On Tue, May 2, 2023 at 3:51 PM wrote:
>
> From: Jerin Jacob
>
> Fix the following build issue by initializing req to NULL for
> the local variable.
>
> In function 'thread_msg_handle', inlined from 'thread_main' at
> ../examples/ip_pipeline/thread.c:3130:6:
> ../examples/ip_pipeline/thread.c:535:
On Thu, May 4, 2023 at 10:54 AM wrote:
>
> From: Jerin Jacob
>
> Fix the following build issue by not allowing nb_ids to be zero.
> nb_ids can be zero based on rte_rawdev_xstats_get() API
> documentation but it is not valid for the case when second
> argument is NULL.
>
> examples/ntb/ntb_fwd.c:
> From: Mattias Rönnblom [mailto:mattias.ronnb...@ericsson.com]
> Sent: Thursday, 11 May 2023 08.43
>
> Use non-burst event enqueue and dequeue calls from burst enqueue and
> dequeue only when the burst size is compile-time constant (and equal
> to one).
>
> Signed-off-by: Mattias Rönnblom
> ---
> -Original Message-
> From: Kevin Traynor
> Sent: Monday, May 8, 2023 11:24 PM
> To: Xu, HailinX ; sta...@dpdk.org
> Cc: Stokes, Ian ; Mcnamara, John
> ; Luca Boccassi ; Xu, Qian Q
> ; Thomas Monjalon ; Peng,
> Yuan ; Chen, Zhaoyan ;
> dev@dpdk.org
> Subject: Re: 21.11.4 patches review an
Add new rte item to match the infiniband BTH in RoCE packets.
Dong Zhou (3):
ethdev: add flow item for RoCE infiniband BTH
net/mlx5: add support for infiniband BTH match
net/mlx5/hws: add support for infiniband BTH match
app/test-pmd/cmdline_flow.c | 58 +++
doc/gu
IB(InfiniBand) is one type of networking used in high-performance
computing with high throughput and low latency. Like Ethernet,
IB defines a layered protocol (Physical, Link, Network, Transport
Layers). IB provides native support for RDMA(Remote DMA), an
extension of the DMA that allows direct acc
This patch adds support to match opcode and dst_qp fields in
infiniband BTH. Currently, only the RoCEv2 packet is supported,
the input BTH match item is defaulted to match one RoCEv2 packet.
Signed-off-by: Dong Zhou
---
drivers/common/mlx5/mlx5_prm.h | 5 +-
drivers/net/mlx5/mlx5_flow.h|
This patch adds support to match opcode and dst_qp fields in
infiniband BTH. Currently, only the RoCEv2 packet is supported,
the input BTH match item is defaulted to match one RoCEv2 packet.
Signed-off-by: Dong Zhou
---
drivers/net/mlx5/hws/mlx5dr_definer.c | 76 ++-
driv
Hi David,
> -Original Message-
> From: David Marchand
> Sent: Thursday, March 23, 2023 1:05 AM
> To: dev@dpdk.org
> Cc: sta...@dpdk.org; Maxime Coquelin ; Xia,
> Chenbo ; Yuanhan Liu
> Subject: [PATCH] vhost: avoid sleeping under mutex
>
> Covscan reported:
>
> 2. dpdk-21.11/lib/vhost
https://bugs.dpdk.org/show_bug.cgi?id=1228
Bug ID: 1228
Summary: [dpdk-21.11.4]pvp_qemu_multi_paths_port_restart:perf_p
vp_qemu_vector_rx_mac: performance drop about 23.5%
when send small packets
Product: DPDK
Ver
> -Original Message-
> From: Boleslav Stankevich
> Sent: Wednesday, March 22, 2023 6:23 PM
> To: dev@dpdk.org
> Cc: Boleslav Stankevich ;
> sta...@dpdk.org; Andrew Rybchenko ; Maxime
> Coquelin ; Xia, Chenbo ;
> Jianfeng Tan ; Yuanhan Liu ;
> Tetsuya Mukawa ; Huawei Xie inc.com>
> Subject
> -Original Message-
> From: Boleslav Stankevich
> Sent: Wednesday, March 22, 2023 6:23 PM
> To: dev@dpdk.org
> Cc: Boleslav Stankevich ;
> sta...@dpdk.org; Andrew Rybchenko ; Maxime
> Coquelin ; Xia, Chenbo ;
> David Marchand ; Hyong Youb Kim
> ; Harman Kalra
> Subject: [PATCH 1/2] net/v
Use non-burst event enqueue and dequeue calls from burst enqueue and
dequeue only when the burst size is compile-time constant (and equal
to one).
Signed-off-by: Mattias Rönnblom
---
v2: Wrap builtin call in __extension__, to avoid compiler warnings if
application is compiled with -pedantic
> From: Mattias Rönnblom [mailto:mattias.ronnb...@ericsson.com]
> Sent: Thursday, 11 May 2023 10.17
>
> Use non-burst event enqueue and dequeue calls from burst enqueue and
> dequeue only when the burst size is compile-time constant (and equal
> to one).
>
> Signed-off-by: Mattias Rönnblom
Acke
This patch is to collect empty poll of 'rte_eth_rx_burst' functions in
dpdk l3fwd application. Empty poll means Rx burst function receives no
pkts in one loop.
Furthermore, we also add 'nic_xstats_display' API to show NIC counters.
Usage:
With this patch, no special settings, just run l3fwd, and
Hi Tianyuan,
> -Original Message-
> From: suntianyuan
> Sent: Tuesday, March 21, 2023 2:24 PM
> To: maxime.coque...@redhat.com; Xia, Chenbo
> Cc: dev@dpdk.org; suntianyuan
> Subject: [PATCH] vhost: fix crash on vhost-user client port deletion
>
> The rte_vhost_driver_unregister() and v
On 2023-05-11 09:09, Morten Brørup wrote:
>> From: Mattias Rönnblom [mailto:mattias.ronnb...@ericsson.com]
>> Sent: Thursday, 11 May 2023 08.43
>>
>> Use non-burst event enqueue and dequeue calls from burst enqueue and
>> dequeue only when the burst size is compile-time constant (and equal
>> to on
Use non-burst event enqueue and dequeue calls from burst enqueue and
dequeue only when the burst size is compile-time constant (and equal
to one).
Signed-off-by: Mattias Rönnblom
---
v3: Actually include the change v2 claimed to contain.
v2: Wrap builtin call in __extension__, to avoid compiler
On Thu, May 11, 2023 at 1:55 PM Feifei Wang wrote:
>
> This patch is to collect empty poll of 'rte_eth_rx_burst' functions in
> dpdk l3fwd application. Empty poll means Rx burst function receives no
> pkts in one loop.
>
> Furthermore, we also add 'nic_xstats_display' API to show NIC counters.
>
>
On Fri, May 5, 2023 at 4:07 PM Bruce Richardson
wrote:
>
> On Thu, May 04, 2023 at 02:37:45PM +0200, Juraj Linkeš wrote:
> > Augment the meson build system with dts api generation. The api docs are
> > generated from Python docstrings in DTS using Sphinx. The format of
> > choice is the Google for
> -Original Message-
> From: Jerin Jacob
> Sent: Thursday, May 11, 2023 4:51 PM
> To: Feifei Wang
> Cc: dev@dpdk.org; nd ; Lijian Zhang
> ; Ruifeng Wang ;
> Honnappa Nagarahalli
> Subject: Re: [PATCH] examples/l3fwd: add hard code to collect empty poll and
> NIC counters
>
> On Thu, M
> -Original Message-
> From: Jiang, Cheng1
> Sent: Monday, May 8, 2023 8:23 PM
> To: Maxime Coquelin ; Xia, Chenbo
>
> Cc: dev@dpdk.org; Hu, Jiayu ; Ding, Xuan
> ; Ma, WenwuX ; Wang, YuanX
> ; He, Xingguang ; David
> Marchand
> Subject: RE: [PATCH 0/2] vhost: add port mirroring function
> From: Mattias Rönnblom [mailto:mattias.ronnb...@ericsson.com]
> Sent: Thursday, 11 May 2023 10.24
>
> Use non-burst event enqueue and dequeue calls from burst enqueue and
> dequeue only when the burst size is compile-time constant (and equal
> to one).
>
> Signed-off-by: Mattias Rönnblom
Acke
> From: Mattias Rönnblom [mailto:mattias.ronnb...@ericsson.com]
> Sent: Thursday, 11 May 2023 10.30
>
> On 2023-05-11 09:09, Morten Brørup wrote:
[...]
> > In my experience using __builtin_constant_p(), you need
> __extension__(__builtin_constant_p(nb_events)) to avoid warnings about using
> thi
Augment the meson build system with dts api generation. The api docs are
generated from Python docstrings in DTS using Sphinx. The format of
choice is the Google format [0].
The guides html sphinx configuration is used to preserve the same style.
The build requires the same Python version and dep
sphinx-build only imports the Python modules when building the
documentation; it doesn't run DTS. This requires changes that make the
code importable without running it. This means:
* properly guarding argument parsing in the if __name__ == '__main__'
block.
* the logger used by DTS runner underw
Sphinx imports every Python module when generating documentation from
docstrings, meaning all dts dependencies, including Python version,
must be satisfied.
By adding Sphinx to dts dependencies we make sure that the proper
Python version and dependencies are used when Sphinx is executed.
Signed-of
The tool used to generate developer docs is sphinx, which is already
used in DPDK. The configuration is kept the same to preserve the style.
Sphinx generates the documentation from Python docstrings. The docstring
format most suitable for DTS seems to be the Google format [0] which
requires the sp
WIP: only one module is reformatted to serve as a demonstration.
The google format is documented here [0].
[0]: https://google.github.io/styleguide/pyguide.html
Signed-off-by: Juraj Linkeš
---
dts/framework/testbed_model/node.py | 152 +++-
1 file changed, 103 insertion
Hi Thomas, dev group:
After I sent out this patch, I can't find it at the patch site:
https://patches.dpdk.org/project/dpdk/list/?param=2&page=1
I would like to ask if there's anything I should do first here.
Thanks,
Best Regards,
Zijin Tao(Trevor Tao, 陶孜谨)
ARM Electronic Technology (Shanghai)
On 5/11/23 10:59, Xia, Chenbo wrote:
-Original Message-
From: Jiang, Cheng1
Sent: Monday, May 8, 2023 8:23 PM
To: Maxime Coquelin ; Xia, Chenbo
Cc: dev@dpdk.org; Hu, Jiayu ; Ding, Xuan
; Ma, WenwuX ; Wang, YuanX
; He, Xingguang ; David
Marchand
Subject: RE: [PATCH 0/2] vhost: add po
>
> On Thu, May 11, 2023 at 1:55 PM Feifei Wang
> wrote:
> >
> > This patch is to collect empty poll of 'rte_eth_rx_burst' functions in
> > dpdk l3fwd application. Empty poll means Rx burst function receives no
> > pkts in one loop.
> >
> > Furthermore, we also add 'nic_xstats_display' API to s
Release status meeting minutes 2023-05-11
=
Agenda:
* Release Dates
* Subtrees
* Roadmaps
* LTS
* Defects
* Opens
Participants:
* AMD
* ARM
* Intel
* Marvell
* Nvidia
* Red Hat
Release Dates
-
The following are the proposed current dates for
From: Saoirse O'Donovan
Allocated memory was being freed using 'free' when it should have been
freed using 'OPENSSL_free'. This has now been modified so that the
correct method is used to free allocated memory.
Coverity issue: 384415
Fixes: 4c7ae22f1f83 ("crypto/openssl: update DSA routine with
https://bugs.dpdk.org/show_bug.cgi?id=1229
Bug ID: 1229
Summary: rte_mempool_avail_count : returns bigger than mempool
size
Product: DPDK
Version: 22.11
Hardware: x86
OS: Linux
Status: UNCONFIRME
On 5/10/2023 1:58 PM, Nipun Gupta wrote:
On 4/24/2023 8:52 PM, David Marchand wrote:
Hello Anatoly,
On Wed, Apr 5, 2023 at 4:17 PM Burakov, Anatoly
wrote:
Could you please provide some steps to reproduce the hotplug issue
you're having? It would be great to have a test case for this patchs
Acked-by: Kai Ji
> -Original Message-
> From: Power, Ciara
> Sent: Thursday, May 11, 2023 3:00 PM
> To: Ji, Kai
> Cc: dev@dpdk.org; O'Donovan, Saoirse ;
> sta...@dpdk.org; Power, Ciara
> Subject: [PATCH] crypto/openssl: fix memory free issue
>
> From: Saoirse O'Donovan
>
> Allocated
Hi all,
A gentle ping on this series.
It touches a lot of files when syncing the PMD with the kernel driver,
and to avoid conflicts with future work, it would be kind if this could
be looked at early in the release cycle.
On 2023-04-10 19:00:02 +0800, Chaoyong He wrote:
> The DPDK NFP PMD shar
The rte_vhost_driver_unregister() and vhost_user_read_cb() can be
called at the same time by 2 threads. reconn may be added back to
reconn_list by vhost_user_read_cb() after rte_vhost_driver_unregister()
removed from reconn_list. Then rte_vhost_driver_unregister free vsocket,
cause vhost_user_clien
On Thu, 11 May 2023 16:25:19 +0800
Feifei Wang wrote:
> This patch is to collect empty poll of 'rte_eth_rx_burst' functions in
> dpdk l3fwd application. Empty poll means Rx burst function receives no
> pkts in one loop.
>
> Furthermore, we also add 'nic_xstats_display' API to show NIC counters.
The tracepoints added are used to track lcore role and status,
as well as service mapping and service runstates. These
tracepoints are then used in analyses in Trace Compass.
Signed-off-by: Arnaud Fiorini
---
.mailmap | 1 +
doc/guides/prog_guide/service_cores.rs
On Thu, 11 May 2023 20:16:28 +
Arnaud Fiorini wrote:
> +The service core library is instrumented with tracepoints using the DPDK
> Trace
> +Library. These tracepoints allow you to track the service and logical cores
> +state. The tracepoints that are on the fast path are compiled out by defa
On 2023-05-11 4:50 p.m., Stephen Hemminger wrote:
On Thu, 11 May 2023 20:16:28 +
Arnaud Fiorini wrote:
+The service core library is instrumented with tracepoints using the DPDK Trace
+Library. These tracepoints allow you to track the service and logical cores
+state. The tracepoints that a
On Thu, 11 May 2023 17:01:48 -0400
Arnaud Fiorini wrote:
> On 2023-05-11 4:50 p.m., Stephen Hemminger wrote:
> > On Thu, 11 May 2023 20:16:28 +
> > Arnaud Fiorini wrote:
> >
> >> +The service core library is instrumented with tracepoints using the DPDK
> >> Trace
> >> +Library. These trac
A gentle ping on this series.
There has a large patch series enhance the bonding PMD depends on this series
is waiting for send out, it would be kind if this could be looked at early in
this release cycle.
> -Original Message-
> From: Chaoyong He
> Sent: Wednesday, March 1, 2023 10:49 A
> -Original Message-
> From: Honnappa Nagarahalli
> Sent: Thursday, May 11, 2023 9:32 PM
> To: Jerin Jacob ; Feifei Wang
>
> Cc: dev@dpdk.org; nd ; Lijian Zhang
> ; Ruifeng Wang ; nd
>
> Subject: RE: [PATCH] examples/l3fwd: add hard code to collect empty poll and
> NIC counters
>
>
>
When process VLAN push flow action, the former logic perform the
mask and shift operations directly on the big endian data, which
cause both the pcp and VLAN id are not correct for offloaded packets.
Fix it by convert the data from big endian to CPU endian before
perform the mask and shift operati
In the tx descriptor free logic of nfd3, the former logic might force
cast a negative number into a very big unsigned number, and which will
cause potential problem in the xmit loop.
The xmit loop will continue in the place where it should break, and will
overwrite the Tx descriptor which is not f
Hi Kumar,
For TCP/IPv4 and TCP/IPv6, the TCP layer is the same and the difference is
the IP layer. So the code used for TCP layer needs to be shared among gro_tcp6.c
and gro_tcp4.c. But there are too much code duplication in gro_tcp4.c and
gro_tcp6.c
in current implementation.
For example, stru
Hi Stephen:
Yes, I think there is no point in doing RSS if only a single queue, but for
the 2nd part:
"|| !local_port_conf.rx_adv_conf.rss_conf.rss_hf"
It's an "or" relationship with the dev_info.max_rx_queues==1, and in my case,
the device(virtio) information showed that
the max_rx_queues is n
The mixed test cases can have keys with different key lengths. The
routine which prepares the session parameters uses same key length for
both cipher & auth keys. Instead allow the caller to use same keys as
required.
Signed-off-by: Anoob Joseph
---
app/test/test_cryptodev.c | 48 +++
Keys are not required for NULL algorithms. Same way IV, digest lengths
should also be set to 0. The values are invalid and any PMD which
validates such parameters would return "-ENOTSUP" for such cases which
would result in false skipping of tests.
Signed-off-by: Anoob Joseph
---
app/test/test_c
52 matches
Mail list logo
