The return from sockunion2hostprefix tells us if the conversion
succeeded or not. There are places in the code where we
always assume that it just `works`, since it can fail
notice and try to do the right thing.
Please note that failure of this function for most cases
of sockunion2hostprefix is highly highly unlikely as that
the sockunion was already created and tested elsewhere
it's just that this function can fail.
Signed-off-by: Donald Sharp <sharpd@nvidia.com>
When setting authentication on a BGP peer in a VRF the listener is
looked up from a global list. However there is no check that the
listener is the one associated with the VRF being configured. This
can result in the wrong listener beiong configured with a password,
leaving the intended listener in an open authentication state.
To simplify this lookup stash a pointer to the bgp instance in
the listener on creating (in the same way as is done for NS-based
VRFS).
Signed-off-by: Pat Ruddy <pat@voltanet.io>
* Fixed integration in FSM and packet handling.
* Added CLI "show" output, incl. JSON.
* For review and testing only.
Signed-off-by: David Schweizer <dschweizer@opensourcerouting.org>
Remove mid-string line breaks, cf. workflow doc:
.. [#tool_style_conflicts] For example, lines over 80 characters are allowed
for text strings to make it possible to search the code for them: please
see `Linux kernel style (breaking long lines and strings)
<https://www.kernel.org/doc/html/v4.10/process/coding-style.html#breaking-long-lines-and-strings>`_
and `Issue #1794 <https://github.com/FRRouting/frr/issues/1794>`_.
Scripted commit, idempotent to running:
```
python3 tools/stringmangle.py --unwrap `git ls-files | egrep '\.[ch]$'`
```
Signed-off-by: David Lamparter <equinox@diac24.net>
bgp_accept() gets called over and over again when a VRF device is
deleted out from under a bgp listener socket that is bound to it.
Prevent this by noting the error and cancelling ourselves, allowing the
vrf status code to clean up the mess when it receives word about the
change from Zebra.
Signed-off-by: Quentin Young <qlyoung@cumulusnetworks.com>
Try to give a bit more useful data about where we
think the connection is trying to come in from.
Hopefully this will let us debug connection issues
a bit faster in cases where there are config issues.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
We are crashing in thread_cancel on shutdown because
the thread pointer is NULL. Use the more appropriate
THREAD_CANCEL macro
Ticket: CM-29873
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
The act of peer_sort() being called always set this value
even when we are just looking it up. We need to seperate
out the idea of lookup from set.
For those places that this is immediately obvious that
this is a lookup switch over to using this function.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
*Adding helper caller hooks function for signalling from BGPD
to ZEBRA to enable or disable GR feature in ZEBRA depending
on bgp per peer gr configuration.
Signed-off-by: Biswajit Sadhu <sadhub@vmware.com>
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.
Signed-off-by: Biswajit Sadhu <sadhub@vmware.com>
Add -s X or --socket_size X to the bgp cli to allow
the end user to specify the outgoing bgp tcp kernel
socket buffer size.
It is recommended that this option is only used on
large scale operations.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
Under high load instances with hundreds of thousands of prefixes this
could result in very unstable systems.
When maximum-prefix is set, but restart timer is not set then the session
flaps between Idle(Pfx) -> Established -> Idle(Pfx) states.
Signed-off-by: Donatas Abraitis <donatas.abraitis@gmail.com>
When using the maximum-prefix restart option with a BGP peer,
if the peer exceeds the limit of prefixes, bgpd causes the
connection to be closed and sets a timer. It will not attempt
to connect to that peer until the timer expires. But if the
peer attempts to connect to it before the timer expires, it
accepts the connection and starts exchanging routes again.
When accepting a connection from a peer, reject the connection
if the max prefix restart timer is set.
Signed-off-by: Matthew Smith <mgsmith@netgate.com>
the vrf_id parameter is replaced by struct vrf * parameter.
this impacts most of the daemons that look for an interface based on the
name and the vrf identifier.
Also, it fixes 2 lookup calls in zebra and sharpd, where the vrf_id was
ignored until now.
Signed-off-by: Philippe Guibert <philippe.guibert@6wind.com>
The SO_MARK socket option was being used pre vrf to allow for the
separation of the front panel -vs- the management port. This
was facilitated by a ip rule. Since this is undocumented anywhere
in our system( other than old commits see
ed40466af8 ). We should remove this
because this will cause interference with people using rules
and are not aware of this offshoot of functionality.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
Co-authored-by: Donald Sharp <sharpd@cumulusnetworks.com>
Co-authored-by: Quentin Young <qlyoung@cumulusnetworks.com>
Signed-off-by: Quentin Young <qlyoung@cumulusnetworks.com>
Now that all daemons receive the VRF backend from zebra, we can get
rid of vrf_is_mapped_on_netns() in favor of using the more convenient
vrf_is_backend_netns() function, which doesn't require any argument.
This commit also fixes the following problem:
debian(config)# ip route 50.0.0.0/8 blackhole vrf FAKE table 2
% table param only available when running on netns-based vrfs
Even when zebra was started with the --vrfwnetns, the error
above would be displayed since the VRF FAKE didn't exist, which
would make vrf_is_mapped_on_netns() return 0 incorrectly. Using
vrf_is_backend_netns() this problem doesn't happen anymore.
Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
if zebra is not started, then vrf identifiers are not available. This
prevents import/exportation to be available. This commit permits having
import/export available, even when zebra is not started.
Signed-off-by: Philippe Guibert <philippe.guibert@6wind.com>
Cleanup calls where we were passing in the su for
peer creation a tiny bit.
Creating a peer from the cli will always have a conf_if *or*
a su but not both. While a doppelganger will have both.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
The peer->nexthop.ifp pointer must be set when parsing the
attributes in bgp_mp_reach_parse, notice this
and fail gracefully.
Rework bgp_nexthop_set to remove the HAVE_CUMULUS and to
fail the nexthop_set when we have a zebra connection and
no ifp pointer, as that not havinga zebra connection and
no ifp pointer is legal.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
Instead of relying on local usage of vrf bind operation, the vrf API for
that usage is done.
Signed-off-by: Philippe Guibert <philippe.guibert@6wind.com>
fixup bgp
Because socket creation is tightly linked with socket binding for vrf
lite, the proposal is made to extend socket creation APIs and to create
a new API called vrf_bind that applies to vrf lite. The passed interface
name is the interface that will be bound to the socket passed.
Signed-off-by: Philippe Guibert <philippe.guibert@6wind.com>
Upon creation of BGP instances, server socket may or may not be created.
In the case of VRF instances, if the VRF backend relies on NETNS, then
a new server socket will be created for each BGP VRF instance. If the
VRF backend relies on VRF LITE, then only one server socket will be
enough. Moreover, At startup, with BGP VRF configuration, a server
socket may not be created if VRF is not the default one or VRF is not
recognized yet.
Signed-off-by: Philippe Guibert <philippe.guibert@6wind.com>
The change contained in this commit does the following:
- discovery of vrf id from zebra daemon, and adaptation of bgp contexts
with BGP.
The list of network addresses contain a reference to the bgp context
supporting the vrf.
The bgp context contains a vrf pointer that gives information about
the netns path in case the vrf is a netns path.
Only some contexts are impacted, namely socket creation, and retrieval
of local IP settings. ( this requires vrf identifier).
Signed-off-by: Philippe Guibert <philippe.guibert@6wind.com>
* Start bit flags at 1, not 2
* Make run-flags atomic for i/o thread
* Remove work_cond mutex, it should no longer be necessary
* Add asserts to ensure proper ordering in bgp_connect()
* Use true/false with booleans, not 1/0
Signed-off-by: Quentin Young <qlyoung@cumulusnetworks.com>
When we fail to bind to port 179 we are left in a situation
where we have not saved the bgp pointer created and when
the bgp cli mode is exited we leak the memory.
Additionally there is no recovery situation here that
could be easily programmed without fundamentally changing
the code.
So let's exit and output to the log file some useful
information to hopefully clue the user in on what is
going wrong.
Fixes: #1130
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
Handle better stress situations when multiple peers are trying to
connect at the same time by bumping the TCP connection backlog limit.
This reduces the convergence time of BGPerf stress test.
Signed-off-by: Rafael Zalamena <rzalamena@opensourcerouting.org>
This reverts commit c14777c6bf.
clang 5 is not widely available enough for people to indent with. This
is particularly problematic when rebasing/adjusting branches.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
The FSF's address changed, and we had a mixture of comment styles for
the GPL file header. (The style with * at the beginning won out with
580 to 141 in existing files.)
Note: I've intentionally left intact other "variations" of the copyright
header, e.g. whether it says "Zebra", "Quagga", "FRR", or nothing.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
Pass pointer to pointer instead of assigning by return value. See
previous commit message.
To ensure that the behavior stays functionally correct, any assignments
with the result of a thread_add* function have been transformed to set
the pointer to null before passing it. These can be removed wherever the
pointer is known to already be null.
Signed-off-by: Quentin Young <qlyoung@cumulusnetworks.com>
The way thread.c is written, a caller who wishes to be able to cancel a
thread or avoid scheduling it twice must keep a reference to the thread.
Typically this is done with a long lived pointer whose value is checked
for null in order to know if the thread is currently scheduled. The
check-and-schedule idiom is so common that several wrapper macros in
thread.h existed solely to provide it.
This patch removes those macros and adds a new parameter to all
thread_add_* functions which is a pointer to the struct thread * to
store the result of a scheduling call. If the value passed is non-null,
the thread will only be scheduled if the value is null. This helps with
consistency.
A Coccinelle spatch has been used to transform code of the form:
if (t == NULL)
t = thread_add_* (...)
to the form
thread_add_* (..., &t)
The THREAD_ON macros have also been transformed to the underlying
thread.c calls.
Signed-off-by: Quentin Young <qlyoung@cumulusnetworks.com>
If we fail to set any socket's buffer size, try again with a smaller value
and keep going until it succeeds. This is better than just giving up or,
even worse, abort the creation of a socket (ospf6d and ripd).
Fix broken ospf6d on FreeBSD.
Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
Description:
We use valgrind memcheck quite a bit to spot leaks in
our work with bgpd. In order to eliminate false positives,
we added code in the exit path to release the remaining
allocated memory.
Bgpd startup log message now includes pid.
Some little tweaks by Paul Jakma <paul.jakma@hpe.com>:
* bgp_mplsvpn.c: (str2prefix_rd) do the cleanup in common code at the end
and goto it.
[DL: dropped several chunks from original commit which are obsolete by
now on this tree.]
If bgp gets inbound connect messages on an interface associated with
a vrf, but the vrf is not defined yet in bgp, the log is filled with
continual error messages. This change moves that error message to a
debug under "debug bgp neighbor-events". Manual testing results
applied to the ticket.
Ticket: CM-10394
Signed-off-by: Don Slice
Reviewed-by: Donald Sharp
Reverts the --enable-bgp-standalone and makes it so that you
need to use --enable-cumulus to get the cumulus behavior.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When getsockopt(...,SO_BINDTODEVICE,...); fails
assume the bgp instance we are interested is the default
one.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
lib/zebra.h has FILTER_X #define's. These do not belong there.
Put them in lib/filter.h where they belong.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
(cherry picked from commit 0490729cc033a3483fc6b0ed45085ee249cac779)
Only Linux has SO_BINDTODEVICE, but that's not a problem since the whole
VRF use case in that instance is currently Linux-specific. Other OS's
VRF implementations will need different code.
Reported-by: Martin Winter <mwinter@opensourcerouting.org>
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
As the comments in if.h, it is better to call ifname2ifindex()
instead of if_nametoindex().
And ifname2ifindex() can work for VRF by appending a parameter
while if_nametoindex() can not.
Signed-off-by: Feng Lu <lu.feng@6wind.com>
Reviewed-by: Alain Ritoux <alain.ritoux@6wind.com>
Signed-off-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
Acked-by: Vincent JARDIN <vincent.jardin@6wind.com>
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
(cherry picked from commit 395828eea809e8b2b8c5824d3639cefedd7aa9f0)
Avoids a dynamic allocation which is usually freed immediate afterwards.
Signed-off-by: Timo Teräs <timo.teras@iki.fi>
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
When you shutdown interfaces with ifdown -a -X eth0
bgp would still attempt to talk to it's configured
neighbors but since the interface is down it would
just complain and complain. Now that we have
somewhat aggressive timers the error message
really starts to show up.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
Reviewed-by: Daniel Walton <dwalton@cumulusnetworks.com>
Basically when modifying the peer->su, we must *always*
release the hash and then re-install it, else
we will cause crashes when we go to look up data
that is not going to be there.
Ticket: CM-10212
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
Reviewed-by: Don Slice <dslice@cumulusnetworks.com>
Reviewed-by: Daniel Walton <dwalton@cumulusnetworks.com>
When creating a 'struct peer' add in the ability to set the peer group
associated with that peer.
Ticket: CM-10184
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
Reviewed-by: Daniel Walton <dwalton@cumulusnetworks.com>
Upon receipt of incoming connection, a peer structure (doppelganger) is
created internally and the connection processed for it. The problem is
that in the case of BGP unnumbered, the sockunion structure within BGP was
being updated (in peer_create()) prior to the peer's flags being updated,
so it didn't take into account the 'v6only' configuration. This results
in subsequent problems when bgp_bind() is done - the socket ends up being
bound to the BGP instance instead of the interface.
In the case of an incoming connection, we should just use the addresses
on which the connection was setup/accepted, there is no need to attempt to
derive it again. Further, there is no need to attempt to update addresses
at the time of peer_create() since that is done when the connection is
attempted in bgp_start().
Signed-off-by: Vivek Venkatraman <vivek@cumulusnetworks.com>
Reviewed-by: Don Slice <dslice@cumulusnetworks.com>
Ticket: CM-10028
Reviewed By: CCR-4373
Testing Done: Manual, bgpsmoke
Code changes to make unnumbered peering work in a VRF. The changes have
to do with locating the interface in the correct VRF (in order to look for
neighbor address) in the case of outgoing connections and when specifying
source address as well as fetching the correct instance for an incoming
connection based on reading the device the socket is bound to (the multi-vrf
socket option in the kernel).
Additionally, for IPv4 unnumbered peering in a VRF (based on /30 or /31
addresses), bind to the VRF rather than the interface.
Signed-off-by: Vivek Venkatraman <vivek@cumulusnetworks.com>
Reviewed-by: Donald Sharp <sharpd@cumulusnetworks.com>
Ticket: CM-9311
Reviewed By: CCR-4192
Testing Done: Manual, bgpsmoke
Added debug requirement to issue the Bind to connect messages that were filling the logs.
Ticket:CM-9463
Signed-off-by: Don Slice
Reviewed-by: Donald Sharp
Ticket: CM-7861
Reviewed by: CCR-3651
Testing: See bug
bgp is using both bm->master and master pointers interchangebly
for thread manipulation. Since they are the same thing consolidate
to one pointer.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
Ticket: CM-6369
Reviewed By: CCR-3318
Testing Done: Manual testing of various password scenarios.
This is a port of patch bgpd-unnumbered-nbr-fix-password.patch from
2.5-br.
In the case of BGP unnumbered, the peer IP address is derived and not
explicitly configured. If there is a password configured for the peer,
it can be set on the listen socket only after the IP address has been
derived and needs to be cleared when the IP address goes away.
Ticket: CM-6883
Reviewed By: CCR-3272
Testing Done: Tested on 2.5.3-SE-1
This commit is a port of the patch bgpd-handle-peer-local-address-failure.patch
from 2.5-br.
When a peering is being established, the IPv4 and IPv6 addresses of the
local end of the connection, as applicable, are obtained and stored in
the peer's 'nexthop' structure to facilitate filling of the NEXT_HOP
field in Update messages among other things. The process of obtaining the
local address involves examination of the list of interfaces to identify
a match corresponding to the socket address of the connection.
There are timing conditions, especially when BGPD starts with a config,
where the interface may not have reached BGP from Zebra at the time a
peering reaches the state to determine the local addresses. The code does
not handle this well and the result could be Updates generated with bad
(Martian) NEXT_HOP values. Resolve the issue by bringing down the connection
in this case as not identifying the local addresses is really an error.
deleted from the listen socket in some situations. This would lead
to incorrect behavior where a BGP connection from a peer that does
not specify the MD5 option would be accepted.
For IPv6, update source sometimes fails at the first attempt. If we continue
ignoring the error, some sessions will not come up. If instead we check for
the error and return connect_error, the FSM will reset its state and try again
till the update source bind succeeds and the session will come up.
This patch adds checking for the result of bind and update_source to return
connect_error or success. The rest of the code handles the situation correctly
after that.
Signed-off-by: Dinesh G Dutt <ddutt@cumulusnetworks.com>
Reviewed-by: Donald Sharp <sharpd@cumulusnetworks.com>
Reviewed-by: Vivek Venkataraman <vivek@cumulusnetworks.com>
BGP: Set SO_MARK on connecting sockets to ensure lookup in right routing table
In the presence of a function such as management VRF/MRF, bgpd needs to be able
to specify that it intends to run in the dataplane and not the front panel.
To ensure this, we add a mark in the connecting socket so that the kernel
does the routing lookup in the right table. This assumes that an appropriate
ip rule has been configured (outside the scope of this package).
While we've forced the mark to be 254 for now, it maybe required to make it
configurable at a later time.
BGP: Fix EBGP multihop transitions correctly
Since BGP connection setup has migrated to using NHT to decide when to bring a
session up, we have to handle ebgp multihop transitions correctly to ensure NHT
registrations are correctly handled.
Signed-off-by: Dinesh G Dutt <ddutt@cumulusnetworks.com>
bgp: Fixup of the remote-as command to allow user to not have to enter an actual as number
Signed-off-by: Donald Sharp<sharpd@cumulusnetworks.com>
Reviewed-by:
peer - one initiated by the local system and the other initiated by the peer.
Enhance key debug logs to also print the socket file descriptor so that it is
clear which events pertain to which connection.
is not activated for any address-family, the connection is accepted without
taking further action. This causes the connection to hang in OpenSent on the
neighbor and can in turn delay the connection setup. Fix to reject incoming
connections when there is no address-family activated for the neighbor.
sockunion_same() and bgp_peer_conf_if_to_su_update() need to use the scope_id
field of the ipv6 address to uniquify/identify the address.
This allows sessions based on link local address when that address is not
unique across peers.
sessions dynamically. The operator configures a range of neighbor addresses
to which peering is allowed. The ranges are configured as subnets and
multiple ranges are allowed. Each range is associated with a peer-group
so that additional parameters can be configured.
BGP neighbor sessions are dynamically created when connections are initiated
by remote neighbors whose addresses fall within a configured range. The
sessions are deleted when the BGP connection terminates.
A limit on the number of neighbors allowed from each range of addresses
can be specified.
IPv4 and IPv6 peering is supported. Over the peering, any of the address
families configured for the peer-group can be negotiated.
This patch implements the 'update-groups' functionality in BGP. This is a
function that can significantly improve BGP performance for Update generation
and resultant network convergence. BGP Updates are formed for "groups" of
peers and then replicated and sent out to each peer rather than being formed
for each peer. Thus major BGP operations related to outbound policy
application, adj-out maintenance and actual Update packet formation
are optimized.
BGP update-groups dynamically groups peers together based on configuration
as well as run-time criteria. Thus, it is more flexible than update-formation
based on peer-groups, which relies on operator configuration.
[Note that peer-group based update formation has been introduced into BGP by
Cumulus but is currently intended only for specific releases.]
From 11098af65b2b8f9535484703e7f40330a71cbae4 Mon Sep 17 00:00:00 2001
Subject: [PATCH] updgrp commits
Summary of changes
- added an option to enable keepalive debugs for a specific peer
- added an option to enable inbound and/or outbound updates debugs for a specific peer
- added an option to enable update debugs for a specific prefix
- added an option to enable zebra debugs for a specific prefix
- combined "deb bgp", "deb bgp events" and "deb bgp fsm" into "deb bgp neighbor-events". "deb bgp neighbor-events" can be enabled for a specific peer.
- merged "deb bgp filters" into "deb bgp update"
- moved the per-peer logging to one central log file. We now have the ability to filter all verbose debugs on a per-peer and per-prefix basis so we no longer need to keep log files per-peer. This simplifies troubleshooting by keeping all BGP logs in one location. The use
r can then grep for the peer IP they are interested in if they wish to see the logs for a specific peer.
- Changed "show debugging" in isis to "show debugging isis" to be consistent with all other protocols. This was very confusing for the user because they would type "show debug" and expect to see a list of debugs enabled across all protocols.
- Removed "undebug" from the parser for BGP. Again this was to be consisten with all other protocols.
- Removed the "all" keyword from the BGP debug parser. The user can now do "no debug bgp" to disable all BGP debugs, before you had to type "no deb all bgp" which was confusing.
The new parse tree for BGP debugging is:
deb bgp as4
deb bgp as4 segment
deb bgp keepalives [A.B.C.D|WORD|X:X::X:X]
deb bgp neighbor-events [A.B.C.D|WORD|X:X::X:X]
deb bgp nht
deb bgp updates [in|out] [A.B.C.D|WORD|X:X::X:X]
deb bgp updates prefix [A.B.C.D/M|X:X::X:X/M]
deb bgp zebra
deb bgp zebra prefix [A.B.C.D/M|X:X::X:X/M]
