paul [Tue, 28 Jun 2005 17:17:12 +0000 (17:17 +0000)]
2005-06-28 Paul Jakma <paul.jakma@sun.com>
* (global) Extern and static'ification, with related fixups
of declarations, ensuring files include their own headers, etc.
if_ioctl.c: (interface_info_ioctl) fix obvious arg mis-order in
list loop
paul [Tue, 28 Jun 2005 12:44:16 +0000 (12:44 +0000)]
2005-06-28 Paul Jakma <paul.jakma@sun.com>
* (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.
paul [Tue, 28 Jun 2005 00:19:48 +0000 (00:19 +0000)]
2005-06-28 Paul Jakma <paul.jakma@sun.com>
* INSTALL.quagga.txt: GNU make is required now, because of manual
automatic rules in solaris/Makefile.am. (If someone knows how
to do these in a better way..).
GNU AWK is required for CVS checkout builds.
hasso [Fri, 24 Jun 2005 08:17:51 +0000 (08:17 +0000)]
* ospf6_message.c: Changed to be insensitive to changes of neighbors'
IP source address in the OSPF packets. It was sometimes problematic
in actual operation (needed some operational cost: restarting
all-neighbor routers when I/F NIC was changed). Due to this change,
a previously safe case, attaching multiple interface to the same
link will now be dengerous and will not work. Remedy to that should
be applied later.
hasso [Fri, 24 Jun 2005 07:50:12 +0000 (07:50 +0000)]
* ospf6_interface.c: fix the way inactivity_timer is called. Because
inactivity_timer() deletes the neighbor from the neighbor_list, it
cannot be called by thread_execute() from inner side of the
neighbor_list for-loop.
(Although crash was already fixed in Quagga, it's better follow the
GNU Zebra logic).
paul [Wed, 15 Jun 2005 19:15:35 +0000 (19:15 +0000)]
2005-06-15 Paul Jakma <paul.jakma@sun.com>
* bgpd.c: (bgp_terminate) workqueue's are lazy allocated and its
possible to terminate bgpd before workqueues were setup, causing
an abort/crash. Reported by Ashish Mehta of Sun.
paul [Wed, 15 Jun 2005 11:00:11 +0000 (11:00 +0000)]
2005-06-15 Paul Jakma <paul.jakma@sun.com>
* quagga.xml.in: Pass the FMRI to the method, removing need for
daemon_name property and deducing the FMRI.
remove the empty config_file and pid_file definitions, they just
make things difficult in the method script, cause it then has to
deal with svcprop returning "" for defined empty properties.
Remove daemon_name everywhere.
* quagga.init.in: Take the FMRI as an argument
Remove deducing the FMRI from the defunct daemon_name property.
Use svcprop -q to test for presence of a property first.
Default vty_port to 0 and vty_addr to 127.1 if equivalent
SMF properties are not set.
Deduce the pid_file, we can infer it from @quagga_statedir@ I
think, it's probably not useful to allow it to be configurable
as a property anyway.
/var/run/ is on tmpfs on Solaris, so we probably will need
to create @quagga_statedir@ first run after boot.
Use @sbindir@, not /usr/local/sbin.
paul [Tue, 14 Jun 2005 14:07:07 +0000 (14:07 +0000)]
2005-06-14 Paul Jakma <paul.jakma@sun.com>
* kernel_socket.c: consolidate the IFAM{ADDR,MASK}GET and
RTM{ADDR,MASK}GET macros into generic rta_addrs macros,
RTA_{ADDR,ATTR}_GET.
(af_check) could use 'inline' attribute
(ifam_read_mesg) remove IFAM{ADDR,MASK}GET macro, change to
generic macro.
(rtm_read_mesg) similar
paul [Mon, 13 Jun 2005 13:57:16 +0000 (13:57 +0000)]
2005-06-13 Paul Jakma <paul.jakma@sun.com>
* ospf_spf.c: Try get more information on a SEGV under
ospf_spf_vertex_add_parent.
(ospf_vertex_free) NULL out the child and nexthop lists
(ospf_vertex_add_parent) nexthop and child can not be NULL
vertex_nexthop's parent->child list can not be NULL
(ospf_spf_next) w and cw are per-loop iteration variables, move
declarations into loop body.
paul [Fri, 3 Jun 2005 18:01:50 +0000 (18:01 +0000)]
2005-06-03 Paul Jakma <paul.jakma@sun.com>
* ripd.c: (rip_create_socket) Make it static.
Remove the getservbyname stuff, as RFC2453 3.9.2 says non-RIP
port messages should be discarded, quagga doesnt accept them,
no need to lookup port.
Take a 'to' argument, if socket should be bound to something else.
setsockopt_so_recvbuf might need privs, move it to the raised
privileges section.
dont forget to close the socket if bind fails.
(rip_send_packet) use strncpy, just in case (address is under
our control anyway, but still).
dont duplicate rip_create_socket - just use it.
(rip_create) rip_create_socket takes an argument now, modify.
paul [Thu, 2 Jun 2005 16:33:53 +0000 (16:33 +0000)]
2005-06-02 Paul Jakma <paul.jakma@sun.com>
* quagga.xml.in: Fix mistake in bgpd definition. Change dependency
on zebra to optional. Remove the duplicated stability statements.
* quagga.init.in: svcprop should check PACKAGE_TARNAME, not _NAME.
no need to check for config file either, there's already a
dependency in the manifest.
paul [Thu, 2 Jun 2005 08:20:53 +0000 (08:20 +0000)]
2005-06-01 Paul Jakma <paul.jakma@sun.com>
* rip_interface.c: Fix authentication, no-auth impossible to specify
(rip_interface_new) default to RIP_NO_AUTH
(rip_interface_reset) ditto
(rip_interface_config_write) write out config for simple
paul [Wed, 1 Jun 2005 11:17:05 +0000 (11:17 +0000)]
2005-06-01 Paul Jakma <paul.jakma@sun.com>
* 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
hasso [Tue, 31 May 2005 10:24:28 +0000 (10:24 +0000)]
* ospf6d.c: No need for double ';'. Fixes parsing "show ipv6 ospf6
database ..." commands for vtysh by extract.pl. Remove duplicate
install_element calls.
paul [Tue, 31 May 2005 08:38:50 +0000 (08:38 +0000)]
2005-05-31 Paul Jakma <paul.jakma@sun.com>
* zserv.c: (zsend_route_multipath) Fix bug if route is sent
with no NEXTHOP_FLAG_FIB nexthops. As ZAPI_MESSAGE_IFINDEX
and ZAPI_MESSAGE_NEXTHOP are always set, clients would try
read non-existent nexthop information and hit stream assert.
Zserv is still broken for multi-nexthop messages, but it always was.
paul [Sun, 29 May 2005 11:27:24 +0000 (11:27 +0000)]
2005-05-29 Paul Jakma <paul@dishone.st>
* ripd.c: (rip_output_process) fix error which crept in my
previous rip auth untanglement commit - it had become impossible
to not have authentication (even for v1).
paul [Wed, 25 May 2005 12:21:13 +0000 (12:21 +0000)]
2005-05-25 Paul Jakma <paul@dishone.st>
* isisd.c: (show_isis_generated_topology) change to _RO version
of list macro. remove the extra listnode variable. one of the
macros had had incorrect number of arguments.
paul [Mon, 23 May 2005 13:42:46 +0000 (13:42 +0000)]
2005-05-23 Paul Jakma <paul@dishone.st>
* workqueue.h: Add a WQ_QUEUE_BLOCKED item_status return code,
to allow a queue function to indicate the queue is not
ready/blocked - rather than any problem with the item at hand.
Add a notion of being able to 'plug' and 'unplug' a queue.
Add helpers to plug/unplug a queue.
Add a completion callback, to be called when a queue is emptied.
* workqueue.c: (work_queue_new) remove useless list_free.
(work_queue_schedule) new internal helper function to schedule
queue, if appropriate.
(work_queue_add) use work_queue_schedule
(show_work_queues) Print 'P' if queue is plugged.
(work_queue_plug) new API function, plug a queue - ie prevent it
from 'drained' / processed / scheduled.
(work_queue_unplug) unplug a queue, allowing it to be drained
/ scheduled / processed again.
(work_queue_run) Add support for WQ_QUEUE_BLOCKED.
Add comment for RETRY_NOW case.
Make hysteris more aggresive in ramping up granularity, improves
performance significantly.
Add support for calling completion callback when queue is emptied,
possibly useful for knowing when to unplug a queue.
paul [Mon, 23 May 2005 12:43:34 +0000 (12:43 +0000)]
2005-05-23 Paul Jakma <paul@dishone.st>
* routemap.c: (rmap_onmatch_goto) fix crash if 'continue' command
is used, which does not supply an argv[0].
this is a backport candidate /iff/ the trailing ; is removed
from VTY_GET_INTEGER_RANGE
* vty.h: fix the VTY_GET macros, do {..} while(0) so they have
correct function like syntax in usage.
paul [Mon, 23 May 2005 12:33:58 +0000 (12:33 +0000)]
2005-05-23 Paul Jakma <paul@dishone.st>
* memtypes.awk: use character classes, which work correctly in
all LC_COLLATE environments, unlike A-Z, which doesnt work in
eg estonian collate order. Reported by Hasso.
paul [Thu, 19 May 2005 02:12:25 +0000 (02:12 +0000)]
2005-05-19 Paul Jakma <paul@dishone.st>
* bgp_network.c: (bgp_accept) use XSTRDUP
* bgpd.c: (peer_delete) XFREE the correct memtype, not free.
(peer_create) use XSTRDUP
* bgp_packet.c: (bgp_stream_dup) deleted, stream_dup should be used
(various) update -> s/bgp_stream_dup/stream_dup
paul [Thu, 19 May 2005 01:50:11 +0000 (01:50 +0000)]
2005-05-19 Paul Jakma <paul@dishone.st>
* bgp_fsm.c: (bgp_stop) use sockunion_free, not XFREE..
* bgp_network.c: (bgp_getsockname) ditto
* bgp_routemap.c: (route_match_peer) ditto, als use a ret value and
remove one sockunion_free.
* bgpd.c: (peer_delete) ditto
paul [Thu, 19 May 2005 01:30:53 +0000 (01:30 +0000)]
2005-05-19 Paul Jakma <paul@dishone.st>
* thread.c: (thread_cancel_event) the number of pending events
cancelled is potentially useful information, dont throw it away,
pass it back to the caller.
paul [Wed, 18 May 2005 23:29:57 +0000 (23:29 +0000)]
2005-05-19 Paul Jakma <paul.jakma@sun.com>
* ospf_interface.c: (ospf_if_table_lookup) Fix a serious bug
a less serious one.
1: this function is supposed to lookup
entries in the oifs ospf_interface route_table and return either
an existing oi or NULL to indicate not found, its caller depends
on this, yet this function uses route_node_get which /always/
returns a route_node - one is created if none exists. Use
route_node_lookup instead. This should fix root cause of the
reports of the (ospf_add_to_if) assert being hit.
2: oi's are inserted into this table with prefixlength set to
/32 (indeed, it should be a hash table, not a route_table),
however prefixlength to lookup was not changed, if no valid entry
can be inserted other than /32, then nothng but /32 should be
looked up. This possibly only worked by fluke..
Fix confirmed by 2 reporters (one list, one IRC), definitely a
backport candidate once it has been incubated in HEAD for a while.
Thanks to Patrick Friedel and Ivan Warren for testing.
paul [Sun, 15 May 2005 14:25:08 +0000 (14:25 +0000)]
2005-05-15 Paul Jakma <paul@dishone.st>
* getopt.h: It's not just __GNU_LIBRARY__ which defines
getopt, eg __EXTENSIONS__ does too on SunOS. It still seems
awfully fragile though.
* getopt.c: include zebra.h after config.h, before including
getopt.h so that things at least are consistent..
* getopt1.c: ditto
paul [Fri, 13 May 2005 20:11:53 +0000 (20:11 +0000)]
- revert the commenting of touch doc/quagga.info, wont work because
defines.texi is autogenerated every configure time so it will always
be newer than quagga.info, and hence require quagga.info to be built. drat.
paul [Fri, 13 May 2005 07:26:07 +0000 (07:26 +0000)]
2005-05-13 Paul Jakma <paul.jakma@sun.com>
* prototype.daemons.in: Move the Solaris 10 SMF specific stuff to
a seperate package.
* depend.smf: dependencies for smf package
* pkginfo.smf.tmpl.in: pkginfo for smf
* prototype.smf.in: prototype for smf
* Makefile.am: Add smf to pkg_names
paul [Fri, 13 May 2005 07:15:35 +0000 (07:15 +0000)]
2005-05-13 Paul Jakma <paul.jakma@sun.com>
* quagga.xml.in: Sigh, XML child elements are order-sensitive.
* prototype.doc.in: info/dir doesnt get created always.. comment
out for now, likely needs to be done with an install method.
paul [Fri, 13 May 2005 07:09:50 +0000 (07:09 +0000)]
2005-05-13 Paul Jakma <paul.jakma@sun.com>
* configure.ac: Comment out the touch doc/quagga.info hack. I'm
hoping it's not needed anymore, cause it breaks out of tree builds
which i've become fond of.. Plus I think we sorted out most of
the original problems (updated texinfo.tex, moved the version
info to an autogenerated by automake version.texi). Dist users
most definitely wont be affected.
paul [Wed, 11 May 2005 18:09:59 +0000 (18:09 +0000)]
2005-05-11 Paul Jakma <paul.jakma@sun.com>
* (general) Fix memory leaks in opaque AS-scope LSAs, reported and
with much debugging done by by scott collins <scollins@agile.tv>.
(possible backport candidate?)
* ospf_lsa.c: (ospf_discard_from_db) dont call
ospf_ase_unregister_external_lsa for opaque-lsa's, opaques are
never registered with ase in the first place.
* ospf_packet.c: (general) Disabuse opaque related code of its
tendency to try gather up things into temporary lists.
(ospf_ls_upd) remove the temporary lists opaque uses, call
opaque functions inline, just like all other types.
(ospf_ls_ack) ditto.
(ospf_recv_packet) fixup sign warning.
* ospf_opaque.c: (general) fix the unneeded use of lists, and
untwist some of the logic.
(ospf_opaque_self_originated_lsa_received) take a single LSA
as argument, not a list of them. Remove the list loop. Logic
otherwise unchanged.
(ospf_opaque_ls_ack_received) Mostly ditto. But untwist the logic,
move the actions up into the switch block, remove the goto's and
sanitise the logic near the end a bit.
* ospf_opaque.h: Adjust definitions of aforementioned functions
in ospf_opaque.c to match.
paul [Sat, 7 May 2005 02:22:51 +0000 (02:22 +0000)]
2005-05-07 Yar Tikhiy <yar@comp.chem.msu.su>
* configure.ac: Check for OSes which support passing ifindex in
struct ip_mreq.
* lib/sockopt.c: Add support for BSD style ifindex in ip_mreq.
* ospfd/ospf_network.c: Log ifindex on multicast membership leave/join
events.
paul [Fri, 6 May 2005 21:37:42 +0000 (21:37 +0000)]
2005-05-06 Paul Jakma <paul.jakma@sun.com>
* (general) extern and static qualifiers added.
unspecified arguments in definitions fixed, typically they should
be 'void'.
function casts added for callbacks.
Guards added to headers which lacked them.
Proper headers included rather than relying on incomplete
definitions.
gcc noreturn function attribute where appropriate.
* ospf_opaque.c: remove the private definition of ospf_lsa's
ospf_lsa_refresh_delay.
* ospf_lsa.h: export ospf_lsa_refresh_delay
* ospf_packet.c: (ospf_make_md5_digest) make *auth_key const,
correct thing to do - removes need for the casts later.
* ospf_vty.c: Use vty.h's VTY_GET_INTEGER rather than ospf_vty's
home-brewed versions, shuts up several warnings.
* ospf_vty.h: remove VTY_GET_UINT32. VTY_GET_IPV4_ADDRESS and
VTY_GET_IPV4_PREFIX moved to lib/vty.h.
* ospf_zebra.c: (ospf_distribute_list_update_timer) hacky
overloading of the THREAD_ARG pointer should at least use
uintptr_t.
paul [Fri, 6 May 2005 21:25:49 +0000 (21:25 +0000)]
2005-05-06 Paul Jakma <paul@dishone.st>
* (general) extern and static'ification of functions in code and
header.
Cleanup any definitions with unspecified arguments.
Add casts for callback assignments where the callback is defined,
typically, as passing void *, but the function being assigned has
some other pointer type defined as its argument, as gcc complains
about casts from void * to X* via function arguments.
Fix some old K&R style function argument definitions.
Add noreturn gcc attribute to some functions, as appropriate.
Add unused gcc attribute to some functions (eg ones meant to help
while debugging)
Add guard defines to headers which were missing them.
* command.c: (install_node) add const qualifier, still doesnt shut
up the warning though, because of the double pointer.
(cmp_node) ditto
* keychain.c: (key_str2time) Add GET_LONG_RANGE() macro, derived
fromn vty.h ones to fix some of the (long) < 0 warnings.
* thread.c: (various) use thread_empty
(cpu_record_hash_key) should cast to uintptr_t, a stdint.h type
* vty.h: Add VTY_GET_IPV4_ADDRESS and VTY_GET_IPV4_PREFIX so they
removed from ospfd/ospf_vty.h
* zebra.h: Move definition of ZEBRA_PORT to here, to remove
dependence of lib on zebra/zserv.h
paul [Tue, 3 May 2005 09:27:23 +0000 (09:27 +0000)]
2005-05-03 Paul Jakma <paul@dishone.st>
* (general) More cleaning up of stream abuse, isisd should be
back to previous functionality. Replace various XMALLOC+memset's
with XCALLOC
* isis_tlv.c: (tlv_add_padding) use stream_put to clear the stream
rather than forward endp, as isisd reuses streams.
* isis_pdu.c: (process_lsp) cleanup direct reference to stream endp
(send_lsp) manual copy of a stream cleaned up to use stream_copy.
* isis_network.c: (isis_recv_pdu_bcast) replace direct memcpy with
stream_write
(isis_recv_pdu_p2p) replace recvfrom directly into stream with
stream_recvfrom. Remove dangerous and now unneeded manual update
of endp.
(isis_recv_pdu_bcast / non-GNU_LINUX) Replace direct memcpy with
stream_write.
(isis_recv_pdu_p2p) replace read direct into stream with
stream_read_try, and hence remove the manual update of endp.
* isis_lsp.c: (lsp_update_data) manual stream dup replaced with
stream_dup.
(lsppdu_realloc) mempcy into stream data replaced with stream_put.
(lsp_build_nonpseudo) remove mysterious stream_forward_endp's -
which were originally stream_set_putp - shouldn't be needed
now that all the manual fiddling of private stream data has been
corrected.
(build_topology_lsp_data) remove unneeded twiddling of endp,
appears to be due to lsppdu_realloc(), but it appears to sort of
do the right thing wrt streams.
paul [Tue, 3 May 2005 09:07:56 +0000 (09:07 +0000)]
2005-05-03 Paul Jakma <paul@dishone.st>
* stream.h: Add comment about the special zero-ing ability of
stream_put.
(stream_recvmsg, stream_write) should return ssize_t and size_t
respectively. Should both be extern linkage.
(stream_recvfrom) Stream aware wrapper around recvfrom, in style
of stream_read_try.
* stream.c: (stream_recvfrom) new function, wrapper around recvfrom.
(stream_recvmsg, stream_write) ssize_t and size_t return values
paul [Fri, 29 Apr 2005 03:16:42 +0000 (03:16 +0000)]
2005-04-29 Paul Jakma <paul@dishone.st>
* defines.texi.in: Add PACKAGE_NAME, PACKAGE_STRING.
* quagga.texi: Use PACKAGE_NAME, this is more to give me an excuse
to update quagga.texi so that the 'UPDATED' in version.texi will
be more recent, the Makefile seems to be dumb and not examine
the texi files from which quagga.texi is generated.
* quagga.info: Fresh copy to match updates.
paul [Thu, 28 Apr 2005 18:21:15 +0000 (18:21 +0000)]
2005-04-28 Paul Jakma <paul@dishone.st>
* quagga.info: updated
* Makefile.am: automake seems to depend doc built stuff on srcdir
for some reason, or maybe its cause we have quagga.info,
quagga.texi in CVS - not sure. Anyway, the static rule for
quagga.pdf should depend on the srcdir quagga.texi in that case.
paul [Thu, 28 Apr 2005 17:35:14 +0000 (17:35 +0000)]
2005-04-28 Paul Jakma <paul.jakma@sun.com>
* rib.h: (struct rib) Add lock field for refcounting.
* zserv.h: (struct zebra_t) Add a ribq workqueue to the zebra
'master' struct.
* zserv.c: (zread_ipv4_add) XMALLOC then memset should be XCALLOC.
* zebra_rib.c: Clean up refcounting of route_node, make struct rib
refcounted and convert rib_process to work-queue. In general,
rib's should be rib_addnode'd and delnode'd to route_nodes, and
these symmetrical functions will manage the locking of referenced
route_node and freeing of struct rib - rather than having users
manage each seperately - with much scope for bugs..
(newrib_free) removed and replaced with rib_lock
(rib_lock) new function, check state of lock and increment.
(rib_unlock) new function, check lock state and decrement. Free
struct rib if refcount hits 0, freeing struct nexthop's, as
newrib_free did.
(rib_addnode) Add RIB to route_node, locking both.
(rib_delnode) Delete RIB from route_node, unlocking each.
(rib_process) Converted to a work-queue work function.
Functional changes are minimal, just arguments, comments and
whitespace.
(rib_queue_add_qnode) Helper function to setup a ribq item.
(rib_queue_add) Helper function, same arguments as old
rib_process, to replace in callers of rib_process.
(rib_queue_qnode_del) ribq deconstructor.
(rib_queue_init) Create the ribq.
(rib_init) call rib_queue_init.
(remainder) Sanitise refcounting of route_node's. Convert to
rib_queue_add, rib_addnode and rib_delnode. Change XMALLOC/memset
to XCALLOC. Remove calls to nexthop_delete and nexthop_free.
2005-04-27 Andrew J. Schorr <ajschorr@alumni.princeton.edu>
Add wall-clock timing statistics to 'show thread cpu' output.
* thread.h: Define struct rusage_t to contain wall-clock time
and cpu time. Change GETRUSAGE macro to collect both pieces
of data. Make appropriate changes to struct cpu_thread_history
to track CPU time and real time. Change proto for
thread_consumed_time to return real and cpu time elapsed.
And declare a new global variable 'struct timeval recent_time'.
* thread.c (struct timeval recent_time): New global timestamp variable.
(timeval_adjust): If timeout is negative, set to 0 (not 10
microseconds). And remove upper bound of 1,000,000 seconds, since
this does not seem to make any sense (and it breaks
funcname_thread_add_timer_timeval).
(timeval_cmp): Should return long, not int.
(vty_out_cpu_thread_history): Show CPU time and real time.
(cpu_record_hash_print): Calculate totals for CPU and real time.
(cpu_record_print): Change 'show thread cpu' title to show CPU and
real time.
(thread_timer_remain_second): Put current time in global recent_time.
(funcname_thread_add_timer_timeval): Fix assert. Replace 2-case
switch assignment with a ternary expression. Use global recent_time
variable. Fix use of timeval_adjust (previously, the value was not
actually being adjusted).
(thread_cancel): Add missing "break" statement in case
THREAD_BACKGROUND.
(thread_timer_wait): Use global recent_time value instead of calling
gettimeofday. And there's no need to check for negative timeouts,
since timeval_subtract already sets these to zero.
(thread_timer_process): Timers are sorted, so bail out once we
encounter a timer that has not yet popped. And remove some
extraneous asserts.
(thread_fetch): Do not process foreground timers before calling
select. Instead, add them to the ready list just after the select.
Also, no need to maintain a count of the number of ready threads,
since we don't care how many there are, just whether there's
one at the head of the ready list (which is easily checked).
Stick current time in global variable recent_time to reduce
the number of calls to gettimeofday. Tighten logic for
calculating the select timeout.
(thread_consumed_time): Now returns real time and puts the elapsed
cpu time in an additional argument.
(thread_should_yield): Use real (wall-clock) time to decide whether
to yield.
(thread_call): Maintain CPU and real time statistics.
* vty.c (vty_command): For slow commands, show real and cpu time.
paul [Wed, 27 Apr 2005 12:39:27 +0000 (12:39 +0000)]
2005-04-27 Paul Jakma <paul.jakma@sun.com>
* workqueue.h: (struct work_queue_item) change retry_count to ran,
its a count of number item has been run.
* workqueue.c: (show_work_queues) Fix formating of slightly
bugfix: fix SIGFPE if wq->runs is 0.
(work_queue_run) retry logic was slightly wrong.
cycles.best is 0 initialy, granularity is 1, so update best
if cycles >= granularity, not just >.
2005-04-26 Andrew J. Schorr <ajschorr@alumni.princeton.edu>
* buffer.c (buffer_write): Comment out call to buffer_flush_available.
This should speed up buffering at the expense of a possible increase
in latency in flushing the data if inside a long-running thread.
paul [Tue, 26 Apr 2005 02:35:03 +0000 (02:35 +0000)]
2005-04-26 Paul Jakma <paul.jakma@sun.com>
* quagga.xml.in: New, better, single integrated manifest for all
the services, courtesy of alan.maguire@sun.com.
* {{ospf,ospf6,rip,ripng,bgp}d,zebra}.xml.in: removed due to
previous
* quagga.init.in: Use the package tarname for properties, use
the correct daemon name variable.
* Makefile.am: Updated to match single manifest
* prototype.daemons.in: ditto
paul [Mon, 25 Apr 2005 16:42:24 +0000 (16:42 +0000)]
2004-05-25 Paul Jakma <paul@dishone.st>
* main.c: new file, common 'main' portion which then calls
test_init.
* heavy.c: remove common portions now in 'main.
* heavy-wq.c: ditto
* heavy-thread.c: ditto
* heavy*.c: Set the slow_function iteration number lower, to 300, it
was several orders of magnitude too 'slow' to be useful. :)
* Makefile.am: main is a source of the tests. heavy should link
to the math library.
paul [Mon, 25 Apr 2005 16:26:42 +0000 (16:26 +0000)]
2005-04-25 Paul Jakma <paul.jakma@sun.com>
* workqueue.{c,h}: Helper API for setting up and running queues via
background threads.
* command.c: install the 'show workqueues' command
* memtypes.c: Add work queue mtypes, and a rib-queue type for
a zebra rib work queue.
* memtypes.h: Updated to match memtypes.c
* Makefile.am: Add new workqueue files to build.
paul [Mon, 25 Apr 2005 00:37:03 +0000 (00:37 +0000)]
2005-04-25 Paul Jakma <paul.jakma@sun.com>
* HACKING: Add some notes about build system changes, to
document common oversights (common for me anyway).
Seperate sections with two newlines, easier to read.
2005-04-22 Andrew J. Schorr <ajschorr@alumni.princeton.edu>
* thread.h: Fix type for struct thread_master add_type: should be
unsigned char. Also, add some documentation of thread_add_background
args. And remove extraneous declaration of
show_thread_work_queues_cmd.
paul [Fri, 22 Apr 2005 00:43:47 +0000 (00:43 +0000)]
2005-04-22 Paul Jakma <paul.jakma@sun.com>
* thread.h: Add background thread type and thread_add_background
macro and accompanying funcname_... function.
export thread_should_yield, background threads can use it.
Lower thread yield time to 10ms, 100ms is noticeable lag and
a thread would only be /starting/ to finish sometime afterward.
* thread.c: (general) Add background thread type and schedule
nearly all thread types through the ready list for fairness.
(timeval_adjust) static qualifier missing
(vty_out_cpu_thread_history) add support for printout of
background threads
(show_thread_cpu) ditto.
(thread_master_debug) add debug of background list
(thread_master_create) fixup long line
(thread_add_unuse) add asserts for required state.
(thread_master_free) free background thread list
(funcname_thread_add_timer_timeval) make generic, able to
support arbitrary timer-like thread types.
(funcname_thread_add_timer) pass thread type to .._add_timer_timeval
(funcname_thread_add_timer_msec) ditto
(funcname_thread_add_background) Add a background thread, with an
optional millisecond delay factor, using .._add_timer_timeval.
(thread_cancel) Add background thread type.
Move the thread_list_delete common to all cases to bottom of
function, after the switch statement..
(thread_cancel_event) indent
(thread_timer_wait) Static qualifier, and make it able to cope
with arbitrary timer-like thread lists, so its of use to
background threads too.
(thread_process_fd) static qualifier. Again, make it take a list
reference rather than thread_master. Fix indentation.
(thread_timer_process) Check for ready timer-like threads in the
given list and move them on to the ready list - code originally
embedded in thread_fetch.
(thread_fetch) Schedule all threads, other than events, through
the ready list, to ensure fairness. Timer readying code moved to
thread_timer_process so it can be reused for background threads.
Remove the unneeded quagga_sigevent_process, as pointed out by
John Lin <john.ch.lin@gmail.com>.
(thread_should_yield) make this available.