freebsd-dev/sys/net/rtsock.c

1998 lines
49 KiB
C
Raw Normal View History

/*-
* SPDX-License-Identifier: BSD-3-Clause
*
1994-05-24 10:09:53 +00:00
* Copyright (c) 1988, 1991, 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.
* 3. Neither the name of the University nor the names of its contributors
1994-05-24 10:09:53 +00:00
* may be used to endorse or promote products derived from this software
* without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*
* @(#)rtsock.c 8.7 (Berkeley) 10/12/95
1999-08-28 01:08:13 +00:00
* $FreeBSD$
1994-05-24 10:09:53 +00:00
*/
#include "opt_mpath.h"
MFp4: Bring in updated jail support from bz_jail branch. This enhances the current jail implementation to permit multiple addresses per jail. In addtion to IPv4, IPv6 is supported as well. Due to updated checks it is even possible to have jails without an IP address at all, which basically gives one a chroot with restricted process view, no networking,.. SCTP support was updated and supports IPv6 in jails as well. Cpuset support permits jails to be bound to specific processor sets after creation. Jails can have an unrestricted (no duplicate protection, etc.) name in addition to the hostname. The jail name cannot be changed from within a jail and is considered to be used for management purposes or as audit-token in the future. DDB 'show jails' command was added to aid debugging. Proper compat support permits 32bit jail binaries to be used on 64bit systems to manage jails. Also backward compatibility was preserved where possible: for jail v1 syscalls, as well as with user space management utilities. Both jail as well as prison version were updated for the new features. A gap was intentionally left as the intermediate versions had been used by various patches floating around the last years. Bump __FreeBSD_version for the afore mentioned and in kernel changes. Special thanks to: - Pawel Jakub Dawidek (pjd) for his multi-IPv4 patches and Olivier Houchard (cognet) for initial single-IPv6 patches. - Jeff Roberson (jeff) and Randall Stewart (rrs) for their help, ideas and review on cpuset and SCTP support. - Robert Watson (rwatson) for lots and lots of help, discussions, suggestions and review of most of the patch at various stages. - John Baldwin (jhb) for his help. - Simon L. Nielsen (simon) as early adopter testing changes on cluster machines as well as all the testers and people who provided feedback the last months on freebsd-jail and other channels. - My employer, CK Software GmbH, for the support so I could work on this. Reviewed by: (see above) MFC after: 3 months (this is just so that I get the mail) X-MFC Before: 7.2-RELEASE if possible
2008-11-29 14:32:14 +00:00
#include "opt_inet.h"
#include "opt_inet6.h"
1994-05-24 10:09:53 +00:00
#include <sys/param.h>
#include <sys/jail.h>
MFp4: Bring in updated jail support from bz_jail branch. This enhances the current jail implementation to permit multiple addresses per jail. In addtion to IPv4, IPv6 is supported as well. Due to updated checks it is even possible to have jails without an IP address at all, which basically gives one a chroot with restricted process view, no networking,.. SCTP support was updated and supports IPv6 in jails as well. Cpuset support permits jails to be bound to specific processor sets after creation. Jails can have an unrestricted (no duplicate protection, etc.) name in addition to the hostname. The jail name cannot be changed from within a jail and is considered to be used for management purposes or as audit-token in the future. DDB 'show jails' command was added to aid debugging. Proper compat support permits 32bit jail binaries to be used on 64bit systems to manage jails. Also backward compatibility was preserved where possible: for jail v1 syscalls, as well as with user space management utilities. Both jail as well as prison version were updated for the new features. A gap was intentionally left as the intermediate versions had been used by various patches floating around the last years. Bump __FreeBSD_version for the afore mentioned and in kernel changes. Special thanks to: - Pawel Jakub Dawidek (pjd) for his multi-IPv4 patches and Olivier Houchard (cognet) for initial single-IPv6 patches. - Jeff Roberson (jeff) and Randall Stewart (rrs) for their help, ideas and review on cpuset and SCTP support. - Robert Watson (rwatson) for lots and lots of help, discussions, suggestions and review of most of the patch at various stages. - John Baldwin (jhb) for his help. - Simon L. Nielsen (simon) as early adopter testing changes on cluster machines as well as all the testers and people who provided feedback the last months on freebsd-jail and other channels. - My employer, CK Software GmbH, for the support so I could work on this. Reviewed by: (see above) MFC after: 3 months (this is just so that I get the mail) X-MFC Before: 7.2-RELEASE if possible
2008-11-29 14:32:14 +00:00
#include <sys/kernel.h>
#include <sys/domain.h>
#include <sys/lock.h>
#include <sys/malloc.h>
1994-05-24 10:09:53 +00:00
#include <sys/mbuf.h>
#include <sys/priv.h>
#include <sys/proc.h>
#include <sys/protosw.h>
#include <sys/rmlock.h>
#include <sys/rwlock.h>
#include <sys/signalvar.h>
1994-05-24 10:09:53 +00:00
#include <sys/socket.h>
#include <sys/socketvar.h>
#include <sys/sysctl.h>
#include <sys/systm.h>
1994-05-24 10:09:53 +00:00
#include <net/if.h>
#include <net/if_var.h>
2008-12-26 19:45:24 +00:00
#include <net/if_dl.h>
#include <net/if_llatbl.h>
#include <net/if_types.h>
#include <net/netisr.h>
1994-05-24 10:09:53 +00:00
#include <net/raw_cb.h>
#include <net/route.h>
#include <net/route_var.h>
#include <net/vnet.h>
1994-05-24 10:09:53 +00:00
#include <netinet/in.h>
#include <netinet/if_ether.h>
#include <netinet/ip_carp.h>
MFp4: Bring in updated jail support from bz_jail branch. This enhances the current jail implementation to permit multiple addresses per jail. In addtion to IPv4, IPv6 is supported as well. Due to updated checks it is even possible to have jails without an IP address at all, which basically gives one a chroot with restricted process view, no networking,.. SCTP support was updated and supports IPv6 in jails as well. Cpuset support permits jails to be bound to specific processor sets after creation. Jails can have an unrestricted (no duplicate protection, etc.) name in addition to the hostname. The jail name cannot be changed from within a jail and is considered to be used for management purposes or as audit-token in the future. DDB 'show jails' command was added to aid debugging. Proper compat support permits 32bit jail binaries to be used on 64bit systems to manage jails. Also backward compatibility was preserved where possible: for jail v1 syscalls, as well as with user space management utilities. Both jail as well as prison version were updated for the new features. A gap was intentionally left as the intermediate versions had been used by various patches floating around the last years. Bump __FreeBSD_version for the afore mentioned and in kernel changes. Special thanks to: - Pawel Jakub Dawidek (pjd) for his multi-IPv4 patches and Olivier Houchard (cognet) for initial single-IPv6 patches. - Jeff Roberson (jeff) and Randall Stewart (rrs) for their help, ideas and review on cpuset and SCTP support. - Robert Watson (rwatson) for lots and lots of help, discussions, suggestions and review of most of the patch at various stages. - John Baldwin (jhb) for his help. - Simon L. Nielsen (simon) as early adopter testing changes on cluster machines as well as all the testers and people who provided feedback the last months on freebsd-jail and other channels. - My employer, CK Software GmbH, for the support so I could work on this. Reviewed by: (see above) MFC after: 3 months (this is just so that I get the mail) X-MFC Before: 7.2-RELEASE if possible
2008-11-29 14:32:14 +00:00
#ifdef INET6
#include <netinet6/ip6_var.h>
MFp4: Bring in updated jail support from bz_jail branch. This enhances the current jail implementation to permit multiple addresses per jail. In addtion to IPv4, IPv6 is supported as well. Due to updated checks it is even possible to have jails without an IP address at all, which basically gives one a chroot with restricted process view, no networking,.. SCTP support was updated and supports IPv6 in jails as well. Cpuset support permits jails to be bound to specific processor sets after creation. Jails can have an unrestricted (no duplicate protection, etc.) name in addition to the hostname. The jail name cannot be changed from within a jail and is considered to be used for management purposes or as audit-token in the future. DDB 'show jails' command was added to aid debugging. Proper compat support permits 32bit jail binaries to be used on 64bit systems to manage jails. Also backward compatibility was preserved where possible: for jail v1 syscalls, as well as with user space management utilities. Both jail as well as prison version were updated for the new features. A gap was intentionally left as the intermediate versions had been used by various patches floating around the last years. Bump __FreeBSD_version for the afore mentioned and in kernel changes. Special thanks to: - Pawel Jakub Dawidek (pjd) for his multi-IPv4 patches and Olivier Houchard (cognet) for initial single-IPv6 patches. - Jeff Roberson (jeff) and Randall Stewart (rrs) for their help, ideas and review on cpuset and SCTP support. - Robert Watson (rwatson) for lots and lots of help, discussions, suggestions and review of most of the patch at various stages. - John Baldwin (jhb) for his help. - Simon L. Nielsen (simon) as early adopter testing changes on cluster machines as well as all the testers and people who provided feedback the last months on freebsd-jail and other channels. - My employer, CK Software GmbH, for the support so I could work on this. Reviewed by: (see above) MFC after: 3 months (this is just so that I get the mail) X-MFC Before: 7.2-RELEASE if possible
2008-11-29 14:32:14 +00:00
#include <netinet6/scope6_var.h>
#endif
#ifdef COMPAT_FREEBSD32
#include <sys/mount.h>
#include <compat/freebsd32/freebsd32.h>
struct if_msghdr32 {
uint16_t ifm_msglen;
uint8_t ifm_version;
uint8_t ifm_type;
int32_t ifm_addrs;
int32_t ifm_flags;
uint16_t ifm_index;
uint16_t _ifm_spare1;
struct if_data ifm_data;
};
struct if_msghdrl32 {
uint16_t ifm_msglen;
uint8_t ifm_version;
uint8_t ifm_type;
int32_t ifm_addrs;
int32_t ifm_flags;
uint16_t ifm_index;
uint16_t _ifm_spare1;
uint16_t ifm_len;
uint16_t ifm_data_off;
uint32_t _ifm_spare2;
struct if_data ifm_data;
};
struct ifa_msghdrl32 {
uint16_t ifam_msglen;
uint8_t ifam_version;
uint8_t ifam_type;
int32_t ifam_addrs;
int32_t ifam_flags;
uint16_t ifam_index;
uint16_t _ifam_spare1;
uint16_t ifam_len;
uint16_t ifam_data_off;
int32_t ifam_metric;
struct if_data ifam_data;
};
#define SA_SIZE32(sa) \
( (((struct sockaddr *)(sa))->sa_len == 0) ? \
sizeof(int) : \
1 + ( (((struct sockaddr *)(sa))->sa_len - 1) | (sizeof(int) - 1) ) )
#endif /* COMPAT_FREEBSD32 */
MALLOC_DEFINE(M_RTABLE, "routetbl", "routing tables");
/* NB: these are not modified */
1995-11-16 19:00:27 +00:00
static struct sockaddr route_src = { 2, PF_ROUTE, };
static struct sockaddr sa_zero = { sizeof(sa_zero), AF_INET, };
/* These are external hooks for CARP. */
int (*carp_get_vhid_p)(struct ifaddr *);
/*
* Used by rtsock/raw_input callback code to decide whether to filter the update
* notification to a socket bound to a particular FIB.
*/
#define RTS_FILTER_FIB M_PROTO8
typedef struct {
int ip_count; /* attached w/ AF_INET */
int ip6_count; /* attached w/ AF_INET6 */
int any_count; /* total attached */
} route_cb_t;
VNET_DEFINE_STATIC(route_cb_t, route_cb);
#define V_route_cb VNET(route_cb)
1994-05-24 10:09:53 +00:00
struct mtx rtsock_mtx;
MTX_SYSINIT(rtsock, &rtsock_mtx, "rtsock route_cb lock", MTX_DEF);
#define RTSOCK_LOCK() mtx_lock(&rtsock_mtx)
#define RTSOCK_UNLOCK() mtx_unlock(&rtsock_mtx)
#define RTSOCK_LOCK_ASSERT() mtx_assert(&rtsock_mtx, MA_OWNED)
static SYSCTL_NODE(_net, OID_AUTO, route, CTLFLAG_RD, 0, "");
1994-05-24 10:09:53 +00:00
struct walkarg {
1995-11-16 19:00:27 +00:00
int w_tmemsize;
int w_op, w_arg;
caddr_t w_tmem;
struct sysctl_req *w_req;
1994-05-24 10:09:53 +00:00
};
static void rts_input(struct mbuf *m);
static struct mbuf *rtsock_msg_mbuf(int type, struct rt_addrinfo *rtinfo);
static int rtsock_msg_buffer(int type, struct rt_addrinfo *rtinfo,
struct walkarg *w, int *plen);
static int rt_xaddrs(caddr_t cp, caddr_t cplim,
struct rt_addrinfo *rtinfo);
2002-03-19 21:54:18 +00:00
static int sysctl_dumpentry(struct radix_node *rn, void *vw);
static int sysctl_iflist(int af, struct walkarg *w);
static int sysctl_ifmalist(int af, struct walkarg *w);
static int route_output(struct mbuf *m, struct socket *so, ...);
static void rt_getmetrics(const struct rtentry *rt, struct rt_metrics *out);
static void rt_dispatch(struct mbuf *, sa_family_t);
static struct sockaddr *rtsock_fix_netmask(struct sockaddr *dst,
struct sockaddr *smask, struct sockaddr_storage *dmask);
1994-05-24 10:09:53 +00:00
Reimplement the netisr framework in order to support parallel netisr threads: - Support up to one netisr thread per CPU, each processings its own workstream, or set of per-protocol queues. Threads may be bound to specific CPUs, or allowed to migrate, based on a global policy. In the future it would be desirable to support topology-centric policies, such as "one netisr per package". - Allow each protocol to advertise an ordering policy, which can currently be one of: NETISR_POLICY_SOURCE: packets must maintain ordering with respect to an implicit or explicit source (such as an interface or socket). NETISR_POLICY_FLOW: make use of mbuf flow identifiers to place work, as well as allowing protocols to provide a flow generation function for mbufs without flow identifers (m2flow). Falls back on NETISR_POLICY_SOURCE if now flow ID is available. NETISR_POLICY_CPU: allow protocols to inspect and assign a CPU for each packet handled by netisr (m2cpuid). - Provide utility functions for querying the number of workstreams being used, as well as a mapping function from workstream to CPU ID, which protocols may use in work placement decisions. - Add explicit interfaces to get and set per-protocol queue limits, and get and clear drop counters, which query data or apply changes across all workstreams. - Add a more extensible netisr registration interface, in which protocols declare 'struct netisr_handler' structures for each registered NETISR_ type. These include name, handler function, optional mbuf to flow ID function, optional mbuf to CPU ID function, queue limit, and ordering policy. Padding is present to allow these to be expanded in the future. If no queue limit is declared, then a default is used. - Queue limits are now per-workstream, and raised from the previous IFQ_MAXLEN default of 50 to 256. - All protocols are updated to use the new registration interface, and with the exception of netnatm, default queue limits. Most protocols register as NETISR_POLICY_SOURCE, except IPv4 and IPv6, which use NETISR_POLICY_FLOW, and will therefore take advantage of driver- generated flow IDs if present. - Formalize a non-packet based interface between interface polling and the netisr, rather than having polling pretend to be two protocols. Provide two explicit hooks in the netisr worker for start and end events for runs: netisr_poll() and netisr_pollmore(), as well as a function, netisr_sched_poll(), to allow the polling code to schedule netisr execution. DEVICE_POLLING still embeds single-netisr assumptions in its implementation, so for now if it is compiled into the kernel, a single and un-bound netisr thread is enforced regardless of tunable configuration. In the default configuration, the new netisr implementation maintains the same basic assumptions as the previous implementation: a single, un-bound worker thread processes all deferred work, and direct dispatch is enabled by default wherever possible. Performance measurement shows a marginal performance improvement over the old implementation due to the use of batched dequeue. An rmlock is used to synchronize use and registration/unregistration using the framework; currently, synchronized use is disabled (replicating current netisr policy) due to a measurable 3%-6% hit in ping-pong micro-benchmarking. It will be enabled once further rmlock optimization has taken place. However, in practice, netisrs are rarely registered or unregistered at runtime. A new man page for netisr will follow, but since one doesn't currently exist, it hasn't been updated. This change is not appropriate for MFC, although the polling shutdown handler should be merged to 7-STABLE. Bump __FreeBSD_version. Reviewed by: bz
2009-06-01 10:41:38 +00:00
static struct netisr_handler rtsock_nh = {
.nh_name = "rtsock",
.nh_handler = rts_input,
.nh_proto = NETISR_ROUTE,
.nh_policy = NETISR_POLICY_SOURCE,
};
static int
sysctl_route_netisr_maxqlen(SYSCTL_HANDLER_ARGS)
{
int error, qlimit;
netisr_getqlimit(&rtsock_nh, &qlimit);
error = sysctl_handle_int(oidp, &qlimit, 0, req);
if (error || !req->newptr)
return (error);
if (qlimit < 1)
return (EINVAL);
return (netisr_setqlimit(&rtsock_nh, qlimit));
}
SYSCTL_PROC(_net_route, OID_AUTO, netisr_maxqlen, CTLTYPE_INT|CTLFLAG_RW,
0, 0, sysctl_route_netisr_maxqlen, "I",
"maximum routing socket dispatch queue length");
static void
vnet_rts_init(void)
{
int tmp;
if (IS_DEFAULT_VNET(curvnet)) {
if (TUNABLE_INT_FETCH("net.route.netisr_maxqlen", &tmp))
rtsock_nh.nh_qlimit = tmp;
netisr_register(&rtsock_nh);
}
#ifdef VIMAGE
else
netisr_register_vnet(&rtsock_nh);
#endif
}
VNET_SYSINIT(vnet_rtsock, SI_SUB_PROTO_DOMAIN, SI_ORDER_THIRD,
vnet_rts_init, 0);
#ifdef VIMAGE
static void
vnet_rts_uninit(void)
{
netisr_unregister_vnet(&rtsock_nh);
}
VNET_SYSUNINIT(vnet_rts_uninit, SI_SUB_PROTO_DOMAIN, SI_ORDER_THIRD,
vnet_rts_uninit, 0);
#endif
static int
raw_input_rts_cb(struct mbuf *m, struct sockproto *proto, struct sockaddr *src,
struct rawcb *rp)
{
int fibnum;
KASSERT(m != NULL, ("%s: m is NULL", __func__));
KASSERT(proto != NULL, ("%s: proto is NULL", __func__));
KASSERT(rp != NULL, ("%s: rp is NULL", __func__));
/* No filtering requested. */
if ((m->m_flags & RTS_FILTER_FIB) == 0)
return (0);
/* Check if it is a rts and the fib matches the one of the socket. */
fibnum = M_GETFIB(m);
if (proto->sp_family != PF_ROUTE ||
rp->rcb_socket == NULL ||
rp->rcb_socket->so_fibnum == fibnum)
return (0);
/* Filtering requested and no match, the socket shall be skipped. */
return (1);
}
static void
rts_input(struct mbuf *m)
{
struct sockproto route_proto;
unsigned short *family;
struct m_tag *tag;
route_proto.sp_family = PF_ROUTE;
tag = m_tag_find(m, PACKET_TAG_RTSOCKFAM, NULL);
if (tag != NULL) {
family = (unsigned short *)(tag + 1);
route_proto.sp_protocol = *family;
m_tag_delete(m, tag);
} else
route_proto.sp_protocol = 0;
raw_input_ext(m, &route_proto, &route_src, raw_input_rts_cb);
}
/*
* It really doesn't make any sense at all for this code to share much
* with raw_usrreq.c, since its functionality is so restricted. XXX
*/
static void
rts_abort(struct socket *so)
{
raw_usrreqs.pru_abort(so);
}
static void
rts_close(struct socket *so)
{
raw_usrreqs.pru_close(so);
}
/* pru_accept is EOPNOTSUPP */
1995-11-16 19:00:27 +00:00
static int
rts_attach(struct socket *so, int proto, struct thread *td)
1994-05-24 10:09:53 +00:00
{
struct rawcb *rp;
int error;
KASSERT(so->so_pcb == NULL, ("rts_attach: so_pcb != NULL"));
/* XXX */
rp = malloc(sizeof *rp, M_PCB, M_WAITOK | M_ZERO);
so->so_pcb = (caddr_t)rp;
Add code to allow the system to handle multiple routing tables. This particular implementation is designed to be fully backwards compatible and to be MFC-able to 7.x (and 6.x) Currently the only protocol that can make use of the multiple tables is IPv4 Similar functionality exists in OpenBSD and Linux. From my notes: ----- One thing where FreeBSD has been falling behind, and which by chance I have some time to work on is "policy based routing", which allows different packet streams to be routed by more than just the destination address. Constraints: ------------ I want to make some form of this available in the 6.x tree (and by extension 7.x) , but FreeBSD in general needs it so I might as well do it in -current and back port the portions I need. One of the ways that this can be done is to have the ability to instantiate multiple kernel routing tables (which I will now refer to as "Forwarding Information Bases" or "FIBs" for political correctness reasons). Which FIB a particular packet uses to make the next hop decision can be decided by a number of mechanisms. The policies these mechanisms implement are the "Policies" referred to in "Policy based routing". One of the constraints I have if I try to back port this work to 6.x is that it must be implemented as a EXTENSION to the existing ABIs in 6.x so that third party applications do not need to be recompiled in timespan of the branch. This first version will not have some of the bells and whistles that will come with later versions. It will, for example, be limited to 16 tables in the first commit. Implementation method, Compatible version. (part 1) ------------------------------- For this reason I have implemented a "sufficient subset" of a multiple routing table solution in Perforce, and back-ported it to 6.x. (also in Perforce though not always caught up with what I have done in -current/P4). The subset allows a number of FIBs to be defined at compile time (8 is sufficient for my purposes in 6.x) and implements the changes needed to allow IPV4 to use them. I have not done the changes for ipv6 simply because I do not need it, and I do not have enough knowledge of ipv6 (e.g. neighbor discovery) needed to do it. Other protocol families are left untouched and should there be users with proprietary protocol families, they should continue to work and be oblivious to the existence of the extra FIBs. To understand how this is done, one must know that the current FIB code starts everything off with a single dimensional array of pointers to FIB head structures (One per protocol family), each of which in turn points to the trie of routes available to that family. The basic change in the ABI compatible version of the change is to extent that array to be a 2 dimensional array, so that instead of protocol family X looking at rt_tables[X] for the table it needs, it looks at rt_tables[Y][X] when for all protocol families except ipv4 Y is always 0. Code that is unaware of the change always just sees the first row of the table, which of course looks just like the one dimensional array that existed before. The entry points rtrequest(), rtalloc(), rtalloc1(), rtalloc_ign() are all maintained, but refer only to the first row of the array, so that existing callers in proprietary protocols can continue to do the "right thing". Some new entry points are added, for the exclusive use of ipv4 code called in_rtrequest(), in_rtalloc(), in_rtalloc1() and in_rtalloc_ign(), which have an extra argument which refers the code to the correct row. In addition, there are some new entry points (currently called rtalloc_fib() and friends) that check the Address family being looked up and call either rtalloc() (and friends) if the protocol is not IPv4 forcing the action to row 0 or to the appropriate row if it IS IPv4 (and that info is available). These are for calling from code that is not specific to any particular protocol. The way these are implemented would change in the non ABI preserving code to be added later. One feature of the first version of the code is that for ipv4, the interface routes show up automatically on all the FIBs, so that no matter what FIB you select you always have the basic direct attached hosts available to you. (rtinit() does this automatically). You CAN delete an interface route from one FIB should you want to but by default it's there. ARP information is also available in each FIB. It's assumed that the same machine would have the same MAC address, regardless of which FIB you are using to get to it. This brings us as to how the correct FIB is selected for an outgoing IPV4 packet. Firstly, all packets have a FIB associated with them. if nothing has been done to change it, it will be FIB 0. The FIB is changed in the following ways. Packets fall into one of a number of classes. 1/ locally generated packets, coming from a socket/PCB. Such packets select a FIB from a number associated with the socket/PCB. This in turn is inherited from the process, but can be changed by a socket option. The process in turn inherits it on fork. I have written a utility call setfib that acts a bit like nice.. setfib -3 ping target.example.com # will use fib 3 for ping. It is an obvious extension to make it a property of a jail but I have not done so. It can be achieved by combining the setfib and jail commands. 2/ packets received on an interface for forwarding. By default these packets would use table 0, (or possibly a number settable in a sysctl(not yet)). but prior to routing the firewall can inspect them (see below). (possibly in the future you may be able to associate a FIB with packets received on an interface.. An ifconfig arg, but not yet.) 3/ packets inspected by a packet classifier, which can arbitrarily associate a fib with it on a packet by packet basis. A fib assigned to a packet by a packet classifier (such as ipfw) would over-ride a fib associated by a more default source. (such as cases 1 or 2). 4/ a tcp listen socket associated with a fib will generate accept sockets that are associated with that same fib. 5/ Packets generated in response to some other packet (e.g. reset or icmp packets). These should use the FIB associated with the packet being reponded to. 6/ Packets generated during encapsulation. gif, tun and other tunnel interfaces will encapsulate using the FIB that was in effect withthe proces that set up the tunnel. thus setfib 1 ifconfig gif0 [tunnel instructions] will set the fib for the tunnel to use to be fib 1. Routing messages would be associated with their process, and thus select one FIB or another. messages from the kernel would be associated with the fib they refer to and would only be received by a routing socket associated with that fib. (not yet implemented) In addition Netstat has been edited to be able to cope with the fact that the array is now 2 dimensional. (It looks in system memory using libkvm (!)). Old versions of netstat see only the first FIB. In addition two sysctls are added to give: a) the number of FIBs compiled in (active) b) the default FIB of the calling process. Early testing experience: ------------------------- Basically our (IronPort's) appliance does this functionality already using ipfw fwd but that method has some drawbacks. For example, It can't fully simulate a routing table because it can't influence the socket's choice of local address when a connect() is done. Testing during the generating of these changes has been remarkably smooth so far. Multiple tables have co-existed with no notable side effects, and packets have been routes accordingly. ipfw has grown 2 new keywords: setfib N ip from anay to any count ip from any to any fib N In pf there seems to be a requirement to be able to give symbolic names to the fibs but I do not have that capacity. I am not sure if it is required. SCTP has interestingly enough built in support for this, called VRFs in Cisco parlance. it will be interesting to see how that handles it when it suddenly actually does something. Where to next: -------------------- After committing the ABI compatible version and MFCing it, I'd like to proceed in a forward direction in -current. this will result in some roto-tilling in the routing code. Firstly: the current code's idea of having a separate tree per protocol family, all of the same format, and pointed to by the 1 dimensional array is a bit silly. Especially when one considers that there is code that makes assumptions about every protocol having the same internal structures there. Some protocols don't WANT that sort of structure. (for example the whole idea of a netmask is foreign to appletalk). This needs to be made opaque to the external code. My suggested first change is to add routing method pointers to the 'domain' structure, along with information pointing the data. instead of having an array of pointers to uniform structures, there would be an array pointing to the 'domain' structures for each protocol address domain (protocol family), and the methods this reached would be called. The methods would have an argument that gives FIB number, but the protocol would be free to ignore it. When the ABI can be changed it raises the possibilty of the addition of a fib entry into the "struct route". Currently, the structure contains the sockaddr of the desination, and the resulting fib entry. To make this work fully, one could add a fib number so that given an address and a fib, one can find the third element, the fib entry. Interaction with the ARP layer/ LL layer would need to be revisited as well. Qing Li has been working on this already. This work was sponsored by Ironport Systems/Cisco Reviewed by: several including rwatson, bz and mlair (parts each) Obtained from: Ironport systems/Cisco
2008-05-09 23:03:00 +00:00
so->so_fibnum = td->td_proc->p_fibnum;
error = raw_attach(so, proto);
rp = sotorawcb(so);
if (error) {
so->so_pcb = NULL;
free(rp, M_PCB);
return error;
}
RTSOCK_LOCK();
switch(rp->rcb_proto.sp_protocol) {
case AF_INET:
V_route_cb.ip_count++;
break;
case AF_INET6:
V_route_cb.ip6_count++;
break;
1994-05-24 10:09:53 +00:00
}
V_route_cb.any_count++;
RTSOCK_UNLOCK();
soisconnected(so);
so->so_options |= SO_USELOOPBACK;
return 0;
}
static int
rts_bind(struct socket *so, struct sockaddr *nam, struct thread *td)
{
return (raw_usrreqs.pru_bind(so, nam, td)); /* xxx just EINVAL */
}
static int
rts_connect(struct socket *so, struct sockaddr *nam, struct thread *td)
{
return (raw_usrreqs.pru_connect(so, nam, td)); /* XXX just EINVAL */
}
/* pru_connect2 is EOPNOTSUPP */
/* pru_control is EOPNOTSUPP */
static void
rts_detach(struct socket *so)
{
struct rawcb *rp = sotorawcb(so);
KASSERT(rp != NULL, ("rts_detach: rp == NULL"));
RTSOCK_LOCK();
switch(rp->rcb_proto.sp_protocol) {
case AF_INET:
V_route_cb.ip_count--;
break;
case AF_INET6:
V_route_cb.ip6_count--;
break;
1994-05-24 10:09:53 +00:00
}
V_route_cb.any_count--;
RTSOCK_UNLOCK();
raw_usrreqs.pru_detach(so);
}
static int
rts_disconnect(struct socket *so)
{
return (raw_usrreqs.pru_disconnect(so));
1994-05-24 10:09:53 +00:00
}
/* pru_listen is EOPNOTSUPP */
static int
rts_peeraddr(struct socket *so, struct sockaddr **nam)
{
return (raw_usrreqs.pru_peeraddr(so, nam));
}
/* pru_rcvd is EOPNOTSUPP */
/* pru_rcvoob is EOPNOTSUPP */
static int
rts_send(struct socket *so, int flags, struct mbuf *m, struct sockaddr *nam,
struct mbuf *control, struct thread *td)
{
return (raw_usrreqs.pru_send(so, flags, m, nam, control, td));
}
/* pru_sense is null */
static int
rts_shutdown(struct socket *so)
{
return (raw_usrreqs.pru_shutdown(so));
}
static int
rts_sockaddr(struct socket *so, struct sockaddr **nam)
{
return (raw_usrreqs.pru_sockaddr(so, nam));
}
static struct pr_usrreqs route_usrreqs = {
.pru_abort = rts_abort,
.pru_attach = rts_attach,
.pru_bind = rts_bind,
.pru_connect = rts_connect,
.pru_detach = rts_detach,
.pru_disconnect = rts_disconnect,
.pru_peeraddr = rts_peeraddr,
.pru_send = rts_send,
.pru_shutdown = rts_shutdown,
.pru_sockaddr = rts_sockaddr,
.pru_close = rts_close,
};
MFp4: Bring in updated jail support from bz_jail branch. This enhances the current jail implementation to permit multiple addresses per jail. In addtion to IPv4, IPv6 is supported as well. Due to updated checks it is even possible to have jails without an IP address at all, which basically gives one a chroot with restricted process view, no networking,.. SCTP support was updated and supports IPv6 in jails as well. Cpuset support permits jails to be bound to specific processor sets after creation. Jails can have an unrestricted (no duplicate protection, etc.) name in addition to the hostname. The jail name cannot be changed from within a jail and is considered to be used for management purposes or as audit-token in the future. DDB 'show jails' command was added to aid debugging. Proper compat support permits 32bit jail binaries to be used on 64bit systems to manage jails. Also backward compatibility was preserved where possible: for jail v1 syscalls, as well as with user space management utilities. Both jail as well as prison version were updated for the new features. A gap was intentionally left as the intermediate versions had been used by various patches floating around the last years. Bump __FreeBSD_version for the afore mentioned and in kernel changes. Special thanks to: - Pawel Jakub Dawidek (pjd) for his multi-IPv4 patches and Olivier Houchard (cognet) for initial single-IPv6 patches. - Jeff Roberson (jeff) and Randall Stewart (rrs) for their help, ideas and review on cpuset and SCTP support. - Robert Watson (rwatson) for lots and lots of help, discussions, suggestions and review of most of the patch at various stages. - John Baldwin (jhb) for his help. - Simon L. Nielsen (simon) as early adopter testing changes on cluster machines as well as all the testers and people who provided feedback the last months on freebsd-jail and other channels. - My employer, CK Software GmbH, for the support so I could work on this. Reviewed by: (see above) MFC after: 3 months (this is just so that I get the mail) X-MFC Before: 7.2-RELEASE if possible
2008-11-29 14:32:14 +00:00
#ifndef _SOCKADDR_UNION_DEFINED
#define _SOCKADDR_UNION_DEFINED
/*
* The union of all possible address formats we handle.
*/
union sockaddr_union {
struct sockaddr sa;
struct sockaddr_in sin;
struct sockaddr_in6 sin6;
};
#endif /* _SOCKADDR_UNION_DEFINED */
static int
rtm_get_jailed(struct rt_addrinfo *info, struct ifnet *ifp,
struct rtentry *rt, union sockaddr_union *saun, struct ucred *cred)
{
#if defined(INET) || defined(INET6)
struct epoch_tracker et;
#endif
MFp4: Bring in updated jail support from bz_jail branch. This enhances the current jail implementation to permit multiple addresses per jail. In addtion to IPv4, IPv6 is supported as well. Due to updated checks it is even possible to have jails without an IP address at all, which basically gives one a chroot with restricted process view, no networking,.. SCTP support was updated and supports IPv6 in jails as well. Cpuset support permits jails to be bound to specific processor sets after creation. Jails can have an unrestricted (no duplicate protection, etc.) name in addition to the hostname. The jail name cannot be changed from within a jail and is considered to be used for management purposes or as audit-token in the future. DDB 'show jails' command was added to aid debugging. Proper compat support permits 32bit jail binaries to be used on 64bit systems to manage jails. Also backward compatibility was preserved where possible: for jail v1 syscalls, as well as with user space management utilities. Both jail as well as prison version were updated for the new features. A gap was intentionally left as the intermediate versions had been used by various patches floating around the last years. Bump __FreeBSD_version for the afore mentioned and in kernel changes. Special thanks to: - Pawel Jakub Dawidek (pjd) for his multi-IPv4 patches and Olivier Houchard (cognet) for initial single-IPv6 patches. - Jeff Roberson (jeff) and Randall Stewart (rrs) for their help, ideas and review on cpuset and SCTP support. - Robert Watson (rwatson) for lots and lots of help, discussions, suggestions and review of most of the patch at various stages. - John Baldwin (jhb) for his help. - Simon L. Nielsen (simon) as early adopter testing changes on cluster machines as well as all the testers and people who provided feedback the last months on freebsd-jail and other channels. - My employer, CK Software GmbH, for the support so I could work on this. Reviewed by: (see above) MFC after: 3 months (this is just so that I get the mail) X-MFC Before: 7.2-RELEASE if possible
2008-11-29 14:32:14 +00:00
/* First, see if the returned address is part of the jail. */
if (prison_if(cred, rt->rt_ifa->ifa_addr) == 0) {
info->rti_info[RTAX_IFA] = rt->rt_ifa->ifa_addr;
return (0);
}
MFp4: Bring in updated jail support from bz_jail branch. This enhances the current jail implementation to permit multiple addresses per jail. In addtion to IPv4, IPv6 is supported as well. Due to updated checks it is even possible to have jails without an IP address at all, which basically gives one a chroot with restricted process view, no networking,.. SCTP support was updated and supports IPv6 in jails as well. Cpuset support permits jails to be bound to specific processor sets after creation. Jails can have an unrestricted (no duplicate protection, etc.) name in addition to the hostname. The jail name cannot be changed from within a jail and is considered to be used for management purposes or as audit-token in the future. DDB 'show jails' command was added to aid debugging. Proper compat support permits 32bit jail binaries to be used on 64bit systems to manage jails. Also backward compatibility was preserved where possible: for jail v1 syscalls, as well as with user space management utilities. Both jail as well as prison version were updated for the new features. A gap was intentionally left as the intermediate versions had been used by various patches floating around the last years. Bump __FreeBSD_version for the afore mentioned and in kernel changes. Special thanks to: - Pawel Jakub Dawidek (pjd) for his multi-IPv4 patches and Olivier Houchard (cognet) for initial single-IPv6 patches. - Jeff Roberson (jeff) and Randall Stewart (rrs) for their help, ideas and review on cpuset and SCTP support. - Robert Watson (rwatson) for lots and lots of help, discussions, suggestions and review of most of the patch at various stages. - John Baldwin (jhb) for his help. - Simon L. Nielsen (simon) as early adopter testing changes on cluster machines as well as all the testers and people who provided feedback the last months on freebsd-jail and other channels. - My employer, CK Software GmbH, for the support so I could work on this. Reviewed by: (see above) MFC after: 3 months (this is just so that I get the mail) X-MFC Before: 7.2-RELEASE if possible
2008-11-29 14:32:14 +00:00
switch (info->rti_info[RTAX_DST]->sa_family) {
#ifdef INET
case AF_INET:
{
struct in_addr ia;
struct ifaddr *ifa;
int found;
MFp4: Bring in updated jail support from bz_jail branch. This enhances the current jail implementation to permit multiple addresses per jail. In addtion to IPv4, IPv6 is supported as well. Due to updated checks it is even possible to have jails without an IP address at all, which basically gives one a chroot with restricted process view, no networking,.. SCTP support was updated and supports IPv6 in jails as well. Cpuset support permits jails to be bound to specific processor sets after creation. Jails can have an unrestricted (no duplicate protection, etc.) name in addition to the hostname. The jail name cannot be changed from within a jail and is considered to be used for management purposes or as audit-token in the future. DDB 'show jails' command was added to aid debugging. Proper compat support permits 32bit jail binaries to be used on 64bit systems to manage jails. Also backward compatibility was preserved where possible: for jail v1 syscalls, as well as with user space management utilities. Both jail as well as prison version were updated for the new features. A gap was intentionally left as the intermediate versions had been used by various patches floating around the last years. Bump __FreeBSD_version for the afore mentioned and in kernel changes. Special thanks to: - Pawel Jakub Dawidek (pjd) for his multi-IPv4 patches and Olivier Houchard (cognet) for initial single-IPv6 patches. - Jeff Roberson (jeff) and Randall Stewart (rrs) for their help, ideas and review on cpuset and SCTP support. - Robert Watson (rwatson) for lots and lots of help, discussions, suggestions and review of most of the patch at various stages. - John Baldwin (jhb) for his help. - Simon L. Nielsen (simon) as early adopter testing changes on cluster machines as well as all the testers and people who provided feedback the last months on freebsd-jail and other channels. - My employer, CK Software GmbH, for the support so I could work on this. Reviewed by: (see above) MFC after: 3 months (this is just so that I get the mail) X-MFC Before: 7.2-RELEASE if possible
2008-11-29 14:32:14 +00:00
found = 0;
MFp4: Bring in updated jail support from bz_jail branch. This enhances the current jail implementation to permit multiple addresses per jail. In addtion to IPv4, IPv6 is supported as well. Due to updated checks it is even possible to have jails without an IP address at all, which basically gives one a chroot with restricted process view, no networking,.. SCTP support was updated and supports IPv6 in jails as well. Cpuset support permits jails to be bound to specific processor sets after creation. Jails can have an unrestricted (no duplicate protection, etc.) name in addition to the hostname. The jail name cannot be changed from within a jail and is considered to be used for management purposes or as audit-token in the future. DDB 'show jails' command was added to aid debugging. Proper compat support permits 32bit jail binaries to be used on 64bit systems to manage jails. Also backward compatibility was preserved where possible: for jail v1 syscalls, as well as with user space management utilities. Both jail as well as prison version were updated for the new features. A gap was intentionally left as the intermediate versions had been used by various patches floating around the last years. Bump __FreeBSD_version for the afore mentioned and in kernel changes. Special thanks to: - Pawel Jakub Dawidek (pjd) for his multi-IPv4 patches and Olivier Houchard (cognet) for initial single-IPv6 patches. - Jeff Roberson (jeff) and Randall Stewart (rrs) for their help, ideas and review on cpuset and SCTP support. - Robert Watson (rwatson) for lots and lots of help, discussions, suggestions and review of most of the patch at various stages. - John Baldwin (jhb) for his help. - Simon L. Nielsen (simon) as early adopter testing changes on cluster machines as well as all the testers and people who provided feedback the last months on freebsd-jail and other channels. - My employer, CK Software GmbH, for the support so I could work on this. Reviewed by: (see above) MFC after: 3 months (this is just so that I get the mail) X-MFC Before: 7.2-RELEASE if possible
2008-11-29 14:32:14 +00:00
/*
* Try to find an address on the given outgoing interface
* that belongs to the jail.
MFp4: Bring in updated jail support from bz_jail branch. This enhances the current jail implementation to permit multiple addresses per jail. In addtion to IPv4, IPv6 is supported as well. Due to updated checks it is even possible to have jails without an IP address at all, which basically gives one a chroot with restricted process view, no networking,.. SCTP support was updated and supports IPv6 in jails as well. Cpuset support permits jails to be bound to specific processor sets after creation. Jails can have an unrestricted (no duplicate protection, etc.) name in addition to the hostname. The jail name cannot be changed from within a jail and is considered to be used for management purposes or as audit-token in the future. DDB 'show jails' command was added to aid debugging. Proper compat support permits 32bit jail binaries to be used on 64bit systems to manage jails. Also backward compatibility was preserved where possible: for jail v1 syscalls, as well as with user space management utilities. Both jail as well as prison version were updated for the new features. A gap was intentionally left as the intermediate versions had been used by various patches floating around the last years. Bump __FreeBSD_version for the afore mentioned and in kernel changes. Special thanks to: - Pawel Jakub Dawidek (pjd) for his multi-IPv4 patches and Olivier Houchard (cognet) for initial single-IPv6 patches. - Jeff Roberson (jeff) and Randall Stewart (rrs) for their help, ideas and review on cpuset and SCTP support. - Robert Watson (rwatson) for lots and lots of help, discussions, suggestions and review of most of the patch at various stages. - John Baldwin (jhb) for his help. - Simon L. Nielsen (simon) as early adopter testing changes on cluster machines as well as all the testers and people who provided feedback the last months on freebsd-jail and other channels. - My employer, CK Software GmbH, for the support so I could work on this. Reviewed by: (see above) MFC after: 3 months (this is just so that I get the mail) X-MFC Before: 7.2-RELEASE if possible
2008-11-29 14:32:14 +00:00
*/
NET_EPOCH_ENTER(et);
ifnet: Replace if_addr_lock rwlock with epoch + mutex Run on LLNW canaries and tested by pho@ gallatin: Using a 14-core, 28-HTT single socket E5-2697 v3 with a 40GbE MLX5 based ConnectX 4-LX NIC, I see an almost 12% improvement in received packet rate, and a larger improvement in bytes delivered all the way to userspace. When the host receiving 64 streams of netperf -H $DUT -t UDP_STREAM -- -m 1, I see, using nstat -I mce0 1 before the patch: InMpps OMpps InGbs OGbs err TCP Est %CPU syscalls csw irq GBfree 4.98 0.00 4.42 0.00 4235592 33 83.80 4720653 2149771 1235 247.32 4.73 0.00 4.20 0.00 4025260 33 82.99 4724900 2139833 1204 247.32 4.72 0.00 4.20 0.00 4035252 33 82.14 4719162 2132023 1264 247.32 4.71 0.00 4.21 0.00 4073206 33 83.68 4744973 2123317 1347 247.32 4.72 0.00 4.21 0.00 4061118 33 80.82 4713615 2188091 1490 247.32 4.72 0.00 4.21 0.00 4051675 33 85.29 4727399 2109011 1205 247.32 4.73 0.00 4.21 0.00 4039056 33 84.65 4724735 2102603 1053 247.32 After the patch InMpps OMpps InGbs OGbs err TCP Est %CPU syscalls csw irq GBfree 5.43 0.00 4.20 0.00 3313143 33 84.96 5434214 1900162 2656 245.51 5.43 0.00 4.20 0.00 3308527 33 85.24 5439695 1809382 2521 245.51 5.42 0.00 4.19 0.00 3316778 33 87.54 5416028 1805835 2256 245.51 5.42 0.00 4.19 0.00 3317673 33 90.44 5426044 1763056 2332 245.51 5.42 0.00 4.19 0.00 3314839 33 88.11 5435732 1792218 2499 245.52 5.44 0.00 4.19 0.00 3293228 33 91.84 5426301 1668597 2121 245.52 Similarly, netperf reports 230Mb/s before the patch, and 270Mb/s after the patch Reviewed by: gallatin Sponsored by: Limelight Networks Differential Revision: https://reviews.freebsd.org/D15366
2018-05-18 20:13:34 +00:00
CK_STAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) {
struct sockaddr *sa;
sa = ifa->ifa_addr;
if (sa->sa_family != AF_INET)
continue;
ia = ((struct sockaddr_in *)sa)->sin_addr;
if (prison_check_ip4(cred, &ia) == 0) {
found = 1;
break;
}
}
NET_EPOCH_EXIT(et);
if (!found) {
MFp4: Bring in updated jail support from bz_jail branch. This enhances the current jail implementation to permit multiple addresses per jail. In addtion to IPv4, IPv6 is supported as well. Due to updated checks it is even possible to have jails without an IP address at all, which basically gives one a chroot with restricted process view, no networking,.. SCTP support was updated and supports IPv6 in jails as well. Cpuset support permits jails to be bound to specific processor sets after creation. Jails can have an unrestricted (no duplicate protection, etc.) name in addition to the hostname. The jail name cannot be changed from within a jail and is considered to be used for management purposes or as audit-token in the future. DDB 'show jails' command was added to aid debugging. Proper compat support permits 32bit jail binaries to be used on 64bit systems to manage jails. Also backward compatibility was preserved where possible: for jail v1 syscalls, as well as with user space management utilities. Both jail as well as prison version were updated for the new features. A gap was intentionally left as the intermediate versions had been used by various patches floating around the last years. Bump __FreeBSD_version for the afore mentioned and in kernel changes. Special thanks to: - Pawel Jakub Dawidek (pjd) for his multi-IPv4 patches and Olivier Houchard (cognet) for initial single-IPv6 patches. - Jeff Roberson (jeff) and Randall Stewart (rrs) for their help, ideas and review on cpuset and SCTP support. - Robert Watson (rwatson) for lots and lots of help, discussions, suggestions and review of most of the patch at various stages. - John Baldwin (jhb) for his help. - Simon L. Nielsen (simon) as early adopter testing changes on cluster machines as well as all the testers and people who provided feedback the last months on freebsd-jail and other channels. - My employer, CK Software GmbH, for the support so I could work on this. Reviewed by: (see above) MFC after: 3 months (this is just so that I get the mail) X-MFC Before: 7.2-RELEASE if possible
2008-11-29 14:32:14 +00:00
/*
* As a last resort return the 'default' jail address.
MFp4: Bring in updated jail support from bz_jail branch. This enhances the current jail implementation to permit multiple addresses per jail. In addtion to IPv4, IPv6 is supported as well. Due to updated checks it is even possible to have jails without an IP address at all, which basically gives one a chroot with restricted process view, no networking,.. SCTP support was updated and supports IPv6 in jails as well. Cpuset support permits jails to be bound to specific processor sets after creation. Jails can have an unrestricted (no duplicate protection, etc.) name in addition to the hostname. The jail name cannot be changed from within a jail and is considered to be used for management purposes or as audit-token in the future. DDB 'show jails' command was added to aid debugging. Proper compat support permits 32bit jail binaries to be used on 64bit systems to manage jails. Also backward compatibility was preserved where possible: for jail v1 syscalls, as well as with user space management utilities. Both jail as well as prison version were updated for the new features. A gap was intentionally left as the intermediate versions had been used by various patches floating around the last years. Bump __FreeBSD_version for the afore mentioned and in kernel changes. Special thanks to: - Pawel Jakub Dawidek (pjd) for his multi-IPv4 patches and Olivier Houchard (cognet) for initial single-IPv6 patches. - Jeff Roberson (jeff) and Randall Stewart (rrs) for their help, ideas and review on cpuset and SCTP support. - Robert Watson (rwatson) for lots and lots of help, discussions, suggestions and review of most of the patch at various stages. - John Baldwin (jhb) for his help. - Simon L. Nielsen (simon) as early adopter testing changes on cluster machines as well as all the testers and people who provided feedback the last months on freebsd-jail and other channels. - My employer, CK Software GmbH, for the support so I could work on this. Reviewed by: (see above) MFC after: 3 months (this is just so that I get the mail) X-MFC Before: 7.2-RELEASE if possible
2008-11-29 14:32:14 +00:00
*/
ia = ((struct sockaddr_in *)rt->rt_ifa->ifa_addr)->
sin_addr;
if (prison_get_ip4(cred, &ia) != 0)
return (ESRCH);
MFp4: Bring in updated jail support from bz_jail branch. This enhances the current jail implementation to permit multiple addresses per jail. In addtion to IPv4, IPv6 is supported as well. Due to updated checks it is even possible to have jails without an IP address at all, which basically gives one a chroot with restricted process view, no networking,.. SCTP support was updated and supports IPv6 in jails as well. Cpuset support permits jails to be bound to specific processor sets after creation. Jails can have an unrestricted (no duplicate protection, etc.) name in addition to the hostname. The jail name cannot be changed from within a jail and is considered to be used for management purposes or as audit-token in the future. DDB 'show jails' command was added to aid debugging. Proper compat support permits 32bit jail binaries to be used on 64bit systems to manage jails. Also backward compatibility was preserved where possible: for jail v1 syscalls, as well as with user space management utilities. Both jail as well as prison version were updated for the new features. A gap was intentionally left as the intermediate versions had been used by various patches floating around the last years. Bump __FreeBSD_version for the afore mentioned and in kernel changes. Special thanks to: - Pawel Jakub Dawidek (pjd) for his multi-IPv4 patches and Olivier Houchard (cognet) for initial single-IPv6 patches. - Jeff Roberson (jeff) and Randall Stewart (rrs) for their help, ideas and review on cpuset and SCTP support. - Robert Watson (rwatson) for lots and lots of help, discussions, suggestions and review of most of the patch at various stages. - John Baldwin (jhb) for his help. - Simon L. Nielsen (simon) as early adopter testing changes on cluster machines as well as all the testers and people who provided feedback the last months on freebsd-jail and other channels. - My employer, CK Software GmbH, for the support so I could work on this. Reviewed by: (see above) MFC after: 3 months (this is just so that I get the mail) X-MFC Before: 7.2-RELEASE if possible
2008-11-29 14:32:14 +00:00
}
bzero(&saun->sin, sizeof(struct sockaddr_in));
saun->sin.sin_len = sizeof(struct sockaddr_in);
saun->sin.sin_family = AF_INET;
saun->sin.sin_addr.s_addr = ia.s_addr;
info->rti_info[RTAX_IFA] = (struct sockaddr *)&saun->sin;
MFp4: Bring in updated jail support from bz_jail branch. This enhances the current jail implementation to permit multiple addresses per jail. In addtion to IPv4, IPv6 is supported as well. Due to updated checks it is even possible to have jails without an IP address at all, which basically gives one a chroot with restricted process view, no networking,.. SCTP support was updated and supports IPv6 in jails as well. Cpuset support permits jails to be bound to specific processor sets after creation. Jails can have an unrestricted (no duplicate protection, etc.) name in addition to the hostname. The jail name cannot be changed from within a jail and is considered to be used for management purposes or as audit-token in the future. DDB 'show jails' command was added to aid debugging. Proper compat support permits 32bit jail binaries to be used on 64bit systems to manage jails. Also backward compatibility was preserved where possible: for jail v1 syscalls, as well as with user space management utilities. Both jail as well as prison version were updated for the new features. A gap was intentionally left as the intermediate versions had been used by various patches floating around the last years. Bump __FreeBSD_version for the afore mentioned and in kernel changes. Special thanks to: - Pawel Jakub Dawidek (pjd) for his multi-IPv4 patches and Olivier Houchard (cognet) for initial single-IPv6 patches. - Jeff Roberson (jeff) and Randall Stewart (rrs) for their help, ideas and review on cpuset and SCTP support. - Robert Watson (rwatson) for lots and lots of help, discussions, suggestions and review of most of the patch at various stages. - John Baldwin (jhb) for his help. - Simon L. Nielsen (simon) as early adopter testing changes on cluster machines as well as all the testers and people who provided feedback the last months on freebsd-jail and other channels. - My employer, CK Software GmbH, for the support so I could work on this. Reviewed by: (see above) MFC after: 3 months (this is just so that I get the mail) X-MFC Before: 7.2-RELEASE if possible
2008-11-29 14:32:14 +00:00
break;
}
#endif
#ifdef INET6
case AF_INET6:
{
struct in6_addr ia6;
struct ifaddr *ifa;
int found;
MFp4: Bring in updated jail support from bz_jail branch. This enhances the current jail implementation to permit multiple addresses per jail. In addtion to IPv4, IPv6 is supported as well. Due to updated checks it is even possible to have jails without an IP address at all, which basically gives one a chroot with restricted process view, no networking,.. SCTP support was updated and supports IPv6 in jails as well. Cpuset support permits jails to be bound to specific processor sets after creation. Jails can have an unrestricted (no duplicate protection, etc.) name in addition to the hostname. The jail name cannot be changed from within a jail and is considered to be used for management purposes or as audit-token in the future. DDB 'show jails' command was added to aid debugging. Proper compat support permits 32bit jail binaries to be used on 64bit systems to manage jails. Also backward compatibility was preserved where possible: for jail v1 syscalls, as well as with user space management utilities. Both jail as well as prison version were updated for the new features. A gap was intentionally left as the intermediate versions had been used by various patches floating around the last years. Bump __FreeBSD_version for the afore mentioned and in kernel changes. Special thanks to: - Pawel Jakub Dawidek (pjd) for his multi-IPv4 patches and Olivier Houchard (cognet) for initial single-IPv6 patches. - Jeff Roberson (jeff) and Randall Stewart (rrs) for their help, ideas and review on cpuset and SCTP support. - Robert Watson (rwatson) for lots and lots of help, discussions, suggestions and review of most of the patch at various stages. - John Baldwin (jhb) for his help. - Simon L. Nielsen (simon) as early adopter testing changes on cluster machines as well as all the testers and people who provided feedback the last months on freebsd-jail and other channels. - My employer, CK Software GmbH, for the support so I could work on this. Reviewed by: (see above) MFC after: 3 months (this is just so that I get the mail) X-MFC Before: 7.2-RELEASE if possible
2008-11-29 14:32:14 +00:00
found = 0;
MFp4: Bring in updated jail support from bz_jail branch. This enhances the current jail implementation to permit multiple addresses per jail. In addtion to IPv4, IPv6 is supported as well. Due to updated checks it is even possible to have jails without an IP address at all, which basically gives one a chroot with restricted process view, no networking,.. SCTP support was updated and supports IPv6 in jails as well. Cpuset support permits jails to be bound to specific processor sets after creation. Jails can have an unrestricted (no duplicate protection, etc.) name in addition to the hostname. The jail name cannot be changed from within a jail and is considered to be used for management purposes or as audit-token in the future. DDB 'show jails' command was added to aid debugging. Proper compat support permits 32bit jail binaries to be used on 64bit systems to manage jails. Also backward compatibility was preserved where possible: for jail v1 syscalls, as well as with user space management utilities. Both jail as well as prison version were updated for the new features. A gap was intentionally left as the intermediate versions had been used by various patches floating around the last years. Bump __FreeBSD_version for the afore mentioned and in kernel changes. Special thanks to: - Pawel Jakub Dawidek (pjd) for his multi-IPv4 patches and Olivier Houchard (cognet) for initial single-IPv6 patches. - Jeff Roberson (jeff) and Randall Stewart (rrs) for their help, ideas and review on cpuset and SCTP support. - Robert Watson (rwatson) for lots and lots of help, discussions, suggestions and review of most of the patch at various stages. - John Baldwin (jhb) for his help. - Simon L. Nielsen (simon) as early adopter testing changes on cluster machines as well as all the testers and people who provided feedback the last months on freebsd-jail and other channels. - My employer, CK Software GmbH, for the support so I could work on this. Reviewed by: (see above) MFC after: 3 months (this is just so that I get the mail) X-MFC Before: 7.2-RELEASE if possible
2008-11-29 14:32:14 +00:00
/*
* Try to find an address on the given outgoing interface
* that belongs to the jail.
MFp4: Bring in updated jail support from bz_jail branch. This enhances the current jail implementation to permit multiple addresses per jail. In addtion to IPv4, IPv6 is supported as well. Due to updated checks it is even possible to have jails without an IP address at all, which basically gives one a chroot with restricted process view, no networking,.. SCTP support was updated and supports IPv6 in jails as well. Cpuset support permits jails to be bound to specific processor sets after creation. Jails can have an unrestricted (no duplicate protection, etc.) name in addition to the hostname. The jail name cannot be changed from within a jail and is considered to be used for management purposes or as audit-token in the future. DDB 'show jails' command was added to aid debugging. Proper compat support permits 32bit jail binaries to be used on 64bit systems to manage jails. Also backward compatibility was preserved where possible: for jail v1 syscalls, as well as with user space management utilities. Both jail as well as prison version were updated for the new features. A gap was intentionally left as the intermediate versions had been used by various patches floating around the last years. Bump __FreeBSD_version for the afore mentioned and in kernel changes. Special thanks to: - Pawel Jakub Dawidek (pjd) for his multi-IPv4 patches and Olivier Houchard (cognet) for initial single-IPv6 patches. - Jeff Roberson (jeff) and Randall Stewart (rrs) for their help, ideas and review on cpuset and SCTP support. - Robert Watson (rwatson) for lots and lots of help, discussions, suggestions and review of most of the patch at various stages. - John Baldwin (jhb) for his help. - Simon L. Nielsen (simon) as early adopter testing changes on cluster machines as well as all the testers and people who provided feedback the last months on freebsd-jail and other channels. - My employer, CK Software GmbH, for the support so I could work on this. Reviewed by: (see above) MFC after: 3 months (this is just so that I get the mail) X-MFC Before: 7.2-RELEASE if possible
2008-11-29 14:32:14 +00:00
*/
NET_EPOCH_ENTER(et);
ifnet: Replace if_addr_lock rwlock with epoch + mutex Run on LLNW canaries and tested by pho@ gallatin: Using a 14-core, 28-HTT single socket E5-2697 v3 with a 40GbE MLX5 based ConnectX 4-LX NIC, I see an almost 12% improvement in received packet rate, and a larger improvement in bytes delivered all the way to userspace. When the host receiving 64 streams of netperf -H $DUT -t UDP_STREAM -- -m 1, I see, using nstat -I mce0 1 before the patch: InMpps OMpps InGbs OGbs err TCP Est %CPU syscalls csw irq GBfree 4.98 0.00 4.42 0.00 4235592 33 83.80 4720653 2149771 1235 247.32 4.73 0.00 4.20 0.00 4025260 33 82.99 4724900 2139833 1204 247.32 4.72 0.00 4.20 0.00 4035252 33 82.14 4719162 2132023 1264 247.32 4.71 0.00 4.21 0.00 4073206 33 83.68 4744973 2123317 1347 247.32 4.72 0.00 4.21 0.00 4061118 33 80.82 4713615 2188091 1490 247.32 4.72 0.00 4.21 0.00 4051675 33 85.29 4727399 2109011 1205 247.32 4.73 0.00 4.21 0.00 4039056 33 84.65 4724735 2102603 1053 247.32 After the patch InMpps OMpps InGbs OGbs err TCP Est %CPU syscalls csw irq GBfree 5.43 0.00 4.20 0.00 3313143 33 84.96 5434214 1900162 2656 245.51 5.43 0.00 4.20 0.00 3308527 33 85.24 5439695 1809382 2521 245.51 5.42 0.00 4.19 0.00 3316778 33 87.54 5416028 1805835 2256 245.51 5.42 0.00 4.19 0.00 3317673 33 90.44 5426044 1763056 2332 245.51 5.42 0.00 4.19 0.00 3314839 33 88.11 5435732 1792218 2499 245.52 5.44 0.00 4.19 0.00 3293228 33 91.84 5426301 1668597 2121 245.52 Similarly, netperf reports 230Mb/s before the patch, and 270Mb/s after the patch Reviewed by: gallatin Sponsored by: Limelight Networks Differential Revision: https://reviews.freebsd.org/D15366
2018-05-18 20:13:34 +00:00
CK_STAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) {
struct sockaddr *sa;
sa = ifa->ifa_addr;
if (sa->sa_family != AF_INET6)
continue;
bcopy(&((struct sockaddr_in6 *)sa)->sin6_addr,
&ia6, sizeof(struct in6_addr));
if (prison_check_ip6(cred, &ia6) == 0) {
found = 1;
break;
}
}
NET_EPOCH_EXIT(et);
if (!found) {
MFp4: Bring in updated jail support from bz_jail branch. This enhances the current jail implementation to permit multiple addresses per jail. In addtion to IPv4, IPv6 is supported as well. Due to updated checks it is even possible to have jails without an IP address at all, which basically gives one a chroot with restricted process view, no networking,.. SCTP support was updated and supports IPv6 in jails as well. Cpuset support permits jails to be bound to specific processor sets after creation. Jails can have an unrestricted (no duplicate protection, etc.) name in addition to the hostname. The jail name cannot be changed from within a jail and is considered to be used for management purposes or as audit-token in the future. DDB 'show jails' command was added to aid debugging. Proper compat support permits 32bit jail binaries to be used on 64bit systems to manage jails. Also backward compatibility was preserved where possible: for jail v1 syscalls, as well as with user space management utilities. Both jail as well as prison version were updated for the new features. A gap was intentionally left as the intermediate versions had been used by various patches floating around the last years. Bump __FreeBSD_version for the afore mentioned and in kernel changes. Special thanks to: - Pawel Jakub Dawidek (pjd) for his multi-IPv4 patches and Olivier Houchard (cognet) for initial single-IPv6 patches. - Jeff Roberson (jeff) and Randall Stewart (rrs) for their help, ideas and review on cpuset and SCTP support. - Robert Watson (rwatson) for lots and lots of help, discussions, suggestions and review of most of the patch at various stages. - John Baldwin (jhb) for his help. - Simon L. Nielsen (simon) as early adopter testing changes on cluster machines as well as all the testers and people who provided feedback the last months on freebsd-jail and other channels. - My employer, CK Software GmbH, for the support so I could work on this. Reviewed by: (see above) MFC after: 3 months (this is just so that I get the mail) X-MFC Before: 7.2-RELEASE if possible
2008-11-29 14:32:14 +00:00
/*
* As a last resort return the 'default' jail address.
MFp4: Bring in updated jail support from bz_jail branch. This enhances the current jail implementation to permit multiple addresses per jail. In addtion to IPv4, IPv6 is supported as well. Due to updated checks it is even possible to have jails without an IP address at all, which basically gives one a chroot with restricted process view, no networking,.. SCTP support was updated and supports IPv6 in jails as well. Cpuset support permits jails to be bound to specific processor sets after creation. Jails can have an unrestricted (no duplicate protection, etc.) name in addition to the hostname. The jail name cannot be changed from within a jail and is considered to be used for management purposes or as audit-token in the future. DDB 'show jails' command was added to aid debugging. Proper compat support permits 32bit jail binaries to be used on 64bit systems to manage jails. Also backward compatibility was preserved where possible: for jail v1 syscalls, as well as with user space management utilities. Both jail as well as prison version were updated for the new features. A gap was intentionally left as the intermediate versions had been used by various patches floating around the last years. Bump __FreeBSD_version for the afore mentioned and in kernel changes. Special thanks to: - Pawel Jakub Dawidek (pjd) for his multi-IPv4 patches and Olivier Houchard (cognet) for initial single-IPv6 patches. - Jeff Roberson (jeff) and Randall Stewart (rrs) for their help, ideas and review on cpuset and SCTP support. - Robert Watson (rwatson) for lots and lots of help, discussions, suggestions and review of most of the patch at various stages. - John Baldwin (jhb) for his help. - Simon L. Nielsen (simon) as early adopter testing changes on cluster machines as well as all the testers and people who provided feedback the last months on freebsd-jail and other channels. - My employer, CK Software GmbH, for the support so I could work on this. Reviewed by: (see above) MFC after: 3 months (this is just so that I get the mail) X-MFC Before: 7.2-RELEASE if possible
2008-11-29 14:32:14 +00:00
*/
ia6 = ((struct sockaddr_in6 *)rt->rt_ifa->ifa_addr)->
sin6_addr;
if (prison_get_ip6(cred, &ia6) != 0)
MFp4: Bring in updated jail support from bz_jail branch. This enhances the current jail implementation to permit multiple addresses per jail. In addtion to IPv4, IPv6 is supported as well. Due to updated checks it is even possible to have jails without an IP address at all, which basically gives one a chroot with restricted process view, no networking,.. SCTP support was updated and supports IPv6 in jails as well. Cpuset support permits jails to be bound to specific processor sets after creation. Jails can have an unrestricted (no duplicate protection, etc.) name in addition to the hostname. The jail name cannot be changed from within a jail and is considered to be used for management purposes or as audit-token in the future. DDB 'show jails' command was added to aid debugging. Proper compat support permits 32bit jail binaries to be used on 64bit systems to manage jails. Also backward compatibility was preserved where possible: for jail v1 syscalls, as well as with user space management utilities. Both jail as well as prison version were updated for the new features. A gap was intentionally left as the intermediate versions had been used by various patches floating around the last years. Bump __FreeBSD_version for the afore mentioned and in kernel changes. Special thanks to: - Pawel Jakub Dawidek (pjd) for his multi-IPv4 patches and Olivier Houchard (cognet) for initial single-IPv6 patches. - Jeff Roberson (jeff) and Randall Stewart (rrs) for their help, ideas and review on cpuset and SCTP support. - Robert Watson (rwatson) for lots and lots of help, discussions, suggestions and review of most of the patch at various stages. - John Baldwin (jhb) for his help. - Simon L. Nielsen (simon) as early adopter testing changes on cluster machines as well as all the testers and people who provided feedback the last months on freebsd-jail and other channels. - My employer, CK Software GmbH, for the support so I could work on this. Reviewed by: (see above) MFC after: 3 months (this is just so that I get the mail) X-MFC Before: 7.2-RELEASE if possible
2008-11-29 14:32:14 +00:00
return (ESRCH);
}
bzero(&saun->sin6, sizeof(struct sockaddr_in6));
saun->sin6.sin6_len = sizeof(struct sockaddr_in6);
saun->sin6.sin6_family = AF_INET6;
bcopy(&ia6, &saun->sin6.sin6_addr, sizeof(struct in6_addr));
if (sa6_recoverscope(&saun->sin6) != 0)
return (ESRCH);
info->rti_info[RTAX_IFA] = (struct sockaddr *)&saun->sin6;
MFp4: Bring in updated jail support from bz_jail branch. This enhances the current jail implementation to permit multiple addresses per jail. In addtion to IPv4, IPv6 is supported as well. Due to updated checks it is even possible to have jails without an IP address at all, which basically gives one a chroot with restricted process view, no networking,.. SCTP support was updated and supports IPv6 in jails as well. Cpuset support permits jails to be bound to specific processor sets after creation. Jails can have an unrestricted (no duplicate protection, etc.) name in addition to the hostname. The jail name cannot be changed from within a jail and is considered to be used for management purposes or as audit-token in the future. DDB 'show jails' command was added to aid debugging. Proper compat support permits 32bit jail binaries to be used on 64bit systems to manage jails. Also backward compatibility was preserved where possible: for jail v1 syscalls, as well as with user space management utilities. Both jail as well as prison version were updated for the new features. A gap was intentionally left as the intermediate versions had been used by various patches floating around the last years. Bump __FreeBSD_version for the afore mentioned and in kernel changes. Special thanks to: - Pawel Jakub Dawidek (pjd) for his multi-IPv4 patches and Olivier Houchard (cognet) for initial single-IPv6 patches. - Jeff Roberson (jeff) and Randall Stewart (rrs) for their help, ideas and review on cpuset and SCTP support. - Robert Watson (rwatson) for lots and lots of help, discussions, suggestions and review of most of the patch at various stages. - John Baldwin (jhb) for his help. - Simon L. Nielsen (simon) as early adopter testing changes on cluster machines as well as all the testers and people who provided feedback the last months on freebsd-jail and other channels. - My employer, CK Software GmbH, for the support so I could work on this. Reviewed by: (see above) MFC after: 3 months (this is just so that I get the mail) X-MFC Before: 7.2-RELEASE if possible
2008-11-29 14:32:14 +00:00
break;
}
#endif
default:
return (ESRCH);
}
return (0);
}
1994-05-24 10:09:53 +00:00
/*ARGSUSED*/
1995-11-16 19:00:27 +00:00
static int
route_output(struct mbuf *m, struct socket *so, ...)
1994-05-24 10:09:53 +00:00
{
RIB_RLOCK_TRACKER;
struct rt_msghdr *rtm = NULL;
struct rtentry *rt = NULL;
struct rib_head *rnh;
1994-05-24 10:09:53 +00:00
struct rt_addrinfo info;
struct sockaddr_storage ss;
#ifdef INET6
struct sockaddr_in6 *sin6;
int i, rti_need_deembed = 0;
#endif
int alloc_len = 0, len, error = 0, fibnum;
struct ifnet *ifp = NULL;
MFp4: Bring in updated jail support from bz_jail branch. This enhances the current jail implementation to permit multiple addresses per jail. In addtion to IPv4, IPv6 is supported as well. Due to updated checks it is even possible to have jails without an IP address at all, which basically gives one a chroot with restricted process view, no networking,.. SCTP support was updated and supports IPv6 in jails as well. Cpuset support permits jails to be bound to specific processor sets after creation. Jails can have an unrestricted (no duplicate protection, etc.) name in addition to the hostname. The jail name cannot be changed from within a jail and is considered to be used for management purposes or as audit-token in the future. DDB 'show jails' command was added to aid debugging. Proper compat support permits 32bit jail binaries to be used on 64bit systems to manage jails. Also backward compatibility was preserved where possible: for jail v1 syscalls, as well as with user space management utilities. Both jail as well as prison version were updated for the new features. A gap was intentionally left as the intermediate versions had been used by various patches floating around the last years. Bump __FreeBSD_version for the afore mentioned and in kernel changes. Special thanks to: - Pawel Jakub Dawidek (pjd) for his multi-IPv4 patches and Olivier Houchard (cognet) for initial single-IPv6 patches. - Jeff Roberson (jeff) and Randall Stewart (rrs) for their help, ideas and review on cpuset and SCTP support. - Robert Watson (rwatson) for lots and lots of help, discussions, suggestions and review of most of the patch at various stages. - John Baldwin (jhb) for his help. - Simon L. Nielsen (simon) as early adopter testing changes on cluster machines as well as all the testers and people who provided feedback the last months on freebsd-jail and other channels. - My employer, CK Software GmbH, for the support so I could work on this. Reviewed by: (see above) MFC after: 3 months (this is just so that I get the mail) X-MFC Before: 7.2-RELEASE if possible
2008-11-29 14:32:14 +00:00
union sockaddr_union saun;
sa_family_t saf = AF_UNSPEC;
struct rawcb *rp = NULL;
struct walkarg w;
1994-05-24 10:09:53 +00:00
fibnum = so->so_fibnum;
1994-05-24 10:09:53 +00:00
#define senderr(e) { error = e; goto flush;}
if (m == NULL || ((m->m_len < sizeof(long)) &&
(m = m_pullup(m, sizeof(long))) == NULL))
1994-05-24 10:09:53 +00:00
return (ENOBUFS);
if ((m->m_flags & M_PKTHDR) == 0)
panic("route_output");
len = m->m_pkthdr.len;
if (len < sizeof(*rtm) ||
len != mtod(m, struct rt_msghdr *)->rtm_msglen)
1994-05-24 10:09:53 +00:00
senderr(EINVAL);
/*
* Most of current messages are in range 200-240 bytes,
* minimize possible re-allocation on reply using larger size
* buffer aligned on 1k boundaty.
*/
alloc_len = roundup2(len, 1024);
if ((rtm = malloc(alloc_len, M_TEMP, M_NOWAIT)) == NULL)
senderr(ENOBUFS);
1994-05-24 10:09:53 +00:00
m_copydata(m, 0, len, (caddr_t)rtm);
bzero(&info, sizeof(info));
bzero(&w, sizeof(w));
1994-05-24 10:09:53 +00:00
if (rtm->rtm_version != RTM_VERSION) {
/* Do not touch message since format is unknown */
free(rtm, M_TEMP);
rtm = NULL;
1994-05-24 10:09:53 +00:00
senderr(EPROTONOSUPPORT);
}
/*
* Starting from here, it is possible
* to alter original message and insert
* caller PID and error value.
*/
1994-05-24 10:09:53 +00:00
rtm->rtm_pid = curproc->p_pid;
info.rti_addrs = rtm->rtm_addrs;
info.rti_mflags = rtm->rtm_inits;
info.rti_rmx = &rtm->rtm_rmx;
/*
* rt_xaddrs() performs s6_addr[2] := sin6_scope_id for AF_INET6
* link-local address because rtrequest requires addresses with
* embedded scope id.
*/
if (rt_xaddrs((caddr_t)(rtm + 1), len + (caddr_t)rtm, &info))
senderr(EINVAL);
if (rtm->rtm_flags & RTF_RNH_LOCKED)
senderr(EINVAL);
info.rti_flags = rtm->rtm_flags;
if (info.rti_info[RTAX_DST] == NULL ||
info.rti_info[RTAX_DST]->sa_family >= AF_MAX ||
(info.rti_info[RTAX_GATEWAY] != NULL &&
info.rti_info[RTAX_GATEWAY]->sa_family >= AF_MAX))
1994-05-24 10:09:53 +00:00
senderr(EINVAL);
saf = info.rti_info[RTAX_DST]->sa_family;
/*
* Verify that the caller has the appropriate privilege; RTM_GET
* is the only operation the non-superuser is allowed.
*/
if (rtm->rtm_type != RTM_GET) {
error = priv_check(curthread, PRIV_NET_ROUTE);
if (error)
senderr(error);
}
/*
* The given gateway address may be an interface address.
* For example, issuing a "route change" command on a route
* entry that was created from a tunnel, and the gateway
* address given is the local end point. In this case the
* RTF_GATEWAY flag must be cleared or the destination will
* not be reachable even though there is no error message.
*/
if (info.rti_info[RTAX_GATEWAY] != NULL &&
info.rti_info[RTAX_GATEWAY]->sa_family != AF_LINK) {
struct rt_addrinfo ginfo;
struct sockaddr *gdst;
bzero(&ginfo, sizeof(ginfo));
bzero(&ss, sizeof(ss));
ss.ss_len = sizeof(ss);
ginfo.rti_info[RTAX_GATEWAY] = (struct sockaddr *)&ss;
gdst = info.rti_info[RTAX_GATEWAY];
/*
* A host route through the loopback interface is
* installed for each interface adddress. In pre 8.0
* releases the interface address of a PPP link type
* is not reachable locally. This behavior is fixed as
* part of the new L2/L3 redesign and rewrite work. The
* signature of this interface address route is the
* AF_LINK sa_family type of the rt_gateway, and the
* rt_ifp has the IFF_LOOPBACK flag set.
*/
if (rib_lookup_info(fibnum, gdst, NHR_REF, 0, &ginfo) == 0) {
if (ss.ss_family == AF_LINK &&
ginfo.rti_ifp->if_flags & IFF_LOOPBACK) {
info.rti_flags &= ~RTF_GATEWAY;
info.rti_flags |= RTF_GWFLAG_COMPAT;
}
rib_free_info(&ginfo);
}
}
1994-05-24 10:09:53 +00:00
switch (rtm->rtm_type) {
struct rtentry *saved_nrt;
1994-05-24 10:09:53 +00:00
case RTM_ADD:
case RTM_CHANGE:
if (rtm->rtm_type == RTM_ADD) {
if (info.rti_info[RTAX_GATEWAY] == NULL)
senderr(EINVAL);
}
saved_nrt = NULL;
2008-12-26 19:45:24 +00:00
/* support for new ARP code */
if (info.rti_info[RTAX_GATEWAY] != NULL &&
info.rti_info[RTAX_GATEWAY]->sa_family == AF_LINK &&
2008-12-26 19:45:24 +00:00
(rtm->rtm_flags & RTF_LLDATA) != 0) {
error = lla_rt_output(rtm, &info);
#ifdef INET6
if (error == 0)
rti_need_deembed = (V_deembed_scopeid) ? 1 : 0;
#endif
break;
}
error = rtrequest1_fib(rtm->rtm_type, &info, &saved_nrt,
fibnum);
if (error == 0 && saved_nrt != NULL) {
#ifdef INET6
rti_need_deembed = (V_deembed_scopeid) ? 1 : 0;
#endif
RT_LOCK(saved_nrt);
rtm->rtm_index = saved_nrt->rt_ifp->if_index;
RT_REMREF(saved_nrt);
RT_UNLOCK(saved_nrt);
1994-05-24 10:09:53 +00:00
}
break;
case RTM_DELETE:
saved_nrt = NULL;
/* support for new ARP code */
if (info.rti_info[RTAX_GATEWAY] &&
2008-12-26 19:45:24 +00:00
(info.rti_info[RTAX_GATEWAY]->sa_family == AF_LINK) &&
(rtm->rtm_flags & RTF_LLDATA) != 0) {
error = lla_rt_output(rtm, &info);
#ifdef INET6
if (error == 0)
rti_need_deembed = (V_deembed_scopeid) ? 1 : 0;
#endif
break;
}
error = rtrequest1_fib(RTM_DELETE, &info, &saved_nrt, fibnum);
if (error == 0) {
RT_LOCK(saved_nrt);
rt = saved_nrt;
goto report;
}
#ifdef INET6
/* rt_msg2() will not be used when RTM_DELETE fails. */
rti_need_deembed = (V_deembed_scopeid) ? 1 : 0;
#endif
1994-05-24 10:09:53 +00:00
break;
case RTM_GET:
rnh = rt_tables_get_rnh(fibnum, saf);
if (rnh == NULL)
senderr(EAFNOSUPPORT);
RIB_RLOCK(rnh);
if (info.rti_info[RTAX_NETMASK] == NULL &&
rtm->rtm_type == RTM_GET) {
/*
* Provide longest prefix match for
* address lookup (no mask).
* 'route -n get addr'
*/
rt = (struct rtentry *) rnh->rnh_matchaddr(
info.rti_info[RTAX_DST], &rnh->head);
} else
rt = (struct rtentry *) rnh->rnh_lookup(
info.rti_info[RTAX_DST],
info.rti_info[RTAX_NETMASK], &rnh->head);
if (rt == NULL) {
RIB_RUNLOCK(rnh);
1994-05-24 10:09:53 +00:00
senderr(ESRCH);
}
#ifdef RADIX_MPATH
/*
* for RTM_CHANGE/LOCK, if we got multipath routes,
* we require users to specify a matching RTAX_GATEWAY.
*
* for RTM_GET, gate is optional even with multipath.
* if gate == NULL the first match is returned.
* (no need to call rt_mpath_matchgate if gate == NULL)
*/
if (rt_mpath_capable(rnh) &&
(rtm->rtm_type != RTM_GET || info.rti_info[RTAX_GATEWAY])) {
rt = rt_mpath_matchgate(rt, info.rti_info[RTAX_GATEWAY]);
if (!rt) {
RIB_RUNLOCK(rnh);
senderr(ESRCH);
}
}
#endif
/*
* If performing proxied L2 entry insertion, and
* the actual PPP host entry is found, perform
* another search to retrieve the prefix route of
* the local end point of the PPP link.
*/
if (rtm->rtm_flags & RTF_ANNOUNCE) {
struct sockaddr laddr;
if (rt->rt_ifp != NULL &&
rt->rt_ifp->if_type == IFT_PROPVIRTUAL) {
struct epoch_tracker et;
struct ifaddr *ifa;
NET_EPOCH_ENTER(et);
ifa = ifa_ifwithnet(info.rti_info[RTAX_DST], 1,
RT_ALL_FIBS);
if (ifa != NULL)
rt_maskedcopy(ifa->ifa_addr,
&laddr,
ifa->ifa_netmask);
NET_EPOCH_EXIT(et);
} else
rt_maskedcopy(rt->rt_ifa->ifa_addr,
&laddr,
rt->rt_ifa->ifa_netmask);
/*
* refactor rt and no lock operation necessary
*/
rt = (struct rtentry *)rnh->rnh_matchaddr(&laddr,
&rnh->head);
if (rt == NULL) {
RIB_RUNLOCK(rnh);
senderr(ESRCH);
}
}
RT_LOCK(rt);
RT_ADDREF(rt);
RIB_RUNLOCK(rnh);
2002-12-24 03:03:39 +00:00
report:
RT_LOCK_ASSERT(rt);
if ((rt->rt_flags & RTF_HOST) == 0
? jailed_without_vnet(curthread->td_ucred)
: prison_if(curthread->td_ucred,
rt_key(rt)) != 0) {
RT_UNLOCK(rt);
senderr(ESRCH);
}
info.rti_info[RTAX_DST] = rt_key(rt);
info.rti_info[RTAX_GATEWAY] = rt->rt_gateway;
info.rti_info[RTAX_NETMASK] = rtsock_fix_netmask(rt_key(rt),
rt_mask(rt), &ss);
info.rti_info[RTAX_GENMASK] = 0;
if (rtm->rtm_addrs & (RTA_IFP | RTA_IFA)) {
ifp = rt->rt_ifp;
if (ifp) {
info.rti_info[RTAX_IFP] =
ifp->if_addr->ifa_addr;
error = rtm_get_jailed(&info, ifp, rt,
&saun, curthread->td_ucred);
if (error != 0) {
RT_UNLOCK(rt);
senderr(error);
}
if (ifp->if_flags & IFF_POINTOPOINT)
info.rti_info[RTAX_BRD] =
rt->rt_ifa->ifa_dstaddr;
rtm->rtm_index = ifp->if_index;
} else {
info.rti_info[RTAX_IFP] = NULL;
info.rti_info[RTAX_IFA] = NULL;
}
} else if ((ifp = rt->rt_ifp) != NULL) {
rtm->rtm_index = ifp->if_index;
}
/* Check if we need to realloc storage */
rtsock_msg_buffer(rtm->rtm_type, &info, NULL, &len);
if (len > alloc_len) {
struct rt_msghdr *new_rtm;
new_rtm = malloc(len, M_TEMP, M_NOWAIT);
if (new_rtm == NULL) {
RT_UNLOCK(rt);
senderr(ENOBUFS);
1994-05-24 10:09:53 +00:00
}
bcopy(rtm, new_rtm, rtm->rtm_msglen);
free(rtm, M_TEMP);
rtm = new_rtm;
alloc_len = len;
1994-05-24 10:09:53 +00:00
}
w.w_tmem = (caddr_t)rtm;
w.w_tmemsize = alloc_len;
rtsock_msg_buffer(rtm->rtm_type, &info, &w, &len);
if (rt->rt_flags & RTF_GWFLAG_COMPAT)
rtm->rtm_flags = RTF_GATEWAY |
(rt->rt_flags & ~RTF_GWFLAG_COMPAT);
else
rtm->rtm_flags = rt->rt_flags;
rt_getmetrics(rt, &rtm->rtm_rmx);
rtm->rtm_addrs = info.rti_addrs;
RT_UNLOCK(rt);
1994-05-24 10:09:53 +00:00
break;
default:
senderr(EOPNOTSUPP);
}
flush:
if (rt != NULL)
RTFREE(rt);
1994-05-24 10:09:53 +00:00
/*
* Check to see if we don't want our own messages.
*/
if ((so->so_options & SO_USELOOPBACK) == 0) {
if (V_route_cb.any_count <= 1) {
if (rtm != NULL)
free(rtm, M_TEMP);
1994-05-24 10:09:53 +00:00
m_freem(m);
return (error);
}
/* There is another listener, so construct message */
rp = sotorawcb(so);
}
if (rtm != NULL) {
#ifdef INET6
if (rti_need_deembed) {
/* sin6_scope_id is recovered before sending rtm. */
sin6 = (struct sockaddr_in6 *)&ss;
for (i = 0; i < RTAX_MAX; i++) {
if (info.rti_info[i] == NULL)
continue;
if (info.rti_info[i]->sa_family != AF_INET6)
continue;
bcopy(info.rti_info[i], sin6, sizeof(*sin6));
if (sa6_recoverscope(sin6) == 0)
bcopy(sin6, info.rti_info[i],
sizeof(*sin6));
}
}
#endif
if (error != 0)
rtm->rtm_errno = error;
else
rtm->rtm_flags |= RTF_DONE;
1994-05-24 10:09:53 +00:00
m_copyback(m, 0, rtm->rtm_msglen, (caddr_t)rtm);
if (m->m_pkthdr.len < rtm->rtm_msglen) {
m_freem(m);
m = NULL;
} else if (m->m_pkthdr.len > rtm->rtm_msglen)
m_adj(m, rtm->rtm_msglen - m->m_pkthdr.len);
free(rtm, M_TEMP);
1994-05-24 10:09:53 +00:00
}
if (m != NULL) {
M_SETFIB(m, fibnum);
m->m_flags |= RTS_FILTER_FIB;
if (rp) {
/*
* XXX insure we don't get a copy by
* invalidating our protocol
*/
unsigned short family = rp->rcb_proto.sp_family;
rp->rcb_proto.sp_family = 0;
rt_dispatch(m, saf);
rp->rcb_proto.sp_family = family;
} else
rt_dispatch(m, saf);
}
1994-05-24 10:09:53 +00:00
return (error);
}
static void
rt_getmetrics(const struct rtentry *rt, struct rt_metrics *out)
{
bzero(out, sizeof(*out));
out->rmx_mtu = rt->rt_mtu;
out->rmx_weight = rt->rt_weight;
out->rmx_pksent = counter_u64_fetch(rt->rt_pksent);
/* Kernel -> userland timebase conversion. */
out->rmx_expire = rt->rt_expire ?
rt->rt_expire - time_uptime + time_second : 0;
}
/*
* Extract the addresses of the passed sockaddrs.
* Do a little sanity checking so as to avoid bad memory references.
* This data is derived straight from userland.
*/
static int
rt_xaddrs(caddr_t cp, caddr_t cplim, struct rt_addrinfo *rtinfo)
1994-05-24 10:09:53 +00:00
{
struct sockaddr *sa;
int i;
1994-05-24 10:09:53 +00:00
for (i = 0; i < RTAX_MAX && cp < cplim; i++) {
1994-05-24 10:09:53 +00:00
if ((rtinfo->rti_addrs & (1 << i)) == 0)
continue;
sa = (struct sockaddr *)cp;
/*
* It won't fit.
*/
if (cp + sa->sa_len > cplim)
return (EINVAL);
/*
* there are no more.. quit now
* If there are more bits, they are in error.
* I've seen this. route(1) can evidently generate these.
* This causes kernel to core dump.
* for compatibility, If we see this, point to a safe address.
*/
if (sa->sa_len == 0) {
rtinfo->rti_info[i] = &sa_zero;
return (0); /* should be EINVAL but for compat */
}
/* accept it */
#ifdef INET6
if (sa->sa_family == AF_INET6)
sa6_embedscope((struct sockaddr_in6 *)sa,
V_ip6_use_defzone);
#endif
rtinfo->rti_info[i] = sa;
cp += SA_SIZE(sa);
1994-05-24 10:09:53 +00:00
}
return (0);
1994-05-24 10:09:53 +00:00
}
/*
* Fill in @dmask with valid netmask leaving original @smask
* intact. Mostly used with radix netmasks.
*/
static struct sockaddr *
rtsock_fix_netmask(struct sockaddr *dst, struct sockaddr *smask,
struct sockaddr_storage *dmask)
{
if (dst == NULL || smask == NULL)
return (NULL);
memset(dmask, 0, dst->sa_len);
memcpy(dmask, smask, smask->sa_len);
dmask->ss_len = dst->sa_len;
dmask->ss_family = dst->sa_family;
return ((struct sockaddr *)dmask);
}
/*
* Writes information related to @rtinfo object to newly-allocated mbuf.
* Assumes MCLBYTES is enough to construct any message.
* Used for OS notifications of vaious events (if/ifa announces,etc)
*
* Returns allocated mbuf or NULL on failure.
*/
1994-05-24 10:09:53 +00:00
static struct mbuf *
rtsock_msg_mbuf(int type, struct rt_addrinfo *rtinfo)
1994-05-24 10:09:53 +00:00
{
struct rt_msghdr *rtm;
struct mbuf *m;
int i;
struct sockaddr *sa;
#ifdef INET6
struct sockaddr_storage ss;
struct sockaddr_in6 *sin6;
#endif
1994-05-24 10:09:53 +00:00
int len, dlen;
switch (type) {
case RTM_DELADDR:
case RTM_NEWADDR:
len = sizeof(struct ifa_msghdr);
break;
case RTM_DELMADDR:
case RTM_NEWMADDR:
len = sizeof(struct ifma_msghdr);
break;
1994-05-24 10:09:53 +00:00
case RTM_IFINFO:
len = sizeof(struct if_msghdr);
break;
case RTM_IFANNOUNCE:
case RTM_IEEE80211:
len = sizeof(struct if_announcemsghdr);
break;
1994-05-24 10:09:53 +00:00
default:
len = sizeof(struct rt_msghdr);
}
/* XXXGL: can we use MJUMPAGESIZE cluster here? */
KASSERT(len <= MCLBYTES, ("%s: message too big", __func__));
if (len > MHLEN)
m = m_getcl(M_NOWAIT, MT_DATA, M_PKTHDR);
else
m = m_gethdr(M_NOWAIT, MT_DATA);
if (m == NULL)
return (m);
1994-05-24 10:09:53 +00:00
m->m_pkthdr.len = m->m_len = len;
rtm = mtod(m, struct rt_msghdr *);
bzero((caddr_t)rtm, len);
for (i = 0; i < RTAX_MAX; i++) {
if ((sa = rtinfo->rti_info[i]) == NULL)
continue;
rtinfo->rti_addrs |= (1 << i);
dlen = SA_SIZE(sa);
#ifdef INET6
if (V_deembed_scopeid && sa->sa_family == AF_INET6) {
sin6 = (struct sockaddr_in6 *)&ss;
bcopy(sa, sin6, sizeof(*sin6));
if (sa6_recoverscope(sin6) == 0)
sa = (struct sockaddr *)sin6;
}
#endif
1994-05-24 10:09:53 +00:00
m_copyback(m, len, dlen, (caddr_t)sa);
len += dlen;
}
if (m->m_pkthdr.len != len) {
m_freem(m);
return (NULL);
}
rtm->rtm_msglen = len;
rtm->rtm_version = RTM_VERSION;
rtm->rtm_type = type;
return (m);
}
/*
* Writes information related to @rtinfo object to preallocated buffer.
* Stores needed size in @plen. If @w is NULL, calculates size without
* writing.
* Used for sysctl dumps and rtsock answers (RTM_DEL/RTM_GET) generation.
*
* Returns 0 on success.
*
*/
1994-05-24 10:09:53 +00:00
static int
rtsock_msg_buffer(int type, struct rt_addrinfo *rtinfo, struct walkarg *w, int *plen)
1994-05-24 10:09:53 +00:00
{
int i;
int len, buflen = 0, dlen;
caddr_t cp = NULL;
struct rt_msghdr *rtm = NULL;
#ifdef INET6
struct sockaddr_storage ss;
struct sockaddr_in6 *sin6;
#endif
#ifdef COMPAT_FREEBSD32
bool compat32 = false;
#endif
1994-05-24 10:09:53 +00:00
switch (type) {
case RTM_DELADDR:
case RTM_NEWADDR:
if (w != NULL && w->w_op == NET_RT_IFLISTL) {
#ifdef COMPAT_FREEBSD32
if (w->w_req->flags & SCTL_MASK32) {
len = sizeof(struct ifa_msghdrl32);
compat32 = true;
} else
#endif
len = sizeof(struct ifa_msghdrl);
} else
len = sizeof(struct ifa_msghdr);
1994-05-24 10:09:53 +00:00
break;
case RTM_IFINFO:
#ifdef COMPAT_FREEBSD32
if (w != NULL && w->w_req->flags & SCTL_MASK32) {
if (w->w_op == NET_RT_IFLISTL)
len = sizeof(struct if_msghdrl32);
else
len = sizeof(struct if_msghdr32);
compat32 = true;
break;
}
#endif
if (w != NULL && w->w_op == NET_RT_IFLISTL)
len = sizeof(struct if_msghdrl);
else
len = sizeof(struct if_msghdr);
1994-05-24 10:09:53 +00:00
break;
case RTM_NEWMADDR:
len = sizeof(struct ifma_msghdr);
break;
1994-05-24 10:09:53 +00:00
default:
len = sizeof(struct rt_msghdr);
}
if (w != NULL) {
rtm = (struct rt_msghdr *)w->w_tmem;
buflen = w->w_tmemsize - len;
cp = (caddr_t)w->w_tmem + len;
}
rtinfo->rti_addrs = 0;
1994-05-24 10:09:53 +00:00
for (i = 0; i < RTAX_MAX; i++) {
struct sockaddr *sa;
1994-05-24 10:09:53 +00:00
if ((sa = rtinfo->rti_info[i]) == NULL)
1994-05-24 10:09:53 +00:00
continue;
rtinfo->rti_addrs |= (1 << i);
#ifdef COMPAT_FREEBSD32
if (compat32)
dlen = SA_SIZE32(sa);
else
#endif
dlen = SA_SIZE(sa);
if (cp != NULL && buflen >= dlen) {
#ifdef INET6
if (V_deembed_scopeid && sa->sa_family == AF_INET6) {
sin6 = (struct sockaddr_in6 *)&ss;
bcopy(sa, sin6, sizeof(*sin6));
if (sa6_recoverscope(sin6) == 0)
sa = (struct sockaddr *)sin6;
}
#endif
1994-05-24 10:09:53 +00:00
bcopy((caddr_t)sa, cp, (unsigned)dlen);
cp += dlen;
buflen -= dlen;
} else if (cp != NULL) {
/*
* Buffer too small. Count needed size
* and return with error.
*/
cp = NULL;
1994-05-24 10:09:53 +00:00
}
1994-05-24 10:09:53 +00:00
len += dlen;
}
if (cp != NULL) {
dlen = ALIGN(len) - len;
if (buflen < dlen)
cp = NULL;
else {
bzero(cp, dlen);
cp += dlen;
buflen -= dlen;
}
1994-05-24 10:09:53 +00:00
}
len = ALIGN(len);
1994-05-24 10:09:53 +00:00
if (cp != NULL) {
/* fill header iff buffer is large enough */
1994-05-24 10:09:53 +00:00
rtm->rtm_version = RTM_VERSION;
rtm->rtm_type = type;
rtm->rtm_msglen = len;
}
*plen = len;
if (w != NULL && cp == NULL)
return (ENOBUFS);
return (0);
1994-05-24 10:09:53 +00:00
}
/*
* This routine is called to generate a message from the routing
* socket indicating that a redirect has occurred, a routing lookup
1994-05-24 10:09:53 +00:00
* has failed, or that a protocol has detected timeouts to a particular
* destination.
*/
void
rt_missmsg_fib(int type, struct rt_addrinfo *rtinfo, int flags, int error,
int fibnum)
1994-05-24 10:09:53 +00:00
{
struct rt_msghdr *rtm;
struct mbuf *m;
1994-05-24 10:09:53 +00:00
struct sockaddr *sa = rtinfo->rti_info[RTAX_DST];
if (V_route_cb.any_count == 0)
1994-05-24 10:09:53 +00:00
return;
m = rtsock_msg_mbuf(type, rtinfo);
if (m == NULL)
1994-05-24 10:09:53 +00:00
return;
if (fibnum != RT_ALL_FIBS) {
KASSERT(fibnum >= 0 && fibnum < rt_numfibs, ("%s: fibnum out "
"of range 0 <= %d < %d", __func__, fibnum, rt_numfibs));
M_SETFIB(m, fibnum);
m->m_flags |= RTS_FILTER_FIB;
}
1994-05-24 10:09:53 +00:00
rtm = mtod(m, struct rt_msghdr *);
rtm->rtm_flags = RTF_DONE | flags;
rtm->rtm_errno = error;
rtm->rtm_addrs = rtinfo->rti_addrs;
rt_dispatch(m, sa ? sa->sa_family : AF_UNSPEC);
1994-05-24 10:09:53 +00:00
}
void
rt_missmsg(int type, struct rt_addrinfo *rtinfo, int flags, int error)
{
rt_missmsg_fib(type, rtinfo, flags, error, RT_ALL_FIBS);
}
1994-05-24 10:09:53 +00:00
/*
* This routine is called to generate a message from the routing
* socket indicating that the status of a network interface has changed.
*/
void
rt_ifmsg(struct ifnet *ifp)
1994-05-24 10:09:53 +00:00
{
struct if_msghdr *ifm;
1994-05-24 10:09:53 +00:00
struct mbuf *m;
struct rt_addrinfo info;
if (V_route_cb.any_count == 0)
1994-05-24 10:09:53 +00:00
return;
bzero((caddr_t)&info, sizeof(info));
m = rtsock_msg_mbuf(RTM_IFINFO, &info);
if (m == NULL)
1994-05-24 10:09:53 +00:00
return;
ifm = mtod(m, struct if_msghdr *);
ifm->ifm_index = ifp->if_index;
ifm->ifm_flags = ifp->if_flags | ifp->if_drv_flags;
if_data_copy(ifp, &ifm->ifm_data);
1994-05-24 10:09:53 +00:00
ifm->ifm_addrs = 0;
rt_dispatch(m, AF_UNSPEC);
1994-05-24 10:09:53 +00:00
}
/*
* Announce interface address arrival/withdraw.
* Please do not call directly, use rt_addrmsg().
* Assume input data to be valid.
* Returns 0 on success.
1994-05-24 10:09:53 +00:00
*/
int
rtsock_addrmsg(int cmd, struct ifaddr *ifa, int fibnum)
1994-05-24 10:09:53 +00:00
{
struct rt_addrinfo info;
struct sockaddr *sa;
int ncmd;
struct mbuf *m;
struct ifa_msghdr *ifam;
1994-05-24 10:09:53 +00:00
struct ifnet *ifp = ifa->ifa_ifp;
struct sockaddr_storage ss;
1994-05-24 10:09:53 +00:00
if (V_route_cb.any_count == 0)
return (0);
1995-05-30 08:16:23 +00:00
ncmd = cmd == RTM_ADD ? RTM_NEWADDR : RTM_DELADDR;
bzero((caddr_t)&info, sizeof(info));
info.rti_info[RTAX_IFA] = sa = ifa->ifa_addr;
info.rti_info[RTAX_IFP] = ifp->if_addr->ifa_addr;
info.rti_info[RTAX_NETMASK] = rtsock_fix_netmask(
info.rti_info[RTAX_IFP], ifa->ifa_netmask, &ss);
info.rti_info[RTAX_BRD] = ifa->ifa_dstaddr;
if ((m = rtsock_msg_mbuf(ncmd, &info)) == NULL)
return (ENOBUFS);
ifam = mtod(m, struct ifa_msghdr *);
ifam->ifam_index = ifp->if_index;
ifam->ifam_metric = ifa->ifa_ifp->if_metric;
ifam->ifam_flags = ifa->ifa_flags;
ifam->ifam_addrs = info.rti_addrs;
if (fibnum != RT_ALL_FIBS) {
M_SETFIB(m, fibnum);
m->m_flags |= RTS_FILTER_FIB;
1994-05-24 10:09:53 +00:00
}
rt_dispatch(m, sa ? sa->sa_family : AF_UNSPEC);
return (0);
1994-05-24 10:09:53 +00:00
}
/*
* Announce route addition/removal.
* Please do not call directly, use rt_routemsg().
* Note that @rt data MAY be inconsistent/invalid:
* if some userland app sends us "invalid" route message (invalid mask,
* no dst, wrong address families, etc...) we need to pass it back
* to app (and any other rtsock consumers) with rtm_errno field set to
* non-zero value.
*
* Returns 0 on success.
*/
int
rtsock_routemsg(int cmd, struct ifnet *ifp, int error, struct rtentry *rt,
int fibnum)
{
struct rt_addrinfo info;
struct sockaddr *sa;
struct mbuf *m;
struct rt_msghdr *rtm;
struct sockaddr_storage ss;
if (V_route_cb.any_count == 0)
return (0);
bzero((caddr_t)&info, sizeof(info));
info.rti_info[RTAX_DST] = sa = rt_key(rt);
info.rti_info[RTAX_NETMASK] = rtsock_fix_netmask(sa, rt_mask(rt), &ss);
info.rti_info[RTAX_GATEWAY] = rt->rt_gateway;
if ((m = rtsock_msg_mbuf(cmd, &info)) == NULL)
return (ENOBUFS);
rtm = mtod(m, struct rt_msghdr *);
rtm->rtm_index = ifp->if_index;
rtm->rtm_flags |= rt->rt_flags;
rtm->rtm_errno = error;
rtm->rtm_addrs = info.rti_addrs;
if (fibnum != RT_ALL_FIBS) {
M_SETFIB(m, fibnum);
m->m_flags |= RTS_FILTER_FIB;
}
rt_dispatch(m, sa ? sa->sa_family : AF_UNSPEC);
return (0);
}
/*
* This is the analogue to the rt_newaddrmsg which performs the same
* function but for multicast group memberhips. This is easier since
* there is no route state to worry about.
*/
void
rt_newmaddrmsg(int cmd, struct ifmultiaddr *ifma)
{
struct rt_addrinfo info;
struct mbuf *m = NULL;
struct ifnet *ifp = ifma->ifma_ifp;
struct ifma_msghdr *ifmam;
if (V_route_cb.any_count == 0)
return;
bzero((caddr_t)&info, sizeof(info));
info.rti_info[RTAX_IFA] = ifma->ifma_addr;
if (ifp && ifp->if_addr)
info.rti_info[RTAX_IFP] = ifp->if_addr->ifa_addr;
else
info.rti_info[RTAX_IFP] = NULL;
/*
* If a link-layer address is present, present it as a ``gateway''
* (similarly to how ARP entries, e.g., are presented).
*/
info.rti_info[RTAX_GATEWAY] = ifma->ifma_lladdr;
m = rtsock_msg_mbuf(cmd, &info);
if (m == NULL)
return;
ifmam = mtod(m, struct ifma_msghdr *);
KASSERT(ifp != NULL, ("%s: link-layer multicast address w/o ifp\n",
__func__));
ifmam->ifmam_index = ifp->if_index;
ifmam->ifmam_addrs = info.rti_addrs;
rt_dispatch(m, ifma->ifma_addr ? ifma->ifma_addr->sa_family : AF_UNSPEC);
}
1995-11-16 19:00:27 +00:00
static struct mbuf *
rt_makeifannouncemsg(struct ifnet *ifp, int type, int what,
struct rt_addrinfo *info)
{
struct if_announcemsghdr *ifan;
struct mbuf *m;
if (V_route_cb.any_count == 0)
return NULL;
bzero((caddr_t)info, sizeof(*info));
m = rtsock_msg_mbuf(type, info);
if (m != NULL) {
ifan = mtod(m, struct if_announcemsghdr *);
ifan->ifan_index = ifp->if_index;
strlcpy(ifan->ifan_name, ifp->if_xname,
sizeof(ifan->ifan_name));
ifan->ifan_what = what;
}
return m;
}
/*
* This is called to generate routing socket messages indicating
* IEEE80211 wireless events.
* XXX we piggyback on the RTM_IFANNOUNCE msg format in a clumsy way.
*/
void
rt_ieee80211msg(struct ifnet *ifp, int what, void *data, size_t data_len)
{
struct mbuf *m;
struct rt_addrinfo info;
m = rt_makeifannouncemsg(ifp, RTM_IEEE80211, what, &info);
if (m != NULL) {
/*
* Append the ieee80211 data. Try to stick it in the
* mbuf containing the ifannounce msg; otherwise allocate
* a new mbuf and append.
*
* NB: we assume m is a single mbuf.
*/
if (data_len > M_TRAILINGSPACE(m)) {
struct mbuf *n = m_get(M_NOWAIT, MT_DATA);
if (n == NULL) {
m_freem(m);
return;
}
bcopy(data, mtod(n, void *), data_len);
n->m_len = data_len;
m->m_next = n;
} else if (data_len > 0) {
bcopy(data, mtod(m, u_int8_t *) + m->m_len, data_len);
m->m_len += data_len;
}
if (m->m_flags & M_PKTHDR)
m->m_pkthdr.len += data_len;
mtod(m, struct if_announcemsghdr *)->ifan_msglen += data_len;
rt_dispatch(m, AF_UNSPEC);
}
}
/*
* This is called to generate routing socket messages indicating
* network interface arrival and departure.
*/
void
rt_ifannouncemsg(struct ifnet *ifp, int what)
{
struct mbuf *m;
struct rt_addrinfo info;
m = rt_makeifannouncemsg(ifp, RTM_IFANNOUNCE, what, &info);
if (m != NULL)
rt_dispatch(m, AF_UNSPEC);
}
static void
rt_dispatch(struct mbuf *m, sa_family_t saf)
{
struct m_tag *tag;
/*
* Preserve the family from the sockaddr, if any, in an m_tag for
* use when injecting the mbuf into the routing socket buffer from
* the netisr.
*/
if (saf != AF_UNSPEC) {
tag = m_tag_get(PACKET_TAG_RTSOCKFAM, sizeof(unsigned short),
M_NOWAIT);
if (tag == NULL) {
m_freem(m);
return;
}
*(unsigned short *)(tag + 1) = saf;
m_tag_prepend(m, tag);
}
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
#ifdef VIMAGE
if (V_loif)
m->m_pkthdr.rcvif = V_loif;
else {
m_freem(m);
return;
}
#endif
netisr_queue(NETISR_ROUTE, m); /* mbuf is free'd on failure. */
}
1994-05-24 10:09:53 +00:00
/*
* This is used in dumping the kernel table via sysctl().
*/
static int
sysctl_dumpentry(struct radix_node *rn, void *vw)
1994-05-24 10:09:53 +00:00
{
struct walkarg *w = vw;
struct rtentry *rt = (struct rtentry *)rn;
1994-05-24 10:09:53 +00:00
int error = 0, size;
struct rt_addrinfo info;
struct sockaddr_storage ss;
1994-05-24 10:09:53 +00:00
NET_EPOCH_ASSERT();
1994-05-24 10:09:53 +00:00
if (w->w_op == NET_RT_FLAGS && !(rt->rt_flags & w->w_arg))
return 0;
if ((rt->rt_flags & RTF_HOST) == 0
? jailed_without_vnet(w->w_req->td->td_ucred)
: prison_if(w->w_req->td->td_ucred, rt_key(rt)) != 0)
return (0);
1994-05-24 10:09:53 +00:00
bzero((caddr_t)&info, sizeof(info));
info.rti_info[RTAX_DST] = rt_key(rt);
info.rti_info[RTAX_GATEWAY] = rt->rt_gateway;
info.rti_info[RTAX_NETMASK] = rtsock_fix_netmask(rt_key(rt),
rt_mask(rt), &ss);
info.rti_info[RTAX_GENMASK] = 0;
if (rt->rt_ifp && !(rt->rt_ifp->if_flags & IFF_DYING)) {
info.rti_info[RTAX_IFP] = rt->rt_ifp->if_addr->ifa_addr;
info.rti_info[RTAX_IFA] = rt->rt_ifa->ifa_addr;
if (rt->rt_ifp->if_flags & IFF_POINTOPOINT)
info.rti_info[RTAX_BRD] = rt->rt_ifa->ifa_dstaddr;
}
if ((error = rtsock_msg_buffer(RTM_GET, &info, w, &size)) != 0)
return (error);
1995-11-16 19:00:27 +00:00
if (w->w_req && w->w_tmem) {
struct rt_msghdr *rtm = (struct rt_msghdr *)w->w_tmem;
1994-05-24 10:09:53 +00:00
bzero(&rtm->rtm_index,
sizeof(*rtm) - offsetof(struct rt_msghdr, rtm_index));
if (rt->rt_flags & RTF_GWFLAG_COMPAT)
rtm->rtm_flags = RTF_GATEWAY |
(rt->rt_flags & ~RTF_GWFLAG_COMPAT);
else
rtm->rtm_flags = rt->rt_flags;
rt_getmetrics(rt, &rtm->rtm_rmx);
1994-05-24 10:09:53 +00:00
rtm->rtm_index = rt->rt_ifp->if_index;
rtm->rtm_addrs = info.rti_addrs;
1995-11-16 19:00:27 +00:00
error = SYSCTL_OUT(w->w_req, (caddr_t)rtm, size);
return (error);
1994-05-24 10:09:53 +00:00
}
return (error);
}
static int
sysctl_iflist_ifml(struct ifnet *ifp, const struct if_data *src_ifd,
struct rt_addrinfo *info, struct walkarg *w, int len)
{
struct if_msghdrl *ifm;
struct if_data *ifd;
ifm = (struct if_msghdrl *)w->w_tmem;
#ifdef COMPAT_FREEBSD32
if (w->w_req->flags & SCTL_MASK32) {
struct if_msghdrl32 *ifm32;
ifm32 = (struct if_msghdrl32 *)ifm;
ifm32->ifm_addrs = info->rti_addrs;
ifm32->ifm_flags = ifp->if_flags | ifp->if_drv_flags;
ifm32->ifm_index = ifp->if_index;
ifm32->_ifm_spare1 = 0;
ifm32->ifm_len = sizeof(*ifm32);
ifm32->ifm_data_off = offsetof(struct if_msghdrl32, ifm_data);
ifm32->_ifm_spare2 = 0;
ifd = &ifm32->ifm_data;
} else
#endif
{
ifm->ifm_addrs = info->rti_addrs;
ifm->ifm_flags = ifp->if_flags | ifp->if_drv_flags;
ifm->ifm_index = ifp->if_index;
ifm->_ifm_spare1 = 0;
ifm->ifm_len = sizeof(*ifm);
ifm->ifm_data_off = offsetof(struct if_msghdrl, ifm_data);
ifm->_ifm_spare2 = 0;
ifd = &ifm->ifm_data;
}
memcpy(ifd, src_ifd, sizeof(*ifd));
return (SYSCTL_OUT(w->w_req, (caddr_t)ifm, len));
}
static int
sysctl_iflist_ifm(struct ifnet *ifp, const struct if_data *src_ifd,
struct rt_addrinfo *info, struct walkarg *w, int len)
{
struct if_msghdr *ifm;
struct if_data *ifd;
ifm = (struct if_msghdr *)w->w_tmem;
#ifdef COMPAT_FREEBSD32
if (w->w_req->flags & SCTL_MASK32) {
struct if_msghdr32 *ifm32;
ifm32 = (struct if_msghdr32 *)ifm;
ifm32->ifm_addrs = info->rti_addrs;
ifm32->ifm_flags = ifp->if_flags | ifp->if_drv_flags;
ifm32->ifm_index = ifp->if_index;
ifm32->_ifm_spare1 = 0;
ifd = &ifm32->ifm_data;
} else
#endif
{
ifm->ifm_addrs = info->rti_addrs;
ifm->ifm_flags = ifp->if_flags | ifp->if_drv_flags;
ifm->ifm_index = ifp->if_index;
ifm->_ifm_spare1 = 0;
ifd = &ifm->ifm_data;
}
memcpy(ifd, src_ifd, sizeof(*ifd));
return (SYSCTL_OUT(w->w_req, (caddr_t)ifm, len));
}
static int
sysctl_iflist_ifaml(struct ifaddr *ifa, struct rt_addrinfo *info,
struct walkarg *w, int len)
{
struct ifa_msghdrl *ifam;
struct if_data *ifd;
ifam = (struct ifa_msghdrl *)w->w_tmem;
#ifdef COMPAT_FREEBSD32
if (w->w_req->flags & SCTL_MASK32) {
struct ifa_msghdrl32 *ifam32;
ifam32 = (struct ifa_msghdrl32 *)ifam;
ifam32->ifam_addrs = info->rti_addrs;
ifam32->ifam_flags = ifa->ifa_flags;
ifam32->ifam_index = ifa->ifa_ifp->if_index;
ifam32->_ifam_spare1 = 0;
ifam32->ifam_len = sizeof(*ifam32);
ifam32->ifam_data_off =
offsetof(struct ifa_msghdrl32, ifam_data);
ifam32->ifam_metric = ifa->ifa_ifp->if_metric;
ifd = &ifam32->ifam_data;
} else
#endif
{
ifam->ifam_addrs = info->rti_addrs;
ifam->ifam_flags = ifa->ifa_flags;
ifam->ifam_index = ifa->ifa_ifp->if_index;
ifam->_ifam_spare1 = 0;
ifam->ifam_len = sizeof(*ifam);
ifam->ifam_data_off = offsetof(struct ifa_msghdrl, ifam_data);
ifam->ifam_metric = ifa->ifa_ifp->if_metric;
ifd = &ifam->ifam_data;
}
bzero(ifd, sizeof(*ifd));
ifd->ifi_datalen = sizeof(struct if_data);
ifd->ifi_ipackets = counter_u64_fetch(ifa->ifa_ipackets);
ifd->ifi_opackets = counter_u64_fetch(ifa->ifa_opackets);
ifd->ifi_ibytes = counter_u64_fetch(ifa->ifa_ibytes);
ifd->ifi_obytes = counter_u64_fetch(ifa->ifa_obytes);
/* Fixup if_data carp(4) vhid. */
if (carp_get_vhid_p != NULL)
ifd->ifi_vhid = (*carp_get_vhid_p)(ifa);
return (SYSCTL_OUT(w->w_req, w->w_tmem, len));
}
static int
sysctl_iflist_ifam(struct ifaddr *ifa, struct rt_addrinfo *info,
struct walkarg *w, int len)
{
struct ifa_msghdr *ifam;
ifam = (struct ifa_msghdr *)w->w_tmem;
ifam->ifam_addrs = info->rti_addrs;
ifam->ifam_flags = ifa->ifa_flags;
ifam->ifam_index = ifa->ifa_ifp->if_index;
ifam->_ifam_spare1 = 0;
ifam->ifam_metric = ifa->ifa_ifp->if_metric;
return (SYSCTL_OUT(w->w_req, w->w_tmem, len));
}
static int
sysctl_iflist(int af, struct walkarg *w)
1994-05-24 10:09:53 +00:00
{
struct ifnet *ifp;
struct ifaddr *ifa;
struct if_data ifd;
struct rt_addrinfo info;
int len, error = 0;
struct sockaddr_storage ss;
struct epoch_tracker et;
1994-05-24 10:09:53 +00:00
bzero((caddr_t)&info, sizeof(info));
bzero(&ifd, sizeof(ifd));
NET_EPOCH_ENTER(et);
CK_STAILQ_FOREACH(ifp, &V_ifnet, if_link) {
1994-05-24 10:09:53 +00:00
if (w->w_arg && w->w_arg != ifp->if_index)
continue;
if_data_copy(ifp, &ifd);
ifa = ifp->if_addr;
info.rti_info[RTAX_IFP] = ifa->ifa_addr;
error = rtsock_msg_buffer(RTM_IFINFO, &info, w, &len);
if (error != 0)
goto done;
info.rti_info[RTAX_IFP] = NULL;
1995-11-16 19:00:27 +00:00
if (w->w_req && w->w_tmem) {
if (w->w_op == NET_RT_IFLISTL)
error = sysctl_iflist_ifml(ifp, &ifd, &info, w,
len);
else
error = sysctl_iflist_ifm(ifp, &ifd, &info, w,
len);
1994-10-08 22:38:27 +00:00
if (error)
goto done;
1994-05-24 10:09:53 +00:00
}
ifnet: Replace if_addr_lock rwlock with epoch + mutex Run on LLNW canaries and tested by pho@ gallatin: Using a 14-core, 28-HTT single socket E5-2697 v3 with a 40GbE MLX5 based ConnectX 4-LX NIC, I see an almost 12% improvement in received packet rate, and a larger improvement in bytes delivered all the way to userspace. When the host receiving 64 streams of netperf -H $DUT -t UDP_STREAM -- -m 1, I see, using nstat -I mce0 1 before the patch: InMpps OMpps InGbs OGbs err TCP Est %CPU syscalls csw irq GBfree 4.98 0.00 4.42 0.00 4235592 33 83.80 4720653 2149771 1235 247.32 4.73 0.00 4.20 0.00 4025260 33 82.99 4724900 2139833 1204 247.32 4.72 0.00 4.20 0.00 4035252 33 82.14 4719162 2132023 1264 247.32 4.71 0.00 4.21 0.00 4073206 33 83.68 4744973 2123317 1347 247.32 4.72 0.00 4.21 0.00 4061118 33 80.82 4713615 2188091 1490 247.32 4.72 0.00 4.21 0.00 4051675 33 85.29 4727399 2109011 1205 247.32 4.73 0.00 4.21 0.00 4039056 33 84.65 4724735 2102603 1053 247.32 After the patch InMpps OMpps InGbs OGbs err TCP Est %CPU syscalls csw irq GBfree 5.43 0.00 4.20 0.00 3313143 33 84.96 5434214 1900162 2656 245.51 5.43 0.00 4.20 0.00 3308527 33 85.24 5439695 1809382 2521 245.51 5.42 0.00 4.19 0.00 3316778 33 87.54 5416028 1805835 2256 245.51 5.42 0.00 4.19 0.00 3317673 33 90.44 5426044 1763056 2332 245.51 5.42 0.00 4.19 0.00 3314839 33 88.11 5435732 1792218 2499 245.52 5.44 0.00 4.19 0.00 3293228 33 91.84 5426301 1668597 2121 245.52 Similarly, netperf reports 230Mb/s before the patch, and 270Mb/s after the patch Reviewed by: gallatin Sponsored by: Limelight Networks Differential Revision: https://reviews.freebsd.org/D15366
2018-05-18 20:13:34 +00:00
while ((ifa = CK_STAILQ_NEXT(ifa, ifa_link)) != NULL) {
1994-05-24 10:09:53 +00:00
if (af && af != ifa->ifa_addr->sa_family)
continue;
if (prison_if(w->w_req->td->td_ucred,
ifa->ifa_addr) != 0)
This Implements the mumbled about "Jail" feature. This is a seriously beefed up chroot kind of thing. The process is jailed along the same lines as a chroot does it, but with additional tough restrictions imposed on what the superuser can do. For all I know, it is safe to hand over the root bit inside a prison to the customer living in that prison, this is what it was developed for in fact: "real virtual servers". Each prison has an ip number associated with it, which all IP communications will be coerced to use and each prison has its own hostname. Needless to say, you need more RAM this way, but the advantage is that each customer can run their own particular version of apache and not stomp on the toes of their neighbors. It generally does what one would expect, but setting up a jail still takes a little knowledge. A few notes: I have no scripts for setting up a jail, don't ask me for them. The IP number should be an alias on one of the interfaces. mount a /proc in each jail, it will make ps more useable. /proc/<pid>/status tells the hostname of the prison for jailed processes. Quotas are only sensible if you have a mountpoint per prison. There are no privisions for stopping resource-hogging. Some "#ifdef INET" and similar may be missing (send patches!) If somebody wants to take it from here and develop it into more of a "virtual machine" they should be most welcome! Tools, comments, patches & documentation most welcome. Have fun... Sponsored by: http://www.rndassociates.com/ Run for almost a year by: http://www.servetheweb.com/
1999-04-28 11:38:52 +00:00
continue;
info.rti_info[RTAX_IFA] = ifa->ifa_addr;
info.rti_info[RTAX_NETMASK] = rtsock_fix_netmask(
ifa->ifa_addr, ifa->ifa_netmask, &ss);
info.rti_info[RTAX_BRD] = ifa->ifa_dstaddr;
error = rtsock_msg_buffer(RTM_NEWADDR, &info, w, &len);
if (error != 0)
goto done;
1995-11-16 19:00:27 +00:00
if (w->w_req && w->w_tmem) {
if (w->w_op == NET_RT_IFLISTL)
error = sysctl_iflist_ifaml(ifa, &info,
w, len);
else
error = sysctl_iflist_ifam(ifa, &info,
w, len);
1994-10-08 22:38:27 +00:00
if (error)
goto done;
1994-05-24 10:09:53 +00:00
}
}
info.rti_info[RTAX_IFA] = NULL;
info.rti_info[RTAX_NETMASK] = NULL;
info.rti_info[RTAX_BRD] = NULL;
1994-05-24 10:09:53 +00:00
}
done:
NET_EPOCH_EXIT(et);
return (error);
1994-05-24 10:09:53 +00:00
}
static int
sysctl_ifmalist(int af, struct walkarg *w)
{
struct rt_addrinfo info;
struct epoch_tracker et;
struct ifaddr *ifa;
struct ifmultiaddr *ifma;
struct ifnet *ifp;
int error, len;
error = 0;
bzero((caddr_t)&info, sizeof(info));
NET_EPOCH_ENTER(et);
CK_STAILQ_FOREACH(ifp, &V_ifnet, if_link) {
if (w->w_arg && w->w_arg != ifp->if_index)
continue;
ifa = ifp->if_addr;
info.rti_info[RTAX_IFP] = ifa ? ifa->ifa_addr : NULL;
ifnet: Replace if_addr_lock rwlock with epoch + mutex Run on LLNW canaries and tested by pho@ gallatin: Using a 14-core, 28-HTT single socket E5-2697 v3 with a 40GbE MLX5 based ConnectX 4-LX NIC, I see an almost 12% improvement in received packet rate, and a larger improvement in bytes delivered all the way to userspace. When the host receiving 64 streams of netperf -H $DUT -t UDP_STREAM -- -m 1, I see, using nstat -I mce0 1 before the patch: InMpps OMpps InGbs OGbs err TCP Est %CPU syscalls csw irq GBfree 4.98 0.00 4.42 0.00 4235592 33 83.80 4720653 2149771 1235 247.32 4.73 0.00 4.20 0.00 4025260 33 82.99 4724900 2139833 1204 247.32 4.72 0.00 4.20 0.00 4035252 33 82.14 4719162 2132023 1264 247.32 4.71 0.00 4.21 0.00 4073206 33 83.68 4744973 2123317 1347 247.32 4.72 0.00 4.21 0.00 4061118 33 80.82 4713615 2188091 1490 247.32 4.72 0.00 4.21 0.00 4051675 33 85.29 4727399 2109011 1205 247.32 4.73 0.00 4.21 0.00 4039056 33 84.65 4724735 2102603 1053 247.32 After the patch InMpps OMpps InGbs OGbs err TCP Est %CPU syscalls csw irq GBfree 5.43 0.00 4.20 0.00 3313143 33 84.96 5434214 1900162 2656 245.51 5.43 0.00 4.20 0.00 3308527 33 85.24 5439695 1809382 2521 245.51 5.42 0.00 4.19 0.00 3316778 33 87.54 5416028 1805835 2256 245.51 5.42 0.00 4.19 0.00 3317673 33 90.44 5426044 1763056 2332 245.51 5.42 0.00 4.19 0.00 3314839 33 88.11 5435732 1792218 2499 245.52 5.44 0.00 4.19 0.00 3293228 33 91.84 5426301 1668597 2121 245.52 Similarly, netperf reports 230Mb/s before the patch, and 270Mb/s after the patch Reviewed by: gallatin Sponsored by: Limelight Networks Differential Revision: https://reviews.freebsd.org/D15366
2018-05-18 20:13:34 +00:00
CK_STAILQ_FOREACH(ifma, &ifp->if_multiaddrs, ifma_link) {
if (af && af != ifma->ifma_addr->sa_family)
continue;
if (prison_if(w->w_req->td->td_ucred,
ifma->ifma_addr) != 0)
continue;
info.rti_info[RTAX_IFA] = ifma->ifma_addr;
info.rti_info[RTAX_GATEWAY] =
(ifma->ifma_addr->sa_family != AF_LINK) ?
ifma->ifma_lladdr : NULL;
error = rtsock_msg_buffer(RTM_NEWMADDR, &info, w, &len);
if (error != 0)
break;
if (w->w_req && w->w_tmem) {
struct ifma_msghdr *ifmam;
ifmam = (struct ifma_msghdr *)w->w_tmem;
ifmam->ifmam_index = ifma->ifma_ifp->if_index;
ifmam->ifmam_flags = 0;
ifmam->ifmam_addrs = info.rti_addrs;
ifmam->_ifmam_spare1 = 0;
error = SYSCTL_OUT(w->w_req, w->w_tmem, len);
if (error != 0)
break;
}
}
if (error != 0)
break;
}
NET_EPOCH_EXIT(et);
return (error);
}
1995-11-16 19:00:27 +00:00
static int
sysctl_rtsock(SYSCTL_HANDLER_ARGS)
1994-05-24 10:09:53 +00:00
{
RIB_RLOCK_TRACKER;
1995-11-16 19:00:27 +00:00
int *name = (int *)arg1;
u_int namelen = arg2;
struct rib_head *rnh = NULL; /* silence compiler. */
int i, lim, error = EINVAL;
int fib = 0;
u_char af;
1994-05-24 10:09:53 +00:00
struct walkarg w;
1995-11-16 19:00:27 +00:00
name ++;
namelen--;
if (req->newptr)
1994-05-24 10:09:53 +00:00
return (EPERM);
if (name[1] == NET_RT_DUMP) {
if (namelen == 3)
fib = req->td->td_proc->p_fibnum;
else if (namelen == 4)
fib = (name[3] == RT_ALL_FIBS) ?
req->td->td_proc->p_fibnum : name[3];
else
return ((namelen < 3) ? EISDIR : ENOTDIR);
if (fib < 0 || fib >= rt_numfibs)
return (EINVAL);
} else if (namelen != 3)
return ((namelen < 3) ? EISDIR : ENOTDIR);
1994-05-24 10:09:53 +00:00
af = name[0];
if (af > AF_MAX)
return (EINVAL);
bzero(&w, sizeof(w));
1994-05-24 10:09:53 +00:00
w.w_op = name[1];
w.w_arg = name[2];
1995-11-16 19:00:27 +00:00
w.w_req = req;
1994-05-24 10:09:53 +00:00
error = sysctl_wire_old_buffer(req, 0);
if (error)
return (error);
/*
* Allocate reply buffer in advance.
* All rtsock messages has maximum length of u_short.
*/
w.w_tmemsize = 65536;
w.w_tmem = malloc(w.w_tmemsize, M_TEMP, M_WAITOK);
1994-05-24 10:09:53 +00:00
switch (w.w_op) {
case NET_RT_DUMP:
case NET_RT_FLAGS:
if (af == 0) { /* dump all tables */
i = 1;
lim = AF_MAX;
} else /* dump only one table */
i = lim = af;
2008-12-26 19:45:24 +00:00
/*
* take care of llinfo entries, the caller must
* specify an AF
*/
if (w.w_op == NET_RT_FLAGS &&
(w.w_arg == 0 || w.w_arg & RTF_LLINFO)) {
2008-12-26 19:45:24 +00:00
if (af != 0)
error = lltable_sysctl_dumparp(af, w.w_req);
else
error = EINVAL;
break;
}
/*
* take care of routing entries
*/
for (error = 0; error == 0 && i <= lim; i++) {
rnh = rt_tables_get_rnh(fib, i);
if (rnh != NULL) {
struct epoch_tracker et;
RIB_RLOCK(rnh);
NET_EPOCH_ENTER(et);
error = rnh->rnh_walktree(&rnh->head,
sysctl_dumpentry, &w);
NET_EPOCH_EXIT(et);
RIB_RUNLOCK(rnh);
} else if (af != 0)
2002-12-24 03:03:39 +00:00
error = EAFNOSUPPORT;
}
1994-05-24 10:09:53 +00:00
break;
case NET_RT_IFLIST:
case NET_RT_IFLISTL:
1994-05-24 10:09:53 +00:00
error = sysctl_iflist(af, &w);
break;
case NET_RT_IFMALIST:
error = sysctl_ifmalist(af, &w);
break;
1994-05-24 10:09:53 +00:00
}
free(w.w_tmem, M_TEMP);
1994-05-24 10:09:53 +00:00
return (error);
}
static SYSCTL_NODE(_net, PF_ROUTE, routetable, CTLFLAG_RD, sysctl_rtsock, "");
1995-11-16 19:00:27 +00:00
1994-05-24 10:09:53 +00:00
/*
* Definitions of protocols supported in the ROUTE domain.
*/
static struct domain routedomain; /* or at least forward */
1994-05-24 10:09:53 +00:00
1995-11-16 19:00:27 +00:00
static struct protosw routesw[] = {
{
.pr_type = SOCK_RAW,
.pr_domain = &routedomain,
.pr_flags = PR_ATOMIC|PR_ADDR,
.pr_output = route_output,
.pr_ctlinput = raw_ctlinput,
.pr_init = raw_init,
.pr_usrreqs = &route_usrreqs
1994-05-24 10:09:53 +00:00
}
};
static struct domain routedomain = {
.dom_family = PF_ROUTE,
.dom_name = "route",
.dom_protosw = routesw,
.dom_protoswNPROTOSW = &routesw[nitems(routesw)]
};
VNET_DOMAIN_SET(route);