freebsd-dev/sys/kern/uipc_domain.c
Marko Zec bfe1aba468 Introduce vnet module registration / initialization framework with
dependency tracking and ordering enforcement.

With this change, per-vnet initialization functions introduced with
r190787 are no longer directly called from traditional initialization
functions (which cc in most cases inlined to pre-r190787 code), but are
instead registered via the vnet framework first, and are invoked only
after all prerequisite modules have been initialized.  In the long run,
this framework should allow us to both initialize and dismantle
multiple vnet instances in a correct order.

The problem this change aims to solve is how to replay the
initialization sequence of various network stack components, which
have been traditionally triggered via different mechanisms (SYSINIT,
protosw).  Note that this initialization sequence was and still can be
subtly different depending on whether certain pieces of code have been
statically compiled into the kernel, loaded as modules by boot
loader, or kldloaded at run time.

The approach is simple - we record the initialization sequence
established by the traditional mechanisms whenever vnet_mod_register()
is called for a particular vnet module.  The vnet_mod_register_multi()
variant allows a single initializer function to be registered multiple
times but with different arguments - currently this is only used in
kern/uipc_domain.c by net_add_domain() with different struct domain *
as arguments, which allows for protosw-registered initialization
routines to be invoked in a correct order by the new vnet
initialization framework.

For the purpose of identifying vnet modules, each vnet module has to
have a unique ID, which is statically assigned in sys/vimage.h.
Dynamic assignment of vnet module IDs is not supported yet.

A vnet module may specify a single prerequisite module at registration
time by filling in the vmi_dependson field of its vnet_modinfo struct
with the ID of the module it depends on.  Unless specified otherwise,
all vnet modules depend on VNET_MOD_NET (container for ifnet list head,
rt_tables etc.), which thus has to and will always be initialized
first.  The framework will panic if it detects any unresolved
dependencies before completing system initialization.  Detection of
unresolved dependencies for vnet modules registered after boot
(kldloaded modules) is not provided.

Note that the fact that each module can specify only a single
prerequisite may become problematic in the long run.  In particular,
INET6 depends on INET being already instantiated, due to TCP / UDP
structures residing in INET container.  IPSEC also depends on INET,
which will in turn additionally complicate making INET6-only kernel
configs a reality.

The entire registration framework can be compiled out by turning on the
VIMAGE_GLOBALS kernel config option.

Reviewed by:	bz
Approved by:	julian (mentor)
2009-04-11 05:58:58 +00:00

518 lines
14 KiB
C

/*-
* Copyright (c) 1982, 1986, 1993
* The Regents of the University of California. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* 4. Neither the name of the University nor the names of its contributors
* may be used to endorse or promote products derived from this software
* without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*
* @(#)uipc_domain.c 8.2 (Berkeley) 10/18/93
*/
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
#include <sys/param.h>
#include <sys/socket.h>
#include <sys/protosw.h>
#include <sys/domain.h>
#include <sys/eventhandler.h>
#include <sys/mbuf.h>
#include <sys/kernel.h>
#include <sys/lock.h>
#include <sys/mutex.h>
#include <sys/socketvar.h>
#include <sys/systm.h>
#include <sys/vimage.h>
#include <vm/uma.h>
/*
* System initialization
*
* Note: domain initialization takes place on a per domain basis
* as a result of traversing a SYSINIT linker set. Most likely,
* each domain would want to call DOMAIN_SET(9) itself, which
* would cause the domain to be added just after domaininit()
* is called during startup.
*
* See DOMAIN_SET(9) for details on its use.
*/
static void domaininit(void *);
SYSINIT(domain, SI_SUB_PROTO_DOMAIN, SI_ORDER_FIRST, domaininit, NULL);
static void domainfinalize(void *);
SYSINIT(domainfin, SI_SUB_PROTO_IFATTACHDOMAIN, SI_ORDER_FIRST, domainfinalize,
NULL);
static vnet_attach_fn net_init_domain;
static struct callout pffast_callout;
static struct callout pfslow_callout;
static void pffasttimo(void *);
static void pfslowtimo(void *);
struct domain *domains; /* registered protocol domains */
int domain_init_status = 0;
static struct mtx dom_mtx; /* domain list lock */
MTX_SYSINIT(domain, &dom_mtx, "domain list", MTX_DEF);
/*
* Dummy protocol specific user requests function pointer array.
* All functions return EOPNOTSUPP.
*/
struct pr_usrreqs nousrreqs = {
.pru_accept = pru_accept_notsupp,
.pru_attach = pru_attach_notsupp,
.pru_bind = pru_bind_notsupp,
.pru_connect = pru_connect_notsupp,
.pru_connect2 = pru_connect2_notsupp,
.pru_control = pru_control_notsupp,
.pru_disconnect = pru_disconnect_notsupp,
.pru_listen = pru_listen_notsupp,
.pru_peeraddr = pru_peeraddr_notsupp,
.pru_rcvd = pru_rcvd_notsupp,
.pru_rcvoob = pru_rcvoob_notsupp,
.pru_send = pru_send_notsupp,
.pru_sense = pru_sense_null,
.pru_shutdown = pru_shutdown_notsupp,
.pru_sockaddr = pru_sockaddr_notsupp,
.pru_sosend = pru_sosend_notsupp,
.pru_soreceive = pru_soreceive_notsupp,
.pru_sopoll = pru_sopoll_notsupp,
};
#ifndef VIMAGE_GLOBALS
vnet_modinfo_t vnet_domain_modinfo = {
.vmi_id = VNET_MOD_DOMAIN,
.vmi_name = "domain",
.vmi_iattach = net_init_domain
};
#endif
static void
protosw_init(struct protosw *pr)
{
struct pr_usrreqs *pu;
pu = pr->pr_usrreqs;
KASSERT(pu != NULL, ("protosw_init: %ssw[%d] has no usrreqs!",
pr->pr_domain->dom_name,
(int)(pr - pr->pr_domain->dom_protosw)));
/*
* Protocol switch methods fall into three categories: mandatory,
* mandatory but protosw_init() provides a default, and optional.
*
* For true protocols (i.e., pru_attach != NULL), KASSERT truly
* mandatory methods with no defaults, and initialize defaults for
* other mandatory methods if the protocol hasn't defined an
* implementation (NULL function pointer).
*/
#if 0
if (pu->pru_attach != NULL) {
KASSERT(pu->pru_abort != NULL,
("protosw_init: %ssw[%d] pru_abort NULL",
pr->pr_domain->dom_name,
(int)(pr - pr->pr_domain->dom_protosw)));
KASSERT(pu->pru_send != NULL,
("protosw_init: %ssw[%d] pru_send NULL",
pr->pr_domain->dom_name,
(int)(pr - pr->pr_domain->dom_protosw)));
}
#endif
#define DEFAULT(foo, bar) if ((foo) == NULL) (foo) = (bar)
DEFAULT(pu->pru_accept, pru_accept_notsupp);
DEFAULT(pu->pru_bind, pru_bind_notsupp);
DEFAULT(pu->pru_connect, pru_connect_notsupp);
DEFAULT(pu->pru_connect2, pru_connect2_notsupp);
DEFAULT(pu->pru_control, pru_control_notsupp);
DEFAULT(pu->pru_disconnect, pru_disconnect_notsupp);
DEFAULT(pu->pru_listen, pru_listen_notsupp);
DEFAULT(pu->pru_peeraddr, pru_peeraddr_notsupp);
DEFAULT(pu->pru_rcvd, pru_rcvd_notsupp);
DEFAULT(pu->pru_rcvoob, pru_rcvoob_notsupp);
DEFAULT(pu->pru_sense, pru_sense_null);
DEFAULT(pu->pru_shutdown, pru_shutdown_notsupp);
DEFAULT(pu->pru_sockaddr, pru_sockaddr_notsupp);
DEFAULT(pu->pru_sosend, sosend_generic);
DEFAULT(pu->pru_soreceive, soreceive_generic);
DEFAULT(pu->pru_sopoll, sopoll_generic);
#undef DEFAULT
if (pr->pr_init)
(*pr->pr_init)();
}
/*
* Add a new protocol domain to the list of supported domains
* Note: you cant unload it again because a socket may be using it.
* XXX can't fail at this time.
*/
static int
net_init_domain(const void *arg)
{
const struct domain *dp = arg;
struct protosw *pr;
if (dp->dom_init)
(*dp->dom_init)();
for (pr = dp->dom_protosw; pr < dp->dom_protoswNPROTOSW; pr++)
protosw_init(pr);
/*
* update global information about maximums
*/
max_hdr = max_linkhdr + max_protohdr;
max_datalen = MHLEN - max_hdr;
if (max_datalen < 1)
panic("%s: max_datalen < 1", __func__);
return (0);
}
/*
* Add a new protocol domain to the list of supported domains
* Note: you cant unload it again because a socket may be using it.
* XXX can't fail at this time.
*/
void
net_add_domain(void *data)
{
struct domain *dp;
dp = (struct domain *)data;
mtx_lock(&dom_mtx);
dp->dom_next = domains;
domains = dp;
KASSERT(domain_init_status >= 1,
("attempt to net_add_domain(%s) before domaininit()",
dp->dom_name));
#ifndef INVARIANTS
if (domain_init_status < 1)
printf("WARNING: attempt to net_add_domain(%s) before "
"domaininit()\n", dp->dom_name);
#endif
#ifdef notyet
KASSERT(domain_init_status < 2,
("attempt to net_add_domain(%s) after domainfinalize()",
dp->dom_name));
#else
if (domain_init_status >= 2)
printf("WARNING: attempt to net_add_domain(%s) after "
"domainfinalize()\n", dp->dom_name);
#endif
mtx_unlock(&dom_mtx);
#ifndef VIMAGE_GLOBALS
vnet_mod_register_multi(&vnet_domain_modinfo, dp, dp->dom_name);
#else
net_init_domain(dp);
#endif
}
static void
socket_zone_change(void *tag)
{
uma_zone_set_max(socket_zone, maxsockets);
}
/* ARGSUSED*/
static void
domaininit(void *dummy)
{
/*
* Before we do any setup, make sure to initialize the
* zone allocator we get struct sockets from.
*/
socket_zone = uma_zcreate("socket", sizeof(struct socket), NULL, NULL,
NULL, NULL, UMA_ALIGN_PTR, UMA_ZONE_NOFREE);
uma_zone_set_max(socket_zone, maxsockets);
EVENTHANDLER_REGISTER(maxsockets_change, socket_zone_change, NULL,
EVENTHANDLER_PRI_FIRST);
if (max_linkhdr < 16) /* XXX */
max_linkhdr = 16;
callout_init(&pffast_callout, CALLOUT_MPSAFE);
callout_init(&pfslow_callout, CALLOUT_MPSAFE);
mtx_lock(&dom_mtx);
KASSERT(domain_init_status == 0, ("domaininit called too late!"));
domain_init_status = 1;
mtx_unlock(&dom_mtx);
}
/* ARGSUSED*/
static void
domainfinalize(void *dummy)
{
mtx_lock(&dom_mtx);
KASSERT(domain_init_status == 1, ("domainfinalize called too late!"));
domain_init_status = 2;
mtx_unlock(&dom_mtx);
callout_reset(&pffast_callout, 1, pffasttimo, NULL);
callout_reset(&pfslow_callout, 1, pfslowtimo, NULL);
}
struct protosw *
pffindtype(int family, int type)
{
struct domain *dp;
struct protosw *pr;
for (dp = domains; dp; dp = dp->dom_next)
if (dp->dom_family == family)
goto found;
return (0);
found:
for (pr = dp->dom_protosw; pr < dp->dom_protoswNPROTOSW; pr++)
if (pr->pr_type && pr->pr_type == type)
return (pr);
return (0);
}
struct protosw *
pffindproto(int family, int protocol, int type)
{
struct domain *dp;
struct protosw *pr;
struct protosw *maybe = 0;
if (family == 0)
return (0);
for (dp = domains; dp; dp = dp->dom_next)
if (dp->dom_family == family)
goto found;
return (0);
found:
for (pr = dp->dom_protosw; pr < dp->dom_protoswNPROTOSW; pr++) {
if ((pr->pr_protocol == protocol) && (pr->pr_type == type))
return (pr);
if (type == SOCK_RAW && pr->pr_type == SOCK_RAW &&
pr->pr_protocol == 0 && maybe == (struct protosw *)0)
maybe = pr;
}
return (maybe);
}
/*
* The caller must make sure that the new protocol is fully set up and ready to
* accept requests before it is registered.
*/
int
pf_proto_register(int family, struct protosw *npr)
{
struct domain *dp;
struct protosw *pr, *fpr;
/* Sanity checks. */
if (family == 0)
return (EPFNOSUPPORT);
if (npr->pr_type == 0)
return (EPROTOTYPE);
if (npr->pr_protocol == 0)
return (EPROTONOSUPPORT);
if (npr->pr_usrreqs == NULL)
return (ENXIO);
/* Try to find the specified domain based on the family. */
for (dp = domains; dp; dp = dp->dom_next)
if (dp->dom_family == family)
goto found;
return (EPFNOSUPPORT);
found:
/* Initialize backpointer to struct domain. */
npr->pr_domain = dp;
fpr = NULL;
/*
* Protect us against races when two protocol registrations for
* the same protocol happen at the same time.
*/
mtx_lock(&dom_mtx);
/* The new protocol must not yet exist. */
for (pr = dp->dom_protosw; pr < dp->dom_protoswNPROTOSW; pr++) {
if ((pr->pr_type == npr->pr_type) &&
(pr->pr_protocol == npr->pr_protocol)) {
mtx_unlock(&dom_mtx);
return (EEXIST); /* XXX: Check only protocol? */
}
/* While here, remember the first free spacer. */
if ((fpr == NULL) && (pr->pr_protocol == PROTO_SPACER))
fpr = pr;
}
/* If no free spacer is found we can't add the new protocol. */
if (fpr == NULL) {
mtx_unlock(&dom_mtx);
return (ENOMEM);
}
/* Copy the new struct protosw over the spacer. */
bcopy(npr, fpr, sizeof(*fpr));
/* Job is done, no more protection required. */
mtx_unlock(&dom_mtx);
/* Initialize and activate the protocol. */
protosw_init(fpr);
return (0);
}
/*
* The caller must make sure the protocol and its functions correctly shut down
* all sockets and release all locks and memory references.
*/
int
pf_proto_unregister(int family, int protocol, int type)
{
struct domain *dp;
struct protosw *pr, *dpr;
/* Sanity checks. */
if (family == 0)
return (EPFNOSUPPORT);
if (protocol == 0)
return (EPROTONOSUPPORT);
if (type == 0)
return (EPROTOTYPE);
/* Try to find the specified domain based on the family type. */
for (dp = domains; dp; dp = dp->dom_next)
if (dp->dom_family == family)
goto found;
return (EPFNOSUPPORT);
found:
dpr = NULL;
/* Lock out everyone else while we are manipulating the protosw. */
mtx_lock(&dom_mtx);
/* The protocol must exist and only once. */
for (pr = dp->dom_protosw; pr < dp->dom_protoswNPROTOSW; pr++) {
if ((pr->pr_type == type) && (pr->pr_protocol == protocol)) {
if (dpr != NULL) {
mtx_unlock(&dom_mtx);
return (EMLINK); /* Should not happen! */
} else
dpr = pr;
}
}
/* Protocol does not exist. */
if (dpr == NULL) {
mtx_unlock(&dom_mtx);
return (EPROTONOSUPPORT);
}
/* De-orbit the protocol and make the slot available again. */
dpr->pr_type = 0;
dpr->pr_domain = dp;
dpr->pr_protocol = PROTO_SPACER;
dpr->pr_flags = 0;
dpr->pr_input = NULL;
dpr->pr_output = NULL;
dpr->pr_ctlinput = NULL;
dpr->pr_ctloutput = NULL;
dpr->pr_init = NULL;
dpr->pr_fasttimo = NULL;
dpr->pr_slowtimo = NULL;
dpr->pr_drain = NULL;
dpr->pr_usrreqs = &nousrreqs;
/* Job is done, not more protection required. */
mtx_unlock(&dom_mtx);
return (0);
}
void
pfctlinput(int cmd, struct sockaddr *sa)
{
struct domain *dp;
struct protosw *pr;
for (dp = domains; dp; dp = dp->dom_next)
for (pr = dp->dom_protosw; pr < dp->dom_protoswNPROTOSW; pr++)
if (pr->pr_ctlinput)
(*pr->pr_ctlinput)(cmd, sa, (void *)0);
}
void
pfctlinput2(int cmd, struct sockaddr *sa, void *ctlparam)
{
struct domain *dp;
struct protosw *pr;
if (!sa)
return;
for (dp = domains; dp; dp = dp->dom_next) {
/*
* the check must be made by xx_ctlinput() anyways, to
* make sure we use data item pointed to by ctlparam in
* correct way. the following check is made just for safety.
*/
if (dp->dom_family != sa->sa_family)
continue;
for (pr = dp->dom_protosw; pr < dp->dom_protoswNPROTOSW; pr++)
if (pr->pr_ctlinput)
(*pr->pr_ctlinput)(cmd, sa, ctlparam);
}
}
static void
pfslowtimo(void *arg)
{
struct domain *dp;
struct protosw *pr;
for (dp = domains; dp; dp = dp->dom_next)
for (pr = dp->dom_protosw; pr < dp->dom_protoswNPROTOSW; pr++)
if (pr->pr_slowtimo)
(*pr->pr_slowtimo)();
callout_reset(&pfslow_callout, hz/2, pfslowtimo, NULL);
}
static void
pffasttimo(void *arg)
{
struct domain *dp;
struct protosw *pr;
for (dp = domains; dp; dp = dp->dom_next)
for (pr = dp->dom_protosw; pr < dp->dom_protoswNPROTOSW; pr++)
if (pr->pr_fasttimo)
(*pr->pr_fasttimo)();
callout_reset(&pffast_callout, hz/5, pffasttimo, NULL);
}