Donald Sharp [Wed, 20 Sep 2017 16:56:38 +0000 (12:56 -0400)]
bgpd: Fix json memory leak
When issuing 'show bgp ...' commands that dump
the entire table, we were dropping the initial
json_paths = json_object_new_object() memory
allocation. Fix this.
Additionally reformat the output to fit
better in 80 columns.
There may be additional memory leaks here
hidden away in how we decide to continue
or not.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
Quentin Young [Wed, 14 Jun 2017 19:47:14 +0000 (19:47 +0000)]
lib: allow 'do' commands in ENABLE_NODE
'do' is syntax sugar that allows the user to execute a command under
ENABLE_NODE when in another CLI node. If the user is already in
ENABLE_NODE, use of 'do' was previously disallowed. This patch allows it
because it makes it easier for us to hack around certain instances of
the node synchronization problem with vtysh.
Also included is a fix for one of these problems.
Signed-off-by: Quentin Young <qlyoung@cumulusnetworks.com>
Once ldpd allocated label 48 for a given FEC, all subsequent requests
for a new label would return the same value (48). The problem is that
we were left shifting an uint32_t value up to 64 times, losing important
information.
David Lamparter [Thu, 7 Sep 2017 12:24:00 +0000 (14:24 +0200)]
bgpd: kill bgp_attr_refcount()
This attempt at optimization has cost us more than a week's worth of
time on several people hunting down the subtle bug that it was missing
an increment on attr->lcommunity.
This is absolutely not worth the maintenance cost.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
Donald Sharp [Fri, 10 Mar 2017 12:59:52 +0000 (07:59 -0500)]
bgpd: Modify attr->flag to be 64 bit
With the some current bgp drafts the Attribute number has
surpassed 32. Which is a bit unfortunate in that we keep
track of the attributes via a bitfield based on the attribute #.
For the moment since I am not aware of Attribute #'s being
greater than 64, convert the flag to 64 bit and allow the
bit shifting to know about it.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
Nigel Kukard [Wed, 30 Aug 2017 09:39:36 +0000 (09:39 +0000)]
bgpd: Fixed incorrect calculation of pointer location
com_index_to_delete[i] holds the location in lcom->val of where the
value is stored, we cannot just increment it by this value on each
iteration as we'll overflow the size of lcom->val.
isisd: don't validate dict integrity for regular builds
This method is intended to be only used for debugging as per the author
and profiling shows we are spending a lot of cycles on it. Remove it for
regular builds by guarding it with a define.
Signed-off-by: Christian Franke <chris@opensourcerouting.org>
David Lamparter [Fri, 11 Aug 2017 16:54:26 +0000 (18:54 +0200)]
lib: fix prefix list trie corruption
The specific code here needs to establish an absolute order of more
specific to less specific possible matches in a prefix list. This is
indirectly checked by an assert on insertion, because the "next best"
entry is required to be consistent even when joining multiple chains
of candidates.
Unfortunately, trie_install_fn() would insert entries too far ahead in
the chain if another entry with higher sequence number was seen. This
breaks the trie and (rightfully) triggers the assertion failure on
insert.
Fixes: #937 Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
David Lamparter [Sun, 6 Aug 2017 03:14:39 +0000 (05:14 +0200)]
zebra: static: update on ifindex changes
Whenever an interface is created or deleted in the system, we need to
check whether we have static routes referencing that interface by name.
If so, we need to [un]install these routes.
This has the unfortunate side effect of making static routes with
non-existent interfaces disappear from "show ip route", but I think
that's acceptable (and I don't see a "good" fix for that).
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
ßingen [Tue, 1 Aug 2017 09:38:19 +0000 (11:38 +0200)]
ospf: Fix segfault if compiled with DEBUG
If OSPF_LS_REFRESH_TIME is 60, min_delay in ospf_refresher_register_lsa
function (ospf_lsa.c) would be negative, so index (which is unsigned)
would be out of range, causing a segfault.
OpenBSD uses ioctl to fetch interface information on startup and the
SIOCGIFMEDIA command is just too cumbersome to use.
The best way to fix the problem for OpenBSD is probably to stop treating
it differently from the other BSDs for no apparent reason. There should
be nothing preventing us to make OpenBSD use the routing socket to fetch
interface information on startup (we already do it to detect runtime
changes). This is something that should be done in a separate commit
after a careful analysis.
In some scenarios, it's possible to send a Label Withdraw to a neighbor
and not receive a corresponding Label Release right away. This can happen
during reconvergence after a network failure or configuration change.
When this happens, the list of upstream mappings of a given FEC might
not be empty even after sending a Label Withdraw to all neighbors. This
situation holds until all neighbors either send a Label Release or are
torn down (e.g. keepalive timeout).
With that said, we shouldn't check for 'RB_EMPTY(&fn->upstream)'
in lde_kernel_update() because it can prevent ldpd from sending label
mappings in such circumstances. This check was introduced to avoid sending
the same label mapping more than once to the same neighbor, but we need
to remove this optimization for now until we find a better solution (which
probably involves refactoring the whole zebra<->ldpd communication).
While here, add a new debug message in lde_send_labelmapping() which
can aid in troubleshooting label problems in the future.
Renato Westphal [Fri, 30 Jun 2017 15:19:39 +0000 (12:19 -0300)]
ldpd: integrate with the pseudowire manager in zebra
If we receive a notification from zebra indicating that the installation
of a pseudowire has failed (e.g. no reachability), send a PW Status
notification to the remote peer (or a Label Withdraw if the remote peer
doesn't support the PW Status TLV).
Renato Westphal [Thu, 8 Jun 2017 23:29:12 +0000 (20:29 -0300)]
zebra: add new flag to detect nexthop label updates
With the introduction of the pseudowire manager, the NHT tracking code
needs to detect label updates as well.
Create a specific nexthop flag for that. We can't reuse the
RIB_ENTRY_NEXTHOPS_CHANGED flag for this porpose because this flag is
always cleared and reevaluated in rib_process(), setting it outside that
function is a nop.
Renato Westphal [Fri, 30 Jun 2017 15:26:04 +0000 (12:26 -0300)]
zebra: add nexthop tracking for pseudowires
If the remote end of a pseudowire becomes unreachable (no route or an
unlabeled route), then it must be uninstalled. In the same way, when
the remote end becomes reachable, the pseudowire must be installed.
ßingen [Mon, 15 May 2017 15:09:28 +0000 (17:09 +0200)]
zebra: add pseudowire manager
Base framework for supporting MPLS pseudowires in FRR.
A consistent zserv interface is provided so that any client daemon
(e.g. ldpd, bgpd) can install/uninstall pseudowires in a standard
way. Static pseudowires can also be implemented by using the same
interface.
When zebra receives a request to install a pseudowire and the installation
in the kernel or hardware fails, a notification is sent back to the
client daemon and a new install attempt is made every 60 seconds (until
it succeeds).
Support for external dataplanes is provided by the use of hooks to
install/uninstall pseudowires.
We were assuming that a neighbor can be deleted only when all of its
adjacencies are dead. This is not the case for dual-stack neighbors. If
the transport-preference is IPv4 and all adjacencies are IPv6 (or
vice-versa), then it should be deleted and everything cleaned-up
accordingly.
Bug exposed by the new RB tree implementation on master, but the fix
also applies to stable/3.0.
Quentin Young [Tue, 20 Jun 2017 23:56:50 +0000 (23:56 +0000)]
*: simplify log message lookup
log.c provides functionality for associating a constant (typically a
protocol constant) with a string and finding the string given the
constant. However this is highly delicate code that is extremely prone
to stack overflows and off-by-one's due to requiring the developer to
always remember to update the array size constant and to do so correctly
which, as shown by example, is never a good idea.b
The original goal of this code was to try to implement lookups in O(1)
time without a linear search through the message array. Since this code
is used 99% of the time for debugs, it's worth the 5-6 additional cmp's
worst case if it means we avoid explitable bugs due to oversights...
Signed-off-by: Quentin Young <qlyoung@cumulusnetworks.com>
projects / matthieu / frr.git / log