summaryrefslogtreecommitdiffstats
path: root/sys/net/if.c
diff options
context:
space:
mode:
authorrwatson <rwatson@FreeBSD.org>2009-07-14 22:48:30 +0000
committerrwatson <rwatson@FreeBSD.org>2009-07-14 22:48:30 +0000
commit57ca4583e728cab422fba8f15de10bd0b637b3dd (patch)
tree13848f891fb2f7a396281b31633563d0f764ff65 /sys/net/if.c
parentef443476d9706035ac219f0280ef0b817dda7a6d (diff)
downloadFreeBSD-src-57ca4583e728cab422fba8f15de10bd0b637b3dd.zip
FreeBSD-src-57ca4583e728cab422fba8f15de10bd0b637b3dd.tar.gz
Build on Jeff Roberson's linker-set based dynamic per-CPU allocator
(DPCPU), as suggested by Peter Wemm, and implement a new per-virtual network stack memory allocator. Modify vnet to use the allocator instead of monolithic global container structures (vinet, ...). This change solves many binary compatibility problems associated with VIMAGE, and restores ELF symbols for virtualized global variables. Each virtualized global variable exists as a "reference copy", and also once per virtual network stack. Virtualized global variables are tagged at compile-time, placing the in a special linker set, which is loaded into a contiguous region of kernel memory. Virtualized global variables in the base kernel are linked as normal, but those in modules are copied and relocated to a reserved portion of the kernel's vnet region with the help of a the kernel linker. Virtualized global variables exist in per-vnet memory set up when the network stack instance is created, and are initialized statically from the reference copy. Run-time access occurs via an accessor macro, which converts from the current vnet and requested symbol to a per-vnet address. When "options VIMAGE" is not compiled into the kernel, normal global ELF symbols will be used instead and indirection is avoided. This change restores static initialization for network stack global variables, restores support for non-global symbols and types, eliminates the need for many subsystem constructors, eliminates large per-subsystem structures that caused many binary compatibility issues both for monitoring applications (netstat) and kernel modules, removes the per-function INIT_VNET_*() macros throughout the stack, eliminates the need for vnet_symmap ksym(2) munging, and eliminates duplicate definitions of virtualized globals under VIMAGE_GLOBALS. Bump __FreeBSD_version and update UPDATING. Portions submitted by: bz Reviewed by: bz, zec Discussed with: gnn, jamie, jeff, jhb, julian, sam Suggested by: peter Approved by: re (kensmith)
Diffstat (limited to 'sys/net/if.c')
-rw-r--r--sys/net/if.c86
1 files changed, 17 insertions, 69 deletions
diff --git a/sys/net/if.c b/sys/net/if.c
index 4803477..daab1ac 100644
--- a/sys/net/if.c
+++ b/sys/net/if.c
@@ -84,7 +84,6 @@
#endif
#ifdef INET
#include <netinet/if_ether.h>
-#include <netinet/vinet.h>
#endif
#if defined(INET) || defined(INET6)
#ifdef DEV_CARP
@@ -94,12 +93,6 @@
#include <security/mac/mac_framework.h>
-#ifndef VIMAGE
-#ifndef VIMAGE_GLOBALS
-struct vnet_net vnet_net_0;
-#endif
-#endif
-
struct ifindex_entry {
struct ifnet *ife_ifnet;
};
@@ -162,37 +155,30 @@ static int vnet_net_iattach(const void *);
static int vnet_net_idetach(const void *);
#endif
-#ifdef VIMAGE_GLOBALS
-struct ifnethead ifnet; /* depend on static init XXX */
-struct ifgrouphead ifg_head;
-int if_index;
-static int if_indexlim;
+VNET_DEFINE(struct ifnethead, ifnet); /* depend on static init XXX */
+VNET_DEFINE(struct ifgrouphead, ifg_head);
+VNET_DEFINE(int, if_index);
+static VNET_DEFINE(int, if_indexlim) = 8;
+
/* Table of ifnet by index. Locked with ifnet_lock. */
-static struct ifindex_entry *ifindex_table;
-#endif
+static VNET_DEFINE(struct ifindex_entry *, ifindex_table);
+
+#define V_if_indexlim VNET_GET(if_indexlim)
+#define V_ifindex_table VNET_GET(ifindex_table)
int ifqmaxlen = IFQ_MAXLEN;
struct rwlock ifnet_lock;
static if_com_alloc_t *if_com_alloc[256];
static if_com_free_t *if_com_free[256];
-#ifndef VIMAGE_GLOBALS
-static struct vnet_symmap vnet_net_symmap[] = {
- VNET_SYMMAP(net, ifnet),
- VNET_SYMMAP_END
-};
-
+#ifdef VIMAGE
static const vnet_modinfo_t vnet_net_modinfo = {
.vmi_id = VNET_MOD_NET,
.vmi_name = "net",
- .vmi_size = sizeof(struct vnet_net),
- .vmi_symmap = vnet_net_symmap,
.vmi_iattach = vnet_net_iattach,
-#ifdef VIMAGE
.vmi_idetach = vnet_net_idetach
-#endif
};
-#endif /* !VIMAGE_GLOBALS */
+#endif
/*
* System initialization
@@ -207,7 +193,6 @@ MALLOC_DEFINE(M_IFMADDR, "ether_multi", "link-level multicast address");
struct ifnet *
ifnet_byindex_locked(u_short idx)
{
- INIT_VNET_NET(curvnet);
if (idx > V_if_index)
return (NULL);
@@ -244,7 +229,6 @@ ifnet_byindex_ref(u_short idx)
static void
ifnet_setbyindex(u_short idx, struct ifnet *ifp)
{
- INIT_VNET_NET(curvnet);
IFNET_WLOCK_ASSERT();
@@ -276,7 +260,7 @@ static void
if_init(void *dummy __unused)
{
-#ifndef VIMAGE_GLOBALS
+#ifdef VIMAGE
vnet_mod_register(&vnet_net_modinfo);
#else
vnet_net_iattach(NULL);
@@ -289,11 +273,6 @@ if_init(void *dummy __unused)
static int
vnet_net_iattach(const void *unused __unused)
{
- INIT_VNET_NET(curvnet);
-
- V_if_index = 0;
- V_ifindex_table = NULL;
- V_if_indexlim = 8;
TAILQ_INIT(&V_ifnet);
TAILQ_INIT(&V_ifg_head);
@@ -306,7 +285,6 @@ vnet_net_iattach(const void *unused __unused)
static int
vnet_net_idetach(const void *unused __unused)
{
- INIT_VNET_NET(curvnet);
VNET_ASSERT(TAILQ_EMPTY(&V_ifnet));
VNET_ASSERT(TAILQ_EMPTY(&V_ifg_head));
@@ -320,7 +298,6 @@ vnet_net_idetach(const void *unused __unused)
void
if_grow(void)
{
- INIT_VNET_NET(curvnet);
u_int n;
struct ifindex_entry *e;
@@ -354,7 +331,6 @@ if_check(void *dummy __unused)
struct ifnet *
if_alloc(u_char type)
{
- INIT_VNET_NET(curvnet);
struct ifnet *ifp;
ifp = malloc(sizeof(struct ifnet), M_IFNET, M_WAITOK|M_ZERO);
@@ -418,7 +394,6 @@ if_alloc(u_char type)
static void
if_free_internal(struct ifnet *ifp)
{
- INIT_VNET_NET(curvnet); /* ifp->if_vnet is already NULL here */
KASSERT((ifp->if_flags & IFF_DYING),
("if_free_internal: interface not dying"));
@@ -544,7 +519,6 @@ if_attach(struct ifnet *ifp)
static void
if_attach_internal(struct ifnet *ifp, int vmove)
{
- INIT_VNET_NET(curvnet);
unsigned socksize, ifasize;
int namelen, masklen;
struct sockaddr_dl *sdl;
@@ -649,7 +623,6 @@ if_attach_internal(struct ifnet *ifp, int vmove)
static void
if_attachdomain(void *dummy)
{
- INIT_VNET_NET(curvnet);
struct ifnet *ifp;
int s;
@@ -769,7 +742,6 @@ if_detach(struct ifnet *ifp)
static void
if_detach_internal(struct ifnet *ifp, int vmove)
{
- INIT_VNET_NET(ifp->if_vnet);
struct ifaddr *ifa;
struct radix_node_head *rnh;
int i, j;
@@ -904,25 +876,17 @@ if_vmove(struct ifnet *ifp, struct vnet *new_vnet)
/*
* Unlink the ifnet from ifindex_table[] in current vnet,
* and shrink the if_index for that vnet if possible.
- * do / while construct below is needed to confine the scope
- * of INIT_VNET_NET().
*/
- {
- INIT_VNET_NET(curvnet);
-
- IFNET_WLOCK();
- ifnet_setbyindex(ifp->if_index, NULL);
- while (V_if_index > 0 && \
- ifnet_byindex_locked(V_if_index) == NULL)
- V_if_index--;
- IFNET_WUNLOCK();
- };
+ IFNET_WLOCK();
+ ifnet_setbyindex(ifp->if_index, NULL);
+ while (V_if_index > 0 && ifnet_byindex_locked(V_if_index) == NULL)
+ V_if_index--;
+ IFNET_WUNLOCK();
/*
* Switch to the context of the target vnet.
*/
CURVNET_SET_QUIET(new_vnet);
- INIT_VNET_NET(new_vnet);
/*
* Try to find an empty slot below if_index. If we fail, take
@@ -956,7 +920,6 @@ if_vmove(struct ifnet *ifp, struct vnet *new_vnet)
int
if_addgroup(struct ifnet *ifp, const char *groupname)
{
- INIT_VNET_NET(ifp->if_vnet);
struct ifg_list *ifgl;
struct ifg_group *ifg = NULL;
struct ifg_member *ifgm;
@@ -1026,7 +989,6 @@ if_addgroup(struct ifnet *ifp, const char *groupname)
int
if_delgroup(struct ifnet *ifp, const char *groupname)
{
- INIT_VNET_NET(ifp->if_vnet);
struct ifg_list *ifgl;
struct ifg_member *ifgm;
@@ -1072,7 +1034,6 @@ if_delgroup(struct ifnet *ifp, const char *groupname)
static void
if_delgroups(struct ifnet *ifp)
{
- INIT_VNET_NET(ifp->if_vnet);
struct ifg_list *ifgl;
struct ifg_member *ifgm;
char groupname[IFNAMSIZ];
@@ -1164,7 +1125,6 @@ if_getgroup(struct ifgroupreq *data, struct ifnet *ifp)
static int
if_getgroupmembers(struct ifgroupreq *data)
{
- INIT_VNET_NET(curvnet);
struct ifgroupreq *ifgr = data;
struct ifg_group *ifg;
struct ifg_member *ifgm;
@@ -1336,7 +1296,6 @@ ifa_free(struct ifaddr *ifa)
static struct ifaddr *
ifa_ifwithaddr_internal(struct sockaddr *addr, int getref)
{
- INIT_VNET_NET(curvnet);
struct ifnet *ifp;
struct ifaddr *ifa;
@@ -1392,7 +1351,6 @@ ifa_ifwithaddr_check(struct sockaddr *addr)
struct ifaddr *
ifa_ifwithbroadaddr(struct sockaddr *addr)
{
- INIT_VNET_NET(curvnet);
struct ifnet *ifp;
struct ifaddr *ifa;
@@ -1426,7 +1384,6 @@ done:
struct ifaddr *
ifa_ifwithdstaddr(struct sockaddr *addr)
{
- INIT_VNET_NET(curvnet);
struct ifnet *ifp;
struct ifaddr *ifa;
@@ -1460,7 +1417,6 @@ done:
struct ifaddr *
ifa_ifwithnet(struct sockaddr *addr)
{
- INIT_VNET_NET(curvnet);
struct ifnet *ifp;
struct ifaddr *ifa;
struct ifaddr *ifa_maybe = NULL;
@@ -1841,7 +1797,6 @@ if_slowtimo(void *arg)
VNET_LIST_RLOCK();
VNET_FOREACH(vnet_iter) {
CURVNET_SET(vnet_iter);
- INIT_VNET_NET(vnet_iter);
TAILQ_FOREACH(ifp, &V_ifnet, if_link) {
if (ifp->if_timer == 0 || --ifp->if_timer)
continue;
@@ -1863,7 +1818,6 @@ if_slowtimo(void *arg)
struct ifnet *
ifunit_ref(const char *name)
{
- INIT_VNET_NET(curvnet);
struct ifnet *ifp;
IFNET_RLOCK();
@@ -1881,7 +1835,6 @@ ifunit_ref(const char *name)
struct ifnet *
ifunit(const char *name)
{
- INIT_VNET_NET(curvnet);
struct ifnet *ifp;
IFNET_RLOCK();
@@ -2478,7 +2431,6 @@ ifpromisc(struct ifnet *ifp, int pswitch)
static int
ifconf(u_long cmd, caddr_t data)
{
- INIT_VNET_NET(curvnet);
struct ifconf *ifc = (struct ifconf *)data;
#ifdef __amd64__
struct ifconf32 *ifc32 = (struct ifconf32 *)data;
@@ -2837,7 +2789,6 @@ if_delmulti(struct ifnet *ifp, struct sockaddr *sa)
int lastref;
#ifdef INVARIANTS
struct ifnet *oifp;
- INIT_VNET_NET(ifp->if_vnet);
IFNET_RLOCK();
TAILQ_FOREACH(oifp, &V_ifnet, if_link)
@@ -2878,9 +2829,6 @@ if_delmulti(struct ifnet *ifp, struct sockaddr *sa)
void
if_delmulti_ifma(struct ifmultiaddr *ifma)
{
-#ifdef DIAGNOSTIC
- INIT_VNET_NET(curvnet);
-#endif
struct ifnet *ifp;
int lastref;
OpenPOWER on IntegriCloud