With a dead interval of 40 seconds, each tests is waiting 40+
seconds for ospf convergence to occurr because the DR is re-elected
Signed-off-by: Donald Sharp <sharpd@nvidia.com>
The idea here is to pass "non-code agreements" through the PR review
mechanism, and have them visible in the git tree.
Two "example" (but real) accords are included, mostly to illustrate the
idea. Both of these should be non-controversial and have had some
previous discussion in random places.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
The default vrf in bgp when created, ends up having the
bgp->name as NULL. This of course crashes the ilk
of `json_object_string_add` when a NULL is passed in.
Go through all the places in bgp_evpn_mh.c and fix
so that it doesn't crash.
Signed-off-by: Donald Sharp <sharpd@nvidia.com>
Add some safe guards to avoid crashes and alert us about programming
errors in packet build.
Signed-off-by: Rafael Zalamena <rzalamena@opensourcerouting.org>
Increase the MSDP peer stream buffer size to handle the whole TLV
(maximum is 65KiB due to 16bit field). If the stream is not resized
there will be a crash in the read function attempting to put more than
9192 (`PIM_MSDP_SA_TLV_MAX_SIZE`) bytes.
According to the RFC 3618 Section 12 we should accept the TLV and we
should not reset the peer connection.
Signed-off-by: Rafael Zalamena <rzalamena@opensourcerouting.org>
This reverts commit 2000ac4075.
There were concerns that ensuring zebra stopped last led to
problems with zebra's "-r" flag, so we'll revert that for the
time being and reconsider this area.
Signed-off-by: Mark Stapp <mjs@labn.net>
Some keys are only present in the JSON data of BGP neighbors are only present if the peer is, or has previously been established.
While they are not present if the peer has never come up.
To keep the data structure aligned, the below keys are added also to the neighbors that BGP adjacency has never been established.
Values of the keys are all set to Unknown
hostname:Unknown,
nexthop:Unknown,
nexthopGlobal:Unknown,
nexthopLocal:Unknown,
bgpConnection:Unknown,
Signed-off-by: Karl Quan <kquan@nvidia.com>
The BGP fsm uses return codes to pass event success/fail
as well as some extra data to the bgp_event_update function.
Convert this to use a enum instead of an int to track the
changes.
Signed-off-by: Donald Sharp <sharpd@nvidia.com>
When actually creating a peer in BGP, tell the creation if
it is a config node or not. There were cases where the
CONFIG_NODE was being set *after* being placed into
the bgp->peerhash, thus causing collisions between the
doppelganger and the peer and eventually use after free's.
Signed-off-by: Donald Sharp <sharpd@nvidia.com>
The bgp->peerhash is made up of the sockunion and the CONFIG_NODE
flag. If the CONFIG_NODE flag is moved around or changed then
we get into a situation where both the doppelganger and the peer
actually hash to the exact same thing. Leading to wrongful deletion
and pointers being used after freed.
Signed-off-by: Donald Sharp <sharpd@nvidia.com>
Currently bgp does not stop any events that are on the thread
system for execution on peer deletion. This is not good.
Stop those events and prevent use after free's.
Signed-off-by: Donald Sharp <sharpd@nvidia.com>
When the decision has been made to copy a peer configuration from
a peer to another peer because one is taking over. Do not automatically
set the CONFIG_NODE flag. Instead we need to handle that appropriately
when the final decision is made to transfer.
Signed-off-by: Donald Sharp <sharpd@nvidia.com>
When changing the peers sockunion structure the bgp->peer
list was not being updated properly. Since the peer's su
is being used for a sorted insert then the change of it requires
that the value be pulled out of the bgp->peer list and then
put back into as well.
Additionally ensure that the hash is always released on peer
deletion.
Lead to this from this decode in a address sanitizer run.
=================================================================
==30778==ERROR: AddressSanitizer: heap-use-after-free on address 0x62a0000d8440 at pc 0x7f48c9c5c547 bp 0x7ffcba272cb0 sp 0x7ffcba272ca8
READ of size 2 at 0x62a0000d8440 thread T0
#0 0x7f48c9c5c546 in sockunion_same lib/sockunion.c:425
#1 0x55cfefe3000f in peer_hash_same bgpd/bgpd.c:890
#2 0x7f48c9bde039 in hash_release lib/hash.c:209
#3 0x55cfefe3373f in bgp_peer_conf_if_to_su_update bgpd/bgpd.c:1541
#4 0x55cfefd0be7a in bgp_stop bgpd/bgp_fsm.c:1631
#5 0x55cfefe4028f in peer_delete bgpd/bgpd.c:2362
#6 0x55cfefdd5e97 in no_neighbor_interface_config bgpd/bgp_vty.c:4267
#7 0x7f48c9b9d160 in cmd_execute_command_real lib/command.c:949
#8 0x7f48c9ba1112 in cmd_execute_command lib/command.c:1009
#9 0x7f48c9ba1573 in cmd_execute lib/command.c:1162
#10 0x7f48c9c87402 in vty_command lib/vty.c:526
#11 0x7f48c9c87832 in vty_execute lib/vty.c:1291
#12 0x7f48c9c8e741 in vtysh_read lib/vty.c:2130
#13 0x7f48c9c7a66d in thread_call lib/thread.c:1585
#14 0x7f48c9bf64e7 in frr_run lib/libfrr.c:1123
#15 0x55cfefc75a15 in main bgpd/bgp_main.c:540
#16 0x7f48c96b009a in __libc_start_main (/lib/x86_64-linux-gnu/libc.so.6+0x2409a)
#17 0x55cfefc787f9 in _start (/usr/lib/frr/bgpd+0xe27f9)
0x62a0000d8440 is located 576 bytes inside of 23376-byte region [0x62a0000d8200,0x62a0000ddd50)
freed by thread T0 here:
#0 0x7f48c9eb9fb0 in __interceptor_free (/lib/x86_64-linux-gnu/libasan.so.5+0xe8fb0)
#1 0x55cfefe3fe42 in peer_free bgpd/bgpd.c:1113
#2 0x55cfefe3fe42 in peer_unlock_with_caller bgpd/bgpd.c:1144
#3 0x55cfefe4092e in peer_delete bgpd/bgpd.c:2457
#4 0x55cfefdd5e97 in no_neighbor_interface_config bgpd/bgp_vty.c:4267
#5 0x7f48c9b9d160 in cmd_execute_command_real lib/command.c:949
#6 0x7f48c9ba1112 in cmd_execute_command lib/command.c:1009
#7 0x7f48c9ba1573 in cmd_execute lib/command.c:1162
#8 0x7f48c9c87402 in vty_command lib/vty.c:526
#9 0x7f48c9c87832 in vty_execute lib/vty.c:1291
#10 0x7f48c9c8e741 in vtysh_read lib/vty.c:2130
#11 0x7f48c9c7a66d in thread_call lib/thread.c:1585
#12 0x7f48c9bf64e7 in frr_run lib/libfrr.c:1123
#13 0x55cfefc75a15 in main bgpd/bgp_main.c:540
#14 0x7f48c96b009a in __libc_start_main (/lib/x86_64-linux-gnu/libc.so.6+0x2409a)
Signed-off-by: Donald Sharp <sharpd@nvidia.com>
When a peer is a peer-group based peer, and the config is inherited
from the peer group, let's ensure that the CONFIG_NODE flag stays
no matter what.
Signed-off-by: Donald Sharp <sharpd@nvidia.com>
Three fixes:
a) When calling bgp_fsm_change_status with `Deleted` do
not add a new event to the peer's t_event because
we are already in the process of deleting everything
b) When bgp_stop decides to delete a peer return a notification
that it is happening to bgp_event_update so that it does not
set the peer state back to idle or do other processing.
c) bgp_event_update can cause a peer deletion, because
the peer can be deleted in the fsm function but the peer
data structure is still pointed to and used after words.
So lock the peer before entering and prevent a use after
free.
Signed-off-by: Donald Sharp <sharpd@nvidia.com>
At some point in the past the peer creation was not
properly setting the su and the code had the release
and re-add when setting the su. Since peer_create
got a bit of code to handle the su properly the
need to release then add it back in is negated
so remove the code.
Signed-off-by: Donald Sharp <sharpd@nvidia.com>
When multiple daemons report a exit line for the same
node, vtysh was just dropping the previous pointer.
Ensure that we free it first.
Signed-off-by: Donald Sharp <sharpd@nvidia.com>
