When n_rx_packets is less then 16(VEC256) or 8(VEC128), code execution
will fall into scalar path of processing packets. But with a wrong
initialization value for n_left set to zero, i in the for-loop will
equal to n_rx_packets. This leads to the bypass of required ip4 checksum
validation and byte count endianness conversion in scalar path.
Besides, refactor the code using while instead of for-loop to keep
consistency with VPP code style.
Type: fix
Fixes: bf93670c515d ("rdma: fix ipv4 checksum check in rdma-input node")
Signed-off-by: Lijian Zhang <lijian.zhang@arm.com>
Signed-off-by: Jieqiang Wang <jieqiang.wang@arm.com>
Change-Id: Ib4e8cb5202735f8b060c99caddf26035657551e1
Type: fix
Fixes: a51f9b3747
Some IPsec message type definitions were not being written to
ipsec.api.vapi.h. These include ipsec_sad_entry_add_del_v3 and
ipsec_sad_entry_add.
The cause appears to be that tunnel_flags, which is defined in
tunnel_types.api is a special case of enum called an enumflag. These do
not appear to have been handled in the code that generates the vapi
header files.
This patch adds processing of enumflag objects for vapi.
Change-Id: Ie506c4fcb5a07fe97a330ba11c252d1df98adfd9
Signed-off-by: Matthew Smith <mgsmith@netgate.com>
Fast path spd was explicitely storing array of policy id vectors.
This information was redundand, as this inofrmation is already stored
in bihash table. This additional array was affecting performance
when adding and removing fast path policies.
The other place that needed refactoring after removing this array was
cli command showing fast path policies.
Type: feature
Signed-off-by: Piotr Bronowski <piotrx.bronowski@intel.com>
Change-Id: I78d45653f71539e7ba90ff5d2834451f83ead4be
Not ideal. The sparse vector used to map ports to next nodes assumes
only a few ports are ever used. When udp transport is enabled this does
not hold and, to make matters worse, ports are consumed in a random
order.
This can lead to a lot of slow updates to internal data structures
which in turn can slow udp connection allocations until all ports are
eventually consumed.
Consequently, reallocate sparse vector, preallocate all ports and have
them point to UDP_NO_NODE_SET. We could consider switching the sparse
vector to a preallocated vector but that would increase memory
consumption for vpp deployments that do not rely on host stack.
For reference, populating one of the v4 or v6 sparse vectors in reverse
order takes about 9.8s on a skylake cpu.
Type: improvement
Signed-off-by: Florin Coras <fcoras@cisco.com>
Change-Id: Id795e1805d0d3ba54f56a152a9506a7a2a06ecbc
Implements the API for SRv6 Path Tracing
Type: feature
Signed-off-by: Julian Klaiber <julian@klaiber.me>
Change-Id: Iefa7e512c8e1894595a9e3f5d42eab4160db1f28
In order to loop over the list of `vl_api_ip_address_details_t`,
increment the pointer by one instead of `i`.
Type: fix
Change-Id: I8554d1388d67bb95e029eddf444d383fb85ecac7
Signed-off-by: Timur Celik <mail@timurcelik.de>
Type: fix
Fib entries for attached routes when sourced from
FIB_SOURCE_API or FIB_SOURCE_CLI
get the FIB_ENTRY_FLAG_ATTACHED flag raised on the source.
Such a route added from linux-cp doesn't get this flag.
Fix this flag for linux-cp sources by passing it to the
fib entry's update/create function in lcp_router_route_add().
Signed-off-by: Alexander Skorichenko <askorichenko@netgate.com>
Change-Id: I24278ef86886cfee8a14acb250fb6992a754cc3c
Type: docs
Several kinds of policers are implemented in VPP.
However, they could differ from the RFCs it is
said they are from.
Additionally, the CLI command's help has been
updated with the current list of acceptable
parameters.
Signed-off-by: Maxime Peim <mpeim@cisco.com>
Change-Id: Ic9bf94e1094bea0fcc87ccaa882c2c5f88824041
We add the crypto key to the vnet crypto library via vnet_crypto_key_add.
However, when the session is disconnected, we don't call
vnet_crypto_key_del and the memory is leaked in vnet_crypto library
as well as in pico tls key store.
It seems dispose crypto is the appropriate place to add
vnet_crypto_key_del.
Type: fix
Signed-off-by: Steven Luong <sluong@cisco.com>
Change-Id: If6d1266baf686fefe5bb81330ce60b35c8ff574e
Plugin checks just for AVX2 instruction set, while the v1.3 of IPsec
Multi-Buffer library checks for both AVX2 and BMI2 sets during init.
VirtualBox VM doesn't provide BMI2 by default to guest operating system.
Result is that VPP plugin decides to use AVX2 initialization and library
then doesn't do it. Since flush_job remains empty, the self-check fails
and with that the whole VPP crashes on start-up.
Type: fix
Signed-off-by: Maros Ondrejicka <maros.ondrejicka@pantheon.tech>
Change-Id: I6b661f2b9bbe6dd03b499c55c38a9b814e6d718a
Make sure they only double in size.
Type: improvement
Signed-off-by: Florin Coras <fcoras@cisco.com>
Change-Id: I18d5508c7f32836deb3b25943e8e3af39d0dbc33
API refactoring moved the address-family tag from rule
level down to prefix level.
This necessarily warrants the check that they are the same.
Also, add a check that the address family is sane.
Change-Id: Ia63b688cc9e7c9e9cc773e89708d9e9f99185fb7
Type: fix
Signed-off-by: Andrew Yourtchenko <ayourtch@gmail.com>
dd2f12ba made use of ip46_address_is_ip4() in order to determine whether
the address is ipv4 or ipv6 within unformat_ip_address - however, its
logic is correct only for some addresses. e.g. a valid IPv6 address of :: (unspecified)
will result in "true" result. This is probably not an issue for most
of the cases (the unspecified address is quite rare),
however if the unformat_ip_address is used as part of the
prefix parsing, the ::/0 is a fairly often utilized construct,
which gets parsed as 0.0.0.0
Solution: return the old logic, but use a temporary
variable to avoid overwriting the target memory on failure.
Type: fix
Fixes: dd2f12ba6ab952d9d66f4d9ba89ffde6309b1ff2.
Change-Id: I272f740dfdf07036cec68516e153f0701a53233d
Signed-off-by: Andrew Yourtchenko <ayourtch@gmail.com>
cli show policer's help info is not consistent
with it's arguments.
Type: improvement
Signed-off-by: Huawei LI <lihuawei_zzu@163.com>
Change-Id: I8332fe97ba343e98511db9ff1bb6afd6f3c657cd
Each NAT44 ED session has a per_vrf_sessions_index referencing
an element in the thread-local vector per_vrf_sessions_vec.
However this index can be possibly invalidated by vec_del1() in
per_vrf_sessions_cleanup(), before a session is registered.
Such a stale index can cause an assertion failure in function
per_vrf_sessions_is_expired() when we use it to locate the
per_vrf_sessions object.
A possible sequence to reproduce is:
1. Create two NAT44 ED sessions s1, s2 so that two per_vrf_sessions are created:
index 0: between VRF pair 10 and 11 (expired=0, ses_count=1)
index 1: between VRF pair 20 and 21 (expired=0, ses_count=1)
For the sessions we have:
s1->per_vrf_sessions_index == 0
s2->per_vrf_sessions_index == 1
2. Delete the first session via CLI, now the two per_vrf_sessions become:
index 0: between VRF pair 10 and 11 (expired=0, ses_count=0)
index 1: between VRF pair 20 and 21 (expired=0, ses_count=1)
For the sessions we have:
s2->per_vrf_sessions_index == 1
3. Delete the VRF 11:
index 0: between VRF pair 10 and 11 (expired=1, ses_count=0)
index 1: between VRF pair 20 and 21 (expired=0, ses_count=1)
For the sessions we have:
s2->per_vrf_sessions_index == 1
4. Create a new session s3 between VRF pair 20 and 21 so that the first
per_vrf_sessions will be deleted:
index 0: between VRF pair 20 and 21 (expired=0, ses_count=2)
For the sessions we have:
s2->per_vrf_sessions_index == 1
s3->per_vrf_sessions_index == 0
Here, note that the actual index of per_vrf_session is changed due
to vec_del1(). The new session is added after the cleanup so it gets
the correct index. But the index held by the existing session is not
updated.
5. Trigger the fast path of the session s2. To achieve this, session
s2 could be created in step 1 by
ping -i20 -Iiface_in_vrf_10 1.1.1.1
and steps 2-4 should then be performed within the 20-second interval.
This patch fixes this by changing per_vrf_sessions_vec to a pool so
that indicies are kept intact.
Type: fix
Signed-off-by: Jing Peng <jing@meter.com>
Change-Id: I4c08f9bfd50134bcb5f08e50ad61af2bddbcb645
fix byte order error about the struct snat_address_t's member net.
for example configurations:
set interface ip table loop1 1
set interface ip addr loop1 10.10.10.2/24
nat44 add address 10.10.10.2 tenant-vrf 1
the snat address's net should be "as_u8 = {0xa, 0xa, 0xa, 0x0}",
but now it's "as_u8 = {0x0, 0xa, 0xa, 0x2}" because of missing
transition of byte order about the member net of snat_address_t.
(gdb) p/x *snat_main->addresses
$3 = {addr = {data = {0xa, 0xa, 0xa, 0x2}, data_u32 = 0x20a0a0a,
as_u8 = {0xa, 0xa, 0xa, 0x2}, as_u16 = {0xa0a, 0x20a},
as_u32 = 0x20a0a0a}, net = {data = {0x0, 0xa, 0xa, 0x2},
data_u32 = 0x20a0a00, as_u8 = {0x0, 0xa, 0xa, 0x2},
as_u16 = {0xa00, 0x20a}, as_u32 = 0x20a0a00},
sw_if_index = 0x3, fib_index = 0x1,addr_len = 0x18}
(gdb)
Type: fix
Signed-off-by: Huawei LI <lihuawei_zzu@163.com>
Change-Id: I4f25f0639ae90a7f2e8715b44f825571283d994d
Type: fix
The router flag on a neighbor advertisement can be used by neighbors to
detect that a router has changed to a host (RFC 4861 section 4.4).
If a neighbor adds routes after receiving a router advertisement sent
by VPP and subsequently receives a neighbor advertisement sent by VPP,
it may remove any routes it added based on the RA if the NA does not
have the router flag set. It appears that this is how windows behaves.
When sending a neighbor advertisement, set the router flag if sending
RAs is enabled on the interface.
Signed-off-by: Matthew Smith <mgsmith@netgate.com>
Change-Id: I1f3e42bbd8ea1a4c116b1ce5a8273652d4cd763d
Otherwise, the newly configured interface will never send RADV's.
See below. In the typical case, suppress = 0 and is_no = 0, which
propagates the current value of radv->send_radv:
radv_info->send_radv =
(suppress != 0) ? ((is_no != 0) ? 1 : 0) : radv_info->send_radv;
No other bit of code will set send_radv, at least in straightforward
ways.
Type:fix
Signed-off-by: Dave Barach <dave@barachs.net>
Change-Id: If9368155f7676460ca1f87729c2b3c453405d08d
When echo client fails to connect to remote, it should quit.
Type: fix
Signed-off-by: Steven Luong <sluong@cisco.com>
Change-Id: I787423bdc61a58eea48bab7bd8b73137626c02b4
Type: refactor
1) Removed deprecated API.
- These specific APIs do not have repleacement
because features that they controled
aren't part of current NAT44-ED
implementation anymore.
2) Removed unused typedef of port allocation funciton.
- Missed left over removed.
Change-Id: Ib3f763449065eda7cdcb2c6565a9cae51baf23d6
Signed-off-by: Filip Varga <filipvarga89@gmail.com>
_clib_error (how_to_die=2, function_name=0x0, line_number=0, fmt=0x7fffb3a7e1b5 "%s:%d (%s) assertion `%s' fails") at src/vppinfra/error.c:143
mbedtls_ctx_get (ctx_index=0) at src/plugins/tlsmbedtls/tls_mbedtls.c:114
tls_ctx_get (ctx_handle=536870912) at src/vnet/tls/tls.c:310
tls_app_session_cleanup (s=0x7fffbf102040, ntf=SESSION_CLEANUP_SESSION) at src/vnet/tls/tls.c:624
app_worker_cleanup_notify (app_wrk=0x7fffbef95f80, s=0x7fffbf102040, ntf=SESSION_CLEANUP_SESSION) at src/vnet/session/application_worker.c:445
session_cleanup_notify (s=0x7fffbf102040, ntf=SESSION_CLEANUP_SESSION) at src/vnet/session/session.c:262
session_free_w_fifos (s=0x7fffbf102040) at src/vnet/session/session.c:268
session_delete (s=0x7fffbf102040) at src/vnet/session/session.c:287
session_transport_delete_notify (tc=0x7fffbdf63c40) at src/vnet/session/session.c:1159
tcp_handle_cleanups (wrk=0x7fffbef46d40, now=133.30033046694487) at src/vnet/tcp/tcp.c:1298
tcp_update_time (now=133.30033046694487, thread_index=2 '\002') at src/vnet/tcp/tcp.c:1309
session_update_time_subscribers (smm=0x7ffff7f75ce0 <session_main>, now=133.30033046694487, thread_index=2) at src/vnet/session/session_node.c:1817
session_queue_node_fn (vm=0x7fffbdfad1c0, node=0x7fffbe0b1340, frame=0x0) at src/vnet/session/session_node.c:1934
dispatch_node (vm=0x7fffbdfad1c0, node=0x7fffbe0b1340, type=VLIB_NODE_TYPE_INPUT, dispatch_state=VLIB_NODE_STATE_POLLING, frame=0x0, last_time_stamp=4722227957546624) at src/vlib/main.c:960
Putting a breakpoint in gdb, I found out ctx was free in mbedtls_app_close.
Looking at app_close function in picotls and openssl, I don't see they
free ctx and they don't crash when processing cleanup. I am inclined to
think that mbedtls_ctx_free should not be called in mbedtls_app_close
at src/plugins/tlsmbedtls/tls_mbedtls.c:92
at src/plugins/tlsmbedtls/tls_mbedtls.c:559
at src/vnet/tls/tls.c:360
thread_index=2) at src/vnet/tls/tls.c:762
conn_index=536870912, thread_index=2 '\002')
at src/vnet/session/transport.c:332
at src/vnet/session/session.c:1608
elt=0x7fffbdfef3dc)
at src/vnet/session/session_node.c:1672
node=0x7fffbe0b1340, frame=0x0)
at src/vnet/session/session_node.c:1966
node=0x7fffbe0b1340, type=VLIB_NODE_TYPE_INPUT,
dispatch_state=VLIB_NODE_STATE_POLLING, frame=0x0,
last_time_stamp=4721919444027682)
at src/vlib/main.c:960
Type: fix
Signed-off-by: Steven Luong <sluong@cisco.com>
Change-Id: Ic5c13e659aee618c8accee42af9f40931b62f467
The tx_fifo of session may not be set up yet, if app request to
disconnect the session, svm_fifo_dequeue_drop_all will crash.
In debug image, ho_session_alloc will do clib_bitmap_validate to
prevent race condition, however the input is not correct which
will make vpp crash.
Type: fix
Change-Id: Ia8bff325d238eacb671e6764ea2a4eecd3fca609
Signed-off-by: Dongya Zhang <fortitude.zhang@gmail.com>
Since vpp_sswan plugin already merged in /vpp/extras/strongswan,
no need to provide additional vpp_sswan source files for docker image.
Type: fix
Signed-off-by: Yulong Pei <yulong.pei@intel.com>
Change-Id: I35bad22b1046e0dddbcf39e1af38d589d1438239
Signed-off-by: Yulong Pei <yulong.pei@intel.com>
