Age | Commit message (Collapse) | Author | Files | Lines |
|
Type: improvement
AN SA is uni-drectional therefore it can be used only for encrypt or
decrypt, not both. So it only needs one thread ID. free up some space on
the 1st cacheline.
Signed-off-by: Neale Ranns <neale@graphiant.com>
Change-Id: I21cb7cff70a763cbe2bffead860b574bc80b3136
|
|
When the counter vectors are validated and they are already long enough
to fit the given index in memory, there is no need to increase the stats
segment epoch. In this case, the counter vectors do not change as a
result of the validation.
This optimization is necessary for the case when the configuration is
changed at multiple thousands per second rate. The counter vectors grow
at the beginning and their size stabilizes after a while. Without this
improvement, it can still take several seconds for a stats reader to
succeed.
Type: improvement
Signed-off-by: Miklos Tirpak <miklos.tirpak@gmail.com>
Change-Id: I5a6c30255832716a1460018d0bd0f63031de102b
|
|
Previous commit broke naming of vxlan interfaces.
Type:fix
Fixes:a4b0541f6
Signed-off-by: Ray Kinsella <mdr@ashroe.eu>
Change-Id: I5e304821be73547b4e47c35ad9632283f153830f
|
|
.api files
Type: feature
Make the auto-endian nature explicit, rather than hidden in the x_api.c
file.
Signed-off-by: Neale Ranns <neale@graphiant.com>
Change-Id: Ibe647117ceeaf6f99a38a96576a5a41a3cbb1615
|
|
enums can return signed or unsigned integers
enumflags are unsigned integers
Type: fix
Change-Id: Iafc8f8f09c96679c5983d2cb807699fcf90ca0d7
Signed-off-by: Paul Vinciguerra <pvinci@vinciconsulting.com>
|
|
Add counters to the policer against each of the 3 possible results:
conform, exceed and violate.
Type: improvement
Signed-off-by: Brian Russell <brian@graphiant.com>
Change-Id: Ia98a2f5655df6873259197d6bbf0ff2709b7d60e
|
|
For high throughput tests the amount of debug output can be
considerable.
Type: improvement
Signed-off-by: Florin Coras <fcoras@cisco.com>
Change-Id: Icc9f2092918ebca432502af960da11b154818a02
|
|
Type: fix
Change-Id: I02e473440a8732ddfb1a13ad6552779adaa67f60
Signed-off-by: Mathias Raoul <mathias.raoul@gmail.com>
|
|
When calling the bihash_add_del... functions some callers add a comment
beside the value to indicate this is the is_add param. Make the code
easier to read by adding defines for add and delete that the callers
can use instead of having to use 0 or 1.
Type: improvement
Signed-off-by: Paul Atkins <patkins@graphiant.com>
Change-Id: Iab5f7c8e8df12ac62fc7e726ca1798622dcdb42c
|
|
Type: docs
Change-Id: Ia541839e1f1ceddfae4579dece43b9cc820702e2
Signed-off-by: Andrew Yourtchenko <ayourtch@gmail.com>
(cherry picked from commit de569048a504cf211d80b86882e6781684301790)
|
|
Type: improvement
1) stack the interpose on any path-extensions (e.g. labels) from the
next best source
2) allow more than 1 source to contribute a DPO for a given prefix
Signed-off-by: Neale Ranns <neale@graphiant.com>
Change-Id: Idc2fbb36cfbd2387081765d8af0f1fbe61612160
|
|
Type: fix
Interrupt are suppressed from kernel on tx path.
Signed-off-by: Mohsin Kazmi <sykazmi@cisco.com>
Change-Id: I9f39f343b7e16bad09910766adf0b09654721f67
|
|
hi->name is a non-NULL-terminated vector.
Type: fix
Change-Id: I1a9f128f24f137b43cb47169677cc4288043fbd7
Signed-off-by: Benoît Ganne <bganne@cisco.com>
|
|
feature is u64. We need to print it with %llx and enough precision to
avoid truncation
Type: fix
Signed-off-by: Steven Luong <sluong@cisco.com>
Change-Id: I9f5dd523400473b4881b01b8e6cecf439a8060d9
|
|
Type: fix
A plugin exit callback function registered via
VLIB_MAIN_LOOP_EXIT_FUNCTION() is called without stopping worker cores
doing data path processing. Following error is observed while exiting
vpp.
error log:
vppctl# q
0: /vpp_master/src/vnet/interface.c:1058 (vnet_delete_hw_interface) assertion `ii == 0 || this_vlib_main->parked_at_barrier == 1' fails
This patch calls all exit() functions after acquiring barrier
Signed-off-by: Kommula Shiva Shankar <kshankar@marvell.com>
Reviewed-by: Nitin Saxena <nsaxena@marvell.com>
Change-Id: I2fd565f0afaa62cb0fb0389922f49e4d382f028c
|
|
Type: improvement
Signed-off-by: Florin Coras <fcoras@cisco.com>
Change-Id: I011faa52939d44bdc90de69ad381dc1eea8c8eb3
|
|
Type: fix
Signed-off-by: Florin Coras <fcoras@cisco.com>
Change-Id: I2d008a12da6c6e68ad8955a638b44d09e4febd8d
|
|
Type: improvement
Signed-off-by: Florin Coras <fcoras@cisco.com>
Change-Id: I1b083ee793a7cf91b1001bfe88353fa5e6515c42
|
|
Type: fix
Signed-off-by: Florin Coras <fcoras@cisco.com>
Change-Id: I4bfdeb4bfbdcf333cfd4047a80edd04e35a371fc
|
|
Type: fix
When a translation gets deleted, hiting a
session pointing to it sefaults. We're better
off directly storing the next node index.
Change-Id: I4f0716d775202b4ecf54d6cdb827bbeebd23056c
Signed-off-by: Nathan Skrzypczak <nathan.skrzypczak@gmail.com>
|
|
Type: fix
This fixes two issues :
- We used a hash to throttle RPC for adding fib entries,
but as we rely on a refcount, we cannot accept loosing an
entry, which could happen in case of a collision.
- On client cleanup we weren't freeing the fib entry correctly
which resulted in crashes when recreating an entry.
Added a test that ensures proper cleanup
Change-Id: Ie6660b0b02241f75092737410ae2299f8710d6b9
Signed-off-by: Nathan Skrzypczak <nathan.skrzypczak@gmail.com>
|
|
Pass packets arriving at the ip[46] punt policer nodes to punt policer
handoff nodes if the worker thread they arrive on is not the same one
configured in the policer. Initially, the policer will be tied to the
worker thread that it first received a packet on. This will be expanded
in future to be a configuration API option.
Type: improvement
Signed-off-by: Brian Russell <brian@graphiant.com>
Change-Id: Ic39d936084c354af1859ad3d946da6cd0f6e34d9
|
|
Add thread handoff for packets being policed. Note that the handoff
currently requires the policer index to be passed in. This is suitable
for use in the ip[46] punt paths where each policer node will only
ever use a single policer. For the more general case, this will be
expanded in future to use a policer index stored in packet metadata.
Type: improvement
Signed-off-by: Brian Russell <brian@graphiant.com>
Change-Id: I85a0ecbcfb025f8844e763224cd3de1561249aca
|
|
The policer is not thread safe. In order that handoff could be
performed, add a frame queue to each of the punt policer nodes.
Type: improvement
Signed-off-by: Brian Russell <brian@graphiant.com>
Change-Id: Iee50267ee7e36f0e6c95b9b43bf651648198b834
|
|
Add a thread index field to the policer structure. The policer is not
thread safe. The thread index will be used to tie it to one worker
thread and other workers can use thread handoff.
Type: improvement
Signed-off-by: Brian Russell <brian@graphiant.com>
Change-Id: I650e983a9ed800bf660d6f06368717484c4a83bf
|
|
Type: feature
A device/router needs to have a unique ID which is included in the flow
has so that flows are not polarised through the network, i.e. each deice
in the network chooses the same nth link for the same flow.
Signed-off-by: Neale Ranns <neale@graphiant.com>
Change-Id: I963e03674adbb085902b4084fdc4886b88f5734c
|
|
Type: fix
Change-Id: I79258a95cfb75587685dffae671255063f57cbef
Signed-off-by: Benoît Ganne <bganne@cisco.com>
|
|
When compiling in -O0, the glibc feature.h ignores _FORTIFY_SOURCE and,
in some distributions, throws a #warning, which fails the build. This
patch removes -D_FORTIFY_SOURCE in all debug configurations.
Change-Id: Ic30229c038353e49ff9419779e082a7083c1ea5f
Type: fix
Signed-off-by: Mohammed Hawari <mohammed@hawari.fr>
|
|
The check args->hwaddr is always true and it always copies the
mac address from args->hwaddr even though none was set.
Check args->use_custom_mac instead.
Type: fix
Fixes: gerrit 29970
Signed-off-by: Steven Luong <sluong@cisco.com>
Change-Id: I0c51bf1ea79b02c4fbdc3c52e694f186bdd96600
|
|
Type: feature
Change-Id: Ic502d806410ea3c8f3f1eac70b694114ccb053bf
Signed-off-by: Filip Tehlar <ftehlar@cisco.com>
|
|
Type: fix
Signed-off-by: Florin Coras <fcoras@cisco.com>
Change-Id: I6db15780d87426eee852aa020c50cbdf45dc1f95
|
|
Change-Id: I85a463b4ca15baf11e3eb70189f5190ba2585170
Type: refactor
Signed-off-by: Mohammed Hawari <mohammed@hawari.fr>
|
|
Add separate queue implementation for the message queue as it's custom
tailored for fifo segments as opposed to binary api.
Also move eventfds to the private data structures.
Type: refactor
Signed-off-by: Florin Coras <fcoras@cisco.com>
Change-Id: I6df0c824ecd94c7904516373f92a9fffc6b04736
|
|
Users tend to miss the warning too often, add it to the end of the
output in hope it would be easier to catch.
Type: improvement
Change-Id: I264df7e3b6ab1ffb1a383807f4444d9cbaf40999
Signed-off-by: Benoît Ganne <bganne@cisco.com>
|
|
Use error counters related to ipsec-mb return codes instead of
'bad-hmac' only.
Type: improvement
Change-Id: I9329da300a70d76b4d4ab30fa45f0a2a85d6519b
Signed-off-by: Benoît Ganne <bganne@cisco.com>
|
|
Propagate the multi-arch variant selection to interfaces.
Type: fix
Signed-off-by: Radu Nicolau <radu.nicolau@intel.com>
Change-Id: I99c4a7896f172f0d14d2ded22a27383825529a7d
|
|
Avoid accessing the private data structure of mutexes which is
implementation-dependent, eg. musl is different from glibc.
Type: improvement
Change-Id: I20ec0c1c9faef0749d89a1969cd2430c80ac04b3
Signed-off-by: Benoît Ganne <bganne@cisco.com>
|
|
- update quic plugin with new quicly/picotls API
- remove packet allocator
- remove crypto batching
- update picotls plugin
- add cli for quicly congestion control configuration
Type: feature
Change-Id: If76ef31c43b430eea2f7674539b2112aee0f351e
Signed-off-by: Mathias Raoul <mathias.raoul@gmail.com>
|
|
We have accumulated several scenarios in prod or wishlists
where it would be useful to have a general infra to say yes/no
about a certain test, and potentially make decisions based on that,
for example:
- runs solo (aka 'time-dependent')
- (wishlist) part of quick smoke-test set
- (wishlist) intermittent failure unrelated to timing
- (wishlist) test broken with a multi-worker config in vpp
Refactor the current "run-solo" code to allow for this extension.
Type: test
Change-Id: Ia5b3810e57c0543753c8e0dc4dc0cfb4a30b36ac
Signed-off-by: Andrew Yourtchenko <ayourtch@gmail.com>
Signed-off-by: Klement Sekera <ksekera@cisco.com>
|
|
Change-Id: Ic977ffe761efc2129c61aec581da5479fe4838da
Type: fix
Signed-off-by: Mohammed Hawari <mohammed@hawari.fr>
|
|
Change-Id: I03e164d8d5a329497f422e99f8b0058135241b4e
Signed-off-by: Mohammed Hawari <mohammed@hawari.fr>
Type: fix
|
|
Change-Id: If8077280cef501599f810ad9255efa2a5a451ced
Signed-off-by: Mohammed Hawari <mohammed@hawari.fr>
Type: improvement
|
|
Take into account pending interrupts when resizing rxq_interrupts
Type: fix
Signed-off-by: Mohammed Hawari <mohammed@hawari.fr>
Change-Id: I64f6f4404738c90ed5a2c10d72c7ce59cf96a644
|
|
Type: improvement
Signed-off-by: Jakub Horn <jhorn@cisco.com>
Change-Id: I06401a8645615d139893dca27d07bb3b3214bea7
|
|
Asking for TSO (TCP Segmentation Offload) on packets that are already
smaller than (headers + MSS) does not make sense and may not work
on some HW.
Fix to only set the TSO flag when a segmentation offload is
really required, i.e when packet is large enough.
Type: improvement
Signed-off-by: Somnath Kotur <somnath.kotur@broadcom.com>
Change-Id: I7830ae8474581c8e518fb4910f7863e10346bb62
Signed-off-by: Somnath Kotur <somnath.kotur@broadcom.com>
|
|
Performance improvement: on supported platforms, currently only Intel Tremont,
use a write combining store to update the tail pointers.
Also, Tremont node variant is added for all.
Type: improvement
Signed-off-by: Radu Nicolau <radu.nicolau@intel.com>
Change-Id: Ie9606e403b7d9655184f778e3ffee3027c8c9edd
|
|
Change-Id: Id539d36635f0ab9625dc2fc73630be39bead09af
Signed-off-by: Mohammed Hawari <mohammed@hawari.fr>
Type: improvement
|
|
Type: fix
Change-Id: If35cf7da0c3d1b92532cc2e45c164d6b5a08592c
Signed-off-by: Benoît Ganne <bganne@cisco.com>
|
|
Change-Id: Ifa8bccd8a34ec1b14e772ee53757e9083373e3de
Signed-off-by: Mohammed Hawari <mohammed@hawari.fr>
Type: feature
|
|
Type: improvement
Change-Id: Icb23af5f5e458a555f416cb0a829e84646b25dd9
Signed-off-by: Benoît Ganne <bganne@cisco.com>
|