ripd and ripngd were leveraging the zclient code to keep track of
the redistribute configuration, which is what most daemons do. The
problem, however, is that the zclient code uses VRF IDs to identify
VRFs, and VRF IDs are unknown until a VRF is enabled (information
received from zebra). This means we can't configure a redistribute
command on a RIP instance when the corresponding VRF is disabled
(doing so leads to a null-dereference crash right now in both ripd
and ripngd).
To fix this, change the rip/ripng data structures so that they
keep track of the full redistribute configuration and not only the
route-map and metric associated to each command. This is similar
to what bgpd and ospfd are doing to solve the same problem. In the
future the zclient code and all daemons need to be refactored to
consolidate the handling of redistribute configuration in a single
place to reduce code duplication. One of the most important changes
to do is to use VRF names and not VRF IDs to identify VRFs.
Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
Description of the new parameter (adapted from the ietf-rip module):
"VRF name identifying a specific RIPng instance.
This leaf is optional for the rpc.
If it is specified, the rpc will clear all routes in the
specified RIPng instance;
if it is not specified, the rpc will clear all routes in
all RIPng instances.";
Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
* Turn the "instance" YANG presence-container into a YANG list keyed
by the new "vrf" leaf. This is a backward incompatible change but
this should be ok for now.
* RIPng VRF instances can be configured even when the corresponding
VRF doesn't exist. And a RIPng VRF instance isn't deleted when
the corresponding VRF is deleted. For this to work, implement the
ripng_instance_enable() and ripng_instance_disable() functions
that are called to enable/disable RIPng routing instances when
necessary. A RIPng routing instance can be enabled only when the
corresponding VRF is enabled (this information comes from zebra
and depends on the underlying VRF backend). Routing instances are
stored in the new ripng_instances rb-tree (global variable).
* Add a vrf pointer to the ripng structure instead of storing vrf_id
only. This is much more convenient than using vrf_lookup_by_id()
every time we need to get the vrf pointer from the VRF ID. The
ripng->vrf pointer is updated whenever the VRF enable/disable hooks
are called.
Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
Description of the new parameter (adapted from the ietf-rip module):
"VRF name identifying a specific RIP instance.
This leaf is optional for the rpc.
If it is specified, the rpc will clear all routes in the
specified RIP instance;
if it is not specified, the rpc will clear all routes in
all RIP instances.";
Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
* Turn the "instance" YANG presence-container into a YANG list keyed
by the new "vrf" leaf. This is a backward incompatible change but
this should be ok for now.
* RIP VRF instances can be configured even when the corresponding
VRF doesn't exist. And a RIP VRF instance isn't deleted when
the corresponding VRF is deleted. For this to work, implement the
rip_instance_enable() and rip_instance_disable() functions that are
called to enable/disable RIP routing instances when necessary. A
RIP routing instance can be enabled only when the corresponding
VRF is enabled (this information comes from zebra and depends on
the underlying VRF backend). Routing instances are stored in the new
rip_instances rb-tree (global variable).
* Add a vrf pointer to the rip structure instead of storing vrf_id
only. This is much more convenient than using vrf_lookup_by_id()
every time we need to get the vrf pointer from the VRF ID. The
rip->vrf pointer is updated whenever the VRF enable/disable hooks
are called.
Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
The ripd code can handle this error just fine (e.g. reject the
configuration transaction), there's no need to exit when a socket
fails to be created.
Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
For unknown reasons, the IP_MULTICAST_IF sockoption doesn't
seem to work when the given socket is bound to a VRF device on
Linux. Switch from sendto() to sendmsg() so that we can specify the
outgoing interface of multicast packets using IP_PKTINFO ancillary
data. This also makes ripd more consistent with ripngd, which also
uses sendmsg() to send packets on the wire.
Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
When lyd_validate() is used with the LYD_OPT_DATA option, full YANG
validation is performed. As a side-effect to this, default nodes are
created, which is not desirable when displaying operational data
since configuration nodes can also be created. Use LYD_OPT_GET
option to resolve this problem.
Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
One use case for the new yang_data_list_find() function is to find
input parameters in RPC northbound callbacks easily, without the
need to iterate over the input parameters manually.
Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
Parameters should be const whenever possible to improve code
readability and remove the need to cast away the constness of
const arguments.
Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
This is the last step to make ripngd ready for multi-instance
support.
Remove the ripng global variable and add a "ripng" parameter
to all functions that need to know the RIPng instance they are
working on. On some functions, retrieve the RIPng instance from
the interface variable when it exists (this assumes interfaces can
pertain to one RIPng instance at most, which is ok for VRF support).
In preparation for the next commits (VRF support), add a "vrd_id"
member to the ripng structure, and use ripng->vrf_id instead of
VRF_DEFAULT wherever possible.
Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
This is an old standing bug where the list of RIPng peers wasn't
cleared after deconfiguring RIPng, which caused the existing peers
to still be present on a newly configured RIPng instance (except
when the timed out after ~3 minutes). Fix this.
Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
* Call ripng_clean() only when RIPng is configured, this way we can
remove one indentation level from this function.
* ripng_redistribute_clean() is only called on shutdown, so there's
no need to call ripng_redistribute_withdraw() there since the RIPng
table is already cleaned up elsewhere.
* Deallocate the ripng structure only at the end of the function. This
prepares the ground for the next commits where all global variables
will be moved to the ripng structure.
Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
This is the last step to make ripd ready for multi-instance support.
Remove the rip global variable and add a "rip" parameter to all
functions that need to know the RIP instance they are working
on. On some functions, retrieve the RIP instance from the interface
variable when it exists (this assumes interfaces can pertain to
one RIP instance at most, which is ok for VRF support).
In preparation for the next commits (VRF support), add a "vrd_id"
member to the rip structure, and use rip->vrf_id instead of
VRF_DEFAULT wherever possible.
Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
This is an old standing bug where the list of RIP peers wasn't
cleared after deconfiguring RIP, which caused the existing peers
to still be present on a newly configured RIP instance (except when
the timed out after ~3 minutes). Fix this.
Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
The only sideeffect of this change is that these counters will be
reset when RIP is deconfigured and then configured again, but this
shouldn't be a problem as the RIP MIB isn't specific about this.
Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
None of these variables or functions were being used since the
initial revision ~16 years ago. It's safe to say we can remove
them now.
Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
* Call rip_clean() only when RIP is configured, this way we can
remove one indentation level from this function.
* rip_redistribute_clean() is only called on shutdown, so there's
no need to call rip_redistribute_withdraw() there since the RIP
table is already cleaned up elsewhere.
* There's no need to clean up the "rip->neighbor" nodes manually before
calling route_table_finish().
* Deallocate the rip structure only at the end of the function. This
prepares the ground for the next commits where all global variables
will be moved to the rip structure.
Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
Fixes the following warning when running ripngd with valgrind:
==38== Syscall param sendmsg(msg.msg_control) points to uninitialised
byte(s)
==38== at 0x5EA1E47: sendmsg (sendmsg.c:28)
==38== by 0x118C48: ripng_send_packet (ripngd.c:226)
==38== by 0x11D1D6: ripng_request (ripngd.c:1924)
==38== by 0x120BD8: ripng_interface_wakeup (ripng_interface.c:666)
==38== by 0x4ECB4B4: thread_call (thread.c:1601)
==38== by 0x4E8D9CE: frr_run (libfrr.c:1011)
==38== by 0x1121C8: main (ripng_main.c:180)
==38== Address 0xffefffc34 is on thread 1's stack
==38== in frame #1, created by ripng_send_packet (ripngd.c:172)
Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
We were sending ZEBRA_INTERFACE_LINK_PARAMS messages under the
following circumstances:
* New interface was created (via kernel or config);
* Interface went from down to up;
* Update in the link-params configuration.
Now also send ZEBRA_INTERFACE_LINK_PARAMS messages whenever a zclient
connects and sends a ZEBRA_INTERFACE_ADD request. Without this fix,
the client daemons don't receive interface link parameters if they
are configured in the zebra startup configuration.
Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
In these two functions, we were using VRF_DEFAULT instead of the
VRF ID passed as a parameter when checking if the given client
subscribed to receive default routes or not. This prevented the
"default-originate" command from ospfd/isisd from working correctly
under specific circumstances.
Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
client->ifinfo is a VRF bitmap, hence we need to use
vrf_bitmap_check() to check if a client is subscribed to receive
interface information for a particular VRF. Just checking if
the client->ifinfo value is set will always succeed since it's
a pointer initialized by zserv_client_create(). With this fix,
we'll stop sending interface messages from all VRFs to all clients,
even those that didn't subscribe to it.
Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
the command was not checking
correctly in all cases whether the virtual link existed. This caused
bugs in some corner cases, e.g. when two virtual links were created,
one of them was deleted, and the second one was reset with no
authentication - this would instead create a new virtual link with
the area in decimal format.
Signed-off-by: Emanuele Di Pascale <emanuele@voltanet.io>
strlen is the same as sizeof when the memory is dynamically allocated
but it is not the same when the memory being looked at is an array.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
The modules.rst documents how to create a module, let's update
it a bit to reflect what an end user needs to do a bit more.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
default-information originate does not work
if config is removed and re-added.
Ticket:CM-20026
Testing Done:
Validate default-information originate config
removed and re-added, check ospf lsa database, and peer
route cache entry for default route.
Signed-off-by: Chirag Shah <chirag@cumulusnetworks.com>
