Paul Jakma [Wed, 24 Jan 2007 14:51:51 +0000 (14:51 +0000)]
[ospfd] Bug #330: SPF must consider that nexthop-calc may fail
2007-01-24 Paul Jakma <paul.jakma@sun.com>
* ospf_spf.c: Bug #330: Nexthop calculation sometimes may fail,
and it needs to indicate this result to SPF.
(ospf_spf_add_parent) Flush of parent list needs to be done here,
for simplicity.
(ospf_nexthop_calculation) Caller needs to know whether
nexthop calculation succeeded. Every return statement must
correctly indicate such.
(ospf_spf_next) Queueing/prioritisation of vertices in SPF
must take into account whether nexthop_calculation succeeded,
or SPF may fail to find best paths.
Andrew J. Schorr [Wed, 13 Dec 2006 15:44:15 +0000 (15:44 +0000)]
[zebra] For solaris IPv6 PtP interfaces, try to support prefixlen != 128
2006-12-13 Andrew J. Schorr <ajschorr@alumni.princeton.edu>
* if_ioctl_solaris.c: (if_get_addr) For IPv6, stop assuming
that all IFF_POINTOPOINT have prefixlen of IPV6_MAX_BITLEN.
Instead, always try the SIOCGLIFSUBNET ioctl; if that fails,
then we fall back to IPV6_MAX_BITLEN for PtP interfaces.
Andrew J. Schorr [Tue, 12 Dec 2006 19:18:21 +0000 (19:18 +0000)]
[PtP over ethernet] New peer flag allows much more addressing flexibility
2006-12-12 Andrew J. Schorr <ajschorr@alumni.princeton.edu>
* if.h: (struct connected) Add new ZEBRA_IFA_PEER flag indicating
whether a peer address has been configured. Comment now shows
the new interpretation of the destination addr: if ZEBRA_IFA_PEER
is set, then it must contain the destination address, otherwise
it may contain the broadcast address or be NULL.
(CONNECTED_DEST_HOST,CONNECTED_POINTOPOINT_HOST) Remove obsolete
macros that were specific to IPv4 and not fully general.
(CONNECTED_PEER) New macro to check ZEBRA_IFA_PEER flag.
(CONNECTED_PREFIX) New macro giving the prefix to insert into
the RIB: if CONNECTED_PEER, then use the destination (peer) address,
else use the address field.
(CONNECTED_ID) New macro to come up with an identifying address
for the struct connected.
* if.c: (if_lookup_address, connected_lookup_address) Streamline
logic with new CONNECTED_PREFIX macro.
* prefix.h: (PREFIX_COPY_IPV4, PREFIX_COPY_IPV6) New macros
for better performance than the general prefix_copy function.
* zclient.c: (zebra_interface_address_read) For non-null destination
addresses, set prefixlen to equal the address prefixlen. This
is needed to get the new CONNECTED_PREFIX macro to work properly.
* connected.c: (connected_up_ipv4, connected_down_ipv4,
connected_up_ipv6, connected_down_ipv6) Simplify logic using the
new CONNECTED_PREFIX macro.
(connected_add_ipv4) Set prefixlen in destination addresses (required
by the CONNECTED_PREFIX macro). Use CONNECTED_PEER macro instead
of testing for IFF_POINTOPOINT. Delete invalid warning message.
Warn about cases where the ZEBRA_IFA_PEER is set but no
destination address has been supplied (and turn off the flag).
(connected_add_ipv6) Add new flags argument so callers may set
the ZEBRA_IFA_PEER flag. If peer/broadcast address satisfies
IN6_IS_ADDR_UNSPECIFIED, then reject it with a warning.
Set prefixlen in destination address so CONNECTED_PREFIX will work.
* connected.h: (connected_add_ipv6) Add new flags argument so
callers may set the ZEBRA_IFA_PEER flag.
* interface.c: (connected_dump_vty) Use CONNECTED_PEER macro
to decide whether the destination address is a peer or broadcast
address (instead of checking IFF_BROADCAST and IFF_POINTOPOINT).
* if_ioctl.c: (if_getaddrs) Instead of setting a peer address
only when the IFF_POINTOPOINT is set, we now accept a peer
address whenever it is available and not the same as the local
address. Otherwise (no peer address assigned), we check
for a broadcast address (regardless of the IFF_BROADCAST flag).
And must now pass a flags value of ZEBRA_IFA_PEER to
connected_add_ipv4 when a peer address is assigned.
The same new logic is used with the IPv6 code as well (and we
pass the new flags argument to connected_add_ipv6).
(if_get_addr) Do not bother to check IFF_POINTOPOINT: just
issue the SIOCGIFDSTADDR ioctl and see if we get back
a peer address not matching the local address (and set
the ZEBRA_IFA_PEER in that case). If there's no peer address,
try to grab SIOCGIFBRDADDR regardless of whether IFF_BROADCAST is set.
* if_ioctl_solaris.c: (if_get_addr) Just try the SIOCGLIFDSTADDR ioctl
without bothering to check the IFF_POINTOPOINT flag. And if
no peer address was found, just try the SIOCGLIFBRDADDR ioctl
without checking the IFF_BROADCAST flag. Call connected_add_ipv4
and connected_add_ipv6 with appropriate flags.
* if_proc.c: (ifaddr_proc_ipv6) Must pass new flags argument to
connected_add_ipv6.
* kernel_socket.c: (ifam_read) Must pass new flags argument to
connected_add_ipv6.
* rt_netlink.c: (netlink_interface_addr) Copy logic from iproute2
to determine local and possible peer address (so there's no longer
a test for IFF_POINTOPOINT). Set ZEBRA_IFA_PEER flag appropriately.
Pass new flags argument to connected_add_ipv6.
(netlink_address) Test !CONNECTED_PEER instead of if_is_broadcast
to determine whether the connected destination address is a
broadcast address.
* bgp_nexthop.c: (bgp_connected_add, bgp_connected_delete)
Simplify logic by using new CONNECTED_PREFIX macro.
* ospf_interface.c: (ospf_if_is_configured, ospf_if_lookup_by_prefix,
ospf_if_lookup_recv_if) Simplify logic using new CONNECTED_PREFIX
macro.
* ospf_lsa.c: (lsa_link_ptop_set) Using the new CONNECTED_PREFIX
macro, both options collapse into the same code.
* ospf_snmp.c: (ospf_snmp_if_update) Simplify logic using new
CONNECTED_ID macro.
(ospf_snmp_is_if_have_addr) Simplify logic using new CONNECTED_PREFIX
macro.
* ospf_vty.c: (show_ip_ospf_interface_sub) Use new CONNECTED_PEER macro
instead of testing the IFF_POINTOPOINT flag.
* ospfd.c: (ospf_network_match_iface) Use new CONNECTED_PEER macro
instead of testing with if_is_pointopoint. And add commented-out
code to implement alternative (in my opinion) more elegant behavior
that has no special-case treatment for PtP addresses.
(ospf_network_run) Use new CONNECTED_ID macro to simplify logic.
* rip_interface.c: (rip_interface_multicast_set) Use new CONNECTED_ID
macro to simplify logic.
(rip_request_interface_send) Fix minor bug: ipv4_broadcast_addr does
not give a useful result if prefixlen is 32 (we require a peer
address in such cases).
* ripd.c: (rip_update_interface) Fix same bug as above.
Paul Jakma [Fri, 8 Dec 2006 00:53:14 +0000 (00:53 +0000)]
[zebra] Changes of nexthops of static routes didnt take effect
2006-12-08 Piotr Chytla <pch@packetconsulting.pl>
* zebra_rib.c: (static_install_ipv{4,6}) Case where existing
RIB is updated must explicitely rib_addqueue the route_node,
to ensure the update actually takes effect.
Paul Jakma [Fri, 8 Dec 2006 00:31:22 +0000 (00:31 +0000)]
[bgpd] Bug #302, bgpd can get stuck in state Clearing
2006-12-07 Paul Jakma <paul.jakma@sun.com>
* bgp_fsm.c: Bug #302 fix, diagnosis, suggestions and testing
by Juergen Kammer <j.kammer@eurodata.de>. Fix follows from
his suggested fix, just made in a slightly different way.
(bgp_event) Transitions into Clearing always must call
bgp_clear_route_all().
(bgp_stop) No need to clear routes here, BGP FSM should do
it.
[ospfd] Consider all connected addresses when creating ospf interfaces
2006-12-04 Andrew J. Schorr <ajschorr@alumni.princeton.edu>
* ospfd.c: (ospf_network_run) Remove an offending 'break' statement.
Previously, after creating a single ospf_interface on a given
network interface, the code would skip to the next interface
without considering other connected addresses on the interface.
After removing the 'break', we now consider all connected addresses.
Andrew J. Schorr [Thu, 30 Nov 2006 16:36:57 +0000 (16:36 +0000)]
[bgpd] Implement 'debug bgp zebra' to log all messages to and from zebra.
2006-11-30 Andrew J. Schorr <ajschorr@alumni.princeton.edu>
* bgp_debug.h: Declare new bgp_debug_zebra conf and term flags,
and define BGP_DEBUG_ZEBRA.
* bgp_debug.c: Declare conf_bgp_debug_zebra and term_bgp_debug_zebra.
(debug_bgp_zebra, no_debug_bgp_zebra, undebug_bgp_zebra) New
functions to enable/disable bgp zebra debugging.
(no_debug_bgp_all) Turn off zebra debugging.
(show_debugging_bgp) Show whether zebra debugging is on.
(bgp_config_write_debug) Add 'debug bgp zebra' if configured.
(bgp_debug_init) Add new zebra debugging commands.
* bgp_zebra.c: (bgp_router_id_update, bgp_interface_add,
bgp_interface_delete, bgp_interface_up, bgp_interface_down,
bgp_interface_address_add, bgp_interface_address_delete,
zebra_read_ipv4, zebra_read_ipv6, bgp_zebra_announce,
bgp_zebra_withdraw, bgp_redistribute_set, bgp_redistribute_unset)
If zebra debugging is enabled, log an appropriate debug message.
Andrew J. Schorr [Thu, 30 Nov 2006 16:17:02 +0000 (16:17 +0000)]
[ospfd] Add debug messages for a few zebra messages that had been overlooked
2006-11-30 Andrew J. Schorr <ajschorr@alumni.princeton.edu>
* ospf_zebra.c: (ospf_router_id_update_zebra,
ospf_interface_address_add, ospf_interface_address_delete)
If (IS_DEBUG_OSPF (zebra, ZEBRA_INTERFACE)) is enabled, then
add a debug message about what Zebra is telling us.
(ospf_zebra_add_discard) Add a debug message matching the one
already in ospf_zebra_delete_discard.
Andrew J. Schorr [Tue, 28 Nov 2006 19:50:46 +0000 (19:50 +0000)]
[bgpd] Fix bug where a deleted route that was quickly re-added was being lost
2006-11-28 Andrew J. Schorr <ajschorr@alumni.princeton.edu>
* bgp_route.c: (bgp_info_restore) New function that undoes
the effects of a previous call to bgp_info_delete. This is
used when a route is deleted and quickly re-added before the
deletion has been processed.
(bgp_static_update_rsclient, bgp_static_update_main,
bgp_redistribute_add) Check whether a pre-existing route
has the BGP_INFO_REMOVED set, and, if so, we need to call
bgp_info_restore to resurrect it.
Andrew J. Schorr [Tue, 28 Nov 2006 16:36:39 +0000 (16:36 +0000)]
[ospfd] Fix bug in passive-interface default commands.
2006-11-28 Andrew J. Schorr <ajschorr@alumni.princeton.edu>
* ospf_vty.c: (ospf_passive_interface_default) Take additional
'newval' arg so we can update ospf->passive_interface_default inside
this function. More importantly, we now call ospf_if_set_multicast
on all ospf_interfaces.
(ospf_passive_interface, no_ospf_passive_interface) Fix bug:
for 'default' case, argv[0] is undefined, so we must test for
(argc == 0) before using argv[0]. And since
ospf_passive_interface_default now calls ospf_if_set_multicast as
needed, we can just return after calling
ospf_passive_interface_default.
Andrew J. Schorr [Tue, 24 Oct 2006 19:04:26 +0000 (19:04 +0000)]
[ospfd] Stop losing subsequent default-information originate 'always' info
2006-10-24 Andrew J. Schorr <ajschorr@alumni.princeton.edu>
* ospf_zebra.c: (ospf_redistribute_default_set) Fix bug where
a new value for ospf->default_originate was being ignored
if a previous 'default-information originate' command
had already been processed.
Paul Jakma [Sun, 22 Oct 2006 20:07:53 +0000 (20:07 +0000)]
[ospfd] Add passive-interface default support
2006-10-22 Yar Tikhiy <yar@comp.chem.msu.su>
* (general) Add support for passive-interface default (with
minor edits by Paul Jakma).
* ospf_interface.h: Add OSPF_IF_PASSIVE_STATUS macro, looking
at configured value, or the global 'default' value, as
required.
* ospf_interface.c: (ospf_if_new_hook) Leave passive
unconfigured per default, allowing global 'default' to
take effect for unconfigured interfaces.
* ospf_packet.c: (various) use OSPF_IF_PASSIVE_STATUS
* ospf_vty.c: (ospf_passive_interface_default) new function,
unset passive from all interfaces if default is enabled, as
the per-iface settings become redundant.
(ospf_passive_interface_update) new func, update passive
setting taking global default into account.
({no,}ospf_passive_interface_addr_cmd) Add support for
'default' variant of command.
(show_ip_ospf_interface_sub) Update to take global
default into account when printing passive status.
(ospf_config_write) ditto.
* ospfd.c: (ospf_new) set global passive-interface default.
* ospfd.h: (struct ospf) Add field for global
passive-interface.
Paul Jakma [Sun, 22 Oct 2006 19:13:07 +0000 (19:13 +0000)]
[bgpd] struct peer must have bgp field valid (redistribute crash)
2006-10-19 Paul Jakma <paul.jakma@sun.com>
* bgpd.c: (peer_new) bgp element of peer absolutely must be
filled in, make peer_new() require it as argument and update
all callers. Fixes a crash reported by Jan 'yanek' Bortl and
Andrew Schorr where bgpd would crash in bgp_pcount_adjust
trying to dereference the bgp member of bgp->peer_self,
triggered through redistribution.
* bgp_route.c: (bgp_pcount_adjust) assert sanity of arguments.
* bgp_packet.c: (bgp_update_packet) adv->rn can not be NULL,
check is bogus - changed to assert(), CID#64.
binfo is checked for NULL, but then dereferenced
unconditionally, fix, CID #63.
(bgp_withdraw_packet) Assert adv->rn is valid, as with
bgp_update_packet().
* bgp_fsm.h: Remove BGP_EVENT_FLUSH_ADD, dangerous and not
needed.
* bgp_fsm.c: (bgp_stop) Move BGP_EVENT_FLUSH to the top of the
of the function, otherwise it could flush a ClearingCompleted
event, bug #302.
* bgp_packet.c: Replace all BGP_EVENT_FLUSH_ADD with
BGP_EVENT_ADD, fixing bug #302.
Paul Jakma [Sun, 15 Oct 2006 23:35:57 +0000 (23:35 +0000)]
[daemon startup] Add --dry-run/-C argument to daemons, to check config file syntax
2006-10-04 Oliver Hookins <ohookins@gmail.com>
* bgpd/bgp_main.c: Add configuration check option, with
'-C' rather than '-c' for consistency between daemons.
* isisd/isis_main.c: ditto
* ospf6d/ospf6_main.c: ditto
* ospfd/ospf_main.c: ditto
* ripngd/ripng_main.c: ditto
* vtysh/vtysh_main.c: ditto
* ripd/rip_main.c: Change the config check option to
'-C' and tidy up the code.
* zebra/main.c: ditto
* ripd/rip_main.c: This trivial patch introduces a new
command-line option '-c', which instructs zebra/ripd
to check its configuration file for validity, print
any error message, and then exit. This is useful when
the configuration file is edited by hand or otherwise,
and you simply want to validate it without any other
effect.
* zebra/main.c: ditto
Paul Jakma [Sun, 15 Oct 2006 23:33:50 +0000 (23:33 +0000)]
[lib] Handle upgrade from SUNWzebra to Quagga for 'interface fooX:Y' commands
2006-10-14 Paul Jakma <paul.jakma@sun.com>
* if.c: (general) Handle upgrades from SUNWzebra, which tried
to track each logical interface as a seperate struct
interface, to Quagga, which assigns only one struct interface
per ifindex.
(if_sunwzebra_get) Try decompose a logical interface name
(fooX:Y) to the 'primary' name (fooX), for Solaris.
(interface_cmd) Use if_sunwzebra_get on Solaris.
* smux.c: (smux_open,smux_trap,smux_register) Fix various
asn_build_* calls to pass the proper length in the final
argument: use sizeof(<variable>) instead of sizeof(<type>),
since there were several inconsistencies between the actual
variable type and the size that was passed. This should
fix some problems on 64-bit architectures where sizeof(int)
!= sizeof(long).
Andrew J. Schorr [Mon, 25 Sep 2006 13:26:14 +0000 (13:26 +0000)]
[ospfd] Improve some warning messages.
2006-09-25 Andrew J. Schorr <ajschorr@alumni.princeton.edu>
* ospf_packet.c: (ospf_packet_dup, ospf_make_md5_digest)
Fix zlog_warn messages to eliminate compiler warnings.
(ospf_hello) Improve warning messages to show why we
are complaining.
Paul Jakma [Thu, 14 Sep 2006 03:38:16 +0000 (03:38 +0000)]
[bgpd] reduce the process queue hold time to something more sensible
2006-09-14 Paul Jakma <paul.jakma@sun.com>
* bgp_route.c: (bgp_process_queue_init) process queue hold time
too high, adds extra memory load. Change to be much lower,
until such time as it's made configurable.
Paul Jakma [Thu, 14 Sep 2006 03:06:54 +0000 (03:06 +0000)]
[bgpd] RIB statistics address space size shouldnt double count space
2006-09-14 Paul Jakma <paul.jakma@sun.com>
* bgp_route.c: (bgp_table_stats_walker) Address space announced
should only count top-level unaggregateable prefixes, to
avoid falling afoul of anti-dodgy-accounting regulations
in various jurisdictions.. ;)
* (general) fix the peer refcount issue exposed by previous, by
just removing refcounting of peer threads, which is mostly
senseless as they're references leading from struct peer,
which peer_free cancels anyway. No need to muck around..
* bgp_fsm.h: Just remove the refcounting from the various
TIMER/READ/WRITE/EVENT ON/OFF/ADD macros.
* bgp_fsm.c: (bgp_stop) use BGP_EVENT_FLUSH, no refcounts attached
to events anymore.
(bgp_event) remove peer_unlock, events not refcounted.
* bgpd.c: (peer_free) flush events before free.
Paul Jakma [Thu, 14 Sep 2006 02:58:49 +0000 (02:58 +0000)]
[bgpd] Fix 0.99 shutdown regression, introduce Clearing and Deleted states
2006-09-14 Paul Jakma <paul.jakma@sun.com>
* (general) Fix some niggly issues around 'shutdown' and clearing
by adding a Clearing FSM wait-state and a hidden 'Deleted'
FSM state, to allow deleted peers to 'cool off' and hit 0
references. This introduces a slow memory leak of struct peer,
however that's more a testament to the fragility of the
reference counting than a bug in this patch, cleanup of
reference counting to fix this is to follow.
* bgpd.h: Add Clearing, Deleted states and Clearing_Completed
and event.
* bgp_debug.c: (bgp_status_msg[]) Add strings for Clearing and
Deleted.
* bgp_fsm.h: Don't allow timer/event threads to set anything
for Deleted peers.
* bgp_fsm.c: (bgp_timer_set) Add Clearing and Deleted. Deleted
needs to stop everything.
(bgp_stop) Remove explicit fsm_change_status call, the
general framework handles the transition.
(bgp_start) Log a warning if a start is attempted on a peer
that should stay down, trying to start a peer.
(struct .. FSM) Add Clearing_Completed
events, has little influence except when in state
Clearing to signal wait-state can end.
Add Clearing and Deleted states, former is a wait-state,
latter is a placeholder state to allow peers to disappear
quietly once refcounts settle.
(bgp_event) Try reduce verbosity of FSM state-change debug,
changes to same state are not interesting (Established->Established)
Allow NULL action functions in FSM.
* bgp_packet.c: (bgp_write) Use FSM events, rather than trying
to twiddle directly with FSM state behind the back of FSM.
(bgp_write_notify) ditto.
(bgp_read) Remove the vague ACCEPT_PEER peer_unlock, or else
this patch crashes, now it leaks instead.
* bgp_route.c: (bgp_clear_node_complete) Clearing_Completed
event, to end clearing.
(bgp_clear_route) See extensive comments.
* bgpd.c: (peer_free) should only be called while in Deleted,
peer refcounting controls when peer_free is called.
bgp_sync_delete should be here, not in peer_delete.
(peer_delete) Initiate delete.
Transition to Deleted state manually.
When removing peer from indices that provide visibility of it,
take great care to be idempotent wrt the reference counting
of struct peer through those indices.
Use bgp_timer_set, rather than replicating.
Call to bgp_sync_delete isn't appropriate here, sync can be
referenced while shutting down and finishing deletion.
(peer_group_bind) Take care to be idempotent wrt list references
indexing peers.
Paul Jakma [Thu, 14 Sep 2006 02:56:07 +0000 (02:56 +0000)]
[bgpd] Add RIB reporting commands, show bgp ... statistics
2006-09-13 Paul Jakma <paul.jakma@sun.com>
* bgp_aspath.c: (aspath_highest) new, return highest ASN in an
aspath.
* bgp_route.c: (bgp_peer_count_walker) new, do the walk done
in bgp_peer_counts as a thread.
(bgp_peer_counts) move walk to previous and call it via
thread_execute so this RIB walk shows up in thread stats.
(bgp_table_stats) New, gather some statistics for a given
RIB.
(bgp_table_stats_walker) New, RIB walker thread for former.
(bgp_table_stats_vty) Parsing front-end for 'show bgp ...',
useful model for future rationalisation of 'show ... bgp'.
(bgp_route_init) Add new RIB stats commands.
Paul Jakma [Thu, 7 Sep 2006 00:24:49 +0000 (00:24 +0000)]
[bgpd] Handle pcount as flags are changed, fixing pcount issues
2006-09-06 Paul Jakma <paul.jakma@sun.com>
* (general) Squash any and all prefix-count issues by
abstracting route flag changes, and maintaining count as and
when flags are modified (rather than relying on explicit
modifications of count being sprinkled in just the right
places throughout the code).
* bgp_route.c: (bgp_pcount_{dec,inc}rement) removed.
(bgp_pcount_adjust) new, update prefix count as
needed for a given route.
(bgp_info_{uns,s}et_flag) set/unset a BGP_INFO route status
flag, calling previous function when appropriate.
(general) Update all set/unsets of flags to use previous.
Remove pcount_{dec,inc}rement calls.
No need to unset BGP_INFO_VALID in places where
bgp_info_delete is called, it does that anyway.
* bgp_{damp,nexthop}.c: Update to use bgp_info_{un,}set_flag.
* bgp_route.h: Export bgp_info_{un,}set_flag.
Add a 'meta' BGP_INFO flag, BGP_INFO_UNUSEABLE.
Move BGP_INFO_HOLDDOWN macro to here from bgpd.h
* bgp_route.c: Add 'show ... bgp ... <neighbour> prefix-count'
commands, to provide detailed counts of prefixes for a peer.
Informative, and should help pin down to pfxcnt drift
problems.
Paul Jakma [Wed, 30 Aug 2006 18:47:37 +0000 (18:47 +0000)]
[ospfd] Fix assertion in DB-exchange fix, hit by ogier-db-ex-opt commit
2006-08-28 Andy Gay <andy@andynet.net>
* ospf_packet.c: (ospf_make_db_desc) Assert added with More-bit
fixes does not hold up with addition of Ogier DB-Exchange
optimisation, which can empty the db-summary list in between
sent DD packets. Remove assert, update More-bit always when
in Exchange.
* ospfd.c: (ospf_finish_final) default redistribute should be
unset too, fixes bug where reconfiguring ospfd completely
can no longer enable default redistribution.
Paul Jakma [Sun, 27 Aug 2006 06:49:29 +0000 (06:49 +0000)]
[ospfd] Bug #134, ospfd should be more robust to backward time change
2006-08-25 Paul Jakma <paul.jakma@sun.com>
* (general) Bug #134. Be more robust to backward time changes,
use the newly added libzebra time functions.
In most cases: recent_time -> recent_relative_time()
gettimeofday -> quagga_gettime (QUAGGA_CLK_MONOTONIC, ..)
time -> quagga_time.
(ospf_make_md5_digest) time() call deliberately not changed.
(ospf_external_lsa_refresh) remove useless gettimeofday, LSA
tv_orig time was already set in ospf_lsa_new, called via
ospf_external_lsa_new.
Paul Jakma [Sun, 27 Aug 2006 06:44:02 +0000 (06:44 +0000)]
[lib] Bug #134: threads should be more robust against backward time jumps
2006-08-25 Paul Jakma <paul.jakma@sun.com>
* thread.c: (general) Add support for monotonic clock, it may still
jump forward by huge amounts, but should be immune to going
backwards. Fixes bug #134.
(quagga_gettimeofday_relative_adjust) helper, does what name
says - adjusts gettimeofday based relative timer.
(quagga_gettimeofday) helper to keep recent_time up to date.
(quagga_get_relative) helper, update and getch the relative
timer using gettimeofday(). POSIX CLOCK_MONOTONIC is also
supported, but the code is not enabled yet nor tested.
(quagga_real_stabilised) helper, retrieve absolute time but
stabilised so as to never decrease.
(quagga_gettime) Exported interface, analogous to POSIX
clock_gettime() in interface, supporting several clocks.
(quagga_time) Exported interface, analogous to traditional
time(), will never decrease.
(recent_relative_time) Convenience function to retrieve
relative_time timeval, similar to existing recent_time absolute
timeval, for when an approximately recent value will do.
(remainder) Update to use above helpers.
(thread_getrusage) Previously was a macro, but needs to be
a function to twiddle with thread.c private stuff.
* thread.c: Point the GETRUSAGE macro at previous function.
Export quagga_gettime, quagga_time and recent_relative_time for
general use.
Paul Jakma [Sun, 27 Aug 2006 06:29:30 +0000 (06:29 +0000)]
[ospfd] Raise ExchangeDone earlier, avoid often needless round of DD packets
2006-08-03 Paul Jakma <paul.jakma@sun.com>
* ospf_packet.c: (ospf_make_db_desc) Unset the DD More bit
after constructing the packet, if appropriate.
(ospf_db_desc_proc) Speed up Exchange, slave should raise
ExchangeDone earlier, as RFC mandates, by forming its reply
before deciding whether both sides are done, avoids a
needless round of empty DD packet exchanges at the end of
Exchange, hence speeding up ExchangeDone.
(ospf_db_desc) use UNSET_FLAG macro.
Paul Jakma [Sun, 27 Aug 2006 06:24:34 +0000 (06:24 +0000)]
[ospfd] trivial: consolidate LSDB delete code into single function
2006-08-04 Paul Jakma <paul.jakma@sun.com>
* ospf_lsdb.c: (ospf_lsdb_delete_entry) new function, consolidate
exact same functionality replicated in other functions.
(ospf_lsdb_add) Strip out code by using ospf_lsdb_delete_entry.
(ospf_lsdb_delete) ditto.
(ospf_lsdb_delete_all) ditto.
Paul Jakma [Sun, 6 Aug 2006 15:57:59 +0000 (15:57 +0000)]
[zebra] fix inconsistencies in ifstat_update_* declarations and definitions
2006-08-06 Paul Jakma <paul.jakma@sun.com>
* interface.h: (ifstat_update_proc) declaration should match
ifstat_update_sysctl really, which is to not return status, as
such status is not used anywhere.
* if_{proc,sysctl}.c: Make ifstat_update_* definitions and return values
consistent with each other and their declarations, ie:
(void) (*) (void).
Paul Jakma [Sun, 6 Aug 2006 15:52:11 +0000 (15:52 +0000)]
[bgpd] aspath_loop_check was broken, fix it and the aspath unit test code.
2006-08-06 Paul Jakma <paul.jakma@sun.com>
* bgp_aspath.c: (aspath_loop_check) Fix the typo-bug which
essentially had disabled this check. Problem reported by
Bartek Kania <mrbk@gnarf.org> in [quagga-users 7388].
* aspath_test.c: (validate) Fix the sense of the aspath_loop_check,
which was the wrong way around and had actually been testing for
aspath_loop_check to be buggy.
Paul Jakma [Fri, 4 Aug 2006 06:18:04 +0000 (06:18 +0000)]
[zebra] IRDP: Move stream_free to where its created, probably fixing a leak
2006-08-01 Paul Jakma <paul.jakma@sun.com>
* irdp_main.c: (irdp_advertisement) free the stream here, when done,
right under where it was allocated so it's blindingly obvious
it's correct. This possibly fixes a very slow leak of streams in
zebra.
* irdp_packet.c: (send_packet) don't free the stream here as
it's hard to tell if right, plus an error case seemed to
returning before free anyway.
Paul Jakma [Fri, 28 Jul 2006 04:42:39 +0000 (04:42 +0000)]
[doc] Expand/cross-ref MD5 commands, tweak anchors to avoid added spacing
2006-07-28 Paul Jakma <paul.jakma@sun.com>
* main.texi: link-detect works on Solaris too.
* ospfd.texi: Twiddle around with anchors a bit more.
Clarify how setting MD5 auth by area and by interface interact,
and add cross-references, as well as to the required
command for setting key material.
Paul Jakma [Thu, 27 Jul 2006 23:30:16 +0000 (23:30 +0000)]
[doc] minor tweaks, refine ospf redist,passive-inter, and abr-type help
2006-07-27 Paul Jakma <paul.jakma@sun.com>
* quagga.texi: Remove unused index definitions
Add an Index node - for the concept index.
* routeserver.texi: Set exampleindex to 0, so the example configs
with long IPv6 addresses stand better chance of fitting.
* overview.texi: 'Supported RFC' -> 'Supported RFCs'
Remove paragraph indentation - texinfo does that.
Revise the supported OS list slightly.
Remove the IPv6 stack list, seems very dated and irrelevant.
Revise the 'How to get Quagga' section.
* ospfd.texi: minor tweaks: add some anchors, fix some minor
format issues.
Revise the help for 'abr-type'.
Note that text authentication is unwise, recc'd MD5.
Add some extra text for redistribute and passive-interface,
about how latter can substitute for redist connected.
Andrew J. Schorr [Thu, 27 Jul 2006 22:29:06 +0000 (22:29 +0000)]
[ospfd] Bug #288: do not change router ID unless forced by manual configuration
2006-07-27 Andrew J. Schorr <ajschorr@alumni.princeton.edu>
* ospfd.c: (ospf_router_id_update) Fix and document the algorithm for
selecting the router ID: if there is not a statically configured ID,
then stick to the most recent value to avoid disruptive changes.
This should fix bug #288.
Paul Jakma [Thu, 27 Jul 2006 21:49:00 +0000 (21:49 +0000)]
[zebra] Bug #268, Fix race between add/delete of routes, sanitise rib queueing
2006-07-27 Paul Jakma <paul.jakma@sun.com>
* rib.h: (struct rib) Add a route_node rn_status flag field,
this has to be copied every time head RIB of a route_node
changes.
Remove the rib lock field, not needed - see below.
Add a status field for RIB-private flags.
* zebra_rib.c: Add a global for the workqueue hold time, useful
for testing.
(general) Fix for bug #268. Problem originally
detailed by Simon Bryden in [quagga-dev 4001].
Essentially, add/delete of a RIB must happen /before/ the
queue. Best-path selection (ie rib_process) and reaping of
freed RIBs can then be done after queueing. Only the route_node
is queued - no important RIB state (i.e. whether a RIB is to be
deleted) is queued.
(struct zebra_queue_node_t) Disappears, no longer need to
track multiple things on the queue, only the route_node.
(rib_{lock,unlock}) removed, RIBs no longer need to be
refcounted, no longer queued.
(rib_queue_qnode_del) Removed, deleted RIBs no longer deleted
via the queue.
(rib_queue_add_qnode) deleted
(rib_queue_add) Only the route_node is queued for best-path
selection, we can check whether it is already queued or
not and avoid queueing same node twice - struct rib * argument
is not needed.
(rib_link/unlink) (un)link RIB from route_node.
(rib_{add,del}node) Front-end to updates of a RIB.
(rib_process) Reap any deleted RIBs via rib_unlink.
Unset the route_node 'QUEUED' flag.
(General) Remove calls to rib_queue_add where add/del node was
called - not needed, update calls where not.
Ignore RIB_ENTRY_REMOVEd ribs in loops through route_nodes
Paul Jakma [Thu, 27 Jul 2006 19:59:58 +0000 (19:59 +0000)]
[zebra] Add test rig code, for testing the zebra RIB
2006-07-27 Paul Jakma <paul.jakma@sun.com>
* {ioctl,kernel}_null.c: Dummy/Null kernel method implementations,
useful for testing zebra code that calls such methods.
* {redistribute,misc}_null.c: Dummy/Null methods, as above. But
for zclient, and for various misc functions.
* test_main.c: Test harness for zebra, currently just to test the
RIB.
* Makefile.am: Build testzebra using above.
* zebra_rib.c: Add a global for the workqueue hold time, useful
for testing.
Andrew J. Schorr [Thu, 27 Jul 2006 18:01:41 +0000 (18:01 +0000)]
[vtysh] Never skip authentication, and add support for multiple -c commands
2006-07-27 Andrew J. Schorr <ajschorr@alumni.princeton.edu>
* vtysh.1: Document new options -d and -E, and note that now multiple
-c options may be supplied, with embedded linefeed now supported.
In BUGS section, remove warning about vtysh causing a daemon
to freeze, since this has been fixed.
* vtysh_main.c: (usage) Add new -d and -E options. And note that
-c can be used multiple times, possibly with embedded linefeeds.
(longopts) Add new -d and -E options.
(main) Add new -d and -E options, and create a linked list to
support multiple -c options. Do not call vtysh_connect_all until
after vtysh_read_config(config_default) and vtysh_auth have
succeeded. This prevents the vtysh.conf file from configuring
any daemons, and it ensures that authentication has been passed
before we send any commands to any daemons. Call vtysh_connect_all
with any daemon name supplied with -d. If it is unable to connect
to any daemons, issue an error message and exit immediately.
When used in -c mode, call vtysh_execute("enable") before
executing the commands in order to match interactive behavior.
And detect embedded linefeed chars in -c commands and break them up
appropriately.
* vtysh.h: (vtysh_connect_all) Fix proto to reflect new
daemon_name argument, and that it now returns an integer -- the
number of daemons to which we were able to connect.
* vtysh.c: (vtysh_connect_all) Add a new daemon_name argument.
If supplied, connect only to that daemon. And return
the number of daemons to which we were able to connect.
(vtysh_prompt): Performance enhancement -- make struct utsname
static so we call uname to get the hostname only once.
Paul Jakma [Thu, 27 Jul 2006 16:11:02 +0000 (16:11 +0000)]
[zebra] Connected routes must always be added to main table
2006-07-27 Rumen Svobodnikov <rumen@telecoms.bg>
* connected.c: (connected_up_ipv4) interface connected routes always
go to table main (or otherwise they cannot be used by linux as
nexthops)
* zserv.c: (zread_ipv4_add) send route to the correct routing table
* zebra_rib.c (static_install_ipv4) set routing table
Paul Jakma [Wed, 26 Jul 2006 09:37:26 +0000 (09:37 +0000)]
[ospfd] Allow ospf_lsa_unlock to NULL out callers' LSA pointers upon free
2006-07-26 Paul Jakma <paul.jakma@sun.com>
* ospf_lsa.{c,h}: (ospf_lsa_unlock) Change to take a double pointer
to the LSA to be 'unlocked', so that, if the LSA is freed, the
callers pointer to the LSA can be NULLed out, allowing any further
use of that pointer to provoke a crash sooner rather than later.
* ospf_*.c: (general) Adjust callers of ospf_lsa_unlock to match
previous. Try annotate 'locking' somewhat to show which 'locks'
are protecting what LSA reference, if not obvious.
* ospf_opaque.c: (ospf_opaque_lsa_install) Trivial: remove useless
goto, replace with return.
* ospf_packet.c: (ospf_make_ls_ack) Trivial: merge two list loops,
the dual-loop predated the delete-safe list-loop macro.
Paul Jakma [Tue, 25 Jul 2006 20:44:12 +0000 (20:44 +0000)]
[ospfd] Additional NSM neighbour state change stats/information
2006-07-25 Paul Jakma <paul.jakma@sun.com>
* ospf_neigbor.h: (struct ospf_neighbor) Add some additional
neighbour state statistics fields, timestamps for progressive
and regressive state changes, and pointer to event string
for the latter state change.
* ospf_nsm.c: (nsm_notice_state_change) Update new state changs
history as required.
* ospf_vty.c: (show_ip_ospf_neighbor_detail_sub) Print out above
new per-neighbour state change stats.
Paul Jakma [Tue, 25 Jul 2006 20:40:40 +0000 (20:40 +0000)]
[lib] Optimise thread_call by caching pointer to thread history in the thread
2006-07-25 Paul Jakma <paul.jakma@sun.com>
* thread.h: (struct thread) Add a cache pointer to the struct
cpu_thread_history, if it is known - saving hash lookup on
each thread_call.
* thread.c: (thread_call) Cache the pointer to the
cpu_thread_history, so that future thread_calls of same
thread can avoid the hash_lookup.
Paul Jakma [Tue, 11 Jul 2006 17:57:25 +0000 (17:57 +0000)]
[ospfd] record timestamp and event of last NSM state change for neighbour
2006-07-10 Paul Jakma <paul.jakma@sun.com>
* ospf_nsm.c: (ospf_nsm_event) Record state change timestamp
and event in nbr struct.
* ospf_neighbor.h: (struct ospf_neighbor) Add fields to record
timestamp of last NSM change and event.
* ospf_vty.c: (show_ip_ospf_neighbor_detail_sub) Print
last state change timestamp and event, if available.
Paul Jakma [Tue, 11 Jul 2006 17:52:53 +0000 (17:52 +0000)]
[ospfd] trivial: NSM AdjChange should print event, declutter core functions
2006-07-10 Paul Jakma <paul.jakma@sun.com>
* ospf_nsm.c: (nsm_notice_state_changes) Move state change
logging code to new func to declutter nsm_change_state and
ospf_nsm_event.
Log event with AdjChange, it's useful to know.
(nsm_change_state) move adjchange and snmp logging to previous.
(ospf_nsm_event) call nsm_notice_state_changes from here.
Move the debug message to entry of function, so it gets out
even if something goes wrong.
Paul Jakma [Tue, 11 Jul 2006 17:49:22 +0000 (17:49 +0000)]
[ospfd] Remove nsm_reset_nbr and a bunch of useless NSM event/action functions
2006-07-10 Paul Jakma <paul.jakma@sun.com>
* ospf_nsm.c: (nsm_change_state) call nsm_clear_adj for all
adjacency losses, hence removing need for nsm_reset_nbr.
(nsm_reset_nbr) kill it, clear_adj in previous does fine.
(nsm_kill_nbr,nsm_start) remove nsm_reset_nbr call.
(ospf_nsm_event) Allow NSM function to be NULL, this along with
removal of nsm_reset_nbr, allows a bunch of now useless functiosn
to be removed.
Remove some useless variables.
(nsm_ignore) now useless, remove.
(nsm_bad_ls_req) ditto
(nsm_seq_number_mismatch) "
(nsm_oneway_received) "
(nsm_inactivity_timer) "
(nsm_ll_down) "
(NSM) replace removed action functions with NULL.
Andrew J. Schorr [Tue, 11 Jul 2006 00:06:49 +0000 (00:06 +0000)]
[lib] Do not call vty_close in vty_log_out to avoid possible free memory access
2006-07-10 Andrew J. Schorr <ajschorr@alumni.princeton.edu>
* vty.c: (vty_log_out) Do not call vty_close, because this could
result in a parent function's accessing the freed memory.
Instead, set status VTY_CLOSE and call shutdown(vty->fd, SHUT_RDWR).
And add a comment on vty_close.
Andrew J. Schorr [Mon, 10 Jul 2006 18:09:42 +0000 (18:09 +0000)]
[lib] Fix infinite recursion for errors on vtys with terminal monitor enabled
2006-07-10 Andrew J. Schorr <ajschorr@alumni.princeton.edu>
* vty.c: (vty_log_out, vty_read, vty_flush, vtysh_flush, vtysh_read)
After an I/O error, must set vty->monitor to 0 before calling
zlog_warn, otherwise an infinite recursion could occur
(since zlog_warn triggers a message to be written to the vty,
and that in turn triggers another error message when it fails, etc.).
Paul Jakma [Mon, 10 Jul 2006 07:45:13 +0000 (07:45 +0000)]
[ospfd] cleanup NSM neighbour delete through a new Deleted NSM state
2006-07-07 Paul Jakma <paul.jakma@sun.com>
* ospf_nsm.h: Add a NSM_Deleted neighbour state, to act as dummy
state indicating the neighbour is to be deleted.
* ospf_nsm.c: (general) Use the NSM_Deleted state to delete
neighbours, thus allowing code to be slightly more obvious
in its flow.
(nsm_timer_set) Add NSM_Deleted. Add another timer the code
missed.
(nsm_kill_nbr) No need for special case call to nsm_change_state
anymore.
Make the assert and error-handling for same case more readable
(Andrew Schorr)
Remove the call to ospf_nbr_delete, nsm_change_state can do
this generally now via NSM_Deleted.
(struct ... NSM) Add the dummy NSM_Deleted state, the 3 events
that can lead to nsm_kill_nbr all now transition the NBR to
NSM_Deleted and the general change_state function can be left
to do the work.
(ospf_nsm_event) Special casing of events and early-return can
be removed now.
On transition into Deleted, delete the nbr.
* ospf_dump.c: (ospf_nsm_state_msg) Add Deleted.
Paul Jakma [Thu, 6 Jul 2006 11:12:39 +0000 (11:12 +0000)]
[ospfd] early-return in nsm_event missed LLDown event, fixes use after free
2006-07-06 Paul Jakma <paul.jakma@sun.com>
* ospf_nsm.c: (ospf_nsm_event) LLDown event also results in nbr
being deleted, requires early-return too. Likely explains
some crash reports after interface events.
Paul Jakma [Tue, 4 Jul 2006 14:41:08 +0000 (14:41 +0000)]
[doc] remove the auto-generated quagga.info file from CVS
2006-07-04 Paul Jakma <paul.jakma@sun.com>
* quagga.info: remove auto-generated file. It will still be
present in dist tarballs, so shouldn't affect anyone but
direct users of CVS. Required texinfo version should be
widely available.
* .cvsignore: ignore quagga.info
Paul Jakma [Tue, 4 Jul 2006 13:57:49 +0000 (13:57 +0000)]
[ospfd] Apply some simplifications to ospf_nsm, from Andrew's suggestions
2006-07-04 Paul Jakma <paul.jakma@sun.com>
* ospf_nsm.c: (general) Various small cleanups from Andrew's
review of last set of patches.
(nsm_timer_set) Loading, Full and default can share
same code too.
(nsm_should_adj) Can just be one big OR.
(nsm_twoway_received) Collapse into return statement.
Paul Jakma [Tue, 4 Jul 2006 13:50:44 +0000 (13:50 +0000)]
[ospfd] Clear adjacency state for NSM changes down to TwoWay/ExStart
2006-07-02 Paul Jakma <paul.jakma@sun.com>
* ospf_nsm.c: (nsm_clear_adj) clear adjacency related state for a
neighbour, needed for some state changes from > ExStart down
to ExStart or less, which need not go through nsm_reset_nbr.
(nsm_reset_nbr) move code to former. Should be static.
(ospf_nsm_event) Remove long dead code.
Use nsm_clear_adj for state changes that take down
adjacencies to TwoWay/ExStart.
(nsm_kill_nbr) Oops, action function shouldn't try return
1 for error.
Paul Jakma [Tue, 4 Jul 2006 13:46:14 +0000 (13:46 +0000)]
[ospfd] trivial cleanup of nsm_timer_set
2006-07-02 Paul Jakma <paul.jakma@sun.com>
* ospf_nsm.c: (nsm_timer_set) ls_req timer should be OFF in
early states.
Compact several identical sections.
Set inactivity timer to OFF for Down, for documentary purposes.
Paul Jakma [Tue, 4 Jul 2006 13:44:19 +0000 (13:44 +0000)]
[ospfd] Ensure NSM state functions can redirect next_state accidently
2006-07-02 Paul Jakma <paul.jakma@sun.com>
* ospf_nsm.c: (ospf_nsm_event) Don't allow action functions to
change next_state if the NSM tables do not indicate
next_state is conditional, log warning if one tries -
existing code appears fine though.
Paul Jakma [Tue, 4 Jul 2006 13:35:24 +0000 (13:35 +0000)]
[ospfd] consolidate adjacency check logic
2006-07-02 Paul Jakma <paul.jakma@sun.com>
* ospf_nsm.c: (nsm_should_adj) New function, just consolidate the
10.4 adjacency check from nsm_twoway_received/nsm_adj_ok.
(nsm_twoway_received/nsm_adj_ok) Use former.
Paul Jakma [Sun, 2 Jul 2006 16:38:54 +0000 (16:38 +0000)]
[zebra] Fix CID #104, check addr for null, and #18, check nexthop type args
2006-07-02 Paul Jakma <paul.jakma@sun.com>
* rt_netlink.c: (netlink_interface_addr) Fix CID #104, can not
proceed if addr is NULL.
* zebra_rib.c: (static_add_ipv6) Fix CID #18, double check
required arguments are supplied for the given nexthop type.
Paul Jakma [Sun, 2 Jul 2006 11:01:50 +0000 (11:01 +0000)]
[bgpd] Fix crash on shutdown of peer
2006-07-02 Paul Jakma <paul.jakma@sun.com>
* bgp_fsm.c: (bgp_{stop,start}) Move clear/free of certain
bits of state from stop to start, as they may be used via
peer references on clearing queues..
Andrew J. Schorr [Fri, 30 Jun 2006 16:53:47 +0000 (16:53 +0000)]
[ospfd] Fix "show ip ospf neighbor A.B.C.D" to show all matches
2006-06-30 Andrew J. Schorr <ajschorr@alumni.princeton.edu>
* ospf_vty.c: (show_ip_ospf_neighbor_id) Should show all instances
of that neighbor (since it may appear on multiple interfaces)
instead of bailing out after showing the first match.
Andrew J. Schorr [Fri, 30 Jun 2006 03:21:05 +0000 (03:21 +0000)]
[ospfd] Avoid getting NSM stuck in ExStart by using local view of DR/BDR
2006-06-29 Andrew J. Schorr <ajschorr@alumni.princeton.edu>
* ospf_nsm.c: (nsm_twoway_received) When deciding whether to
change from state Init to ExStart, the test for whether the
neighboring router is DR or BDR should be against the
local router's notion of DR/BDR, not the neighbor's view.
Paul Jakma [Tue, 27 Jun 2006 07:52:03 +0000 (07:52 +0000)]
[lib] Centralise Zserv route type information, auto-generate redist strings
2006-05-23 Paul Jakma <paul.jakma@sun.com>
* route_types.txt: New file, table of ZEBRA_ROUTE definitions.
* route_types.awk: New script, to parse previous and generate
(for now) redistribute string defines.
* Makefile.am: build route_types.h using previous two, ala
memtypes.h, include the script and table file in EXTRA_DIST.
* command.h: pull in route_types.h, add a REDIST_STR define.
Paul Jakma [Mon, 26 Jun 2006 12:55:58 +0000 (12:55 +0000)]
[doc] OSPF MD5 keyid documentation, fix texinfo warnings in bgpd.texi
2006-06-26 Paul Jakma <paul.jakma@sun.com>
* ospfd.texi: Document that MD5 keyid is part of the protocol.
* bgpd.texi: shut texinfo warnings up by replacing brackets in
variable with angle brackets.
Paul Jakma [Mon, 26 Jun 2006 12:50:06 +0000 (12:50 +0000)]
[ospfd] NSSA translate-enabled ABR should declare itself as ASBR
2006-06-26 Paul Jakma <paul.jakma@sun.com>
* ospf_abr.c: (general) NSSA translate-candidate ABRs need to
be ASBRs, or other routers may rightfully refuse to install
translated type-5s LSAs. reported by dendroot@gmail.com.
(ospf_abr_nssa_check_status) Detect change in translator
state when ABR, and inc/dec redistribute count as when we
leave/enter the disabled state - so that translate-enabled
ABR properly sets ASBR bit on non-NSSA areas.
Run the resulting function through indent to clean it up.
* ospf_lsa.c: (router_lsa_flags) For purposes of ASBR bit,
NSSA area is same as stub area.
Andrew J. Schorr [Sun, 18 Jun 2006 00:45:48 +0000 (00:45 +0000)]
[ospfd] Fix ospf passive-interface subcommand to create interface if necessary
2006-06-17 Andrew J. Schorr <ajschorr@alumni.princeton.edu>
* ospf_vty.c: ({no_,}ospf_passive_interface) Replace if_lookup_by_name
with a call to if_get_by_name -- if the interface does not exist
already, it should be created. And remove the obsolete warning
message.
2006-06-15 Andrew J. Schorr <ajschorr@alumni.princeton.edu>
* redhat/quagga.spec.in: Get default distro automatically
by using rpm to query the fedora-release version.
And fix a typo (should be default_dist, not dist_default).
Paul Jakma [Thu, 15 Jun 2006 18:40:49 +0000 (18:40 +0000)]
[ospfd] Fix multicast membership drop bug
2006-06-15 Paul Jakma <paul.jakma@sun.com>
* Reported by Milan Koci
* ospf_interface.h: (struct ospf_if_info) Add reference counts
for multicast group memberships. Add various macros to help
manipulate/check membership state.
* ospf_interface.c: (ospf_if_set_multicast) Maintain the
ospf_if_info reference counts, and only actually drop
memberships if it hits 0, to avoid losing membership when
OSPF is disabled on an interface with multiple active OSPF
interfaces.
* ospf_packet.c: (ospf_{hello,read}) Use the new macros to
check/set
multicast membership.
* ospf_vty.c: (show_ip_ospf_interface_sub) ditto.