Quentin Young [Tue, 15 Nov 2016 22:15:18 +0000 (22:15 +0000)]
lib: Fix nondeterministic command matches in rare cases
When a user erroneously defines two commands which can
match the same input and at least one of the tokens defined
last in the command is a selector or option, the matcher
does not detect an ambiguous match and matches the command
installed first (leftmost in the graph).
Fix is to do a full walkthrough of the follow set when
matching the final token in a command to check that there
is exactly one possible match, and to throw an ambiguity
error otherwise.
Signed-off-by: Quentin Young <qlyoung@cumulusnetworks.com>
Problem reported that in certain configs, when a router is initially
booted and the link is bounced, we can end up with a bogus static route
in the table. This was due to the assumption in zebra_rnh that a static
route would not be recursively resolved through another static route with
a different next-hop. This fix changes this assumption. Tested manually
and bgp-min, ospf-min, and vrf-min run with no new failures.
Ticket: CM-13328 Signed-off-by: Don Slice Reviewed-by: CCR-5338
Quentin Young [Sat, 12 Nov 2016 01:06:32 +0000 (01:06 +0000)]
lib, vtysh: Fix memory leaks, change cmd_element to const
Fix a few memory issues:
* Not freeing tab-completions upon input match failure
* Invalid write when null-terminating tab-completions
* Not freeing argv[] itself in additinon to elements
* Use XFREE() instead of free() as appropriate
* Not freeing final token of an [option] during parsing
Make a few minor changes to CLI internals:
* Improve documentation on matching & completion functions
* Only make one copy of cmd_token's when building argv,
instead of three
* Don't make a copy of the matching cmd_element
Make one major(ish) change to CLI internals:
* Change all pointers to struct cmd_element to const
Code outside of the core CLI units should never have an
occasion to modify the internal state of the command system.
Doing so could easily amount to having a CLI interface that
changes during runtime, and could conceivably lead to security
issues. Explicitly disallowing this removes any chance of
confusion.
Signed-off-by: Quentin Young <qlyoung@cumulusnetworks.com>
Quentin Young [Thu, 10 Nov 2016 23:17:07 +0000 (23:17 +0000)]
lib: Implement hidden and deprecated commands
Each token now knows whether it is part of a hidden
or deprecated command. Command completion logic hides
such tokens when generating completions. Command
matching logic works as before and will still match on
hidden and deprecated commands.
Signed-off-by: Quentin Young <qlyoung@cumulusnetworks.com>
David Lamparter [Wed, 9 Nov 2016 13:42:47 +0000 (14:42 +0100)]
lib: add minimal no-config VTY mode
This silences the following warning from watchquagga:
"Can't save to configuration file, using vtysh."
which otherwise appears when doing a "write file" in vtysh when no
integrated-config is in use.
Also make "show memory" available in watchquagga.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
Donald Sharp [Wed, 9 Nov 2016 15:22:22 +0000 (10:22 -0500)]
watchquagga: Signal when we are actually up and running
When Quagga is starting up, it is returning immediately.
This is leaving us in a state where systemd believes
Quagga is up and running, while the sytem might actually
not have restarted the code yet.
Modify the code so that when watchquagga starts up
it doesn't start communicating with systemd until
such time that it detects that all daemons are
running.
Additionally modify watchquagga to touch a
file in /var/run/quagga/ that the /usr/lib/quagga/quagga
script looks for for 10 seconds. If it finds this
Quagga started file then we know watchquagga
has successfully communicated with all daemons.
If after 10 seconds we haven't communicated
with Quagga, continue on for the start and let the
normal start failure code work.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
David Lamparter [Tue, 8 Nov 2016 22:36:16 +0000 (23:36 +0100)]
vtysh: funnel integrated write through watchquagga
Running vtysh as normal user won't have permissions to write
Quagga.conf. If we're connected to watchquagga, try "write integrated"
first. In all cases if something fails, try directly.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
David Lamparter [Tue, 8 Nov 2016 18:41:48 +0000 (19:41 +0100)]
vtysh: add watchquagga to target list
Also tag some commands as VTYSH_REALLYALL; these are absolutely
neccessary for correct vtysh operation and will cause "interesting"
breakage if not present on all daemons.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
David Lamparter [Wed, 9 Nov 2016 13:15:34 +0000 (14:15 +0100)]
watchquagga: add "write integrated"
This new command - available for internal use by vtysh and explicit
usage by users - calls "vtysh -w" from watchquagga. This ensures vtysh
is run with privileges to actually write the integrated-config file.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
David Lamparter [Tue, 8 Nov 2016 22:36:01 +0000 (23:36 +0100)]
vtysh: set config file permissions
As vtysh may hopefully be running as root from watchquagga here, let's
try to fix up ownership and permissions for Quagga.conf. Doing
chown/chmod instead of changing the process's user/group IDs has the
advantage of fixing up preexisting misconfigurations.
Note errors in chmod/chown will print a message but the config is
already written at that point.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
David Lamparter [Tue, 8 Nov 2016 18:01:06 +0000 (19:01 +0100)]
vtysh: add -w option for integrated-config write
This new option is intended to be used both by watchquagga as well as
directly by users. It performs the collect-configuration operation and
writes out Quagga.conf, regardless of whether integrated-config is
enabled or not.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
David Lamparter [Tue, 8 Nov 2016 17:22:30 +0000 (18:22 +0100)]
vtysh: detangle configuration writes
vtysh has a very convoluted and confusing setup where it isn't even
clear which files are written where (since some filenames come
indirectly from loading config). Detangle.
This also removes writing vtysh.conf. The file is intended to be
manually edited since it has some vague security concerns (if PAM is
used).
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
David Lamparter [Tue, 8 Nov 2016 18:42:01 +0000 (19:42 +0100)]
lib: privs: always look up VTY group
Even if we're running without user switch, we should still try to honor
the VTY group. This applies both to watchquagga (which always runs as
root) as well as "no-userswitch" configurations for other daemons.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
David Lamparter [Tue, 8 Nov 2016 19:46:05 +0000 (20:46 +0100)]
lib: add and use set_cloexec()
watchquagga is already leaking an open file descriptor on its pid file
on fork+exec() invocations; next up is adding vtysh support with even
more fds. Mark things CLOEXEC before going there.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
Quentin Young [Sat, 29 Oct 2016 04:43:04 +0000 (04:43 +0000)]
lib: Various minor improvements & bugfixes to CLI backend
- Do not allow tab-completion on anything except words
- Rewrite cmd_make_strvec to use strsep
- Remove a few trailing whitespaces
- Remove cmd_complete_command_lib
Signed-off-by: Quentin Young <qlyoung@cumulusnetworks.com>
Paul Jakma [Fri, 22 Apr 2016 11:48:49 +0000 (12:48 +0100)]
bgpd: Squash spurious "unknown afi" log messages
* bgp_packet.c: (bgp_update_receive) doesn't differentiate between NLRIs that
are 0 AFI/SAFI cause they weren't set, and those because a peer sent a
bogus AFI/SAFI, before sending sending what may be a misleading, spurious
log message. Check the .nlri pointer is set and avoid this.
Incorporating a suggestion from: G. Paul Ziemba <unp@ziemba.us>
Paul Jakma [Thu, 4 Feb 2016 17:00:18 +0000 (17:00 +0000)]
bgpd: Remove the double-pass parsing of NLRIs
* bgpd parses NLRIs twice, a first pass "sanity check" and then a second pass
that changes actual state. For most AFI/SAFIs this is done by
bgp_nlri_sanity_check and bgp_nlri_parse, which are almost identical.
As the required action on a syntactic error in an NLRI is to NOTIFY and
shut down the session, it should be acceptable to just do a one pass
parse. There is no need to atomically handle the NLRIs.
* bgp_route.h: (bgp_nlri_sanity_check) Delete
* bgp_route.c: (bgp_nlri_parse) Make the prefixlen size check more general
and don't hard-code AFI/SAFI details, e.g. use prefix_blen library function.
Add error logs consistent with bgp_nlri_sanity_check as much as possible.
Add a "defense in depth" type check of the prefixlen against the sizeof
the (struct prefix) storage - ala bgp_nlri_parse_vpn.
Update standards text from draft RFC4271 to the actual RFC4271 text.
Extend the semantic consistency test of IPv6. E.g. it should skip mcast
NLRIs for unicast safi as v4 does.
* bgp_mplsvpn.{c,h}: Delete bgp_nlri_sanity_check_vpn and make
bgp_nlri_parse_vpn_body the bgp_nlri_parse_vpn function again.
(bgp_nlri_parse_vpn) Remove the notifies. The sanity checks were
responsible for this, but bgp_update_receive handles sending NOTIFY
generically for bgp_nlri_parse.
* bgp_attr.c: (bgp_mp_reach_parse,bgp_mp_unreach_parse) Delete sanity check.
NLRI parsing done after attr parsing by bgp_update_receive.
Arising out of discussions on the need for two-pass NLRI parse with:
Lou Berger <lberger@labn.net>
Donald Sharp <sharpd@cumulusnetworks.com>
Paul Jakma [Thu, 4 Feb 2016 13:27:04 +0000 (13:27 +0000)]
bgpd: Regularise bgp_update_receive, add missing notifies and checks
* bgp_packet.c: (bgp_update_receive) Lots of repeated code, doing same
thing for each AFI/SAFI. Except when it doesn't, e.g. the IPv4/VPN
case was missing the EoR bgp_clear_stale_route call - the only action
really needed for EoR.
Make this function a lot more regular, using common, AFI/SAFI
independent blocks so far as possible.
Replace the 4 separate bgp_nlris with an array, indexed by an enum.
The distinct blocks that handle calling bgp_nlri_parse for each
different AFI/SAFI can now be replaced with a loop.
Transmogrify the nlri SAFI from the SAFI_MPLS_LABELED_VPN code-point
used on the wire, to the SAFI_MPLS_VPN safi_t enum we use internally
as early as possible.
The existing code was not necessarily sending a NOTIFY for NLRI
parsing errors, if they arose via bgp_nlri_sanity_check. Send the
correct NOTIFY - INVAL_NETWORK for the classic NLRIs and OPT_ATTR_ERR
for the MP ones.
EoR can now be handled in one block. The existing code seemed broken
for EoR recognition in a number of ways:
1. A v4/unicast EoR should be an empty UPDATE. However, it seemed
to be treating an UPDATE with attributes, inc. MP REACH/UNREACH,
but no classic NLRIs, as a v4/uni EoR.
2. For other AFI/SAFIs, it was treating UPDATEs with no classic
withraw and with a zero-length MP withdraw as EoRs. However, that
would mean an UPDATE packet _with_ update NLRIs and a 0-len MP
withdraw could be classed as an EoR.
This seems to be loose coding leading to ambiguous protocol
situations and likely incorrect behaviour, rather than simply being
liberal. Be more strict about checking that an UPDATE really is an
EoR and definitely is not trying to update any NLRIs.
This same loose EoR parsing was noted by Chris Hall previously on
list.
(bgp_nlri_parse) Front end NLRI parse function, to fan-out to the correct
parser for the AFI/SAFI.
* bgp_route.c: (bgp_nlri_sanity_check) We try convert NLRI safi to
internal code-point ASAP, adjust switch for that. Leave the wire
code point in for defensive coding.
(bgp_nlri_parse) rename to bgp_nlri_parse_ip.
* tests/bgp_mp_attr_test.c: Can just use bgp_nlri_parse frontend.
Paul Jakma [Wed, 27 Jan 2016 16:37:33 +0000 (16:37 +0000)]
bgpd: Regularise BGP NLRI sanity checks a bit
* bgp_route.h: (bgp_nlri_sanity_check) The bulk of the args are equivalent
to a (struct bgp_nlri), consolidate.
* bgp_route.c: (bgp_nlri_sanity_check) Make this a frontend for all afi/safis.
Including SAFI_MPLS_LABELED_VPN.
(bgp_nlri_sanity_check_ip) Regular IP NLRI sanity check based on the
existing code, and adjusted for (struct bgp_nlri *) arg.
* bgp_attr.c: (bgp_mp_reach_parse) Adjust for passing (struct bgp_nlri *)
to bgp_nlri_sanity_check.
Get rid of special-casing to not sanity check VPN.
(bgp_mp_unreach_parse) Ditto.
* bgp_mplsvpn.c: Use the same VPN parsing code for both the sanity
check and the actual parse.
(bgp_nlri_parse_vpn) renamed to bgp_nlri_parse_vpn_body and made
internal.
(bgp_nlri_parse_vpn_body) Added (bool) argument to control whether it
is sanity checking or whether it should update routing state for each
NLRI. Send a NOTIFY and reset the session, if there's a parsing
error, as bgp_nlri_sanity_check_ip does, and as is required by the
RFC.
(bgp_nlri_parse_vpn) now a wrapper to call _body with update.
(bgp_nlri_sanity_check_vpn) wrapper to call parser without
updating.
* bgp_mplsvpn.h: (bgp_nlri_sanity_check_vpn) export for
bgp_nlri_sanity_check.
* bgp_packet.c: (bgp_update_receive) Adjust for bgp_nlri_sanity_check
argument changes.
* test/bgp_mp_attr_test.c: Extend to also test the NLRI parsing functions,
if the initial MP-attr parsing has succeeded. Fix the NLRI in the
VPN cases. Add further VPN tests.
* tests/bgpd.tests/testbgpmpattr.exp: Add the new test cases.
This commit a joint effort of:
Lou Berger <lberger@labn.net>
Donald Sharp <sharpd@cumulusnetworks.com>
Paul Jakma <paul.jakma@hpe.com> / <paul@jakma.org>
Paul Jakma [Fri, 5 Feb 2016 14:57:17 +0000 (14:57 +0000)]
bgpd: make bgp_nlri_parse_encap conform with other nlri_parse funcs
* bgp_encap.{c,h} (bgp_nlri_parse_encap) afi is already in the NLRI argument.
update or withdraw is signalled by attr being non-NULL or NULL.
* bgp_packet.c: (update_receive) fixup to match, and also make the attr
argument conform with NLRI_ATTR_ARG for correct error handling on
optional, transitive, partial, attributes.
Donald Sharp [Tue, 25 Oct 2016 18:25:29 +0000 (14:25 -0400)]
zebra: Disable mpls slightly different
When mpls is not turned on in the kernel, we
are not installing the mpls commands into the cli.
This results in vtysh attempting to run the command
and receiving a 'WTF is this command' back from zebra.
Modify the mpls code to install commands and to check
to see if the command should be accepted based
upon mpls working or not.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com> Reviewed-by: Daniel Walton <dwalton@cumulusnetworks.com>
Daniel Walton [Tue, 25 Oct 2016 14:16:31 +0000 (14:16 +0000)]
bgpd: dynamically grow 'show ip bgp summ' Neighbor column width
Signed-off-by: Daniel Walton <dwalton@cumulusnetworks.com> Reviewed-by: Donald Sharp <sharpd@cumulusnetworks.com>
Now that we display hostnames in 'show ip bgp summary' it is really easy
to have the first column be greater than 16 characters which causes a
line wrap. The line wrap makes the output difficult to read.
Before
======
superm-redxp-05# show ip bgp summ
BGP router identifier 6.0.0.11, local AS number 65001 vrf-id 0
BGP table version 56
RIB entries 19, using 2280 bytes of memory
Peers 2, using 41 KiB of memory
Peer groups 1, using 64 bytes of memory
After
=====
superm-redxp-05# show ip bgp summ
BGP router identifier 6.0.0.11, local AS number 65001 vrf-id 0
BGP table version 10
RIB entries 19, using 2280 bytes of memory
Peers 2, using 41 KiB of memory
Peer groups 1, using 64 bytes of memory
projects / matthieu / frr.git / log