FRR needs to properly include the FreeBSD headers for
compilation on FreeBSD. I have setup v6 as well
but I have not even tested it. Since I know
that the form is the same I think this is ok
at the moment. This is a step forward.
Because of this change *clearly* no-one is even
using pim on FreeBSD. <look at the MRT_XXX values
to prove to yourself>. In any event this is a step
in the direction of getting that working again.
Signed-off-by: Donald Sharp <sharpd@freebsd.network>
Topology:
TOR11 (FHR) --- LEAF-11---SPINE1 (RP)MSDP SPINE-2(RP)MSDP --- LEAF-12 -- TOR12 (LHR)
| | | | |
| -----------------------------------------------------(ECMP) |
| | | |
-----------------------------------------------------------------------(ECMP)
Issue:
In some triggers, S,G upstream is preserved even with the PP timer expiry, resulting
in S,G with NULL OILS. This could be because we create a dummy S,G upstream and
dummy channel_oif for *,G, where RPF is UNKNOWN. As a result, PIM+VXLAN traffic is never
forwarded downstream to LHR.
Fix:
when the S,G stream is running, Determine if a reevaluation of the outgoing interface
list (OIL) is required. S,G upstream should then inherit the OIL from *,G.
Testing:
- Evpn pim tests - TestEvpnPimSingleVtepOneMdt.test_02_broadcast_traffic_spt_zero
- pim-smoke
Ticket: #
Signed-off-by: Rajesh Varatharaj <rvaratharaj@nvidia.com>
Create a single registry of default port values that daemons
are using. Most of these are vty ports, but there are some
others for features like ospfapi and zebra FPM.
Signed-off-by: Mark Stapp <mjs@labn.net>
clang-format doesn't understand FRR_DAEMON_INFO is a long macro where
laying out items semantically makes sense.
(Also use only one `FRR_DAEMON_INFO(` in isisd so editors don't get
confused with the mismatching `( ( )`.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
Currently, when editing a leaf-list, `nb_candidate_edit` expects to
receive it's xpath without a predicate and the value in a separate
argument, and then creates the full xpath. This hack is complicated,
because it depends on the operation and on the caller being a backend or
not. Instead, let's require to always include the predicate in a
leaf-list xpath. Update all the usages in the code accordingly.
Signed-off-by: Igor Ryzhov <iryzhov@nfware.com>
Command: `ip ssmpingd 1.1.1.1`
Backtrace:
```
__GI_raise (sig=sig@entry=6) at ../sysdeps/unix/sysv/linux/raise.c:50
0x00007fd1d3b02859 in __GI_abort () at abort.c:79
0x00007fd1d3e323e1 in yang_dnode_xpath_get_canon (dnode=<optimized out>, xpath_fmt=<optimized out>, ap=<optimized out>) at lib/yang_wrappers.c:61
0x00007fd1d3e34f41 in yang_dnode_get_ipv4 (addr=addr@entry=0x7ffc368554d4, dnode=<optimized out>, xpath_fmt=xpath_fmt@entry=0x5556af8680d4 "./source-addr") at lib/yang_wrappers.c:826
0x00005556af8216d3 in routing_control_plane_protocols_control_plane_protocol_pim_address_family_ssm_pingd_source_ip_create (args=0x7ffc36855530) at pimd/pim_nb_config.c:925
0x00007fd1d3dec13f in nb_callback_create (nb_node=0x5556b197ea40, nb_node=0x5556b197ea40, errmsg_len=8192, errmsg=0x7ffc36855a90 "", resource=0x5556b18fa6f8, dnode=0x5556b1ad7a10, event=NB_EV_APPLY, context=0x5556b1ad75c0) at lib/northbound.c:1260
nb_callback_configuration (context=0x5556b1ad75c0, event=NB_EV_APPLY, change=<optimized out>, errmsg=0x7ffc36855a90 "", errmsg_len=8192) at lib/northbound.c:1648
0x00007fd1d3deca6c in nb_transaction_process (event=event@entry=NB_EV_APPLY, transaction=transaction@entry=0x5556b1ad75c0, errmsg=errmsg@entry=0x7ffc36855a90 "", errmsg_len=errmsg_len@entry=8192) at lib/northbound.c:1779
0x00007fd1d3decdd6 in nb_candidate_commit_apply (transaction=0x5556b1ad75c0, save_transaction=save_transaction@entry=true, transaction_id=transaction_id@entry=0x0, errmsg=errmsg@entry=0x7ffc36855a90 "", errmsg_len=errmsg_len@entry=8192) at lib/northbound.c:1129
0x00007fd1d3decf15 in nb_candidate_commit (context=..., candidate=<optimized out>, save_transaction=save_transaction@entry=true, comment=comment@entry=0x0, transaction_id=transaction_id@entry=0x0, errmsg=0x7ffc36855a90 "", errmsg_len=8192) at lib/northbound.c:1162
0x00007fd1d3ded4af in nb_cli_classic_commit (vty=vty@entry=0x5556b1ada2a0) at lib/northbound_cli.c:50
0x00007fd1d3df025f in nb_cli_apply_changes_internal (vty=vty@entry=0x5556b1ada2a0, xpath_base=xpath_base@entry=0x7ffc36859b50 ".", clear_pending=clear_pending@entry=false) at lib/northbound_cli.c:177
0x00007fd1d3df06ad in nb_cli_apply_changes (vty=vty@entry=0x5556b1ada2a0, xpath_base_fmt=xpath_base_fmt@entry=0x0) at lib/northbound_cli.c:233
0x00005556af80fdd5 in pim_process_ssmpingd_cmd (vty=0x5556b1ada2a0, operation=NB_OP_CREATE, src_str=0x5556b1ad9630 "1.1.1.1") at pimd/pim_cmd_common.c:3423
0x00007fd1d3da7b0e in cmd_execute_command_real (vline=vline@entry=0x5556b1ac9520, vty=vty@entry=0x5556b1ada2a0, cmd=cmd@entry=0x0, up_level=up_level@entry=0) at lib/command.c:982
0x00007fd1d3da7cb1 in cmd_execute_command (vline=vline@entry=0x5556b1ac9520, vty=vty@entry=0x5556b1ada2a0, cmd=0x0, vtysh=vtysh@entry=0) at lib/command.c:1040
0x00007fd1d3da7e50 in cmd_execute (vty=vty@entry=0x5556b1ada2a0, cmd=cmd@entry=0x5556b1ae0a30 "ip ssmpingd 1.1.1.1", matched=matched@entry=0x0, vtysh=vtysh@entry=0) at lib/command.c:1207
0x00007fd1d3e278be in vty_command (vty=vty@entry=0x5556b1ada2a0, buf=<optimized out>) at lib/vty.c:591
0x00007fd1d3e27afd in vty_execute (vty=0x5556b1ada2a0) at lib/vty.c:1354
0x00007fd1d3e2bb23 in vtysh_read (thread=<optimized out>) at lib/vty.c:2362
0x00007fd1d3e22254 in event_call (thread=thread@entry=0x7ffc3685cd80) at lib/event.c:2003
0x00007fd1d3dce9e8 in frr_run (master=0x5556b183c830) at lib/libfrr.c:1218
0x00005556af803653 in main (argc=6, argv=<optimized out>, envp=<optimized out>) at pimd/pim_main.c:162
```
Signed-off-by: Igor Ryzhov <iryzhov@nfware.com>
When turning on debug pim trace, there are lots of messages
surrounding the timing of rpf lookup. 99% of the time
no-one cares about these anymore. Let's make them
not seen unless we turn up debugs
Signed-off-by: Donald Sharp <sharpd@nvidia.com>
The Join or Prune messages require you to turn on `trace`
but this is part of Join/Prune processing of the packet
let's use PIM_DEBUG_PIM_J_P instead of TRACE here.
Signed-off-by: Donald Sharp <sharpd@nvidia.com>
When doing pim vxlan multicast bum handling, setup
the register to send up to 10 null registers on
immediate startup. If the null register packet
gets dropped this delays the formation of the
S,G tree from the RP towards the FHR.
Signed-off-by: Donald Sharp <sharpd@nvidia.com>
Practically no-one uses this and ioctls are pretty much
wrappered. Further wrappering could make this even better.
Signed-off-by: Donald Sharp <sharpd@nvidia.com>
... and use it instead of fiddling with the `.synchronous` field.
(Make it const while at it.)
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
Replace `struct list *` with `DLIST(if_connected, ...)`.
NB: while converting this, I found multiple places using connected
prefixes assuming they were IPv4 without checking:
- vrrpd/vrrp.c: vrrp_socket()
- zebra/irdp_interface.c: irdp_get_prefix(), irdp_if_start(),
irdp_advert_off()
(these fixes are really hard to split off into separate commits as that
would require going back and reapplying the change but with the old list
handling)
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
INTERFACE_NAMSIZ is just a redefine of IFNAMSIZ and IFNAMSIZ
is the standard for interface name length on all platforms
that FRR currently compiles on.
Signed-off-by: Donald Sharp <sharpd@nvidia.com>
Problem Statement:
========================
Mentioning few of the leaks here:
=3843268== 6 bytes in 3 blocks are still reachable in loss record 1 of 29
==3843268== at 0x483C855: malloc (vg_replace_malloc.c:381)
==3843268== by 0x489ED0E: qmalloc (memory.c:106)
==3843268== by 0x48DE8DB: redist_add_instance (zclient.c:125)
==3843268== by 0x48DF561: zclient_init (zclient.c:647)
==3843268== by 0x14FFA3: pim_zebra_init (pim_zebra.c:527)
==3843268== by 0x11D021: main (pim6_main.c:178)
==3843268==
==3843268== 24 bytes in 1 blocks are still reachable in loss record 2 of 29
==3843268== at 0x484147B: calloc (vg_replace_malloc.c:1328)
==3843268== by 0x489EE03: qcalloc (memory.c:111)
==3843268== by 0x4878DDE: buffer_new (buffer.c:72)
==3843268== by 0x48DE7BF: zclient_new (zclient.c:75)
==3843268== by 0x14FF1D: pim_zebra_init (pim_zebra.c:516)
==3843268== by 0x11D021: main (pim6_main.c:178)
==3843268==
==3843268== 24 bytes in 1 blocks are still reachable in loss record 3 of 29
==3843268== at 0x484147B: calloc (vg_replace_malloc.c:1328)
==3843268== by 0x489EE03: qcalloc (memory.c:111)
==3843268== by 0x4878DDE: buffer_new (buffer.c:72)
==3843268== by 0x48DE7BF: zclient_new (zclient.c:75)
==3843268== by 0x150A3D: zclient_lookup_new (pim_zlookup.c:131)
==3843268== by 0x11D021: main (pim6_main.c:178)
RCA:
=======================
Memory is allocated when the daemon started but
it is not freed when terminated.
Fix:
=======================
Freeing the memory when daemon goes down.
Signed-off-by: Mobashshera Rasool <mrasool@vmware.com>
About to use this for sockunion, which is not a prefix. `uniontype`
makes more sense, the macros are for defining transparent unions after
all.
(clang-format off thrown in as it otherwise wrecks formatting.)
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
...so that multiple functions can be subscribed.
The create/destroy hooks are renamed to real/unreal because that's what
they *actually* signal.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
In the scenario on an intermediate router where a *,G join has
been received and a S,G stream is being sent through that router
on the *,G stream, there exists a situation when the *,G in has been pruned
but the stream is still being received on on incoming interface towards
the RP for the *,G. In this situation PIM will see the S,G stream
initially as a NOCACHE from the dataplane, PIM will then do a RPF
for the S and notice that it is supposed to be coming in on adifferent
interface. In this case PIM the original PIM code would create
a blackhole mroute towards the RPF of the *,G( the interface the
stream is being received on ). The original reason for this is that
if there is a scenario where this particular S1,G stream is sending
at basically line rate, and there also happens to be a different
S2,G stream that is sending at a very low rate. With certain
dataplanes there is no way to really rate limit the S1 -vs- S2
stream and the S1 stream completely overwhelms the S2 stream
for sending up to the control plane for proper pim handling.
The problem then becomes that FRR never properly responds
to the situation where the *,G is rereceived and the S,G
stream switches back over to the SPT for itself and FRR ends
up with a dead mroute that stops everything from working properly.
This code change, installs the blackhole mroute with the RPF
towards the RP for the G and then resets the RPF to the correct
RPF for the Stream but does not modify the mroute. When the
*,G is rereceived and we attempt to transition to the S,G stream
this now works.
As a note: Both David L and myself do not necessarily believe
we fully understand the problem yet. What this does do is fix
all the inconsistent CI issues we are seeing in the topotests
at this time. Internally I am seeing other test failures
in PIM that I don't fully understand and we suspect that
there are other problems in the state machine. We plan to
revisit this problem as we are able to debug the issue better.
In the meantime both David and Myself agree that this gets
the CI working again and Streams end up in the right state.
Signed-off-by: Donald Sharp <sharpd@nvidia.com>
Modify empty json object to take input obj
instead of allocating always one.
There are situation where in error condition or no data
case print empty json (`{}`) with already allocated
Signed-off-by: Chirag Shah <chirag@nvidia.com>
Problem Statement:
===================
Syscall param sendmsg(msg.msg_iov[0]) points to uninitialised byte(s)
at 0x4975157: sendmsg (sendmsg.c:28)
==2263111== by 0x1413BE: pim_msg_send_frame (pim_pim.c:629)
==2263111== by 0x1413BE: pim_msg_send (pim_pim.c:743)
==2263111== by 0x1425DC: pim_register_send (pim_register.c:332)
==2263111== by 0x1427EE: pim_null_register_send (pim_register.c:443)
==2263111== by 0x14D228: pim_upstream_register_stop_timer (pim_upstream.c:1608)
==2263111== by 0x48CE6DF: thread_call (thread.c:1693)
==2263111== by 0x4899EFF: frr_run (libfrr.c:1068)
==2263111== by 0x11D035: main (pim6_main.c:190)
==2263111== Address 0x1ffeffdcb1 is on thread 1's stack
==2263111== in frame #2, created by pim_register_send (pim_register.c:273)
==2263111== Uninitialised value was created by a stack allocation
==2263111== at 0x142690: pim_null_register_send (pim_register.c:389)
RCA:
====================
1. All members of struct pim_msg_header were not initiliased while sending
null register packet. Therefore when the pointers are assigned while
sending the msg via sendmsg, it complains the pointer points to
uninitialised byte.
2. struct ipv6_ph ph was also not initialised.
Fix:
====================
Initialised all the members using memset.
Signed-off-by: Mobashshera Rasool <mrasool@vmware.com>
Also:
- replace all /* fallthrough */ comments with portable fallthrough;
pseudo keyword to accomodate both gcc and clang
- add missing break; statements as required by older versions of gcc
- cleanup some code to remove unnecessary fallthrough
Signed-off-by: Igor Ryzhov <iryzhov@nfware.com>
Currently when one interface changes its VRF, zebra will send these messages to
all daemons in *order*:
1) `ZEBRA_INTERFACE_DELETE` ( notify them delete from old VRF )
2) `ZEBRA_INTERFACE_VRF_UPDATE` ( notify them move from old to new VRF )
3) `ZEBRA_INTERFACE_ADD` ( notify them added into new VRF )
When daemons deal with `VRF_UPDATE`, they use
`zebra_interface_vrf_update_read()->if_lookup_by_name()`
to check the interface exist or not in old VRF. This check will always return
*NULL* because `DELETE` ( deleted from old VRF ) is already done, so can't
find this interface in old VRF.
Send `VRF_UPDATE` is redundant and unuseful. `DELETE` and `ADD` are enough,
they will deal with RB tree, so don't send this `VRF_UPDATE` message when
vrf changes.
Since all daemons have good mechanism to deal with changing vrf, and don't
use this `VRF_UPDATE` mechanism. So, it is safe to completely remove
all the code with `VRF_UPDATE`.
Signed-off-by: anlan_cs <anlan_cs@tom.com>
Indicating the configured PIM Rendezvous Point (RP) in the MSDP SA
message
The RFC-3618, section 12.2.1, describes the fields included in the MSDP
SA message. The "RP address" field is "the address of the RP in the
domain the source has become active in".
In the most common case, we will establish an MSDP connection from RP to
RP. However, there are cases where we want to establish a MSDP
connection from an interface/address that is not the RP. Section 3 of
RFC-3618 describes that scenario as "intermediate MSDP peer". Moreover,
the RP could be another router in the PIM domain - not the one
establishing the MSDP connection.
The current implementation could be problematic even with a single
router per PIM domain. Consider the following scenario:
* There are two PIM domains, each one with a single router.
* The two routers are connected via two independent networks. Let's say
that is to provide redundancy.
* The routers are configured to establish two MSDP connections, one on
each network (redundancy again).
* A multicast source becomes active on the router 1. It will be
communicated to router 2 via two independent MSDP SA messages, one per
MSDP connection.
* Without these changes, each MSDP SA message will indicate a different
RP.
* Both RP addresses will pass the RPF check, and both MSDP sources will
be accepted.
* If the router has clients interested in that multicast group, it will
send PIM Join messages to both RPs and start receiving the multicast
traffic from both.
With the changes included in this commit, the multicast source available
in router 1 would still be communicated to router 2 twice. But both MSDP
SA messages would indicate the same RP, and one of them would be
discarded due to failure in the RPF-check failure. Also, the changes
allow us to define the RP that will be included in the MSDP SA message,
and it could be one of the interfaces used to establish the MSDP
connection, some other interface on the router, a loopback interface, or
another router in the PIM domain.
These changes should not create compatibility issues. As I mentioned, we
usually establish MSDP connections from RP to RP. In this case, the
result will be the same. We would still indicate the address used to
establish the MSDP connection if the RP is not set - I wonder if that
should even be a valid configuration.
Signed-off-by: Adriano Marto Reis <adrianomarto@gmail.com>
The socket has been closed in `ssmpingd_setsockopt()` in the wrong cases,
so remove the redundant closing socket from outer layer.
Signed-off-by: anlan_cs <anlan_cs@tom.com>
Use oil_incoming_vif instead of oil_parent. I had
to go look this up as that I failed to remember that
the linux kernel calls this parent for some bizarre
reason.
Signed-off-by: Donald Sharp <sharpd@nvidia.com>
When debugging and outputting the oil_parent() let's just
convert it to a string that is useful for people trying
to debug pim
Signed-off-by: Donald Sharp <sharpd@nvidia.com>
When receiving a register message for a Group, that the group has no
associated RP specified. Prevent a crash from happening.
Signed-off-by: Donald Sharp <sharpd@nvidia.com>
The YANG specification currently designates a uint8 data type for the hello interval,
despite the CLI documentation (ip pim hello (1-65535) [(1-65535)]) indicating a maximum value of 65535.
To address this inconsistency, updating the data type to uint16 allowing for a maximum value for hello intervals.
Signed-off-by: Sai Gomathi N <nsaigomathi@vmware.com>
