Merge branch 'pr/2986'

Conflicts:
bgpd/bgp_network.c (BGP_ERR_* -> EC_BGP_*)

Signed-off-by: David Lamparter <equinox@diac24.net>

Merge pull request #3040 from pacovn/static_analysis__drop_const_1

bgpd isisd ldpd lib: const drop fixes (SA)

bgpd isisd ldpd lib: const drop fixes (SA)

Can be detected with e.g. ./configure CFLAGS=-Wcast-qual CC=clang

Signed-off-by: F. Aragon <paco@voltanet.io>

Merge pull request #3031 from pacovn/static_analysis__Wcomma

bgpd isisd zebra: misuse of comma operator

Merge pull request #3036 from donaldsharp/kernel_is_a_mernel

Kernel is a mernel

Merge pull request #3035 from donaldsharp/the_table_in_all_of_us

staticd: Allow table_id to be a distinguisher for installation

zebra: Newer kernels send data about mroute changes

Newer linux kernels apparently send data down the netlink
bus for the creation of mroutes.  Add a bit of code
to notice this and to handle it appropriately( ie do
nothing at this point in time ) as that the correct
place to do this is in the pim socket in pimd.

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

zebra: Print out netlink message type instead of number

When we are displaying data about a netlink message
in debugs or errors, print out the message type
as a string instead of a number.

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

staticd: Allow table_id to be a distinguisher for installation

The table_id should be a discriminator in the installation of
static routes into zebra from staticd.  Add this to allow the end
user to do something like this:

ip route 4.5.6.7/32 192.168.209.44
ip route 4.5.6.7/32 192.168.209.44 table 3000
ip route 4.5.6.7/32 192.168.209.45 table 3000

robot# show ip route
Codes: K - kernel route, C - connected, S - static, R - RIP,
       O - OSPF, I - IS-IS, B - BGP, E - EIGRP, N - NHRP,
       T - Table, v - VNC, V - VNC-Direct, A - Babel, D - SHARP,
       F - PBR, f - OpenFabric,
       > - selected route, * - FIB route

K>* 0.0.0.0/0 [0/101] via 192.168.201.1, enp0s3, 00:01:40
S>* 4.3.2.1/32 [1/0] via 192.168.210.4, enp0s10, 00:01:35
S>* 4.3.2.2/32 [1/0] via 192.168.209.4, enp0s9, 00:01:35
S>* 4.5.6.0/26 [1/0] via 192.168.210.4, enp0s10, 00:01:35
S>* 4.5.6.7/32 [1/0] via 192.168.209.44, enp0s9, 00:01:35
C>* 192.168.201.0/24 is directly connected, enp0s3, 00:01:40
C>* 192.168.208.0/24 is directly connected, enp0s8, 00:01:40
C>* 192.168.209.0/24 is directly connected, enp0s9, 00:01:40
C>* 192.168.210.0/24 is directly connected, enp0s10, 00:01:40
robot# show ip route table 3000
Codes: K - kernel route, C - connected, S - static, R - RIP,
       O - OSPF, I - IS-IS, B - BGP, E - EIGRP, N - NHRP,
       T - Table, v - VNC, V - VNC-Direct, A - Babel, D - SHARP,
       F - PBR, f - OpenFabric,
       > - selected route, * - FIB route

S>* 4.5.6.7/32 [1/0] via 192.168.209.44, enp0s9, 00:00:55
  *                  via 192.168.209.45, enp0s9, 00:00:55
robot#

Fixes: #2954
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>

Merge pull request #3018 from donaldsharp/twheel_names

Twheel names

Merge pull request #3030 from pacovn/fp_type_fixes1

bgpd: type fix

Merge pull request #3023 from qlyoung/ultimate-warning-reference-cards-rename

warning reference cards rename

bgpd isisd zebra: misuse of comma operator

Detected using ./configure CFLAGS=-Wcomma CC=clang

Signed-off-by: F. Aragon <paco@voltanet.io>

bgpd: type fix

For tracking the last state of the penalty (struct bgp_damp_info) a 'double'
type was used instead of using the 'unsigned int' being used in the structure.

Detected using ./configure CFLAGS=-Wfloat-equal CC=clang

Signed-off-by: F. Aragon <paco@voltanet.io>

*: style for EC replacements

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

*: LIB_[ERR|WARN] -> EC_LIB

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

zebra: ZEBRA_[ERR|WARN] -> EC_ZEBRA

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

watchfrr: WATCHFRR_[ERR|WARN] -> EC_WATCHFRR

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

ripd: RIP_[ERR|WARN] -> EC_RIP

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

pimd: PIM_[ERR|WARN] -> EC_PIM

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

isisd: ISIS_[ERR|WARN] -> EC_ISIS

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

eigrpd: EIGRP_[ERR|WARN] -> EC_EIGRP

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

babeld: BABEL_[ERR|WARN] -> EC_BABEL

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

nhrpd: NHRP_[ERR|WARN] -> EC_NHRP

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

ospfd: OSPF_[ERR|WARN] -> EC_OSPF

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

bgpd: BGP_[WARN|ERR] -> EC_BGP

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

Merge pull request #3019 from pacovn/static_analysis__shadow_variables4

bgpd lib ospf6d pbrd tests zebra: shadowing fixes

bgpd lib ospf6d pbrd tests zebra: shadowing fixes

This fixes all remaining local variable shadowing cases

Signed-off-by: F. Aragon <paco@voltanet.io>

lib, pimd: Convert timer_wheel to use thread_execute_name

Allow at timer wheel creation time the ability to specify a
name for what we want the 'show thread cpu' to show up as.

Modify pim to note this.

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

lib: Add thread_execute_name

Allow the user to specify a run name for display in
'show thread cpu' that is different than the function
name we are calling.

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

Merge pull request #3007 from pacovn/static_analysis__shadow_variables2

lib vtysh zebra: variable shadowing fixes

Merge pull request #3006 from pacovn/static_analysis__shadow_variables1

bgpd isisd ospf6d ospfd: variable shadowing fixes

Merge pull request #3009 from pacovn/static_analysis__shadow_variables3

pimd ripd ripngd: variable shadowing fixes

Merge pull request #3013 from pacovn/static_analysis__conditional_omitted_operand1

nhrpd: missing operand added

nhrpd: missing operand added

Detected using -Wgnu-conditional-omitted-operand

Signed-off-by: F. Aragon <paco@voltanet.io>

Merge branch 'warnings'

Signed-off-by: David Lamparter <equinox@diac24.net>

Merge remote-tracking branch 'frr/master' into warnings

Conflicts:
zebra/if_ioctl_solaris.c
zebra/rtread_getmsg.c

Signed-off-by: David Lamparter <equinox@diac24.net>

bgpd isisd ospf6d ospfd: variable shadowing fixes

Signed-off-by: F. Aragon <paco@voltanet.io>

bgpd: Prevent possible crash when parsing v6 attributes

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>

pimd ripd ripngd: variable shadowing fixes

Signed-off-by: F. Aragon <paco@voltanet.io>

Merge pull request #3005 from patrasar/indentation_warnings

pimd: fix indentation warnings

lib vtysh zebra: variable shadowing fixes

Signed-off-by: F. Aragon <paco@voltanet.io>

pimd: fix indentation warnings

Signed-off-by: Sarita Patra <saritap@vmware.com>

Merge pull request #2944 from thbtcllt/master

fix zebra crash when a vrf interface changes with netns implementation for vrf

Merge branch 'pr2983'

lib: whitespace/spelling fix

Signed-off-by: David Lamparter <equinox@diac24.net>

Merge pull request #2985 from patrasar/Fix_1636

pimd: create a new command "ip pim" configuring pim sm

pimd: create a new command "ip pim" configuring pim sm

A new command "ip pim" is created to configure pim sm on an
interface, which replaces the existing commands "ip pim sm"
and "ip pim ssm" and make "ip pim sm" and "ip pim ssm" as
hidden commands. The command "ip multicast-routing" is removed
since it is already enabled on FRR by default.

Signed-off-by: Sarita Patra saritap@vmware.com

Merge pull request #3001 from donaldsharp/bgp_origin_honor

bgpd: Honor origin change in bgp aggregates

Merge pull request #2999 from donaldsharp/consistent

doc: Fixup to use consistent frrvty group name

Merge pull request #2979 from qlyoung/improve-readme

Improve readme

frr: add README.md to EXTRA_DIST

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

frr: improve README

Friendly READMEs are all the rage!

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

frr: remove stale files, allow non-GNU Automake

* Add 'foreign' option to allow Automake to ignore missing GNU standard
  files
* Remove AUTHORS
* Remove NEWS
* Remove ChangeLog

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

bgpd: Honor origin change in bgp aggregates

When the origin changed we must honor and update the aggregate
to the peer.  This code adds a bit of code to the bgp_aggregate_info_same
code to see if the origin has changed and to indicate that it has.

Fixes: #2993
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>

doc: Fixup to use consistent frrvty group name

In some places we were using `frrvt` instead of `frrvty`.  Make it consistent
with every other place and use frrvty.

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

Merge pull request #2977 from opensourcerouting/doc-fix

build: fix not building docs w/o sphinx

Merge pull request #2965 from opensourcerouting/buildfoo-20180904

more build fixes & warning-free build

build: move SPHINXBUILD= to configure in rpm build

Need to pass this on configure now to work properly.

Signed-off-by: David Lamparter <equinox@diac24.net>

build: fix not building docs w/o sphinx

Can't build manpages without sphinx-build, oops...

Signed-off-by: David Lamparter <equinox@diac24.net>

Merge pull request #2862 from opensourcerouting/non-recursive

final non-recursive make

Merge pull request #2995 from wwijkander/docpatch

doc: Update RPM building instructions

Merge pull request #2994 from opensourcerouting/sa-warnings

fix remaining SA warnings

lib: early-include "config.h" in flex lexers

This is neccessary to get _FILE_OFFSET_BITS right.

Signed-off-by: David Lamparter <equinox@diac24.net>

ospf6d: remove extra struct in ospf6_lsa_handler

This serves no other purpose than to generate stupid warnings for
overwritten initializers on old gcc versions.

Signed-off-by: David Lamparter <equinox@diac24.net>

*: fix some random warnings

Signed-off-by: David Lamparter <equinox@diac24.net>

lib: make pthread_set[_]name_np test OS agnostic

FreeBSD supports pthread_set_name_np() too.  Also, pthread_set_name_np()
returns void.  And NetBSD has pthread_setname_np() with an extra arg...

Signed-off-by: David Lamparter <equinox@diac24.net>

build: check {malloc,pthread}_np.h for *BSD

FreeBSD has malloc_usable_size() in malloc_np.h
OpenBSD has pthread_set_name_np() in pthread_np.h

Signed-off-by: David Lamparter <equinox@diac24.net>

Update RPM building instructions

build: move RFPLDADD to bgpd/rfp-example/librfp

This makes it slightly easier to replace the stub RFP code with an
external implementation.

Signed-off-by: David Lamparter <equinox@diac24.net>

build: use _POSIX_C_SOURCE

Need this to get CMSG_SPACE/CMSG_LEN on Solaris.

Also, AC_GNU_SOURCE is deprecated, AC_USE_SYSTEM_EXTENSIONS does that.

Signed-off-by: David Lamparter <equinox@diac24.net>

vtysh: make RPKI node non-conditional

Whether or not RPKI is enabled during build shouldn't really influence
vtysh; the user can always manually install bgpd_rpki.so later and it
should work.  This also means that the behaviour of "RPKI module not
loaded" is consistent regardless of whether it was a compile-time or
runtime decision.

Signed-off-by: David Lamparter <equinox@diac24.net>

*: fix some solaris warnings

Signed-off-by: David Lamparter <equinox@diac24.net>

*: cleanup .gitignore files

Signed-off-by: David Lamparter <equinox@diac24.net>

*: fix config.h/zebra.h include order

config.h (or, transitively, zebra.h) must be the first include file
listed for autoconf things like _GNU_SOURCE and _POSIX_C_SOURCE to work
correctly.

Signed-off-by: David Lamparter <equinox@diac24.net>

build: fix & clean up *SAN flags

ASAN/MSAN/TSAN flags need to be in CFLAGS and LDFLAGS; the latter links
the correct compiler-dependent library.  Also, the configure switch was
broken (--disable-... would enable the sanitizer.)

Signed-off-by: David Lamparter <equinox@diac24.net>

build: make pkgconfig configure output useful

The variable name is not exactly the most helpful thing there.

Signed-off-by: David Lamparter <equinox@diac24.net>

build: clean up protobuf build integration

We were linking all libs and binaries against libprotobuf-c if the
option was enabled... that makes no sense at all.

Signed-off-by: David Lamparter <equinox@diac24.net>

build: use -export-dynamic

Signed-off-by: David Lamparter <equinox@diac24.net>

tests: properly locate files in builddir

test_cli.refout is written by configure into the build directory, thus
we need a little special glue to find it correctly.

Signed-off-by: David Lamparter <equinox@diac24.net>

build: move vtysh & manpage listings to subdir.am

Since we're now building through one large Makefile, we can easily put
things with their daemons and crossreference nicely.

Signed-off-by: David Lamparter <equinox@diac24.net>

build: remove common.am

Fold things into where they make sense.

Signed-off-by: David Lamparter <equinox@diac24.net>

build: include helper Makefiles in dist

Need these to have "make" work in subdirectories.

Signed-off-by: David Lamparter <equinox@diac24.net>

build: non-recursive tests

May SUBDIRS rest in pieces... er, peace.

Signed-off-by: David Lamparter <equinox@diac24.net>

build: non-recursive vtysh

Signed-off-by: David Lamparter <equinox@diac24.net>

build: non-recursive bgpd & rfp

Note: no more --with-rfp-path on configure - badly messing with the
build system like this really isn't how to do a conditional external
dependency.

Signed-off-by: David Lamparter <equinox@diac24.net>

bfdd: fix garbage "port" string

bfd_recv_ipv4() is getting an uninitialized buffer passed in as port,
and then checks it without clearing it first.  Thus we can end up
leaving garbage data in it.

Signed-off-by: David Lamparter <equinox@diac24.net>

*: fix clang-6 SA warnings

I don't see these in CI, but my local clang-6 does emit warnings for
these.

Signed-off-by: David Lamparter <equinox@diac24.net>

lib: fix SA warning in skiplist code

Clang was thinking the random level could be negative.  (And, no, I
couldn't figure that out by reading its output... trial and error this
was.)

Signed-off-by: David Lamparter <equinox@diac24.net>

lib, ldpd: fix SA warnings from TAILQ oddness

Add a TAILQ_POP_FIRST so Clang understands it's the same item that is
getting removed from the list.

Signed-off-by: David Lamparter <equinox@diac24.net>

Merge pull request #2875 from opensourcerouting/fabricd

OpenFabric support

Merge pull request #2991 from donaldsharp/mac_compiling

Mac compiling cleanups

isisd: silence SA warnings

As suggested by Renato Westphal.

Signed-off-by: David Lamparter <equinox@diac24.net>

vtysh: Actually make the new_completion function match

The new_completion function was not declared the same
way the rl_attempted_completion_function pointer was.
The only difference was a 'const char *' -vs- 'char *'
So convert it over.

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

vtysh: Add code to isolate append_history

The append_history function in lib readline appears to
not be universally available across all of the esoteric
platforms we may want to compile on.  As such
provide a way to gracefully do nothing.

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

lib: Detect if pthread_condattr_setclock is available

Auto-detect if pthread_condattr_setclock is available and if
it is not allow the code to compile around the issue.

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

lib: Cleanup include of link.h

We need link.h for the HAVE_DLINFO_LINKMAP, so include
it if we need it for that code path.

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

Merge pull request #2989 from donaldsharp/static_debug

Static debug

Merge pull request #2988 from donaldsharp/more_debugs

zebra: Modify nexthop checks to report inactive a bit more

Merge pull request #2981 from donaldsharp/v6_vxlan_bug

zebra: Fix RB-Tree storage comparison function for v6