16 02:10:08 +00:00
|
|
|
bcopy((caddr_t) &replay,
|
|
|
|
mtod(mo, caddr_t) + roff + sizeof(u_int32_t),
|
|
|
|
sizeof(u_int32_t));
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Add padding -- better to do it ourselves than use the crypto engine,
|
|
|
|
* although if/when we support compression, we'd have to do that.
|
|
|
|
*/
|
|
|
|
pad = (u_char *) m_pad(m, padding + alen);
|
|
|
|
if (pad == NULL) {
|
2003-09-29 22:57:43 +00:00
|
|
|
DPRINTF(("%s: m_pad failed for SA %s/%08lx\n", __func__,
|
2002-10-16 02:10:08 +00:00
|
|
|
ipsec_address(&saidx->dst), (u_long) ntohl(sav->spi)));
|
|
|
|
m = NULL; /* NB: free'd by m_pad */
|
|
|
|
error = ENOBUFS;
|
|
|
|
goto bad;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Add padding: random, zero, or self-describing.
|
|
|
|
* XXX catch unexpected setting
|
|
|
|
*/
|
|
|
|
switch (sav->flags & SADB_X_EXT_PMASK) {
|
|
|
|
case SADB_X_EXT_PRAND:
|
|
|
|
(void) read_random(pad, padding - 2);
|
|
|
|
break;
|
|
|
|
case SADB_X_EXT_PZERO:
|
|
|
|
bzero(pad, padding - 2);
|
|
|
|
break;
|
|
|
|
case SADB_X_EXT_PSEQ:
|
|
|
|
for (i = 0; i < padding - 2; i++)
|
|
|
|
pad[i] = i+1;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Fix padding length and Next Protocol in padding itself. */
|
|
|
|
pad[padding - 2] = padding - 2;
|
|
|
|
m_copydata(m, protoff, sizeof(u_int8_t), pad + padding - 1);
|
|
|
|
|
|
|
|
/* Fix Next Protocol in IPv4/IPv6 header. */
|
|
|
|
prot = IPPROTO_ESP;
|
|
|
|
m_copyback(m, protoff, sizeof(u_int8_t), (u_char *) &prot);
|
|
|
|
|
|
|
|
/* Get crypto descriptors. */
|
|
|
|
crp = crypto_getreq(esph && espx ? 2 : 1);
|
|
|
|
if (crp == NULL) {
|
2003-09-29 22:57:43 +00:00
|
|
|
DPRINTF(("%s: failed to acquire crypto descriptors\n",
|
|
|
|
__func__));
|
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_espstat.esps_crypto++;
|
2002-10-16 02:10:08 +00:00
|
|
|
error = ENOBUFS;
|
|
|
|
goto bad;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (espx) {
|
|
|
|
crde = crp->crp_desc;
|
|
|
|
crda = crde->crd_next;
|
|
|
|
|
|
|
|
/* Encryption descriptor. */
|
|
|
|
crde->crd_skip = skip + hlen;
|
|
|
|
crde->crd_len = m->m_pkthdr.len - (skip + hlen + alen);
|
|
|
|
crde->crd_flags = CRD_F_ENCRYPT;
|
|
|
|
crde->crd_inject = skip + hlen - sav->ivlen;
|
|
|
|
|
|
|
|
/* Encryption operation. */
|
|
|
|
crde->crd_alg = espx->type;
|
2006-03-25 13:38:52 +00:00
|
|
|
crde->crd_key = sav->key_enc->key_data;
|
2002-10-16 02:10:08 +00:00
|
|
|
crde->crd_klen = _KEYBITS(sav->key_enc);
|
|
|
|
/* XXX Rounds ? */
|
|
|
|
} else
|
|
|
|
crda = crp->crp_desc;
|
|
|
|
|
|
|
|
/* IPsec-specific opaque crypto info. */
|
|
|
|
tc = (struct tdb_crypto *) malloc(sizeof(struct tdb_crypto),
|
|
|
|
M_XDATA, M_NOWAIT|M_ZERO);
|
|
|
|
if (tc == NULL) {
|
|
|
|
crypto_freereq(crp);
|
2003-09-29 22:57:43 +00:00
|
|
|
DPRINTF(("%s: failed to allocate tdb_crypto\n", __func__));
|
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_espstat.esps_crypto++;
|
2002-10-16 02:10:08 +00:00
|
|
|
error = ENOBUFS;
|
|
|
|
goto bad;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Callback parameters */
|
|
|
|
tc->tc_isr = isr;
|
2011-03-31 15:23:32 +00:00
|
|
|
KEY_ADDREFSA(sav);
|
|
|
|
tc->tc_sav = sav;
|
2002-10-16 02:10:08 +00:00
|
|
|
tc->tc_spi = sav->spi;
|
|
|
|
tc->tc_dst = saidx->dst;
|
|
|
|
tc->tc_proto = saidx->proto;
|
|
|
|
|
|
|
|
/* Crypto operation descriptor. */
|
|
|
|
crp->crp_ilen = m->m_pkthdr.len; /* Total input length. */
|
2003-06-30 05:09:32 +00:00
|
|
|
crp->crp_flags = CRYPTO_F_IMBUF | CRYPTO_F_CBIFSYNC;
|
2002-10-16 02:10:08 +00:00
|
|
|
crp->crp_buf = (caddr_t) m;
|
|
|
|
crp->crp_callback = esp_output_cb;
|
|
|
|
crp->crp_opaque = (caddr_t) tc;
|
|
|
|
crp->crp_sid = sav->tdb_cryptoid;
|
|
|
|
|
|
|
|
if (esph) {
|
|
|
|
/* Authentication descriptor. */
|
|
|
|
crda->crd_skip = skip;
|
|
|
|
crda->crd_len = m->m_pkthdr.len - (skip + alen);
|
|
|
|
crda->crd_inject = m->m_pkthdr.len - alen;
|
|
|
|
|
|
|
|
/* Authentication operation. */
|
|
|
|
crda->crd_alg = esph->type;
|
2006-03-25 13:38:52 +00:00
|
|
|
crda->crd_key = sav->key_auth->key_data;
|
2002-10-16 02:10:08 +00:00
|
|
|
crda->crd_klen = _KEYBITS(sav->key_auth);
|
|
|
|
}
|
|
|
|
|
|
|
|
return crypto_dispatch(crp);
|
|
|
|
bad:
|
|
|
|
if (m)
|
|
|
|
m_freem(m);
|
|
|
|
return (error);
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* ESP output callback from the crypto driver.
|
|
|
|
*/
|
|
|
|
static int
|
|
|
|
esp_output_cb(struct cryptop *crp)
|
|
|
|
{
|
|
|
|
struct tdb_crypto *tc;
|
|
|
|
struct ipsecrequest *isr;
|
|
|
|
struct secasvar *sav;
|
|
|
|
struct mbuf *m;
|
2003-09-01 05:35:55 +00:00
|
|
|
int err, error;
|
2002-10-16 02:10:08 +00:00
|
|
|
|
|
|
|
tc = (struct tdb_crypto *) crp->crp_opaque;
|
2003-09-29 22:57:43 +00:00
|
|
|
IPSEC_ASSERT(tc != NULL, ("null opaque data area!"));
|
2002-10-16 02:10:08 +00:00
|
|
|
m = (struct mbuf *) crp->crp_buf;
|
|
|
|
|
|
|
|
isr = tc->tc_isr;
|
2003-09-29 22:57:43 +00:00
|
|
|
IPSECREQUEST_LOCK(isr);
|
2011-03-31 15:23:32 +00:00
|
|
|
sav = tc->tc_sav;
|
|
|
|
/* With the isr lock released SA pointer can be updated. */
|
|
|
|
if (sav != isr->sav) {
|
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_espstat.esps_notdb++;
|
2003-09-29 22:57:43 +00:00
|
|
|
DPRINTF(("%s: SA gone during crypto (SA %s/%08lx proto %u)\n",
|
|
|
|
__func__, ipsec_address(&tc->tc_dst),
|
2002-10-16 02:10:08 +00:00
|
|
|
(u_long) ntohl(tc->tc_spi), tc->tc_proto));
|
|
|
|
error = ENOBUFS; /*XXX*/
|
|
|
|
goto bad;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Check for crypto errors. */
|
|
|
|
if (crp->crp_etype) {
|
|
|
|
/* Reset session ID. */
|
|
|
|
if (sav->tdb_cryptoid != 0)
|
|
|
|
sav->tdb_cryptoid = crp->crp_sid;
|
|
|
|
|
|
|
|
if (crp->crp_etype == EAGAIN) {
|
2003-09-29 22:57:43 +00:00
|
|
|
IPSECREQUEST_UNLOCK(isr);
|
2006-03-23 23:26:34 +00:00
|
|
|
error = crypto_dispatch(crp);
|
|
|
|
return error;
|
2002-10-16 02:10:08 +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_espstat.esps_noxform++;
|
2003-09-29 22:57:43 +00:00
|
|
|
DPRINTF(("%s: crypto error %d\n", __func__, crp->crp_etype));
|
2002-10-16 02:10:08 +00:00
|
|
|
error = crp->crp_etype;
|
|
|
|
goto bad;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Shouldn't happen... */
|
|
|
|
if (m == 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
|
|
|
V_espstat.esps_crypto++;
|
2003-09-29 22:57:43 +00:00
|
|
|
DPRINTF(("%s: bogus returned buffer from crypto\n", __func__));
|
2002-10-16 02:10:08 +00:00
|
|
|
error = EINVAL;
|
|
|
|
goto bad;
|
|
|
|
}
|
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_espstat.esps_hist[sav->alg_enc]++;
|
2002-10-16 02:10:08 +00:00
|
|
|
if (sav->tdb_authalgxform != 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
|
|
|
V_ahstat.ahs_hist[sav->alg_auth]++;
|
2002-10-16 02:10:08 +00:00
|
|
|
|
|
|
|
/* Release crypto descriptors. */
|
|
|
|
free(tc, M_XDATA);
|
|
|
|
crypto_freereq(crp);
|
|
|
|
|
2006-04-10 15:04:36 +00:00
|
|
|
#ifdef REGRESSION
|
2006-04-09 19:11:45 +00:00
|
|
|
/* Emulate man-in-the-middle attack when ipsec_integrity is TRUE. */
|
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_ipsec_integrity) {
|
2011-02-18 09:40:13 +00:00
|
|
|
static unsigned char ipseczeroes[AH_HMAC_MAXHASHLEN];
|
2006-04-09 19:11:45 +00:00
|
|
|
struct auth_hash *esph;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Corrupt HMAC if we want to test integrity verification of
|
|
|
|
* the other side.
|
|
|
|
*/
|
|
|
|
esph = sav->tdb_authalgxform;
|
|
|
|
if (esph != NULL) {
|
2011-02-18 09:40:13 +00:00
|
|
|
int alen;
|
|
|
|
|
|
|
|
switch (esph->type) {
|
|
|
|
case CRYPTO_SHA2_256_HMAC:
|
|
|
|
case CRYPTO_SHA2_384_HMAC:
|
|
|
|
case CRYPTO_SHA2_512_HMAC:
|
|
|
|
alen = esph->hashsize/2;
|
|
|
|
break;
|
|
|
|
default:
|
|
|
|
alen = AH_HMAC_HASHLEN;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
m_copyback(m, m->m_pkthdr.len - alen,
|
|
|
|
alen, ipseczeroes);
|
2006-04-09 19:11:45 +00:00
|
|
|
}
|
|
|
|
}
|
2006-04-10 15:04:36 +00:00
|
|
|
#endif
|
2006-04-09 19:11:45 +00:00
|
|
|
|
2002-10-16 02:10:08 +00:00
|
|
|
/* NB: m is reclaimed by ipsec_process_done. */
|
|
|
|
err = ipsec_process_done(m, isr);
|
|
|
|
KEY_FREESAV(&sav);
|
2003-09-29 22:57:43 +00:00
|
|
|
IPSECREQUEST_UNLOCK(isr);
|
2002-10-16 02:10:08 +00:00
|
|
|
return err;
|
|
|
|
bad:
|
|
|
|
if (sav)
|
|
|
|
KEY_FREESAV(&sav);
|
2003-09-29 22:57:43 +00:00
|
|
|
IPSECREQUEST_UNLOCK(isr);
|
2002-10-16 02:10:08 +00:00
|
|
|
if (m)
|
|
|
|
m_freem(m);
|
|
|
|
free(tc, M_XDATA);
|
|
|
|
crypto_freereq(crp);
|
|
|
|
return error;
|
|
|
|
}
|
|
|
|
|
|
|
|
static struct xformsw esp_xformsw = {
|
|
|
|
XF_ESP, XFT_CONF|XFT_AUTH, "IPsec ESP",
|
|
|
|
esp_init, esp_zeroize, esp_input,
|
|
|
|
esp_output
|
|
|
|
};
|
|
|
|
|
|
|
|
static void
|
|
|
|
esp_attach(void)
|
|
|
|
{
|
|
|
|
#define MAXIV(xform) \
|
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 (xform.blocksize > V_esp_max_ivlen) \
|
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_esp_max_ivlen = xform.blocksize \
|
2002-10-16 02:10:08 +00:00
|
|
|
|
|
|
|
MAXIV(enc_xform_des); /* SADB_EALG_DESCBC */
|
|
|
|
MAXIV(enc_xform_3des); /* SADB_EALG_3DESCBC */
|
|
|
|
MAXIV(enc_xform_rijndael128); /* SADB_X_EALG_AES */
|
|
|
|
MAXIV(enc_xform_blf); /* SADB_X_EALG_BLOWFISHCBC */
|
|
|
|
MAXIV(enc_xform_cast5); /* SADB_X_EALG_CAST128CBC */
|
|
|
|
MAXIV(enc_xform_skipjack); /* SADB_X_EALG_SKIPJACK */
|
|
|
|
MAXIV(enc_xform_null); /* SADB_EALG_NULL */
|
2007-05-09 19:37:02 +00:00
|
|
|
MAXIV(enc_xform_camellia); /* SADB_X_EALG_CAMELLIACBC */
|
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
|
|
|
|
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
|
|
|
xform_register(&esp_xformsw);
|
|
|
|
#undef MAXIV
|
2002-10-16 02:10:08 +00:00
|
|
|
}
|
2004-01-27 17:43:49 +00:00
|
|
|
SYSINIT(esp_xform_init, SI_SUB_PROTO_DOMAIN, SI_ORDER_MIDDLE, esp_attach, NULL);
|