Age | Commit message (Collapse) | Author | Files | Lines |
|
Change-Id: I2040b560b2a00f8bd176ae6ad46035678a2b249e
Type: improvement
Signed-off-by: Mohammed Hawari <mohammed@hawari.fr>
|
|
Encode entropy value in UDP source port when requested per RFC 7510.
CLI already has "src-port-is-entropy", use zero UDP source port in API
to avoid breaking changes, since zero port is not something to be used
in wild.
Also, mark UDP encapsualtion API as mp-safe as already done for CLI.
Type: feature
Change-Id: Ieb61ee11e058179ed566ff1f251a3391eb169d52
Signed-off-by: Vladislav Grishenko <themiron@yandex-team.ru>
|
|
Type: refactor
Change-Id: I7b7ca9ec62cb70243c5b7e87968eab1338d67ec8
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
default RLIMIT_MEMLOCK is 64. if we use multi af_xdp interfaces or
load complex bpf program, libbpf will return permission error.
root cause is default 64 is not large enough. So we change it before
load bpf program.
Type: fix
Change-Id: Ia6aed19c9256c498cf1155586a54a32b3f444105
Signed-off-by: Chen Yahui <goodluckwillcomesoon@gmail.com>
|
|
Type: fix
Currently, when a wg interface is administratively disabled initially or
during operation, handshake packets continue to be sent. Data packets
stop being sent because routes pointing to the wg interface will not be
used. But data keys remain.
With this fix, when a wg interface is administratively disabled during
peer creation, avoid connection initialization to the peer. Data keys
and timers should be empty at this point. When a wg interface is
disabled during operation, disable all peers (i.e. stop all timers,
clear data keys, etc.). Thus, state should be identical in both cases.
When a wg interface is administratively enabled, enable all peers (i.e.
get ready to exchange data packets and initiate a connection). Also,
cover these scenarios with tests.
Signed-off-by: Alexander Chernavin <achernavin@netgate.com>
Change-Id: Ie9a620077e55d519d21b0abc8c0d3c87b378bca3
|
|
Internaly, vpp uses it's own padding, so all the data
is padded using blocksize in /src/vnet/ipsec/ipsec.c
Openssl should add it's own padding, but the data
is already padded. So on decrypt stage when padding
should be removed, it can't be done. And it produces
error `bad decrypt`
Previous versions of openSSL decrypted data almost
at the beginning of EVP_DecryptUpdate/EVP_DecryptFinal_ex
and produced the same error, but data was already decrypted.
Now it's not, so some algorithms could have some problems
with it
PS. openSSL 3.x.x
Type: fix
Signed-off-by: Vladimir Ratnikov <vratnikov@netgate.com>
Change-Id: If715a80228548b4e588cee222968d9da9024c438
|
|
Type: fix
Signed-off-by: Chen Yahui <goodluckwillcomesoon@gmail.com>
Change-Id: Ia447420f692f1487d343886845d648d766e43c27
Signed-off-by: Chen Yahui <goodluckwillcomesoon@gmail.com>
|
|
Type: fix
As per the protocol:
A handshake initiation is retried after "REKEY_TIMEOUT + jitter" ms,
if a response has not been received...
Currently, if retransmit handshake timer is started, it will trigger
after "REKEY_TIMEOUT + jitter" ms and will try to send a handshake
initiation via wg_send_handshake() given that no responses have been
received. wg_send_handshake() will verify that time stored in
REKEY_TIMEOUT has passed since last handshake initiation sending and if
has, will send a handshake initiation. Time when a handshake initiation
was last sent is stored in last_sent_handshake.
The problem is that last_sent_handshake is not only updated in
wg_send_handshake() when sending handshake initiations but also in
wg_send_handshake_response() when sending handshake responses. When
retransmit handshake timer triggers and a handshake response has been
sent recently, a handshake initiation will not be sent because for
wg_send_handshake() it will look like that time stored in REKEY_TIMEOUT
has not passed yet. Also, the timer will not be restarted.
wg_send_handshake_response() must not update last_sent_handshake,
because this time is used only when sending handshake intitiations. And
the protocol does not say that handshake initiation retransmission and
handshake response sending (i.e. replying to authenticated handshake
initiations) must coordinate.
With this fix, stop updating last_sent_handshake in
wg_send_handshake_response().
Also, this fixes tests that used to wait for "REKEY_TIMEOUT + 1" seconds
and did not receive any handshake initiations. Then they fail.
Also, long-running tests that send wrong packets and do not expect
anything in reply may now receive handshake intiations, consider them as
replies to the wrond packets, and fail. Those are updated to filter out
handshake initiations in such verifications. Moreover, after sending
wrong packets, error counters are already inspected there to confirm
packet processing was unsuccessful.
Signed-off-by: Alexander Chernavin <achernavin@netgate.com>
Change-Id: I43c428c97ce06cb8a79d239453cb5f6d1ed609d6
|
|
Type: improvement
Change-Id: I7f7050c19453a69a7fb6c5e62f8f57db847d9144
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
Type: docs
Change-Id: Icfa2bdc9367f8438b53da7c89caec263ed6ab056
Signed-off-by: Filip Varga <fivarga@cisco.com>
Signed-off-by: Dave Wallace <dwallacelf@gmail.com>
|
|
Type: fix
Change-Id: Ib127331507724f853071e66ca1ddfc773a8ed200
Signed-off-by: Nathan Skrzypczak <nathan.skrzypczak@gmail.com>
|
|
Also check for non-zero rpath length in CLI cmd.
While there, no need to use "else" after a return.
Also while there, notice and fix numerous input_line
buffer leaks and fix them.
Type: fix
Fixes: 669d07dc016757b856e1014a415996cf9f0ebc58
Signed-off-by: Jon Loeliger <jdl@netgate.com>
Change-Id: I18ea44b7b82e8938c3e793e7c2a04dfe157076d8
|
|
Type: fix
If unrecognized input was provided to the commands which add or delete a
pair, the error message was being created incorrectly and only displayed
something like "unknown input `'". Provide the correct argument to
format_unformat_error so that the actual unrecognized input is printed.
There also was no error or useful information printed if only the base
command were provided without any additional arguments. This should
print a warning about what required data was missing. Reorganize code to
handle this and to make sure that memory gets freed appropriately.
Change-Id: If454714f50cf41b3b56cfadfbf017f1d160e13a4
Signed-off-by: Matthew Smith <mgsmith@netgate.com>
|
|
Handle the case of the mapping not being found by GID.
Type: fix
Change-Id: Ibce3b9e8419c0dddca97b4d0d5a71f25dfd529d8
Signed-off-by: Andrew Yourtchenko <ayourtch@gmail.com>
|
|
If one attempts to add a pattern with zero length, first time
it will succeed, and the second time it will cause an invalid memcmp call.
Solution: do not allow to add zero-length patterns.
Type: fix
Signed-off-by: Andrew Yourtchenko <ayourtch@gmail.com>
Change-Id: Ic08e021486153be605a4b12a2fe4422307bf68d2
|
|
The number of available dynamic ports is set to (0xffff - 1024) =
64511, which is not divisable by the pow2 number of workers - the
only integer divisors are 31 and 2081.
So, total dynamic port range of all workers will be less than it:
1 wrk: n = (port_per_thread = 64511/1)*1 = 64511 + 1025 = 65536
2 wrk: n = (port_per_thread = 64511/2)*2 = 64510 + 1025 = 65535
4 wrk: n = (port_per_thread = 64511/4)*4 = 64508 + 1025 = 65533
8 wrk: n = (port_per_thread = 64511/8)*8 = 64504 + 1025 = 65529
...
As seen, with multiple workers there are unused trailing ports for every
nat pool address and that is the reason of out-of-bound index in the
worker array on out2in path due (port - 1024) / port_per_thread math.
This was fixed in 5c9f9968de63fa627b4a72b344df36cdc686d18a, so packets
to unused ports will go to existing worker and dropped there.
Per RFC 6335 https://www.rfc-editor.org/rfc/rfc6335#section-6:
6. Port Number Ranges
o the System Ports, also known as the Well Known Ports, from 0-1023
(assigned by IANA)
o the User Ports, also known as the Registered Ports, from 1024-
49151 (assigned by IANA)
o the Dynamic Ports, also known as the Private or Ephemeral Ports,
from 49152-65535 (never assigned)
According that let's allocate dynamic ports from 1024 and have full port
range with a wide range of the workers number - 64 integer divisors in
total, including pow2 ones:
1 wrk: n = (port_per_thread = 64512/1)*1 = 64512 + 1024 = 65536
2 wrk: n = (port_per_thread = 64512/2)*2 = 64512 + 1024 = 65536
3 wrk: n = (port_per_thread = 64512/3)*3 = 64512 + 1024 = 65536
4 wrk: n = (port_per_thread = 64512/4)*4 = 64512 + 1024 = 65536
5 wrk: n = (port_per_thread = 64512/5)*5 = 64510 + 1024 = 65534
6 wrk: n = (port_per_thread = 64512/6)*6 = 64512 + 1024 = 65536
7 wrk: n = (port_per_thread = 64512/7)*7 = 64512 + 1024 = 65536
8 wrk: n = (port_per_thread = 64512/8)*8 = 64512 + 1024 = 65536
...
Modulo from 5c9f9968de63fa627b4a72b344df36cdc686d18a is still required
when the numbers of workers is not the integer divisor of 64512.
Type: fix
Fixes: 5c9f9968de63fa627b4a72b344df36cdc686d18a
Change-Id: I9edaea07e58ff4888812b0d86cbf41a3784b189e
Signed-off-by: Vladislav Grishenko <themiron@yandex-team.ru>
|
|
Free node frames in worker mains on refork. Otherwise these frames are
never returned to free pool and it causes massive memory leaks if
performed under traffic load
Type: fix
Signed-off-by: Dmitry Valter <d-valter@yandex-team.ru>
Change-Id: I15cbf024a3f4b4082445fd5e5aaa10bfcf77f363
|
|
When a session is found expired, the next node of in2out fast path
should be in2out slow path instead of out2in slow path.
Type: fix
Signed-off-by: Jing Peng <jing@meter.com>
Change-Id: If1dd920502089c25b33bea5434823b0496a44499
|
|
Type: improvement
Roaming functionality allows the peer address to change. The main thread
was being called to update a peer's address if necessary after
processing a received packet. Check in the worker whether this is
necessary before incurring the overhead of the RPC to the main thread.
Signed-off-by: Matthew Smith <mgsmith@netgate.com>
Change-Id: I02184b92dc658e0f57dd39993a3b2f9944187b45
|
|
The ability to modify the vlan setting must be checked prior to using
VIRTCHNL_OP_DISABLE_VLAN_STRIPPING_V2 both for inner and outer vlan
stripping
Change-Id: Iffe306c34b81a6077ad6ba5deb3f5b61b5475897
Type: fix
Signed-off-by: Mohammed Hawari <mohammed@hawari.fr>
|
|
Originally initialization cryptodev device(s) calls double request
to enabled async mode and increased ref count twice for async mode.
Due to this cannot be change any assigned async handlers to other
async crypto engine.
The fixes reduce double request to enable async mode in initialization
cryptodev device(s) and VPP can be change assigned async handlers
to other crypto engine after disabled all async feature, for example:
ipsec, wireguard.
Type: fix
Signed-off-by: Gabriel Oginski <gabrielx.oginski@intel.com>
Change-Id: If22e682c3c10de781d05c2e09b5420f75be151c3
|
|
test output before fix:
DBGvpp# vrrp proto start sw_if_index 1 vr_id 1
vrrp proto: unknown input `sw_if_index 1 vr_id 1'
DBGvpp# vrrp vr track-if add sw_if_index 1 vr_id 1 track-index 1
priority 30
vrrp vr track-if: Please specify an interface
Type: fix
Signed-off-by: luoyaozu <luoyaozu@foxmail.com>
Change-Id: Ib8ba67e920b23008d9246318ec8f8f17bf0bea95
|
|
Zero-initialize a variable.
Type: fix
Signed-off-by: Andrew Yourtchenko <ayourtch@gmail.com>
Change-Id: Iccf2eb4bf26755d6cd93fc70df3c5481d69ce7eb
|
|
Zero-initialize the variable
Type: fix
Signed-off-by: Andrew Yourtchenko <ayourtch@gmail.com>
Change-Id: I4ee127ac3e2a3beffa11bbc96db1f3254b3f7c5d
|
|
Initialize the session index in case of error to ~0,
so is defined in case trace needs to copy it.
Type: fix
Signed-off-by: Andrew Yourtchenko <ayourtch@gmail.com>
Change-Id: Iddf6df42c09d2abc11e5821944eb4f41692e6e3e
|
|
The unformat type for "%d" should be u32 or int, otherwise the 'did' in
high stack address will be overflow to zero by the 'qid' which is in the
low stack address.
Like input "dev wq3.2" will return "did=0, qid=2".
Type: fix
Signed-off-by: Haiyue Wang <haiyue.wang@intel.com>
Change-Id: I0fe1d5b03e2c47e0a7925193e2c2f1ccc31d3e90
|
|
When there is a bad descriptor, it may in the beginning, in the middle,
or at the end of the batch if the batch has more than 3 descriptors.
When processing a bad descriptor is encountered in the batch, we need to
rollback n_buffers in memif_process_desc(), or the statement in the same
function
memif_add_copy_op (ptd, mb0 + src_off, bytes_to_copy,
dst_off, n_buffers - 1);
is wrong because it picks up the wrong buffer_vec_index of the bad
descriptor while parsing a good descriptor immediately following the
bad descriptor. n_buffers was incremented in the beginning of
while (n_left) loop.
The other problem is we should count the number of bad packets and
reduce ptd->n_packets to get the correct number of packets for subsequent
processing in device_input.
The last fix is to check if n_buffers == 0 in device_input and skip
doing any descriptor copy. This case can happen when all the descriptors
are bad in the batch.
Type: fix
Signed-off-by: Steven Luong <sluong@cisco.com>
Change-Id: I28ed1d87236b045657006755747b5750a9a733be
|
|
In case of a bad packet, the bihash kv is not initialized before being
copied in the trace. Make sure it is initialized to 0.
Type: fix
Change-Id: I22fcfe99f3586d0fa128493059547a56557b8fb5
Signed-off-by: Benoît Ganne <bganne@cisco.com>
|
|
Initialize the host_sw_if_index to ~0 so in the error cases
the variable is set to something predictable.
Type: fix
Change-Id: Ic55e4f0cbfa286e85dfb54b89b5321af18a439a1
Signed-off-by: Andrew Yourtchenko <ayourtch@gmail.com>
|
|
Zero-initialize the temporary struct, else a->map.adj_index is being used unset.
Type: fix
Signed-off-by: Andrew Yourtchenko <ayourtch@gmail.com>
Change-Id: Ia02636ea1e911250d6aa5e413de48e1e09863880
|
|
This patch introduces DMA infrastructure into vlib. This is well known
that large amount of memory movements will drain core resource. Nowadays
more and more hardware accelerators were designed out for freeing core
from this burden. Meanwhile some restrictions still remained when
utilizing hardware accelerators, e.g. cross numa throughput will have a
significant drop compared to same node. Normally the number of hardware
accelerator instances will less than cores number, not to mention that
applications number will even beyond the number of cores. Some hardware
may support share virtual address with cores, while others are not.
Here we introduce new DMA infrastructure which can fulfill the
requirements of vpp applications like session and memif and in the
meantime dealing with hardware limitations.
Here is some design backgrounds:
Backend is the abstract of resource which allocated from DMA device
and can do some basic operations like configuration, DMA copy and
result query.
Config is the abstract of application DMA requirement. Application
need to request an unique config index from DMA infrastructure. This
unique config index is associated with backend resource. Two options
cpu fallback and barrier before last can be specified in config.
DMA transfer will be performed by CPU when backend is busy if cpu
fallback option is enabled. DMA transfer callback will be in order
if barrier before last option is enabled.
We constructs all the stuffs that DMA transfer request needed into
DMA batch. It contains the pattern of DMA descriptors and function
pointers for submission and callback. One DMA transfer request need
multiple times batch update and one time batch submission.
DMA backends will assigned to config's workers threads equally. Lock
will be used for thread-safety if same backends assigned to multiple
threads. Backend node will check all the pending requests in worker
thread and do callback with the pointer of DMA batch if transfer
completed. Application can utilize cookie in DMA batch for selves
usage.
DMA architecture:
+----------+ +----------+ +----------+ +----------+
| Config1 | | Config2 | | Config1 | | Config2 |
+----------+ +----------+ +----------+ +----------+
|| || || ||
+-------------------------+ +-------------------------+
| DMA polling thread A | | DMA polling thread B |
+-------------------------+ +-------------------------+
|| ||
+----------+ +----------+
| Backend1 | | Backend2 |
+----------+ +----------+
Type: feature
Signed-off-by: Marvin Liu <yong.liu@intel.com>
Change-Id: I1725e0c26687985aac29618c9abe4f5e0de08ebf
|
|
The ipv6 header length should not be counted in the ipv6 payload length.
This is similar to https://gerrit.fd.io/r/c/vpp/+/36945.
Type: fix
Change-Id: I22de0ff828175829102a85288513ee3f55709108
Signed-off-by: Aloys Augustin <aloaugus@cisco.com>
|
|
Type: improvement
VPP crashes when a linux-cp tap is added to a bridge on the host system
because rtnl_neigh_get_dst() returns NULL for the neighbor message that
is sent by the kernel.
Check for NULL before trying to use the address from a neighbor in a
netlink message.
Signed-off-by: Matthew Smith <mgsmith@netgate.com>
Change-Id: I8a683d815a09620df9c0cc76e18df39828428e2c
Signed-off-by: Matthew Smith <mgsmith@netgate.com>
|
|
Add the error checks in parsing, aimed to avoid parser walking past the end of packet in case the data
is garbage.
Type: fix
Signed-off-by: Andrew Yourtchenko <ayourtch@gmail.com>
Change-Id: I9541b555a18baf63cb8081bcd7a4c2750f2ed012
|
|
flags is u64, makes sure we do not overflow when shifting.
Type: fix
Change-Id: Ieea34187c0b568dc4d24c9415b9cff36907a5a87
Signed-off-by: Benoît Ganne <bganne@cisco.com>
|
|
rather than using obfuscated macro hacery, simplify
the per-protocol data management by directly using
an array of NAT protocol types.
Type: refactor
Signed-off-by: Jon Loeliger <jdl@netgate.com>
Change-Id: I6fe987556ac9f402f8d490da0740e2b91440304c
|
|
Type: improvement
If an SA protecting an IPv6 tunnel interface has UDP encapsulation
enabled, the code in esp_encrypt_inline() inserts a UDP header but does
not set the next protocol or the UDP payload length, so the peer that
receives the packet drops it. Set the next protocol field and the UDP
payload length correctly.
The port(s) for UDP encapsulation of IPsec was not registered for IPv6.
Add this registration for IPv6 SAs when UDP encapsulation is enabled.
Add punt handling for IPv6 IKE on NAT-T port.
Add registration of linux-cp for the new punt reason.
Add unit tests of IPv6 ESP w/ UDP encapsulation on tun protect
Signed-off-by: Matthew Smith <mgsmith@netgate.com>
Change-Id: Ibb28e423ab8c7bcea2c1964782a788a0f4da5268
|
|
Free up the vapi context in case of failure.
Type: fix
Signed-off-by: Andrew Yourtchenko <ayourtch@gmail.com>
Change-Id: I4f64e8718014d714f1b82877e69d2354b5fa44fb
|
|
Crypto backend errors should not be using the same error as missing
keypair.
Type: fix
Change-Id: I78c2b3df3f08a354463b7824349b08627f2b023c
Signed-off-by: Benoît Ganne <bganne@cisco.com>
|
|
IPv6 payload length should not include the size of the IPv6 header.
Type: fix
Change-Id: Iedcd17d0af8d72d9b5f8f9b605da7c99e151bc9d
Signed-off-by: Benoît Ganne <bganne@cisco.com>
|
|
Previously, each address maintained an array of 32-bit
reference counts for each of 65K possible ports for each
of 4 NAT protocols. Totalling 1MB per address. Wow.
A close read of the code shows that an "is used" check
precedes each attempted reference count increment.
That means the refcount never actually gets above 1.
That in turn means algorithmically, a bit vector is
sufficient. And one need not be allocated for more
than the highest validated port referenced.
These changes introduce a dynamically sized bit vector
replacing the reference counts, for a maximum of 32K
if all 4 protocols use port 65535. In fact, protocol
OTHER is never used, so at most 24K will be used, and
none of it will be "statically" allocated per address.
Type: fix
Fixes: 85bee7548bc5a360851d92807dae6d4159b68314
Change-Id: I7fd70050e7bf4871692a862231f8f38cf0158132
Signed-off-by: Jon Loeliger <jdl@netgate.com>
|
|
Zero-initialize the temporary struct on stack.
Type: fix
Signed-off-by: Andrew Yourtchenko <ayourtch@gmail.com>
Change-Id: I89ced4cca8e832827fe054e2e60986de5910360c
|
|
Zero-initialize the temporary struct on stack.
Type: fix
Change-Id: I651f87deeb79c6c073d5c510435fa268893a3b0e
Signed-off-by: Andrew Yourtchenko <ayourtch@gmail.com>
|
|
In RFC 7296, CREATE_CHILD_SA Exchange may contain the KE payload
to enable stronger guarantees of forward secrecy.
When the KEi payload is included in the CREATE_CHILD_SA request,
responder should reply with the KEr payload and complete the key
exchange, in accordance with the RFC.
Type: improvement
Signed-off-by: Atzm Watanabe <atzmism@gmail.com>
Change-Id: I13cf6cf24359c11c3366757e585195bb7e999638
|
|
Type: fix
Signed-off-by: Atzm Watanabe <atzmism@gmail.com>
Change-Id: Icbd452b43ecaafe46def1276c98f7e8cbf761e51
|
|
We validate each descriptor via memif_validate_desc_data and set
desc_status to non-zero for the corresponding descriptor when
the descriptor is bad. However, desc_status is not propagated back to
xor_status in memif_validate_desc_data which eventually sets
ptd->xor_status.
Not setting ptd->xor_status causes us to treat all descriptors as
"simple". In that case, when we try to copy also the bad descriptors to
the buffers, it results a crash since desc_data is not set to point
to the correct memory in the descriptor.
The fix is to set xor_status in memif_validate_desc_data such that if
there is a bad descriptor in the frame, "is_simple" is set to false and
we have to selectively copy only the good descriptors to the buffers.
Type: fix
Signed-off-by: Steven Luong <sluong@cisco.com>
Change-Id: I780f51a42aa0f8745edcddebbe02b2961c183598
|
|
Type: fix
After peers roaming support addition, FIB entry tracking stopped
working. For example, it can be observed when an adjacency is stacked on
a FIB entry by the plugin and the FIB entry hasn't got ARP resolution
yet. Once the FIB entry gets ARP resolution, the adjacency is not
re-stacked as it used to. This results in endless ARP requests when a
traffic is sent via the adjacency.
This is broken because the plugin stopped using "midchain delegate" with
peers roaming support addition. The reason is that "midchain delegate"
didn't support stacking on a different FIB entry which is needed when
peer's endpoint changes. Now it is supported there (added in 36892).
With this fix, start using "midchane delegate" again and thus, fix FIB
entry tracking. Also, cover this in tests.
Signed-off-by: Alexander Chernavin <achernavin@netgate.com>
Change-Id: Iea91f38739ab129e601fd6567b52565dbd649371
|
|
In several NAT submodules, the number of available ports (0xffff - 1024)
may not be divisible by the number of workers, so port_per_thread is
determined by integer division, which is the floor of the quotient.
Later when a worker index is needed, dividing the port with port_per_thread
may yield an out-of-bound array index into the workers array.
As an example, assume 2 workers are configured, then port_per_thread
will be (0xffff - 1024) / 2, which is 32255. When we compute a worker
index with port 0xffff, we get (0xffff - 1024) / 32255, which is 2,
but since we only have 2 workers, only 0 and 1 are valid indices.
This patch fixes the problem by adding a modulo at the end of the division.
Type: fix
Signed-off-by: Jing Peng <pj.hades@gmail.com>
Change-Id: Ieae3d5faf716410422610484a68222f1c957f3f8
|
|
we need cancel vrrp_vr_timer when deleting vrrp vr
Type: fix
Signed-off-by: luoyaozu <luoyaozu@chinatelecom.cn>
Change-Id: I8ea01f1943d6e3e60c4990c5be945de613bc8b53
|
|
Type: fix
Signed-off-by: Florin Coras <fcoras@cisco.com>
Change-Id: I18b9d0d67f5fe4c1714427259df29026153d8dd1
|