Introduce enhancements and fixes for flow-perf:
1- Fix decap data for raw-decap actions.
2- Fix setting IPv6 source.
3- Add first flow latency calculation.
4- fixes in flow-perf: CPU and reporting.
5- Use more generic wrapper for cycles.
6- Use unique data on the fly.
--
v2:
* Add first insert
rdtsc() is x86 related, while this might fail for other archs,
so it's better to use more generic API for cycles measurement.
Signed-off-by: Wisam Jaddo
Acked-by: Alexander Kozyrev
---
app/test-flow-perf/main.c | 24
1 file changed, 12 insertions(+), 12 deletions(-)
di
Current support for unique data is to compile with config.h
var as FIXED_VALUES as 0, and this is only supported on
compilation time, as a result the user may use only single
mode for each compilation.
Starting with this commit the user will have the ability to
use this feature on the fly by using
create_flow and create_meter are not correct names since those
are records that contain creation and deletion, which makes
them more of a record for such data.
Fixes: d8099d7ecbd0 ("app/flow-perf: split dump functions")
Cc: dongz...@nvidia.com
Signed-off-by: Wisam Jaddo
Acked-by: Alexander Kozyr
Take into consideration that the user may call portmask for
any run, thus the app should always check if port is needed
to collect and report or not.
Fixes: 070316d01d3e ("app/flow-perf: add multi-core rule insertion and
deletion")
Fixes: d8099d7ecbd0 ("app/flow-perf: split dump functions")
Cc: d
Currently the memset() will not set a correct src ip that represent
the incremental value of the counter.
This commit will fix this and each flow will have correct IPv6.src
that it's incremental from previous flow and equal to the decimal
values.
Fixes: bf3688f1e816 ("app/flow-perf: add insertion
Starting from this commit the app will always
report the first flow latency.
This is useful in debugging to check the first
flow insertion before any caching effect.
Signed-off-by: Wisam Jaddo
Acked-by: Alexander Kozyrev
---
app/test-flow-perf/main.c | 11 +++
1 file changed, 11 insert
When using decap actions it's been set to the data to decap
into the encap_data instead of decap_data, as a results we end
up with bad encap and decap data in many cases.
Fixes: 0c8f1f4ab90e ("app/flow-perf: support raw encap/decap actions")
Cc: sta...@dpdk.org
Signed-off-by: Wisam Jaddo
Acked-b
Hi,
We believe we found a behavior inconsistency based of different source mbuf
layout in the QAT PMD code.
file link: https://github.com/DPDK/dpdk/blob/main/drivers/crypto/qat/qat_sym.c
The inconsistent behavior happens when:
* symmetric operation
* out-of-place operation
There are couple of issues when building with Clang:
1. vector is a keyword and should not be used in code. I undefined it,
but it would probably be better to just change the variable name.
2. vector long is deprecated by Clang and should not be used. I switched
here to vector int.
3. Additionally,
Hi All,
I'm having an issue with the following scenario with SPDK-based applications
built w/DPDK
using CryptoDev in a primary/secondary process setup.
Primary Process: An application that does crypto and properly initializes
aesni_mb
Secondary Process: An application that does not use crypto s
This patch adds support for the timestamp format settings for
the receive and send queues. If the firmware version x.30.1000
or above is installed and the NIC timestamps are configured
with the real-time format, the default zero values for newly
added fields cause the queue creation to fail.
The p
This patch handles the NIC-supported timestamp formats via DevX.
Two different timestamp formats can be provided potentially.
The free-running format provides opaque values captured from
the internal clock counter fed by some independent oscillator.
The free-running frequency is not pre-defined and
This patch adds support for the timestamp format settings for
the receive and send queues. If the firmware version x.30.1000
or above is installed and the NIC timestamps are configured
with the real-time format, the default zero values for newly
added fields cause the queue creation to fail.
The p
This patch adds support for the timestamp format settings for
the receive and send queues. If the firmware version x.30.1000
or above is installed and the NIC timestamps are configured
with the real-time format, the default zero values for newly
added fields cause the queue creation to fail.
The p
This patch adds support for the timestamp format settings for
the receive and send queues. If the firmware version x.30.1000
or above is installed and the NIC timestamps are configured
with the real-time format, the default zero values for newly
added fields cause the queue creation to fail.
The p
There are two different timestamp formats can be provided potentially
by mlx5 supported hardware.
The free-running format provides some opaque values captured from
internal clock counter clocked by some independent oscillator.
The free-running frequency is not pre-defined and should be queried
fro
This patch adds support for the timestamp format settings for
the receive and send queues. If the firmware version x.30.1000
or above is installed and the NIC timestamps are configured
with the real-time format, the default zero values for newly
added fields cause the queue creation to fail.
The p
There are two different timestamp formats can be provided potentially
by mlx5 supported hardware.
The free-running format provides some opaque values captured from
internal clock counter clocked by some independent oscillator.
The free-running frequency is not pre-defined and should be queried
fro
This patch adds support for the timestamp format settings for
the receive and send queues. If the firmware version x.30.1000
or above is installed and the NIC timestamps are configured
with the real-time format, the default zero values for newly
added fields cause the queue creation to fail.
The p
This patch adds support for the timestamp format settings for
the receive and send queues. If the firmware version x.30.1000
or above is installed and the NIC timestamps are configured
with the real-time format, the default zero values for newly
added fields cause the queue creation to fail.
The p
This patch handles the NIC-supported timestamp formats via DevX.
Two different timestamp formats can be provided potentially.
The free-running format provides opaque values captured from
the internal clock counter fed by some independent oscillator.
The free-running frequency is not pre-defined and
This patch adds support for the timestamp format settings for
the receive and send queues. If the firmware version x.30.1000
or above is installed and the NIC timestamps are configured
with the real-time format, the default zero values for newly
added fields cause the queue creation to fail.
The p
In cryptography, a block cipher is a deterministic algorithm operating
on fixed-length groups of bits, called blocks.
A block cipher consists of two paired algorithms, one for encryption
and the other for decryption. Both algorithms accept two inputs:
an input block of size n bits and a key of siz
I may have been unclear.
This affects ppc64le overall.
The reason is that vector is a keyword on ppc64(le).
For some reason, GCC does not have an issue with it, but some future version
might.
So the current issues happen when using clang (with glibc or not).
The issue that happens when attempti
Arming periodic timers are not supported by the event timer adapter
right now. This series adds that support. A timer adapter configured in
periodic mode can be used to arm periodic timers.
First patch adds a flag to expose periodic mode capability of an adapter
and flag to configure the adapter
A timer adapter in periodic mode can be used to arm periodic timers.
This patch adds flags used to advertise capability and configure timer
adapter in periodic mode. Capability flag should be set for adapters
which support periodic mode.
Below is a programming sequence on the usage:
/* che
Add tests to arm and cancel periodic timer.
Signed-off-by: Shijith Thotton
---
app/test/test_event_timer_adapter.c | 136 +---
1 file changed, 123 insertions(+), 13 deletions(-)
diff --git a/app/test/test_event_timer_adapter.c
b/app/test/test_event_timer_adapter.c
index
Add support for periodic mode in event timer adapter.
Signed-off-by: Shijith Thotton
---
drivers/event/octeontx2/otx2_tim_evdev.c | 29
drivers/event/octeontx2/otx2_tim_evdev.h | 1 +
2 files changed, 25 insertions(+), 5 deletions(-)
diff --git a/drivers/event/octeontx
On Fri, Mar 12, 2021 at 03:30:11PM +, Carrillo, Erik G wrote:
> Hi Shijith,
>
> Some comments in-line:
>
> > -Original Message-
> > From: Shijith Thotton
> > Sent: Monday, March 8, 2021 2:46 PM
> > To: Carrillo, Erik G
> > Cc: Shijith Thotton ; Pavan Nikhilesh
> > ; Jerin Jacob ;
>
This patch change judgment statements to disable RSS for vf
when users need to disable RSS or RSS hash function configured
is not supported.
Fixes: 95f2f0e9fc2a ("net/iavf: improve default RSS")
Cc: sta...@dpdk.org
Signed-off-by: Wenjun Wu
---
v4: clear existing RSS when rss_hf = 0.
v3: fix the
Updates the documentation for supported sample actions in the NIC Rx
and E-Switch steering flow.
Signed-off-by: Jiawei Wang
Acked-by: Viacheslav Ovsiienko
---
doc/guides/nics/mlx5.rst | 14 +++---
1 file changed, 11 insertions(+), 3 deletions(-)
diff --git a/doc/guides/nics/mlx5.rst b/
Hi,
> -Original Message-
> From: Thomas Monjalon
> Sent: Saturday, March 13, 2021 1:47
> To: dev@dpdk.org
> Cc: Ori Kam ; Ajit Khaparde
> ; Somnath Kotur
> ; Chas Williams ; Min Hu
> (Connor) ; Rahul Lakkireddy
> ; Hemant Agrawal
> ; Sachin Saxena ;
> Guo, Jia ; Wang, Haiyue ;
> Daley, Jo
Minutes of Technical Board Meeting, 2021-02-24
Members Attending: 11/12
- Aaron Conole
- Bruce Richardson
- Ferruh Yigit
- Hemant Agrawal (Chair)
- Honnappa Nagarahalli
- Jerin Jacob
- Kevin Traynor
- Konstantin Ananyev
- Olivier Matz
- Stephen Hemminger
34 matches
Mail list logo