Age | Commit message (Collapse) | Author | Files | Lines |
|
Change-Id: I30470e8a1404632b0350056769338dc69a94c105
Signed-off-by: Florin Coras <fcoras@cisco.com>
|
|
Change-Id: I579eef6d70b94b20e22b7b116ac63b2d7af49b22
Signed-off-by: Florin Coras <fcoras@cisco.com>
|
|
If uio_hv_generic is not loaded, then the startup code will
fallback to the older failsafe/tap method of initialization
in DPDK. Therefore don't put out scary message in the log.
Also, reorder startup to avoid manipulating lower device until/unless
uio is going to work.
Change-Id: Ie1cc77b4b5359c04f00a93d01a772eccf3bbab37
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
|
|
Use of scalar AVX512 instructions impacts the maximum CPU frequency
as they require CPU to request level 1 power license.
So sparse use of AVX512 like we do, typically causes more harm than
improvement.
This patch makes AVX2 higher priority than AVX512, but still there
is option to change it on per-node basis:
vpp# set node function ethernet-input avx512
Change-Id: I42e03510a6efc1756e22fbb70e8c76d7f74f59d7
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
Change-Id: I2f1fa15a99163b9c105707484503dc9502265c52
Signed-off-by: Florin Coras <fcoras@cisco.com>
|
|
And shuffle some of the other connection variables.
Change-Id: Ib5361d0a45537a3bd2421b8e4145ea0f0e525d91
Signed-off-by: Florin Coras <fcoras@cisco.com>
|
|
The log-level dpdk config value should be transparently
forwarded to DPDK via EAL argument. Since DPDK now supports
naming log-levels, VPP no longer needs to parse and call
rte_set_loglevel().
This was the other part of the DPDK log-level change.
It must have got missed during my initial checkin.
Without it passing dynamic log-level values like are silently
ignored.
Fixes: 6ca6ac6c887e ("dpdk: support passing log-level")
Change-Id: I732cec5f638c9924e3ffb04c4753f957e3633d64
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
|
|
Change-Id: I5cfa0f6eee67156bf87907fcf8a39f16d68a0905
Signed-off-by: Dave Barach <dave@barachs.net>
|
|
Change-Id: Id55ec87724e421d5b722314f9302c6ade7545306
Signed-off-by: Filip Tehlar <ftehlar@cisco.com>
|
|
Change-Id: I8370c06150ce4499475e9d6dc6b3ab8be2016202
Signed-off-by: Filip Tehlar <ftehlar@cisco.com>
|
|
Change-Id: Iac105e09bb88434fe55365fc1546d769f034e0e2
Signed-off-by: Filip Tehlar <ftehlar@cisco.com>
|
|
Change-Id: I0562d597fd45c7ddcb6db42cf17d3ffb569eb140
Signed-off-by: Neale Ranns <nranns@cisco.com>
|
|
Change-Id: I719882acb59bd069fd88b10989f11085a0c41ae6
Signed-off-by: Neale Ranns <nranns@cisco.com>
Signed-off-by: Mohsin Kazmi <sykazmi@cisco.com>
|
|
Change-Id: Ib27952935393163eaabf005c69b1cbc2feca2b98
Signed-off-by: Neale Ranns <nranns@cisco.com>
|
|
VAPI does not handle two VLAs in one struct.
Change-Id: I259c998bef4398ead2bbb9e788350d50c2f05694
Signed-off-by: Neale Ranns <nranns@cisco.com>
|
|
Previous code was walked buffer chain, effectively trying to "truncate" the chain, reset the
length of first buffer and reused that as the ICMP error message. That could have issues in cases
there were other users of the buffer chain. Update to clone the first buffer in chain, and
use that for the ICMP error message instead.
Change-Id: Ibc1a0bf2d854dae41874808c8297028ed93dd69d
Signed-off-by: Ole Troan <ot@cisco.com>
|
|
when adding a recursive path the table is locked
so that it can be removed when the last recursive path
is removed. however, not all RR source'd prefixs use
a recursive path. so flushing the table of all RR source'd
entries is not correct.
Change-Id: Id4010774011046e66ddc443ac83cb8e9245313dd
Signed-off-by: Neale Ranns <nranns@cisco.com>
|
|
Change-Id: Iffb4b314be3ded0d9c6acb77ec1c6f22778f301d
Signed-off-by: Klement Sekera <ksekera@cisco.com>
|
|
Change-Id: Icfba7ae9cbf504a8966923eac06c9eb0e11899d2
Signed-off-by: Florin Coras <fcoras@cisco.com>
|
|
this is a workaround that avoids building/linking other archs for error-node.
Combination error-node + avx512 arch triggers a crash in pg during
test run.
Change-Id: I6ed39fdf0a58f82d951082b1cf91ab2a57d8108c
Signed-off-by: Filip Tehlar <ftehlar@cisco.com>
|
|
Change-Id: I6f0830b786ac46c69c867f73b044aab174e6210b
Signed-off-by: Ole Troan <ot@cisco.com>
|
|
Change-Id: I3043112c3e7584f61e64dc6d20d57604ebceb76a
Signed-off-by: Filip Tehlar <ftehlar@cisco.com>
|
|
Change-Id: I7114cd55ce1fc320d17b36cab731546fe4c58a72
Signed-off-by: Filip Tehlar <ftehlar@cisco.com>
|
|
- disconnect sessions through session layer apis instead of directly
notifying the app worker.
- improve cli
- increase fifo sizes for tls app
Change-Id: I8a7d2865b3b00724e2a9da29fa4a906ea867da9b
Signed-off-by: Florin Coras <fcoras@cisco.com>
|
|
Change-Id: I42addcd413094a8b6e9b2858ba2a292bd295ec82
Signed-off-by: Florin Coras <fcoras@cisco.com>
|
|
Close the mmap file descriptor at first opportunity.
Change-Id: If8faa1de0a57dc7d8abb3eed97861813b841bd92
Signed-off-by: Ole Troan <ot@cisco.com>
|
|
It turns out that for scalar sizes 0..24, frames are always the same
size. That range includes all current use-cases - and then some - so
get rid of the hash table. Old code preserved under #ifdef
VLIB_SUPPORTS_ARBITRARY_SCALAR_SIZES.
Change-Id: Ic005c7143c9639f77d1a0fadd2fc0e90dccb68c1
Signed-off-by: Dave Barach <dbarach@cisco.com>
|
|
Change-Id: I0d56d2cee012b46da848362f322d6f246f8bf4c2
Signed-off-by: wuzhouhui <wuzhouhui14@mails.ucas.ac.cn>
|
|
Change-Id: Ie0680330dd619cfdcb72ae8fb7e2463f732cff8c
Signed-off-by: Neale Ranns <nranns@cisco.com>
|
|
Change-Id: I3cecffae2e30233d4725bd3eaf5ff4a8ff194fae
Signed-off-by: Florin Coras <fcoras@cisco.com>
|
|
Change-Id: Idc7dfe743399dd8dee0f6b3ec83f194f3fca580b
Signed-off-by: Florin Coras <fcoras@cisco.com>
|
|
Change-Id: Idf060f385f4d9b2f42fdab6a1d372727beb8b19e
Signed-off-by: Florin Coras <fcoras@cisco.com>
|
|
Change-Id: I7f3b015ea6750c9773e4285bb63b0d44fa2177b9
Signed-off-by: Florin Coras <fcoras@cisco.com>
|
|
Change-Id: I9c45f390a6454c114f12f9c46c3a93fcecffa73f
Signed-off-by: Neale Ranns <nranns@cisco.com>
|
|
fix a debug CLI scripting bug: cp_ip6_address_add_del_command_function
ate any subsequent commands, yielding indigestion.
Change-Id: Iaca7bed5687759da36ae91dc658e758549b71796
Signed-off-by: Dave Barach <dave@barachs.net>
|
|
The stdlib introduced IntEnum in python 3.4 and IntFlag in python 3.6.
Change-Id: I3ac278a9d5a97eefa9fc4f1491f0cd030e40c3b2
Signed-off-by: Paul Vinciguerra <pvinci@vinciconsulting.com>
|
|
Avoid the cache miss consequences of spraying [functionally harmless]
junk into un-prefetched rewrite space. As things stand, several tunnel
encap rewrites set rewrite data_bytes = 0, and take a performance hit
due to unwanted speculative copying.
Should be performance-neutral in speed-path cases, which won't execute
the added check.
Change-Id: Id83c0325e58c0f31631b4bae5a06457dfc7ed567
Signed-off-by: Dave Barach <dave@barachs.net>
|
|
Also, increase ip6 address field width in "show ip6 conn verbose"
output
Change-Id: If7d1bcd439d94a22d1f1c6c1298cb30aba13d0a2
Signed-off-by: Dave Barach <dave@barachs.net>
|
|
Change-Id: I352f4a4adcf8771c21530657efcaecb532416612
Signed-off-by: Joe Zhou <zhouyoucn@qq.com>
|
|
memif, lacp, nsh and cdp used local REPLY_MACROs.
Remove and use those in api_helper.h
Change-Id: Ib01d6ae5cff0b6f1cef90996a54b3177f0c53463
Signed-off-by: Ole Troan <ot@cisco.com>
|
|
1. fix wrong assignemnt of lik/rik
2. keys initialized to 0, to avoid using random data
in stack. could cause memory overlapped then crash
3. show sa->id in hex format
Change-Id: Id0430aa49bb55c27cee4f97f8c0e4ec87515dcd2
Signed-off-by: Kingwel Xie <kingwel.xie@ericsson.com>
|
|
Change-Id: I77a43bfb37d827727c331cd65eee77536cc15953
Signed-off-by: Ole Troan <ot@cisco.com>
|
|
Change-Id: I952e6aec6487270a79c4d92cfe828cc55d42d536
Signed-off-by: Neale Ranns <nranns@cisco.com>
|
|
gdb_show_traces() dumps buffer traces.
Ease gdb debugging when vpp crashed...
Change-Id: Ib24314832386ee4defc2d31cbb4c05d293fb3338
Signed-off-by: Kingwel Xie <kingwel.xie@ericsson.com>
|
|
Change-Id: I01730ec9eb8033074c8710daf0848c3573293aeb
Signed-off-by: Filip Tehlar <ftehlar@cisco.com>
|
|
Change-Id: I88d2632fa451dbafbc212dd142a67fe5ec4cd610
Signed-off-by: Filip Tehlar <ftehlar@cisco.com>
|
|
Coverity complains about resource leak after open when fd gets 0 with below
warning.
off_by_one: Testing whether handle tfd is strictly greater than zero is
suspicious. tfd leaks when it is zero.
It is right. 0 is a valid fd. -1 is not.
Change-Id: I22c2eb75b99bb6209921b9f874190cbbdf10e6ce
Signed-off-by: Steven Luong <sluong@cisco.com>
|
|
Change-Id: I61d8c35f48a059968909fc8523bd313fc4799389
Signed-off-by: Hongjun Ni <hongjun.ni@intel.com>
|
|
Change-Id: Idbf7f3a57dc399798b8dba9463daeb7d66470ab1
Signed-off-by: Florin Coras <fcoras@cisco.com>
|
|
1. changed to vlib_buffer_enqueue_to_next
2. error counter fixes; stats added to last_sw_if_index
when interface changed
3. udp-encap support
Change-Id: I70b0814aa37181fea4d70fa3c96c608adb5afe49
Signed-off-by: Kingwel Xie <kingwel.xie@ericsson.com>
|