freebsd-dev/sys/netgraph/ng_iface.c

821 lines
20 KiB
C
Raw Normal View History

/*
* ng_iface.c
*/
/*-
* Copyright (c) 1996-1999 Whistle Communications, Inc.
* All rights reserved.
*
* Subject to the following obligations and disclaimer of warranty, use and
* redistribution of this software, in source or object code forms, with or
* without modifications are expressly permitted by Whistle Communications;
* provided, however, that:
* 1. Any and all reproductions of the source or object code must include the
* copyright notice above and the following disclaimer of warranties; and
* 2. No rights are granted, in any manner or form, to use Whistle
* Communications, Inc. trademarks, including the mark "WHISTLE
* COMMUNICATIONS" on advertising, endorsements, or otherwise except as
* such appears in the above copyright notice or in the software.
*
* THIS SOFTWARE IS BEING PROVIDED BY WHISTLE COMMUNICATIONS "AS IS", AND
* TO THE MAXIMUM EXTENT PERMITTED BY LAW, WHISTLE COMMUNICATIONS MAKES NO
* REPRESENTATIONS OR WARRANTIES, EXPRESS OR IMPLIED, REGARDING THIS SOFTWARE,
* INCLUDING WITHOUT LIMITATION, ANY AND ALL IMPLIED WARRANTIES OF
* MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE, OR NON-INFRINGEMENT.
* WHISTLE COMMUNICATIONS DOES NOT WARRANT, GUARANTEE, OR MAKE ANY
* REPRESENTATIONS REGARDING THE USE OF, OR THE RESULTS OF THE USE OF THIS
* SOFTWARE IN TERMS OF ITS CORRECTNESS, ACCURACY, RELIABILITY OR OTHERWISE.
* IN NO EVENT SHALL WHISTLE COMMUNICATIONS BE LIABLE FOR ANY DAMAGES
* RESULTING FROM OR ARISING OUT OF ANY USE OF THIS SOFTWARE, INCLUDING
* WITHOUT LIMITATION, ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY,
* PUNITIVE, OR CONSEQUENTIAL DAMAGES, PROCUREMENT OF SUBSTITUTE GOODS OR
* SERVICES, LOSS OF USE, DATA OR PROFITS, HOWEVER CAUSED AND UNDER 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 WHISTLE COMMUNICATIONS IS ADVISED OF THE POSSIBILITY
* OF SUCH DAMAGE.
*
* Author: Archie Cobbs <archie@freebsd.org>
*
* $FreeBSD$
* $Whistle: ng_iface.c,v 1.33 1999/11/01 09:24:51 julian Exp $
*/
/*
* This node is also a system networking interface. It has
* a hook for each protocol (IP, AppleTalk, etc). Packets
* are simply relayed between the interface and the hooks.
*
* Interfaces are named ng0, ng1, etc. New nodes take the
* first available interface name.
*
* This node also includes Berkeley packet filter support.
*/
#include "opt_inet.h"
#include "opt_inet6.h"
#include <sys/param.h>
#include <sys/systm.h>
#include <sys/errno.h>
#include <sys/kernel.h>
#include <sys/lock.h>
#include <sys/malloc.h>
#include <sys/mbuf.h>
#include <sys/errno.h>
#include <sys/proc.h>
#include <sys/random.h>
#include <sys/rmlock.h>
#include <sys/sockio.h>
#include <sys/socket.h>
#include <sys/sysctl.h>
#include <sys/syslog.h>
#include <sys/libkern.h>
#include <net/if.h>
#include <net/if_var.h>
#include <net/if_types.h>
#include <net/bpf.h>
#include <net/netisr.h>
#include <net/route.h>
#include <net/vnet.h>
#include <netinet/in.h>
#include <netgraph/ng_message.h>
#include <netgraph/netgraph.h>
#include <netgraph/ng_parse.h>
#include <netgraph/ng_iface.h>
#ifdef NG_SEPARATE_MALLOC
static MALLOC_DEFINE(M_NETGRAPH_IFACE, "netgraph_iface", "netgraph iface node");
#else
#define M_NETGRAPH_IFACE M_NETGRAPH
#endif
static SYSCTL_NODE(_net_graph, OID_AUTO, iface, CTLFLAG_RW | CTLFLAG_MPSAFE, 0,
"Point to point netgraph interface");
VNET_DEFINE_STATIC(int, ng_iface_max_nest) = 2;
#define V_ng_iface_max_nest VNET(ng_iface_max_nest)
SYSCTL_INT(_net_graph_iface, OID_AUTO, max_nesting, CTLFLAG_VNET | CTLFLAG_RW,
&VNET_NAME(ng_iface_max_nest), 0, "Max nested tunnels");
/* This struct describes one address family */
struct iffam {
sa_family_t family; /* Address family */
const char *hookname; /* Name for hook */
};
typedef const struct iffam *iffam_p;
/* List of address families supported by our interface */
const static struct iffam gFamilies[] = {
{ AF_INET, NG_IFACE_HOOK_INET },
{ AF_INET6, NG_IFACE_HOOK_INET6 },
{ AF_ATM, NG_IFACE_HOOK_ATM },
{ AF_NATM, NG_IFACE_HOOK_NATM },
};
#define NUM_FAMILIES nitems(gFamilies)
/* Node private data */
struct ng_iface_private {
struct ifnet *ifp; /* Our interface */
int unit; /* Interface unit number */
node_p node; /* Our netgraph node */
hook_p hooks[NUM_FAMILIES]; /* Hook for each address family */
struct rmlock lock; /* Protect private data changes */
};
typedef struct ng_iface_private *priv_p;
#define PRIV_RLOCK(priv, t) rm_rlock(&priv->lock, t)
#define PRIV_RUNLOCK(priv, t) rm_runlock(&priv->lock, t)
#define PRIV_WLOCK(priv) rm_wlock(&priv->lock)
#define PRIV_WUNLOCK(priv) rm_wunlock(&priv->lock)
/* Interface methods */
static void ng_iface_start(struct ifnet *ifp);
static int ng_iface_ioctl(struct ifnet *ifp, u_long cmd, caddr_t data);
static int ng_iface_output(struct ifnet *ifp, struct mbuf *m0,
const struct sockaddr *dst, struct route *ro);
static void ng_iface_bpftap(struct ifnet *ifp,
struct mbuf *m, sa_family_t family);
static int ng_iface_send(struct ifnet *ifp, struct mbuf *m,
sa_family_t sa);
#ifdef DEBUG
static void ng_iface_print_ioctl(struct ifnet *ifp, int cmd, caddr_t data);
#endif
/* Netgraph methods */
static int ng_iface_mod_event(module_t, int, void *);
static ng_constructor_t ng_iface_constructor;
static ng_rcvmsg_t ng_iface_rcvmsg;
static ng_shutdown_t ng_iface_shutdown;
static ng_newhook_t ng_iface_newhook;
static ng_rcvdata_t ng_iface_rcvdata;
static ng_disconnect_t ng_iface_disconnect;
/* Helper stuff */
static iffam_p get_iffam_from_af(sa_family_t family);
static iffam_p get_iffam_from_hook(priv_p priv, hook_p hook);
static iffam_p get_iffam_from_name(const char *name);
static hook_p *get_hook_from_iffam(priv_p priv, iffam_p iffam);
/* List of commands and how to convert arguments to/from ASCII */
static const struct ng_cmdlist ng_iface_cmds[] = {
{
NGM_IFACE_COOKIE,
NGM_IFACE_GET_IFNAME,
"getifname",
NULL,
&ng_parse_string_type
},
{
NGM_IFACE_COOKIE,
NGM_IFACE_POINT2POINT,
"point2point",
NULL,
NULL
},
{
NGM_IFACE_COOKIE,
NGM_IFACE_BROADCAST,
"broadcast",
NULL,
NULL
},
{
NGM_IFACE_COOKIE,
NGM_IFACE_GET_IFINDEX,
"getifindex",
NULL,
&ng_parse_uint32_type
},
{ 0 }
};
/* Node type descriptor */
static struct ng_type typestruct = {
.version = NG_ABI_VERSION,
.name = NG_IFACE_NODE_TYPE,
.mod_event = ng_iface_mod_event,
.constructor = ng_iface_constructor,
.rcvmsg = ng_iface_rcvmsg,
.shutdown = ng_iface_shutdown,
.newhook = ng_iface_newhook,
.rcvdata = ng_iface_rcvdata,
.disconnect = ng_iface_disconnect,
.cmdlist = ng_iface_cmds,
};
NETGRAPH_INIT(iface, &typestruct);
VNET_DEFINE_STATIC(struct unrhdr *, ng_iface_unit);
#define V_ng_iface_unit VNET(ng_iface_unit)
/************************************************************************
HELPER STUFF
************************************************************************/
/*
* Get the family descriptor from the family ID
*/
static __inline iffam_p
get_iffam_from_af(sa_family_t family)
{
iffam_p iffam;
int k;
for (k = 0; k < NUM_FAMILIES; k++) {
iffam = &gFamilies[k];
if (iffam->family == family)
return (iffam);
}
return (NULL);
}
/*
* Get the family descriptor from the hook
*/
static __inline iffam_p
get_iffam_from_hook(priv_p priv, hook_p hook)
{
int k;
for (k = 0; k < NUM_FAMILIES; k++)
if (priv->hooks[k] == hook)
return (&gFamilies[k]);
return (NULL);
}
/*
* Get the hook from the iffam descriptor
*/
static __inline hook_p *
get_hook_from_iffam(priv_p priv, iffam_p iffam)
{
return (&priv->hooks[iffam - gFamilies]);
}
/*
* Get the iffam descriptor from the name
*/
static __inline iffam_p
get_iffam_from_name(const char *name)
{
iffam_p iffam;
int k;
for (k = 0; k < NUM_FAMILIES; k++) {
iffam = &gFamilies[k];
if (!strcmp(iffam->hookname, name))
return (iffam);
}
return (NULL);
}
/************************************************************************
INTERFACE STUFF
************************************************************************/
/*
* Process an ioctl for the virtual interface
*/
static int
ng_iface_ioctl(struct ifnet *ifp, u_long command, caddr_t data)
{
struct ifreq *const ifr = (struct ifreq *) data;
2011-03-19 19:37:53 +00:00
int error = 0;
#ifdef DEBUG
ng_iface_print_ioctl(ifp, command, data);
#endif
switch (command) {
/* These two are mostly handled at a higher layer */
case SIOCSIFADDR:
ifp->if_flags |= IFF_UP;
ifp->if_drv_flags |= IFF_DRV_RUNNING;
ifp->if_drv_flags &= ~(IFF_DRV_OACTIVE);
break;
case SIOCGIFADDR:
break;
/* Set flags */
case SIOCSIFFLAGS:
/*
* If the interface is marked up and stopped, then start it.
* If it is marked down and running, then stop it.
*/
if (ifr->ifr_flags & IFF_UP) {
if (!(ifp->if_drv_flags & IFF_DRV_RUNNING)) {
ifp->if_drv_flags &= ~(IFF_DRV_OACTIVE);
ifp->if_drv_flags |= IFF_DRV_RUNNING;
}
} else {
if (ifp->if_drv_flags & IFF_DRV_RUNNING)
ifp->if_drv_flags &= ~(IFF_DRV_RUNNING |
IFF_DRV_OACTIVE);
}
break;
/* Set the interface MTU */
case SIOCSIFMTU:
if (ifr->ifr_mtu > NG_IFACE_MTU_MAX
|| ifr->ifr_mtu < NG_IFACE_MTU_MIN)
error = EINVAL;
else
ifp->if_mtu = ifr->ifr_mtu;
break;
/* Stuff that's not supported */
case SIOCADDMULTI:
case SIOCDELMULTI:
error = 0;
break;
case SIOCSIFPHYS:
error = EOPNOTSUPP;
break;
default:
error = EINVAL;
break;
}
return (error);
}
/*
* This routine is called to deliver a packet out the interface.
* We simply look at the address family and relay the packet to
* the corresponding hook, if it exists and is connected.
*/
static int
ng_iface_output(struct ifnet *ifp, struct mbuf *m,
const struct sockaddr *dst, struct route *ro)
{
uint32_t af;
int error;
/* Check interface flags */
if (!((ifp->if_flags & IFF_UP) &&
(ifp->if_drv_flags & IFF_DRV_RUNNING))) {
m_freem(m);
return (ENETDOWN);
}
/* Protect from deadly infinite recursion. */
error = if_tunnel_check_nesting(ifp, m, NGM_IFACE_COOKIE,
V_ng_iface_max_nest);
if (error) {
m_freem(m);
return (error);
}
/* BPF writes need to be handled specially. */
if (dst->sa_family == AF_UNSPEC)
bcopy(dst->sa_data, &af, sizeof(af));
else
af = dst->sa_family;
/* Berkeley packet filter */
ng_iface_bpftap(ifp, m, af);
if (ALTQ_IS_ENABLED(&ifp->if_snd)) {
M_PREPEND(m, sizeof(sa_family_t), M_NOWAIT);
if (m == NULL) {
if_inc_counter(ifp, IFCOUNTER_OQDROPS, 1);
return (ENOBUFS);
}
*(sa_family_t *)m->m_data = af;
error = (ifp->if_transmit)(ifp, m);
} else
error = ng_iface_send(ifp, m, af);
return (error);
}
/*
* Start method is used only when ALTQ is enabled.
*/
static void
ng_iface_start(struct ifnet *ifp)
{
struct mbuf *m;
sa_family_t sa;
KASSERT(ALTQ_IS_ENABLED(&ifp->if_snd), ("%s without ALTQ", __func__));
for(;;) {
IFQ_DRV_DEQUEUE(&ifp->if_snd, m);
if (m == NULL)
break;
sa = *mtod(m, sa_family_t *);
m_adj(m, sizeof(sa_family_t));
ng_iface_send(ifp, m, sa);
}
}
/*
* Flash a packet by the BPF (requires prepending 4 byte AF header)
* Note the phoney mbuf; this is OK because BPF treats it read-only.
*/
static void
ng_iface_bpftap(struct ifnet *ifp, struct mbuf *m, sa_family_t family)
{
KASSERT(family != AF_UNSPEC, ("%s: family=AF_UNSPEC", __func__));
2006-06-02 23:14:40 +00:00
if (bpf_peers_present(ifp->if_bpf)) {
int32_t family4 = (int32_t)family;
bpf_mtap2(ifp->if_bpf, &family4, sizeof(family4), m);
}
}
/*
* This routine does actual delivery of the packet into the
* netgraph(4). It is called from ng_iface_start() and
* ng_iface_output().
*/
static int
ng_iface_send(struct ifnet *ifp, struct mbuf *m, sa_family_t sa)
{
struct rm_priotracker priv_tracker;
const priv_p priv = (priv_p) ifp->if_softc;
const iffam_p iffam = get_iffam_from_af(sa);
hook_p hook;
int error;
int len;
/* Check address family to determine hook (if known) */
if (iffam == NULL) {
m_freem(m);
log(LOG_WARNING, "%s: can't handle af%d\n", ifp->if_xname, sa);
return (EAFNOSUPPORT);
}
/* Copy length before the mbuf gets invalidated. */
len = m->m_pkthdr.len;
PRIV_RLOCK(priv, &priv_tracker);
hook = *get_hook_from_iffam(priv, iffam);
if (hook == NULL) {
NG_FREE_M(m);
PRIV_RUNLOCK(priv, &priv_tracker);
return ENETDOWN;
}
NG_HOOK_REF(hook);
PRIV_RUNLOCK(priv, &priv_tracker);
NG_OUTBOUND_THREAD_REF();
NG_SEND_DATA_ONLY(error, hook, m);
NG_OUTBOUND_THREAD_UNREF();
NG_HOOK_UNREF(hook);
/* Update stats. */
if (error == 0) {
if_inc_counter(ifp, IFCOUNTER_OBYTES, len);
if_inc_counter(ifp, IFCOUNTER_OPACKETS, 1);
}
return (error);
}
#ifdef DEBUG
/*
* Display an ioctl to the virtual interface
*/
static void
ng_iface_print_ioctl(struct ifnet *ifp, int command, caddr_t data)
{
char *str;
switch (command & IOC_DIRMASK) {
case IOC_VOID:
str = "IO";
break;
case IOC_OUT:
str = "IOR";
break;
case IOC_IN:
str = "IOW";
break;
case IOC_INOUT:
str = "IORW";
break;
default:
str = "IO??";
}
log(LOG_DEBUG, "%s: %s('%c', %d, char[%d])\n",
ifp->if_xname,
str,
IOCGROUP(command),
command & 0xff,
IOCPARM_LEN(command));
}
#endif /* DEBUG */
/************************************************************************
NETGRAPH NODE STUFF
************************************************************************/
/*
* Constructor for a node
*/
static int
ng_iface_constructor(node_p node)
{
struct ifnet *ifp;
priv_p priv;
/* Allocate node and interface private structures */
priv = malloc(sizeof(*priv), M_NETGRAPH_IFACE, M_WAITOK | M_ZERO);
ifp = if_alloc(IFT_PROPVIRTUAL);
if (ifp == NULL) {
free(priv, M_NETGRAPH_IFACE);
return (ENOMEM);
}
rm_init(&priv->lock, "ng_iface private rmlock");
/* Link them together */
ifp->if_softc = priv;
priv->ifp = ifp;
/* Get an interface unit number */
priv->unit = alloc_unr(V_ng_iface_unit);
/* Link together node and private info */
NG_NODE_SET_PRIVATE(node, priv);
priv->node = node;
/* Initialize interface structure */
if_initname(ifp, NG_IFACE_IFACE_NAME, priv->unit);
ifp->if_output = ng_iface_output;
ifp->if_start = ng_iface_start;
ifp->if_ioctl = ng_iface_ioctl;
ifp->if_mtu = NG_IFACE_MTU_DEFAULT;
ifp->if_flags = (IFF_SIMPLEX|IFF_POINTOPOINT|IFF_NOARP|IFF_MULTICAST);
ifp->if_type = IFT_PROPVIRTUAL; /* XXX */
ifp->if_addrlen = 0; /* XXX */
ifp->if_hdrlen = 0; /* XXX */
ifp->if_baudrate = 64000; /* XXX */
IFQ_SET_MAXLEN(&ifp->if_snd, ifqmaxlen);
ifp->if_snd.ifq_drv_maxlen = ifqmaxlen;
IFQ_SET_READY(&ifp->if_snd);
/* Give this node the same name as the interface (if possible) */
if (ng_name_node(node, ifp->if_xname) != 0)
log(LOG_WARNING, "%s: can't acquire netgraph name\n",
ifp->if_xname);
/* Attach the interface */
if_attach(ifp);
bpfattach(ifp, DLT_NULL, sizeof(u_int32_t));
/* Done */
return (0);
}
/*
* Give our ok for a hook to be added
*/
static int
ng_iface_newhook(node_p node, hook_p hook, const char *name)
{
const iffam_p iffam = get_iffam_from_name(name);
const priv_p priv = NG_NODE_PRIVATE(node);
hook_p *hookptr;
if (iffam == NULL)
return (EPFNOSUPPORT);
PRIV_WLOCK(priv);
hookptr = get_hook_from_iffam(priv, iffam);
if (*hookptr != NULL) {
PRIV_WUNLOCK(priv);
return (EISCONN);
}
*hookptr = hook;
NG_HOOK_HI_STACK(hook);
NG_HOOK_SET_TO_INBOUND(hook);
PRIV_WUNLOCK(priv);
return (0);
}
/*
* Receive a control message
*/
static int
ng_iface_rcvmsg(node_p node, item_p item, hook_p lasthook)
{
const priv_p priv = NG_NODE_PRIVATE(node);
struct ifnet *const ifp = priv->ifp;
struct ng_mesg *resp = NULL;
int error = 0;
struct ng_mesg *msg;
NGI_GET_MSG(item, msg);
switch (msg->header.typecookie) {
case NGM_IFACE_COOKIE:
switch (msg->header.cmd) {
case NGM_IFACE_GET_IFNAME:
NG_MKRESPONSE(resp, msg, IFNAMSIZ, M_NOWAIT);
if (resp == NULL) {
error = ENOMEM;
break;
}
strlcpy(resp->data, ifp->if_xname, IFNAMSIZ);
break;
case NGM_IFACE_POINT2POINT:
case NGM_IFACE_BROADCAST:
{
/* Deny request if interface is UP */
if ((ifp->if_flags & IFF_UP) != 0)
return (EBUSY);
/* Change flags */
switch (msg->header.cmd) {
case NGM_IFACE_POINT2POINT:
ifp->if_flags |= IFF_POINTOPOINT;
ifp->if_flags &= ~IFF_BROADCAST;
break;
case NGM_IFACE_BROADCAST:
ifp->if_flags &= ~IFF_POINTOPOINT;
ifp->if_flags |= IFF_BROADCAST;
break;
}
break;
}
case NGM_IFACE_GET_IFINDEX:
NG_MKRESPONSE(resp, msg, sizeof(uint32_t), M_NOWAIT);
if (resp == NULL) {
error = ENOMEM;
break;
}
*((uint32_t *)resp->data) = priv->ifp->if_index;
break;
default:
error = EINVAL;
break;
}
break;
case NGM_FLOW_COOKIE:
switch (msg->header.cmd) {
case NGM_LINK_IS_UP:
if_link_state_change(ifp, LINK_STATE_UP);
break;
case NGM_LINK_IS_DOWN:
if_link_state_change(ifp, LINK_STATE_DOWN);
break;
default:
break;
}
break;
default:
error = EINVAL;
break;
}
NG_RESPOND_MSG(error, node, item, resp);
NG_FREE_MSG(msg);
return (error);
}
/*
* Recive data from a hook. Pass the packet to the correct input routine.
*/
static int
ng_iface_rcvdata(hook_p hook, item_p item)
{
const priv_p priv = NG_NODE_PRIVATE(NG_HOOK_NODE(hook));
const iffam_p iffam = get_iffam_from_hook(priv, hook);
struct ifnet *const ifp = priv->ifp;
Widen NET_EPOCH coverage. When epoch(9) was introduced to network stack, it was basically dropped in place of existing locking, which was mutexes and rwlocks. For the sake of performance mutex covered areas were as small as possible, so became epoch covered areas. However, epoch doesn't introduce any contention, it just delays memory reclaim. So, there is no point to minimise epoch covered areas in sense of performance. Meanwhile entering/exiting epoch also has non-zero CPU usage, so doing this less often is a win. Not the least is also code maintainability. In the new paradigm we can assume that at any stage of processing a packet, we are inside network epoch. This makes coding both input and output path way easier. On output path we already enter epoch quite early - in the ip_output(), in the ip6_output(). This patch does the same for the input path. All ISR processing, network related callouts, other ways of packet injection to the network stack shall be performed in net_epoch. Any leaf function that walks network configuration now asserts epoch. Tricky part is configuration code paths - ioctls, sysctls. They also call into leaf functions, so some need to be changed. This patch would introduce more epoch recursions (see EPOCH_TRACE) than we had before. They will be cleaned up separately, as several of them aren't trivial. Note, that unlike a lock recursion the epoch recursion is safe and just wastes a bit of resources. Reviewed by: gallatin, hselasky, cy, adrian, kristof Differential Revision: https://reviews.freebsd.org/D19111
2019-10-07 22:40:05 +00:00
struct epoch_tracker et;
struct mbuf *m;
int isr;
NGI_GET_M(item, m);
NG_FREE_ITEM(item);
/* Sanity checks */
KASSERT(iffam != NULL, ("%s: iffam", __func__));
M_ASSERTPKTHDR(m);
if ((ifp->if_flags & IFF_UP) == 0) {
NG_FREE_M(m);
return (ENETDOWN);
}
/* Update interface stats */
if_inc_counter(ifp, IFCOUNTER_IPACKETS, 1);
if_inc_counter(ifp, IFCOUNTER_IBYTES, m->m_pkthdr.len);
/* Note receiving interface */
m->m_pkthdr.rcvif = ifp;
/* Berkeley packet filter */
ng_iface_bpftap(ifp, m, iffam->family);
/* Send packet */
switch (iffam->family) {
#ifdef INET
case AF_INET:
isr = NETISR_IP;
break;
#endif
#ifdef INET6
case AF_INET6:
isr = NETISR_IPV6;
break;
#endif
default:
m_freem(m);
return (EAFNOSUPPORT);
}
random_harvest_queue(m, sizeof(*m), RANDOM_NET_NG);
M_SETFIB(m, ifp->if_fib);
Widen NET_EPOCH coverage. When epoch(9) was introduced to network stack, it was basically dropped in place of existing locking, which was mutexes and rwlocks. For the sake of performance mutex covered areas were as small as possible, so became epoch covered areas. However, epoch doesn't introduce any contention, it just delays memory reclaim. So, there is no point to minimise epoch covered areas in sense of performance. Meanwhile entering/exiting epoch also has non-zero CPU usage, so doing this less often is a win. Not the least is also code maintainability. In the new paradigm we can assume that at any stage of processing a packet, we are inside network epoch. This makes coding both input and output path way easier. On output path we already enter epoch quite early - in the ip_output(), in the ip6_output(). This patch does the same for the input path. All ISR processing, network related callouts, other ways of packet injection to the network stack shall be performed in net_epoch. Any leaf function that walks network configuration now asserts epoch. Tricky part is configuration code paths - ioctls, sysctls. They also call into leaf functions, so some need to be changed. This patch would introduce more epoch recursions (see EPOCH_TRACE) than we had before. They will be cleaned up separately, as several of them aren't trivial. Note, that unlike a lock recursion the epoch recursion is safe and just wastes a bit of resources. Reviewed by: gallatin, hselasky, cy, adrian, kristof Differential Revision: https://reviews.freebsd.org/D19111
2019-10-07 22:40:05 +00:00
NET_EPOCH_ENTER(et);
netisr_dispatch(isr, m);
Widen NET_EPOCH coverage. When epoch(9) was introduced to network stack, it was basically dropped in place of existing locking, which was mutexes and rwlocks. For the sake of performance mutex covered areas were as small as possible, so became epoch covered areas. However, epoch doesn't introduce any contention, it just delays memory reclaim. So, there is no point to minimise epoch covered areas in sense of performance. Meanwhile entering/exiting epoch also has non-zero CPU usage, so doing this less often is a win. Not the least is also code maintainability. In the new paradigm we can assume that at any stage of processing a packet, we are inside network epoch. This makes coding both input and output path way easier. On output path we already enter epoch quite early - in the ip_output(), in the ip6_output(). This patch does the same for the input path. All ISR processing, network related callouts, other ways of packet injection to the network stack shall be performed in net_epoch. Any leaf function that walks network configuration now asserts epoch. Tricky part is configuration code paths - ioctls, sysctls. They also call into leaf functions, so some need to be changed. This patch would introduce more epoch recursions (see EPOCH_TRACE) than we had before. They will be cleaned up separately, as several of them aren't trivial. Note, that unlike a lock recursion the epoch recursion is safe and just wastes a bit of resources. Reviewed by: gallatin, hselasky, cy, adrian, kristof Differential Revision: https://reviews.freebsd.org/D19111
2019-10-07 22:40:05 +00:00
NET_EPOCH_EXIT(et);
return (0);
}
/*
* Shutdown and remove the node and its associated interface.
*/
static int
ng_iface_shutdown(node_p node)
{
const priv_p priv = NG_NODE_PRIVATE(node);
/*
* The ifnet may be in a different vnet than the netgraph node,
* hence we have to change the current vnet context here.
*/
CURVNET_SET_QUIET(priv->ifp->if_vnet);
2000-03-21 01:42:56 +00:00
bpfdetach(priv->ifp);
if_detach(priv->ifp);
if_free(priv->ifp);
CURVNET_RESTORE();
priv->ifp = NULL;
free_unr(V_ng_iface_unit, priv->unit);
rm_destroy(&priv->lock);
free(priv, M_NETGRAPH_IFACE);
NG_NODE_SET_PRIVATE(node, NULL);
NG_NODE_UNREF(node);
return (0);
}
/*
* Hook disconnection. Note that we do *not* shutdown when all
* hooks have been disconnected.
*/
static int
ng_iface_disconnect(hook_p hook)
{
const priv_p priv = NG_NODE_PRIVATE(NG_HOOK_NODE(hook));
const iffam_p iffam = get_iffam_from_hook(priv, hook);
if (iffam == NULL)
panic("%s", __func__);
PRIV_WLOCK(priv);
*get_hook_from_iffam(priv, iffam) = NULL;
PRIV_WUNLOCK(priv);
return (0);
}
/*
* Handle loading and unloading for this node type.
*/
static int
ng_iface_mod_event(module_t mod, int event, void *data)
{
int error = 0;
switch (event) {
case MOD_LOAD:
case MOD_UNLOAD:
break;
default:
error = EOPNOTSUPP;
break;
}
return (error);
}
static void
vnet_ng_iface_init(const void *unused)
{
V_ng_iface_unit = new_unrhdr(0, 0xffff, NULL);
}
VNET_SYSINIT(vnet_ng_iface_init, SI_SUB_PSEUDO, SI_ORDER_ANY,
vnet_ng_iface_init, NULL);
static void
vnet_ng_iface_uninit(const void *unused)
{
delete_unrhdr(V_ng_iface_unit);
}
Get closer to a VIMAGE network stack teardown from top to bottom rather than removing the network interfaces first. This change is rather larger and convoluted as the ordering requirements cannot be separated. Move the pfil(9) framework to SI_SUB_PROTO_PFIL, move Firewalls and related modules to their own SI_SUB_PROTO_FIREWALL. Move initialization of "physical" interfaces to SI_SUB_DRIVERS, move virtual (cloned) interfaces to SI_SUB_PSEUDO. Move Multicast to SI_SUB_PROTO_MC. Re-work parts of multicast initialisation and teardown, not taking the huge amount of memory into account if used as a module yet. For interface teardown we try to do as many of them as we can on SI_SUB_INIT_IF, but for some this makes no sense, e.g., when tunnelling over a higher layer protocol such as IP. In that case the interface has to go along (or before) the higher layer protocol is shutdown. Kernel hhooks need to go last on teardown as they may be used at various higher layers and we cannot remove them before we cleaned up the higher layers. For interface teardown there are multiple paths: (a) a cloned interface is destroyed (inside a VIMAGE or in the base system), (b) any interface is moved from a virtual network stack to a different network stack ("vmove"), or (c) a virtual network stack is being shut down. All code paths go through if_detach_internal() where we, depending on the vmove flag or the vnet state, make a decision on how much to shut down; in case we are destroying a VNET the individual protocol layers will cleanup their own parts thus we cannot do so again for each interface as we end up with, e.g., double-frees, destroying locks twice or acquiring already destroyed locks. When calling into protocol cleanups we equally have to tell them whether they need to detach upper layer protocols ("ulp") or not (e.g., in6_ifdetach()). Provide or enahnce helper functions to do proper cleanup at a protocol rather than at an interface level. Approved by: re (hrs) Obtained from: projects/vnet Reviewed by: gnn, jhb Sponsored by: The FreeBSD Foundation MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6747
2016-06-21 13:48:49 +00:00
VNET_SYSUNINIT(vnet_ng_iface_uninit, SI_SUB_INIT_IF, SI_ORDER_ANY,
vnet_ng_iface_uninit, NULL);