freebsd-skq/sys/net/if_clone.c

601 lines
14 KiB
C
Raw Normal View History

/*-
* Copyright (c) 1980, 1986, 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.
*
* @(#)if.c 8.5 (Berkeley) 1/9/95
* $FreeBSD$
*/
#include <sys/param.h>
#include <sys/malloc.h>
#include <sys/limits.h>
#include <sys/lock.h>
#include <sys/mutex.h>
#include <sys/kernel.h>
#include <sys/systm.h>
#include <sys/types.h>
#include <sys/socket.h>
#include <net/if.h>
#include <net/if_clone.h>
#if 0
#include <net/if_dl.h>
#endif
#include <net/if_types.h>
#include <net/if_var.h>
#include <net/radix.h>
#include <net/route.h>
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
#include <net/vnet.h>
static void if_clone_free(struct if_clone *ifc);
static int if_clone_createif(struct if_clone *ifc, char *name, size_t len,
caddr_t params);
static struct mtx if_cloners_mtx;
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, if_cloners_count);
VNET_DEFINE(LIST_HEAD(, if_clone), if_cloners);
#define V_if_cloners_count VNET(if_cloners_count)
#define V_if_cloners VNET(if_cloners)
#define IF_CLONERS_LOCK_INIT() \
mtx_init(&if_cloners_mtx, "if_cloners lock", NULL, MTX_DEF)
#define IF_CLONERS_LOCK_ASSERT() mtx_assert(&if_cloners_mtx, MA_OWNED)
#define IF_CLONERS_LOCK() mtx_lock(&if_cloners_mtx)
#define IF_CLONERS_UNLOCK() mtx_unlock(&if_cloners_mtx)
#define IF_CLONE_LOCK_INIT(ifc) \
mtx_init(&(ifc)->ifc_mtx, "if_clone lock", NULL, MTX_DEF)
#define IF_CLONE_LOCK_DESTROY(ifc) mtx_destroy(&(ifc)->ifc_mtx)
#define IF_CLONE_LOCK_ASSERT(ifc) mtx_assert(&(ifc)->ifc_mtx, MA_OWNED)
#define IF_CLONE_LOCK(ifc) mtx_lock(&(ifc)->ifc_mtx)
#define IF_CLONE_UNLOCK(ifc) mtx_unlock(&(ifc)->ifc_mtx)
#define IF_CLONE_ADDREF(ifc) \
do { \
IF_CLONE_LOCK(ifc); \
IF_CLONE_ADDREF_LOCKED(ifc); \
IF_CLONE_UNLOCK(ifc); \
} while (0)
#define IF_CLONE_ADDREF_LOCKED(ifc) \
do { \
IF_CLONE_LOCK_ASSERT(ifc); \
KASSERT((ifc)->ifc_refcnt >= 0, \
("negative refcnt %ld", (ifc)->ifc_refcnt)); \
(ifc)->ifc_refcnt++; \
} while (0)
#define IF_CLONE_REMREF(ifc) \
do { \
IF_CLONE_LOCK(ifc); \
IF_CLONE_REMREF_LOCKED(ifc); \
} while (0)
#define IF_CLONE_REMREF_LOCKED(ifc) \
do { \
IF_CLONE_LOCK_ASSERT(ifc); \
KASSERT((ifc)->ifc_refcnt > 0, \
("bogus refcnt %ld", (ifc)->ifc_refcnt)); \
if (--(ifc)->ifc_refcnt == 0) { \
IF_CLONE_UNLOCK(ifc); \
if_clone_free(ifc); \
} else { \
/* silently free the lock */ \
IF_CLONE_UNLOCK(ifc); \
} \
} while (0)
#define IFC_IFLIST_INSERT(_ifc, _ifp) \
LIST_INSERT_HEAD(&_ifc->ifc_iflist, _ifp, if_clones)
#define IFC_IFLIST_REMOVE(_ifc, _ifp) \
LIST_REMOVE(_ifp, if_clones)
static MALLOC_DEFINE(M_CLONE, "clone", "interface cloning framework");
void
vnet_if_clone_init(void)
{
LIST_INIT(&V_if_cloners);
}
void
if_clone_init(void)
{
IF_CLONERS_LOCK_INIT();
}
/*
* Lookup and create a clone network interface.
*/
int
if_clone_create(char *name, size_t len, caddr_t params)
{
struct if_clone *ifc;
/* Try to find an applicable cloner for this request */
IF_CLONERS_LOCK();
LIST_FOREACH(ifc, &V_if_cloners, ifc_list) {
if (ifc->ifc_match(ifc, name)) {
break;
}
}
#ifdef VIMAGE
if (ifc == NULL && !IS_DEFAULT_VNET(curvnet)) {
CURVNET_SET_QUIET(vnet0);
LIST_FOREACH(ifc, &V_if_cloners, ifc_list) {
if (ifc->ifc_match(ifc, name))
break;
}
CURVNET_RESTORE();
}
#endif
IF_CLONERS_UNLOCK();
if (ifc == NULL)
return (EINVAL);
return (if_clone_createif(ifc, name, len, params));
}
/*
* Create a clone network interface.
*/
static int
if_clone_createif(struct if_clone *ifc, char *name, size_t len, caddr_t params)
{
int err;
struct ifnet *ifp;
if (ifunit(name) != NULL)
return (EEXIST);
err = (*ifc->ifc_create)(ifc, name, len, params);
if (!err) {
ifp = ifunit(name);
if (ifp == NULL)
panic("%s: lookup failed for %s", __func__, name);
if_addgroup(ifp, ifc->ifc_name);
IF_CLONE_LOCK(ifc);
IFC_IFLIST_INSERT(ifc, ifp);
IF_CLONE_UNLOCK(ifc);
}
return (err);
}
/*
* Lookup and destroy a clone network interface.
*/
int
if_clone_destroy(const char *name)
{
struct if_clone *ifc;
struct ifnet *ifp;
ifp = ifunit(name);
if (ifp == NULL)
return (ENXIO);
/* Find the cloner for this interface */
IF_CLONERS_LOCK();
LIST_FOREACH(ifc, &V_if_cloners, ifc_list) {
if (strcmp(ifc->ifc_name, ifp->if_dname) == 0) {
break;
}
}
#ifdef VIMAGE
if (ifc == NULL && !IS_DEFAULT_VNET(curvnet)) {
CURVNET_SET_QUIET(vnet0);
LIST_FOREACH(ifc, &V_if_cloners, ifc_list) {
if (ifc->ifc_match(ifc, name))
break;
}
CURVNET_RESTORE();
}
#endif
IF_CLONERS_UNLOCK();
if (ifc == NULL)
return (EINVAL);
return (if_clone_destroyif(ifc, ifp));
}
/*
* Destroy a clone network interface.
*/
int
if_clone_destroyif(struct if_clone *ifc, struct ifnet *ifp)
{
int err;
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
if (ifc->ifc_destroy == NULL)
return(EOPNOTSUPP);
/*
* Given that the cloned ifnet might be attached to a different
* vnet from where its cloner was registered, we have to
* switch to the vnet context of the target vnet.
*/
CURVNET_SET_QUIET(ifp->if_vnet);
IF_CLONE_LOCK(ifc);
IFC_IFLIST_REMOVE(ifc, ifp);
IF_CLONE_UNLOCK(ifc);
if_delgroup(ifp, ifc->ifc_name);
err = (*ifc->ifc_destroy)(ifc, ifp);
if (err != 0) {
if_addgroup(ifp, ifc->ifc_name);
IF_CLONE_LOCK(ifc);
IFC_IFLIST_INSERT(ifc, ifp);
IF_CLONE_UNLOCK(ifc);
}
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
CURVNET_RESTORE();
return (err);
}
/*
* Register a network interface cloner.
*/
void
if_clone_attach(struct if_clone *ifc)
{
int len, maxclone;
/*
* Compute bitmap size and allocate it.
*/
maxclone = ifc->ifc_maxunit + 1;
len = maxclone >> 3;
if ((len << 3) < maxclone)
len++;
ifc->ifc_units = malloc(len, M_CLONE, M_WAITOK | M_ZERO);
ifc->ifc_bmlen = len;
IF_CLONE_LOCK_INIT(ifc);
IF_CLONE_ADDREF(ifc);
IF_CLONERS_LOCK();
LIST_INSERT_HEAD(&V_if_cloners, ifc, ifc_list);
V_if_cloners_count++;
IF_CLONERS_UNLOCK();
LIST_INIT(&ifc->ifc_iflist);
if (ifc->ifc_attach != NULL)
(*ifc->ifc_attach)(ifc);
EVENTHANDLER_INVOKE(if_clone_event, ifc);
}
/*
* Unregister a network interface cloner.
*/
void
if_clone_detach(struct if_clone *ifc)
{
struct ifc_simple_data *ifcs = ifc->ifc_data;
IF_CLONERS_LOCK();
LIST_REMOVE(ifc, ifc_list);
V_if_cloners_count--;
IF_CLONERS_UNLOCK();
/* Allow all simples to be destroyed */
if (ifc->ifc_attach == ifc_simple_attach)
ifcs->ifcs_minifs = 0;
/* destroy all interfaces for this cloner */
while (!LIST_EMPTY(&ifc->ifc_iflist))
if_clone_destroyif(ifc, LIST_FIRST(&ifc->ifc_iflist));
IF_CLONE_REMREF(ifc);
}
static void
if_clone_free(struct if_clone *ifc)
{
for (int bytoff = 0; bytoff < ifc->ifc_bmlen; bytoff++) {
KASSERT(ifc->ifc_units[bytoff] == 0x00,
("ifc_units[%d] is not empty", bytoff));
}
KASSERT(LIST_EMPTY(&ifc->ifc_iflist),
("%s: ifc_iflist not empty", __func__));
IF_CLONE_LOCK_DESTROY(ifc);
free(ifc->ifc_units, M_CLONE);
}
/*
* Provide list of interface cloners to userspace.
*/
int
if_clone_list(struct if_clonereq *ifcr)
{
char *buf, *dst, *outbuf = NULL;
struct if_clone *ifc;
int buf_count, count, err = 0;
if (ifcr->ifcr_count < 0)
return (EINVAL);
IF_CLONERS_LOCK();
/*
* Set our internal output buffer size. We could end up not
* reporting a cloner that is added between the unlock and lock
* below, but that's not a major problem. Not caping our
* allocation to the number of cloners actually in the system
* could be because that would let arbitrary users cause us to
* allocate abritrary amounts of kernel memory.
*/
buf_count = (V_if_cloners_count < ifcr->ifcr_count) ?
V_if_cloners_count : ifcr->ifcr_count;
IF_CLONERS_UNLOCK();
outbuf = malloc(IFNAMSIZ*buf_count, M_CLONE, M_WAITOK | M_ZERO);
IF_CLONERS_LOCK();
ifcr->ifcr_total = V_if_cloners_count;
if ((dst = ifcr->ifcr_buffer) == NULL) {
/* Just asking how many there are. */
goto done;
}
count = (V_if_cloners_count < buf_count) ?
V_if_cloners_count : buf_count;
for (ifc = LIST_FIRST(&V_if_cloners), buf = outbuf;
ifc != NULL && count != 0;
ifc = LIST_NEXT(ifc, ifc_list), count--, buf += IFNAMSIZ) {
strlcpy(buf, ifc->ifc_name, IFNAMSIZ);
}
done:
IF_CLONERS_UNLOCK();
if (err == 0)
err = copyout(outbuf, dst, buf_count*IFNAMSIZ);
if (outbuf != NULL)
free(outbuf, M_CLONE);
return (err);
}
/*
* A utility function to extract unit numbers from interface names of
* the form name###.
*
* Returns 0 on success and an error on failure.
*/
int
ifc_name2unit(const char *name, int *unit)
{
const char *cp;
int cutoff = INT_MAX / 10;
int cutlim = INT_MAX % 10;
for (cp = name; *cp != '\0' && (*cp < '0' || *cp > '9'); cp++);
if (*cp == '\0') {
*unit = -1;
} else if (cp[0] == '0' && cp[1] != '\0') {
/* Disallow leading zeroes. */
return (EINVAL);
} else {
for (*unit = 0; *cp != '\0'; cp++) {
if (*cp < '0' || *cp > '9') {
/* Bogus unit number. */
return (EINVAL);
}
if (*unit > cutoff ||
(*unit == cutoff && *cp - '0' > cutlim))
return (EINVAL);
*unit = (*unit * 10) + (*cp - '0');
}
}
return (0);
}
int
ifc_alloc_unit(struct if_clone *ifc, int *unit)
{
int wildcard, bytoff, bitoff;
int err = 0;
IF_CLONE_LOCK(ifc);
bytoff = bitoff = 0;
wildcard = (*unit < 0);
/*
* Find a free unit if none was given.
*/
if (wildcard) {
while ((bytoff < ifc->ifc_bmlen)
&& (ifc->ifc_units[bytoff] == 0xff))
bytoff++;
if (bytoff >= ifc->ifc_bmlen) {
err = ENOSPC;
goto done;
}
while ((ifc->ifc_units[bytoff] & (1 << bitoff)) != 0)
bitoff++;
*unit = (bytoff << 3) + bitoff;
}
if (*unit > ifc->ifc_maxunit) {
err = ENOSPC;
goto done;
}
if (!wildcard) {
bytoff = *unit >> 3;
bitoff = *unit - (bytoff << 3);
}
if((ifc->ifc_units[bytoff] & (1 << bitoff)) != 0) {
err = EEXIST;
goto done;
}
/*
* Allocate the unit in the bitmap.
*/
KASSERT((ifc->ifc_units[bytoff] & (1 << bitoff)) == 0,
("%s: bit is already set", __func__));
ifc->ifc_units[bytoff] |= (1 << bitoff);
IF_CLONE_ADDREF_LOCKED(ifc);
done:
IF_CLONE_UNLOCK(ifc);
return (err);
}
void
ifc_free_unit(struct if_clone *ifc, int unit)
{
int bytoff, bitoff;
/*
* Compute offset in the bitmap and deallocate the unit.
*/
bytoff = unit >> 3;
bitoff = unit - (bytoff << 3);
IF_CLONE_LOCK(ifc);
KASSERT((ifc->ifc_units[bytoff] & (1 << bitoff)) != 0,
("%s: bit is already cleared", __func__));
ifc->ifc_units[bytoff] &= ~(1 << bitoff);
IF_CLONE_REMREF_LOCKED(ifc); /* releases lock */
}
void
ifc_simple_attach(struct if_clone *ifc)
{
int err;
int unit;
char name[IFNAMSIZ];
struct ifc_simple_data *ifcs = ifc->ifc_data;
KASSERT(ifcs->ifcs_minifs - 1 <= ifc->ifc_maxunit,
("%s: %s requested more units than allowed (%d > %d)",
__func__, ifc->ifc_name, ifcs->ifcs_minifs,
ifc->ifc_maxunit + 1));
for (unit = 0; unit < ifcs->ifcs_minifs; unit++) {
snprintf(name, IFNAMSIZ, "%s%d", ifc->ifc_name, unit);
err = if_clone_createif(ifc, name, IFNAMSIZ, NULL);
KASSERT(err == 0,
("%s: failed to create required interface %s",
__func__, name));
}
}
int
ifc_simple_match(struct if_clone *ifc, const char *name)
{
const char *cp;
int i;
/* Match the name */
for (cp = name, i = 0; i < strlen(ifc->ifc_name); i++, cp++) {
if (ifc->ifc_name[i] != *cp)
return (0);
}
/* Make sure there's a unit number or nothing after the name */
for (; *cp != '\0'; cp++) {
if (*cp < '0' || *cp > '9')
return (0);
}
return (1);
}
int
ifc_simple_create(struct if_clone *ifc, char *name, size_t len, caddr_t params)
{
char *dp;
int wildcard;
int unit;
int err;
struct ifc_simple_data *ifcs = ifc->ifc_data;
err = ifc_name2unit(name, &unit);
if (err != 0)
return (err);
wildcard = (unit < 0);
err = ifc_alloc_unit(ifc, &unit);
if (err != 0)
return (err);
err = ifcs->ifcs_create(ifc, unit, params);
if (err != 0) {
ifc_free_unit(ifc, unit);
return (err);
}
/* In the wildcard case, we need to update the name. */
if (wildcard) {
for (dp = name; *dp != '\0'; dp++);
if (snprintf(dp, len - (dp-name), "%d", unit) >
len - (dp-name) - 1) {
/*
* This can only be a programmer error and
* there's no straightforward way to recover if
* it happens.
*/
panic("if_clone_create(): interface name too long");
}
}
return (0);
}
int
ifc_simple_destroy(struct if_clone *ifc, struct ifnet *ifp)
{
int unit;
struct ifc_simple_data *ifcs = ifc->ifc_data;
unit = ifp->if_dunit;
if (unit < ifcs->ifcs_minifs)
return (EINVAL);
ifcs->ifcs_destroy(ifp);
ifc_free_unit(ifc, unit);
return (0);
}