Jerry/bird
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity
bird/nest/iface.c

567 lines
12 KiB
C
Raw Normal View History

Implemented scanning of network interfaces. Mostly very ugly code due to terrible kernel interface (SIOGIFCONF and friends).
1998-05-27 05:42:05 +08:00
/*
- Rewrote whole interface logic. Removed support for multiple addresses per interface since it makes much trouble everywhere. Instead, we understand secondary addresses as subinterfaces. - In case interface addresses or basic flags change, we simply convert it to a down/up sequence. - Implemented the universal neighbour cache. (Just forget what did previous includes say of neighbour caching, this one is brand new.)
1998-06-02 05:41:11 +08:00
* BIRD -- Management of Interfaces and Neighbor Cache
Implemented scanning of network interfaces. Mostly very ugly code due to terrible kernel interface (SIOGIFCONF and friends).
1998-05-27 05:42:05 +08:00
*
Reimplemented neighbor cache. Now uses real hashing.
2000-03-01 22:51:47 +08:00
* (c) 1998--2000 Martin Mares <mj@ucw.cz>
Implemented scanning of network interfaces. Mostly very ugly code due to terrible kernel interface (SIOGIFCONF and friends).
1998-05-27 05:42:05 +08:00
*
* Can be freely distributed and used under the terms of the GNU GPL.
*/
Added tracing of interface events.
2000-03-13 04:50:35 +08:00
#undef LOCAL_DEBUG
- Rewrote whole interface logic. Removed support for multiple addresses per interface since it makes much trouble everywhere. Instead, we understand secondary addresses as subinterfaces. - In case interface addresses or basic flags change, we simply convert it to a down/up sequence. - Implemented the universal neighbour cache. (Just forget what did previous includes say of neighbour caching, this one is brand new.)
1998-06-02 05:41:11 +08:00
Implemented scanning of network interfaces. Mostly very ugly code due to terrible kernel interface (SIOGIFCONF and friends).
1998-05-27 05:42:05 +08:00
#include "nest/bird.h"
#include "nest/iface.h"
- Rewrote whole interface logic. Removed support for multiple addresses per interface since it makes much trouble everywhere. Instead, we understand secondary addresses as subinterfaces. - In case interface addresses or basic flags change, we simply convert it to a down/up sequence. - Implemented the universal neighbour cache. (Just forget what did previous includes say of neighbour caching, this one is brand new.)
1998-06-02 05:41:11 +08:00
#include "nest/protocol.h"
Added few basic commands: show status, show interfaces [summary], show protocols (incomplete).
1999-11-25 23:35:30 +08:00
#include "nest/cli.h"
Implemented scanning of network interfaces. Mostly very ugly code due to terrible kernel interface (SIOGIFCONF and friends).
1998-05-27 05:42:05 +08:00
#include "lib/resource.h"
Added functions for manipulating interface name pattern lists: o iface_patt_match(list, iface) -- match interface against list o iface_patts_equal(a, b, c) -- compare whether two pattern lists are equal or not. c(x,y) is called for comparison of protocol-dependent data.
1998-11-30 06:01:03 +08:00
#include "lib/string.h"
config->router_id works again.
1999-02-14 02:42:00 +08:00
#include "conf/conf.h"
Implemented scanning of network interfaces. Mostly very ugly code due to terrible kernel interface (SIOGIFCONF and friends).
1998-05-27 05:42:05 +08:00
static pool *if_pool;
config->router_id works again.
1999-02-14 02:42:00 +08:00
static void auto_router_id(void);
- Rewrote whole interface logic. Removed support for multiple addresses per interface since it makes much trouble everywhere. Instead, we understand secondary addresses as subinterfaces. - In case interface addresses or basic flags change, we simply convert it to a down/up sequence. - Implemented the universal neighbour cache. (Just forget what did previous includes say of neighbour caching, this one is brand new.)
1998-06-02 05:41:11 +08:00
list iface_list;
I rewrote the interface handling code, so that it supports multiple addresses per interface (needed for example for IPv6 support). Visible changes: o struct iface now contains a list of all interface addresses (represented by struct ifa), iface->addr points to the primary address (if any). o Interface has IF_UP set iff it's up and it has a primary address. o IF_UP is now independent on IF_IGNORED (i.e., you need to test IF_IGNORED in the protocols; I've added this, but please check). o The if_notify_change hook has been simplified (only one interface pointer etc.). o Introduced a ifa_notify_change hook. (For now, only the Direct protocol does use it -- it's wise to just listen to device routes in all other protocols.) o Removed IF_CHANGE_FLAGS notifier flag (it was meaningless anyway). o Updated all the code except netlink (I'll look at it tomorrow) to match the new semantics (please look at your code to ensure I did it right). Things to fix: o Netlink. o Make krt-iface interpret "eth0:1"-type aliases as secondary addresses.
1999-05-07 05:38:11 +08:00
void
ifa_dump(struct ifa *a)
{
Rewrote interface type detection logic. The `unnumbered' flag is now per address, not per interface (hence it's ifa->flags & IA_UNNUMBERED) and should be set reliably. IF_MULTIACCESS should be fixed now, but it isn't wise to rely on it on interfaces configured with /30 prefix.
2000-03-01 07:19:52 +08:00
debug("\t%I, net %I/%-2d bc %I -> %I%s%s%s\n", a->ip, a->prefix, a->pxlen, a->brd, a->opposite,
I rewrote the interface handling code, so that it supports multiple addresses per interface (needed for example for IPv6 support). Visible changes: o struct iface now contains a list of all interface addresses (represented by struct ifa), iface->addr points to the primary address (if any). o Interface has IF_UP set iff it's up and it has a primary address. o IF_UP is now independent on IF_IGNORED (i.e., you need to test IF_IGNORED in the protocols; I've added this, but please check). o The if_notify_change hook has been simplified (only one interface pointer etc.). o Introduced a ifa_notify_change hook. (For now, only the Direct protocol does use it -- it's wise to just listen to device routes in all other protocols.) o Removed IF_CHANGE_FLAGS notifier flag (it was meaningless anyway). o Updated all the code except netlink (I'll look at it tomorrow) to match the new semantics (please look at your code to ensure I did it right). Things to fix: o Netlink. o Make krt-iface interpret "eth0:1"-type aliases as secondary addresses.
1999-05-07 05:38:11 +08:00
(a->flags & IF_UP) ? "" : " DOWN",
Rewrote interface type detection logic. The `unnumbered' flag is now per address, not per interface (hence it's ifa->flags & IA_UNNUMBERED) and should be set reliably. IF_MULTIACCESS should be fixed now, but it isn't wise to rely on it on interfaces configured with /30 prefix.
2000-03-01 07:19:52 +08:00
(a->flags & IA_PRIMARY) ? "" : " SEC",
(a->flags & IA_UNNUMBERED) ? " UNNUM" : "");
I rewrote the interface handling code, so that it supports multiple addresses per interface (needed for example for IPv6 support). Visible changes: o struct iface now contains a list of all interface addresses (represented by struct ifa), iface->addr points to the primary address (if any). o Interface has IF_UP set iff it's up and it has a primary address. o IF_UP is now independent on IF_IGNORED (i.e., you need to test IF_IGNORED in the protocols; I've added this, but please check). o The if_notify_change hook has been simplified (only one interface pointer etc.). o Introduced a ifa_notify_change hook. (For now, only the Direct protocol does use it -- it's wise to just listen to device routes in all other protocols.) o Removed IF_CHANGE_FLAGS notifier flag (it was meaningless anyway). o Updated all the code except netlink (I'll look at it tomorrow) to match the new semantics (please look at your code to ensure I did it right). Things to fix: o Netlink. o Make krt-iface interpret "eth0:1"-type aliases as secondary addresses.
1999-05-07 05:38:11 +08:00
}
- Rewrote whole interface logic. Removed support for multiple addresses per interface since it makes much trouble everywhere. Instead, we understand secondary addresses as subinterfaces. - In case interface addresses or basic flags change, we simply convert it to a down/up sequence. - Implemented the universal neighbour cache. (Just forget what did previous includes say of neighbour caching, this one is brand new.)
1998-06-02 05:41:11 +08:00
void
if_dump(struct iface *i)
{
I rewrote the interface handling code, so that it supports multiple addresses per interface (needed for example for IPv6 support). Visible changes: o struct iface now contains a list of all interface addresses (represented by struct ifa), iface->addr points to the primary address (if any). o Interface has IF_UP set iff it's up and it has a primary address. o IF_UP is now independent on IF_IGNORED (i.e., you need to test IF_IGNORED in the protocols; I've added this, but please check). o The if_notify_change hook has been simplified (only one interface pointer etc.). o Introduced a ifa_notify_change hook. (For now, only the Direct protocol does use it -- it's wise to just listen to device routes in all other protocols.) o Removed IF_CHANGE_FLAGS notifier flag (it was meaningless anyway). o Updated all the code except netlink (I'll look at it tomorrow) to match the new semantics (please look at your code to ensure I did it right). Things to fix: o Netlink. o Make krt-iface interpret "eth0:1"-type aliases as secondary addresses.
1999-05-07 05:38:11 +08:00
struct ifa *a;
Implemented scanning of network interfaces. Mostly very ugly code due to terrible kernel interface (SIOGIFCONF and friends).
1998-05-27 05:42:05 +08:00
debug("IF%d: %s", i->index, i->name);
if (i->flags & IF_ADMIN_DOWN)
debug(" ADMIN-DOWN");
if (i->flags & IF_UP)
debug(" UP");
Interface logic changes: o Introduced IF_LINK_UP flag corresponding to real link state. o Allowed addressless interfaces. o IF_UP is now automatically calculated and set iff the interface is administratively up, has link up and has an IP address assigned. It may be IF_IGNORED, though (as in case of the loopback). o Any changes which include up/down transition are considered small enough to not provoke artificial upping and downing of the interface. o When an interface disappears (i.e., it wasn't seen in the last scan), we announce this change only once. o IF_LOOPBACK implies IF_IGNORE.
1999-03-03 01:28:06 +08:00
else
debug(" DOWN");
if (i->flags & IF_LINK_UP)
debug(" LINK-UP");
Implemented scanning of network interfaces. Mostly very ugly code due to terrible kernel interface (SIOGIFCONF and friends).
1998-05-27 05:42:05 +08:00
if (i->flags & IF_MULTIACCESS)
debug(" MA");
if (i->flags & IF_BROADCAST)
debug(" BC");
if (i->flags & IF_MULTICAST)
debug(" MC");
if (i->flags & IF_LOOPBACK)
debug(" LOOP");
if (i->flags & IF_IGNORE)
debug(" IGN");
I rewrote the interface handling code, so that it supports multiple addresses per interface (needed for example for IPv6 support). Visible changes: o struct iface now contains a list of all interface addresses (represented by struct ifa), iface->addr points to the primary address (if any). o Interface has IF_UP set iff it's up and it has a primary address. o IF_UP is now independent on IF_IGNORED (i.e., you need to test IF_IGNORED in the protocols; I've added this, but please check). o The if_notify_change hook has been simplified (only one interface pointer etc.). o Introduced a ifa_notify_change hook. (For now, only the Direct protocol does use it -- it's wise to just listen to device routes in all other protocols.) o Removed IF_CHANGE_FLAGS notifier flag (it was meaningless anyway). o Updated all the code except netlink (I'll look at it tomorrow) to match the new semantics (please look at your code to ensure I did it right). Things to fix: o Netlink. o Make krt-iface interpret "eth0:1"-type aliases as secondary addresses.
1999-05-07 05:38:11 +08:00
if (i->flags & IF_TMP_DOWN)
debug(" TDOWN");
Implemented scanning of network interfaces. Mostly very ugly code due to terrible kernel interface (SIOGIFCONF and friends).
1998-05-27 05:42:05 +08:00
debug(" MTU=%d\n", i->mtu);
I rewrote the interface handling code, so that it supports multiple addresses per interface (needed for example for IPv6 support). Visible changes: o struct iface now contains a list of all interface addresses (represented by struct ifa), iface->addr points to the primary address (if any). o Interface has IF_UP set iff it's up and it has a primary address. o IF_UP is now independent on IF_IGNORED (i.e., you need to test IF_IGNORED in the protocols; I've added this, but please check). o The if_notify_change hook has been simplified (only one interface pointer etc.). o Introduced a ifa_notify_change hook. (For now, only the Direct protocol does use it -- it's wise to just listen to device routes in all other protocols.) o Removed IF_CHANGE_FLAGS notifier flag (it was meaningless anyway). o Updated all the code except netlink (I'll look at it tomorrow) to match the new semantics (please look at your code to ensure I did it right). Things to fix: o Netlink. o Make krt-iface interpret "eth0:1"-type aliases as secondary addresses.
1999-05-07 05:38:11 +08:00
WALK_LIST(a, i->addrs)
{
ifa_dump(a);
ASSERT((a != i->addr) == !(a->flags & IA_PRIMARY));
}
Implemented scanning of network interfaces. Mostly very ugly code due to terrible kernel interface (SIOGIFCONF and friends).
1998-05-27 05:42:05 +08:00
}
void
if_dump_all(void)
{
struct iface *i;
Now sending IF_CHANGE_CREATE when a new interface appears and IF_CHANGE_UP only if it's really up.
1998-06-05 04:30:11 +08:00
debug("Known network interfaces:\n");
Implemented scanning of network interfaces. Mostly very ugly code due to terrible kernel interface (SIOGIFCONF and friends).
1998-05-27 05:42:05 +08:00
WALK_LIST(i, iface_list)
if_dump(i);
config->router_id works again.
1999-02-14 02:42:00 +08:00
debug("Router ID: %08x\n", config->router_id);
Implemented scanning of network interfaces. Mostly very ugly code due to terrible kernel interface (SIOGIFCONF and friends).
1998-05-27 05:42:05 +08:00
}
I rewrote the interface handling code, so that it supports multiple addresses per interface (needed for example for IPv6 support). Visible changes: o struct iface now contains a list of all interface addresses (represented by struct ifa), iface->addr points to the primary address (if any). o Interface has IF_UP set iff it's up and it has a primary address. o IF_UP is now independent on IF_IGNORED (i.e., you need to test IF_IGNORED in the protocols; I've added this, but please check). o The if_notify_change hook has been simplified (only one interface pointer etc.). o Introduced a ifa_notify_change hook. (For now, only the Direct protocol does use it -- it's wise to just listen to device routes in all other protocols.) o Removed IF_CHANGE_FLAGS notifier flag (it was meaningless anyway). o Updated all the code except netlink (I'll look at it tomorrow) to match the new semantics (please look at your code to ensure I did it right). Things to fix: o Netlink. o Make krt-iface interpret "eth0:1"-type aliases as secondary addresses.
1999-05-07 05:38:11 +08:00
static inline unsigned
if_what_changed(struct iface *i, struct iface *j)
Implemented scanning of network interfaces. Mostly very ugly code due to terrible kernel interface (SIOGIFCONF and friends).
1998-05-27 05:42:05 +08:00
{
I rewrote the interface handling code, so that it supports multiple addresses per interface (needed for example for IPv6 support). Visible changes: o struct iface now contains a list of all interface addresses (represented by struct ifa), iface->addr points to the primary address (if any). o Interface has IF_UP set iff it's up and it has a primary address. o IF_UP is now independent on IF_IGNORED (i.e., you need to test IF_IGNORED in the protocols; I've added this, but please check). o The if_notify_change hook has been simplified (only one interface pointer etc.). o Introduced a ifa_notify_change hook. (For now, only the Direct protocol does use it -- it's wise to just listen to device routes in all other protocols.) o Removed IF_CHANGE_FLAGS notifier flag (it was meaningless anyway). o Updated all the code except netlink (I'll look at it tomorrow) to match the new semantics (please look at your code to ensure I did it right). Things to fix: o Netlink. o Make krt-iface interpret "eth0:1"-type aliases as secondary addresses.
1999-05-07 05:38:11 +08:00
unsigned c;
if (((i->flags ^ j->flags) & ~(IF_UP | IF_ADMIN_DOWN | IF_UPDATED | IF_LINK_UP | IF_TMP_DOWN | IF_JUST_CREATED))
|| i->index != j->index)
return IF_CHANGE_TOO_MUCH;
c = 0;
if ((i->flags ^ j->flags) & IF_UP)
c |= (i->flags & IF_UP) ? IF_CHANGE_DOWN : IF_CHANGE_UP;
if (i->mtu != j->mtu)
c |= IF_CHANGE_MTU;
return c;
Implemented scanning of network interfaces. Mostly very ugly code due to terrible kernel interface (SIOGIFCONF and friends).
1998-05-27 05:42:05 +08:00
}
static inline void
- Rewrote whole interface logic. Removed support for multiple addresses per interface since it makes much trouble everywhere. Instead, we understand secondary addresses as subinterfaces. - In case interface addresses or basic flags change, we simply convert it to a down/up sequence. - Implemented the universal neighbour cache. (Just forget what did previous includes say of neighbour caching, this one is brand new.)
1998-06-02 05:41:11 +08:00
if_copy(struct iface *to, struct iface *from)
Implemented scanning of network interfaces. Mostly very ugly code due to terrible kernel interface (SIOGIFCONF and friends).
1998-05-27 05:42:05 +08:00
{
I rewrote the interface handling code, so that it supports multiple addresses per interface (needed for example for IPv6 support). Visible changes: o struct iface now contains a list of all interface addresses (represented by struct ifa), iface->addr points to the primary address (if any). o Interface has IF_UP set iff it's up and it has a primary address. o IF_UP is now independent on IF_IGNORED (i.e., you need to test IF_IGNORED in the protocols; I've added this, but please check). o The if_notify_change hook has been simplified (only one interface pointer etc.). o Introduced a ifa_notify_change hook. (For now, only the Direct protocol does use it -- it's wise to just listen to device routes in all other protocols.) o Removed IF_CHANGE_FLAGS notifier flag (it was meaningless anyway). o Updated all the code except netlink (I'll look at it tomorrow) to match the new semantics (please look at your code to ensure I did it right). Things to fix: o Netlink. o Make krt-iface interpret "eth0:1"-type aliases as secondary addresses.
1999-05-07 05:38:11 +08:00
to->flags = from->flags | (to->flags & IF_TMP_DOWN);
- Rewrote whole interface logic. Removed support for multiple addresses per interface since it makes much trouble everywhere. Instead, we understand secondary addresses as subinterfaces. - In case interface addresses or basic flags change, we simply convert it to a down/up sequence. - Implemented the universal neighbour cache. (Just forget what did previous includes say of neighbour caching, this one is brand new.)
1998-06-02 05:41:11 +08:00
to->mtu = from->mtu;
Implemented scanning of network interfaces. Mostly very ugly code due to terrible kernel interface (SIOGIFCONF and friends).
1998-05-27 05:42:05 +08:00
}
Added tracing of interface events.
2000-03-13 04:50:35 +08:00
static inline void
ifa_send_notify(struct proto *p, unsigned c, struct ifa *a)
{
if (p->ifa_notify)
{
if (p->debug & D_IFACES)
log(L_TRACE "%s < %s address %I/%d on interface %s %s",
p->name, (a->flags & IA_PRIMARY) ? "primary" : "secondary",
a->prefix, a->pxlen, a->iface->name,
(c & IF_CHANGE_UP) ? "added" : "removed");
p->ifa_notify(p, c, a);
}
}
I rewrote the interface handling code, so that it supports multiple addresses per interface (needed for example for IPv6 support). Visible changes: o struct iface now contains a list of all interface addresses (represented by struct ifa), iface->addr points to the primary address (if any). o Interface has IF_UP set iff it's up and it has a primary address. o IF_UP is now independent on IF_IGNORED (i.e., you need to test IF_IGNORED in the protocols; I've added this, but please check). o The if_notify_change hook has been simplified (only one interface pointer etc.). o Introduced a ifa_notify_change hook. (For now, only the Direct protocol does use it -- it's wise to just listen to device routes in all other protocols.) o Removed IF_CHANGE_FLAGS notifier flag (it was meaningless anyway). o Updated all the code except netlink (I'll look at it tomorrow) to match the new semantics (please look at your code to ensure I did it right). Things to fix: o Netlink. o Make krt-iface interpret "eth0:1"-type aliases as secondary addresses.
1999-05-07 05:38:11 +08:00
static void
ifa_notify_change(unsigned c, struct ifa *a)
Implemented scanning of network interfaces. Mostly very ugly code due to terrible kernel interface (SIOGIFCONF and friends).
1998-05-27 05:42:05 +08:00
{
I rewrote the interface handling code, so that it supports multiple addresses per interface (needed for example for IPv6 support). Visible changes: o struct iface now contains a list of all interface addresses (represented by struct ifa), iface->addr points to the primary address (if any). o Interface has IF_UP set iff it's up and it has a primary address. o IF_UP is now independent on IF_IGNORED (i.e., you need to test IF_IGNORED in the protocols; I've added this, but please check). o The if_notify_change hook has been simplified (only one interface pointer etc.). o Introduced a ifa_notify_change hook. (For now, only the Direct protocol does use it -- it's wise to just listen to device routes in all other protocols.) o Removed IF_CHANGE_FLAGS notifier flag (it was meaningless anyway). o Updated all the code except netlink (I'll look at it tomorrow) to match the new semantics (please look at your code to ensure I did it right). Things to fix: o Netlink. o Make krt-iface interpret "eth0:1"-type aliases as secondary addresses.
1999-05-07 05:38:11 +08:00
struct proto *p;
Implemented scanning of network interfaces. Mostly very ugly code due to terrible kernel interface (SIOGIFCONF and friends).
1998-05-27 05:42:05 +08:00
Added tracing of interface events.
2000-03-13 04:50:35 +08:00
DBG("IFA change notification (%x) for %s:%I\n", c, a->iface->name, a->ip);
Reworked proto lists -- each proto is now in two lists: the global one (proto_list) and per-type one (original lists). A lot of things simplified. Implemented `disable', `enable' and `restart' CLI commands.
2000-01-17 07:30:06 +08:00
WALK_LIST(p, active_proto_list)
Added tracing of interface events.
2000-03-13 04:50:35 +08:00
ifa_send_notify(p, c, a);
}
static inline void
if_send_notify(struct proto *p, unsigned c, struct iface *i)
{
if (p->if_notify)
{
if (p->debug & D_IFACES)
log(L_TRACE "%s < interface %s %s", p->name, i->name,
(c & IF_CHANGE_UP) ? "goes up" :
(c & IF_CHANGE_DOWN) ? "goes down" :
(c & IF_CHANGE_MTU) ? "changes MTU" :
(c & IF_CHANGE_CREATE) ? "created" :
"sends unknown event");
p->if_notify(p, c, i);
}
Implemented scanning of network interfaces. Mostly very ugly code due to terrible kernel interface (SIOGIFCONF and friends).
1998-05-27 05:42:05 +08:00
}
static void
I rewrote the interface handling code, so that it supports multiple addresses per interface (needed for example for IPv6 support). Visible changes: o struct iface now contains a list of all interface addresses (represented by struct ifa), iface->addr points to the primary address (if any). o Interface has IF_UP set iff it's up and it has a primary address. o IF_UP is now independent on IF_IGNORED (i.e., you need to test IF_IGNORED in the protocols; I've added this, but please check). o The if_notify_change hook has been simplified (only one interface pointer etc.). o Introduced a ifa_notify_change hook. (For now, only the Direct protocol does use it -- it's wise to just listen to device routes in all other protocols.) o Removed IF_CHANGE_FLAGS notifier flag (it was meaningless anyway). o Updated all the code except netlink (I'll look at it tomorrow) to match the new semantics (please look at your code to ensure I did it right). Things to fix: o Netlink. o Make krt-iface interpret "eth0:1"-type aliases as secondary addresses.
1999-05-07 05:38:11 +08:00
if_notify_change(unsigned c, struct iface *i)
Implemented scanning of network interfaces. Mostly very ugly code due to terrible kernel interface (SIOGIFCONF and friends).
1998-05-27 05:42:05 +08:00
{
Protocol hooks. All of them may be NULL.
1998-06-03 16:40:10 +08:00
struct proto *p;
I rewrote the interface handling code, so that it supports multiple addresses per interface (needed for example for IPv6 support). Visible changes: o struct iface now contains a list of all interface addresses (represented by struct ifa), iface->addr points to the primary address (if any). o Interface has IF_UP set iff it's up and it has a primary address. o IF_UP is now independent on IF_IGNORED (i.e., you need to test IF_IGNORED in the protocols; I've added this, but please check). o The if_notify_change hook has been simplified (only one interface pointer etc.). o Introduced a ifa_notify_change hook. (For now, only the Direct protocol does use it -- it's wise to just listen to device routes in all other protocols.) o Removed IF_CHANGE_FLAGS notifier flag (it was meaningless anyway). o Updated all the code except netlink (I'll look at it tomorrow) to match the new semantics (please look at your code to ensure I did it right). Things to fix: o Netlink. o Make krt-iface interpret "eth0:1"-type aliases as secondary addresses.
1999-05-07 05:38:11 +08:00
struct ifa *a;
Protocol hooks. All of them may be NULL.
1998-06-03 16:40:10 +08:00
I rewrote the interface handling code, so that it supports multiple addresses per interface (needed for example for IPv6 support). Visible changes: o struct iface now contains a list of all interface addresses (represented by struct ifa), iface->addr points to the primary address (if any). o Interface has IF_UP set iff it's up and it has a primary address. o IF_UP is now independent on IF_IGNORED (i.e., you need to test IF_IGNORED in the protocols; I've added this, but please check). o The if_notify_change hook has been simplified (only one interface pointer etc.). o Introduced a ifa_notify_change hook. (For now, only the Direct protocol does use it -- it's wise to just listen to device routes in all other protocols.) o Removed IF_CHANGE_FLAGS notifier flag (it was meaningless anyway). o Updated all the code except netlink (I'll look at it tomorrow) to match the new semantics (please look at your code to ensure I did it right). Things to fix: o Netlink. o Make krt-iface interpret "eth0:1"-type aliases as secondary addresses.
1999-05-07 05:38:11 +08:00
if (i->flags & IF_JUST_CREATED)
{
i->flags &= ~IF_JUST_CREATED;
c |= IF_CHANGE_CREATE | IF_CHANGE_MTU;
}
Added tracing of interface events.
2000-03-13 04:50:35 +08:00
DBG("Interface change notification (%x) for %s\n", c, i->name);
I rewrote the interface handling code, so that it supports multiple addresses per interface (needed for example for IPv6 support). Visible changes: o struct iface now contains a list of all interface addresses (represented by struct ifa), iface->addr points to the primary address (if any). o Interface has IF_UP set iff it's up and it has a primary address. o IF_UP is now independent on IF_IGNORED (i.e., you need to test IF_IGNORED in the protocols; I've added this, but please check). o The if_notify_change hook has been simplified (only one interface pointer etc.). o Introduced a ifa_notify_change hook. (For now, only the Direct protocol does use it -- it's wise to just listen to device routes in all other protocols.) o Removed IF_CHANGE_FLAGS notifier flag (it was meaningless anyway). o Updated all the code except netlink (I'll look at it tomorrow) to match the new semantics (please look at your code to ensure I did it right). Things to fix: o Netlink. o Make krt-iface interpret "eth0:1"-type aliases as secondary addresses.
1999-05-07 05:38:11 +08:00
if_dump(i);
- Rewrote whole interface logic. Removed support for multiple addresses per interface since it makes much trouble everywhere. Instead, we understand secondary addresses as subinterfaces. - In case interface addresses or basic flags change, we simply convert it to a down/up sequence. - Implemented the universal neighbour cache. (Just forget what did previous includes say of neighbour caching, this one is brand new.)
1998-06-02 05:41:11 +08:00
if (c & IF_CHANGE_UP)
I rewrote the interface handling code, so that it supports multiple addresses per interface (needed for example for IPv6 support). Visible changes: o struct iface now contains a list of all interface addresses (represented by struct ifa), iface->addr points to the primary address (if any). o Interface has IF_UP set iff it's up and it has a primary address. o IF_UP is now independent on IF_IGNORED (i.e., you need to test IF_IGNORED in the protocols; I've added this, but please check). o The if_notify_change hook has been simplified (only one interface pointer etc.). o Introduced a ifa_notify_change hook. (For now, only the Direct protocol does use it -- it's wise to just listen to device routes in all other protocols.) o Removed IF_CHANGE_FLAGS notifier flag (it was meaningless anyway). o Updated all the code except netlink (I'll look at it tomorrow) to match the new semantics (please look at your code to ensure I did it right). Things to fix: o Netlink. o Make krt-iface interpret "eth0:1"-type aliases as secondary addresses.
1999-05-07 05:38:11 +08:00
neigh_if_up(i);
if (c & IF_CHANGE_DOWN)
WALK_LIST(a, i->addrs)
{
a->flags = (i->flags & ~IA_FLAGS) | (a->flags & IA_FLAGS);
ifa_notify_change(IF_CHANGE_DOWN, a);
}
- Rewrote whole interface logic. Removed support for multiple addresses per interface since it makes much trouble everywhere. Instead, we understand secondary addresses as subinterfaces. - In case interface addresses or basic flags change, we simply convert it to a down/up sequence. - Implemented the universal neighbour cache. (Just forget what did previous includes say of neighbour caching, this one is brand new.)
1998-06-02 05:41:11 +08:00
Reworked proto lists -- each proto is now in two lists: the global one (proto_list) and per-type one (original lists). A lot of things simplified. Implemented `disable', `enable' and `restart' CLI commands.
2000-01-17 07:30:06 +08:00
WALK_LIST(p, active_proto_list)
Added tracing of interface events.
2000-03-13 04:50:35 +08:00
if_send_notify(p, c, i);
- Rewrote whole interface logic. Removed support for multiple addresses per interface since it makes much trouble everywhere. Instead, we understand secondary addresses as subinterfaces. - In case interface addresses or basic flags change, we simply convert it to a down/up sequence. - Implemented the universal neighbour cache. (Just forget what did previous includes say of neighbour caching, this one is brand new.)
1998-06-02 05:41:11 +08:00
I rewrote the interface handling code, so that it supports multiple addresses per interface (needed for example for IPv6 support). Visible changes: o struct iface now contains a list of all interface addresses (represented by struct ifa), iface->addr points to the primary address (if any). o Interface has IF_UP set iff it's up and it has a primary address. o IF_UP is now independent on IF_IGNORED (i.e., you need to test IF_IGNORED in the protocols; I've added this, but please check). o The if_notify_change hook has been simplified (only one interface pointer etc.). o Introduced a ifa_notify_change hook. (For now, only the Direct protocol does use it -- it's wise to just listen to device routes in all other protocols.) o Removed IF_CHANGE_FLAGS notifier flag (it was meaningless anyway). o Updated all the code except netlink (I'll look at it tomorrow) to match the new semantics (please look at your code to ensure I did it right). Things to fix: o Netlink. o Make krt-iface interpret "eth0:1"-type aliases as secondary addresses.
1999-05-07 05:38:11 +08:00
if (c & IF_CHANGE_UP)
WALK_LIST(a, i->addrs)
{
a->flags = (i->flags & ~IA_FLAGS) | (a->flags & IA_FLAGS);
ifa_notify_change(IF_CHANGE_UP, a);
}
- Rewrote whole interface logic. Removed support for multiple addresses per interface since it makes much trouble everywhere. Instead, we understand secondary addresses as subinterfaces. - In case interface addresses or basic flags change, we simply convert it to a down/up sequence. - Implemented the universal neighbour cache. (Just forget what did previous includes say of neighbour caching, this one is brand new.)
1998-06-02 05:41:11 +08:00
if (c & IF_CHANGE_DOWN)
I rewrote the interface handling code, so that it supports multiple addresses per interface (needed for example for IPv6 support). Visible changes: o struct iface now contains a list of all interface addresses (represented by struct ifa), iface->addr points to the primary address (if any). o Interface has IF_UP set iff it's up and it has a primary address. o IF_UP is now independent on IF_IGNORED (i.e., you need to test IF_IGNORED in the protocols; I've added this, but please check). o The if_notify_change hook has been simplified (only one interface pointer etc.). o Introduced a ifa_notify_change hook. (For now, only the Direct protocol does use it -- it's wise to just listen to device routes in all other protocols.) o Removed IF_CHANGE_FLAGS notifier flag (it was meaningless anyway). o Updated all the code except netlink (I'll look at it tomorrow) to match the new semantics (please look at your code to ensure I did it right). Things to fix: o Netlink. o Make krt-iface interpret "eth0:1"-type aliases as secondary addresses.
1999-05-07 05:38:11 +08:00
neigh_if_down(i);
Implemented scanning of network interfaces. Mostly very ugly code due to terrible kernel interface (SIOGIFCONF and friends).
1998-05-27 05:42:05 +08:00
}
I rewrote the interface handling code, so that it supports multiple addresses per interface (needed for example for IPv6 support). Visible changes: o struct iface now contains a list of all interface addresses (represented by struct ifa), iface->addr points to the primary address (if any). o Interface has IF_UP set iff it's up and it has a primary address. o IF_UP is now independent on IF_IGNORED (i.e., you need to test IF_IGNORED in the protocols; I've added this, but please check). o The if_notify_change hook has been simplified (only one interface pointer etc.). o Introduced a ifa_notify_change hook. (For now, only the Direct protocol does use it -- it's wise to just listen to device routes in all other protocols.) o Removed IF_CHANGE_FLAGS notifier flag (it was meaningless anyway). o Updated all the code except netlink (I'll look at it tomorrow) to match the new semantics (please look at your code to ensure I did it right). Things to fix: o Netlink. o Make krt-iface interpret "eth0:1"-type aliases as secondary addresses.
1999-05-07 05:38:11 +08:00
static unsigned
if_recalc_flags(struct iface *i, unsigned flags)
{
if ((flags & (IF_ADMIN_DOWN | IF_TMP_DOWN)) ||
!(flags & IF_LINK_UP) ||
!i->addr)
flags &= ~IF_UP;
else
flags |= IF_UP;
return flags;
}
static void
if_change_flags(struct iface *i, unsigned flags)
{
unsigned of = i->flags;
i->flags = if_recalc_flags(i, flags);
if ((i->flags ^ of) & IF_UP)
if_notify_change((i->flags & IF_UP) ? IF_CHANGE_UP : IF_CHANGE_DOWN, i);
}
struct iface *
Implemented scanning of network interfaces. Mostly very ugly code due to terrible kernel interface (SIOGIFCONF and friends).
1998-05-27 05:42:05 +08:00
if_update(struct iface *new)
{
struct iface *i;
I rewrote the interface handling code, so that it supports multiple addresses per interface (needed for example for IPv6 support). Visible changes: o struct iface now contains a list of all interface addresses (represented by struct ifa), iface->addr points to the primary address (if any). o Interface has IF_UP set iff it's up and it has a primary address. o IF_UP is now independent on IF_IGNORED (i.e., you need to test IF_IGNORED in the protocols; I've added this, but please check). o The if_notify_change hook has been simplified (only one interface pointer etc.). o Introduced a ifa_notify_change hook. (For now, only the Direct protocol does use it -- it's wise to just listen to device routes in all other protocols.) o Removed IF_CHANGE_FLAGS notifier flag (it was meaningless anyway). o Updated all the code except netlink (I'll look at it tomorrow) to match the new semantics (please look at your code to ensure I did it right). Things to fix: o Netlink. o Make krt-iface interpret "eth0:1"-type aliases as secondary addresses.
1999-05-07 05:38:11 +08:00
struct ifa *a, *b;
- Rewrote whole interface logic. Removed support for multiple addresses per interface since it makes much trouble everywhere. Instead, we understand secondary addresses as subinterfaces. - In case interface addresses or basic flags change, we simply convert it to a down/up sequence. - Implemented the universal neighbour cache. (Just forget what did previous includes say of neighbour caching, this one is brand new.)
1998-06-02 05:41:11 +08:00
unsigned c;
Implemented scanning of network interfaces. Mostly very ugly code due to terrible kernel interface (SIOGIFCONF and friends).
1998-05-27 05:42:05 +08:00
WALK_LIST(i, iface_list)
if (!strcmp(new->name, i->name))
{
I rewrote the interface handling code, so that it supports multiple addresses per interface (needed for example for IPv6 support). Visible changes: o struct iface now contains a list of all interface addresses (represented by struct ifa), iface->addr points to the primary address (if any). o Interface has IF_UP set iff it's up and it has a primary address. o IF_UP is now independent on IF_IGNORED (i.e., you need to test IF_IGNORED in the protocols; I've added this, but please check). o The if_notify_change hook has been simplified (only one interface pointer etc.). o Introduced a ifa_notify_change hook. (For now, only the Direct protocol does use it -- it's wise to just listen to device routes in all other protocols.) o Removed IF_CHANGE_FLAGS notifier flag (it was meaningless anyway). o Updated all the code except netlink (I'll look at it tomorrow) to match the new semantics (please look at your code to ensure I did it right). Things to fix: o Netlink. o Make krt-iface interpret "eth0:1"-type aliases as secondary addresses.
1999-05-07 05:38:11 +08:00
new->addr = i->addr;
new->flags = if_recalc_flags(new, new->flags);
c = if_what_changed(i, new);
if (c & IF_CHANGE_TOO_MUCH) /* Changed a lot, convert it to down/up */
Implemented scanning of network interfaces. Mostly very ugly code due to terrible kernel interface (SIOGIFCONF and friends).
1998-05-27 05:42:05 +08:00
{
- Rewrote whole interface logic. Removed support for multiple addresses per interface since it makes much trouble everywhere. Instead, we understand secondary addresses as subinterfaces. - In case interface addresses or basic flags change, we simply convert it to a down/up sequence. - Implemented the universal neighbour cache. (Just forget what did previous includes say of neighbour caching, this one is brand new.)
1998-06-02 05:41:11 +08:00
DBG("Interface %s changed too much -- forcing down/up transition\n", i->name);
I rewrote the interface handling code, so that it supports multiple addresses per interface (needed for example for IPv6 support). Visible changes: o struct iface now contains a list of all interface addresses (represented by struct ifa), iface->addr points to the primary address (if any). o Interface has IF_UP set iff it's up and it has a primary address. o IF_UP is now independent on IF_IGNORED (i.e., you need to test IF_IGNORED in the protocols; I've added this, but please check). o The if_notify_change hook has been simplified (only one interface pointer etc.). o Introduced a ifa_notify_change hook. (For now, only the Direct protocol does use it -- it's wise to just listen to device routes in all other protocols.) o Removed IF_CHANGE_FLAGS notifier flag (it was meaningless anyway). o Updated all the code except netlink (I'll look at it tomorrow) to match the new semantics (please look at your code to ensure I did it right). Things to fix: o Netlink. o Make krt-iface interpret "eth0:1"-type aliases as secondary addresses.
1999-05-07 05:38:11 +08:00
if_change_flags(i, i->flags | IF_TMP_DOWN);
Implemented scanning of network interfaces. Mostly very ugly code due to terrible kernel interface (SIOGIFCONF and friends).
1998-05-27 05:42:05 +08:00
rem_node(&i->n);
Netlink support for secondary interface addresses.
1999-05-07 21:46:16 +08:00
new->addr = i->addr;
memcpy(&new->addrs, &i->addrs, sizeof(i->addrs));
memcpy(i, new, sizeof(*i));
- Rewrote whole interface logic. Removed support for multiple addresses per interface since it makes much trouble everywhere. Instead, we understand secondary addresses as subinterfaces. - In case interface addresses or basic flags change, we simply convert it to a down/up sequence. - Implemented the universal neighbour cache. (Just forget what did previous includes say of neighbour caching, this one is brand new.)
1998-06-02 05:41:11 +08:00
goto newif;
Implemented scanning of network interfaces. Mostly very ugly code due to terrible kernel interface (SIOGIFCONF and friends).
1998-05-27 05:42:05 +08:00
}
I rewrote the interface handling code, so that it supports multiple addresses per interface (needed for example for IPv6 support). Visible changes: o struct iface now contains a list of all interface addresses (represented by struct ifa), iface->addr points to the primary address (if any). o Interface has IF_UP set iff it's up and it has a primary address. o IF_UP is now independent on IF_IGNORED (i.e., you need to test IF_IGNORED in the protocols; I've added this, but please check). o The if_notify_change hook has been simplified (only one interface pointer etc.). o Introduced a ifa_notify_change hook. (For now, only the Direct protocol does use it -- it's wise to just listen to device routes in all other protocols.) o Removed IF_CHANGE_FLAGS notifier flag (it was meaningless anyway). o Updated all the code except netlink (I'll look at it tomorrow) to match the new semantics (please look at your code to ensure I did it right). Things to fix: o Netlink. o Make krt-iface interpret "eth0:1"-type aliases as secondary addresses.
1999-05-07 05:38:11 +08:00
else if (c)
{
if_copy(i, new);
if_notify_change(c, i);
}
o The if_change_too_big_p change was too high-spirited. Fixed. o Introduced if_find_by_index() o Recognizing two types of interface updates: full update (starting with if_start_update(), ending with if_end_update(), guaranteed to see all existing interfaces) and a partial update (only if_update(), usually due to asynchronous interface notifications).
1999-03-03 02:36:09 +08:00
i->flags |= IF_UPDATED;
I rewrote the interface handling code, so that it supports multiple addresses per interface (needed for example for IPv6 support). Visible changes: o struct iface now contains a list of all interface addresses (represented by struct ifa), iface->addr points to the primary address (if any). o Interface has IF_UP set iff it's up and it has a primary address. o IF_UP is now independent on IF_IGNORED (i.e., you need to test IF_IGNORED in the protocols; I've added this, but please check). o The if_notify_change hook has been simplified (only one interface pointer etc.). o Introduced a ifa_notify_change hook. (For now, only the Direct protocol does use it -- it's wise to just listen to device routes in all other protocols.) o Removed IF_CHANGE_FLAGS notifier flag (it was meaningless anyway). o Updated all the code except netlink (I'll look at it tomorrow) to match the new semantics (please look at your code to ensure I did it right). Things to fix: o Netlink. o Make krt-iface interpret "eth0:1"-type aliases as secondary addresses.
1999-05-07 05:38:11 +08:00
return i;
Implemented scanning of network interfaces. Mostly very ugly code due to terrible kernel interface (SIOGIFCONF and friends).
1998-05-27 05:42:05 +08:00
}
- Rewrote whole interface logic. Removed support for multiple addresses per interface since it makes much trouble everywhere. Instead, we understand secondary addresses as subinterfaces. - In case interface addresses or basic flags change, we simply convert it to a down/up sequence. - Implemented the universal neighbour cache. (Just forget what did previous includes say of neighbour caching, this one is brand new.)
1998-06-02 05:41:11 +08:00
i = mb_alloc(if_pool, sizeof(struct iface));
memcpy(i, new, sizeof(*i));
I rewrote the interface handling code, so that it supports multiple addresses per interface (needed for example for IPv6 support). Visible changes: o struct iface now contains a list of all interface addresses (represented by struct ifa), iface->addr points to the primary address (if any). o Interface has IF_UP set iff it's up and it has a primary address. o IF_UP is now independent on IF_IGNORED (i.e., you need to test IF_IGNORED in the protocols; I've added this, but please check). o The if_notify_change hook has been simplified (only one interface pointer etc.). o Introduced a ifa_notify_change hook. (For now, only the Direct protocol does use it -- it's wise to just listen to device routes in all other protocols.) o Removed IF_CHANGE_FLAGS notifier flag (it was meaningless anyway). o Updated all the code except netlink (I'll look at it tomorrow) to match the new semantics (please look at your code to ensure I did it right). Things to fix: o Netlink. o Make krt-iface interpret "eth0:1"-type aliases as secondary addresses.
1999-05-07 05:38:11 +08:00
init_list(&i->addrs);
Netlink support for secondary interface addresses.
1999-05-07 21:46:16 +08:00
newif:
Fix stupid bug in neighbor cache.
2000-04-10 18:40:00 +08:00
init_list(&i->neighbors);
I rewrote the interface handling code, so that it supports multiple addresses per interface (needed for example for IPv6 support). Visible changes: o struct iface now contains a list of all interface addresses (represented by struct ifa), iface->addr points to the primary address (if any). o Interface has IF_UP set iff it's up and it has a primary address. o IF_UP is now independent on IF_IGNORED (i.e., you need to test IF_IGNORED in the protocols; I've added this, but please check). o The if_notify_change hook has been simplified (only one interface pointer etc.). o Introduced a ifa_notify_change hook. (For now, only the Direct protocol does use it -- it's wise to just listen to device routes in all other protocols.) o Removed IF_CHANGE_FLAGS notifier flag (it was meaningless anyway). o Updated all the code except netlink (I'll look at it tomorrow) to match the new semantics (please look at your code to ensure I did it right). Things to fix: o Netlink. o Make krt-iface interpret "eth0:1"-type aliases as secondary addresses.
1999-05-07 05:38:11 +08:00
i->flags |= IF_UPDATED | IF_TMP_DOWN; /* Tmp down as we don't have addresses yet */
Implemented scanning of network interfaces. Mostly very ugly code due to terrible kernel interface (SIOGIFCONF and friends).
1998-05-27 05:42:05 +08:00
add_tail(&iface_list, &i->n);
I rewrote the interface handling code, so that it supports multiple addresses per interface (needed for example for IPv6 support). Visible changes: o struct iface now contains a list of all interface addresses (represented by struct ifa), iface->addr points to the primary address (if any). o Interface has IF_UP set iff it's up and it has a primary address. o IF_UP is now independent on IF_IGNORED (i.e., you need to test IF_IGNORED in the protocols; I've added this, but please check). o The if_notify_change hook has been simplified (only one interface pointer etc.). o Introduced a ifa_notify_change hook. (For now, only the Direct protocol does use it -- it's wise to just listen to device routes in all other protocols.) o Removed IF_CHANGE_FLAGS notifier flag (it was meaningless anyway). o Updated all the code except netlink (I'll look at it tomorrow) to match the new semantics (please look at your code to ensure I did it right). Things to fix: o Netlink. o Make krt-iface interpret "eth0:1"-type aliases as secondary addresses.
1999-05-07 05:38:11 +08:00
return i;
Implemented scanning of network interfaces. Mostly very ugly code due to terrible kernel interface (SIOGIFCONF and friends).
1998-05-27 05:42:05 +08:00
}
o The if_change_too_big_p change was too high-spirited. Fixed. o Introduced if_find_by_index() o Recognizing two types of interface updates: full update (starting with if_start_update(), ending with if_end_update(), guaranteed to see all existing interfaces) and a partial update (only if_update(), usually due to asynchronous interface notifications).
1999-03-03 02:36:09 +08:00
void
if_start_update(void)
{
struct iface *i;
I rewrote the interface handling code, so that it supports multiple addresses per interface (needed for example for IPv6 support). Visible changes: o struct iface now contains a list of all interface addresses (represented by struct ifa), iface->addr points to the primary address (if any). o Interface has IF_UP set iff it's up and it has a primary address. o IF_UP is now independent on IF_IGNORED (i.e., you need to test IF_IGNORED in the protocols; I've added this, but please check). o The if_notify_change hook has been simplified (only one interface pointer etc.). o Introduced a ifa_notify_change hook. (For now, only the Direct protocol does use it -- it's wise to just listen to device routes in all other protocols.) o Removed IF_CHANGE_FLAGS notifier flag (it was meaningless anyway). o Updated all the code except netlink (I'll look at it tomorrow) to match the new semantics (please look at your code to ensure I did it right). Things to fix: o Netlink. o Make krt-iface interpret "eth0:1"-type aliases as secondary addresses.
1999-05-07 05:38:11 +08:00
struct ifa *a;
o The if_change_too_big_p change was too high-spirited. Fixed. o Introduced if_find_by_index() o Recognizing two types of interface updates: full update (starting with if_start_update(), ending with if_end_update(), guaranteed to see all existing interfaces) and a partial update (only if_update(), usually due to asynchronous interface notifications).
1999-03-03 02:36:09 +08:00
WALK_LIST(i, iface_list)
I rewrote the interface handling code, so that it supports multiple addresses per interface (needed for example for IPv6 support). Visible changes: o struct iface now contains a list of all interface addresses (represented by struct ifa), iface->addr points to the primary address (if any). o Interface has IF_UP set iff it's up and it has a primary address. o IF_UP is now independent on IF_IGNORED (i.e., you need to test IF_IGNORED in the protocols; I've added this, but please check). o The if_notify_change hook has been simplified (only one interface pointer etc.). o Introduced a ifa_notify_change hook. (For now, only the Direct protocol does use it -- it's wise to just listen to device routes in all other protocols.) o Removed IF_CHANGE_FLAGS notifier flag (it was meaningless anyway). o Updated all the code except netlink (I'll look at it tomorrow) to match the new semantics (please look at your code to ensure I did it right). Things to fix: o Netlink. o Make krt-iface interpret "eth0:1"-type aliases as secondary addresses.
1999-05-07 05:38:11 +08:00
{
i->flags &= ~IF_UPDATED;
WALK_LIST(a, i->addrs)
a->flags &= ~IF_UPDATED;
}
}
void
if_end_partial_update(struct iface *i)
{
if (i->flags & IF_TMP_DOWN)
if_change_flags(i, i->flags & ~IF_TMP_DOWN);
o The if_change_too_big_p change was too high-spirited. Fixed. o Introduced if_find_by_index() o Recognizing two types of interface updates: full update (starting with if_start_update(), ending with if_end_update(), guaranteed to see all existing interfaces) and a partial update (only if_update(), usually due to asynchronous interface notifications).
1999-03-03 02:36:09 +08:00
}
Implemented scanning of network interfaces. Mostly very ugly code due to terrible kernel interface (SIOGIFCONF and friends).
1998-05-27 05:42:05 +08:00
void
if_end_update(void)
{
struct iface *i, j;
I rewrote the interface handling code, so that it supports multiple addresses per interface (needed for example for IPv6 support). Visible changes: o struct iface now contains a list of all interface addresses (represented by struct ifa), iface->addr points to the primary address (if any). o Interface has IF_UP set iff it's up and it has a primary address. o IF_UP is now independent on IF_IGNORED (i.e., you need to test IF_IGNORED in the protocols; I've added this, but please check). o The if_notify_change hook has been simplified (only one interface pointer etc.). o Introduced a ifa_notify_change hook. (For now, only the Direct protocol does use it -- it's wise to just listen to device routes in all other protocols.) o Removed IF_CHANGE_FLAGS notifier flag (it was meaningless anyway). o Updated all the code except netlink (I'll look at it tomorrow) to match the new semantics (please look at your code to ensure I did it right). Things to fix: o Netlink. o Make krt-iface interpret "eth0:1"-type aliases as secondary addresses.
1999-05-07 05:38:11 +08:00
struct ifa *a, *b;
Implemented scanning of network interfaces. Mostly very ugly code due to terrible kernel interface (SIOGIFCONF and friends).
1998-05-27 05:42:05 +08:00
config->router_id works again.
1999-02-14 02:42:00 +08:00
if (!config->router_id)
auto_router_id();
Implemented scanning of network interfaces. Mostly very ugly code due to terrible kernel interface (SIOGIFCONF and friends).
1998-05-27 05:42:05 +08:00
WALK_LIST(i, iface_list)
I rewrote the interface handling code, so that it supports multiple addresses per interface (needed for example for IPv6 support). Visible changes: o struct iface now contains a list of all interface addresses (represented by struct ifa), iface->addr points to the primary address (if any). o Interface has IF_UP set iff it's up and it has a primary address. o IF_UP is now independent on IF_IGNORED (i.e., you need to test IF_IGNORED in the protocols; I've added this, but please check). o The if_notify_change hook has been simplified (only one interface pointer etc.). o Introduced a ifa_notify_change hook. (For now, only the Direct protocol does use it -- it's wise to just listen to device routes in all other protocols.) o Removed IF_CHANGE_FLAGS notifier flag (it was meaningless anyway). o Updated all the code except netlink (I'll look at it tomorrow) to match the new semantics (please look at your code to ensure I did it right). Things to fix: o Netlink. o Make krt-iface interpret "eth0:1"-type aliases as secondary addresses.
1999-05-07 05:38:11 +08:00
{
if (!(i->flags & IF_UPDATED))
if_change_flags(i, (i->flags & ~IF_LINK_UP) | IF_ADMIN_DOWN);
else
{
WALK_LIST_DELSAFE(a, b, i->addrs)
if (!(a->flags & IF_UPDATED))
ifa_delete(a);
if_end_partial_update(i);
}
}
Implemented scanning of network interfaces. Mostly very ugly code due to terrible kernel interface (SIOGIFCONF and friends).
1998-05-27 05:42:05 +08:00
}
Solve chicken-and-egg problems with protocol startup. We now queue all inactive protocols and don't send route/interface updates to them and when they come up, we resend the whole route/interface tables privately. Removed the "scan interface list after protocol start" work-around.
1998-10-17 19:05:18 +08:00
void
if_feed_baby(struct proto *p)
{
struct iface *i;
I rewrote the interface handling code, so that it supports multiple addresses per interface (needed for example for IPv6 support). Visible changes: o struct iface now contains a list of all interface addresses (represented by struct ifa), iface->addr points to the primary address (if any). o Interface has IF_UP set iff it's up and it has a primary address. o IF_UP is now independent on IF_IGNORED (i.e., you need to test IF_IGNORED in the protocols; I've added this, but please check). o The if_notify_change hook has been simplified (only one interface pointer etc.). o Introduced a ifa_notify_change hook. (For now, only the Direct protocol does use it -- it's wise to just listen to device routes in all other protocols.) o Removed IF_CHANGE_FLAGS notifier flag (it was meaningless anyway). o Updated all the code except netlink (I'll look at it tomorrow) to match the new semantics (please look at your code to ensure I did it right). Things to fix: o Netlink. o Make krt-iface interpret "eth0:1"-type aliases as secondary addresses.
1999-05-07 05:38:11 +08:00
struct ifa *a;
Solve chicken-and-egg problems with protocol startup. We now queue all inactive protocols and don't send route/interface updates to them and when they come up, we resend the whole route/interface tables privately. Removed the "scan interface list after protocol start" work-around.
1998-10-17 19:05:18 +08:00
Added tracing of interface events.
2000-03-13 04:50:35 +08:00
if (!p->if_notify && !p->ifa_notify) /* shortcut */
Solve chicken-and-egg problems with protocol startup. We now queue all inactive protocols and don't send route/interface updates to them and when they come up, we resend the whole route/interface tables privately. Removed the "scan interface list after protocol start" work-around.
1998-10-17 19:05:18 +08:00
return;
Added tracing of interface events.
2000-03-13 04:50:35 +08:00
DBG("Announcing interfaces to new protocol %s\n", p->name);
Solve chicken-and-egg problems with protocol startup. We now queue all inactive protocols and don't send route/interface updates to them and when they come up, we resend the whole route/interface tables privately. Removed the "scan interface list after protocol start" work-around.
1998-10-17 19:05:18 +08:00
WALK_LIST(i, iface_list)
I rewrote the interface handling code, so that it supports multiple addresses per interface (needed for example for IPv6 support). Visible changes: o struct iface now contains a list of all interface addresses (represented by struct ifa), iface->addr points to the primary address (if any). o Interface has IF_UP set iff it's up and it has a primary address. o IF_UP is now independent on IF_IGNORED (i.e., you need to test IF_IGNORED in the protocols; I've added this, but please check). o The if_notify_change hook has been simplified (only one interface pointer etc.). o Introduced a ifa_notify_change hook. (For now, only the Direct protocol does use it -- it's wise to just listen to device routes in all other protocols.) o Removed IF_CHANGE_FLAGS notifier flag (it was meaningless anyway). o Updated all the code except netlink (I'll look at it tomorrow) to match the new semantics (please look at your code to ensure I did it right). Things to fix: o Netlink. o Make krt-iface interpret "eth0:1"-type aliases as secondary addresses.
1999-05-07 05:38:11 +08:00
{
Added tracing of interface events.
2000-03-13 04:50:35 +08:00
if_send_notify(p, IF_CHANGE_CREATE | ((i->flags & IF_UP) ? IF_CHANGE_UP : 0), i);
if (i->flags & IF_UP)
I rewrote the interface handling code, so that it supports multiple addresses per interface (needed for example for IPv6 support). Visible changes: o struct iface now contains a list of all interface addresses (represented by struct ifa), iface->addr points to the primary address (if any). o Interface has IF_UP set iff it's up and it has a primary address. o IF_UP is now independent on IF_IGNORED (i.e., you need to test IF_IGNORED in the protocols; I've added this, but please check). o The if_notify_change hook has been simplified (only one interface pointer etc.). o Introduced a ifa_notify_change hook. (For now, only the Direct protocol does use it -- it's wise to just listen to device routes in all other protocols.) o Removed IF_CHANGE_FLAGS notifier flag (it was meaningless anyway). o Updated all the code except netlink (I'll look at it tomorrow) to match the new semantics (please look at your code to ensure I did it right). Things to fix: o Netlink. o Make krt-iface interpret "eth0:1"-type aliases as secondary addresses.
1999-05-07 05:38:11 +08:00
WALK_LIST(a, i->addrs)
Added tracing of interface events.
2000-03-13 04:50:35 +08:00
ifa_send_notify(p, IF_CHANGE_CREATE | IF_CHANGE_UP, a);
I rewrote the interface handling code, so that it supports multiple addresses per interface (needed for example for IPv6 support). Visible changes: o struct iface now contains a list of all interface addresses (represented by struct ifa), iface->addr points to the primary address (if any). o Interface has IF_UP set iff it's up and it has a primary address. o IF_UP is now independent on IF_IGNORED (i.e., you need to test IF_IGNORED in the protocols; I've added this, but please check). o The if_notify_change hook has been simplified (only one interface pointer etc.). o Introduced a ifa_notify_change hook. (For now, only the Direct protocol does use it -- it's wise to just listen to device routes in all other protocols.) o Removed IF_CHANGE_FLAGS notifier flag (it was meaningless anyway). o Updated all the code except netlink (I'll look at it tomorrow) to match the new semantics (please look at your code to ensure I did it right). Things to fix: o Netlink. o Make krt-iface interpret "eth0:1"-type aliases as secondary addresses.
1999-05-07 05:38:11 +08:00
}
Solve chicken-and-egg problems with protocol startup. We now queue all inactive protocols and don't send route/interface updates to them and when they come up, we resend the whole route/interface tables privately. Removed the "scan interface list after protocol start" work-around.
1998-10-17 19:05:18 +08:00
}
o The if_change_too_big_p change was too high-spirited. Fixed. o Introduced if_find_by_index() o Recognizing two types of interface updates: full update (starting with if_start_update(), ending with if_end_update(), guaranteed to see all existing interfaces) and a partial update (only if_update(), usually due to asynchronous interface notifications).
1999-03-03 02:36:09 +08:00
struct iface *
if_find_by_index(unsigned idx)
{
struct iface *i;
WALK_LIST(i, iface_list)
if (i->index == idx)
return i;
return NULL;
}
I rewrote the interface handling code, so that it supports multiple addresses per interface (needed for example for IPv6 support). Visible changes: o struct iface now contains a list of all interface addresses (represented by struct ifa), iface->addr points to the primary address (if any). o Interface has IF_UP set iff it's up and it has a primary address. o IF_UP is now independent on IF_IGNORED (i.e., you need to test IF_IGNORED in the protocols; I've added this, but please check). o The if_notify_change hook has been simplified (only one interface pointer etc.). o Introduced a ifa_notify_change hook. (For now, only the Direct protocol does use it -- it's wise to just listen to device routes in all other protocols.) o Removed IF_CHANGE_FLAGS notifier flag (it was meaningless anyway). o Updated all the code except netlink (I'll look at it tomorrow) to match the new semantics (please look at your code to ensure I did it right). Things to fix: o Netlink. o Make krt-iface interpret "eth0:1"-type aliases as secondary addresses.
1999-05-07 05:38:11 +08:00
struct iface *
if_find_by_name(char *name)
{
struct iface *i;
WALK_LIST(i, iface_list)
if (!strcmp(i->name, name))
return i;
return NULL;
}
static int
ifa_recalc_primary(struct iface *i)
{
struct ifa *a, *b = NULL;
int res;
WALK_LIST(a, i->addrs)
{
if (!(a->flags & IA_SECONDARY) && (!b || a->scope > b->scope))
b = a;
a->flags &= ~IA_PRIMARY;
}
res = (b != i->addr);
i->addr = b;
if (b)
{
b->flags |= IA_PRIMARY;
rem_node(&b->n);
add_head(&i->addrs, &b->n);
}
return res;
}
struct ifa *
ifa_update(struct ifa *a)
{
struct iface *i = a->iface;
struct ifa *b;
WALK_LIST(b, i->addrs)
if (ipa_equal(b->ip, a->ip))
{
if (ipa_equal(b->prefix, a->prefix) &&
b->pxlen == a->pxlen &&
ipa_equal(b->brd, a->brd) &&
ipa_equal(b->opposite, a->opposite) &&
Rewrote interface type detection logic. The `unnumbered' flag is now per address, not per interface (hence it's ifa->flags & IA_UNNUMBERED) and should be set reliably. IF_MULTIACCESS should be fixed now, but it isn't wise to rely on it on interfaces configured with /30 prefix.
2000-03-01 07:19:52 +08:00
b->scope == a->scope &&
!((b->flags ^ a->flags) & IA_UNNUMBERED))
I rewrote the interface handling code, so that it supports multiple addresses per interface (needed for example for IPv6 support). Visible changes: o struct iface now contains a list of all interface addresses (represented by struct ifa), iface->addr points to the primary address (if any). o Interface has IF_UP set iff it's up and it has a primary address. o IF_UP is now independent on IF_IGNORED (i.e., you need to test IF_IGNORED in the protocols; I've added this, but please check). o The if_notify_change hook has been simplified (only one interface pointer etc.). o Introduced a ifa_notify_change hook. (For now, only the Direct protocol does use it -- it's wise to just listen to device routes in all other protocols.) o Removed IF_CHANGE_FLAGS notifier flag (it was meaningless anyway). o Updated all the code except netlink (I'll look at it tomorrow) to match the new semantics (please look at your code to ensure I did it right). Things to fix: o Netlink. o Make krt-iface interpret "eth0:1"-type aliases as secondary addresses.
1999-05-07 05:38:11 +08:00
{
b->flags |= IF_UPDATED;
return b;
}
ifa_delete(b);
break;
}
Rewrote interface type detection logic. The `unnumbered' flag is now per address, not per interface (hence it's ifa->flags & IA_UNNUMBERED) and should be set reliably. IF_MULTIACCESS should be fixed now, but it isn't wise to rely on it on interfaces configured with /30 prefix.
2000-03-01 07:19:52 +08:00
if (!(i->flags & IF_MULTIACCESS) && a->pxlen < BITS_PER_IP_ADDRESS - 2)
log(L_WARN "Strange prefix length %d for point-to-point interface %s", a->pxlen, i->name);
if ((i->flags & IF_BROADCAST) && !ipa_nonzero(a->brd))
log(L_ERR "Missing broadcast address for interface %s", i->name);
I rewrote the interface handling code, so that it supports multiple addresses per interface (needed for example for IPv6 support). Visible changes: o struct iface now contains a list of all interface addresses (represented by struct ifa), iface->addr points to the primary address (if any). o Interface has IF_UP set iff it's up and it has a primary address. o IF_UP is now independent on IF_IGNORED (i.e., you need to test IF_IGNORED in the protocols; I've added this, but please check). o The if_notify_change hook has been simplified (only one interface pointer etc.). o Introduced a ifa_notify_change hook. (For now, only the Direct protocol does use it -- it's wise to just listen to device routes in all other protocols.) o Removed IF_CHANGE_FLAGS notifier flag (it was meaningless anyway). o Updated all the code except netlink (I'll look at it tomorrow) to match the new semantics (please look at your code to ensure I did it right). Things to fix: o Netlink. o Make krt-iface interpret "eth0:1"-type aliases as secondary addresses.
1999-05-07 05:38:11 +08:00
b = mb_alloc(if_pool, sizeof(struct ifa));
memcpy(b, a, sizeof(struct ifa));
add_tail(&i->addrs, &b->n);
b->flags = (i->flags & ~IA_FLAGS) | (a->flags & IA_FLAGS);
if ((!i->addr || i->addr->scope < b->scope) && ifa_recalc_primary(i))
if_change_flags(i, i->flags | IF_TMP_DOWN);
if (b->flags & IF_UP)
ifa_notify_change(IF_CHANGE_CREATE | IF_CHANGE_UP, b);
return b;
}
void
ifa_delete(struct ifa *a)
{
struct iface *i = a->iface;
struct ifa *b;
WALK_LIST(b, i->addrs)
if (ipa_equal(b->ip, a->ip))
{
rem_node(&b->n);
if (b->flags & IF_UP)
{
b->flags &= ~IF_UP;
ifa_notify_change(IF_CHANGE_DOWN, b);
}
if (b->flags & IA_PRIMARY)
{
if_change_flags(i, i->flags | IF_TMP_DOWN);
ifa_recalc_primary(i);
}
mb_free(b);
Netlink support for secondary interface addresses.
1999-05-07 21:46:16 +08:00
return;
I rewrote the interface handling code, so that it supports multiple addresses per interface (needed for example for IPv6 support). Visible changes: o struct iface now contains a list of all interface addresses (represented by struct ifa), iface->addr points to the primary address (if any). o Interface has IF_UP set iff it's up and it has a primary address. o IF_UP is now independent on IF_IGNORED (i.e., you need to test IF_IGNORED in the protocols; I've added this, but please check). o The if_notify_change hook has been simplified (only one interface pointer etc.). o Introduced a ifa_notify_change hook. (For now, only the Direct protocol does use it -- it's wise to just listen to device routes in all other protocols.) o Removed IF_CHANGE_FLAGS notifier flag (it was meaningless anyway). o Updated all the code except netlink (I'll look at it tomorrow) to match the new semantics (please look at your code to ensure I did it right). Things to fix: o Netlink. o Make krt-iface interpret "eth0:1"-type aliases as secondary addresses.
1999-05-07 05:38:11 +08:00
}
}
config->router_id works again.
1999-02-14 02:42:00 +08:00
static void
Basic support for IPv6. The system-dependent part doesn't work yet, but the core routines are there and seem to be working. o lib/ipv6.[ch] written o Lexical analyser recognizes IPv6 addresses and when in IPv6 mode, treats pure IPv4 addresses as router IDs. o Router ID must be configured manually on IPv6 systems. o Added SCOPE_ORGANIZATION for org-scoped IPv6 multicasts. o Fixed few places where ipa_(hton|ntoh) was called as a function returning converted address.
1999-08-04 03:36:06 +08:00
auto_router_id(void)
Generate router_id automatically if possible (standard "smallest of local regular interface addresses" rule). Protocols should NOT rely on router_id existence -- when router ID is not available, the router_id variable is set to zero and protocols requiring valid router ID should just refuse to start, reporting such error to the log.
1998-10-20 02:13:36 +08:00
{
Basic support for IPv6. The system-dependent part doesn't work yet, but the core routines are there and seem to be working. o lib/ipv6.[ch] written o Lexical analyser recognizes IPv6 addresses and when in IPv6 mode, treats pure IPv4 addresses as router IDs. o Router ID must be configured manually on IPv6 systems. o Added SCOPE_ORGANIZATION for org-scoped IPv6 multicasts. o Fixed few places where ipa_(hton|ntoh) was called as a function returning converted address.
1999-08-04 03:36:06 +08:00
#ifndef IPV6
Generate router_id automatically if possible (standard "smallest of local regular interface addresses" rule). Protocols should NOT rely on router_id existence -- when router ID is not available, the router_id variable is set to zero and protocols requiring valid router ID should just refuse to start, reporting such error to the log.
1998-10-20 02:13:36 +08:00
struct iface *i, *j;
j = NULL;
WALK_LIST(i, iface_list)
I rewrote the interface handling code, so that it supports multiple addresses per interface (needed for example for IPv6 support). Visible changes: o struct iface now contains a list of all interface addresses (represented by struct ifa), iface->addr points to the primary address (if any). o Interface has IF_UP set iff it's up and it has a primary address. o IF_UP is now independent on IF_IGNORED (i.e., you need to test IF_IGNORED in the protocols; I've added this, but please check). o The if_notify_change hook has been simplified (only one interface pointer etc.). o Introduced a ifa_notify_change hook. (For now, only the Direct protocol does use it -- it's wise to just listen to device routes in all other protocols.) o Removed IF_CHANGE_FLAGS notifier flag (it was meaningless anyway). o Updated all the code except netlink (I'll look at it tomorrow) to match the new semantics (please look at your code to ensure I did it right). Things to fix: o Netlink. o Make krt-iface interpret "eth0:1"-type aliases as secondary addresses.
1999-05-07 05:38:11 +08:00
if ((i->flags & IF_LINK_UP) &&
Rewrote interface type detection logic. The `unnumbered' flag is now per address, not per interface (hence it's ifa->flags & IA_UNNUMBERED) and should be set reliably. IF_MULTIACCESS should be fixed now, but it isn't wise to rely on it on interfaces configured with /30 prefix.
2000-03-01 07:19:52 +08:00
!(i->flags & (IF_IGNORE | IF_ADMIN_DOWN)) &&
I rewrote the interface handling code, so that it supports multiple addresses per interface (needed for example for IPv6 support). Visible changes: o struct iface now contains a list of all interface addresses (represented by struct ifa), iface->addr points to the primary address (if any). o Interface has IF_UP set iff it's up and it has a primary address. o IF_UP is now independent on IF_IGNORED (i.e., you need to test IF_IGNORED in the protocols; I've added this, but please check). o The if_notify_change hook has been simplified (only one interface pointer etc.). o Introduced a ifa_notify_change hook. (For now, only the Direct protocol does use it -- it's wise to just listen to device routes in all other protocols.) o Removed IF_CHANGE_FLAGS notifier flag (it was meaningless anyway). o Updated all the code except netlink (I'll look at it tomorrow) to match the new semantics (please look at your code to ensure I did it right). Things to fix: o Netlink. o Make krt-iface interpret "eth0:1"-type aliases as secondary addresses.
1999-05-07 05:38:11 +08:00
i->addr &&
Rewrote interface type detection logic. The `unnumbered' flag is now per address, not per interface (hence it's ifa->flags & IA_UNNUMBERED) and should be set reliably. IF_MULTIACCESS should be fixed now, but it isn't wise to rely on it on interfaces configured with /30 prefix.
2000-03-01 07:19:52 +08:00
!(i->addr->flags & IA_UNNUMBERED) &&
I rewrote the interface handling code, so that it supports multiple addresses per interface (needed for example for IPv6 support). Visible changes: o struct iface now contains a list of all interface addresses (represented by struct ifa), iface->addr points to the primary address (if any). o Interface has IF_UP set iff it's up and it has a primary address. o IF_UP is now independent on IF_IGNORED (i.e., you need to test IF_IGNORED in the protocols; I've added this, but please check). o The if_notify_change hook has been simplified (only one interface pointer etc.). o Introduced a ifa_notify_change hook. (For now, only the Direct protocol does use it -- it's wise to just listen to device routes in all other protocols.) o Removed IF_CHANGE_FLAGS notifier flag (it was meaningless anyway). o Updated all the code except netlink (I'll look at it tomorrow) to match the new semantics (please look at your code to ensure I did it right). Things to fix: o Netlink. o Make krt-iface interpret "eth0:1"-type aliases as secondary addresses.
1999-05-07 05:38:11 +08:00
(!j || ipa_to_u32(i->addr->ip) < ipa_to_u32(j->addr->ip)))
Generate router_id automatically if possible (standard "smallest of local regular interface addresses" rule). Protocols should NOT rely on router_id existence -- when router ID is not available, the router_id variable is set to zero and protocols requiring valid router ID should just refuse to start, reporting such error to the log.
1998-10-20 02:13:36 +08:00
j = i;
Implemented new configuration/reconfiguration interface and defined protocol state machines. Full explanation will follow soon.
1999-02-06 05:37:34 +08:00
if (!j)
config->router_id works again.
1999-02-14 02:42:00 +08:00
die("Cannot determine router ID (no suitable network interface found), please configure it manually");
Great cleanup of debug messages. LOCAL_DEBUG turned off in most modules, several debug() calls converted to DBG().
2000-03-13 05:01:38 +08:00
log(L_INFO "Guessed router ID %I according to interface %s", j->addr->ip, j->name);
I rewrote the interface handling code, so that it supports multiple addresses per interface (needed for example for IPv6 support). Visible changes: o struct iface now contains a list of all interface addresses (represented by struct ifa), iface->addr points to the primary address (if any). o Interface has IF_UP set iff it's up and it has a primary address. o IF_UP is now independent on IF_IGNORED (i.e., you need to test IF_IGNORED in the protocols; I've added this, but please check). o The if_notify_change hook has been simplified (only one interface pointer etc.). o Introduced a ifa_notify_change hook. (For now, only the Direct protocol does use it -- it's wise to just listen to device routes in all other protocols.) o Removed IF_CHANGE_FLAGS notifier flag (it was meaningless anyway). o Updated all the code except netlink (I'll look at it tomorrow) to match the new semantics (please look at your code to ensure I did it right). Things to fix: o Netlink. o Make krt-iface interpret "eth0:1"-type aliases as secondary addresses.
1999-05-07 05:38:11 +08:00
config->router_id = ipa_to_u32(j->addr->ip);
Basic support for IPv6. The system-dependent part doesn't work yet, but the core routines are there and seem to be working. o lib/ipv6.[ch] written o Lexical analyser recognizes IPv6 addresses and when in IPv6 mode, treats pure IPv4 addresses as router IDs. o Router ID must be configured manually on IPv6 systems. o Added SCOPE_ORGANIZATION for org-scoped IPv6 multicasts. o Fixed few places where ipa_(hton|ntoh) was called as a function returning converted address.
1999-08-04 03:36:06 +08:00
#endif
Generate router_id automatically if possible (standard "smallest of local regular interface addresses" rule). Protocols should NOT rely on router_id existence -- when router ID is not available, the router_id variable is set to zero and protocols requiring valid router ID should just refuse to start, reporting such error to the log.
1998-10-20 02:13:36 +08:00
}
Implemented scanning of network interfaces. Mostly very ugly code due to terrible kernel interface (SIOGIFCONF and friends).
1998-05-27 05:42:05 +08:00
void
if_init(void)
{
if_pool = rp_new(&root_pool, "Interfaces");
init_list(&iface_list);
Reimplemented neighbor cache. Now uses real hashing.
2000-03-01 22:51:47 +08:00
neigh_init(if_pool);
Implemented scanning of network interfaces. Mostly very ugly code due to terrible kernel interface (SIOGIFCONF and friends).
1998-05-27 05:42:05 +08:00
}
Added functions for manipulating interface name pattern lists: o iface_patt_match(list, iface) -- match interface against list o iface_patts_equal(a, b, c) -- compare whether two pattern lists are equal or not. c(x,y) is called for comparison of protocol-dependent data.
1998-11-30 06:01:03 +08:00
/*
* Interface Pattern Lists
*/
struct iface_patt *
iface_patt_match(list *l, struct iface *i)
{
struct iface_patt *p;
WALK_LIST(p, *l)
{
char *t = p->pattern;
int ok = 1;
Cleaned up handling of interface patterns: o Parsing of interface patterns moved to generic code, introduced this_ipatt which works similarly to this_iface. o Interface patterns now support selection by both interface names and primary IP addresses. o Proto `direct' updated. o RIP updated as well, it also seems the memory corruption bug there is gone.
1999-08-04 03:30:49 +08:00
if (t)
Added functions for manipulating interface name pattern lists: o iface_patt_match(list, iface) -- match interface against list o iface_patts_equal(a, b, c) -- compare whether two pattern lists are equal or not. c(x,y) is called for comparison of protocol-dependent data.
1998-11-30 06:01:03 +08:00
{
Cleaned up handling of interface patterns: o Parsing of interface patterns moved to generic code, introduced this_ipatt which works similarly to this_iface. o Interface patterns now support selection by both interface names and primary IP addresses. o Proto `direct' updated. o RIP updated as well, it also seems the memory corruption bug there is gone.
1999-08-04 03:30:49 +08:00
if (*t == '-')
{
t++;
ok = 0;
}
if (!patmatch(t, i->name))
continue;
Added functions for manipulating interface name pattern lists: o iface_patt_match(list, iface) -- match interface against list o iface_patts_equal(a, b, c) -- compare whether two pattern lists are equal or not. c(x,y) is called for comparison of protocol-dependent data.
1998-11-30 06:01:03 +08:00
}
Cleaned up handling of interface patterns: o Parsing of interface patterns moved to generic code, introduced this_ipatt which works similarly to this_iface. o Interface patterns now support selection by both interface names and primary IP addresses. o Proto `direct' updated. o RIP updated as well, it also seems the memory corruption bug there is gone.
1999-08-04 03:30:49 +08:00
if (!i->addr || !ipa_in_net(i->addr->ip, p->prefix, p->pxlen))
continue;
return ok ? p : NULL;
Added functions for manipulating interface name pattern lists: o iface_patt_match(list, iface) -- match interface against list o iface_patts_equal(a, b, c) -- compare whether two pattern lists are equal or not. c(x,y) is called for comparison of protocol-dependent data.
1998-11-30 06:01:03 +08:00
}
return NULL;
}
int
iface_patts_equal(list *a, list *b, int (*comp)(struct iface_patt *, struct iface_patt *))
{
struct iface_patt *x, *y;
x = HEAD(*a);
y = HEAD(*b);
while (x->n.next && y->n.next)
{
Cleaned up handling of interface patterns: o Parsing of interface patterns moved to generic code, introduced this_ipatt which works similarly to this_iface. o Interface patterns now support selection by both interface names and primary IP addresses. o Proto `direct' updated. o RIP updated as well, it also seems the memory corruption bug there is gone.
1999-08-04 03:30:49 +08:00
if (strcmp(x->pattern, y->pattern) ||
!ipa_equal(x->prefix, y->prefix) ||
x->pxlen != y->pxlen ||
comp && !comp(x, y))
Added functions for manipulating interface name pattern lists: o iface_patt_match(list, iface) -- match interface against list o iface_patts_equal(a, b, c) -- compare whether two pattern lists are equal or not. c(x,y) is called for comparison of protocol-dependent data.
1998-11-30 06:01:03 +08:00
return 0;
x = (void *) x->n.next;
y = (void *) y->n.next;
}
return (!x->n.next && !y->n.next);
}
Added few basic commands: show status, show interfaces [summary], show protocols (incomplete).
1999-11-25 23:35:30 +08:00
/*
* CLI commands.
*/
static void
if_show_addr(struct ifa *a)
{
byte broad[STD_ADDRESS_P_LENGTH + 16];
byte opp[STD_ADDRESS_P_LENGTH + 16];
if (ipa_nonzero(a->brd))
bsprintf(broad, ", broadcast %I", a->brd);
else
broad[0] = 0;
if (ipa_nonzero(a->opposite))
bsprintf(opp, ", opposite %I", a->opposite);
else
opp[0] = 0;
Rewrote interface type detection logic. The `unnumbered' flag is now per address, not per interface (hence it's ifa->flags & IA_UNNUMBERED) and should be set reliably. IF_MULTIACCESS should be fixed now, but it isn't wise to rely on it on interfaces configured with /30 prefix.
2000-03-01 07:19:52 +08:00
cli_msg(-1003, "\t%I/%d (%s%s%s, scope %s%s)",
Added few basic commands: show status, show interfaces [summary], show protocols (incomplete).
1999-11-25 23:35:30 +08:00
a->ip, a->pxlen,
(a->flags & IA_PRIMARY) ? "Primary" : (a->flags & IA_SECONDARY) ? "Secondary" : "???",
broad, opp,
Rewrote interface type detection logic. The `unnumbered' flag is now per address, not per interface (hence it's ifa->flags & IA_UNNUMBERED) and should be set reliably. IF_MULTIACCESS should be fixed now, but it isn't wise to rely on it on interfaces configured with /30 prefix.
2000-03-01 07:19:52 +08:00
ip_scope_text(a->scope),
(a->flags & IA_UNNUMBERED) ? ", unnumbered" : "");
Added few basic commands: show status, show interfaces [summary], show protocols (incomplete).
1999-11-25 23:35:30 +08:00
}
void
if_show(void)
{
struct iface *i;
struct ifa *a;
char *type;
WALK_LIST(i, iface_list)
{
`show interfaces' and `show protocols' works.
1999-11-30 20:57:14 +08:00
cli_msg(-1001, "%s %s (index=%d)", i->name, (i->flags & IF_UP) ? "up" : "DOWN", i->index);
Rewrote interface type detection logic. The `unnumbered' flag is now per address, not per interface (hence it's ifa->flags & IA_UNNUMBERED) and should be set reliably. IF_MULTIACCESS should be fixed now, but it isn't wise to rely on it on interfaces configured with /30 prefix.
2000-03-01 07:19:52 +08:00
if (!(i->flags & IF_MULTIACCESS))
Added few basic commands: show status, show interfaces [summary], show protocols (incomplete).
1999-11-25 23:35:30 +08:00
type = "PtP";
else
type = "MultiAccess";
`show interfaces' and `show protocols' works.
1999-11-30 20:57:14 +08:00
cli_msg(-1004, "\t%s%s%s Admin%s Link%s%s%s MTU=%d",
Added few basic commands: show status, show interfaces [summary], show protocols (incomplete).
1999-11-25 23:35:30 +08:00
type,
(i->flags & IF_BROADCAST) ? " Broadcast" : "",
(i->flags & IF_MULTICAST) ? " Multicast" : "",
(i->flags & IF_ADMIN_DOWN) ? "Down" : "Up",
(i->flags & IF_LINK_UP) ? "Up" : "Down",
(i->flags & IF_LOOPBACK) ? " Loopback" : "",
(i->flags & IF_IGNORE) ? " Ignored" : "",
i->mtu);
if (i->addr)
if_show_addr(i->addr);
WALK_LIST(a, i->addrs)
if (a != i->addr)
if_show_addr(a);
}
cli_msg(0, "");
}
void
if_show_summary(void)
{
struct iface *i;
byte addr[STD_ADDRESS_P_LENGTH + 16];
`show interfaces' and `show protocols' works.
1999-11-30 20:57:14 +08:00
cli_msg(-2005, "interface state address");
Added few basic commands: show status, show interfaces [summary], show protocols (incomplete).
1999-11-25 23:35:30 +08:00
WALK_LIST(i, iface_list)
{
if (i->addr)
bsprintf(addr, "%I/%d", i->addr->ip, i->addr->pxlen);
else
addr[0] = 0;
`show interfaces' and `show protocols' works.
1999-11-30 20:57:14 +08:00
cli_msg(-1005, "%-9s %-5s %s", i->name, (i->flags & IF_UP) ? "up" : "DOWN", addr);
Added few basic commands: show status, show interfaces [summary], show protocols (incomplete).
1999-11-25 23:35:30 +08:00
}
cli_msg(0, "");
}
Reference in a new issue Copy permalink