Type: improvement
Currently, if an error happens on the netlink notification socket, the
socket gets reopened. A portion of notification messages have been lost
because of this and will never arrive on the socket. VPP will be out of
sync with the kernel networking state.
With this change, synchronize VPP state and the kernel networking state
if there was an error polling or reading the notification socket.
Signed-off-by: Alexander Chernavin <achernavin@netgate.com>
Change-Id: I8cdff89b505cd227535a0acaadb6ee757204c45e
Type: refactor
check for pool expansion in the DPO allocation, just in case.
Signed-off-by: Neale Ranns <neale@graphiant.com>
Change-Id: I6ba7dd66313630d3f24a51700ab4486ba43d856b
Type: improvement
For some message types, timestamps are checked on netlink message to
decide whether the message should be applied. For notification messages
timestamps are expected to be always available.
With this change, before accessing the timestamp, make sure the message
info object that carries it is not null. If it is null, pass the check.
This is to be ready to process dump replies that will not need the
timestamp check and will have the message info object set to null.
Signed-off-by: Alexander Chernavin <achernavin@netgate.com>
Change-Id: Ic7211c0d451d72f6a5248898b3a8f8e0bca8f7aa
Type: fix
Currently, a tap interface created to be a member of a linux-cp pair has
default link state (down) and default link speed (10Mb/s). Then the
plugin monitors the link state of the paired hardware interface and if
it changes, the new link state is reflected on the tap interface. And
when the new link state is "up", the link speed is also reflected on the
tap interface.
The problem is that this scheme implies that the hardware interface's
link state is "down" at the moment of the linux-cp pair creation and
then changes. But there are cases when the link state is already "up" at
that moment. If that is the case, the link speed on the tap interface
will remain the default one until the link comes down and then comes up.
With this fix, when a linux-cp pair is created, reflect current link
state of the hardware interface being paired on the created tap
interface.
Signed-off-by: Alexander Chernavin <achernavin@netgate.com>
Change-Id: I73664d753f4daaa6d439c9ca898fb7363d21c06d
Some possible side effects for multicast routes appears to be in
lcp_router_table_add_or_lock. so ff00/8 route will be processed
for ospf purposes the right way
Ignore IPv6 kernel routes
Skip adding auto routes into FIB
Type: fix
Signed-off-by: Vladimir Ratnikov <vratnikov@netgate.com>
Change-Id: I35f73d629a7fffca7f7d4547adc2549b72c2048f
Include the user_instance in wireguard interface details.
In addition to dumping all wireguard interface details,
also allow selective dumping of just one interface.
Type: improvement
Signed-off-by: Jon Loeliger <jdl@netgate.com>
Change-Id: Iaf1093c6ae3eb00a685f34b2e0171285b02fae2b
When building a VPP library out of tree using add_vpp_library, the build
sometimes fails because the library API files are not necessarily
generated before compiling this library.
This is fixed by adding the lib API files as dependencies of the
lib compilation.
Type: fix
Signed-off-by: Guillaume Solignac <gsoligna@cisco.com>
Change-Id: I69dffaecbfd547f10115504494a47358c4624258
- cli_inband is missing from vat2
Type: improvement
Change-Id: I1f22dee3ee29f3cf0f1f7c6076d5f2b2b2bf969d
Signed-off-by: Dave Wallace <dwallacelf@gmail.com>
- A call to vapi_msg_control_ping_hton() is generated in
a static inline function in vlib.api.vapi.h, which is
defined/generated in memclnt.api.vapi.h without
including memclnt.api.vapi.h in vlib.api.vapi.h.
This breaks the compilation of plugins which include
only vlib.api.vapi.h (e.g. hicn_plugin from the HICN
project).
Type: fix
Fixes: a1400cecb
Signed-off-by: Dave Wallace <dwallacelf@gmail.com>
Change-Id: I5574f4ed496183ea93265f493bf3624254a865a2
The app keeps on using the default app ns but each listen will be done
in the vrf configured.
Type: improvement
Signed-off-by: Florin Coras <fcoras@cisco.com>
Change-Id: I0947e03188d55231b299916351115038e0b1f5da
This fixes an issue caused by session layer reading expected part of
data (cert + key) before the client actually sends it.
Type: fix
Signed-off-by: Filip Tehlar <ftehlar@cisco.com>
Change-Id: I6ddddb08f9576211b302e814d7c2b040383e5fb7
- HICN project's hicn_plugin requires vnet header files
fib/fib_entry_track.h and udp/udp_encap.h to be
included in build-root/install-vpp*-*/vpp/include/vnet
Type: fix
Signed-off-by: Dave Wallace <dwallacelf@gmail.com>
Change-Id: Iabd3f8fe0aee8d727758fc6ef202e859d68d63a3
Type: fix
Fixed coverity-issue CID 248517.
Originally possible passing null pointer to one function and directly
dereferences it.
This patch fixes the issue by add a new condition to check this
pointer.
Change-Id: If506abaf08c9f003860b641971af291f68613c18
Signed-off-by: Gabriel Oginski <gabrielx.oginski@intel.com>
Type: fix
Fixes: 39e9428b90
When a VRRP advertisement is received by a worker thread, the worker
calls vl_api_rpc_call_main_thread() so the main thread will process the
packet and make adjustments to VR state if necessary.
The data being passed to the main thread included a pointer to the VRRP
header in the received packet buffer. Since the main thread processes
the RPC request asynchronously from the worker thread, it's possible for
the worker to drop the packet and for the buffer to be overwritten before
the main thread can process it.
Copy the fields which may be needed by the main thread into a struct
instead of passing a pointer to a packet buffer.
Change-Id: I4e899e967df5a54776b521825a80e9cce1a94f5f
Signed-off-by: Matthew Smith <mgsmith@netgate.com>
Integer math on 32 bytes of base64 data might yield 33 bytes
of data in some poorly formed user input of private key values.
Rather than smashing the stack (detected) and aborting, simply
allow for the possible yet irrelevant 33-rd byte of data.
Type: fix
Fixes: edca1325cf
Change-Id: I42acfbf3e8fbb3d517e21c53d4f80459d4800e9d
Signed-off-by: Jon Loeliger <jdl@netgate.com>
Type: fix
otherwise the feature is enabled n times for n lcp-pairs and the packets go n times through the feature.
Signed-off-by: Neale Ranns <neale@graphiant.com>
Change-Id: I8e47e4a293d6e2711f54aa09e9545e5e07728026
adj_get() crashed in worker thread. The index is valid but was freed.
0x00007fba5fb07921 in __GI_abort () at abort.c:79
0x00007fba528de7c3 in os_panic () at /development/libvpp/src/vpp/vnet/main.c:618
0x00007fba506f2a19 in debugger () at /development/libvpp/src/vppinfra/error.c:84
0x00007fba506f2797 in _clib_error (how_to_die=2, function_name=0x0, line_number=0, fmt=0x7fba521cecd9 "%s:%d (%s) assertion `%s' fails") at /development/libvpp/src/vppinfra/error.c:143
0x00007fba51d88254 in adj_get (adj_index=79) at /development/libvpp/src/vnet/adj/adj.h:472
0x00007fba51d87642 in icmp6_neighbor_solicitation_or_advertisement (vm=0x7fb9f19a0580, node=0x7fb9bb820340, frame=0x7fb9b718bc40, is_solicitation=0) at /development/libvpp/src/vnet/ip6-nd/ip6_nd.c:134
0x00007fba51d8f479 in icmp6_neighbor_advertisement (vm=0x7fb9f19a0580, node=0x7fb9bb820340, frame=0x7fb9b718bc40) at /development/libvpp/src/vnet/ip6-nd/ip6_nd.c:348
Type: fix
Signed-off-by: Rajith P R <rajith@rtbrick.com>
Change-Id: Ibe29854137ee9680f7947450eb7e256b8c0ade31
mpls fib DB size was 2^20 instead of intended 2^21.
Therefore large mpls labels caused DB to overflow and write
to other tables or some random objects. Or crash with ASAN.
Sometimes.
Type: fix
Signed-off-by: Dmitry Valter <d-valter@yandex-team.ru>
Change-Id: I6db65680037a266fe47e8213464a0c16de63c06c
TCP/UDP packets which are dropped are not counted towards TCP/UDP
counters. Apply same behaviour to ICMP packets.
Type: improvement
Signed-off-by: Klement Sekera <klement.sekera@gmail.com>
Change-Id: I487fa7135ac8e49431a621fac213638d2dab31c8
Breakdown the distribution of uops delivered to the frontend.
Collerates directly with the source of the uops.
Type: improvement
Signed-off-by: Ray Kinsella <mdr@ashroe.eu>
Change-Id: I93a57dbe56dfa0f378527844aa4e63f45a548e55
Type: improvement
remove the [un]lock logs, they are not useful.
Signed-off-by: Neale Ranns <neale@graphiant.com>
Change-Id: I903d3088b8ed9831f931208aeb6b6862a945550c
This is required after distinguishing between max_frame_size and MTU
Type: fix
Signed-off-by: Artem Glazychev <artem.glazychev@xored.com>
Change-Id: Ie642bee4e30ca76903bb8be5eeb6914c2c09bf35
Type: fix
Fixed coverity-issue CID 248517.
Originally possibly passing null pointer to one function and
directly dereferences it.
This patch fixes the problem by add a new condition.
Signed-off-by: Gabriel Oginski <gabrielx.oginski@intel.com>
Change-Id: I02fc6fb5d1cfd6138ea4ba2b1946fd8a7ef34d3b
Originally after remove the policy entry in spd, macro "vec_del1"
can change localization of the last entry in vector and finally the
entry list has not been sorted.
This patch fixes this issue by change executed macro "vec_delete"
instead of "vec_del1".
Type: fix
Signed-off-by: Gabriel Oginski <gabrielx.oginski@intel.com>
Change-Id: I396591cbbe17646e1d243aedb4cdc272ed4d5e25
If the user supplies an instance number when creating a wireguard
interface, it should be used in the device name. If no user-instance
is given, the device instnce (tunnel number) is used.
For example:
vpp# wireguard create instance 17 listen-port 23023 src 1.2.3.4
private-key MDEwMjAzMDQwNTA2MDcwODA5MTAxMTEyMTMxNDE1MTY=
wg17
Type: fix
Fixes: edca1325cf
Signed-off-by: Jon Loeliger <jdl@netgate.com>
Change-Id: Ibf51868bf7b4b9e8a3f85557d05667207873bc91
Supported only when eventfd option is enabled.
Type: feature
Change-Id: Ic9d6e38604e978f7bc8e54d74fe9b8f3fc53622d
Signed-off-by: Filip Tehlar <ftehlar@cisco.com>
Type: refactor
To be consistent with the location of the IPv4 responder
Signed-off-by: Neale Ranns <neale@graphiant.com>
Change-Id: Ie3a5c3ecc10755317591d7ff57b74770c2798e77
Type: improvement
Ethernet frames on the wire are a minimum of 64 bytes, so use the length in the UDP header to determine if the ESP payload is one bytes of the special SPI, rather than the buffer's size (which will include the ethernet header's padding).
In the case of drop advance the packet back to the IP header so the ipx-drop node sees a sane packet.
Signed-off-by: Neale Ranns <neale@graphiant.com>
Change-Id: Ic3b75487919f0c77507d6f725bd11202bc5afee8
Originally XDP_FLAGS_REPLACE was passed to the function.
For kernels not defining this macro (for example 5.4) in if_link.h vpp
sets it to 0.
If kernel has this macro problem appears, replace flag requires
specifying the program to be replaced. bpf_set_link_xdp_fd wraps
around __bpf_set_link_xdp_fd_replace and passes 0 as old_fd, it leads to
an error while assessing for replacement (if 0 is passed
there is no assessing for replacement).
To address this issue no flag is passed to the function, only 0.
Type: improvement
Signed-off-by: Dastin Wilski <dastin.wilski@gmail.com>
Change-Id: I3689ce7eb8c71c699f0e589111929979c2bbe213
Type: improvement
When the adj subsystem is notified of a BFD session, it attempts to find the appropriate adjacency from the session's key.
This could lead to a mismatch between the adj used by BFD and that of FIB. The BFD session stores the adj it is using, so FIB uses that instead.
Since adj is now using the same adj as BFD, it does not need to maintain its own locks.
In BFD it is necessary to initialise the adj index used in INVALID and ensure it is not unlock before listeners are notified of the session delete.
Signed-off-by: Neale Ranns <neale@graphiant.com>
Change-Id: I9630867b10bb18969475299a0c754942a8df0f44
Type: fix
The adjacency used is then the same one as that used by routes in the FIB and so the BFD protection/fast-failover works for thise route, since they are children of the BFD protected adjacency.
Signed-off-by: Neale Ranns <neale@graphiant.com>
Change-Id: I72e10b3074697cba8a002a4c1acf690983985157
When the ipfix address was changed to be an ip_address instead of
an ip4_address the output when creating an exporter via the cli
should have been modified to take the address of the v4 part of
the addr.
Type: fix
Signed-off-by: Paul Atkins <patkins@graphiant.com>
Change-Id: I141456cd9092c861a4c4aefba4035dbde23efcd6
Alexander Chernavin