Age | Commit message (Collapse) | Author | Files | Lines |
|
Type: refactor
Change-Id: I5235bf3e9aff58af6ba2c14e8c6529c4fc9ec86c
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
This change aims to affect crypto_sw_scheduler behavior,
but all the edits end up in vnet/crypto.
After 9a9604b introduced adaptive mode for crypto dispatch,
the performance of async mode at lower rate got worse.
A work around for CSIT test is done by changing dispatch mode via explicit API call
in https://github.com/FDio/vpp/commit/139aba204780f6cc2845b311820a0b4c47517d02
In this change, the CLI is brought back to allow user to fix the
dispatch mode.
set crypto async dispatch mode <polling|interrupt|adaptive>
Type: improvement
Change-Id: I029e98aa25889eddcf62e75a6c78926cdee862ef
Signed-off-by: Niyaz Murshed <niyaz.murshed@arm.com>
|
|
This is a prerequisite patch for the following openssl API optimization
patch, which tries to offload openssl ctx init and key expansion work to
the initialization stage.
Wireguard adds crypto keys via vnet_crypto_key_add (), and whenever it
modifies the keys, the underneath openssl crypto engine shoud be informed
of the changes to update the openssl ctx.
Type: feature
Signed-off-by: Lijian Zhang <Lijian.Zhang@arm.com>
Change-Id: I3e8f033f3f77eebcecfbd06e8e3bbbfdc95a50e2
|
|
Check if crypto ops vector is matching actual ops instead if blindly
dereferencing it.
Type: fix
Signed-off-by: Dmitry Valter <d-valter@yandex-team.com>
Change-Id: Ib88ab44137d9360ee96228e72349a62b2fa7a7e0
|
|
This change aims to affect crypto_sw_scheduler behavior,
but all the edits end up in vnet/crypto.
Previous release CSIT tests were testing async crypto in polling mode.
After 9a9604b09f15691d7c4ddf29afd99a31e7e31eed introduced adaptive mode
for crypto dispatch, the CSIT performance got way worse.
Possibly, there is another VPP bug related to adaptive mode
(it should not lose as many packets as seen in CSIT),
but the next release is too close for trying to fix that.
This change (instead of fixing adaptive mode)
allows CSIT to continue testing polling mode (after explicit API call),
while keeping the adaptive mode as default behavior.
The deprecated crypto_set_async_dispatch always disable adaptive mode,
crypto_set_async_dispatch_v2 has parameter to enable or disable it.
The mode parameter is still used for the inital state of adaptive mode.
Type: feature
Change-Id: Ib98080eefb4be291207af543884f2c3837f92f59
Signed-off-by: Vratko Polak <vrpolak@cisco.com>
|
|
Type: improvement
Change-Id: I830f7a2ea3ac0aff5185698b9fa7a278c45116b0
Signed-off-by: Benoît Ganne <bganne@cisco.com>
|
|
The async frames pool may be resized once drained. This will cause 2 problems: original pool pointer is invalidated and pool size changed, both problems will confuse the crypto infra user graph nodes (like IPsec and Wireguard) and crypto engines if they expect the pool pointers always valid and the pool size never changed (for performance reason).
This patch introduces fixed size of the async frames pool. This helps zeroing surprise to the components shown above and avoiding segmentation fault when pool resizing happened. In addition, the crypto engine may take advantage of the feature to sync its own pool/vector with crypto infra.
Type: improvement
Signed-off-by: Gabriel Oginski <gabrielx.oginski@intel.com>
Signed-off-by: Piotr Bronowski <piotrx.bronowski@intel.com>
Change-Id: I2a71783b90149fa376848b9c4f84ce8c6c034bef
|
|
This patch can make crypto dispatch node adaptively switching
between pooling and interrupt mode, and improve vpp overall
performance.
Type: improvement
Signed-off-by: Xiaoming Jiang <jiangxiaoming@outlook.com>
Change-Id: I845ed1d29ba9f3c507ea95a337f6dca7f8d6e24e
|
|
IV requirements vary wildly with the selected mode of operation. For
example, for AES-CBC the IV must be unpredictable whereas for AES
counter mode (CTR or GCM), it can be predictable but reusing an IV with
the same key material is catastrophic.
Because of that, it is hard to generate IV in a generic way, and it is
better left to the crypto user (eg. IPsec).
Type: improvement
Change-Id: I32689c591d8c6572b8d37c4d24f175ea6132d3ec
Signed-off-by: Benoît Ganne <bganne@cisco.com>
|
|
Type: improvement
Change-Id: Ib98eba146e24e659acf3b9a228b81fcd641f4c67
Signed-off-by: Benoît Ganne <bganne@cisco.com>
|
|
For AES-CBC, the IV must be unpredictable (see NIST SP800-38a Appendix
C). Chaining IVs like is done by ipsecmb and native backends for the
VNET_CRYPTO_OP_FLAG_INIT_IV is fully predictable.
Encrypt a counter as part of the message, making the (predictable)
counter-generated IV unpredictable.
Fixes: VPP-2037
Type: fix
Change-Id: If4f192d62bf97dda553e7573331c75efa11822ae
Signed-off-by: Benoît Ganne <bganne@cisco.com>
|
|
Pass vec_foreach args in the right order
Type: fix
Signed-off-by: Dmitry Valter <d-valter@yandex-team.ru>
Change-Id: I90d3dfb70ee2dbd9a48ae9ecfd2c1526a74d0786
|
|
If async engines are disbaled and async is turned on vpp tries to
enqueue frame with nonexisting handler which leads to segfault.
This patch checks for handler and drops the frame in case it doesn't
exist.
Type: fix
Signed-off-by: Dastin Wilski <dastin.wilski@gmail.com>
Change-Id: I67211867ee29dc41cc9f0733e8e0b3ea86677f85
|
|
Type: refactor
Change-Id: I3625eacf9e04542ca8778df5d46075a8654642c7
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
Fix for the crash when both crypto_native_plugin and DPDK QAT
are enabled in startup conf.
Type: fix
Signed-off-by: mgovind <govindarajan.mohandoss@arm.com>
Change-Id: Ib020ed7130a99080a093c70c06d47bcacd6d23b1
|
|
Originally wireguard doesn't support async mode for encryption packets.
This patch add async mode for encryption in wireguard and also adds
support chacha20-poly1305 algorithm in cryptodev for async handler.
In addition it contains new command line to activate async mode for wireguard:
set wireguard async mode on|off
and also add new command to check active mode for wireguard:
show wireguard mode
Type: improvement
Signed-off-by: Gabriel Oginski <gabrielx.oginski@intel.com>
Change-Id: I141d48b42ee8dbff0112b8542ab5205268089da6
|
|
This fix adds check that will omit loop iteration
in case dequeue handler is zero.
Type: fix
Signed-off-by: Dastin Wilski <dastin.wilski@gmail.com>
Change-Id: I7526e3fe7d8c8da9662b4e9204efd5e2d8be1908
|
|
Type: improvement
Previously multiple sw crypto scheduler queues per core design
caused unaverage frame processing rate for each async op ID –
the lower the op ID is the highly likely they are processed first.
For example, when a RX core is feeding both encryption and
decryption jobs of the same crypto algorithm to the queues at a
high rate, in the mean time the crypto cores have no enough
cycles to process all: the jobs in the decryption queue are less
likely being processed, causing packet drop.
To improve the situation this patch makes every core only owning
a two queues, one for encrypt operations and one for decrypt.
The queue is changed either after checking each core
or after founding a frame to process.
All crypto jobs with different algorithm are pushed to
thoses queues and are treated evenly.
In addition, the crypto async infra now uses unified dequeue handler,
one per engine. Only the active engine will be registered its
dequeue handler in crypto main.
Signed-off-by: DariuszX Kazimierski <dariuszx.kazimierski@intel.com>
Signed-off-by: PiotrX Kleski <piotrx.kleski@intel.com>
Signed-off-by: Fan Zhang <roy.fan.zhang@intel.com>
Signed-off-by: Jakub Wysocki <jakubx.wysocki@intel.com>
Change-Id: I517ee8e31633980de5e0dd4b05e1d5db5dea760e
|
|
Originally the pool of keys can be expand and cache with pointer for
key can be invalid.
For example in Wireguard during handshake process this pool can be
expand and pointer for these keys in cache can be invalid for workers
or can has incorrect value (poison memory).
The fixes add barrier if the pool needs be to expand to ensure that
cache in function will be valid and avoid situation when cache has
invalid pointer for these keys.
Type: fix
Signed-off-by: Gabriel Oginski <gabrielx.oginski@intel.com>
Change-Id: Ida8f300213dfebb91ecaf1937fb08de81c20ba7b
|
|
vlib_buffer_enqueue_to_next() requires to allow overflow of up to 63
elements of 'buffer' and 'nexts' array.
- add helper to compute the minimum size
- fix occurences in session and async crypto
Type: fix
Change-Id: If8d7eebc5bf9beba71ba194aec0f79b8eb6d5843
Signed-off-by: Benoît Ganne <bganne@cisco.com>
|
|
- Generate copyright year and version
instead of using hard-coded data
Type: refactor
Signed-off-by: Dave Wallace <dwallacelf@gmail.com>
Change-Id: I6058f5025323b3aa483f5df4a2c4371e27b5914e
|
|
Type: feature
Change-Id: I36041fe5c5f0ff129aee42516189807e96f62123
Signed-off-by: Filip Tehlar <ftehlar@cisco.com>
|
|
With this change, add support for 3DES and MD5 in IPsec async mode.
After changes in foreach_crypto_link_async_alg, the last combination in
the list (aes-256-cbc-hmac-sha-512) started to fail during decription.
That was also fixed by proper vector size validation.
Type: improvement
Signed-off-by: Alexander Chernavin <achernavin@netgate.com>
Change-Id: I660657bdab62ea9cf031c3e43d99f2317e5f74d7
|
|
Type: improvement
This patch combined cryptodev op and cryptodev raw API data
paths into one and makes the engine run-timely select which
data path is used: if all cryptodev devices support
RTE_CRYPTODEV_FF_SYM_RAW_DP feature flag, the raw data path
API is used, otherwise the traditional data path is used.
Signed-off-by: Fan Zhang <roy.fan.zhang@intel.com>
Change-Id: Ibbd7c4405288bd9a48a34878954fd3040df7b4ad
|
|
Type: improvement
Change-Id: If3da7d4338470912f37ff1794620418d928fb77f
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
Type: improvement
In the current scheme an async frame is submitted each time the crypto
op changes. thus happens each time a different SA is used and thus
potentially many times per-node. thi can lead to the submision of many
partially filled frames.
change the scheme to construct as many full frames as possible in the
node and submit them all at the end. the frame owner ship is passed to
the user so that there can be more than one open frame per-op at any
given time.
Signed-off-by: Neale Ranns <neale@graphiant.com>
Change-Id: Ic2305581d7b5aa26133f52115e0cd28ba956ed55
|
|
Type: feature
Change-Id: I9d4f90bc701d2b9b903a018f8d27cec5e129d7be
Signed-off-by: Benoît Ganne <bganne@cisco.com>
|
|
Fixes coverity issue CID 218445 (#1 of 1): Logically dead code
(DEADCODE) dead_error_line: Execution cannot reach this statement:
return 4294967295U;.
Type: fix
Signed-off-by: Piotr Bronowski <piotrx.bronowski@intel.com>
Change-Id: Ibf8ee0458320d20c3adca2efa2a4bfad7c190dbe
|
|
Type: improvement
Also:
- state as enum so my GDB life is easier
- typo; s/indice/indices/;
Signed-off-by: Neale Ranns <neale@graphiant.com>
Change-Id: I3320f5ef1ccd7d042071ef336488a41adfad7463
|
|
Type: fix
vnet_crypto_is_set_handler() was checking if op handler is set by
using alg instead of op id.
This fix changes the check to use op ids from crypto_main's alg_data
for provided alg.
Signed-off-by: PiotrX Kleski <piotrx.kleski@intel.com>
Change-Id: I5f0f00a889ddada44d9912af175ebad66677037a
|
|
Type: refactor
Change-Id: I077110e1a422722e20aa546a6f3224c06ab0cde5
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
Type: fix
Changed vnet_crypto_async_reset_frame assert to expect also
ERROR state frames.
Signed-off-by: PiotrX Kleski <piotrx.kleski@intel.com>
Change-Id: I3abc29f3f9642027aee38a59a932e54c90da859d
|
|
Type: fix
To avoid race condition happening in async crypto engines,
async frame state and thread index set should happen before enqueue.
In addition as the enqueue handler already returns the enqueue status,
when an enqueue is failed, the async crypto engine shall not worry
about setting the async frame state but let the submit_open_frame function
to do just that.
Signed-off-by: PiotrX Kleski <piotrx.kleski@intel.com>
Reviewed-by: Fan Zhang <roy.fan.zhang@intel.com>
Change-Id: Ic1b0c94478b3cfd5fab98657218bbd70c46a220a
|
|
- Refactor make test code to be co-located with
the vpp feature source code
Type: test
Signed-off-by: Dave Wallace <dwallacelf@gmail.com>
Change-Id: I089bda44c31cbb217132e5b385cd9ea96ea5239e
|
|
This reverts commit 8c91b2ae2b32d428ef35605707788fe064621cb3,
but keep a comment fix.
Type: fix
Signed-off-by: Yulong Pei <yulong.pei@intel.com>
Change-Id: Ia66941bf18d3efac96f41bdf905d877cfb3ab211
|
|
Type: improvement
Signed-off-by: Yulong Pei <yulong.pei@intel.com>
Change-Id: I43556f8c76c7aae64d9c927e1fda3c1774d7e49d
|
|
Type: feature
This patch updateds cryptodev engine uses new DPDK Cryptodev
API planned to be upstreamed in DPDK 20.11.
Signed-off-by: Fan Zhang <roy.fan.zhang@intel.com>
Signed-off-by: Piotr Bronowski <piotrX.bronowski@intel.com>
Change-Id: I8dd1a8ac643f1e952deb787e466b76ea7aa5f420
|
|
Type: feature
This adds api calls for the following CLIs:
* set sw_scheuduler worker <N> crypto on|off
* set crypto async dispatch polling|interrupt
* set crypto handler
* set crypto async handler
Change-Id: Ic701d149c440e42ea4575da42b9f69e4c8759602
Signed-off-by: Nathan Skrzypczak <nathan.skrzypczak@gmail.com>
|
|
Type: feature
This patch adds new sw_scheduler async crypto engine.
The engine transforms async frames info sync crypto ops and
delegates them to active sync engines. With the patch it
is possible to increase the single worker crypto throughput
by offloading the crypto workload to multiple workers.
By default all workers in the system will attend the crypto
workload processing. However a worker's available cycles
are limited. To avail more cycles to one worker to process
other workload (e.g. the worker core that handles the RX/TX
and IPSec stack processing), a useful cli command is added
to remove itself (or add it back later) from the heavy
crypto workload but only let other workers to process the
crypto. The command is:
- set sw_scheduler worker <idx> crypto <on|off>
It also adds new interrupt mode to async crypto dispatch node.
This mode signals the node when new frames are enqueued
as opposed to polling mode that continuously calls dispatch node.
New cli commands:
- set crypto async dispatch [polling|interrupt]
- show crypto async status (displays mode and nodes' states)
Signed-off-by: PiotrX Kleski <piotrx.kleski@intel.com>
Signed-off-by: DariuszX Kazimierski <dariuszx.kazimierski@intel.com>
Reviewed-by: Fan Zhang <roy.fan.zhang@intel.com>
Change-Id: I332655f347bb9e3bc9c64166e86e393e911bdb39
|
|
Type: feature
Signed-off-by: Artem Glazychev <artem.glazychev@xored.com>
Change-Id: I3697cf7fab7abb7c3d2f61ef326c9116bc1eed66
|
|
Fix coverity issues in crypto framework and cryptodev
engine.
Type: fix
Signed-off-by: Fan Zhang <roy.fan.zhang@intel.com>
Change-Id: Ib261da0163c8182c803600db22c5a6dad5a19999
|
|
Type: feature
Signed-off-by: Damjan Marion <damarion@cisco.com>
Signed-off-by: Filip Tehlar <ftehlar@cisco.com>
Signed-off-by: Fan Zhang <roy.fan.zhang@intel.com>
Signed-off-by: Piotr Bronowski <piotrx.bronowski@intel.com>
Signed-off-by: Dariusz Kazimierski <dariuszx.kazimierski@intel.com>
Signed-off-by: Piotr Kleski <piotrx.kleski@intel.com>
Change-Id: I4c3fcccf55c36842b7b48aed260fef2802b5c54b
|
|
Type: fix
Change-Id: I9e168a87aff1882636840eb94e5509cd08885fb5
Signed-off-by: Filip Tehlar <ftehlar@cisco.com>
|
|
Type: improvement
Show simple and chained handler details on a single line.
Change-Id: I5ad807a4d3608fc38c1134f564755d5751c30070
Signed-off-by: Filip Tehlar <ftehlar@cisco.com>
|
|
Type: feature
Change-Id: Ie072a7c2bbb1e4a77f7001754f01897efd30fc53
Signed-off-by: Filip Tehlar <ftehlar@cisco.com>
|
|
Do not overflow ops_handlers vector.
Type: fix
Change-Id: I8d5e7fb8125a7bd87ecfe6f4f1390fb9f43dad8f
Signed-off-by: Benoît Ganne <bganne@cisco.com>
|
|
opt_data is defined as a array, while in some code, e.g., function
vnet_crypto_get_op_type, it's used as vec.
vec api is not applicable to static arraies.
src/vnet/crypto/crypto.h:234:70: error: address of array 'cm->opt_data' will always evaluate to 'true' [-Werror,-Wpointer-bool-conversion]
vnet_crypto_op_data_t *od = ({ do { if ((0 > 0) && ! ((id) < ((cm->opt_data) ? (((vec_header_t *) (cm->opt_data) - 1)->len) : 0)))
Type: fix
Change-Id: I0b6754406e4216ca975bc1da4b5d4ce293a9bb45
Signed-off-by: Lijian Zhang <Lijian.Zhang@arm.com>
|
|
Type: fix
Several Fixes:
1 - Anti-replay did not work with GCM becuase it overwrote the sequence
number in the ESP header. To fix i added the seq num to the per-packet
data so it is preserved
2 - The high sequence number was not byte swapped during ESP encrypt.
3 - openssl engine was the only one to return FAIL_DECRYPT for bad GCM
the others return BAD_HMAC. removed the former
4 - improved tracing to show the low and high seq numbers
5 - documented the anti-replay window checks
6 - fixed scapy patch for ESN support for GCM
7 - tests for anti-reply (w/ and w/o ESN) for each crypto algo
Change-Id: Id65d96b6d1d4dd821b2ab557e87468fff6d70e5b
Signed-off-by: Neale Ranns <nranns@cisco.com>
|
|
requested alogrithm.
Type: feature
Change-Id: I19a9c14b2bb52ba2fc66246845b7ada73d5095d1
Signed-off-by: Neale Ranns <nranns@cisco.com>
|
|
there's no use case to just change the key of an SA. instead the SA
should be renegociated and the new SA applied to the existing SPD entry
or tunnel.
the set_key functions were untested.
Type: refactor
Change-Id: Ib096eebaafb20be7b5501ece5a24aea038373002
Signed-off-by: Neale Ranns <nranns@cisco.com>
|