Age | Commit message (Collapse) | Author | Files | Lines |
|
Change-Id: I7a9a436d00d04e68c36284de0897e62b838f8506
Signed-off-by: Szymon Sliwa <szs@semihalf.com>
|
|
Change-Id: I89cdac40c9f768f42443406cebf7257c3e139b47
Signed-off-by: Szymon Sliwa <szs@semihalf.com>
|
|
1) renamed ipsec_api to enable_odp_ipsec, for consistency
2) added crude ASSERTs to make sure the configuration is sane
3) removed passing ipsec_api as argument, as this is a global flag now
4) reorder code - put ipsec initialization before putting
workers in polling mode
5) remove not used recycle vector from esp_encrypt.c/esp_decrypt.c
6) change clib_error_return to clib_error in ipsec.c, as
clib_error_return silently passes when 0 is the first argument
Change-Id: Id2c6da985e872f12c8409918fb9cc6113b486d10
Signed-off-by: Szymon Sliwa <szs@semihalf.com>
|
|
Change-Id: I1f71fae3af5e36b710c8c978d8b9b74ea802e012
Signed-off-by: Szymon Sliwa <szs@semihalf.com>
|
|
Change-Id: Ib32d9df19bb9b0db1b3a0d30d991f1a1f11d66c7
Signed-off-by: Szymon Sliwa <szs@semihalf.com>
|
|
To use inline mode put inline flag in the odp section
of the startup.conf file, like this:
odp {
enable-odp-ipsec
inline
}
Falls back to lookaside mode.
Change-Id: I1292a7254b25a15b25285773a43bae112394827d
Signed-off-by: Szymon Sliwa <szs@semihalf.com>
|
|
To use asynchronous mode put async in the odp section
of the startup.conf file, like this:
odp {
enable-odp-ipsec
async
}
Falls back to synchronous mode.
Change-Id: Iacfdc03158228f4385511e6736fb24a97284320c
Signed-off-by: Szymon Sliwa <szs@semihalf.com>
|
|
Created two new nodes for encryption/decryption,
based on user preferences we can use them to accellerate
IPsec by offloading it to hardware or optimized software
implementations
To use the new nodes two flags are needed in the odp
section of the startup.conf file, like this:
To use the new nodes put the enable-odp-ipsec flag to
startup.conf file, like this:
odp {
enable-odp-ipsec
}
Change-Id: Ib4bbc481efad7e90d63994580b57849b74400947
Signed-off-by: Szymon Sliwa <szs@semihalf.com>
|
|
1) added casts on assignment to post_crypto.next_index
as the type of the field changed
2) removed checking result when crypto operation is
asynchronous as it make no sense
3) fixed bug which could cause issues when some of the
operations are done synchronously and some
asynchronously
Change-Id: I123d70402c0210b11dfc2d0fff2c72aa100ce838
Signed-off-by: Szymon Sliwa <szs@semihalf.com>
|
|
Change ethertype from double u8 to u16 for consistency, and next_index to u8,
for consistency with next_index defined int the ipsec field of this union.
Change-Id: I9315bbdd95e23290b37da53267b94a5ddacc85d7
Signed-off-by: Szymon Sliwa <szs@semihalf.com>
|
|
Add a check if crypto operation was successful
Change-Id: Iaa7c4df89aab309a4c85300240d245b4cdb30533
Signed-off-by: Szymon Sliwa <szs@semihalf.com>
|
|
1) add footer to check style
2) fix style
3) change the TODO note to a more descriptive one
Change-Id: I232c937c13bc0d5f206ebb4426b50bebc512a192
Signed-off-by: Szymon Sliwa <szs@semihalf.com>
|
|
In case of trying to add GRE to the IPsec implementation, be careful as the
post_crypto and ipsec fields of the vnet_buffer union overlap, and both are
needed for the ODP based ipsec with GRE.
Change-Id: If20fa9d78f2879264c02922f21bc204fba1ab616
Signed-off-by: Szymon Sliwa <szs@semihalf.com>
|
|
Change-Id: Ib0c3e2ceebb3b3929f3f0b96b8a40d07c896d70f
Signed-off-by: Szymon Sliwa <szs@semihalf.com>
|
|
Change-Id: I229128ccd88c1cf608e20507ca26752b9c0495f5
Signed-off-by: Szymon Sliwa <szs@semihalf.com>
|
|
By default ipsec picks asynchronuous crypto. After the operation it may
turn out that the operation was performed synchronously anyways, in such
case the packet is send further by the esp_* node because there will be
no notification event sent about the crypto completion.
To use asynchronous mode put async in the odp section
of the startup.conf file,
like this:
odp {
async
}
Falls back to synchronous mode.
Change-Id: I5301df5f1c93a5ccd53a9c0ed2c4cacb9ca5fdd4
Signed-off-by: Szymon Sliwa <szs@semihalf.com>
|
|
Change-Id: Ib865a59c1124a50f0bd7c6e597f45190c26fa5f9
Signed-off-by: Szymon Sliwa <szs@semihalf.com>
|
|
Change-Id: I8be5f8aa63da8fdf4b2043ba9cd048f2269e4e99
Signed-off-by: Szymon Sliwa <szs@semihalf.com>
|
|
Change-Id: I5d71505e221e837c7569963fda674e970e5ee31d
Signed-off-by: Szymon Sliwa <szs@semihalf.com>
|
|
Based heavily on the src/vnet/ipsec.
Reuses the pure software VPP IPsec implementation
CLI command. To use enable-odp-crypto needs to be
present in the odp section of startup configuration,
like this:
odp {
enable-odp-crypto
}
Currenlty only IPv4 is supported, and only in transport
mode.
Authentication headers are not supported, this limitation
is inherited from VPP.
Supported crypto algorithms are
aes-cbc-128 for encryption, and
sha-512-256, sha-256-128, sha1-96 for authentication
Change-Id: I08c66f96a73d8cb4ef1095f181ddf47506abc39a
Signed-off-by: Szymon Sliwa <szs@semihalf.com>
|
|
1) Remove function vlib_buffer_set_known_state because
it is not multi-thread safe.
2) Stop all worker threads before deleting interfaces.
3) Fix style of odp_packet_exit function.
4) Fix infinite loop in odp_config.
Change-Id: I4463b4e416e440ac326ddaac8e46a95d17b202d2
Signed-off-by: Michal Mazur <mkm@semihalf.com>
|
|
Change-Id: I5abeddbc20da4528e811537d071de96a2800e016
Signed-off-by: Michal Mazur <mkm@semihalf.com>
|
|
Adjust ODP plugin for Caterpillar release.
Address range of packet pool can be obtained using odp_pool_info API.
Change-Id: I0b0175cce4d9e44ca27709d1dbca560a4a72b5a1
Signed-off-by: Michal Mazur <mkm@semihalf.com>
|
|
Buffers with VLIB_BUFFER_RECYCLE flag set should not be freed.
They are replicated on send and sent back to vlib for processing.
Change-Id: If381884fa7c898253f35dc3a39feeb4eb1529286
Signed-off-by: Michal Mazur <mkm@semihalf.com>
|
|
Free lists are required by some VPP nodes e.g. IPv6.
Update copyrigths as lots of code were imported from src/vlib/buffer.c
Change-Id: Ie4c56d3a3104624fe77a04069fe3ba1281d3f0cb
Signed-off-by: Michal Mazur <mkm@semihalf.com>
|
|
Packet pool cannot be destroyed when last interface is deleted
because it will not be reallocated on creation of a new interface.
Change-Id: Iec7b03b4413fc6602a7a1493ffd706f4c4dd1309
Signed-off-by: Michal Mazur <mkm@semihalf.com>
|
|
Remove odp-linux from supported implementations.
Change-Id: Ie83ab7314bcbd359bb3071743b9eb2b8d3df6a5e
Signed-off-by: Michal Mazur <mkm@semihalf.com>
|
|
ODP platform can now be configured using environment variable
ODP_PLATFORM_PARAMS. Remove the option from startup config file.
Change-Id: I9d550dabd32c691a031bb88d5453c8146d5fbce8
Signed-off-by: Michal Mazur <mkm@semihalf.com>
|
|
Environment variables may contain configuration for platform drivers.
They were lost when starting VPP using 'make run/debug' commands.
Change-Id: I47d71b1d47a33cb8386f106bac215208e7030b1d
Signed-off-by: Michal Mazur <mkm@semihalf.com>
|
|
Fix support of atomic mode and add two new modes: parallel and ordered.
Three modes of synchronization between threads are supported now:
- parallel - no synchronization between events
- ordered - relative sequence of events is restored in output queues
- atomic - only a single thread may process events from a single queue
Change-Id: I9399aa601f0927042ac0f7f1e5f265f4d47724b8
Signed-off-by: Michal Mazur <mkm@semihalf.com>
|
|
1) Parse multiple output packets in loop.
2) Remove checking for error flags, they are never set because
ODP parser is disabled.
3) Do not call vlib_buffer_advance if not necessary.
4) Remove lock in TX path. Queues are synchronized by ODP.
5) Optimize detection of packet type.
6) Create a wrapper for call to odp_packet_user_area.
Change-Id: Ib50c9be9a62d67824b58ef2cd443b7fc59471655
Signed-off-by: Michal Mazur <mkm@semihalf.com>
|
|
1) Allow to setup size of allocated memory, number of packets in the
pool, number of packets in transmit burst and delay in scheduler
mode from VPP startup file.
More details are in template: src/vpp/conf/startup.conf
2) Separate Rx mode from Tx mode. Traffic Manager is not supported.
3) Devices defined in VPP startup are automatically created at start.
Change-Id: I1f8a3c2900f56cb23ce6997d311cfe55e1c2078c
Signed-off-by: Michal Mazur <mkm@semihalf.com>
|
|
1) Incoming packets can be spread across multiple worker
threads based on IP and TCP/UDP headers.
2) Multiple output queues are used in Burst mode if supported
by hardware (checked in interface capabilities)
3) Synchronization of output traffic can be disabled due to
multiple Tx queues - one for each thread.
Change-Id: Ib5ee18103c860eae3b56ffc453a5953c729bb521
Signed-off-by: Michal Mazur <mkm@semihalf.com>
|
|
1) Handle multiple ODP packets at once in receive loop
2) Wait to collect as many RX buffers as possible in single vector
3) Add prefetch of received and transmitted buffers
4) Disable parser, classifier and synchronization of RX queues
Synchronization of Tx queues can also be disabled if 2 ports used.
Change-Id: I65ed49ef2b60278022712e10a83f6ca24360694e
Signed-off-by: Michal Mazur <mkm@semihalf.com>
|
|
This patch allows VPP to run multiple worker threads compatible with ODP.
By default a single thread is assigned to each interface but this
can be changed using "set interface rx-placement" command.
Only Direct input/output mode without RSS is supported.
Change-Id: Ia6af7e5af4950159eb28b7d56bc9c8d34b80f8d4
Signed-off-by: Michal Mazur <mkm@semihalf.com>
|
|
This structure contains data shared among multiple threads and
must be globally accessible.
Change-Id: Ibde8222e9ed3f61561f7902750a1716da5c0bd81
Signed-off-by: Michal Mazur <mkm@semihalf.com>
|
|
Based on patch from Sreejith Surendran Nair.
Change-Id: I32ff73871ce0439378a7e3f0f9a93ac169e770cb
Signed-off-by: Michal Mazur <mkm@semihalf.com>
Signed-off-by: Sreejith Surendran Nair <srsurend@cisco.com>
Signed-off-by: Sachin Saxena <sachin.saxena@nxp.com>
|
|
Change-Id: I97617d84eb5ba6403bc2d8f87d8c76a086413094
Signed-off-by: Michal Mazur <mkm@semihalf.com>
|
|
Change-Id: I5c86cd0a1150ba2e915ce3413c67c504740356b3
Signed-off-by: Michal Mazur <mkm@semihalf.com>
|
|
|
|
Change-Id: I364fa14ade5282ec828662894041f9cc784ace7c
Signed-off-by: srsurend <srsurend@cisco.com>
Signed-off-by: sachin saxena <sachin.saxena@nxp.com>
|
|
Change-Id: I48adc6605d97821eb3f02f10a0afa4a7ab7a8a59
Signed-off-by: Sachin Saxena <sachin.saxena@nxp.com>
|
|
The previous iteration of the code used the L2 classifier tables for session storage,
as a result, the table allocations were pretty big. The new ACL plugin
datapath uses the tables just as a redirection mechanism, without adding any
entries. Thus, the tables can be much smaller.
Change-Id: Ieec4a5abf0abda6e513ab4e675f912f14d47e671
Signed-off-by: Andrew Yourtchenko <ayourtch@gmail.com>
|
|
Best guess, tested carefully, should do no harm. Clang doesn't
complain either way; it's not certain that this patch will make the
coverity warnings in tw_timer_template.c disappear.
Change-Id: I75aa0cfd8970751e823a1165df2a755e947c4cf9
Signed-off-by: Dave Barach <dave@barachs.net>
|
|
l2_len was not updated for the third tag
as the ethernet node retracts by the vlan count after parse_header (using
ethernet_buffer_header_size) it ends up pointing before the ethernet header
+ some minor cleanups
Change-Id: I4ccaedd33928912e5d837376f146503b27071741
Signed-off-by: Eyal Bari <ebari@cisco.com>
|
|
Updated test to reduce tests runtime.
Change-Id: I7904628fc19d349d6c26502e49d4c990cb8816ff
Signed-off-by: Pavel Kotucek <pkotucek@cisco.com>
|
|
Change-Id: I7dedf283c83c7f0e0b7642f095b68bc0b40898cf
Signed-off-by: Neale Ranns <nranns@cisco.com>
|
|
Change-Id: I5077fcf3671a6116b475f87e43120efc10ecaa08
Signed-off-by: Matus Fabian <matfabia@cisco.com>
|
|
Change-Id: Ide6d26d6fcc81be6f26ac0abe2cd0d6a0838cfe6
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
after deleting a sub interface it's l2fib entries are left with a dangling
sw_if_index (while waiting for the ager to delete them).
changed "show l2fib" to reflect that state with "Deleted" as the interface name.
added sleep in test_l2_fib as a workaround for packets still passing after flush
will investigate...
Change-Id: Id998d7d3c6a073ef5005c5f3009e1cfb7febf7db
Signed-off-by: Eyal Bari <ebari@cisco.com>
|