l2_len and l3_len fields are related to Tx offloading features.
It is undefined in the DPDK API what those fields may contain if an
application did not request a Tx offload.
Skip reading them if no Tx offloads has been requested.
Fixes: 1e728b01120c ("net/iavf: rework Tx path")
Cc: sta...@dpdk.o
Hi Anoob,
> -Original Message-
> From: Anoob Joseph
> Sent: Wednesday, August 23, 2023 11:17 AM
> To: Gujjar, Abhinandan S
> Cc: Jerin Jacob Kollanukkaran ; Shijith Thotton
> ; dev@dpdk.org; Aakash Sasidharan
> ; Akhil Goyal
> Subject: RE: [PATCH 3/3] test/event_crypto_adapter: fix cryp
The sa to port map is updated when the sa is created. In this patch the
portid field is also added to macsec event callback function.
Signed-off-by: Ankur Dwivedi
---
drivers/net/cnxk/cnxk_ethdev_mcs.c | 14 +-
1 file changed, 13 insertions(+), 1 deletion(-)
diff --git a/drivers/net
Adds sa to port mapping in roc mcs. This helps to propagate the tx and
rx pn threshold events to the correct ethernet device.
Signed-off-by: Ankur Dwivedi
---
drivers/common/cnxk/roc_mcs.c | 22 +++---
drivers/common/cnxk/roc_mcs.h | 8 ++--
drivers/common/cn
This patch series adds support for packet number threshold events to be
received on the respective ethernet port. For this a sa to port mapping
is created during sa creation. When the pn threshold event is received
then the check for port id is made to receive the event on the correct
port.
Ankur
Hi Abhinandan,
Please see inline.
Thanks,
Anoob
> -Original Message-
> From: Gujjar, Abhinandan S
> Sent: Wednesday, August 23, 2023 10:19 AM
> To: Aakash Sasidharan ; Akhil Goyal
>
> Cc: Jerin Jacob Kollanukkaran ; Shijith Thotton
> ; Anoob Joseph ;
> dev@dpdk.org
> Subject: [EXT] RE:
> -Original Message-
> From: Wenjing Qiao
> Sent: Monday, August 21, 2023 5:52 PM
> To: Wu, Jingjing ; Xing, Beilei
> Cc: dev@dpdk.org; Huang, ZhiminX ; Qiao, Wenjing
> ; sta...@dpdk.org
> Subject: [PATCH v4] net/iavf: fix VLAN offload strip flag inconsistency
>
> For i40e kernel driv
> -Original Message-
> From: Aakash Sasidharan
> Sent: Tuesday, August 22, 2023 6:47 PM
> To: Gujjar, Abhinandan S ; Akhil Goyal
>
> Cc: jer...@marvell.com; sthot...@marvell.com; ano...@marvell.com;
> dev@dpdk.org; asasidha...@marvell.com
> Subject: [PATCH 3/3] test/event_crypto_adapte
> -Original Message-
> From: Aakash Sasidharan
> Sent: Tuesday, August 22, 2023 6:47 PM
> To: Gujjar, Abhinandan S
> Cc: jer...@marvell.com; sthot...@marvell.com; ano...@marvell.com;
> dev@dpdk.org; asasidha...@marvell.com
> Subject: [PATCH 2/3] test/event_crypto_adapter: skip unsuppor
>
> Skip the test if rte_event_crypto_adapter_caps_get()
> returns -ENOTSUP.
>
> Signed-off-by: Aakash Sasidharan
> ---
> app/test/test_event_crypto_adapter.c | 17 +
> 1 file changed, 17 insertions(+)
>
Acked-by: Anoob Joseph
> Create crypto null device only if no other crypto devices were found.
>
> Fixes: 3c2c535ecfc0 ("test: add event crypto adapter auto-test")
> Cc: abhinandan.guj...@intel.com
>
> Signed-off-by: Aakash Sasidharan
> ---
> app/test/test_event_crypto_adapter.c | 7 +++
> 1 file changed, 3 inser
>
> crypto adapter capability now returns -ENOTSUP instead of -EINVAL when
[Anoob] Better to start with caps (crypto -> Crypto). May be reword to,
When event_cnxk or crypto_cnxk PMDs are not available, return -ENOTSUP instead
of -EINVAL.
> event_cnxk or crypto_cnxk device is not available. Thi
Good evening,
Tomorrow is our biweekly Tech Board Meeting.
A read-only agenda can be found here:
https://annuel.framapad.org/p/r.0c3cc4d1e011214183872a98f6b5c7db
And our login link:
https://meet.jit.si/dpdk
See you soon
Thanks,
Nathan
Nathan C. Southern, Project Coordinator
Data Plane Deve
22/08/2023 08:27, Feifei Wang пишет:
Define specific function implementation for i40e driver.
Currently, mbufs recycle mode can support 128bit
vector path and avx2 path. And can be enabled both in
fast free and no fast free mode.
Suggested-by: Honnappa Nagarahalli
Signed-off-by: Feifei Wang
Re
Hi Feifei,
Add 'rte_eth_recycle_rx_queue_info_get' and 'rte_eth_recycle_mbufs'
APIs to recycle used mbufs from a transmit queue of an Ethernet device,
and move these mbufs into a mbuf ring for a receive queue of an Ethernet
device. This can bypass mempool 'put/get' operations hence saving CPU
cy
18/08/2023 14:48, Bruce Richardson пишет:
On Fri, Aug 18, 2023 at 02:25:14PM +0100, Ferruh Yigit wrote:
On 8/17/2023 3:18 PM, Konstantin Ananyev wrote:
Caution: This message originated from an External Source. Use proper caution
when opening attachments, clicking links, or responding.
On We
On Fri, 16 Jul 2021 15:27:59 +0100
Ferruh Yigit wrote:
> @@ -271,6 +319,10 @@ static const struct eth_dev_ops
> virtual_ethdev_default_dev_ops = {
> .dev_stop = virtual_ethdev_stop,
> .dev_close = virtual_ethdev_close,
> .dev_infos_get = virtual_ethdev_info_get,
> + .rx_que
On Fri, 16 Jul 2021 15:28:00 +0100
Ferruh Yigit wrote:
> +/**
> + * Device flags set on ``eth_dev->data->dev_flags`` by drivers.
> + * These values can be received via ``rte_eth_dev_info_get()``
> + */
Already done in later patches.
> @@ -3036,7 +3050,7 @@ int rte_eth_macaddr_get(uint16_t port_i
Looking at this old patch.
On Fri, 16 Jul 2021 15:27:54 +0100
Ferruh Yigit wrote:
> diff --git a/app/test/virtual_pmd.c b/app/test/virtual_pmd.c
> index 6098e633f35a..17f28c5a304c 100644
> --- a/app/test/virtual_pmd.c
> +++ b/app/test/virtual_pmd.c
> @@ -355,8 +355,8 @@ virtual_ethdev_rx_burst_s
Adapt bpf for EAL optional atomics API changes
Signed-off-by: Tyler Retzlaff
Reviewed-by: Morten Brørup
Acked-by: Konstantin Ananyev
---
lib/bpf/bpf_pkt.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/lib/bpf/bpf_pkt.c b/lib/bpf/bpf_pkt.c
index ffd2db7..7a8e4a6 1006
Refrain from using compiler __atomic_xxx builtins DPDK now requires
the use of rte_atomic__explicit macros when operating on DPDK
atomic variables.
Signed-off-by: Tyler Retzlaff
Suggested-by: Morten Brørup
Acked-by: Bruce Richardson
Acked-by: Morten Brørup
Acked-by: Konstantin Ananyev
---
de
Adapt the EAL public headers to use rte optional atomics API instead of
directly using and exposing toolchain specific atomic builtin intrinsics.
Signed-off-by: Tyler Retzlaff
Reviewed-by: Morten Brørup
Acked-by: Konstantin Ananyev
---
app/test/test_mcslock.c| 6 ++--
lib/eal/
Adapt distributor for EAL optional atomics API changes
Signed-off-by: Tyler Retzlaff
Reviewed-by: Morten Brørup
Acked-by: Konstantin Ananyev
---
lib/distributor/distributor_private.h| 2 +-
lib/distributor/rte_distributor_single.c | 44
2 files changed, 23
Introduce __rte_atomic qualifying casts in rte_optional atomics inline
functions to prevent cascading the need to pass __rte_atomic qualified
arguments.
Warning, this is really implementation dependent and being done
temporarily to avoid having to convert more of the libraries and tests in
DPDK in
This series introduces API additions prefixed in the rte namespace that allow
the optional use of stdatomics.h from C11 using enable_stdatomics=true for
targets where enable_stdatomics=false no functional change is intended.
Be aware this does not contain all changes to use stdatomics across the D
Provide API for atomic operations in the rte namespace that may
optionally be configured to use C11 atomics with meson
option enable_stdatomics=true
Signed-off-by: Tyler Retzlaff
Reviewed-by: Morten Brørup
Acked-by: Konstantin Ananyev
---
config/meson.build | 1 +
doc/api
Recheck-request: iol-sample-apps-testing,
This email is a test of the community lab's retesting system, please
ignore. Testing superseded reply on a patch.
On Thu, Aug 3, 2023 at 3:51 AM David Marchand
wrote:
> Define some PCI command constants and use them in existing drivers.
>
> Signed-off-b
On Fri, Aug 18, 2023 at 09:13:29AM +0200, Morten Brørup wrote:
> > From: Tyler Retzlaff [mailto:roret...@linux.microsoft.com]
> > Sent: Thursday, 17 August 2023 21.14
> >
> > On Thu, Aug 17, 2023 at 01:57:01PM +0200, Morten Brørup wrote:
> > > > From: Tyler Retzlaff [mailto:roret...@linux.microsof
In case of failure of a test in macsec autotest,
the buffers were not getting cleaned.
Added appropriate code to clean the buffers.
Fixes: 993ea577a006 ("test/security: add inline MACsec cases")
Cc: sta...@dpdk.org
Signed-off-by: Akhil Goyal
---
app/test/test_security_inline_macsec.c | 70 +
As done in FLR, similar to CQ disable, in process of
aura disable we needs explicitly clear BP_ENA in order to
deassert backpressure if it was earlier asserted.
Signed-off-by: Nithin Dabilpuram
Signed-off-by: Ashwin Sekhar T K
---
drivers/common/cnxk/roc_npa.c | 2 ++
1 file changed, 2 insertio
From: Sibaranjan Pattnayak
Support for CNF10KA SoC by adding its PCI device ID.
Signed-off-by: Sibaranjan Pattnayak
Signed-off-by: Ashwin Sekhar T K
---
drivers/mempool/cnxk/cnxk_mempool.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/drivers/mempool/cnxk/cnxk_mempool.c
b/drivers/mem
For non-EAL pthreads, rte_lcore_id() will not be valid.
So, batch allocation cannot be used as we won't have a
dedicated alloc buffer for the thread. So, fallback to
bulk alloc in such cases.
Fixes: 91531e63f43b ("mempool/cnxk: add cn10k batch dequeue")
Signed-off-by: Ashwin Sekhar T K
---
driv
From: Harman Kalra
For non-EAL pthreads, rte_lcore_id() will not be valid.
So, batch free cannot be used as those threads won't have
dedicated lmtlines. So, fallback to bulk alloc in such cases.
Fixes: ecbc731a2286 ("mempool/cnxk: add cn10k batch enqueue")
Signed-off-by: Harman Kalra
Signed-of
Currently mempool_cnxk driver uses asynchronous allocation
for all pools. This asynchronous allocation can result in
local caching of additional 512 objects on a single core
even when cache is disabled. This will eventually lead to
starvation on pools where the number of objects is very less.
This
Currently mempool_cnxk driver uses asynchronous allocation
for all pools. This asynchronous allocation can result in
local caching of additional 512 objects on a single core
even when cache is disabled. This will eventually lead to
starvation on pools where the number of objects is very less.
This
Recheck-request: iol-sample-apps-testing,
This is a test email of the retesting framework, please ignore.
On Thu, Aug 3, 2023 at 3:50 AM David Marchand
wrote:
> Rather than rely on Linux headers to find some PCI(e) standard constants
> or reinvent the same PCI capability helper, this series com
> >> Subject: Re: [RFC] lib/st_ring: add single thread ring
> >>
> >> On 2023-08-21 08:04, Honnappa Nagarahalli wrote:
> >>> Add a single thread safe and multi-thread unsafe ring data structure.
> >>
> >> One must have set the bar very low, if one needs to specify that an
> >> API is single-threa
Recheck-request: iol-sample-apps-testing
This is a test email of the retesting framework, please ignore. Testing the
behavior of replying to a superseded patch.
On Thu, Aug 3, 2023 at 3:50 AM David Marchand
wrote:
> Rather than rely on Linux headers to find some PCI(e) standard constants
> or r
On Tue, Aug 22, 2023 at 10:55 AM Olivier Matz wrote:
>
> Hi David,
>
> On Tue, Aug 22, 2023 at 09:32:44AM +0200, David Marchand wrote:
> > The L4 checksum offloading API does not require l4_len to be set.
> > Make the driver discover the L4 headers size by itself.
> >
> > Fixes: 6546e76056e3 ("net
Recheck-request: iol-unit-arm64-testing, iol-broadcom-Performance,
iol-compile-amd64-testing, iol-sample-apps-testing
This email is just a test of the new email based retesting request
framework.
Add support for matching "ttl" field in MPLS item.
Signed-off-by: Michael Baum
---
app/test-pmd/cmdline_flow.c | 9 +
1 file changed, 9 insertions(+)
diff --git a/app/test-pmd/cmdline_flow.c b/app/test-pmd/cmdline_flow.c
index 94827bcc4a..95c0a19beb 100644
--- a/app/test-pmd/cmdline_flo
Hi Ivan,
On Tue, 22 Aug 2023, Ivan Malov wrote:
>
> Hi Michael,
>
> Very interesting proposal. However, could you please provide a use case
> example for this feature
I can think about 2 different use cases:
1. Sampling - when application wants to sample certain percentage of the flow,
it c
Create crypto null device only if no other crypto devices
were found.
Fixes: 3c2c535ecfc0 ("test: add event crypto adapter auto-test")
Cc: abhinandan.guj...@intel.com
Signed-off-by: Aakash Sasidharan
---
app/test/test_event_crypto_adapter.c | 7 +++
1 file changed, 3 insertions(+), 4 deleti
Skip the test if rte_event_crypto_adapter_caps_get()
returns -ENOTSUP.
Signed-off-by: Aakash Sasidharan
---
app/test/test_event_crypto_adapter.c | 17 +
1 file changed, 17 insertions(+)
diff --git a/app/test/test_event_crypto_adapter.c
b/app/test/test_event_crypto_adapter.c
ind
crypto adapter capability now returns -ENOTSUP instead
of -EINVAL when event_cnxk or crypto_cnxk device is not
available. This would allow unit test to be skipped
when the devices are not available.
Fixes: 19f81cb59cb4 ("event/cnxk: add crypto adapter operations")
Signed-off-by: Aakash Sasidharan
From: Pavan Nikhilesh
Add support to get the remaining ticks to expire for a
given event timer.
Signed-off-by: Pavan Nikhilesh
---
doc/guides/rel_notes/release_23_11.rst | 5
drivers/event/cnxk/cnxk_tim_evdev.c| 1 +
drivers/event/cnxk/cnxk_tim_evdev.h| 3 +++
drivers/event/cnx
Hi Michael,
Very interesting proposal. However, could you please
provide a use case example for this feature and also
provide an example of where such value could come
from, theoretically.
Thank you.
On Tue, 22 Aug 2023, Michael Baum wrote:
Add support for matching random value using new "rte
On Tue, Aug 22, 2023 at 2:12 PM Mattias Rönnblom wrote:
>
> On 2023-08-18 08:09, Jerin Jacob wrote:
> > On Fri, Jun 16, 2023 at 1:17 PM Mattias Rönnblom
> > wrote:
> >>
> >
> > Hi Mattais,
> >
> > Finally, got some time to review this series, and thanks for excellent
> > documentation.
> >
> > I
From: Erez Shitrit
The HW adds a random number per each hash, this value can be used for
statistic calculation over the packets, for example by setting one bit in
the mask of that filed we will get half of the traffic in the flow, and
so on with the rest of the mask.
Signed-off-by: Erez Shitrit
Add support for random item in HWS mode.
Signed-off-by: Michael Baum
---
doc/guides/nics/features/mlx5.ini | 1 +
doc/guides/nics/mlx5.rst | 10 +-
doc/guides/rel_notes/release_23_11.rst | 4
drivers/net/mlx5/mlx5_flow_dv.c| 5 +
drivers/net/mlx5/ml
Add support for matching random value using the "rte_flow_item_random"
structure.
Depends-on: series-29307 ("ethdev: add random item support")
Erez Shitrit (1):
net/mlx5/hws: add support for random number match
Michael Baum (1):
net/mlx5: add random item support
doc/guides/nics/features/ml
Minutes of Technical Board Meeting, 2023-August -9
Members Attending
-
-Aaron
-Bruce
-Hemant
-Honnappa
-Jerin (Chair)
-Konstantin
-Stephen
-Thomas
-Tyler
-Morten
NOTE: The technical board meetings every second Wednesday at
https://meet.jit.si/DPDK at 3 pm UTC.
Meetings are publ
> -Original Message-
> From: Zhang, Qi Z
> Sent: Tuesday, August 22, 2023 3:59 PM
> To: 'David Marchand'
> Cc: dev@dpdk.org; echau...@redhat.com; m...@redhat.com;
> sta...@dpdk.org; Wu, Jingjing ; Xing, Beilei
> ; Doherty, Declan ; Sinha,
> Abhijit ; Nicolau, Radu
> Subject: RE: [PATCH]
Add support for random item, usage example:
pattern random spec value 0x1 mask value 0x3 / eth / end
Flow rule with above pattern matching 25% of the traffic, it hits only
when random value suffix is "01" and miss the others ("00", "10", "11").
Signed-off-by: Michael Baum
---
app/test-
Add support for a new item type "RTE_FLOW_ITEM_TYPE_RANDOM".
This item enables to match on some random value as a part of flow rule.
Signed-off-by: Michael Baum
---
doc/guides/nics/features/default.ini | 1 +
doc/guides/prog_guide/rte_flow.rst | 11 +
doc/guides/rel_notes/release_
Add support for matching random value using new "rte_flow_item_random"
structure.
This random value is not based on the packet data/headers.
Application shouldn't assume that this value is kept during the life
time of the packet.
Michael Baum (2):
ethdev: add random item support
app/testpmd: a
Hi David,
On Tue, Aug 22, 2023 at 09:32:44AM +0200, David Marchand wrote:
> The L4 checksum offloading API does not require l4_len to be set.
> Make the driver discover the L4 headers size by itself.
>
> Fixes: 6546e76056e3 ("net/tap: calculate checksums of multi segs packets")
> Cc: sta...@dpdk.
On 2023-08-18 08:09, Jerin Jacob wrote:
On Fri, Jun 16, 2023 at 1:17 PM Mattias Rönnblom
wrote:
The purpose of the event dispatcher is to help reduce coupling in an
Eventdev-based DPDK application.
In addition, the event dispatcher also provides a convenient and
flexible way for the applicati
Hello Julien,
On Tue, Aug 22, 2023 at 08:34:53AM +0200, jhascoet wrote:
> From: Julien Hascoet
>
> In case of ring full state, we retry the enqueue
> operation in order to avoid mbuf loss.
>
> Fixes: af75078fece ("first public release")
>
> Signed-off-by: Julien Hascoet
> ---
> app/test/test
From: Julien Hascoet
In case of ring full state, we retry the enqueue
operation in order to avoid mbuf loss.
Fixes: af75078fece ("first public release")
Signed-off-by: Julien Hascoet
---
app/test/test_mbuf.c | 15 ---
1 file changed, 12 insertions(+), 3 deletions(-)
diff --git a/
Hello DPDK mentors,
We are facing an issue, KNI core stuck after 4-5 days of traffic. Can
anybody guide us on this?
We have set isolcpu, rcu_nocbs in grub for given cores.
Aug 22 01:38:29 kernel: INFO: rcu_sched self-detected stall on CPU { 2}
(t=60001 jiffies g=426362199 c=426362198 q=150932)
On 2023-08-22 07:43, Honnappa Nagarahalli wrote:
-Original Message-
From: Mattias Rönnblom
Sent: Monday, August 21, 2023 4:14 PM
To: Honnappa Nagarahalli ;
jack...@nvidia.com; konstantin.v.anan...@yandex.ru
Cc: dev@dpdk.org; Ruifeng Wang ; Aditya
Ambadipudi ; Wathsala Wathawana
Vithan
> -Original Message-
> From: David Marchand
> Sent: Tuesday, August 22, 2023 3:40 PM
> To: Zhang, Qi Z
> Cc: dev@dpdk.org; echau...@redhat.com; m...@redhat.com;
> sta...@dpdk.org; Wu, Jingjing ; Xing, Beilei
> ; Doherty, Declan ; Sinha,
> Abhijit ; Nicolau, Radu
> Subject: Re: [PATCH]
The cn10k cross build file sets cpu to 'armv8.6-a' while
N2 is armv8.5-a arch.
The cpu field in the cross file doesn't take effect as
config/arm/meson.build controls machine_args for march.
Then correct the value in arm cross files to 'auto'.
Signed-off-by: Joyce Kong
Reviewed-by: Ruifeng Wang
-
On Tue, Aug 22, 2023 at 9:33 AM Zhang, Qi Z wrote:
> > If the driver reads l2_len or l3_len, this is an undefined behavior:
> > for example, OVS might have been using l2_len or l3_len for its internal
> > uses
> > (though I agree it would be risky for an application to do so).
> >
> > We probably
> -Original Message-
> From: David Marchand
> Sent: Tuesday, August 22, 2023 2:12 PM
> To: Zhang, Qi Z
> Cc: dev@dpdk.org; echau...@redhat.com; m...@redhat.com;
> sta...@dpdk.org; Wu, Jingjing ; Xing, Beilei
> ; Doherty, Declan ; Sinha,
> Abhijit ; Nicolau, Radu
> Subject: Re: [PATCH]
Hi, Ferruh
Would you please give some comments on these patches?
If no comments, would mbufs recycle mode be merged in dpdk-next branch?
Thanks very much.
Best Regards
Feifei
> -Original Message-
> From: Feifei Wang
> Sent: Tuesday, August 22, 2023 3:27 PM
> Cc: dev@dpdk.org; nd ; Feif
The L4 checksum offloading API does not require l4_len to be set.
Make the driver discover the L4 headers size by itself.
Fixes: 6546e76056e3 ("net/tap: calculate checksums of multi segs packets")
Cc: sta...@dpdk.org
Signed-off-by: David Marchand
Tested-by: Ales Musil
---
.mailmap
Define specific function implementation for ixgbe driver.
Currently, recycle buffer mode can support 128bit
vector path. And can be enabled both in fast free and
no fast free mode.
Suggested-by: Honnappa Nagarahalli
Signed-off-by: Feifei Wang
Reviewed-by: Ruifeng Wang
Reviewed-by: Honnappa Naga
Add recycle mbufs engine for testpmd. This engine forward pkts with
I/O forward mode. But enable mbufs recycle feature to recycle used
txq mbufs for rxq mbuf ring, which can bypass mempool path and save
CPU cycles.
Suggested-by: Jerin Jacob
Signed-off-by: Feifei Wang
Reviewed-by: Ruifeng Wang
-
Define specific function implementation for i40e driver.
Currently, mbufs recycle mode can support 128bit
vector path and avx2 path. And can be enabled both in
fast free and no fast free mode.
Suggested-by: Honnappa Nagarahalli
Signed-off-by: Feifei Wang
Reviewed-by: Ruifeng Wang
Reviewed-by: H
Add 'rte_eth_recycle_rx_queue_info_get' and 'rte_eth_recycle_mbufs'
APIs to recycle used mbufs from a transmit queue of an Ethernet device,
and move these mbufs into a mbuf ring for a receive queue of an Ethernet
device. This can bypass mempool 'put/get' operations hence saving CPU
cycles.
For eac
Currently, the transmit side frees the buffers into the lcore cache and
the receive side allocates buffers from the lcore cache. The transmit
side typically frees 32 buffers resulting in 32*8=256B of stores to
lcore cache. The receive side allocates 32 buffers and stores them in
the receive side
73 matches
Mail list logo
