This feature adds an L3 & L2 VPN application that makes use of the VPN
and Encap SAFIs. This code is currently used to support IETF NVO3 style
operation. In NVO3 terminology it provides the Network Virtualization
Authority (NVA) and the ability to import/export IP prefixes and MAC
addresses from Network Virtualization Edges (NVEs). The code supports
per-NVE tables.
The NVE-NVA protocol used to communicate routing and Ethernet / Layer 2
(L2) forwarding information between NVAs and NVEs is referred to as the
Remote Forwarder Protocol (RFP). OpenFlow is an example RFP. For
general background on NVO3 and RFP concepts see [1]. For information on
Openflow see [2].
RFPs are integrated with BGP via the RF API contained in the new "rfapi"
BGP sub-directory. Currently, only a simple example RFP is included in
Quagga. Developers may use this example as a starting point to integrate
Quagga with an RFP of their choosing, e.g., OpenFlow. The RFAPI code
also supports the ability import/export of routing information between
VNC and customer edge routers (CEs) operating within a virtual
network. Import/export may take place between BGP views or to the
default zebera VRF.
BGP, with IP VPNs and Tunnel Encapsulation, is used to distribute VPN
information between NVAs. BGP based IP VPN support is defined in
RFC4364, BGP/MPLS IP Virtual Private Networks (VPNs), and RFC4659,
BGP-MPLS IP Virtual Private Network (VPN) Extension for IPv6 VPN . Use
of both the Encapsulation Subsequent Address Family Identifier (SAFI)
and the Tunnel Encapsulation Attribute, RFC5512, The BGP Encapsulation
Subsequent Address Family Identifier (SAFI) and the BGP Tunnel
Encapsulation Attribute, are supported. MAC address distribution does
not follow any standard BGB encoding, although it was inspired by the
early IETF EVPN concepts.
The feature is conditionally compiled and disabled by default.
Use the --enable-bgp-vnc configure option to enable.
The majority of this code was authored by G. Paul Ziemba
<paulz@labn.net>.
[1] http://tools.ietf.org/html/draft-ietf-nvo3-nve-nva-cp-req
[2] https://www.opennetworking.org/sdn-resources/technical-library
Now includes changes needed to merge with cmaster-next.
FIB override routes are for routing protocols that establish
shortcut routes, or establish point-to-point routes that should
not be redistributed. Namely this is useful NHRP daemon to come.
Zebra is extended to select two entries from RIB the "best" entry
from routing protocols, and the FIB entry to install to kernel.
FIB override routes are never selected as best entry, and thus
are never adverticed to other routing daemons. The best FIB
override, or if it does not exist the otherwise best RIB is
selected as FIB entry to be installed.
Signed-off-by: Timo Teräs <timo.teras@iki.fi>
Acked-by: Donald Sharp <sharpd@cumulusnetworks.com>
[CF: Massage to fit cumulus tree]
Signed-off-by: Christian Franke <chris@opensourcerouting.org>
When compiling with the json library instead of the json-c
library allow it to switch back to the old way of printing.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
Whenever a routing daemon advertises a new version of a route to zebra,
zebra removes the old version of this route (implicit withdraw) and then
create a new 'rib' structure for the new version of the route.
In this process, the previously received label(s) from ldpd are lost. This
is because upon receiving a ZEBRA_MPLS_LABELS_ADD message, zebra only
adds a label to a nexthop of an existing route. And routes are volatile,
they can be removed while being updated.
To workaround this issue, this patch makes ldpd always advertise the
appropriate labels whenever it receives a redistributed route, even
if it was already received before (an older version). This way, when
ldpd receives the updated version of a route, it will readvertise the
appropriate label(s) and zebra will reinstall them.
Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
Commit 5048fe changed the way zebra behave when a route is updated. Now,
whenever a route is changed, zebra advertises its new version without
withdrawing the old one. This patch adapts ldpd to understand this new
behavior. After processing a ZEBRA_REDISTRIBUTE_IPV[46]_ADD message,
we need to check for nexthops that were removed and, for each of them
(if any), withdraw the associated labels from zebra.
Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
zclient_stop(), which is used as antagonist to zclient_init(), needs to
undo the vrf_bitmap allocation. Otherwise zclient_init() will leak the
allocated memory, for example when zclient_reset() is used.
Reported-by: Lou Berger <lberger@labn.net>
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
Signed-off-by: Daniel Walton <dwalton@cumulusnetworks.com>
Reviewed-by: Donald Sharp <sharpd@cumulusnetworks.com>
NCLU imports quagga-reload.py and uses its Config class to parse
Quagga.conf. The Config class will call 'vtysh -m -f Quagga.conf" and
if that exited with an error Config would call sys.exit(1) which in my
cases causes the NCLU daemon to exit which is bad. The fix is to have
the Config class raise an exception instead of exiting, then NCLU can
catch the exception, log it and move on.
(cherry picked from commit 276887bb1c)
Signed-off-by: Daniel Walton <dwalton@cumulusnetworks.com>
Reviewed-by: Don Slice <dslice@cumulusnetworks.com>
Ticket: CM-13020
Now that we have evpn we have the following AFIs
/* Address family numbers from RFC1700. */
typedef enum {
AFI_IP = 1,
AFI_IP6 = 2,
AFI_ETHER = 3, /* RFC 1700 has "6" for 802.* */
AFI_MAX = 4
} afi_t;
The import-table code was treating the afi as a flag which was fine
before when the only choices were 1 and 2 but now that we have #3 that
doesn't work. The fix is to change zebra_import_table_used to a
[AFI_MAX][ZEBRA_KERNEL_TABLE_MAX] array to track if import-table is
enabled.
Signed-off-by: Daniel Walton <dwalton@cumulusnetworks.com>
Reviewed-by: Donald Sharp <sharpd@cumulusnetworks.com>
NCLU imports quagga-reload.py and uses its Config class to parse
Quagga.conf. The Config class will call 'vtysh -m -f Quagga.conf" and
if that exited with an error Config would call sys.exit(1) which in my
cases causes the NCLU daemon to exit which is bad. The fix is to have
the Config class raise an exception instead of exiting, then NCLU can
catch the exception, log it and move on.
When displaying thread time for long running/busy
protocols, the space allocated may not be sufficient.
Allow the runtime to take a bit more space.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
distribute.c doesn't allow to manage both v4 and v6 distribute lists. This
patch fix this problem by having 4 DISTRIBUTE* values in the enumeration instead
of two. The code in all daemons using distribute.c is adapted.
* zclient.c: prefix length on router-id and interface address add
messages not sanity checked. fix.
* */*_zebra.c: Prefix length on zebra route read was not checked, and
clients use it to write to storage. An evil zebra could overflow
client structures by sending overly long prefixlen.
Prompted by discussions with:
Donald Sharp <sharpd@cumulusnetworks.com>
Linux/glibc sys/param.h has definitions of MIN/MAX that result in
multiple evaluations of its parameters. Force local definitions.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
BABEL was removed, ifname nexthops were removed, additional includes
were needed, and lastly the protobuf enum-handling triggers a warning.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
Add test functions for the zebra code that interfaces with the
Forwarding Plane Manager. These functions can be invoked in a
development build via the recently-added 'invoke' command.
For example:
# invoke zebra function zfpm_dt_benchmark_protobuf_encode 100000
Changes:
* zebra/zebra_fpm_dt.c
Add the following functions. Each function encodes or decodes a
route in a particular FPM format a specified number of times.
- zfpm_dt_benchmark_netlink_encode()
- zfpm_dt_benchmark_protobuf_encode()
- zfpm_dt_benchmark_protobuf_decode()
* zebra/Makefile.am
Compile zebra_fpm_dt when building a development build.
Signed-off-by: Avneesh Sachdev <avneesh@sproute.com>
* configure.ac
Add the --enable-dev-build flag. It controls the DEV_BUILD
define for autoconf and automake, which can be used to
conditionally build in code that is only intended for development..
Signed-off-by: Avneesh Sachdev <avneesh@sproute.com>
Change zebra so that it can optionally use protobuf serialization when
communicating with a Forwarding Plane Manager component.
* zebra/main.c
Add the --fpm-format/-F command line option. This allows the user
to control the format (protbuf|netlink) that is used to
communicate with the FPM.
* zebra/zebra_fpm.c
- zebra_init_msg_format(),
This new function is invoked on process startup to determine the
FPM format that should be used.
- zfpm_init()
Change to accept any 'FPM message format' specified by the user
(via the new command line flag).
- zebra_encode_route()
Tweak to use the selected FPM format.
* zebra_fpm_protobuf.c
New code to build protobuf messages to be sent to the FPM.
* zebra/Makefile.am
- Include common.am
- Build new file zebra_fpm_protobuf.c when protobuf is available.
- Link with the fpm_pb library.
Signed-off-by: Avneesh Sachdev <avneesh@sproute.com>
Simplify licensing for protobuf files based on discussion on the
quagga-dev mailing list.
Previously, users could choose between the GPL and the ISC license.
Signed-off-by: Avneesh Sachdev <avneesh@sproute.com>
Signed-off-by: Avneesh Sachdev <avneesh@sproute.com>
Code that allows a client to convey routes to a Forwarding Plane
Manager component using protobuf instead of netlink..
* fpm/fpm.proto
Protobuf definitions pertaining to the Forwarding Plane Manager.
In particular, this file defines the AddRoute and DeleteRoute
messages.
* fpm/fpm.h
Tweak FPM message header definition to also allow messages to be
encoded in protobuf format.
* fpm/{fpm_pb.h,.gitignore,.Makefile.am}
Add the fpm_pb library, which contains code for interfacing with
the FPM using protobuf.
* configure.ac
Generate fpm/Makefile.
* Makefile.am
Add fpm subdirectory to build.
* common.am
Add flags to be used by clients of the fpm_pb library.
Signed-off-by: Avneesh Sachdev <avneesh@sproute.com>
Modify configure.ac to disable portability warnings for automake --
our automake code (in particular common.am) uses some constructs
specific to gmake.
Signed-off-by: Avneesh Sachdev <avneesh@sproute.com>
Infrastructure that allows protocol buffers to be used in Quagga. The
changes below comprise of:
- Build hooks
- Protobuf definitions for common types.
- Library routines for working with protobuf, including functions
that help translate between common quagga types and their protobuf
equivalents.
Changes:
* qpb/{Makefile.am,README.txt,qpb.h,.gitignore}
Add the qpb library, which provides shared code and definitions
for using protocol buffers in quagga code.
* qpb/qpb.proto
Protobuf definitions that can be shared by all of quagga.
* qpb/linear_allocator.h
An allocator that allocates memory by walking down towards the end
of a buffer. This is used to cheaply allocate/deallocate memory on
the stack for protobuf operations.
* qpb/qpb_allocator.[ch]
Thin layer that allows a linear allocator to be used with the
protobuf-c library.
* common.am
This is an automake fragment that is intended to be shared by
Makefile.am files in the tree. It currently includes definitions
related to protobuf.
* configure.ac
- Add logic to optionally build protobuf code.
By default, protobuf support is enabled if the protobuf C
compiler (protoc-c) is available, and the associated header
files/library can be found.
The user can choose to override this behavior via the new
--disable-protobuf/--enable-protobuf flags.
- Include the quagga protobuf library (qpb) in the build.
* .gitignore
Ignore source code generated by protobuf compiler.
* Makefile.am
Add 'qpb' to the list of subdirectories.
Signed-off-by: Avneesh Sachdev <avneesh@sproute.com>
Edited: Paul Jakma <paul.jakma@hpe.com>: Change the sense of the
configure enable option to require explicit specifying, as
an experimental feature.
icc (the Intel C Compiler) "knows" that assert() can be disabled by
setting specific optimisation flags, and therefore emits a warning about
missing a return value after an "always-error" assert.
Workaround by returning a value - this probably needs discussion and a
better fix (for all places where the code needs to abort due to internal
errors).
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
FPM aims to provide cross platform mechanism to support the scenario
where the router has forwarding path distinct fromt the kernel.Commonly
Hardware based fast path.Hence it is non-configurable paramter.This
limits us to use funcationality to update FIB information to remote
hosts, like SDN controller.
This implementation provides the CLI to configure remote hosts and port
information of remote fpm controller.Otherwise default fpm server will
be localhost and default fpm port will be well know port 2620.
* zebra_fpm.c: added fpm_server paramter to zfpm_global_t handler.
Implemented CLI for configuring the fpm server and no fpm
command to revert back to default configuration.
* zserv.c: Install zebra node to write fpm configuration info
on console/config file.
Further documentation supplied:
-------------------------------
ZEBRA : CLI CONFIGURATION FOR FPM MODULE
========================================================
1. INTRODUCTION
================================
1.1 scope
This memo discusses the configuration option for zebra to update
FIB information to local and remote modules.
This will also helps to address the issue associated with CORD project.
https://jira.onosproject.org/browse/CORD-411
2. REFERENCE
================================
Quagga version 99.24+ ( main branch committed on 29-sep-2015)
3. PROBLEM DESCRIPTION
================================
Once FPM is enabled, Quagga periodically tries to initiate fpm
connection to localhost:2620. These values are non configurable in
existing implementation. There is no CLI available to configure
"host:port". hence limits us to use it for hardware based fast path
modules only.
4. PROPOSED CHANGES
================================
Following changes are done to the quagga code
a) Added new CLI to configure "host address : port".
The CLI format
<conf t>
$ fpm connection ip <ipv4 address> port <tcp port num>
and no fpm command to revert back to default
<conf t>
$ no fpm connection ip <ipv4 address> port <tcp port num>
b) Allowed values are ipv4 address and tcp port range <1-65535>
c) FPM initialization code has been enhanced to pick the "host
address : port" values from zebra.conf. if not found then
default values as localhost:2620 will be used. and updated the
information on to config file on write config command
5. FILES MODIFIED
================================
1) fpm/fpm.h :
a) Added MACRO to represent network order loopback ip
2) zebra/zebra_fpm.h :
a) introduced fpm_server variable in zfpm_glob_t handler to hold
the remote fpm server address
b) Hooked 'fpm_remote_ip_cmd' and 'no_fpm_remote_ip_cmd' at CONFIG
node to configure remote fpm detail and to revert back to
default respectively
3) zebra/zserv.c :
a) Hooked 'config_write_fpm' callback function, at ZEBRA_NODE to
display the fpm connection details on console on entering
command
$ show running_config
and to write to configuration file on entering command
$ write config
6. TESTING DETAILS
================================
6.1. default behavior
In default configuration FPM will attempt to connect to
localhost:2620
6.2. update fpm info
a) Using CLI command user can configure fpm host:port details
and can be able to write to config file(zebra.conf) using
write config command. this parameters has no
dependency/impact on other parameters of config file
b) show running-config/write config will display the fpm
information if configured. and will not display any
information related to fpm for default configuration
c) these configured information will be stored to config file.
only on write config command.
6.3 loading from config file
a) zebra attempts to connect to fpm server if fpm parameter
found in config file.else connects to default parameters.
b) if fpm connection drops, fpm will periodically attempts to
connect to remote server.
c) if fpm connections already established. then newly
configured fpm parameters will not disconnect the existing
connection. new connection to the different fpm server will
happen only after existing connection closes by either of
the end.
fix fpm prototype
I have a fix for 2 segmentation fault scenarios on the isis daemon:
1. When running a command "isis passive" on an interface in the
following context:
"end"
"configure terminal "
"interface dummy0"
"isis passive"
The trace back collected:
isis_adjacency.c:521
family=2,
root_sysid=0x20aee6d0 "", parent=0x20af4d68) at isis_spf.c:999
sysid=0x20aee6d0 "")
at isis_spf.c:1217
isis_spf.c:1372
isis_lsp.c:416
isis_lsp.c:1660
isis_main.c:368
The fix location:
file name: isisd/isis_adjacency.c
routine name: isis_adj_build_up_list
2. When deleting the existing isis router instance:
"end"
"configure terminal "
"no router isis DEAD"
The fix location:
isisd/isis_events.c, routine circuit_resign_level
isisd/isis_lsp.c, routine lsp_destroy
isisd/isis_route.c, isis_route_validate
The trace back collection:
"DEAD") at isisd.c:252
argc=1, argv=0xbfc39054) at isisd.c:1520
vty=0x20d6f528, cmd=0x0) at command.c:2121
cmd=0x0, vtysh=0) at command.c:2155
isis DEAD") at vty.c:433
isis_main.c:368
and
"DEAD") at isisd.c:260
argc=1, argv=0xbfd6cf54) at isisd.c:1520
vty=0x208cb528, cmd=0x0) at command.c:2121
cmd=0x0, vtysh=0) at command.c:2155
isis DEAD") at vty.c:433
isis_main.c:368
The patch is included.
patchwork #833: http://patchwork.quagga.net/patch/833/
On OpenIndiana/Solaris the build fails with "unsupported multicast API".
It's only in the IPv4 part where setsockopt IP_MULTICAST_IF needs a
local address and not the index (IPv6 wants the index).
The following code walks the list of interfaces until it finds the matching
index and uses the interface's local address for the setsockopt call.
I don't know if it works on Solaris < 10 (I guess yes, but I don't have
any machine to verify it).
[NB: this breaks unnumbered setups that use the same IPv4 address on
multiple interfaces. -- equinox@opensourcerouting.org]
Reported-by: Brian Utterback <brian.utterback@oracle.com>
Signed-off-by: Christian Franke <chris@opensourcerouting.org>
Patchwork #762
* rip_interface.c: Default for split_horizon_default differed between
rip_interface_new and rip_interface_reset, causing at least some issues
after interface events. See patchwork #604. Fix, and consolidate code.
(rip_interface_{reset,clean}) rename these to 'interface', as that's more
appropriate. Spin the ri specific bodies of these functions out to
rip_interface_{reset,clean} helpers. Factor out the overlaps, so
rip_interface_reset uses rip_interface_clean.
(rip_interface_new) just use rip_interface_reset.
* ripd.h: Update for (rip_interface_{reset,clean})
Reported by xufeng zhang, with a suggested fix on which this commit expands.
See patchwork #604. This commit addresses only the split-horizon
discrepency, issue #2. The other issue they reported, #1, is not addressed,
though suggested fix seems inappropriate.
Cc: xufeng.zhang@windriver.com
"DEFUN (" could only match with exactly one space between DEFUN and
opening brace. Allow any amount of space.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
ICC thinks we're storing a pointer and might be losing bits at the top:
error #810: conversion from "void *" to "route_map_event_t={enum
<unnamed>}" may lose significant bits
Build is warning-free on ICC 14.0.3 (for Linux x86_64) with this.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
