| Age | Commit message (Collapse) | Author |
|---|
|
I've spent the last several weeks working on stability fixes to bgpd.
These patches fix all of the numerous crashes, assertion failures, memory
leaks and memory stomping I could find. Valgrind was used extensively.
Added new function bgp_exit() to help catch problems. If "debug bgp" is
configured and bgpd exits with status of 0, statistics on remaining
lib/memory.c allocations are printed to stderr. It is my hope that other
developers will use this to stay on top of memory issues.
Example questionable exit:
bgpd: memstats: Current memory utilization in module LIB:
bgpd: memstats: Link List : 6
bgpd: memstats: Link Node : 5
bgpd: memstats: Hash : 8
bgpd: memstats: Hash Bucket : 2
bgpd: memstats: Hash Index : 8
bgpd: memstats: Work queue : 3
bgpd: memstats: Work queue item : 2
bgpd: memstats: Work queue name string : 3
bgpd: memstats: Current memory utilization in module BGP:
bgpd: memstats: BGP instance : 1
bgpd: memstats: BGP peer : 1
bgpd: memstats: BGP peer hostname : 1
bgpd: memstats: BGP attribute : 1
bgpd: memstats: BGP extra attributes : 1
bgpd: memstats: BGP aspath : 1
bgpd: memstats: BGP aspath str : 1
bgpd: memstats: BGP table : 24
bgpd: memstats: BGP node : 1
bgpd: memstats: BGP route : 1
bgpd: memstats: BGP synchronise : 8
bgpd: memstats: BGP Process queue : 1
bgpd: memstats: BGP node clear queue : 1
bgpd: memstats: NOTE: If configuration exists, utilization may be expected.
Example clean exit:
bgpd: memstats: No remaining tracked memory utilization.
This patch fixes bug #397: "Invalid free in bgp_announce_check()".
This patch fixes bug #492: "SIGBUS in bgpd/bgp_route.c:
bgp_clear_route_node()".
My apologies for not separating out these changes into individual patches.
The complexity of doing so boggled what is left of my brain. I hope this
is all still useful to the community.
This code has been production tested, in non-route-server-client mode, on
a linux 32-bit box and a 64-bit box.
Release/reset functions, used by bgp_exit(), added to:
bgpd/bgp_attr.c,h
bgpd/bgp_community.c,h
bgpd/bgp_dump.c,h
bgpd/bgp_ecommunity.c,h
bgpd/bgp_filter.c,h
bgpd/bgp_nexthop.c,h
bgpd/bgp_route.c,h
lib/routemap.c,h
File by file analysis:
* bgpd/bgp_aspath.c: Prevent re-use of ashash after it is released.
* bgpd/bgp_attr.c: #if removed uncalled cluster_dup().
* bgpd/bgp_clist.c,h: Allow community_list_terminate() to be called from
bgp_exit().
* bgpd/bgp_filter.c: Fix aslist->name use without allocation check, and
also fix memory leak.
* bgpd/bgp_main.c: Created bgp_exit() exit routine. This function frees
allocations made as part of bgpd initialization and, to some extent,
configuration. If "debug bgp" is configured, memory stats are printed
as described above.
* bgpd/bgp_nexthop.c: zclient_new() already allocates stream for
ibuf/obuf, so bgp_scan_init() shouldn't do it too. Also, made it so
zlookup is global so bgp_exit() can use it.
* bgpd/bgp_packet.c: bgp_capability_msg_parse() call to bgp_clear_route()
adjusted to use new BGP_CLEAR_ROUTE_NORMAL flag.
* bgpd/bgp_route.h: Correct reference counter "lock" to be signed.
bgp_clear_route() now accepts a bgp_clear_route_type of either
BGP_CLEAR_ROUTE_NORMAL or BGP_CLEAR_ROUTE_MY_RSCLIENT.
* bgpd/bgp_route.c:
- bgp_process_rsclient(): attr was being zero'ed and then
bgp_attr_extra_free() was being called with it, even though it was
never filled with valid data.
- bgp_process_rsclient(): Make sure rsclient->group is not NULL before
use.
- bgp_processq_del(): Add call to bgp_table_unlock().
- bgp_process(): Add call to bgp_table_lock().
- bgp_update_rsclient(): memset clearing of new_attr not needed since
declarationw with "= { 0 }" does it. memset was already commented
out.
- bgp_update_rsclient(): Fix screwed up misleading indentation.
- bgp_withdraw_rsclient(): Fix screwed up misleading indentation.
- bgp_clear_route_node(): Support BGP_CLEAR_ROUTE_MY_RSCLIENT.
- bgp_clear_node_queue_del(): Add call to bgp_table_unlock() and also
free struct bgp_clear_node_queue used for work item.
- bgp_clear_node_complete(): Do peer_unlock() after BGP_EVENT_ADD() in
case peer is released by peer_unlock() call.
- bgp_clear_route_table(): Support BGP_CLEAR_ROUTE_MY_RSCLIENT. Use
struct bgp_clear_node_queue to supply data to worker. Add call to
bgp_table_lock().
- bgp_clear_route(): Add support for BGP_CLEAR_ROUTE_NORMAL or
BGP_CLEAR_ROUTE_MY_RSCLIENT.
- bgp_clear_route_all(): Use BGP_CLEAR_ROUTE_NORMAL.
Bug 397 fixes:
- bgp_default_originate()
- bgp_announce_table()
* bgpd/bgp_table.h:
- struct bgp_table: Added reference count. Changed type of owner to be
"struct peer *" rather than "void *".
- struct bgp_node: Correct reference counter "lock" to be signed.
* bgpd/bgp_table.c:
- Added bgp_table reference counting.
- bgp_table_free(): Fixed cleanup code. Call peer_unlock() on owner if
set.
- bgp_unlock_node(): Added assertion.
- bgp_node_get(): Added call to bgp_lock_node() to code path that it was
missing from.
* bgpd/bgp_vty.c:
- peer_rsclient_set_vty(): Call peer_lock() as part of peer assignment
to owner. Handle failure gracefully.
- peer_rsclient_unset_vty(): Add call to bgp_clear_route() with
BGP_CLEAR_ROUTE_MY_RSCLIENT purpose.
* bgpd/bgp_zebra.c: Made it so zclient is global so bgp_exit() can use it.
* bgpd/bgpd.c:
- peer_lock(): Allow to be called when status is "Deleted".
- peer_deactivate(): Supply BGP_CLEAR_ROUTE_NORMAL purpose to
bgp_clear_route() call.
- peer_delete(): Common variable listnode pn. Fix bug in which rsclient
was only dealt with if not part of a peer group. Call
bgp_clear_route() for rsclient, if appropriate, and do so with
BGP_CLEAR_ROUTE_MY_RSCLIENT purpose.
- peer_group_get(): Use XSTRDUP() instead of strdup() for conf->host.
- peer_group_bind(): Call bgp_clear_route() for rsclient, and do so with
BGP_CLEAR_ROUTE_MY_RSCLIENT purpose.
- bgp_create(): Use XSTRDUP() instead of strdup() for peer_self->host.
- bgp_delete(): Delete peers before groups, rather than after. And then
rather than deleting rsclients, verify that there are none at this
point.
- bgp_unlock(): Add assertion.
- bgp_free(): Call bgp_table_finish() rather than doing XFREE() itself.
* lib/command.c,h: Compiler warning fixes. Add cmd_terminate(). Fixed
massive leak in install_element() in which cmd_make_descvec() was being
called more than once for the same cmd->strvec/string/doc.
* lib/log.c: Make closezlog() check fp before calling fclose().
* lib/memory.c: Catch when alloc count goes negative by using signed
counts. Correct #endif comment. Add log_memstats_stderr().
* lib/memory.h: Add log_memstats_stderr().
* lib/thread.c: thread->funcname was being accessed in thread_call() after
it had been freed. Rearranged things so that thread_call() frees
funcname. Also made it so thread_master_free() cleans up cpu_record.
* lib/vty.c,h: Use global command_cr. Add vty_terminate().
* lib/zclient.c,h: Re-enable zclient_free().
|
|
* bgp_route.c: (bgp_{input,output}_filter) Log a debug warning if a route is
received or sent and a filter name is configured for a prefix, as or
distribute list but none is found - guaranteed configuration mistake.
|
|
* bgp_route.c: Was missing these commands.
|
|
* bgpd/bgp_damp.c: Make bgp_damp_reuse_time_vty() accept a buffer and
length, rather than returning a local var buffer whose contents can get
trounced. Remove duplicate BGP_UPTIME_LEN define.
* bgpd/bgp_damp.h: bgp_damp_reuse_time_vty() prototype change.
* bgpd/bgp_route.c: Provide bgp_damp_reuse_time_vty() with a buffer and
length. Remove duplicate BGP_UPTIME_LEN define.
This problem was noticed in 2005...
http://hibernia.jakma.org/~paul/patches/quagga-test.diff
...but the fix didn't make it into the code.
Signed-off-by: Chris Caputo <ccaputo@alt.net>
|
|
When a BGP instance is deleted with lots of routes and neighbors
it is possible for the peer rsclient queue to run after
bgp_delete has been called. This would lead to bgpd crashing,
see https://bugzilla.vyatta.com/show_bug.cgi?id=3436
The fix is to add reference counting to the BGP instance and defer
actual freeing until all references are gone.
This patch also fixes a memory leak where the self-reference
peer instance was being created but never freed.
The check in bgp_clear_route is no longer valid because it is possible
for it to be called when peer is in Deleted state during cleanup.
|
|
These variables are const.
|
|
Use Ansi-C prototypes rather than old K&R method of declaring
function without arguments
|
|
Simple conversion of XMALLOC/memset to XCALLOC
|
|
* bgpd/bgp_aspath.c: (aspath_make_str_count) "assert (len < str_size)" was
getting hit under certain 4-byte ASN conditions. New realloc strategy.
* bgpd/bgp_aspath.c: (aspath_key_make) const warning fix.
"%d" -> "%u" 4-byte ASN corrections. Prevent negative number when ASN is
above 2^31.
|
|
* bgp_route.c: (show_ip_bgp_view_rsclient_route_cmd) Add the missing "ip"
to the command string.
Signed-off-by: Paul Jakma <paul@quagga.net>
|
|
* lib/command.h: Add a RESTRICTED_NODE, intended for use with
anonymous, 'no login' vtys, to provide a subset of 'view' mode
commands.
* lib/command.c: Add RESTRICTED_NODE bits, nothing special, just
following VIEW_NODE.
* lib/vty.c: (vty_auth) enable authentication should fall back to
restricted/view node as appropriate.
(vty_create) init vty's to restricted/view node as appropriate,
for the 'no login' case.
(vty_{no_,}restricted_mode_cmd) config commands to enable
'anonymous restricted' in vty configuration.
(vty_config_write) 'anonymous restricted' config.
(vty_init) Install some commands to restricted mode, and the
'anonymous restricted' config commands into VTY_NODE.
* bgpd/*.c: Install some of the safe(r) BGP commands into
'restricted mode', i.e. lookup commands of non-sensitive data.
Useful with looking-glass route-servers.
|
|
2008-08-15 Paul Jakma <paul.jakma@sun.com>
* */*: Fix various problems flagged by Sun Studio compiler.
- '<qualifier> <storage>' obsolescent in declarations
- empty statements (';' after ALIAS definitions)
- implicit declarations (e.g printstack in lib/log.c)
- "\%" in printf string instead of "%%"
- loops that return on the first iteration (legitimately, but
compiler can't really know), e.g. bgp_routemap.c
- internal declarations which mask prototypes.
|
|
2008-07-22 Paul Jakma <paul.jakma@sun.com>
* HACKING: Document preference for compiler conditional code, over
cpp conditional.
* configure.ac: DISABLE_BGP_ANNOUNCE always should be defined.
* bgp_{packet,route,advertise}.c: change to compiler testing of
DISABLE_BGP_ANNOUNCE, rather than cpp.
2008-07-22 MIYAJIMA Mitsuharu <miyajima.mitsuharu@anchor.jp>
* bgp_packet.c: (bgp_update_packet_eor) Fix crash triggerable
if a bgpd was compiled with --disable-bgp-announce and if GR is
advertised by peer.
|
|
|
|
2007-08-27 Paul Jakma <paul.jakma@sun.com>
* bgp_route.c: (bgp_announce_check) Fix bug #398, slight
modification of Vladimir Ivanov's suggested fix - to keep
memory alloc conditional.
(bgp_process_announce_selected) Don't take struct attr as
argument, none of the callers need it and it needlessly
distances allocation from use.
Free the extended attr, the attr itself is on the stack.
Fix bad indentation.
* bgp_attr.c: (bgp_packet_attribute) Remove incorrect assert,
and adjust conditional to test attr->extra, diagnosis by
Vladimir Ivanov in bug #398.
2007-08-27 Vladimir Ivanov <wawa@yandex-team.ru>
* bgp_route.c: (bgp_announce_check_rsclient) copy of
ri->attr is no longer deep enough, due to addition of
attr->extra. It should use bgp_attr_dup, as
bgp_announce_check() does.
|
|
2007-07-31 Paul Jakma <paul.jakma@sun.com>
* (general) Support for draft-ietf-idr-as-pathlimit-03.
* bgp_attr.h: (struct attr) Add pathlimit struct
bgp_attr.c: (attr_str) Add BGP_ATTR_AS_PATHLIMIT string.
(attrhash_key_make) tally pathlimit too
(attrhash_cmp) cmp pathlimit attr
(bgp_attr_aspathlimit) New, parse AS_PATHLIMIT attr.
(bgp_attr_parse) ditto
(bgp_packet_attribute) Write out AS_PATHLIMIT when set
(bgp_dump_routes_attr) ditto
* bgp_route.h: (struct bgp_static) Add TTL field
* bgp_route.c: (bgp_announce_check) Drop paths that are over
their hop-count TTL before sending via EBGP.
Mangle ASN in pathlimit for confeds/private as best we can.
(bgp_static_update_{rsclient,main}) Add any configure pathlimit
information.
(bgp_pathlimit_update_parents) New, update atomic-aggr setting for
parents of an aspathlimit'ed static.
(bgp_static_set) Add TTL argument, for all the 'bgp network'
commands.
Call previous for TTL changed statics.
(bgp_static_unset) Call pathlimit_update_parents.
(various bgp network commands) Add 'pathlimit <0-255>' qualifier
to all the various forms, bar route-map - which can set ttl
itself.
* bgp_routemap.c: (general) Add support for 'set pathlimit ttl' and
'match pathlimit as'.
* doc/bgpd.texi: Document 'network ... pathlimit <ttl>'
|
|
2007-05-25 Paul Jakma <paul.jakma@sun.com>
* bgp_route.c: (bgp_default_originate) Sanity check added
previously was broken and always failed, thus this function
never could run, bug #370.
|
|
2007-05-03 Paul Jakma <paul.jakma@sun.com>
* bgp_route.h: (struct info) Move less frequently used
fields to a lazily allocated struct info_extra.
Export bgp_info_extra_get
* bgp_route.c: (bgp_info_extra_new) allocate extra
(bgp_info_extra_free) Free damp info and the info_extra.
(bgp_info_extra_get) Retrieve the info_extra of a struct
info, allocating as required.
(generally) adjust to use info->extra
* bgp_damp.c: (generally) use bgp_info_extra_get to access
dampinfo
* bgp_attr.h: Move rarely allocated attributes from struct attr
to a struct attr_extra, for a substantial saving in size of
struct attr.
* bgp_attr.c: (bgp_attr_extra_{new,free}), new, self-explanatory.
(bgp_attr_extra_get) Get the attr_extra for a given struct
attr, allocating it if needs be.
(bgp_attr_dup) Shallow copy the struct attr and its attr_extra.
(generally) adjust to know about attr->extra.
* bgp_debug.c: (bgp_dump_attr) ditto
* bgp_vty.c: (show_bgp_memory) print attr and info extra sizes.
* bgp_nexthop.c: (generally) adjust to know about attr->extra
and info->extra.
* bgp_{packet,routemap,snmp,zebra}.c: ditto
* lib/memtypes.c: Add MTYPE_ATTR_EXTRA and MTYPE_BGP_ROUTE_EXTRA
|
|
2007-04-08 Paul Jakma <paul.jakma@sun.com>
* bgp_route.c: (general) Same bug as fixed on 2006-11-28 by ajs for
bgp static routes, but for main BGP RIB: Quick 'flap' of routes
can cause bgpd to mistake a new route for a duplicate route,
due to presence of removed, but not yet processed BGP RIB entry.
(bgp_update_rsclient) Ignore REMOVED bgp_info for duplicate,
restore route instead.
(bgp_update_main) Ditto.
|
|
2007-02-22 Paul Jakma <paul.jakma@sun.com>
* bgp_fsm.c: (bgp_fsm_change_status) Handle state change into
clearing or greater here. Simpler.
(bgp_event) Clearing state change work moved to previous
* bgp_route.c: (bgp_clear_route_node) Clearing adj-in here
is too late, as it leaves a race between a peer being deleted
and an identical peer being configured before clearing
completes, leading to a crash.
Simplest fix is to clean peers Adj-in up-front, rather than
queueing such work.
(bgp_clear_route_table) Clear peer's Adj-In and Adj-Out
up-front here, rather than queueing such work.
Extensive comment added on the various bits of indexed data
that exist and how they need to be dealt with.
(bgp_clear_route) Update comment.
|
|
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.
|
|
2006-10-27 Paul Jakma <paul.jakma@sun.com>
* bgp_route.c: (bgp_table_stats) oops, u_intXX_t should be
uintXX_t
|
|
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.
|
|
emtpy table
2006-10-15 Paul Jakma <paul.jakma@sun.com>
* bgp_route.c: (bgp_table_stats_walker) NULL deref if table is
empty, bgp_table_top may return NULL, Coverity CID#73.
|
|
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.
|
|
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.. ;)
|
|
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.
|
|
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.
|
|
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
|
|
2006-09-03 Paul Jakma <paul.jakma@sun.com>
* 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.
|
|
2006-05-12 Paul Jakma <paul.jakma@sun.com>
* bgp_aspath.{c,h}: (aspath_print_vty) take a format string,
so as to reduce burden on callers, all in bgp_route.c
* bgp_route.c: (route_vty_out{,tmp}) Update to match
aspath_print_vty, simplifying checks needed to get spacing
right. CID #4,#5.
({damp,flap}_route_vty_out) Ditto, CID #9, #10
|
|
2006-05-12 Paul Jakma <paul.jakma@sun.com>
* bgp_route.c: (bgp_static_update_rsclient) Remove useless NULL
check, code already assumes bgp_static can not be NULL,
fixes CID #6.
|
|
2006-05-12 Paul Jakma <paul.jakma@sun.com>
* bgp_route.c: (bgp_static_update_main) Remove useless NULL
check, code already assumes bgp_static can not be NULL,
fixes CID #7.
|
|
2006-05-08 Paul Jakma <paul.jakma@sun.com>
* bgp_route.c: (bgp_afi_node_get) given table should never be
NULL, check/assert this.
(bgp_static_update) Bug #240. Rsclients should only be passed
the static update if they are configured for the afi,safi.
|
|
2006-05-04 Paul Jakma <paul.jakma@sun.com>
* bgp_route.c: (general) Fix logical bug in clearing, noted
by Chris Caputo in [quagga-users 6728] - clearing depended on
at least one route being added to workqueue, in order for
workqueue completion function to restart FSM. However, if no
routes are cleared, then the completion function never is
called, it needs to be called manually if the workqueue
didn't get scheduled.
Finally, clearing is per-peer-session, not per AFI/SAFI, so
the FSM synchronisation should be in bgp_clear_route_table.
(bgp_clear_route_table) Wrong place for FSM/clearing
synchronisation, move to..
(bgp_clear_route) FSM/clearing synchronisation should be
here.
If no routes were cleared, no workqueue scheduled, call
the completion func to ensure FSM kicks off again.
|
|
2006-03-03 Paul Jakma <paul.jakma@sun.com>
* bgp_route.c: (bgp_clear_node_complete) Doh. When clearing
is complete we need to kick off FSM again.
|
|
2006-02-21 Paul Jakma <paul.jakma@sun.com>
* bgpd.h: move the clear_node_queue to be peer specific.
Add a new peer status flag, PEER_STATUS_CLEARING.
* bgp_table.h: (struct bgp_table) Add fields to record afi,
safi of the table.
(bgp_table_init) Take afi and safi to create table for.
* bgp_table.c: (bgp_table_init) record the afi and safi.
* bgp_nexthop.c: Update all calls to bgp_table_init.
* bgp_vty.c: ditto.
* bgpd.c: ditto.
* bgp_fsm.c: (bgp_timer_set) dont bring up a session which is
clearing.
* bgp_route.c: (general) Update all bgp_table_init calls.
(bgp_process_{rsclient,main}) clear_node is serialised
via PEER_STATUS_CLEARING and fsm now.
(struct bgp_clear_node_queue) can be removed. struct bgp_node
can be the queue item data directly, as struct peer can be
kept in the new wq global user data and afi/safi can be
retrieved via bgp_node -> bgp_table.
(bgp_clear_route_node) fix to get peer via wq->spec.data,
afi/safi via bgp_node->bgp_table.
(bgp_clear_node_queue_del) no more item data to delete, only
unlock the bgp_node.
(bgp_clear_node_complete) only need to unset CLEARING flag
and unlock struct peer.
(bgp_clear_node_queue_init) queue attaches to struct peer
now. record peer name as queue name.
(bgp_clear_route_table) If queue transitions to active,
serialise clearing by setting PEER_STATUS_CLEARING rather
than plugging process queue, and lock peer while queue
active.
Update to pass only bgp_node as per-queue-item specific data.
|
|
2006-02-18 Paul Jakma <paul.jakma@sun.com>
* bgp_route.c: (bgp_announce_check) trivial, move declaration
of two local variables into the only block where they are
used, to aid the reader.
|
|
2006-02-05 Paul Jakma <paul.jakma@sun.com>
* bgp_route.h: Add BGP_INFO_COUNTED to track whether
prefix has been counted or not.
* bgp_route.c: (bgp_pcount_{inc,dec}rement) new helpers, to
centralise inc/dec of prefix-count,
(bgp_rib_remove) Remove pcount decrement, use helper.
(bgp_rib_withdraw) ditto, additionally use previous function
too.
(bgp_update_main) Use pcount helpers.
(bgp_clear_route_node) ditto, aslo REMOVED routes don't need
clearing.
|
|
2006-02-02 Paul Jakma <paul.jakma@sun.com>
* workqueue.h: (struct work_queue) Remove the delay field.
It served no purpose and just introduced bad behaviour.
Should be excised before its allowed to escape into 1.0.
This removes need for the 'flood' and runs_since_clear
fields.
* workqueue.c: (general) excise delay factor between queue
runs, hence the 'flood' crap too.. See above.
* bgp_route.c: (bgp_{clear_node,process}_queue_init) delay
field is removed from workqueue spec.
|
|
2006-01-19 Paul Jakma <paul.jakma@sun.com>
* (general) various miscellaneous compiler warning fixes.
Remove redundant break statements from switch clauses
which return.
return from main, not exit, cause it annoys SOS.
Remove stray semi-colons which cause empty-statement
warnings.
* zebra/main.c: (sighup) remove private declaration of external
function.
|
|
* zebra_rib.c: (rib_process) convert to new workqueue specs and
shut up gcc, which complains about cast from void via
function parameters, for some dumb reason. Do the cast
inside the function instead.
(rib_queue_qnode_del) ditto.
(rib_queue_init) no need for the casts anymore.
|
|
* Makefile.am: bgpd shouldn't list libgp's sources as its own.
Use LDADD.
* bgp_aspath.h:
(struct assegment) New struct, abstract representation of a
list of AS_PATH segments and the contained ASNs.
(struct aspath) Remove the raw-data related
fields, reference the abstract struct assegment instead.
Remove several other computed fields, it's just a
headache to maintain them and they're cheap to compute from
struct assegment.
(aspath_parse) parse a stream, not a pointer to raw data.
(aspath_count_{hops,confeds,size}) helpers to access
information formerly directly contained in struct aspath.
(aspath_snmp_pathseg) Helper for SNMP, BGP MIB wants
to be able to output hex representation of raw data.
* bgp_aspath.c: (general) partial-rewrite. Store aspath data
as an abstract singly-linked list of abstract segments,
rather than storing the raw data, and parsing it each and
every time. Remove several count/size fields which are cheap
to compute from the abstract segment structure.
(global) Include stream.h, needed for aspath_parse, and
others. Couple of helper macros added.
(struct assegment_header) Just the header, and only the
header.
(assegment_data_{new,free}) convenience functions for
AS_SEG_DATA allocation, the dynamic, per-segment array of
ASNs.
(assegment_{new,free,free_all,dup,dup_all}) convenience
functions for creating struct assegments. The _all forms will
follow the entire chain of segments from the given segment.
(assegment_prepend_asns) new function, prepend an ASN N times
to segment.
(assegment_append_asns) Append a list (array) of ASNs to
segment.
(int_cmp) convenience function for the aspath hash.
(assegment_normalise) new function. Normalise the given
segment chain to meet expectations of Quagga, and to
eliminate differing raw representations of the same paths.
Merge 'runs' of SEQUENCEs into one segment as our internal
segment is not limited by the protocol AS_PATH segment
length. Sort ASNs in SETs.
(aspath_new) Take void argument to quell warnings. Use the
assegment convenience functions.
(assegment_count_{asns,confeds,hops}) new functions to
compute at runtime values previously held in struct aspath.
(aspath_size) ditto.
(aspath_make_str_count) rewritten to stringify new
representation, and to be slightly easier to understand
hopefully.
(aspath_str_update) convenience function, update the aspath
str. Should investigate removing maintained string from
struct aspath, just run-time compute it, as per other fields.
It's just a maintenance headache, would save noticeable
amount of RAM with possibly not much extra run-time cost.
(aspath_dup) use the assegment dup functions.
(aspath_hash_alloc) Take void * argument to satisfy gcc. Use
the proper helper functions to dup data.
(assegments_parse) new function. parse raw AS_PATH data into
struct assegments. Normalise and return the head of the list.
(aspath_parse) Parse a stream, not pointer to raw data and
use assegments_parse to do it.
(assegment_data_put) Write out a single segment data in protocol
form to stream.
(assegment_header_put) ditto but for segment header.
(aspath_put) new function. As per previous but for an entire
struct aspath.
(aspath_snmp_pathseg) wrapper around aspath_put for
bgp_snmp.c. Uses a static buffer sadly.
(aspath_aggregate_as_set_add) rewritten to use assegments.
(aspath_aggregate) ditto
(aspath_{firstas,loop,private_as}_check) ditto
(aspath_{merge,prepend,add_one_as}) ditto
(aspath_cmp_left{_confed}) ditto
(aspath_delete_confed_seq) ditto, plus fixed to properly
delete all leading confed segments.
(aspath_as_add) Just use assegment_append_asns.
(aspath_segment_add) updated to use assegments.
(enum as_token) Add values for confeds
(aspath_gettoken) Add support for confeds
(aspath_str2aspath) ditto
(aspath_key_make) updated to use as_segments. Also, add
segment type into the hash value as appropriate.
(aspath_cmp) updated to use as_segments.
(aspath_print) don't segfault on NULL argument.
* bgp_attr.c: (bgp_attr_aspath) aspath_parse wants the stream
now. No need for manual forwarding of stream.
(bgp_packet_attribute) empty aspath is now denoted by NULL
segment field, length is gone.
Use aspath_size() to determine size.
(bgp_attr_init) Fix declaration, explicitely specify void
arg.
(bgp_dump_routes_attr) Use aspath_size() to determine size.
* bgp_route.c: (bgp_info_cmp) use the aspath_count_* functions.
(bgp_rib_withdraw) remove unused variable. Use
aspath_count_hops.
* bgp_snmp.c: (bgp4PathAttrTable) raw data is gone, use
aspath_snmp_pathseg to get the representation.
|
|
introduced long time ago with route server patch. Hopefully
it's last case to fix - route-server client not in peer group.
[backport candidate]
|
|
* bgp_route.h: (struct bgp_info) add a new flag, BGP_INFO_REMOVED.
BGP_INFO_VALID is already overloaded, don't care to do same thing
to STALE or HISTORY.
* bgpd.h: (BGP_INFO_HOLDDOWN) Add INFO_REMOVED to the macro, as a
route which should generally be ignored.
* bgp_route.c: (bgp_info_delete) Just set the REMOVE flag, rather
than doing actual work, so that bgp_process (called directly,
or indirectly via the scanner) can catch withdrawn routes.
(bgp_info_reap) Actually remove the route, what bgp_info_delete
used to do, only for use by bgp_process.
(bgp_best_selection) reap any REMOVED routes, other than the old
selected route.
(bgp_process_rsclient) reap the old-selected route, if appropriate
(bgp_process_main) ditto
(bgp_rib_withdraw, bgp_rib_remove) make them more consistent with
each other. Don't play games with the VALID flag, bgp_process
is async now, so it didn't make a difference anyway.
Remove the 'force' argument from bgp_rib_withdraw, withdraw+force
is equivalent to bgp_rib_remove. Update all its callers.
(bgp_update_rsclient) bgp_rib_withdraw and force set is same as
bgp_rib_remove.
(route_vty_short_status_out) new helper to print the leading
route-status string used in many command outputs. Consolidate.
(route_vty_out, route_vty_out_tag, damp_route_vty_out,
flap_route_vty_out) use route_vty_short_status_out rather than
duplicate.
(route_vty_out_detail) print state of REMOVED flag.
(BGP_SHOW_SCODE_HEADER) update for Removed flag.
|
|
* (global) The great bgpd extern and static'ification.
* bgp_routemap.c: remove unused ROUTE_MATCH_ASPATH_OLD code
(route_set_metric_compile) fix u_int32_t to ULONG_MAX comparison
warnings.
* bgp_route.h: (bgp_process, bgp_withdraw, bgp_update) export these
used by various files which had their own private declarations,
in the case of mplsvpn - incorrect.
|
|
* bgpd/(general) refcount struct peer and bgp_info, hence allowing us
add work_queues for bgp_process.
* bgpd/bgp_route.h: (struct bgp_info) Add 'lock' field for refcount.
Add bgp_info_{lock,unlock} helper functions.
Add bgp_info_{add,delete} helpers, to remove need for
users managing locking/freeing of bgp_info and bgp_node's.
* bgpd/bgp_table.h: (struct bgp_node) Add a flags field, and
BGP_NODE_PROCESS_SCHEDULED to merge redundant processing of
nodes.
* bgpd/bgp_fsm.h: Make the ON/OFF/ADD/REMOVE macros lock and unlock
peer reference as appropriate.
* bgpd/bgp_damp.c: Remove its internal prototypes for
bgp_info_delete/free. Just use bgp_info_delete.
* bgpd/bgpd.h: (struct bgp_master) Add work_queue pointers.
(struct peer) Add reference count 'lock'
(peer_lock,peer_unlock) New helpers to take/release reference
on struct peer.
* bgpd/bgp_advertise.c: (general) Add peer and bgp_info refcounting
and balance how references are taken and released.
(bgp_advertise_free) release bgp_info reference, if appropriate
(bgp_adj_out_free) unlock peer
(bgp_advertise_clean) leave the adv references alone, or else
call bgp_advertise_free cant unlock them.
(bgp_adj_out_set) lock the peer on new adj's, leave the reference
alone otherwise. lock the new bgp_info reference.
(bgp_adj_in_set) lock the peer reference
(bgp_adj_in_remove) and unlock it here
(bgp_sync_delete) make hash_free on peer conditional, just in
case.
* bgpd/bgp_fsm.c: (general) document that the timers depend on
bgp_event to release a peer reference.
(bgp_fsm_change_status) moved up the file, unchanged.
(bgp_stop) Decrement peer lock as many times as cancel_event
canceled - shouldnt be needed but just in case.
stream_fifo_clean of obuf made conditional, just in case.
(bgp_event) always unlock the peer, regardless of return value
of bgp_fsm_change_status.
* bgpd/bgp_packet.c: (general) change several bgp_stop's to BGP_EVENT's.
(bgp_read) Add a mysterious extra peer_unlock for ACCEPT_PEERs
along with a comment on it.
* bgpd/bgp_route.c: (general) Add refcounting of bgp_info, cleanup
some of the resource management around bgp_info. Refcount peer.
Add workqueues for bgp_process and clear_table.
(bgp_info_new) make static
(bgp_info_free) Ditto, and unlock the peer reference.
(bgp_info_lock,bgp_info_unlock) new exported functions
(bgp_info_add) Add a bgp_info to a bgp_node in correct fashion,
taking care of reference counts.
(bgp_info_delete) do the opposite of bgp_info_add.
(bgp_process_rsclient) Converted into a work_queue work function.
(bgp_process_main) ditto.
(bgp_processq_del) process work queue item deconstructor
(bgp_process_queue_init) process work queue init
(bgp_process) call init function if required, set up queue item
and add to queue, rather than calling process functions directly.
(bgp_rib_remove) let bgp_info_delete manage bgp_info refcounts
(bgp_rib_withdraw) ditto
(bgp_update_rsclient) let bgp_info_add manage refcounts
(bgp_update_main) ditto
(bgp_clear_route_node) clear_node_queue work function, does
per-node aspects of what bgp_clear_route_table did previously
(bgp_clear_node_queue_del) clear_node_queue item delete function
(bgp_clear_node_complete) clear_node_queue completion function,
it unplugs the process queues, which have to be blocked while
clear_node_queue is being processed to prevent a race.
(bgp_clear_node_queue_init) init function for clear_node_queue
work queues
(bgp_clear_route_table) Sets up items onto a workqueue now, rather
than clearing each node directly. Plugs both process queues to
avoid potential race.
(bgp_static_withdraw_rsclient) let bgp_info_{add,delete} manage
bgp_info refcounts.
(bgp_static_update_rsclient) ditto
(bgp_static_update_main) ditto
(bgp_static_update_vpnv4) ditto, remove unneeded cast.
(bgp_static_withdraw) see bgp_static_withdraw_rsclient
(bgp_static_withdraw_vpnv4) ditto
(bgp_aggregate_{route,add,delete}) ditto
(bgp_redistribute_{add,delete,withdraw}) ditto
* bgpd/bgp_vty.c: (peer_rsclient_set_vty) lock rsclient list peer
reference
(peer_rsclient_unset_vty) ditto, but unlock same reference
* bgpd/bgpd.c: (peer_free) handle frees of info to be kept for lifetime
of struct peer.
(peer_lock,peer_unlock) peer refcount helpers
(peer_new) add initial refcounts
(peer_create,peer_create_accept) lock peer as appropriate
(peer_delete) unlock as appropriate, move out some free's to
peer_free.
(peer_group_bind,peer_group_unbind) peer refcounting as
appropriate.
(bgp_create) check CALLOC return value.
(bgp_terminate) free workqueues too.
* lib/memtypes.c: Add MTYPE_BGP_PROCESS_QUEUE and
MTYPE_BGP_CLEAR_NODE_QUEUE
|
|
* bgp_aspath.[ch], bgp_route.c, bgp_vty.c, bgpd.[ch]: Allow to enable
the length of confederation path segments to be included during the
as-path length check in the best path decision.
|
|
* (global): Fix up list loops to match changes in lib/linklist,
and some basic auditing of usage.
* configure.ac: define QUAGGA_NO_DEPRECATED_INTERFACES
* HACKING: Add notes about deprecating interfaces and commands.
* lib/linklist.h: Add usage comments.
Rename getdata macro to listgetdata.
Rename nextnode to listnextnode and fix its odd behaviour to be
less dangerous.
Make listgetdata macro assert node is not null, NULL list entries
should be bug condition.
ALL_LIST_ELEMENTS, new macro, forward-referencing macro for use
with for loop, Suggested by Jim Carlson of Sun.
Add ALL_LIST_ELEMENTS_RO for cases which obviously do not need the
"safety" of previous macro.
LISTNODE_ADD and DELETE macros renamed to ATTACH, DETACH, to
distinguish from the similarly named functions, and reflect their
effect better.
Add a QUAGGA_NO_DEPRECATED_INTERFACES define guarded section
with the old defines which were modified above,
for backwards compatibility - guarded to prevent Quagga using it..
* lib/linklist.c: fix up for linklist.h changes.
* ospf6d/ospf6_abr.c: (ospf6_abr_examin_brouter) change to a single
scan of the area list, rather than scanning all areas first for
INTER_ROUTER and then again for INTER_NETWORK. According to
16.2, the scan should be area specific anyway, and further
ospf6d does not seem to implement 16.3 anyway.
|
|
no particular afi/safi configured.
[backport candidate]
|
