pimd/TODO

   1 T1 DONE Implement debug command
   2    test pim receive join
   3
   4 T2 DONE Implement debug command
   5    test pim receive prune
   6
   7 T3 DONE Per-interface Downstream (S,G) state machine
   8    (RFC 4601 4.5.3. Receiving (S,G) Join/Prune Messages)
   9
  10 T4 DONE Upstream (S,G) state machine
  11    (RFC 4601 4.5.7. Sending (S,G) Join/Prune Messages)
  12
  13 T5 DONE Verify Data Packet Forwarding Rules
  14    RFC 4601 4.2.  Data Packet Forwarding Rules
  15    RFC 4601 4.8.2.  PIM-SSM-Only Routers
  16
  17    Additionally, the Packet forwarding rules of Section 4.2 can be
  18    simplified in a PIM-SSM-only router:
  19
  20      iif is the incoming interface of the packet.
  21      oiflist = NULL
  22      if (iif == RPF_interface(S) AND UpstreamJPState(S,G) == Joined) {
  23        oiflist = inherited_olist(S,G)
  24      } else if (iif is in inherited_olist(S,G)) {
  25        send Assert(S,G) on iif
  26      }
  27      oiflist = oiflist (-) iif
  28      forward packet on all interfaces in oiflist
  29
  30    Macro:
  31      inherited_olist(S,G) =
  32        joins(S,G) (+) pim_include(S,G) (-) lost_assert(S,G)
  33
  34 T6 DONE Implement (S,G) Assert state machine (RFC 4601, section 4.6.1).
  35    Changes in pim_ifchannel.ifassert_winner should trigger
  36    pim_upstream_update_join_desired().
  37    Depends on TODO T27.
  38    Depends on TODO T33.
  39    See also CAVEAT C7.
  40    See also: RFC 4601 4.5.7. Sending (S,G) Join/Prune Messages
  41     Transitions from Joined State
  42      RPF'(S,G) changes due to an Assert
  43
  44    http://www.hep.ucl.ac.uk/~ytl/multi-cast/pim-dm_01.html:
  45
  46      The PIM Assert mechanism is used to shutoff duplicate flows onto
  47      the same multiaccess network. Routers detect this condiction when
  48      they receive an (S,G) packet via a multi-access interface that is
  49      in the (S,G) OIL. This causes the routers to send Assert
  50      Messages.
  51
  52    Note that neighbors will not accept Join/Prune or Assert messages
  53    from a router unless they have first heard a Hello message from that
  54    router.  Thus, if a router needs to send a Join/Prune or Assert
  55    message on an interface on which it has not yet sent a Hello message
  56    with the currently configured IP address, then it MUST immediately
  57    send the relevant Hello message without waiting for the Hello Timer
  58    to expire, followed by the Join/Prune or Assert message.
  59
  60 T7 DONE Implement hello option: LAN Prune Delay
  61
  62 T8 DONE Implement J/P_Override_Interval(I)
  63    Depends on TODO T7.
  64    See pim_ifchannel.c, pim_ifchannel_prune(), jp_override_interval.
  65
  66 T9 DONE Detect change in IGMPv3 RPF interface/next-hop for S and update.
  67    channel_oil vif index accordingly ?
  68    Beware accidentaly adding looped MFC entries (IIF=OIF).
  69
  70 T10 DONE React to (S,G) join directed to another upstream address. See
  71     also:
  72
  73     RFC 4601: 4.5.7.  Sending (S,G) Join/Prune Messages
  74
  75     If a router wishes to propagate a Join(S,G) upstream, it must also
  76     watch for messages on its upstream interface from other routers on
  77     that subnet, and these may modify its behavior.  If it sees a
  78     Join(S,G) to the correct upstream neighbor, it should suppress its
  79     own Join(S,G).  If it sees a Prune(S,G), Prune(S,G,rpt), or
  80     Prune(*,G) to the correct upstream neighbor towards S, it should
  81     be prepared to override that prune by scheduling a Join(S,G) to be
  82     sent almost immediately.
  83
  84 T11 DONE Review protocol modifications for SSM
  85     (RFC 4601 4.8.1.  Protocol Modifications for SSM Destination
  86     Addresses)
  87
  88 T12 DONE Review updates of RPF entries.
  89     FIXME pim_upstream.c send_join():
  90     Currently only one upstream state is affected by detection of RPF change.
  91     RPF change should affect all upstream states sharing the RPF cache.
  92
  93 T13 DONE Check that RFC macros using S,G,RPF_interface(S) are actually
  94     implemented with this strategy:
  95     rpf_ifch=find_ifch(up->rpf->interface).
  96     See pim_rpf.c pim_rpf_find_rpf_addr() for a correct example.
  97
  98     $ grep -i macro pimd/*.c
  99     pimd/pim_iface.c:  RFC 4601: 4.1.6.  State Summarization Macros
 100     pimd/pim_ifchannel.c:    RFC 4601: 4.6.5.  Assert State Macros
 101     pimd/pim_ifchannel.c:  RFC 4601: 4.1.6.  State Summarization Macros
 102     pimd/pim_ifchannel.c:  RFC 4601: 4.1.6.  State Summarization Macros
 103     pimd/pim_ifchannel.c:  RFC 4601: 4.6.5.  Assert State Macros
 104     pimd/pim_ifchannel.c:  Macro:
 105     pimd/pim_rpf.c:  RFC 4601: 4.1.6.  State Summarization Macros
 106
 107 T14 DONE Send Assert(S,G) on iif as response to WRONGVIF kernel upcall.
 108     See pim_mroute.c mroute_msg().
 109
 110 T15 DONE Interface command to statically join (S,G).
 111     interface eth0
 112      ip igmp join-group 239.1.1.1 source 1.1.1.1
 113
 114 T16 DONE RPF'(S,G) lookup is not working for S reachable with default route.
 115     See "RPF'(S,G) not found" in pim_rpf_update() from pim_rpf.c.
 116     Zebra daemon RIB is not reflecting changes in kernel routes
 117     accurately?
 118
 119 T17 DONE Prevent CLI from creating bogus interfaces.
 120     Example:
 121     conf t
 122      interface xxx
 123
 124 T18 Consider reliable pim solution (refresh reduction)
 125     A Reliable Transport Mechanism for PIM
 126     http://tools.ietf.org/wg/pim/draft-ietf-pim-port/
 127     PORT=PIM-Over-Reliable-Transport
 128
 129 T19 DONE Fix self as neighbor
 130     See mailing list post:
 131     http://lists.gnu.org/archive/html/qpimd-users/2009-04/msg00000.html
 132
 133 T20 DONE Fix debug message: "pim_neighbor_update: internal error:
 134     trying to replace same prefix list"
 135     See mailing list post:
 136     http://lists.gnu.org/archive/html/qpimd-users/2009-04/msg00000.html
 137
 138 T21 DONE Clean-up PIM/IGMP interface mismatch debugging
 139     See option PIM_CHECK_RECV_IFINDEX_SANITY in pimd/Makefile.am
 140     See mailing list post:
 141     http://lists.nongnu.org/archive/html/qpimd-users/2009-04/msg00003.html
 142
 143 T22 DONE IGMP must be protected against adding looped MFC entries
 144     created by both source and receiver attached to the same
 145     interface.
 146
 147 T23 DONE libfrr crash after zclient_lookup_nexthop.
 148     See mailing list post:
 149     http://lists.nongnu.org/archive/html/qpimd-users/2009-04/msg00008.html
 150
 151 T24 DONE zserv may return recursive routes:
 152      - nexthop type is set to ZEBRA_NEXTHOP_IPV4
 153      - ifindex is not reported
 154      - calls expecting ifindex (fib_lookup_if_vif_index) are disrupted
 155     See also this mailing list post:
 156     [PATCH 21/21] Link detect and recursive routes
 157     http://www.gossamer-threads.com/lists/quagga/dev/17564
 158
 159 T25 DONE Zclient nexthop lookup missing OSPF route to 1.1.1.1/32
 160     See also:
 161     pim_zlookup.c zclient_lookup_nexthop misses OSPF 1.1.1.1/32
 162     zebra/zebra_vty.c show_ip_route_addr_cmd hits OSPF 1.1.1.1/32
 163
 164 T26 DONE Zebra daemon is marking recursive static route as inactive.
 165
 166     FIXED: zebra daemon was incorrectly marking recursive routes
 167     pointing to kernel routes as inactive:
 168       zebra/zebra_rib.c nexthop_active_ipv4:
 169         -- Original:
 170           else if (CHECK_FLAG (rib->flags, ZEBRA_FLAG_INTERNAL))
 171         -- Fixed:
 172           else if (CHECK_FLAG (rib->flags, ZEBRA_FLAG_INTERNAL) ||
 173                    match->type == ZEBRA_ROUTE_KERNEL)
 174
 175     Old problem description:
 176
 177     This prevents rib_match_ipv4 from returning its nexthop:
 178     client: pim_zlookup.c zclient_read_nexthop
 179     server: zebra/zserv.c zsend_ipv4_nexthop_lookup_v2 -> rib_match_ipv4
 180
 181     Kernel route is injected into zebra in zebra_rib.c rib_add_ipv4
 182     Examples:
 183     rt_netlink.c:726: rib_add_ipv4 (ZEBRA_ROUTE_KERNEL, flags, &p, gate, src, index, table, metric, 0);
 184     rt_netlink.c:864: rib_add_ipv4 (ZEBRA_ROUTE_KERNEL, 0, &p, gate, src, index, table, 0, 0);
 185
 186     This patch didn't fix the issue:
 187     [PATCH 21/21] Link detect and recursive routes
 188     http://www.gossamer-threads.com/lists/quagga/dev/17564
 189
 190     See the example below for the route 2.2.2.2.
 191
 192 bash# route add -host 1.1.1.1 gw 127.0.0.1
 193 bash# route add -host 2.2.2.2 gw 1.1.1.1
 194 bash# netstat -nvr
 195 Kernel IP routing table
 196 Destination     Gateway         Genmask         Flags   MSS Window  irtt Iface
 197 2.2.2.2         1.1.1.1         255.255.255.255 UGH       0 0          0 lo
 198 1.1.1.1         127.0.0.1       255.255.255.255 UGH       0 0          0 lo
 199 192.168.0.0     0.0.0.0         255.255.255.0   U         0 0          0 eth0
 200 0.0.0.0         192.168.0.2     0.0.0.0         UG        0 0          0 eth0
 201 bash#
 202
 203 zebra# sh ip route
 204 Codes: K - kernel route, C - connected, S - static, R - RIP, O - OSPF,
 205        I - ISIS, B - BGP, > - selected route, * - FIB route
 206
 207 K>* 0.0.0.0/0 via 192.168.0.2, eth0
 208 K>* 1.1.1.1/32 via 127.0.0.1, lo
 209 K * 2.2.2.2/32 via 1.1.1.1, lo inactive
 210 C>* 127.0.0.0/8 is directly connected, lo
 211 C>* 192.168.0.0/24 is directly connected, eth0
 212
 213 quagga-pimd-router# sh ip route 1.1.1.1
 214 Address         NextHop         Interface Metric Preference
 215 1.1.1.1         127.0.0.1       lo             0          0
 216 quagga-pimd-router#
 217 quagga-pimd-router# sh ip route 2.2.2.2
 218 Address         NextHop         Interface Metric Preference
 219 2.2.2.2         192.168.0.2     eth0           0          0
 220 quagga-pimd-router#
 221
 222 T27 DONE Implement debug command
 223     test pim receive assert
 224     See also TODO T6: (S,G) Assert state machine.
 225
 226 T28 DONE Bad IPv4 address family=02 in Join/Prune dump
 227     Reported by Andrew Lunn <andrew.lunn@ascom.ch>
 228
 229     # 58-byte pim v2 Join/Prune dump
 230     # ------------------------------
 231     # IPv4 address family=02 is wrong, correct IPv4 address family is 01
 232     # See http://www.iana.org/assignments/address-family-numbers
 233     #
 234     c8XX YY03 : ip src 200.xx.yy.3
 235     e000 000d : ip dst 224.0.0.13
 236     9404 0000 : ip router alert option 148.4.0.0
 237     2300 ab13 : pimv2,type=3 res=00 checksum=ab13
 238     0200      : upstream family=02, encoding=00
 239     c8XX YY08 : upstream 200.xx.yy.8
 240     0001 00d2 : res=00 groups=01 holdtime=00d2
 241     0200 0020 : group family=02, encoding=00, res=00, mask_len=20
 242     ef01 0101 : group address 239.1.1.1
 243     0001 0000 : joined=0001 pruned=0000
 244     0200 0020 : source family=02, encoding=00, res=00, mask_len=20
 245     0101 0101 : source address 1.1.1.1
 246
 247 T29 DONE Reset interface PIM-hello-sent counter when primary address changes
 248     See pim_ifp->pim_ifstat_hello_sent
 249
 250     RFC 4601: 4.3.1.  Sending Hello Messages
 251
 252     Thus, if a router needs to send a Join/Prune or Assert message on
 253     an interface on which it has not yet sent a Hello message with the
 254     currently configured IP address, then it MUST immediately send the
 255     relevant Hello message without waiting for the Hello Timer to
 256     expire, followed by the Join/Prune or Assert message.
 257
 258 T30 DONE Run interface DR election when primary address changes
 259     Reported by Andrew Lunn <andrew.lunn@ascom.ch>
 260     See pim_if_dr_election().
 261
 262 T31 If an interface changes one of its secondary IP addresses, a Hello
 263     message with an updated Address_List option and a non-zero
 264     HoldTime should be sent immediately.
 265     See also detect_secondary_address_change
 266     See also CAVEAT C15.
 267     See also RFC 4601: 4.3.1.  Sending Hello Messages
 268
 269 T32 FIXED Detection of interface primary address changes may fail when
 270     there are multiple addresses.
 271     See also CAVEAT C14.
 272
 273     pim_find_primary_addr() should return interface primary address
 274     from connected list. Currently it returns the first address.
 275
 276     Zebra daemon "show int" is able to keep the primary address as
 277     first address.
 278
 279 T33 DONE Implement debug command: test pim receive upcall
 280     See also TODO T6: (S,G) Assert state machine.
 281
 282 T34 DONE assert_action_a1
 283
 284 T35 DONE Review macros depending on interface I.
 285
 286     See also: grep ,I\) pimd/*.c
 287
 288     For the case (S,G,I) check if I is either
 289     1) interface attached to this per-interface S,G state (don't think so)
 290     or
 291     2) an arbitrary interface (most probably)
 292
 293     For the arbitrary interface case (2), consider representing
 294     interface ifp as its primary address (struct in_addr ifaddr).  The
 295     benefit is in_addr does not need to be dereferenced, so it does
 296     not demand protection against crashes.
 297
 298 T36 DONE React to zebra daemon link-detect up/down notification.
 299     pim_ifp->primary_address is managed by detect_primary_address_change()
 300     depending on to ifp->connected (managed by zebra_interface_address_read()).
 301
 302 T37 DONE Review list of variables which may affect pim_upstream.c
 303     pim_upstream_evaluate_join_desired().
 304     Call pim_upstream_update_join_desired() accordingly.
 305
 306     See the order of invokation:
 307       pim_if_dr_election(ifp);
 308       pim_if_update_join_desired(pim_ifp); /* depends on DR */
 309       pim_if_update_could_assert(ifp); /* depends on DR */
 310       pim_if_update_my_assert_metric(ifp); /* depends on could_assert */
 311
 312     join_desired depends on:
 313       pim_ifp->primary_address
 314       pim_ifp->pim_dr_addr
 315       ch->ifassert_winner_metric
 316       ch->ifassert_winner
 317       ch->local_ifmembership
 318       ch->ifjoin_state
 319       ch->upstream->rpf.source_nexthop.mrib_metric_preference
 320       ch->upstream->rpf.source_nexthop.mrib_route_metric
 321       ch->upstream->rpf.source_nexthop.interface
 322
 323 T38 DONE Detect change in AssertTrackingDesired(S,G,I)
 324
 325     See the order of invokation:
 326       dr_election: none
 327       update_join_desired: depends on DR
 328       update_tracking_desired: depends on DR, join_desired
 329
 330     AssertTrackingDesired(S,G,I) depends on:
 331       pim_ifp->primary_address
 332       pim_ifp->pim_dr_addr
 333       ch->local_ifmembership
 334       ch->ifassert_winner
 335       ch->ifjoin_state
 336       ch->upstream->rpf.source_nexthop.interface
 337       PIM_UPSTREAM_FLAG_TEST_DR_JOIN_DESIRED(ch->upstream->flags)
 338
 339 T39 DONE AssertTrackingDesired: flags is not matching evaluation
 340
 341     # show ip pim assert-internal
 342     CA:   CouldAssert
 343     ECA:  Evaluate CouldAssert
 344     ATD:  AssertTrackingDesired
 345     eATD: Evaluate AssertTrackingDesired
 346
 347     Interface Address         Source          Group           CA  eCA ATD eATD
 348     eth0      192.168.1.100   1.1.1.1         239.1.1.1       no  no  no  yes
 349     #
 350
 351 T40 Lightweight MLDv2
 352     http://tools.ietf.org/html/draft-ietf-mboned-lightweight-igmpv3-mldv2-05
 353     http://www.ietf.org/internet-drafts/draft-ietf-mboned-lightweight-igmpv3-mldv2-05.txt
 354     http://www.ietf.org/html.charters/mboned-charter.html
 355
 356 T41 DONE ssmping support
 357
 358     See also:
 359       http://www.venaas.no/multicast/ssmping/
 360       draft-ietf-mboned-ssmping-07
 361       http://tools.ietf.org/html/draft-ietf-mboned-ssmping-07
 362
 363     Example:
 364
 365     debug ssmpingd
 366
 367     conf t
 368      ip ssmpingd 1.1.1.1
 369
 370     show ip ssmpingd
 371
 372 T42 Static igmp join fails when loading config at boot time
 373
 374     ! Wrong behavior seen at boot time:
 375     !
 376     2010/02/22 08:59:00 PIM: igmp_source_forward_start: ignoring request for
 377     looped MFC entry (S,G)=(3.3.3.3,239.3.3.3): igmp_sock=12 oif=eth0 vif_index=2
 378
 379     ! Correct behavior seen later:
 380     !
 381     2010/02/22 09:03:16 PIM: igmp_source_forward_start: ignoring request for
 382     looped MFC entry (S,G)=(2.2.2.2,239.2.2.2): igmp_sock=17 oif=lo vif_index=1
 383
 384     ! To see the wrong message at boot:
 385     !
 386     debug igmp trace
 387     !
 388     interface lo
 389      ip igmp
 390      ip igmp join 239.2.2.2 2.2.2.2
 391      ip igmp join 239.3.3.3 3.3.3.3
 392     !
 393
 394     ! Interfaces indexes:
 395     Interface Address         ifi Vif  PktsIn PktsOut    BytesIn   BytesOut
 396     eth0      200.202.112.3     2   2       0       0          0          0
 397     lo        127.0.0.1         1   1       0       0          0          0
 398
 399 T43 PIM Neighbor Reduction
 400     https://datatracker.ietf.org/doc/draft-wijnands-pim-neighbor-reduction/
 401
 402     "In a transit LAN (no directly connected source or receiver), many
 403     of the PIM procedures don't apply. (...) This proposal describes
 404     a procedure to reduce the amount of neighbors established over a
 405     transit LAN."
 406
 407 T44 Single Stream Multicast Fast Reroute (SMFR) Method
 408     https://datatracker.ietf.org/doc/draft-liu-pim-single-stream-multicast-frr/
 409
 410     "This document proposes an IP multicast fast convergence method
 411     based on differentiating primary and backup PIM join."
 412
 413 T45 RFC5384 - The Join Attribute Format
 414     "This document describes a modification of the Join message that
 415     allows a node to associate attributes with a particular tree."
 416
 417 T46 PIM Multi-Topology ID (MT-ID) Join-Attribute
 418     http://tools.ietf.org/html/draft-cai-pim-mtid-00
 419     Depends on T45.
 420
 421     "This draft introduces a new type of PIM Join Attribute used to
 422     encode the identity of the topology PIM uses for RPF."
 423
 424 -x-