Anoob Joseph [Tue, 25 Jun 2019 05:39:24 +0000 (11:09 +0530)]
cryptodev: add feature flags to disable
Adding a new field, ff_disable, to allow applications to control the
features enabled on the crypto device. This would allow for efficient
usage of HW/SW offloads.
Signed-off-by: Anoob Joseph <anoobj@marvell.com> Acked-by: Akhil Goyal <akhil.goyal@nxp.com>
Added info about socket id used by compression perf
test process and by the driver
Signed-off-by: Tomasz Cel <tomaszx.cel@intel.com> Signed-off-by: Fiona Trahe <fiona.trahe@intel.com> Acked-by: Artur Trybula <arturx.trybula@intel.com>
- Include list of supported adapters.
- Include list of supported features.
- Document requirements for vector mode PMD.
- Remove obsolete limitation (scatter rx has been supported for
some time).
- Fixed broken links.
When adding offloads from commandline, not only port
related configuration bits should be set, but also queue
related offloads configuration bits, or it will cause error.
For example, test in this process for ixgbe:
(1)./x86_64-native-linuxapp-gcc/app/testpmd -c 0x6 -n 4
-- -i --portmask=0x1 --port-topology=loop --disable-crc-strip
(2)port stop all
(3)port config all crc-strip on
(4)port start all
we will see "Fail to configure port 0 rx queues" of warning info.
Fixes: 0074d02fca21 ("app/testpmd: convert to new Rx offloads API") Cc: stable@dpdk.org Signed-off-by: Wei Zhao <wei.zhao1@intel.com> Acked-by: Bernard Iremonger <bernard.iremonger@intel.com>
Haiyue Wang [Tue, 25 Jun 2019 10:38:03 +0000 (18:38 +0800)]
app/testpmd: fix parsing RSS queue rule
Declare a local variable 'arg' to hold the pointer to ARGS_ENTRY_ARB,
otherwise, the pop 'arg' in 'parse_int' holds wrong data like 'arg->
offset = 5912737, arg->size = 0', this caused the parse failure.
If calling like 'push_args(ctx, ARGS_ENTRY_ARB..)', the below code for
assigning 'offset & size' will be missed for some gcc optimization.
The above assembly code is from this modification with:
gcc version 9.1.1 20190503 (Red Hat 9.1.1-1) (GCC)
Fedora release 30 (Thirty)
5.1.11-300.fc30.x86_64
Fixes: d0ad8648b1c5 ("app/testpmd: fix RSS flow action configuration") Cc: stable@dpdk.org Signed-off-by: Haiyue Wang <haiyue.wang@intel.com> Reviewed-by: Xiaolong Ye <xiaolong.ye@intel.com> Acked-by: Bernard Iremonger <bernard.iremonger@intel.com>
Asaf Penso [Wed, 19 Jun 2019 09:52:45 +0000 (09:52 +0000)]
net/mlx5: fix condition for link update fallback
mlx5_link_update uses the newer ethtool command
ETHTOOL_GLINKSETTINGS to determine interface capabilities but falls
back to the older (deprecated) ETHTOOL_GSET command if the new
method fails for any reason.
The older method only supports reporting of capabilities up to 40G.
However, mlx5_link_update_unlocked_gs can return a failure for a
number of reasons (including the link being down).
Using the older method in cases of transient failure of the method
can result in reporting of reduced capabilities to the application.
The older method (mlx5_link_update_unlocked_gset) should only be
invoked if the newer method returns EOPNOTSUPP.
This commit removes the support of configuring the device E-switch
using TCF since it is now possible to configure it via DR (direct
verbs rules), and by that to also remove the PMD dependency in libmnl.
mlx5 implements mlx5_flow_null_drv_ops to be used when a specific
flow typei/driver is not available or invalid.
This routines return error without modifying the rte_flow_error
parameter passed to them which causes testpmd, for example, to crash.
This commit addresses the issue by modifying the rte_flow_error
parameter in theses routines.
Vivek Sharma [Sun, 30 Jun 2019 02:25:19 +0000 (07:55 +0530)]
net/octeontx2: support VLAN offloads
Support configuring VLAN offloads for an ethernet device and
dynamic promiscuous mode configuration for VLAN filters where
filters are updated according to promiscuous mode of the device.
Signed-off-by: Vivek Sharma <viveksharma@marvell.com> Acked-by: Jerin Jacob <jerinj@marvell.com>
Jerin Jacob [Thu, 30 May 2019 06:54:11 +0000 (12:24 +0530)]
net/octeontx2: support packet type
The fields from CQE needs to be converted to
ptype and rx ol flags in mbuf. This patch adds
create lookup memory for those items to be
used in Fastpath.
Signed-off-by: Jerin Jacob <jerinj@marvell.com> Signed-off-by: Kiran Kumar K <kirankumark@marvell.com> Signed-off-by: Harman Kalra <hkalra@marvell.com>
Krzysztof Kanas [Sun, 26 May 2019 17:34:42 +0000 (23:04 +0530)]
net/octeontx2: enable Tx through traffic manager
This patch enables pkt transmit through traffic manager
hierarchy by clearing software XOFF on the nodes and linking
tx queues to corresponding leaf nodes.
It also adds support to start and stop tx queue using
traffic manager.
This patch sets up and configure hierarchy in hw
nodes. Since all the registers are with RVU AF,
register configuration is also done using mbox
communication.
Signed-off-by: Nithin Dabilpuram <ndabilpuram@marvell.com> Signed-off-by: Krzysztof Kanas <kkanas@marvell.com>
Introduce traffic manager infra and default hierarchy
creation.
Upon ethdev configure, a default hierarchy is
created with one-to-one mapped tm nodes. This topology
will be overridden when user explicitly creates and commits
a new hierarchy using rte_tm interface.
Signed-off-by: Nithin Dabilpuram <ndabilpuram@marvell.com> Signed-off-by: Krzysztof Kanas <kkanas@marvell.com>
Vamsi Attunuru [Wed, 29 May 2019 16:45:32 +0000 (22:15 +0530)]
net/octeontx2: handle port reconfigure
setup tx & rx queues with the previous configuration during
port reconfig, it handles cases like port reconfigure without
reconfiguring tx & rx queues.
Jerin Jacob [Wed, 29 May 2019 03:38:14 +0000 (09:08 +0530)]
net/octeontx2: add device configure operation
Add device configure operation. This would call lf_alloc
mailbox to allocate a NIX LF and upon return, AF will
return the attributes for the select LF.
Andy Pei [Thu, 4 Jul 2019 01:38:39 +0000 (09:38 +0800)]
net/i40e: fix crash when TxQ/RxQ set to 0 in VF
Testpmd would stuck and result in core dump when user specifies an
invalid VF queue number, for example when setting txq or rxq to 0.
When txq or rxq is set to 0, pointer of pointer rxq or txq in
function i40evf_configure_vsi_queues is NULL. The usage of txq[i]
or rxq[0] is valid. This patch fixes this issue.
Fixes: d6b19729093e ("i40evf: support configurable crc stripping") Cc: stable@dpdk.org Signed-off-by: Andy Pei <andy.pei@intel.com> Acked-by: Qi Zhang <qi.z.zhang@intel.com>
The definition of ipsec and rsc offloads was incorrect.
Neither one of these is used by current driver, but the values
should be fixed now for future RSC support.
Due to position change, cannot be a full time maintainer anymore.
The git tree dpdk-next-net-mlx will be managed by Raslan,
as approved by the DPDK Technical Board.
For mlx5 PMD, Slava is added as additional maintainer.
Will continue to be involved in the DPDK community in different ways.
Having version and ihl set to 0 would result in a
corrupted/incorrect encapsulating packet for MPLS,
this commit sets these fields to 4 and 5 respectively
as defined in librte_net/rte_ip.h.
Define IPv4 Minimum IHL and VHL according to rfc791 (see [1])
"The Version field indicates the format of the
internet header."
"Internet Header Length (ihl) is the length of the
internet header in 32 bit words, and thus points
to the beginning of the data. Note that
the minimum value for a correct header is 5."
In current implementation, an action which requires parameters
must accept them enclosed in a structure.
Some actions require a single, trivial type parameter, but it still
must be enclosed in a structure.
This obligation results in multiple, action-specific structures, each
containing a single trivial type parameter.
This patch introduces a new approach, allowing an action configuration
object of any type, trivial or a structure.
Krzysztof Kanas [Mon, 24 Jun 2019 14:32:00 +0000 (16:32 +0200)]
net/af_packet: set default block size to page size
Kernel validates block size to be aligned to page size. Code works on
platforms with page size 4k, but on others e.g. Arm64 fails with default
parameters.
Signed-off-by: Krzysztof Kanas <kkanas@marvell.com> Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>
This patch implements additional actions of packet header
modifications.
Add actions:
- INC_TCP_SEQ - Increase sequence number in the outermost TCP header.
- DEC_TCP_SEQ - Decrease sequence number in the outermost TCP header.
- INC_TCP_ACK - Increase acknowledgment number in the outermost TCP
header.
- DEC_TCP_ACK - Decrease acknowledgment number in the outermost TCP
header.
app/testpmd: add actions to modify TCP header fields
Add actions:
- INC_TCP_SEQ - Increase sequence number in the outermost TCP header.
- DEC_TCP_SEQ - Decrease sequence number in the outermost TCP header.
- INC_TCP_ACK - Increase acknowledgment number in the outermost TCP
header.
- DEC_TCP_ACK - Decrease acknowledgment number in the outermost TCP
header.
Original work by Xiaoyu Min.
This patch uses the new approach introduced by [1], using a new
macro ARG_ENTRY_HTON to pass a single integer argument to each of
the new actions.
Add actions:
- INC_TCP_SEQ - Increase sequence number in the outermost TCP header.
- DEC_TCP_SEQ - Decrease sequence number in the outermost TCP header.
- INC_TCP_ACK - Increase acknowledgment number in the outermost TCP
header.
- DEC_TCP_ACK - Decrease acknowledgment number in the outermost TCP
header.
Original work by Xiaoyu Min.
This patch uses the new approach introduced by [1], using a simple
integer instead of using an action-specific structure for each of
the new actions.
Andrew Rybchenko [Fri, 21 Jun 2019 09:40:40 +0000 (10:40 +0100)]
net/sfc: ensure that device is closed on removal
PCI device may be unplugged when ethdev is not closed yet.
Fixes: aaa3f5f0f79d ("net/sfc: add configure and close stubs") Cc: stable@dpdk.org Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com> Reviewed-by: Ivan Malov <ivan.malov@oktetlabs.ru>
Andrew Rybchenko [Fri, 21 Jun 2019 09:40:39 +0000 (10:40 +0100)]
ethdev: avoid error on PCI unplug of closed port
If PCI Ethernet device driver removes it on close
(RTE_ETH_DEV_CLOSE_REMOVE) and later PCI device itself is unplugged,
it should not fail because of Ethernet device is already removed.
Fixes: 23ea57a2a0ce ("ethdev: complete closing of port") Cc: stable@dpdk.org Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com> Reviewed-by: Ivan Malov <ivan.malov@oktetlabs.ru> Reported-by: David Marchand <david.marchand@redhat.com> Acked-by: Ferruh Yigit <ferruh.yigit@intel.com>
git.droids-corp.org / dpdk.git / log