The FRR bgp topotests are employing a luCommand that looks for bgp peering
to be up on the first router with a `wait` sub-command. Please note that
a variety of tests are using this. This wait command has a variety of time
outs being used `30`, `90`, and `300`. BGP peering with how we compile
it have very long timers and 30( and possibly 90) seconds is clearly not enough when
we are waiting, given the nature of our test beds. Additionally we were employing a model
where once the first summary command succeeded we automatically assumed that all
subsuquent summary commands( to look at other routers ) would not need to
possibly wait. This is insufficient in that if I have multiple peerings in
multiple vrf's there is no guarantee that one router peers being up will
be sufficient information to know that all the other routers peers are up.
Modify the test cases to be a bit more conformant about this and to
allow peer checks to actually wait a reasonable amount of time for
all peers to have a chance to come up.
Signed-off-by: Donald Sharp <sharpdc@cumulusnetworks.com>
Allow a local build of a frr docker container to be built with
`--enable-dev-build`. This allows better decodes of symbols
which could be useful when you are trying to fix something
that is broken inside the docker container.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
Arm platforms are crashing in our topotests with this callstack;
50 ../sysdeps/unix/sysv/linux/raise.c: No such file or directory.
[Current thread is 1 (Thread 0xffffabb591d0 (LWP 18947))]
(gdb) bt
file=file@entry=0xaaaadfed1e48 "lib/memory.c", line=line@entry=80,
function=function@entry=0xaaaadfed1db8 <__func__.10514> "mt_count_free") at lib/log.c:837
(gdb)
So we are crashing because we are attempting to free a mtype that has no allocations
associated with it.
I added this debug code:
@@ -227,7 +230,9 @@ static void rcu_bump(void)
struct rcu_next *rn;
rn = XMALLOC(MTYPE_RCU_NEXT, sizeof(*rn));
-
+ zlog_debug("RCU_BUMP");
+ mtype_dump(MTYPE_RCU_THREAD);
+ mtype_dump(MTYPE_RCU_NEXT);
/* note: each RCUA_NEXT item corresponds to exactly one seqno bump.
* This means we don't need to communicate which seqno is which
* RCUA_NEXT, since we really don't care.
and added a mtype_dump function:
+void mtype_dump(struct memtype *mt)
+{
+ zlog_debug("%s: %d", mt->name, (int)mt->n_alloc);
+}
Which resulted in this output:
2019/08/28 15:41:11 BGP: RCU_BUMP
2019/08/28 15:41:11 BGP: RCU thread: 3
2019/08/28 15:41:11 BGP: RCU thread: 3
If we look at the defintion of the two static memory types:
DEFINE_MTYPE_STATIC(LIB, RCU_THREAD, "RCU thread")
DEFINE_MTYPE_STATIC(LIB, RCU_NEXT, "RCU sequence barrier")
I would have expected the output to be:
RCU_BUMP
RCU thread: 3
RCU sequence barrier: X
instead.
As a thought experiment I reduced the number of static memory types
to 1 in the file and the crash stopped happening.
I suspect we have a systematic error on arm in lib/memory.h
due to the asm code. I am going to leave that alone for the
moment ( and leave the crash issue open ), but see if we
can get this code change into the system so that our CI
system becomes happy again.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
even if vty commands were available, the default resolution command was
working only for the first vrf configured. others were ignored. Also,
for nexthop, resolution was working for all vrfs, and not the specific
one.
Signed-off-by: Philippe Guibert <philippe.guibert@6wind.com>
We have this crash:
2019-08-18T07:58:44.831656-04:00 rch2-140-fwK2b bgpd[1791]: %NOTIFICATION: sent to neighbor 10.73.248.8 4/0 (Hold Timer Expired) 0 bytes
2019-08-18T07:58:44.832164-04:00 rch2-140-fwK2b bgpd[1791]: Assertion `!((peer->thread_flags) & ((1 << 0)))' failed in file bgpd.c, line 2173, function peer_delete
2019-08-18T07:58:44.832548-04:00 rch2-140-fwK2b bgpd[1791]: Backtrace for 11 stack frames:
2019-08-18T07:58:44.832942-04:00 rch2-140-fwK2b bgpd[1791]: [bt 0] /usr/lib/libfrr.so.0(zlog_backtrace+0x3a) [0x7f5503c7c31a]
2019-08-18T07:58:44.833311-04:00 rch2-140-fwK2b bgpd[1791]: [bt 1] /usr/lib/libfrr.so.0(_zlog_assert_failed+0x61) [0x7f5503c7c891]
2019-08-18T07:58:44.833684-04:00 rch2-140-fwK2b bgpd[1791]: [bt 2] /usr/lib/frr/bgpd(peer_delete+0x4d5) [0x1432ceea15]
2019-08-18T07:58:44.834095-04:00 rch2-140-fwK2b bgpd[1791]: [bt 3] /usr/lib/frr/bgpd(+0x430e9) [0x1432cfc0e9]
2019-08-18T07:58:44.834479-04:00 rch2-140-fwK2b bgpd[1791]: [bt 4] /usr/lib/frr/bgpd(bgp_event_update+0x121) [0x1432cfe1c1]
2019-08-18T07:58:44.834852-04:00 rch2-140-fwK2b bgpd[1791]: [bt 5] /usr/lib/frr/bgpd(+0x453f1) [0x1432cfe3f1]
2019-08-18T07:58:44.835388-04:00 rch2-140-fwK2b bgpd[1791]: [bt 6] /usr/lib/libfrr.so.0(thread_call+0x60) [0x7f5503c9e3c0]
2019-08-18T07:58:44.835829-04:00 rch2-140-fwK2b bgpd[1791]: [bt 7] /usr/lib/libfrr.so.0(frr_run+0xb8) [0x7f5503c79de8]
2019-08-18T07:58:44.836292-04:00 rch2-140-fwK2b bgpd[1791]: [bt 8] /usr/lib/frr/bgpd(main+0x229) [0x1432ce4a69]
2019-08-18T07:58:44.836729-04:00 rch2-140-fwK2b bgpd[1791]: [bt 9] /lib/x86_64-linux-gnu/libc.so.6(__libc_start_main+0xf5) [0x7f550271bb45]
2019-08-18T07:58:44.837198-04:00 rch2-140-fwK2b bgpd[1791]: [bt 10] /usr/lib/frr/bgpd(+0x2cefc) [0x1432ce5efc]
2019-08-18T07:58:44.837670-04:00 rch2-140-fwK2b bgpd[1791]: Current thread function (bgp_holdtime_timer), scheduled from file bgp_fsm.c, line 380
This is the code:
bgp_reads_off(peer);
bgp_writes_off(peer);
assert(!CHECK_FLAG(peer->thread_flags, PEER_THREAD_WRITES_ON));
assert(!CHECK_FLAG(peer->thread_flags, PEER_THREAD_READS_ON));
The line crashing is the first assert. We know in bgp_writes_off we unset this flag:
void bgp_writes_off(struct peer *peer)
{
struct frr_pthread *fpt = bgp_pth_io;
assert(fpt->running);
thread_cancel_async(fpt->master, &peer->t_write, NULL);
THREAD_OFF(peer->t_generate_updgrp_packets);
UNSET_FLAG(peer->thread_flags, PEER_THREAD_WRITES_ON);
}
We also know that the keepalives are not being turned off until we call
bgp_fsm_change_status(peer, Deleted);
later in the function. We know that the keepalive pthread will
write to individual peers and issue a bgp_write_on(), which sets
this flag.
Modify the code base so that we explicitly turn off the keepalives
immediately before the turning of writes off.
Ticket: CM-26119
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
The FRR community has run into an issue where keeping up our
CI system to work with solaris has become a fairly large burden.
We have also sent emails and asked around and have not found
anyone standing up saying that they are using Solaris.
Given the fact that we do not have any comprehensive testing
being done w/ solaris and the fact that we are getting a steady
stream of new features that will never work on solaris and
we cannot find anyone to say that they are using it. Let's
start the drawn out process of deprecating the code.
If in the mean-time someone comes forward with the fact that
they are using it we can then not deprecate it.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
Problem:
With advertise_svi_ip knob enabled per vni.
Post vni flap, svi MAC-IP route are not originated.
Fix:
When a vni is flapped, upon re-add
send advetise_svi_ip knob to zebra.
Workaround:
re-configure advertise-svi-ip under l2vpn/evpn.
Ticket:CM-26001
Reviewed By:CCR-9118
Testing Done:
With advertise-svi-ip enabled under l2vpn/evpn
in bgp default instance.
Validated vni del/create post ifdown vxlan device
followed by ifup vxlan device.
Signed-off-by: Chirag Shah <chirag@cumulusnetworks.com>
Evpn extended communities like auto rts (import/export) should
check if its present in list before adding it, to avoid duplicate
addition.
L3vni_add callback from zebra to bgp may see updates to vnis.
The auto import/export rt derivation may call multiple times.
Testing Done:
Before:
TORC11# show bgp l2vpn evpn vni 4001
VNI: 4001 (known to the kernel)
Type: L3
Tenant VRF: vrf1
RD: 45.0.2.2:3
...
Import Route Target:
5546:4001
5546:4001
Export Route Target:
5546:4001
5546:4001
After:
VNI: 4001 (known to the kernel)
Type: L3
Tenant VRF: vrf1
RD: 45.0.2.2:3
...
Import Route Target:
5546:4001
Export Route Target:
5546:4001
Signed-off-by: Chirag Shah <chirag@cumulusnetworks.com>
as part of the 'ip router isis TAG' command we were not validating
the MTU of the interface against the minimum LSP MTU of the area.
This could cause an assertion when the circuit is created in the
APPLY phase.
Fixes issue #4825
Signed-off-by: Emanuele Di Pascale <emanuele@voltanet.io>
In a number of places, the JSON output had invalid key names for
AFI/SAFI. For example, the key name in JSON was "IPv4 Unicast" which
is invalid as a JSON Key name. Many JSON tools such as those used in
Ansible, jq etc. all fail to parse the output in these scenarios. The
valid name is ipv4Unicast. There's already a routine afi_safi_json()
defined to handle this change, but it was not consistently called.
The non-JSON version was called afi_safi_print() and it merely returned
the CLI version of the string, didn't print anything.
This patch deals with this issue by:
- Renaming afi_safi_print to get_afi_safi_str()
- get_afi_safi_str takes an additional param, for_json which if true
will return the JSON-valid string
- Renaming afi_safi_json to get_afi_safi_json_str()
- Creating a new routine get_afi_safi_vty_str() for printing to vty
- Consistently using get_afi_safi_str() with the appropriate for_json
value
Signed-off-by: Dinesh G Dutt <5016467+ddutt@users.noreply.github.com>
Original Idea is to display normal & detailed debugs when detailed
debug alone is configured. because of this "sh debugs" are showing
wrong Information, because same macro is used to disply the configured
debugs.
that means even if Normal debug is configured, detailed macro returns
TRUE. To avoid this ambiguity check whetehr detailed debug is configured
or not during dumping configured debugs. In all other places using
old macro.
Signed-off-by: Satheesh Kumar K <sathk@cumulusnetworks.com>
The `set as-path prepend last-as X` command had no, 'no' form
of the command. Add this into the cli.
Testing:
!
route-map BLARBLE permit 10
set as-path prepend last-as 3
!
!
router bgp 9999
neighbor 10.50.12.118 remote-as external
neighbor 10.50.12.118 ebgp-multihop 30
!
address-family ipv4 unicast
neighbor 10.50.12.118 route-map BLARBLE in
!
!
eva# show bgp ipv4 uni 4.4.4.4
BGP routing table entry for 4.4.4.4/32
Paths: (1 available, best #1, table default)
Advertised to non peer-group peers:
10.50.12.118
999 999 999 999
10.50.12.118 from 10.50.12.118 (10.50.12.118)
Origin incomplete, metric 0, valid, external, best (First path received)
Last update: Mon Aug 26 09:47:17 2019
eva# conf
eva(config)# route-map BLARBLE permit 10
eva(config-route-map)# no set as-path prepend last-as 3
eva(config-route-map)# end
eva# clear bgp ipv4 uni *
eva# show bgp ipv4 uni 4.4.4.4
BGP routing table entry for 4.4.4.4/32
Paths: (1 available, best #1, table default)
Advertised to non peer-group peers:
10.50.12.118
999
10.50.12.118 from 10.50.12.118 (10.50.12.118)
Origin incomplete, metric 0, valid, external, best (First path received)
Last update: Mon Aug 26 09:48:31 2019
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
FRR has two implementations of VRF, one backed by netns and the other by
the proper VRF implementation in the Linux kernel. In certain places, the
code assumes that a VRF is netns and so lookups fail. One example of this
is in IPv6 RA code. This causes functionality such as Unnumbered BGP to
fail. To fix this, this patch makes if_lookup_by_index handle the
behavior based on the backend, similar to if_get_by_index. For the two
places in if.c that were calling if_lookup_by_index to be specific to
the VRF, I renamed the existing code, if_lookup_by_ifindex and made it a
static function that is never exposed or called by any routine outside of
if.c.
Signed-off-by: Dinesh G Dutt <5016467+ddutt@users.noreply.github.com>
When user wants to dump individual large-community-list with the name
then bgp throws an error. It is due to command to dump the bgp RIB routes
having a particular large-community-list values. To segregate both the
commands this fix has added the detail keyword in the below command.
show bgp large-community-list <(1-500)|WORD> detail
The same code change is applicable for community-list also.
Signed-off-by: vishaldhingra<vdhingra@vmware.com>
Couple code paths end up trying to dereference vty->of which can be null
in one special case.
Signed-off-by: Quentin Young <qlyoung@cumulusnetworks.com>
The `show bgp l2vpn evpn route type <es|prefix>` commands
only accepted 2 letters. You could not type `show bgp l2vpn evpn route type e`
or `show bgp l2vpn evpn route type p` although both are technically legal
since nothing overlaps with them.
Ticket: CM-25988
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
Move neighbor programming to the dataplane; remove
old apis; remove some ifdef'd use of direct netlink
code points, using neutral values outside of the netlink-
specific files.
Signed-off-by: Mark Stapp <mjs@voltanet.io>
re->nexthop_num and re->nexthop_active_num are calculated while rib
processing. Also It helps in encoding the ZAPI message.
It's good to dump these parameters also, when the system is in
abnormal state.
Signed-off-by: vishaldhingra<vdhingra@vmware.com>
Make isisd create BFD sessions over IPv6 when IS-IS is configured
for IPv6 operation only.
When IS-IS is enabled for both IPv4 and IPv6 on a given interface,
prefer creating a BFD session over IPv6 to avoid having two BFD
sessions protecting the same IS-IS adjacency.
Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
