- ip post-rewrite feature subgraph arc support
Change-Id: Ia4b07197463021ade916326231af246e2559a290
Signed-off-by: Dave Barach <dave@barachs.net>
Signed-off-by: Keith Burns (alagalah) <alagalah@gmail.com>
If a custom fib ID is used (different from ~0), the associated
fib is used to forward outgoing encapsulated packets.
Otherwise, the fib used is the same as for any packet
received on the original RX interface (L2TP does not modify RX interface index).
Change-Id: I4533d5f7fa432c78c937d3acdd802d0d1c92a0c7
Signed-off-by: Pierre Pfister <ppfister@cisco.com>
L2TP tunnels use virtual interfaces but directly send
packets to l2-input node (not ethernet-input).
This node requires a bridge-domain to be associated with
the interface.
Past code was immediatly turning the interface up, but
some packets could be sent to l2-input without bridge domain
between interface creation and association with a bridge domain.
The tunnel is now created as down and has to be set up later
(typically after being associated with a bridge-domain).
Another option would have been to change the api and enforce
a bridge-domain to be specified before the tunnel, but this
is less flexible for the user.
Change-Id: I26d1f36bb4f327d9fa1c8044023f2210c4117904
Signed-off-by: Pierre Pfister <ppfister@cisco.com>
When receiving a local ipv6 packet containing an l2tp packet not associated
with any session, l2tp node was handling the packet as if provided by an ipv6 feature,
hence crashing.
This patch fixes the issue by correctly dropping the packet instead.
This patch also fixes a typo from commit d65346098d.
Change-Id: I1b377fc5685568c16831920227671feffac64287
Signed-off-by: Pierre Pfister <ppfister@cisco.com>
This change-set enables plugins to add themselves to the ip4/ip6
feature subgraphs without having to modify core vpp engine code
at all. Add VNET_IP4/IP6_UNICAST/MULTICAST_FEATURE_INIT macros
which express the required ordering constraints, and off you go.
Along the way, added an implementation of Warshall's algorithm to
vppinfra; to compute the positive transitive closure of a relation. In
this case, the relation is "feature A runs before feature B."
With that in hand, ip_feature_init_cast(...) computes a partial order
across the set of configured feature subgraph nodes.
In unit-testing, we discovered VPP-145 - ip4/6 inacl wiped out
vnet_buffer(b)->ip>current_config_index, which exists in main. So, we
fixed that by moving b->trace_index, adding b->current_config_index,
and removing the ip opaque union current_config_index.
Change-Id: Iff132116f66413dc6b31ac3377198c7a32d51f48
Signed-off-by: Dave Barach <dave@barachs.net>
When the CLI is used to create an interface, and whend the operation
succeeds, the newly created interface name is printed-out.
The patch includes the following interfaces types:
- AF_PACKET
- Vhost User
- Netmap
- GRE
- L2TP
- MPLS-GRE
- Loopback
Change-Id: Id518c139ec63a261eae81d2ed95c4cd1f10b5157
Signed-off-by: Pierre Pfister <ppfister@cisco.com>
* compiler -march= parameter is changed from native to corei7
so code is always genereted with instructions which are available
on the Nehalem microarchitecture (up to SSE4.2)
* compiler -mtune= parameter is added so code is optimized for
corei7-avx which equals to Sandy Bridge microarchitecture
* set of macros is added which allows run-time detection of available
cpu instructions (e.g. clib_cpu_supports_avx())
* set of macros is added which allows us to clone graph node funcitons
where cloned function is optmized for different microarchitecture
Those macros are using following attributes:
__attribute__((flatten))
__attribute__((target("arch=core-avx2)))
I.e. If applied to foo_node_fn() macro will generate cloned
functions foo_node_fn_avx2() and foo_node_fn_avx512() (future)
It will also generate function void * foo_node_fn_multiarch_select()
which detects available instruction set and returns pointer to the
best matching function clone.
Change-Id: I2dce0ac92a5ede95fcb56f47f3d1f3c4c040bac0
Signed-off-by: Damjan Marion <damarion@cisco.com>
I got many warnings at the link step with gcc version 5.3.1 20160225 (Ubuntu 5.3.1-10ubuntu2):
/usr/bin/ld: Warning: size of symbol `cop_input_node' changed from 112 in vnet/cop/.libs/cop.o to 168 in vnet/cop/.libs/node1.o
/usr/bin/ld: Warning: size of symbol `ethernet_input_node' changed from 112 in vnet/.libs/interface.o to 136 in vnet/ethernet/.libs/node.o
/usr/bin/ld: Warning: size of symbol `l2output_node' changed from 112 in vnet/l2/.libs/l2_efp_filter.o to 120 in vnet/l2/.libs/l2_output.o
/usr/bin/ld: Warning: size of symbol `srp_input_node' changed from 112 in vnet/srp/.libs/format.o to 136 in vnet/srp/.libs/node.o
/usr/bin/ld: Warning: size of symbol `vxlan_encap_node' changed from 112 in vnet/vxlan/.libs/vxlan.o to 128 in vnet/vxlan/.libs/encap.o
/usr/bin/ld: Warning: size of symbol `vxlan_input_node' changed from 112 in vnet/vxlan/.libs/vxlan.o to 144 in vnet/vxlan/.libs/decap.o
...
Looking at vlib_node_registration_t, I think the reason is that
the char * next_nodes[] could be bigger where the variable is defined
in .c file.
We should mark global variables as external in header files.
Some of them can be made static.
Change-Id: Ieb6961fd08180c9a69e1d884852703f3eb23f23f
Signed-off-by: Jean-Mickael Guerin <jean-mickael.guerin@6wind.com>
Dave Barach