According to the description in the startup.conf, the assignment
of worker threads starts with the lcore following the main_lcore.
A non-zero skip_cores will correctly achieve this assignment.
However, prior to this patch when workers are assigned, the code
picks up and assigns core 0 even thought it shouldn't.
This patch determins if a non-zero number of workers are desired
and if so, marks CPU unavailable for a worker assignment.
Type: fix
Change-Id: I1fdf73a6f218dcbf146fda2efc90c553f7cd6d20
Signed-off-by: Jon Loeliger <jdl@netgate.com>
(cherry picked from commit 4a06846dd668d7f687e6770215c38e8feb5f1740)
L4 checksums for IPv6 should be calculated using a pseudo header that
includes the source/destination addresses, payload length, and payload
protocol.
ip6_tcp_udp_icmp_compute_checksum() was using the payload length and
protocol from the IPv6 header. If there is a hop-by-hop header (or any
other extension header), the payload length used for the pseudo header
should only include the upper layer header and payload and not the
extension header bytes. Same deal with the protocol, the upper layer
next header value should be used instead of the extension header.
Type: fix
Fixes: cb9cadad57
Change-Id: Ifa2c9ad41c0fc4eea674f0671255b637c8e01f71
Signed-off-by: Matthew Smith <mgsmith@netgate.com>
(cherry picked from commit 97677a26f7c857e7ee0acbdb2c13eef214aa70a8)
Using random addresses can confuse AddressSanitizer
Type: fix
Change-Id: I44368093f899672ac4d511cc5a01ed87c988e63a
Signed-off-by: Benoît Ganne <bganne@cisco.com>
(cherry picked from commit dc90c719bca27856101f758939dcfb0b67124775)
Avoid using vec_del1() for directory vector to keep indexes valid all
the time.
There are state counters for each slave in LACP bond mode which can be
dynamically created and removed. Vector index is used to access these
counters. But also vec_del1() is used to remove counter from vector.
This function changes the index of the last element, so after this we
are unable to access ex-last element using old index.
As a result it is not possible to add-del-add two interfaces to the LACP
bond:
DBGvpp# create bond mode lacp
BondEthernet0
DBGvpp# create packet-generator interface pg1
DBGvpp# create packet-generator interface pg2
DBGvpp# bond add BondEthernet0 pg1
DBGvpp# bond add BondEthernet0 pg2
DBGvpp# bond del pg1
DBGvpp# bond del pg2
DBGvpp# bond add BondEthernet0 pg1
DBGvpp# bond add BondEthernet0 pg2
bond add: /if/lacp/1/3/partner-state is already register
Type: fix
Signed-off-by: Vladimir Isaev <visaev@netgate.com>
Change-Id: I2c86e13905eefdef6233369cd4ab5c1b53d123bd
(cherry picked from commit 72e31bc2d9b910147c09e1c329713fccc873a018)
Type: fix
Some fiber ports that are managed by the ixgbe PMD have the
possibility to get into a state where link can never be brought up.
This patch should fix it and will be submitted to upstream DPDK.
Change-Id: Ia4d0df2e70d098b2151e513b96e8bd742151e8ce
Signed-off-by: Matthew Smith <mgsmith@netgate.com>
(cherry picked from commit 0860b2e19365c092f10dd1ce639caaded0e87ded)
Recent modification added a call to vnet_gso_header_offset_parser in the
beginning of vhost_user_handle_tx_offload. The former routine may set tcp or
udp->checksum to 0. While it is appropriate to set it to 0 for the GSO packet,
it is broken and causes checksum error if the aformentiooned routine is called
by a non-GSO packet. The fix is to not call vhost_user_handle_tx_offload
if the buffer does not indicate checksum offload is needed.
Type: fix
Signed-off-by: Steven Luong <sluong@cisco.com>
Change-Id: I6e699d7a40b7887ff149cd8f77e8f0fa9374ef19
(cherry picked from commit 564e1672917e205d7ae79525bb937df18f8d764b)
Prevent malicious packets with spoofed embedded IPv4 addresses
by limiting the IPv6 ingress packets to known MAP-T domains.
Drop spoofed packets.
Add several tests that ensure spoofing isn't allowed.
Type: fix
Fixes: fc7344f9be
Change-Id: I80a5dd10d5fe7492e3a1b04de389d649a78065e2
Signed-off-by: Jon Loeliger <jdl@netgate.com>
(cherry picked from commit 65866f03d96bd41b99b1c823ea6f38cd77fac58c)
For ip4 tcp, ip6 tcp, and ip6 udp packet, we set checksum = 0 prior to
computing the checksum. We missed ip4 udp case. This oversight requires all
clients to set udp->checksum = 0 if ip4 udp checksum offload is needed.
Type: fix
Signed-off-by: Steven Luong <sluong@cisco.com>
Change-Id: Ic608811e82099f3bec469e123671e9b281f38d76
(cherry picked from commit 03328ec8bb86b93fa70bb6b2a9b37c40e686a1f7)
The 'tag' parameter is expected to be a NULL-terminated C-string in
callees:
- make sure it is null-terminated in both API and CLI cases
- do not allocate & copy the string into a non-NULL-terminated vector
in API case
- fix leak in CLI case
Type: fix
Change-Id: I221a489a226240548cdeb5e3663bbfb94eee4600
Signed-off-by: Benoît Ganne <bganne@cisco.com>
(cherry picked from commit 3b37125bdb0251181f90a429a4532b339711cf89)
Call session queue node with the right node runtime instead of the
pre-input node runtime.
Type: fix
Change-Id: I43d20bed4930fc877b187ce7ecdce62034b393c5
Signed-off-by: Florin Coras <fcoras@cisco.com>
(cherry picked from commit 2d8829cbb5f3d214fbc09bf4258573659e0c5e60)
Currently if user want to set ip4 address to the api, it must convert to ip6
format, e.g. user want to ip4 "90.1.2.1" but must convert to "::5A01:0201",
it is not acceptable, this fix solved the issue.
Ticket: FDIO-753
Type: fix
Change-Id: I2ffa5a3d38400ee176cf601421074f71fc395f03
Signed-off-by: Yulong Pei <yulong.pei@intel.com>
(cherry picked from commit db43bb6af78c33e47d29889b047cced4b11fe4d7)
Fixing compilation issuues for 32-bit also setting init flag for shm based bihash
Type: fix
Signed-off-by: Vijayabhaskar Katamreddy <vkatamre@cisco.com>
Change-Id: Ic2072c5ba7fc77d061ca9f1b844a71f6e22e58b2
(cherry picked from commit f0bae64f6fd4c410c19f6ece688443f389932688)
Two modules in vnet include vpp/app/version.h but there is
no explicit build dependency for this generated file. This
leaves a race condition in the build system that the Coverity
build has recently started triggering.
Change-Id: I8e2bb32feeb16e1bdd8efb0d2633cfdba60f51aa
Type: fix
Signed-off-by: Chris Luke <chrisy@flirble.org>
(cherry picked from commit c171d01cdb5183c8bf640951e94af6b1fd5e3efc)
All translated ICMPv6 packets that exceed the minimal IPv6 MTU get
truncated but according to RFC 4443 2.4 only ICMPv6 error messages
(type < 128) need to be truncated.
With this commit, truncate only ICMPv6 error messages.
Type: fix
Change-Id: Ic455352de2ff4ff6aa3421b46a2a54923f2d3f80
Signed-off-by: Alexander Chernavin <achernavin@netgate.com>
(cherry picked from commit 180210f99b74b97b127b7800bdc7bd243713cbf4)
n_retry was never decremented and so never enforced.
Type: fix
Change-Id: I71d60a72c156286f7e5b82b1c77a723361317c69
Signed-off-by: Benoît Ganne <bganne@cisco.com>
(cherry picked from commit 207a1633094526697729f322269b937f841aaf47)
Type: fix
Fixes: be2286b0
This patch does following:
- If terminating frame has consecutive packets with same source IP, this patch
avoids fib lookup for those packets in ip4-local node. This drops cycle count
for ip4-local node on both ARM and x86. It being done by enabling dead code in
else {} case of ip4_local_check_src_x2() and ip4_local_check_src() functions.
- In case all packets in terminating frame have unique source IP (e.g:
incrementing), ip4-local is costlier by 2 cycles (broadwell)
Change-Id: I472ddc324716cec8bfe601568b8aeb7565f97ab3
Signed-off-by: Nitin Saxena <nsaxena@marvell.com>
(cherry picked from commit 2d18d2ea9f0e3d6c47d365ec135af651b14e8165)
Type: fix
Persist and retransmit can pop at the same time.
Signed-off-by: Florin Coras <fcoras@cisco.com>
Change-Id: Ia9530645cc84c83a881a75d7b4627197dc50ed29
(cherry picked from commit a6696719cb9fcd8ab54a5007e91dac6aeffe4e70)
The path must be next-to-impossible to hit, because the code has been
wrong for at least 5 years.
Type: fix
Signed-off-by: Dave Barach <dave@barachs.net>
Change-Id: I23b8c4e1631827e7931f353c561c1e19c596c598
(cherry picked from commit 5c944eef7012e7c5b363399ed92966fb659196b7)
