We've already allocated the mp data structure
and am using hash_alloc_intern for the hash_get
function. This will return the passed
in data structure. There is no possibility
of mp being NULL.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
We need to react to route changes/Interface up/down events
faster in PIM. Reduce the timer down to 50ms.
Ticket:CM-13549
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When we get a callback for a specific (S,G) and
we have no ifchannel for it, see if we have a
(*,G) channel and handle it appropriately.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
1. Added a new MSDP source reference flag for creating (S,G) entries
based on the SA-cache. The RFC recommends treating as SA like rxing
a (S, G) join (which is a bit different then treating like a traffic
stream).
2. SA-SPT is only setup if we are RP for the group and a corresponding
(*,G) exists with a non-empty OIL.
3. When an SA is moved we need to let the SPT live if it is active (this
change will come in a subsequent CL).
Testing done:
1. SA first; SPT setup whenever (*, G) comes around.
2. (*, G) first. As soon as SA is added SPT is setup.
3. (*, G) del with valid SA entries around.
Ticket: CM-13306
Signed-off-by: Anuradha Karuppiah <anuradhak@cumulusnetworks.com>
Acked-by: Donald Sharp <sharpd@cumulusnetworks.com>
When determining the inherited_olist(S,G) add
the determination that we have received a
prune(S,G,rpt) from a neighbor.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
Add some more code to handle the prune(S,G) with the
sptbit set. Turns this ifchannel into a (S,G,rpt).
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When leaving a channel, there exists a possibility
that we have not created the channel oil yet.
Ensure that we have channel oil before dereferencing
Ticket: CM-13522
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When we handle the thread arguments,
there is no need to assert. As that
if they are wrong, we are going down
shortly anyways.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
According to Figure 5( Downstream per-interface (S,G,rpt)
state when we receive a (*,G) we need to move (S,G,rpt)
children of the (*,G) into different states. This
implements that.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When we called pim_parse_addr_group, don't reinitialize
the 'struct prefix_sg' *after* we've parsed the group.
Ensure in other places that we do this work, we initialize
prior as well.
Ticket: CM-13510
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When we have multiple incoming joins for
a single group on a interface, we need to
allow proper output.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
If you add to the json by grp, if you have multiple
ifchannels for that group, on the the first one will
be displayed.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
Don't delete the ifchannel if the expiry timer is still running.
We might need to add the prune pending timer as well? Not sure
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When we read in the S,G from the join/prune message,
convert it into a 'struct prefix_sg' at an
earlier point in time.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
Modify the pim_parse_addr_source function to take
a 'struct prefix_sg' and to fill in the src data.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
Modify the pim_parse_addr_group to use 'struct prefix_sg sg'.
This is the first of 2 commits to clean up this parsing to
be a bit better.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
Add the ability for pim ifchannels *,G's to know their
corresponding S,G's. This will facilitate handling
S,G,rpt state information better.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
This commit includes -
1. Maintaining SA cache with local and remote entries.
2. Local SA entries - there are two cases where we pick up these -
- We are RP and got a source-register from the FHR.
- We are RP and FHR and learnt a new directly connected source on a
DR interface.
3. Local entries are pushed to peers immediately on addition and
periodically. An immediate push is also done when peer session is
established.
4. Remote SA entries - from other peers in the mesh group and passed
peer-RPF checks.
5. Remote entries are aged out. No other way to del them
currently. In the future we may add a knob to flush entries on
peer-down.
Testing done -
Misc topologies with CL routers plus basic interop with another vendor (
we can process their SA updates and they ours).
Sample output -
root@rp:~# vtysh -c "show ip msdp sa"
Source Group RP Uptime
33.1.1.1 239.1.1.2 local 00:02:34
33.1.1.1 239.1.1.3 local 00:02:19
44.1.1.1 239.1.1.4 100.1.3.1 00:01:12
44.1.1.1 239.1.1.5 100.1.3.1 00:00:55
root@rp:~#
Ticket: CM-13306
Signed-off-by: Anuradha Karuppiah <anuradhak@cumulusnetworks.com>
Acked-by: Donald Sharp <sharpd@cumulusnetworks.com>
This fix handles two issues:
1) Searching entire vrf_iflist to get per interface pim_ifchannel_list
2) Display of ifchannel information in pim not being ordered correctly.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When a 'show ip pim join' is issued and we have
(S,G,rpt) being sent back up to us. We need
to order correctly for this situation.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When we decide to ignore a incoming packet, allow detailed
debugging to give a pointer to where to go to understand the
issue.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
The (S,G,rpt) prune inclusion was incorrectly considering
if the RPF' was the same for (S,G) and (*,G).
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
Do not reset the time the mroute has been in it's
current state if we get a transition to the same state.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
Add logic to show that we are making decisions about
the s,g,rpt prune send. We are not actually sending
anything yet.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
Some igmpv2 messages are sent to 224.0.0.2. We were
not listening for these messages and as such we
were ignoring some withdrawals.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When a static igmp join is issued, before routing
has come up, the ability to recover was accidently
removed from the code.
Ticket: CM-13379
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When the nexthop lookup fails when establishing the
upstream state as part of a register receive, kill
the upstream state.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When we have a FHR/RP/LHR all on the same box, we were
experiencing a situation where we were not sending
a register stop nor where we setting the sptbit.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
fd
When we get a request to look up the mroute statistics
from the kernel, ensure that the interface returned
is a valid usable interface.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
This commit includes the following changes -
1. Support for MSDP peer DB (hash and sorted list).
2. Support for the following timers - keepalive, connect-retry, hold.
3. TCP session management (lower-ip is active, higher-ip is passive).
4. MSDP KA packet rx/tx.
5. Limited temporary config (will be replaced with the more automation
friendly RP-set).
Testing done -
Peer bringup/deletion (including interop with another vendor)
Sample out -
root@dell-s6000-04:~# sudo vtysh -c "show ip msdp peer"
Peer Local Mesh-group State Uptime
100.1.1.1 100.1.2.1 default established 00:07:27
100.1.3.1 100.1.2.1 default established 00:31:50
root@dell-s6000-04:~#
Coming soon -
1. part-2: SA cache management.
2. part-3: SPT setup using source in SA cache.
3. part-4: CLI cleanup.
Ticket: CM-13306
Signed-off-by: Anuradha Karuppiah <anuradhak@cumulusnetworks.com>
Acked-by: Donald Sharp <sharpd@cumulusnetworks.com>
Pass in the upstream data structure to pim_msg_join_prune_encode
so it can decide to send (S,G,rpt) information if it wants to or
not.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
The test commands are not being maintained and
are not out of date with the rest of the system. There
are better ways to test code and in addition these
commands if entered by a user could seriously impact
their running system.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When sending a join/prune send in the upstream pointer.
this will allow us to implement some of the other state
machines necessary.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
We are seeing situations where nexthop lookups are failing
unexpectedly. Don't consider the lookup to have succeeded
in this case to allow the next lookup to work?
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When we are looking up the incoming interface for the ipmr
call back socket from the kernel, look to see if the src
is connected as well as has a valid pim_ifp to use.
This is to allow vrr configuration in a mlag env.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When we print a debug in the pim_upstream_add, there
exists failure cases for pim_upstream_new where
we cannot create a upstream data structure. Make
sure the debug handles it right.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
Cache the last time we looked up the nexthop for this particular
address. Store time to usec accuracy.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
Fix possible double free of upstream and in
addition add some debug code to help find
where the problem is coming from.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When we receive a igmp message through the kernel upcall, make
sure that we are configured to work on that interface via pim/igmp
before attempting to use that interface.
Ticket: CM-13338
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
Signed-off-by: Daniel Walton <dwalton@cumulusnetworks.com>
This cleans up the following so that they always collect the data to
display in json structures and then either dumps the json output or
traverses it to produce the non-json output.
show ip pim local-membership
show ip pim interface
When handling a igmp request, fill in group address before
we attempt to output debug information on it.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When outputting data about ifchannels due to debug
include the interface this is happening on.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
In some situations, the deletion of the ifchannel would
not ocurr if the last flag cleared was the assert flag.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When entering debug commands under 'conf t' mode
allow the debugs to be saved for future fun and
adventure.
Ticket: CM-13213
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
If the incoming interface comes back, reinstall the channel oil
if the mroute is not installed.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When looking up the channel_oil use a hash
to find it. Keep the list around for quick
walks of the channel oils.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When we have pim_mroute.c or pim_register.c create
the upstream state, only delete it then when
the KAT timer expires, else we will not have
the refcount right and we will do bad things.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When receiving callbacks from the kernel allow bigger
packet sizes than 3k to be handled appropriately.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When looping over S,G's associated with a *,G,
there can exist situations where we have not created
the channel oil for the S,G yet. Don't allow this
to crash the system.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When we have intentionally not installed a mroute( for whatever
reason ), do not ask for information about that mroute from the
kernel when it happens.
Ticket: CM-12986
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
Pim is outputting a bunch of unprotected debugs.
In a system with a high # of events even with
no logging we will receive lots of messages.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
The kernel now reports lastused as the time in hz since
we last saw any packets. So let's start tracking it
that way.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When you send a register stop it is not necessary to
have a neighbor out the choosen interface.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When we receive a join/prune for a upstream
that we are unable to create, safely ignore
the request until the situation resolves itself.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When starting up allow the 1st query to go out immediately,
well ok 1 second, and then the next queries to go out in
query_interval/4 seconds until startup mode is finished.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When we are disabling pim, there exists some
race conditions where we are attempting
to send a register stop out a interface that
is not setup for pim yet.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When we do a RPF lookup return a path that we have
neighbors for in those cases where we need to have
a neighbor to pass along the SG state via a pim
join/prune message.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When we reject a register message from someone, give some reasoning
as to the why of it being rejected to help in debugging the situation.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
Pim sometimes needs the upstream rpf lookup to
only take into account if we have a nbr out
the selected interface or not. Move
the code for this to a better spot so
we can make a more intelligent decision
here.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When doing a rpf check ensure that
if we've considered the RP to be a loopback
allow there to be no neighbor on the other
side.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When looking up nexthops for a path, we
should only allow nexthops that we
actually have neighbors formed for. Otherwise
when we send join/prune messages they will
do nothing.
Ticket:CM-12754
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
Do not assert when looking up a neighbor, we don't know
if we have a neighbor don't punish us.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
NEXTHOP_TYPE_IPV4 has the ifindex of the route. Pass it
along so the other side can use it if it is needed.
This will make pim much happier in that we will need to do less
recursive lookups.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When looking up nexthops for a address, just
use the MULTIPATH_NUM as the number allowed
to receive instead of an arbitrary limit of 20 paths.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
We need the ability to know where upstream state
comes from and to do the right thing from there.
ticket: CM-12771
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When rebooting a switch, if under unnumbered config
there exists a situation where some interfaces
may not create the vif in the kernel.
Ticket: CM-12830
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When pim starts up there is no need to assert in pim_init
if state is not set as wanted. Especially for qpim_debugs
as that I would like to turn on debugs before we even start
doing anything in the code.
As for the PIM_MROUTE_IS_ENABLED, the variable is set
right above it, so no need to assert on it.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When a interface goes down we accidently put ourselves into
an infinite loop.
Ticket: CM-12803
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
In pim_rpf.c when we receive multiple paths for the
rpf we are complaining that we are not sure what
to do with it yet. Change the _info to _debug
and don't print out unless we are debugging
stuff.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When we get a v6 address/ifindex combo back from
a v4 lookup. Check to see if the outgoing
interface only has one neighbor and if so
use that neighbors loopback address as
the route to use.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When a interface is configured as BGP unnumbered, it
has a v6 LL address as well as no v4 addresses. In
this case let's look at the lo's ip address as the
primary address to use.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
rebase
The recent conversion of in_addr_t to struct prefix
caused a display issue in the 'show run' of the rp.
Ticket: CM-12752
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
The pim_find_primary_addr function just called 1 static function
that was called no where else.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
Restructure code to remove SA warning from clang build.
For some reason, the Clang SA system thought that si.sin_port and
si.sin_addr where not being set to anything. Fix this problem.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When I converted over to using 'struct prefix'
I broke the initialization of the rp.
In addition, I used the wrong AFI type
to switch on in pim_rpf.c
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
mrib_nexthop_addr and rpf_addr should be 'struct prefix'
so that we can safely handle unnumbered data from a nexthop
lookup in zebra
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
This is setup code to allow us to receive v6 addresses from
a nexthop lookup from zebra. This will allow us to work
with unnumbered interfaces.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
In the nexthop lookup routine, fix the duplicate code
to not check for too many indexes earlier.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
Specify the ifindex of the interface that we want
to turn on IP_MULTICAST_IF on instead of the
address. If we pass in the address then the
fact that we have multiple interfaces with the
same ip address causes confusion in the kernel.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
This has 2 fixes:
On nocache event, crate the channel oil first in
case we don't need to actually create the upstream
information
on wrvifwhole, create the channel oil and install
it, then only create the upstream information
if we are connected to the source.
Ticket: CM-12593
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
If a neighbor comes up *after* igmp joins and *after* all
the routes are installed into the kernel, then we never
go back and fix up the rpf cache information. So joins
never go out for those igmp joins.
Ticket: CM-12613
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
The rpf debugs really should be zebra debugs
instead of trace debugs.
Additionally fix a return 0; to return NULL;
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
The original modification to allow the RP to receive
packets on the incoming interface when it has
no outgoing interfaces was a bit aggressive and
turned all register events into a join event.
This further divides up the code to do the
right thing in each situation.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When starting up with a large # of streams
it is possible to have a timing where we
do not yet have a pim_ifp.
Ticket: CM-12636
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When installing a mroute on rpf change make sure
that we have oil for it at this time.
Ticket: CM-12636
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
If we receive a join for a upstream interface that hasn't
yet been added to pim for working under it, we never
go back and add the forwarding on when we get the
interface added to pim.
Ticket: CM-11800
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When the RP's address in PIM is not under the loopback,
we can have a situation where this causes PIM to crash.
Check for NULL pointers.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
List the RP information we have configured.
Fix bug where we were not properly initializing some code
Ticket: CM-12617
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
If on the FHR we have an established stream between
interested parties and we stop the stream and restart
quagga we were not coming back up into the correct
state. For the multicast stream.
Ticket: CM-12462
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When register stop was received but the upstream
creation fails we are not handling it properly.
Note and return.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
Move the pim_mroute_conected_to_source function to pim_iface.c
because this is a interface question not a mroute question.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
On the RP when we receive a register packet for the
same interface that we received the *,G join on
allow the RP to receive the packets but to do
nothing with them.
Additionally check the last_used values even
if the ioctl fails.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When sending register packets to the RP from the FHR
we should be using the ip address of the incoming interface
that received the mcast packet.
Ticket: CM-12445
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
This fix allows the RP to transition from a Prune'd to Join
state on reception of proper control messages.
Additionally it was noticed on if down -> if up transitions
the prune then join message was going out the wrong interface.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
I tried to be smart and skirt around rpf lookup if I knew
the incoming interface. This turns out to be not necessarily
a good thing because we can easily have asymetrical routing.
This fix removes the attempt to cache the ifp we received
the incoming packet on and just lets the lookup work like
it should.
Additionally it removes the weird hardcoding of the rpf
interface from the register stuff.
Ticket: CM-12530
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
reb
Ensure that when an upstream is removed that we actually
remove the route
Also when we get a callback for whole packet before
we send a register stop make sure we are actually on the
RP.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When we are a FHR, and the upstream state is in Prune or Prune Pending
allow the join because we know we want to forward traffic to the RP.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
Just pass in the actual state to allow us to translate
the state to a string even if we don't have a upstream.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When the Keep alive timer Pops stop the timer
and send a prune upstream if we need to
remove the state.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
The RP needs to send the join when we want to send it
Also make sure the ka_timer is reset properly.
Ticket: CM-12473
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When there is allot of IGMP activity for the
kernel upcall that we are ignoring, just
ignore it for the moment as well.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When on a intermediate switch between the FHR and RP
when Quagga is restarted. Pimd is coming up so fast
and getting the S,G join before networking is properly
working on the switch. We would never install the
mroute in this case.
What was happening is that the scan of the S,G was
occurring and we were figuring out who we should talk
to but we were never installing the route for it.
Notice that we've never installed the route and install it
if the mroute has changed.
Ticket: CM-12460
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When the keep alive timer finds that we are no longer
sending packets for that particular S,G, remove
the upstream information but use the function that
makes sure we check the ref count.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When we receive a (*,G) route decide if we
need to cause any output interfaces to be
added to any S,G route associated.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When on the RP, if we receive the first correctly routed
mroute from upstream, Set the SPT bit so that the
state machine will work properly.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
1) Ensure Timer is actually off
2) I missread the rfc, when we are in prune state, just send the
register stop. Do not add the tunnel device yet.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
The RP was not properly handling the series of events:
1) When a WRVIFWHOLE is received if we are the RP,
send a pim register stop to the FHR.
2) When a register was received we were sending
a join (S,G) towards the S, then a immediate prune (S,G)
followed by another join (S,G). Just send the first join
3) Save whom we received the S,G register from so we
can use it later
4) Allow a join timer to restart itself instead
of causing a crash.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
pim_zebra.c was checking to see if a ifp pointer was null
in some debug statements. This caused the SA code
to assume that if we were at this point then the ifp
could be null and subsuquent code that doesn't check
for it to be wrong.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
SA found these two issues:
1) Check of ifp non-null caused other non-checks down code
to assume ifp could be null. This is not the case.
If we get a callback at this point we must have a valid ifp pointer.
2) We were inadvertently calling igmp_source_reset_gmi with
a possible NULL source pointer.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
This commit addresses some SA issues found:
1) double set variables
2) value set but never used.
3) wrong return path used.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When we receive a new genid from a neighbor, we need
to form the neigbhor relationship before the join/prune
messages are sent to the neighbor.
Additionally we were calling the
pim_upstream_rpf_genid_changed function 2 times
in a row. This function just spun throught the upstream
list and marked all relevant upstreams to be sent
immediately
Ticket:CM-11979
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
Allow the user to specify multiple rp commands.
'ip pim rp A.B.C.D' -> translates to 'ip pim rp A.G.C.D 224.0.0.0/24'
ip pim rp A.B.C.D A.B.C.D/M
First is the rp, second is the group with mask.
Groups and masks cannot be over each other except 224.0.0.0/24 which
is the fallback if used.
Ticket: CM-7860
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
Refactor the qpim_rp into pim_rp.c so that the global data
is protected. This will allow us to easily add the group
ranges.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When we are evaluating if we have a inherited_olist, include
the *,G state in determing what to do.
Additionally when we receive a join or prune of *,G
do the right thing.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
Fix the struct prefix to be an actual struct prefix_sg.
This cleans up a bunch of code to make it look nicer.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
RFC 4601 4.5.7:
Joined
The downstream state machines and local membership information
indicate that the router should join the shortest-path tree for
this (S,G).
The code has a upstream state that is in 'J' state already
due to a S,G multicast packet received on an incoming
interface. This packet has been forwarded to the RP.
The RP sees this and immediately sends a join towards
the S,G.
The code as originally written assumed that you could
not transition from a J state to a J state. This is not
true.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
The PIMREG ifchannel state is an evolutionary dead-end.
I should not have introduced this code.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When we receive a NOCACHE upcall, add the pimreg interface
to the oil so that we can receive the data.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When a *,G is received, and we have S,G that is in a
non-joined state, Send the join to the S,G.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
We were erroneously sending a prune *,G immediately followed
by a join *,G under certain situations
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
Each interface needs the ability to send raw packets
at the RP. Create the data structure to do so.
Further commits will create the socket as necessary
and send the packet.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
Given a S,G ifchannel, create a parent pointer for any
*,G state for that interface. When removing a *,G
remove S,G pointers.
We need to think about the case where their is a *,G
but no S,G for a interface when we are determing.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When we call pim_str_sg_dump we cannot call the variable twice
for the same debug. This fixes this issue in the 2 places it
was borked.
Ticket: CM-12045
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When a register is received, forward the packet as appropriate.
This is the infrastructure to make this happen.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When a *,G report is sent, the RPT and WC bits
are set as well as the source is the RP address
for the group.
If this routers idea of the RP for this group is
different than the idea of the RP from the sender
than that particular *,G can be dropped.
Ticket: CM-12031
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
The kernel upcalls for control messages should
be PIM_DEBUG_MROUTE instead of PIM_DEBUG_PIM_TRACE
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
For some reason Join/Prune debugs were being
turned on by trace not by the 'debug pim packets join'
command.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
Add some missing 'show debug pim' output so that I can
see what I think I'm debugging.
In general the pim debugs need another round of fixing.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
We were not correctly determing the inherited olist and
applying it to the s,g.
Start the work to do this.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When we received a pim_register_recv we were shortcutting
the algorithm to determine what to do when we receive
a register pim message.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
The pim_upstream_find function should not be finding a *,G
Find what is asked of it.
The switch_to_spt_desired should return true on the RP currently.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
Convert more of the code to pass around a 'struct prefix sg'
instead of individual struct addr's.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
Convert ifchannel to use a prefix to store (s,g) information.
Additionally create pim_str_sg_dump function to output
data when using debugs.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
pim was sending the join from the RP towards the FHR with
a address that was the source instead of the nexthop. This join
was treated as a non-local and nothing was done with it.
Ticket: CM-11945
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When we receive a register for a (S,G) that we have no interest in
prune it off and send the register stop.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
Allow pim_upstream_switch to be called from outside of pim_upstream.c
Also change the assert to an if test to just correctly stop the
timer.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
Change the pim message type to an enum and add the ability
to output a string based upon message type.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
Implement the pim_register_stop state machine. There are still a few
bugs still but this is enough to get us rolling a bit more.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
Remove the prune_echo function as that it just took some
data and passed it into another function.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When a register stop is received, create the register stop
timer and associated state machine.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When sending a pim register message to an RP, we need
the ability to set the null bit in the message. This
adds the ability to do so.
Additionally we need to switch the ip_hdr to a const uint8_t *buf
because pim_register_send really shouldn't know or care about
the fact it's a 'struct ip'.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
For a given interface remove the interface from the list
of outgoing interfaces for a mroute and reinstall it
into the kernel.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
When we receive a register stop we will have created a
timer for this. As such if the t_rs_timer pointer will
be non-NULL at this point.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
We were using a variety of techniques to handle
incoming pim packets. Refactor to use a switch
statement to handle the incoming packets.
Also add the ability to notice that we are getting
a register stop.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
The RP now has the ability to send a register stop message
to the originator of the packet in order to tell it to
stop encapsulating packets to the RP.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
This replaces Quagga -> FRR in most configure.ac settings as well as
a handful of preprocessor macros in the source code.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
This moves all install_element calls into the file where the DEFUNs are
located. This fixes several small related bugs:
- ospf6d wasn't installing a "no interface FOO" command
- zebra had a useless copy of "interface FOO"
- pimd's copy of "interface FOO" was not setting qobj_index, which means
"description LINE" commands would fail with an error
The next commit will do the actual act of making "foo_cmd" static.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
When displaying drpriority you can enter unsigned
integer values from 1-2^32. The display was
turning the unsigned value into a signed value
and thus we were displaying the wrong value.
Ticket: CM-13787
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
Reviewed-by: David Ahern <dsa@cumulusnetworks.com>
(cherry picked from commit db17265f1025e3ec578998ad537c33b49ed33cde)
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
Linux, FreeBSD and NetBSD (and possibly others too) accept both uint8_t
and int for the IP_MULTICAST_LOOP sockoption. OpenBSD, in the other hand,
accepts only uint8_t. To make setting IP_MULTICAST_LOOP work on every
supported platform, always pass a uint8_t variable as a parameter.
Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
Since we have autoconf results from a wide swath of target platforms, we
can go remove checks that have the same result on all systems.
This also removes several "fallback" implementations of functions that,
at some point in the history, weren't available on all target platforms.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
original_s_route is allocated on the heap and was not freed during the
error case.
Signed-off-by: Christian Franke <chris@opensourcerouting.org>
Signed-off-by: Christian Franke <chris@opensourcerouting.org>
Acked-by: Donald Sharp <sharpd@cumulusnetworks.com>
Consolidate the routemap initialization into one
function.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
Reviewed-by: Daniel Walton <dwalton@cumulusnetworks.com>
If a command is put into the VIEW_NODE, it is going into the
ENABLE_NODE as well. This is especially true for show commands.
As such if a command is in both consolidate it down to VIEW_NODE.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
This patch improves zebra,ripd,ripngd,ospfd and bgpd so that they can
make use of 32-bit route tags in the case of zebra,ospf,bgp or 16-bit
route-tags in the case of ripd,ripngd.
It is based on the following patch:
commit d25764028829a3a30cdbabe85f32408a63cccadf
Author: Paul Jakma <paul.jakma@hpe.com>
Date: Fri Jul 1 14:23:45 2016 +0100
*: Widen width of Zserv routing tag field.
But also contains the changes which make this actually useful for all
the daemons.
Signed-off-by: Christian Franke <chris@opensourcerouting.org>
