New constructor can construct the l3 rule
using all or partial paratmeters.
Change-Id: I828ec1c4713decb5824e4a73c3692cebc2324cc2
Signed-off-by: Mohsin Kazmi <sykazmi@cisco.com>
Taken from DPDK, also AVX2 variant updated to be in sync with DPDK
version.
Change-Id: I8a42e4141a5a1a8cfbee328b07bd0c9b38a9eb05
Signed-off-by: Damjan Marion <damarion@cisco.com>
Elastic Network Adapter PMD for newer AWS instance types
Change-Id: Ic7f6ac4a261ccc7af3ffb2ed8950274532e3feae
Signed-off-by: Matthew Smith <mgsmith@netgate.com>
The "test-all" target is still never called as part of any continuous
test (as it probably should) but at least it can now be expected to
succeed.
VXLAN-GPE:
* decapsulate Ethernet to "l2-input" instead of "ethernet-input"
otherwise the inner mac address get checked against the interface one
(external) and packet gets dropped (mac mismatch)
* set packet input sw_if_index to unicast vxlan tunnel for learning
TEST:
* VXLAN:
* reduce the number of share tunnels:
=> reduce test duration by half
=> no functional change
* VXLAN-GPE:
* fix test TearDown() cli: command is "show vxlan-gpe" only
* remove vxlan-gpe specific tests as the were a duplicated of the
BridgeDomain one and already inherited.
* disable test_mcast_rcv() and test_mcast_flood() tests
* P2PEthernetAPI:
* remove test: "create 100k of p2p subifs"
there already is a "create 1k p2p subifs" so this one is a load test
and not a unit test.
See: lists.fd.io/pipermail/vpp-dev/2017-November/007280.html
Change-Id: Icafb83769eb560cbdeb3dc6d1f1d3c23c0901cd9
Signed-off-by: Gabriel Ganne <gabriel.ganne@enea.com>
- compute session type out of transport and network protos
- make session, session lookup and session queue code network protocol
agnostic
This does not update the session layer to support non-ip network layer
protocols
Change-Id: Ifc2f92845e158b649d59462eb7d51c12af536691
Signed-off-by: Florin Coras <fcoras@cisco.com>
The classifier tables layout might (and most always will) change during the MACIP ACL modification.
Furthermore, vnet_set_input_acl_intfc() is quite a picky creature - it quietly does nothing
if there is an existing inacl applied, even if the number is different, so a simple "reapply"
does not work. So, cleanly remove inacl, then reapply when the new tables are ready.
Also, fix the testcase which was supposed to test this exact behavior.
Thanks to Jon Loeliger for spotting this issue.
Change-Id: I7e4bd8023d9de7e914448bb4466c1b0ef6940f58
Signed-off-by: Andrew Yourtchenko <ayourtch@gmail.com>
* app_index has to be valid
* vnet_session needs to be enabled before binding
Change-Id: I410a25153e650d8bc332131a81bf46a0175566d4
Signed-off-by: Andreas Schultz <aschultz@warp10.net>
On the first session server, the first payload packet gets queued
but the Rx callback will only be invoked when the second packet
is received.
By initializing enqueue_epoch in the session, we can make sure that
the RX callback will be invoked for the first packet as well.
Change-Id: I60e2cf8d4d2ca71d461ca003d459c8e6e6956a4d
Signed-off-by: Andreas Schultz <aschultz@warp10.net>
as decsribed in section 2.2
ihttps://tools.ietf.org/html/draft-ietf-bier-mpls-encapsulation-10
with BIFT encoding from:
https://tools.ietf.org/html/draft-wijnandsxu-bier-non-mpls-bift-encoding-00
changes:
1 - introduce the new BIFT lookup table. BIER tables that have an associated
MPLS label are added to the MPLS-FIB. Those that don't are added to the
BIER table
2 - BIER routes that have no associated output MPLS label will add a BIFT label.
3 - The BIER FMask has a path-list as a member to resolve via any possible path.
Change-Id: I1fd4d9dbd074f0e855c16e9329b81460ebe1efce
Signed-off-by: Neale Ranns <nranns@cisco.com>
Also fixes old ixge driver, so it works with recent physmem changes and
vfio.
Change-Id: Id4be74b34daed47cd281a77eec43d6692340d882
Signed-off-by: Damjan Marion <damarion@cisco.com>
JVpp maps request messages with replies
for Java API user convenience, e.g.:
- do not polute send APIs with messages other than requests/dumps,
- allow callback registration only for replies/details and events.
Since there are no conventions for event message naming
(https://wiki.fd.io/view/VPP/API_Concepts#API_Conventions),
jvpp should not limit events to messages
that end with 'event' or 'counters' suffix.
Instead jvpp should treat all messages
except for requests/dumps as potential events.
Such behaviour was introduced on Java API level by
https://gerrit.fd.io/r/#/c/8377/
in order support reusing
details messages as events (e.g. BFD events).
This patch goes one step forward by
relaxing rules at jvpp generation level.
Change-Id: I2a35e9eb2a288b2cf02d36ca95e6cb13e76e19e3
Signed-off-by: Marek Gradzki <mgradzki@cisco.com>
