Merge pull request #762 from bingen/mutiple_nh_recursive_levels

Allow for more than 1 NH recursion level

Merge pull request #804 from opensourcerouting/fixes/tests/prng

tests: update prng to return better pseudo random numbers

Merge pull request #799 from dwalton76/bgpd-show-labeled-unicast

bgpd: show bgp ipv4 label-unicast does not provide any output

Merge pull request #808 from qlyoung/vtysh-termcols

lib, vtysh: pretty-print variable autocompletions

Merge pull request #800 from qlyoung/fix-old-vpn-commands

bgpd: fix old-style vpn commands

Merge pull request #815 from dwalton76/CMD_WARNING_take2

vtysh: return non-zero for configuration failures

Merge pull request #817 from donaldsharp/babel_redist

babeld: Allow babel to redistribute v6 routes

Merge pull request #814 from qlyoung/vtysh-password-commands

vtysh: fixup password commands

babeld: Allow babel to redistribute v6 routes

Turn on the ability in babel to receive v6 routes when requested.

robot# show babel route
192.168.201.0/24 metric 0 (exported)
2606:a000:111d:803e::/64 metric 0 (exported)
robot# conf t
robot(config)# int enp3s0
robot(config-if)# ipv6 addr 2404:4000:991d:804c:c32:ee94:742c:4d5/73
robot(config-if)# end
robot# show babel route
192.168.201.0/24 metric 0 (exported)
2606:a000:111d:803e::/64 metric 0 (exported)
2404:4000:991d:804c:c00::/73 metric 0 (exported)
robot#

Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>

vtysh: return non-zero for configuration failures

Signed-off-by: Daniel Walton <dwalton@cumulusnetworks.com>
This allows frr-reload.py (or anything else that scripts via vtysh)
to know if the vtysh command worked or hit an error.

vtysh: fixup password commands

`password` and `enable password` commands got a bit broken in an earlier
commit, fix 'em up.

Signed-off-by: Quentin Young <qlyoung@cumulusnetworks.com>

bgpd: show bgp ipv4 label-unicast does not provide any output

Signed-off-by: Daniel Walton <dwalton@cumulusnetworks.com>

lib, vtysh: pretty-print variable autocompletions

Pretty-prints variable autocompletions by breaking them up into multiple
lines, indenting them consistently and respecting the column width of
the terminal.

Signed-off-by: Quentin Young <qlyoung@cumulusnetworks.com>

Merge pull request #807 from Jafaral/qua2frr

Find/Replace Quagga/FRR

zebra: Fix compilation warning

Signed-off-by: ßingen <bingen@voltanet.io>

lib, test: Add test case for ALL_NEXTHOPS_RO

Without it, tests were passing before commit 9fb47c0, so a case was
missing (the one which that commit fixes).

Signed-off-by: ßingen <bingen@voltanet.io>

Merge pull request #798 from dwalton76/CMD_RANGE_STR

bgpd: --enable-multipath=256 setting is ignored

Merge pull request #796 from qlyoung/fix-hash-stats

lib: use 32-bit atomics, s/pow/mul

doc: update a few places to say FRR

Signed-off-by: Jafar Al-Gharaibeh <jafar@atcorp.com>

zebra: rename to frrconfdir

Signed-off-by: Jafar Al-Gharaibeh <jafar@atcorp.com>

tools: update to use FRR name

Signed-off-by: Jafar Al-Gharaibeh <jafar@atcorp.com>

gitignore: update references to FRR

Signed-off-by: Jafar Al-Gharaibeh <jafar@atcorp.com>

tests: update prng to return better pseudo random numbers

Signed-off-by: Christian Franke <chris@opensourcerouting.org>

lib: fix error message to reference FRR

Signed-off-by: Jafar Al-Gharaibeh <jafar@atcorp.com>

lib: Change comment to reference watchfrr

Signed-off-by: Jafar Al-Gharaibeh <jafar@atcorp.com>

watchfrr: fix error message

Signed-off-by: Jafar Al-Gharaibeh <jafar@atcorp.com>

zebra: prettier format for multilevel-recursive routes

Signed-off-by: Christian Franke <chris@opensourcerouting.org>

zebra: use copy_nexthops in route_entry_copy_nexthops

Signed-off-by: Christian Franke <chris@opensourcerouting.org>

lib: copy rmap_src too in copy_nexthops

Signed-off-by: Christian Franke <chris@opensourcerouting.org>

lib,zebra: Fix ALL_NEXTHOPS iterator

Signed-off-by: Christian Franke <chris@opensourcerouting.org>

config: switch a few references to say FRR

Signed-off-by: Jafar Al-Gharaibeh <jafar@atcorp.com>

Merge pull request #791 from LabNConsulting/working/master/patch-set/style

A couple of compatibility/style fixes

Merge pull request #793 from opensourcerouting/hash-double-init

lib: init & cleanup fixes

Merge pull request #784 from Jafaral/debpkgfix

Debian pkg cleanup

Update comments for ALL_NEXTHOPS_RO definition

Signed-off-by: ßingen <bingen@voltanet.io>

bgpd: fix old-style vpn commands

Remove:
-------
* set vpnv4 next-hop X:X::X:X
* set vpnv6 next-hop A.B.C.D
* no set vpn next-hop <A.B.C.D|X:X::X:X>

Add:
----
* no set vpnv4 next-hop A.B.C.D
* no set vpnv6 next-hop X:X::X:X

Signed-off-by: Quentin Young <qlyoung@cumulusnetworks.com>

bgpd: show bgp ipv4 label-unicast does not provide any output

Signed-off-by: Daniel Walton <dwalton@cumulusnetworks.com>
Reviewed-by: Donald Sharp <sharpd@cumulusnetworks.com>
labeled-unicast routes are installed in the unicast table, to make
troubleshooting easier we need "show bgp ipv4 labeled-unicast x.x.x.x"
to show the x.x.x.x route in the unicast table.

doc: add Backwards Compatibility section to COMMUNITY.md

Signed-off-by: Lou Berger <lberger@labn.net>

lib: use 32-bit atomics, s/pow/mul

Some platforms don't support 64-bit atomics, missed converting a
floating point pow() to an integral mul when changing SD algo.

Signed-off-by: Quentin Young <qlyoung@cumulusnetworks.com>

vty: add some more defines for backwards compatibility

Signed-off-by: Lou Berger <lberger@labn.net>

bgpd: --enable-multipath=256 setting is ignored

Signed-off-by: Daniel Walton <dwalton@cumulusnetworks.com>
Reviewed-by: Donald Sharp <sharpd@cumulusnetworks.com>
Restoring some code that was unintentionally removed when we were
migrating to Quentin's parser.

Merge pull request #795 from opensourcerouting/build-fix-openbsd-tree-path

minor build-fixes

lib: grammar_sandbox: don't pre-setup nodegraph

This shows up as still reachable allocation on exit, which is somewhat
misleading.  Note this only affects --enable-dev-build.

Signed-off-by: David Lamparter <equinox@opensourcerouting.org>

lib: drop off "masters" list on master_free()

Don't leave a dangling leaked reference.

Signed-off-by: David Lamparter <equinox@opensourcerouting.org>

build: fix ttable test

test_ttable.refout wasn't in the dist package.

Signed-off-by: David Lamparter <equinox@opensourcerouting.org>

lib: fix out-of-tree build

"$(top_srcdir)" is not on the include path, but "$(top_srcdir)/lib" is.
This is relevant when building with a separate build directory.

Signed-off-by: David Lamparter <equinox@opensourcerouting.org>

lib: hash: don't double-init hash list

hash_cmd_init will overwrite _hashes with a new list, while _hashes
already has been initialised from cmd_init(), thread_master_create(), or
any other function that may have created a hash.

Found while valgrind'ing ospf6d/test_lsdb.

Signed-off-by: David Lamparter <equinox@opensourcerouting.org>

vty: add some defines for backwards compatibility

Signed-off-by: Lou Berger <lberger@labn.net>

bgpd: have labeled unicast print consistent with other safis

Signed-off-by: Lou Berger <lberger@labn.net>

Merge branch 'evpn-prep'

First 12-and-a-half commits from PR #619

Signed-off-by: David Lamparter <equinox@opensourcerouting.org>

Merge pull request #786 from dwalton76/bgpd-peergroup-activate

bgpd: peer-group members 'activate' when they shouldn't

Avoid to make indentation even worse

lib: Remove typedef from ipaddr

The ipaddr_t type was conflicting with code on omnios.

Remove the typedef

Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>

Merge commit '3d22338f04d9554fa' into evpn-prep

Conflicts:
lib/Makefile.am

Signed-off-by: David Lamparter <equinox@opensourcerouting.org>

Remove recursive param from netlink_route_info_add

..._nh

Rename ALL_NEXTHOPS_RO to ALL_NEXTHOPS

Merge pull request #733 from dwalton76/bgpd-ipv4-plus-label-misc3

bgpd: Install SAFI_LABELED_UNICAST routes in SAFI_UNICAST table

bgpd: Install SAFI_LABELED_UNICAST routes in SAFI_UNICAST table

Signed-off-by: Daniel Walton <dwalton@cumulusnetworks.com>

Fix bug with NH macro in zebra_fpm_protobuf

One occurrence was missing to be adpated the new macro definition for
multiple NH recursion levels.

Merge branch 'master' of https://github.com/dwalton76/frr into bgpd-ipv4-plus-label-misc3

Conflicts:
bgpd/bgp_route.c
bgpd/bgp_zebra.c

Fix tests for NH multiple recursion levels

Tests were no passing. Besides, added cases to test more than 1 level
depth in nexthop recursion.

Merge pull request #750 from donaldsharp/bgp_buffer

Bgp buffer

bgpd: peer-group members 'activate' when they shouldn't

Signed-off-by: Daniel Walton <dwalton@cumulusnetworks.com>
Reviewed-by: Donald Sharp <sharpd@cumulusnetworks.com>
Today if you configure the following where the 'fabric' peer-group has
been deactivated for ipv4 unicast and then assign swp1/swp2 to that
peer-group we end up activating those two peers for ipv4 unicast.

conf t
no router bgp 100
router bgp 100
  neighbor fabric peer-group
  neighbor fabric capability extended-nexthop
  neighbor fabric remote-as external
  !
  address-family ipv4 unicast
    no neighbor fabric activate
  !
  neighbor swp1 interface peer-group fabric
  neighbor swp2 interface peer-group fabric
  neighbor 1.1.1.1 peer-group fabric
end

cel-redxp-10# show run bgp
!
router bgp 100
 neighbor fabric peer-group
 neighbor fabric remote-as external
 neighbor fabric capability extended-nexthop
 neighbor swp1 interface peer-group fabric
 neighbor swp2 interface peer-group fabric
 neighbor 1.1.1.1 peer-group fabric
 !
 address-family ipv4 unicast
  no neighbor fabric activate
  neighbor swp1 activate
  neighbor swp2 activate
 exit-address-family
!
cel-redxp-10#

With the patch we do not activate swp1/swp2

cel-redxp-10# show run bgp
!
router bgp 100
 neighbor fabric peer-group
 neighbor fabric remote-as external
 neighbor fabric capability extended-nexthop
 neighbor swp1 interface peer-group fabric
 neighbor swp2 interface peer-group fabric
 neighbor 1.1.1.1 peer-group fabric
 !
 address-family ipv4 unicast
  no neighbor fabric activate
 exit-address-family
!
cel-redxp-10#

config: No need to test these if enabled. Plays havoc with make distribution.

Signed-off-by: Ryan Hagelstrom <rhagelstrom@gmail.com>

docs: Changed the manpage section from 1 to 8 so it conforms with its definition

Signed-off-by: Ryan Hagelstrom <rhagelstrom@gmail.com>

tools: etc needed in extra dist to make distribution correctly

Signed-off-by: Ryan Hagelstrom <rhagelstrom@gmail.com>

build: Hard directory instead of dynamic so make distribution works correctly

Merge pull request #778 from qlyoung/fix-excess-docstring

Fix excess docstring

Merge pull request #779 from qlyoung/update-docs

doc: update build docs for ubuntu

lib: fix more docstrings

Signed-off-by: Quentin Young <qlyoung@cumulusnetworks.com>

Merge pull request #777 from qlyoung/fix-mpls-str2label

zebra: fix mpls_str2label()

Merge pull request #781 from opensourcerouting/vty-out-bufresize

lib: fix vty_out with >1024 bytes of output

lib: fix vty_out with >1024 bytes of output

Consuming va_args modifies its internal bits, hence the need to copy
it... but the copying wasn't quite right just yet.

Fixes: 4d5f445 ("lib: add vty_outln()")
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>

Merge pull request #780 from dwalton76/bgpd-address-family-vty-out

bgpd: "address-family" not displayed in configuration

bgpd: Install SAFI_LABELED_UNICAST routes in SAFI_UNICAST table

Signed-off-by: Daniel Walton <dwalton@cumulusnetworks.com>

bgpd: "address-family" not displayed in configuration

Signed-off-by: Daniel Walton <dwalton@cumulusnetworks.com>
Reviewed-by: Donald Sharp <sharpd@cumulusnetworks.com>

Merge branch 'master' of https://github.com/dwalton76/frr into bgpd-ipv4-plus-label-misc3

*: fix excess docstring

Signed-off-by: Quentin Young <qlyoung@cumulusnetworks.com>

doc: update build docs for ubuntu

cumulus/etc --> tools/etc

Signed-off-by: Quentin Young <qlyoung@cumulusnetworks.com>

lib: warn about too much docstring

Signed-off-by: Quentin Young <qlyoung@cumulusnetworks.com>

zebra: fix mpls_str2label()

When making improvements to error handling in this code I accidentally
introduced an off-by-one. Fix it.

Signed-off-by: Quentin Young <qlyoung@cumulusnetworks.com>

Merge pull request #742 from qlyoung/hashstats

Hashtable statistics

lib, vtysh: rebase hashstats

Signed-off-by: Quentin Young <qlyoung@cumulusnetworks.com>

lib: use doubles instead of long doubles

NetBSD can't take the square root of a long double
and we should be fine just using a double here anyway

Signed-off-by: Quentin Young <qlyoung@cumulusnetworks.com>

*: update hash_create(), hash_create_size()

Signed-off-by: Quentin Young <qlyoung@cumulusnetworks.com>

lib, vtysh: hashtable statistics

Adds the ability to name hash tables, and a new cli command that will
show various summary statistics for named hash tables.

Statistics computed are
  - load factor
  - full load factor (see comments)
  - stddev of full load factor

Standard deviation is computed by storing the sum of squares of bucket
lengths. This is somewhat susceptible to overflow. On platforms where a
double is 32 bits, placing 65535 or more elements into a hash table
opens up the potential for overflow, depending on how they are arranged
in buckets (which depends on the hash function). For example, placing
65535 elements into one hash bucket would cause ssq overflow, but
distributing 40000000 elements evenly among 400000 buckets (100 elements
per bucket) would not.

These cases are extremely degenerate, so the vague possibility of
overflow in an informational command is deemed an acceptable tradeoff
for constant time calculation of variance without locks or compromising
efficiency of actual table operations.

Signed-off-by: Quentin Young <qlyoung@cumulusnetworks.com>

lib: add statistics for hash tables

Adds a function that calculates various statistics on our implementation
of a hash table. These are useful for evaluating performance.

Signed-off-by: Quentin Young <qlyoung@cumulusnetworks.com>

Merge pull request #771 from qlyoung/printf-madness

lib: printf bugfixes & improvement

bgpd: Install SAFI_LABELED_UNICAST routes in SAFI_UNICAST table

Signed-off-by: Daniel Walton <dwalton@cumulusnetworks.com>

Merge branch 'master' of https://github.com/dwalton76/frr into bgpd-ipv4-plus-label-misc3

Conflicts:
bgpd/bgp_route.c

Merge pull request #740 from donaldsharp/ospf_commands

ospfd: Fix 'advanced' type data base show

lib: printf bugfixes & improvement

* Fix bad format specifier in thread.[ch]
* Move PRINTF_ATTRIBUTE macro to zebra.h
* Use PRINTF_ATTRIBUTE on termtable printers

Signed-off-by: Quentin Young <qlyoung@cumulusnetworks.com>

Merge pull request #768 from qlyoung/fix-gitignore2

lib, bgpd: fix .gitignore

Merge pull request #770 from donaldsharp/zclient

Zclient

ospfd: Fix 'advanced' type data base show

Two things:

1) Some advanced 'show ip ospf database'... code
was not working properly.  Fix the cli to handle
this appropriately

2) Consolidate two cli functions into one.

Signed-off-by: Donald Sharp <sharpd@cumulusnetworkscom>

Merge pull request #723 from qlyoung/fix-static-pipe-poker-buf

lib: more multithreading infra work

Merge pull request #769 from qlyoung/fix-json-int

lib: always use 64-bit integers for json

ospf6d: Fix zclient cleanup on shutdown

Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>

ospfd: Fix zclient cleanup on shutdown

Signed-off-by: Donald Sharp <sharpd@cumulunsnetworks.com>

pimd: Fix zclient cleanup on shutdown

Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>