| Age | Commit message (Collapse) | Author |
|---|
|
2006-04-24 Paul Jakma <paul.jakma@sun.com>
* (general) More Virtual-link fixes, again with much help in
testing / debug from Juergen Kammer. Primarily in SPF.
* ospf_spf.h: Add guard. ospf_interface.h will include this
header.
* ospf_interface.h: Modify ospf_vl_lookup definition to take
struct ospf as argument, so as to allow for NULL area
argument.
(struct ospf_vl_data) Remove out_oi, instead add a struct
vertex_nexthop, to use as initial nexthop for backbone paths
through a vlink.
* ospf_interface.c: (ospf_vl_lookup) Modified to allow
NULL area to be passed to indicate "any" (first) area.
Add extra debug.
(ospf_vl_set_params) vl_oi -> nexthop. Add extra debug.
(ospf_vl_up_check) Fix debug, inet_ntoa returns a static
buffer..
* ospf_route.c: (ospf_intra_add_router) Vlinks dont go through
backbone, don't bother checking.
* ospf_spf.c: (static struct list vertex_list) Record vertices
that will need to be freed.
(cmp) Order network before router vertices, as required,
wasn't implemented.
(vertex_nexthop_free) Mild additional robustness check.
(vertex_parent_free) Take void argument, as this function
is passed as list deconstructor for vertex parent list.
(ospf_vertex_new) More debug. Set deconstructor for parent
list. Track allocated vertices on the vertex_list.
(ospf_vertex_free) Get rid of the tricky recursive cleanup of
vertices. Now frees only the given vertex.
(ospf_vertex_add_parent) Fix assert.
(ospf_nexthop_calculation) Fix calculation of nexthop for
VLink vertices, lookup the vl_data and use its previously
recorded nexthop information.
(ospf_spf_calculate) Vertices are freed simply by deleting
vertex_list nodes and letting ospf_vertex_free as deconstructor
work per-node.
(ospf_spf_calculate_timer) Trivial optimisation, leave
backbone SPF calculation till last to reduce SPF churn on
VLink updates.
* ospf_vty.c: (ospf_find_vl_data) update call to ospf_vl_lookup
(no_ospf_area_vlink_cmd) ditto.
(show_ip_ospf_interface_sub) For Vlinks, the peer address is
more interesting than the output interface.
|
|
2006-04-28 Andrew J. Schorr <ajschorr@alumni.princeton.edu>
* ripd.c: (rip_update_process) Try to fix the logic for sending
an updated on each connected network. The new code will
attempt to send the update on each connected network, whereas
the previous code seemed to be attempting to avoid sending
more than one RIPv1 update on a given interface, but was coded
incorrectly. The actual effect of the old code was to send
an update only on the first connected address in the cases
where the interface is not multicast, or RIPv2 is not being used.
|
|
2006-04-03 Paul Jakma <paul.jakma@sun.com>
* quagga.xml.in: oops, ipvX-routing services don't exist yet in
any Solaris builds.
|
|
2006-04-03 Paul Jakma <paul.jakma@sun.com>
* (general) Fix issues with handling of Vlinks and entries
in the nbrs route-table which were highlighted by the
nsm/nbr_self fixes from bug #234. Many thanks to Juergen
Kammer for his help and efforts in testing out debug patches to
pinpoint the issue.
* ospf_interface.c: (ospf_vl_new) Add nbr_self for Vlink.
* ospf_neighbor.c: (ospf_nbr_key) new static function, helper
to create key in nbrs table for a given nbr.
(ospf_nbr_delete) Use ospf_nbr_key. Add an assert() to
document an expected state.
(ospf_nbr_add_self) Ditto.
(ospf_nbr_lookup_by_addr) Add an assert.
* ospf_nsm.c: (nsm_kill_nbr) Can never kill the nbr_self
psuedo-neighbour.
|
|
2006-03-31 Paul Jakma <paul.jakma@sun.com>
* fig*.txt: New files, txt versions of the diagrammes for the
info file.
* protocol.texi: Document the new version 1 header.
* routeserver.texi: Explicitly specify scaling widths for the
figures, only way to make them look reasonable, sadly.
Try fix some formatting and explicitely note lists as being
bulleted.
* Makefile.am: convert seems to do png to pdf just fine. No
need for intermediate eps.
Add txt figure expansions and list as dependencies and
EXTRA_DIST files.
|
|
|
|
2006-03-14 Paul Jakma <paul.jakma@sun.com>
* privs.c: (zprivs_caps_init) Change user IDs before lowering
privileges, while this seems to work on Linux, on Solaris
it rightfully refuses due to PRIV_PROC_SETID having been
dropped.
* command.h: Add the struct host global exported from command.c
|
|
2006-03-30 Paul Jakma <paul.jakma@sun.com>
* workqueue.c: (work_queue_run) fix line length of comment
|
|
2006-03-30 Paul Jakma <paul.jakma@sun.com>
* command.h: (DEFUN_CMD_FUNC_TEXT) Annotate arguments as
potentially being unused.
|
|
2006-03-30 Paul Jakma <paul.jakma@sun.com>
* bgp_community.c: (community_gettoken) Unknown token should
return NULL, to give a strong indication to callers that
the token no longer can be parsed, otherwise callers looping
on this function may have a hard time ending their loop.
(community_str2com) While loop around community_gettoken appears
to have been coded thinking that break statement would break
from the while{}, hence it could never exit for unknown token
case. Fix it to do..while, so it can use the NULL result from
community_gettoken easily.
|
|
2006-03-30 Paul Jakma <paul.jakma@sun.com>
* TODO: Add reminder for useful MED functionality we should
implement.
|
|
2006-03-27 Paul Jakma <paul.jakma@sun.com>
* ospf_lsa.c: (ospf_lsa_checksum) Add an explicit cast to avoid
the ambiguities of ANSI and C99 C with respect to type
conversion. Detailed problem report and test case with
example data supplied by Dmitry Ivanov <dimss@telecentrs.lv>.
|
|
2006-03-27 Paul Jakma <paul.jakma@sun.com>
* memtypes.awk: Fix gensub call, g should be a string..
|
|
2006-03-25 Paul Jakma <paul.jakma@sun.com>
* workqueue.h: (struct work_queue) Remove status field and
state flag, no longer used.
|
|
2006-03-25 Paul Jakma <paul.jakma@sun.com>
* ospf_interface.c: (ospf_if_lookup_recv_if) Ignore loopbacks,
we can never ever receive packets on those. Should fix
case where CARP is run with address in same subnet as real
interface. Problem report and diagnosis thanks to:
Landon Fuller <landonf@opendarwin.org>.
However, ospf_read() still can't deal deterministically with
multiple interfaces in same subnet.
|
|
2006-03-23 Steve Lawson <steve.lawson@aheadcomusa.com>
* ospf_lsa.c: (ospf_lsa_install) Fix incorrect byte-order
conversion of OSPF_MAX_SEQUENCE_NUMBER
|
|
2006-03-22 Paul Jakma <paul.jakma@sun.com>
* bgpd.c: (peer_free) release the per-peer workqueue when
freeing the peer.
|
|
2006-03-19 Paul Jakma <paul.jakma@sun.com>
* memtypes.c: Add MTYPE_BGP_SYNCHRONISE.
* bgp_advertise.c: Use a distinct memory type for struct
bgp_synchronize.
|
|
2006-03-19 Paul Jakma <paul.jakma@sun.com>
* bgpd/bgp_vty.c: Add includes to get several structs we want
to provide usage statistics on.
(show_bgp_memory_cmd) Show memory usage stats for various
notable fixed size objects. Using mtype_stats_alloc and
mtype_memstr recently added to memory.c.
(bgp_show_summary) Report some additional stats specific to
the given BGP instance and/or AFI/SAFI such as table counts,
peers, rsclients and peer-groups.
(bgp_vty_init) Install show_bgp_memory_cmd.
|
|
2006-03-19 Paul Jakma <paul.jakma@sun.com>
* bgp_nexthop.h: Include if.h as a dependent header, for struct
connected.
|
|
2006-03-16 Paul Jakma <paul.jakma@sun.com>
* Makefile.am: Fix -version-info argument.
|
|
2006-02-15 Paul Jakma <paul.jakma@sun.com>
* configure.ac: Check for mallinfo, being careful to link test
so we can detect things like umem being used (which doesn't
provide a mallinfo).
* lib/memory.c: (mtype_memstr) new helper function to
return human friendly string for a byte count.
(mtype_stats_alloc) new function, for users to retrieve
number of objects allocated.
(show_memory_mallinfo) New function, show mallinfo statistics
if available.
(show_memory_all_cmd) Call show_memory_mallinfo, if mallinfo
is available.
* lib/memory.h: Export mtype_memstr and mtype_stats_alloc.
Provide a define for a reasonable buffer size for
mtype_memstr.
|
|
2006-03-16 Paul Jakma <paul.jakma@sun.com>
* prototype.smf.in: Oops, continue to install manifest to
var/svc/manifest/network for now.
* Changelog: fix lack of trailing newline
|
|
2006-03-16 Paul Jakma <paul.jakma@sun.com>
* heavy-wq.c: (heavy_wq_init) delay is gone from workqueue
spec.
* aspath_test.c: (test_segments) Add an AS_SET with redundant
ASNs.
({empty_prepend,prepend,parse}_test) add static qualifiers
|
|
2006-03-13 Alan Maguire <alan.maguire@sun.com>
* quagga.xml.in: Change the naming of services and instances.
network/routing/quagga:<daemon> may be neat and tidy, but
it's at odds with SMF convention elsewhere that demands
the service be named for what it does, and the instance name
describe the implementation (e.g. network/smtp:sendmail).
Remove call to method script for 'stop', SMF can do that
itself.
Enumerate the privileges required in the SMF manifest, with the
method_credential element.
Dont try provide seperate properties for each argument, it's
just tedious, particularly when they can no longer be
inherited from a common quagga service - use a single
"daemon-args" property, defaulting to -P 0.
Specify authorisation for the 'routeadm' utility and for
RBAC.
2006-03-13 Paul Jakma <paul.jakma@sun.com>
* quagga.init.in: Update to match SMF manifest changes.
If run from an SMF environment, assume method is start.
Offer a more tradition init script interface for pre-SMF.
Enable zebra for the link-state daemons.
Don't allow daemons to start in non-global zones, except
for bgpd - it doesn't have to care about underlying
networking much.
Remove all the svcprop calls, SMF or the user will supply any
arguments on the command-line.
* depend.smf.in: Add dependency on SUNWroute, which provides
routeadm, which provides /var/svc/manifest/network/routing
* prototype.smf.in: Install manifest to
var/svc/manifest/network/routing.
|
|
2006-02-09 Paul Jakma <paul.jakma@sun.com>
* rib.h: (struct {rib,nexthop}) Rearrange fields to avoid
needless padding.
(struct rib) Remove the indirect pointer, not used anywhere!
|
|
2006-03-14 Paul Jakma <paul.jakma@sun.com>
* privs.c: (zprivs_caps_init) Change user IDs before lowering
privileges, while this seems to work on Linux, on Solaris
it rightfully refuses due to PRIV_PROC_SETID having been
dropped.
|
|
2006-03-12 Paul Jakma <paul.jakma@sun.com>
* bgp_attr.h: (struct attr) rearrange fields to avoid
wasted padding between them as much as possible.
(attr_count,attr_unknown_count) export new functions to
return number of counts of cached attributes.
* bgp_attr.c: (attr_count,attr_unknown_count) new functions to
return number of counts of cached attributes.
* bgp_route.h: (struct bgp_info) rearrange fields to avoid
wasted padding.
* bgp_table.h: (struct bgp_table) Add a count field, of number
of nodes in the table.
(struct bgp_node) rearrange fields to avoid
wasted padding between them, though I don't think there
was any in this case.
* bgp_table.c: (bgp_node_{delete,get}) Maintain the table node count.
(bgp_table_count) new function to access the table count.
|
|
2006-03-06 Paul Jakma <paul.jakma@sun.com>
* if.h: export show_address_cmd, for anyone who wishes to use
it.
|
|
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.
|
|
2006-03-03 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>
* sockunion.c: (sockunion_{su2str,log}) Use XSTRDUP.
Particularly with _su2str, as that string gets XFREEd,
which can be annoying if run debug code in memory.c.
|
|
2006-02-21 Paul Jakma <paul.jakma@sun.com>
* sockunion.c: (sockunion_{su2str,log}) Use XSTRDUP.
Particularly with _su2str, as that string gets XFREEd,
which can be annoying if run debug code in memory.c.
|
|
2006-02-21 Paul Jakma <paul.jakma@sun.com>
* bgp_vty.c: (bgp_show_summary) Add a 'Clrng' sub-description
to state of peer while it's suppressed due to clearing.
|
|
2006-02-21 Paul Jakma <paul.jakma@sun.com>
* bgp_vty.c: (bgp_show_summary) Add a 'Clrng' sub-description
to state of peer while it's suppressed due to clearing.
|
|
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.
|
|
2006-02-21 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.
|
|
2006-02-19 Paul Jakma <paul.jakma@sun.com>
* depend.*: renamed to depend.*.in.
* depend.*.in: Add version dependency.
* Makefile.am: Generate depend.* from depend.*.in
Distribute the depend.*.in files rather than the depend.*
files.
* quagga.init.in: Enable zebra if one of the link-state
daemons is started.
Restart method is not required - SMF handles that.
* quagga.xml.in: Remove restart method details.
|
|
2006-02-19 Paul Jakma <paul.jakma@sun.com>
* quagga.info: update auto-built file.
* ChangeLog: Fix old, existing entry for snmptrap.texi addition to
credit the author, who got in touch with me.
* snmptrap.texi: Add comment line with author's details.
|
|
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.
|
|
2006-02-18 Paul Jakma <paul.jakma@sun.com>
* bgp_routemap.c: (route_set_community) Quick, very hacky, fix
for the set-community leak, bug #89. True fix will be to
detangle the web of *_intern caching and provide saner object
caching for Quagga, future work.
|
|
2006-02-05 Paul Jakma <paul.jakma@sun.com>
* bgp_aspath.c: (aspath_gettoken) fix gcc warning about
possible uninitialised usage.
|
|
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.
|
|
2006-02-02 Paul Jakma <paul.jakma@sun.com>
* interface.c: (if_dump_vty) move flags to their line, neater.
|
|
2006-01-31 Paul Jakma <paul.jakma@sun.com>
* configure.ac: Cleanup the hideous {net,ucd}-snmp section
by removing ucd-snmp. Hence fixing detection where
net-snmp is installed in /usr/local (Boris Kovalenko reported
the problem).
|
|
over connected routes
|
|
2006-01-30 Paul Jakma <paul.jakma@sun.com>
* zebra_rib.c: (rib_process) Fourth time lucky on this jinxed
commit, last commit had a hole that could allow connected
route selection to escape beyond the connected route logic.
This time I cross-checked with Gunnar first. ;)
|
|
2006-01-25 Gunnar Stigen <gunnar.stigen@axxessit.no>
* zebra_rib.c: (rib_process) Application of Gunnar's earlier
metric selection RIB change included incorrect tidy-ups made
by commiter. Fix. (NB: any errors here are again due to paul).
|
|
2006-01-25 Paul Jakma <paul.jakma@sun.com>
* (general) More solaris PF_ROUTE hacks. The IFF_UP mangling
for solaris was incomplete on the PF_ROUTE side. fix it.
This changeset generally uglifies things. For some future
work I'd like to see the state changes seperated out from
the details of the code. Differences between systems might
then be slightly easier to implement without convoluted
hacks.
Changes should be specific to Solaris mostly, however
also tested on FreeBSD 6.
* if_ioctl_solaris.c: (interface_list_ioctl) ignore ~IFF_UP
interfaces, we'll hear about them when/if interface goes up
through NEWADDR.
Update flags explicitely at end of it to kick mangling.
* ioctl_solaris.c: (if_mangle_up) removed to interface.c, in
kind.
(lifreq_set_name) more convenient to take the string, than
the ifp.
(if_get_flags_direct) new convenience function, returns
the actual flags. Used during bootstrap in if_ioctl_solaris.c
to peek at flags of logical interfaces to see whether or
not to ignore them.
(if_get_flags) ENXIO means it's gone, poke out IFF_UP and
kick flags update.
(if_{un,}set_flags) flags argument should be 64bit.
* ioctl.{c,h}: flags argument should be 64bit.
* interface.h: Add a 'primary_state' flag to struct zebra_if on
SUNOS_5.
Export if_flags_update.
* interface.c: (if_flags_mangle) moved over in kind from
ioctl_solaris.c. Nasty kludge to try get IFF_UP right, as
much as is possible. Also keep track of the actual IFF_UP
value for the primary interface, so we can know when the ifp
must be deleted.
(if_flags_update) Take a new interface flags value, apply it
to the interface, and take whatever actions are required due
to flag transitions.
(if_refresh) flag state change logic is moved out to
previous. Just call if_get_flags, which will end up using
previous to effect the update of flags.
(if_flag_dump_vty) IFF_IPV{4,6} aren't interesting, VIRTUAL
and NOXMIT are though.
* kernel_socket.c: (ifm_read) Down->Down transitions shouldn't
create ifp, for non-IFANNOUNCE systems.
Use if_flags_update to update flags.
flag transition logic is now handled automatically through
if_flags_update.
(ifam_read) Better to call if_refresh *after* adding
connected addresses, as connected count affects IFF_UP on
IFF_UP-mangled systems.
On Solaris, Up->Down due to DELADDR means we need to delete
the ifp - the IFINFO might already have been and gone.
* rt.h: include other dependent headers.
|
