bisdhdh [Fri, 25 Oct 2019 15:42:39 +0000 (21:12 +0530)]
bgpd: Adding header files for BGPD-ZEBRA integration for GR.
Data Structures, function declaration and Macros forSignalling
from BGPD to ZEBRA to enable or disable GR feature in ZEBRA
depending on bgp per peer gr configuration.
bisdhdh [Thu, 24 Oct 2019 14:59:43 +0000 (20:29 +0530)]
bgpd: Restarting node does not send EOR after the convergence.
*After a restarting router comes up and the bgp session is
successfully established with the peer. If the restarting
router doesn’t have any route to send, it send EOR to
the peer immediately before receiving updates from its peers.
*Instead the restarting router should send EOR, if the
selection deferral timer is not running OR count of eor received
and eor required are matches then send EOR.
bisdhdh [Thu, 24 Oct 2019 10:21:18 +0000 (15:51 +0530)]
bgpd: Added hidden CLI command to disable sending of End-of-Rib.
BGP disable EOR sending is a useful command for testing various
scenarios of BGP graceful restart.
* Added the hidden CLI command : bgp graceful-restart disable-eor
* The CLI will not be displayed in "show running-config" and will not
be stored in configuration file.
* When enabled, EOR will not be sent to peer
Signed-off-by: Biswajit Sadhu <sadhub@vmware.com> Signed-off-by: Soman K S <somanks@vmware.com>
bisdhdh [Thu, 24 Oct 2019 04:40:53 +0000 (10:10 +0530)]
bgpd: BGP tcp session failed to apply GR configuration on the transferred
bgp tcp connection.
When the BGP peer is configured between two bgp routes both routers would create
peer structure , when they receive each other’s open message. In this event both
speakers, open duplicate TCP sessions and send OPEN messages on each socket
simultaneously, the BGP Identifier is used to resolve which socket should be closed.
If BGP GR is enabled the old tcp session is dumped and the new session is retained.
So while this transfer of connection is happening, if all the bgp gr config
is not migrated to the new connection, the new bgp gr mode will never get applied.
Fix Summary:
1. Replicate GR configuration from the old session to the new session in bgp_accept().
2. Replicate GR configuration from stub to full-fledged peer in bgp_establish().
3. Disable all NSF flags, clear stale routes (if present), stop restart & stale timers
(if they are running) when the bgp GR mode is changed to “Disabled”.
4. Disable R-bit in cap, if it is not set the received open message.
bisdhdh [Thu, 24 Oct 2019 03:34:05 +0000 (09:04 +0530)]
bgpd: show BGP GR Neighbor mode as “NotApplicable”,when local mode is “Disable”.
BGP GR Neighbor mode is showing the default string as “NotRecieved”,
as the bgp gr neighbour capability was not processed,
since the local mode is “Disable”.
However now it would be changed to “NotApplicable”.
bisdhdh [Wed, 23 Oct 2019 19:15:43 +0000 (00:45 +0530)]
bgpd: Fix for BGP core when connected routes are redistributed
& GR is enabled.
When GR with deferral is enabled and connected routes are
distributed then in one race condition route node gets added
in to both deferred queue and work queue. If deferred queue
gets processed first then it ends up delete only flag while
leaving the entry in the work queue as it is. When a new update
comes for the same route node next time from peer then it hits
assert. Assert check is added to ensure we don’t add to work queue
again while it is already present.
So, check before adding in to deferred queue if it is already present
in work queue and bail if so.
bisdhdh [Wed, 23 Oct 2019 19:12:10 +0000 (00:42 +0530)]
bgpd: Adding BGP GR change mode config apply on notification sent & received.
* Changing GR mode on a router needs a session reset from the
SAME router to negotiate new GR capability.
* The present GR implementation needs a session reset after every
new BGP GR mode change.
* When BGP session reset happens due to sending or receiving BGP
notification after changing BGP GR mode, there is no need of
explicit session reset.
bisdhdh [Wed, 23 Oct 2019 19:10:06 +0000 (00:40 +0530)]
bgpd: Adding BGP GR Neighbour mode show, when local mode is “Disable”.
* BGP GR Neighbour mode in show command would show as
“NotApplicable”, when local mode is “Disable”. As the bgp
gr neighbour capability was not processed, since the local mode
is “Disable”.
* Minor changes in show Selection Deferral Time.
* Selection Deferral Timer for Graceful Restart.
* Added selection deferral timer handling function.
* Route marking as selection defer when update message is received.
* Staggered processing of routes which are pending best selection.
* Fix for multi-path test case.
bisdhdh [Wed, 23 Oct 2019 10:02:48 +0000 (15:32 +0530)]
bgpd: Adding BGP GR Per Neighbor show commands.
* Added new show command to show the graceful restart
information for each neighbor.
Cmd: show bgp [<ipv4|ipv6>] neighbors [<A.B.C.D|X:X::X:X|WORD>] graceful-restart
* Changes to show neighbors commands for displaying
graceful restart information.
Cmd :show [ip] bgp [<view|vrf> VIEWVRFNAME] [<ipv4|ipv6>] neighbors [<A.B.C.D|X:X::X:X|
bisdhdh [Wed, 23 Oct 2019 05:40:22 +0000 (11:10 +0530)]
bgpd: Adding BGP GR Global & Per Neighbour FSM changes
* Added FSM for peer and global configuration for graceful restart
* Added debug option BGP_GRACEFUL_RESTART for logs specific to
graceful restart processing
Donald Sharp [Sat, 18 Jan 2020 14:16:10 +0000 (09:16 -0500)]
tests: Add another router to the basic pim tests
Add an additional router to the basic pim tests.
1) This test will add a link between r1 and a new rp
2) This test will ensure that r1 and rp have the expected
who is the rp.
3) This test will ensure that the rp has received the upstream
data for the multicast stream that is started.
Ostensibly commit 68a02e06e5f103048d947262c08c569056f74d1c is the first bad commit
commit 68a02e06e5f103048d947262c08c569056f74d1c
Author: Mark Stapp <mjs@voltanet.io>
Date: Wed Nov 13 16:06:06 2019 -0500
*: revise zapi nexthop encoding
Use a per-nexthop flag to indicate the presence of labels; add
some utility zapi encode/decode apis for nexthops; use the zapi
apis more consistently. Signed-off-by: Mark Stapp <mjs@voltanet.io>
Sparked this commit in that it broke nexthop reporting to upper
level protocols. Ensure that this expectation stays working
in the future.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
This code combined the different types of nexthop encoding
being done in the zapi protocol. What was missed that
resolved nexthops of type NEXTHOP_TYPE_IPV4|6 have an ifindex
value that was not being reported. This commit ensures
that we always send this data( even if it is 0).
The following test commit will ensure that this stays working
as is expected by an upper level protocol.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
Quentin Young [Wed, 15 Jan 2020 18:00:34 +0000 (13:00 -0500)]
bgpd: fix memory leak when parsing capabilities
Duplicated domain name capability messages cause memory leak. The amount
of leaked memory is proportional to the size of the duplicated
capabilities. This bug was introduced in 2015.
To hit this, a BGP OPEN message must contain multiple FQDN capabilities.
Memory is leaked when the hostname portion of the capability is of
length 0, but the domainname portion is not, for any of the duplicated
capabilities beyond the first one.
Don Slice [Mon, 23 Dec 2019 16:18:50 +0000 (08:18 -0800)]
tools: improve frr-reload delete performance for some commands
Problem seen when deleting many static routes or access-lists due
to frr-reload.py issuing individual vtysh -c commands for every
line. On slow switches, this can take long enough for systemd to
time out the reload process and restart frr. This fix uses add
logic for static routes, prefix-lists, and access-lists to gang
the changes together.
Signed-off-by: Don Slice <dslice@cumulusnetworks.com>
Ticket: CM-27856
Stephen Worley [Tue, 31 Dec 2019 22:10:58 +0000 (17:10 -0500)]
tests: add basic nexthop group functionality test
Add a very basic nexthop group functionality test.
This test creates a 2-way ecmp group and installs a route
with it using sharpd. Then we check to see that the nexthop
groups are marked valid/installed in zebra.
Signed-off-by: Stephen Worley <sworley@cumulusnetworks.com>
Quentin Young [Tue, 14 Jan 2020 06:18:44 +0000 (01:18 -0500)]
zebra: fix iptable install heap UAF
My previous patch to fix a memory leak, caused by not properly freeing
the iptable iface list on stream parse failure, created/exposed a heap
use after free because we were not doing a deep copy
Signed-off-by: Quentin Young <qlyoung@cumulusnetworks.com>
Stephen Worley [Mon, 13 Jan 2020 19:28:29 +0000 (14:28 -0500)]
lib: assert if someone adds to nexthop list to nhg
If someone tries to add a nexthop with a list of nexthops
already attached to it, let's just assert. This standardizes
the API to say we assume this is an individual nexthop
you are appending to a group.
Signed-off-by: Stephen Worley <sworley@cumulusnetworks.com>
Stephen Worley [Fri, 3 Jan 2020 17:35:15 +0000 (12:35 -0500)]
zebra: just set nexthop member in handle_recursive_depend()
With recent changes to the lib nexthop_group
APIs (e1f3a8eb193267da195088cc515b598ae5a92a12), we are making
new assumptions that this should be adding a single nexthop
to a group, not a list of nexthops.
This broke the case of a recursive nexthop resolving to a group:
```
D> 2.2.2.1/32 [150/0] via 1.1.1.1 (recursive), 00:00:09
* via 1.1.1.1, dummy1 onlink, 00:00:09
via 1.1.1.2 (recursive), 00:00:09
* via 1.1.1.2, dummy2 onlink, 00:00:09
D> 3.3.3.1/32 [150/0] via 2.2.2.1 (recursive), 00:00:04
* via 1.1.1.1, dummy1 onlink, 00:00:04
K * 10.0.0.0/8 [0/1] via 172.27.227.148, tun0, 00:00:21
```
This group can instead just directly point to the nh that was passed.
Its only being used for a lookup (the memory gets copied and used
elsewhere if the nexthop is not found).
Signed-off-by: Stephen Worley <sworley@cumulusnetworks.com>
Stephen Worley [Mon, 13 Jan 2020 18:29:58 +0000 (13:29 -0500)]
lib,zebra: tighten up the nexthop_copy/nexthop_dup APIs
Make the nexthop_copy/nexthop_dup APIs more consistent by
adding a secondary, non-recursive, version of them. Before,
it was inconsistent whether the APIs were expected to copy
recursive info or not. Make it clear now that the default is
recursive info is copied unless the _no_recurse() version is
called. These APIs are not heavily used so it is fine to
change them for now.
Signed-off-by: Stephen Worley <sworley@cumulusnetworks.com>
!
nexthop-group red
nexthop 1.1.1.1
nexthop 1.1.1.2
!
sharp install routes 8.8.8.1 nexthop-group red 1
=========================================
==11898== Invalid write of size 8
==11898== at 0x48E53B4: _nexthop_add_sorted (nexthop_group.c:254)
==11898== by 0x48E5336: nexthop_group_add_sorted (nexthop_group.c:296)
==11898== by 0x453593: handle_recursive_depend (zebra_nhg.c:481)
==11898== by 0x451CA8: zebra_nhg_find (zebra_nhg.c:572)
==11898== by 0x4530FB: zebra_nhg_find_nexthop (zebra_nhg.c:597)
==11898== by 0x4536B4: depends_find (zebra_nhg.c:1065)
==11898== by 0x453526: depends_find_add (zebra_nhg.c:1087)
==11898== by 0x451C4D: zebra_nhg_find (zebra_nhg.c:567)
==11898== by 0x4519DE: zebra_nhg_rib_find (zebra_nhg.c:1126)
==11898== by 0x452268: nexthop_active_update (zebra_nhg.c:1729)
==11898== by 0x461517: rib_process (zebra_rib.c:1049)
==11898== by 0x4610C8: process_subq_route (zebra_rib.c:1967)
==11898== Address 0x0 is not stack'd, malloc'd or (recently) free'd
Zebra crashes because we weren't handling the case of the depend nexthop
being recursive.
For this case, we cannot make the function more efficient. A nexthop
could resolve to a group of any size, thus we need allocs/frees.
To solve this and retain the goal of the original patch, we separate out the
two cases so it will still be more efficient if the nexthop is not recursive.
Signed-off-by: Stephen Worley <sworley@cumulusnetworks.com>
Mitchell Skiba [Thu, 9 Jan 2020 19:46:13 +0000 (11:46 -0800)]
bgpd: add addpath ID to adj_out tree sort
When withdrawing addpaths, adj_lookup was called to find the path that
needed to be withdrawn. It would lookup in the RB tree based on subgroup
pointer alone, often find the path with the wrong addpath ID, and return
null. Only the path highest in the tree sent to the subgroup could be
found, thus withdrawn.
Adding the addpath ID to the sort criteria for the RB tree allows us to
simplify the logic for adj_lookup, and address this problem. We are able
to remove the logic around non-addpath subgroups because the addpath ID
is consistently 0 for non-addpath adj_outs, so special logic to skip
matching the addpath ID isn't required. (As a side note, addpath will
also never use ID 0, so there won't be any ambiguity when looking at the
structure content.)
Donald Sharp [Wed, 15 Jan 2020 02:03:11 +0000 (21:03 -0500)]
zebra: Do not accept illegal safi's for route installation
The only two safi's that are usable for zebra for installation
of routes into the rib are SAFI_UNICAST and SAFI_MULTICAST.
The acceptance of other safi's is causing a memory leak:
Direct leak of 56 byte(s) in 1 object(s) allocated from:
#0 0x5332f2 in calloc (/usr/lib/frr/zebra+0x5332f2)
#1 0x7f594adc29db in qcalloc /opt/build/frr/lib/memory.c:110:27
#2 0x686849 in zebra_vrf_get_table_with_table_id /opt/build/frr/zebra/zebra_vrf.c:390:11
#3 0x65a245 in rib_add_multipath /opt/build/frr/zebra/zebra_rib.c:2591:10
#4 0x7211bc in zread_route_add /opt/build/frr/zebra/zapi_msg.c:1616:8
#5 0x73063c in zserv_handle_commands /opt/build/frr/zebra/zapi_msg.c:2682:2
Collapse
Sequence of events:
Upon vrf creation there is a zvrf->table[afi][safi] data structure
that tables are auto created for. These tables only create SAFI_UNICAST
and SAFI_MULTICAST tables. Since these are the only safi types that
are zebra can actually work on. zvrf data structures also have a
zvrf->otable data structure that tracks in a RB tree other tables
that are created ( say you have routes stuck in any random table
in the 32bit route table space in linux ). This data structure is
only used if the lookup in zvrf->table[afi][safi] fails.
After creation if we pass a route down from an upper level protocol
that has non unicast or multicast safi *but* has the actual
tableid of the vrf we are in, the initial lookup will always
return NULL leaving us to look in the otable. This will create
a data structure to track this data.
If after this event you pass in a second route with the same
afi/safi/table_id, the otable will be created and attempted
to be stored, but the RB_TREE_UNIQ data structure when it sees
this will return the original otable returned and the lookup function
zebra_vrf_get_table_with_table_id will just drop the second otable.
Signed-off-by: Quentin Young <qlyoung@cumulusnetworks.com> Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
Donald Sharp [Wed, 15 Jan 2020 03:38:49 +0000 (22:38 -0500)]
zebra: Fix label manager memory leak
==25402==ERROR: LeakSanitizer: detected memory leaks
Direct leak of 16 byte(s) in 1 object(s) allocated from:
#0 0x533302 in calloc (/usr/lib/frr/zebra+0x533302)
#1 0x7fee84cdc80b in qcalloc /home/qlyoung/frr/lib/memory.c:110:27
#2 0x5a3032 in create_label_chunk /home/qlyoung/frr/zebra/label_manager.c:188:3
#3 0x5a3c2b in assign_label_chunk /home/qlyoung/frr/zebra/label_manager.c:354:8
#4 0x5a2a38 in label_manager_get_chunk /home/qlyoung/frr/zebra/label_manager.c:424:9
#5 0x5a1412 in hook_call_lm_get_chunk /home/qlyoung/frr/zebra/label_manager.c:60:1
#6 0x5a1412 in lm_get_chunk_call /home/qlyoung/frr/zebra/label_manager.c:81:2
#7 0x72a234 in zread_get_label_chunk /home/qlyoung/frr/zebra/zapi_msg.c:2026:2
#8 0x72a234 in zread_label_manager_request /home/qlyoung/frr/zebra/zapi_msg.c:2073:4
#9 0x73150c in zserv_handle_commands /home/qlyoung/frr/zebra/zapi_msg.c:2688:2
When creating label chunk that has a specified base, we eventually are
calling assign_specific_label_chunk. This function finds the appropriate
list node and deletes it from the lbl_mgr.lc_list but since
the function uses list_delete_node() the deletion function that is
specified for lbl_mgr.lc_list is not called thus dropping the memory.
Signed-off-by: Quentin Young <qlyoung@cumulusnetworks.com> Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
Donald Sharp [Fri, 10 Jan 2020 20:13:36 +0000 (15:13 -0500)]
zebra: Actually add the NLA_F_NESTED flag to our code
The existing usage of the rta_nest and addattr_nest
functions were not adding the NLA_F_NESTED flag
to the type. As such the new nexthop functionality was
actually looking for this flag, while apparently older
code did not.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
projects / mirror_frr.git / log