[AMD Official Use Only - General]
Hi Ferruh,
> -Original Message-
> From: Yigit, Ferruh
> Sent: Tuesday, February 27, 2024 9:48 PM
> To: Tummala, Sivaprasad ;
> david.h...@intel.com; anatoly.bura...@intel.com; jer...@marvell.com;
> radu.nico...@intel.com; gak...@marvell.com; cristian.dum
On 2024/2/28 21:07, Ferruh Yigit wrote:
> On 2/28/2024 2:27 AM, huangdengdui wrote:
>>
>>
>> On 2024/2/27 0:43, Ferruh Yigit wrote:
>>> On 2/26/2024 3:16 AM, Jie Hai wrote:
On 2024/2/23 21:53, Ferruh Yigit wrote:
> On 2/20/2024 3:58 AM, Jie Hai wrote:
>> Hi, Ferruh,
>>
>> Th
"flow update" updates a flow rule specified by a rule ID with a
new action list by making a call to "rte_flow_actions_update()":
flow update {port_id} {rule_id}
actions {action} [/ {action} [...]] / end [user_id]
Creating, updating and destroying a flow rule:
testpmd> flow create
Add SipHash which is a fast and cryptographicly sound hash
created by Jean-Philippe Aumasson and Daniel J. Bernstein.
Siphash is widely used by Linux, FreeBSD, OpenBSD and other
projects because it is fast and resistant to DoS attacks.
This version is designed to be useful as alternative hash
with
On Tue, 27 Feb 2024 19:15:49 +0800
Chaoyong He wrote:
> From: Peng Zhang
>
> Add the elf module, which can get mip information from the
> firmware ELF file.
>
> Signed-off-by: Peng Zhang
> Reviewed-by: Chaoyong He
> Reviewed-by: Long Wu
> ---
Why are you rolling your own ELF parser?
There
mlx5 PMD maintains a global counter pool and per-queue counter cache,
which are used to allocate COUNT flow action objects.
Whenever an empty cache is accessed, it is replenished
with a pre-defined number of counters.
If number of configured counters was sufficiently small, then
it might have happ
In case of an error when an asynchronous flow create operation was
enqueued, test-pmd attempted to enqueue a flow destroy operation
of that flow rule.
However, this was incorrect because:
- Flow rule index was used to enqueue a flow destroy operation.
This flow rule index was not yet initialized
Add __rte_pure for __attribute__((pure)) to permit elimination of
direct use of __attribute__((pure)) in other libs.
Signed-off-by: Tyler Retzlaff
Reviewed-by: Morten Brørup
---
lib/eal/include/rte_common.h | 9 +
1 file changed, 9 insertions(+)
diff --git a/lib/eal/include/rte_common.
Use existing __rte_may_alias macro from rte_common.h instead of
directly using __attribute__((__may_alias__)).
Signed-off-by: Tyler Retzlaff
Reviewed-by: Morten Brørup
---
lib/rcu/rcu_qsbr_pvt.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/lib/rcu/rcu_qsbr_pvt.h b/lib/rcu
Clean up direct uses of GCC __attribute__ in libs outside of EAL. A
checkpatch check already warns for new additions.
v2:
* fix typo __rte_pure -> __rte_may_alias in rcu commit message
Tyler Retzlaff (3):
eal: add rte pure attribute macro
lpm: use rte macro instead of GCC attribute
rcu: u
Use newly added __rte_pure macro from rte_common.h instead of directly
using __attribute__((pure)).
Signed-off-by: Tyler Retzlaff
Reviewed-by: Morten Brørup
---
lib/lpm/rte_lpm.c | 4 ++--
lib/lpm/rte_lpm6.c | 2 +-
2 files changed, 3 insertions(+), 3 deletions(-)
diff --git a/lib/lpm/rte_lpm
On Wed, Feb 28, 2024 at 06:29:14PM +0100, David Marchand wrote:
> On Mon, Feb 26, 2024 at 9:20 PM Tyler Retzlaff
> wrote:
> >
> > RTE_LOG_LINE cannot be augmented with a prefix format and arguments
> > without the user of RTE_LOG_LINE using the args... and ## args compiler
> > extension to conditi
Updates the tag of Arm IPsec-MB library to SECLIB-IPSEC-2023.10.13
in snow3g and zuc documentation. Tag SECLIB-IPSEC-2023.10.13 updates
IPSec-MB library version to 1.4.
Signed-off-by: Wathsala Vithanage
Reviewed-by: Ruifeng Wang
Reviewed-by: Honnappa Nagarahalli
Reviewed-by: Jack Bond-Preston
> -Original Message-
> From: Haifei Luo
> Sent: Monday, February 26, 2024 2:04 PM
> To: Ori Kam ; Slava Ovsiienko ;
> Matan Azrad ; Shahaf Shuler ;
> Dariusz Sosnowski ; Suanming Mou
>
> Cc: dev@dpdk.org; NBU-Contact-Thomas Monjalon (EXTERNAL)
> ; Wisam Jaddo ; Raslan
> Darawsheh ; Roni B
On 2/23/2024 4:19 AM, Oleksandr Kolomeiets wrote:
> "flow update" updates a flow rule specified by a rule ID with a
> new action list by making a call to "rte_flow_actions_update()":
>
> flow update {port_id} {rule_id}
> actions {action} [/ {action} [...]] / end [user_id]
>
> Creating
On 2/22/2024 1:46 PM, Loftus, Ciara wrote:
>> Subject: [v9 2/3] net/af_xdp: fix multi interface support for K8s
>>
>> The original 'use_cni' implementation, was added
>> to enable support for the AF_XDP PMD in a K8s env
>> without any escalated privileges.
>> However 'use_cni' used a hardcoded sock
On Wed, Feb 28, 2024 at 10:29:11AM +0100, Morten Brørup wrote:
> > From: Tyler Retzlaff [mailto:roret...@linux.microsoft.com]
> > Sent: Wednesday, 28 February 2024 00.08
> >
> > Use existing __rte_pure macro from rte_common.h instead of directly
>
> Typo: __rte_pure -> __rte_may_alias
>
> The co
On Mon, Feb 26, 2024 at 9:20 PM Tyler Retzlaff
wrote:
>
> RTE_LOG_LINE cannot be augmented with a prefix format and arguments
> without the user of RTE_LOG_LINE using the args... and ## args compiler
> extension to conditionally remove trailing comma when the macro receives
> only a single argumen
On Wed, Feb 28, 2024 at 05:27:54PM +, Konstantin Ananyev wrote:
>
>
> > On Wed, Feb 28, 2024 at 01:34:59PM +, Konstantin Ananyev wrote:
> > >
> > > >
> > > > On Mon, Feb 26, 2024 at 9:20 PM Tyler Retzlaff
> > > > wrote:
> > > > >
> > > > > Use RTE_LOG_LINE_PREFIX instead of RTE_LOG_LINE
> On Wed, Feb 28, 2024 at 01:34:59PM +, Konstantin Ananyev wrote:
> >
> > >
> > > On Mon, Feb 26, 2024 at 9:20 PM Tyler Retzlaff
> > > wrote:
> > > >
> > > > Use RTE_LOG_LINE_PREFIX instead of RTE_LOG_LINE in macro expansions
> > > > which allow a prefix and arguments to be inserted into the
On 2/10/2024 2:09 AM, Stephen Hemminger wrote:
> On Sat, 10 Feb 2024 00:54:15 +
> Ferruh Yigit wrote:
>
>> On 2/8/2024 7:05 PM, Stephen Hemminger wrote:
>>> Rewrite the BPF program used to do queue based RSS.
>>> Important changes:
>>> - uses newer BPF map format BTF
>>> - accepts key
On Wed, Feb 28, 2024 at 12:45:04PM +0100, David Marchand wrote:
> On Tue, Feb 27, 2024 at 7:15 PM Tyler Retzlaff
> wrote:
> >
> > On Mon, Feb 26, 2024 at 12:54:36PM -0800, Stephen Hemminger wrote:
> > > On Mon, 26 Feb 2024 12:19:30 -0800
> > > Tyler Retzlaff wrote:
> > >
> > > > RTE_LOG_LINE cann
On Wed, Feb 28, 2024 at 01:34:59PM +, Konstantin Ananyev wrote:
>
> >
> > On Mon, Feb 26, 2024 at 9:20 PM Tyler Retzlaff
> > wrote:
> > >
> > > Use RTE_LOG_LINE_PREFIX instead of RTE_LOG_LINE in macro expansions
> > > which allow a prefix and arguments to be inserted into the log line
> > >
On Wed, Feb 28, 2024 at 03:18:40PM +0100, David Marchand wrote:
> On Tue, Feb 27, 2024 at 6:44 AM Tyler Retzlaff
> wrote:
> >
> > RTE_MARKER typedefs are a GCC extension unsupported by MSVC. Remove
> > RTE_MARKER fields from rte_mbuf struct.
> >
> > Maintain alignment of fields after removed cache
> -Original Message-
> From: Kundapura, Ganapati
> Sent: Wednesday, February 28, 2024 4:09 PM
> To: dev@dpdk.org; jer...@marvell.com; Jayatheerthan, Jay
>
> Cc: Naga Harish K, S V ; Gujjar, Abhinandan S
>
> Subject: [PATCH v1] eventdev/crypto: fix enqueueing invalid ops
>
> When tail
Each time a flow is allocated in mlx5 PMD the whole buffer,
both rte_flow_hw and mlx5dr_rule parts, are zeroed.
This introduces some wasted work because:
- mlx5dr layer does not assume that mlx5dr_rule must be initialized,
- flow action translation in mlx5 PMD does not need most of the fields
of
Some of the flow fields are either not always required
or are used very rarely, e.g.:
- AGE action reference,
- direct METER/METER_MARK action reference,
- matcher selector for resizable tables.
This patch moves these fields to rte_flow_hw_aux struct in order to
reduce the overall size of the flo
While processing async flow operations in mlx5 PMD,
mlx5_hw_q_job struct is used to hold the following data
related to the ongoing operation.
- operation type,
- user data,
- flow reference.
Job itself is then passed to mlx5dr layer as its "user data".
Other types of data required during flow ope
From: Bing Zhao
The flow rule can be inserted even before the device started. The
only exception is for a queue or RSS action.
For the other interfaces of template API, the start status is not
checked. The checking would cause some cache miss or eviction since
the flag locates on some other cach
mlx5_hw_q_job struct held references to buffers which contained:
- modify header commands array,
- encap/decap data buffer,
- IPv6 routing data buffer.
These buffers were passed as parameters to HWS layer during rule
creation. They were needed only during the call to HWS layer
when flow operation
mlx5_hw_q_job struct held a reference to a temporary flow rule struct,
used during flow rule update operation. It serves as a container for
flow actions data calculated during actions construction.
After flow rule update operation succeeds, data from temporary flow rule
is copied over to original f
From: Bing Zhao
The resource index was introduced to decouple the flow rule and its
resources used by hardware steering. This is needed only when a rule
update is supported.
In some cases, the update is not supported on a table(matcher). E.g.:
* Table is resizable
* FW gets involved
* Root
mlx5_hw_q_job struct held a reference to temporary flow rule pattern
and contained temporary REPRESENTED_PORT and TAG items structs.
They are used whenever it is required to prepend a flow rule pattern,
provided by the application with one of such items.
If prepending is required, then flow rule pa
From: Erez Shitrit
The user want to know before trying to update a rule if that matcher that
keeps the original rule supports updating.
Signed-off-by: Erez Shitrit
---
drivers/net/mlx5/hws/mlx5dr.h | 8
drivers/net/mlx5/hws/mlx5dr_matcher.c | 12
2 files changed,
From: Erez Shitrit
The function returns true if that matcher can contain complicated rule,
which means rule that needs more than one writing to the HW in order to
have it.
Signed-off-by: Erez Shitrit
---
drivers/net/mlx5/hws/mlx5dr.h | 8
drivers/net/mlx5/hws/mlx5dr_action.c
Goal of this is to remove the unnecessary copying of precalculated
mlx5dr_rule_action structures used to create HWS flow rules.
Before this patch, during template table creation an array of these
structures was calculated for each actions template used.
Each of these structures contained either fu
Goal of this patchset is to improve the throughput of flow insertion
and deletion in mlx5 PMD when HW Steering flow engine is used.
- Patch 1 - Use preallocated per-queue, per-actions template buffer
for storing translated flow actions, instead of allocating and
filling it on demand, on each f
On 2/27/2024 11:15 AM, Chaoyong He wrote:
> From: Peng Zhang
>
> Add an option to force reload the firmware.
>
> This option can be ignored in some case, for example:
> When using a 2 port NFP card and both with this reload firmware option,
> only the first one will cause the firmware reload and
On 2/27/2024 11:15 AM, Chaoyong He wrote:
> From: Peng Zhang
>
> Add the elf module, which can get mip information from the
> firmware ELF file.
>
What is MIP?
Is it in the .note section of the ELF binary?
How it is used?
Source code always refers it as MIP, can you please clarify it in the
com
On 2/27/2024 7:38 PM, Harman Kalra wrote:
> Fixing uninitialized value access.
>
> Coverity issue: 414934
> Fixes: 9a9eb104edf6 ("ethdev: parse multiple representor devargs")
>
> Signed-off-by: Harman Kalra
>
Acked-by: Ferruh Yigit
Applied to dpdk-next-net/main, thanks.
https://bugs.dpdk.org/show_bug.cgi?id=1389
Bug ID: 1389
Summary: Investigate automatically building config schema
documentation
Product: DPDK
Version: unspecified
Hardware: All
OS: All
Status: UN
On Wed, Feb 28, 2024 at 09:28:22AM +0100, Morten Brørup wrote:
> +To: Thomas, previously joined the discussion
>
> > From: Tyler Retzlaff [mailto:roret...@linux.microsoft.com]
> > Sent: Tuesday, 27 February 2024 18.17
> >
> > On Tue, Feb 27, 2024 at 10:10:03AM +0100, Morten Brørup wrote:
> > > >
https://bugs.dpdk.org/show_bug.cgi?id=1388
Bug ID: 1388
Summary: Add dependency documentation for TG Node
Product: DPDK
Version: unspecified
Hardware: All
OS: All
Status: UNCONFIRMED
Severity: normal
Hi,
> -Original Message-
> From: Gregory Etelson
> Sent: Wednesday, February 28, 2024 3:33 PM
> To: dev@dpdk.org
> Cc: Gregory Etelson ; Maayan Kashani
> ; Raslan Darawsheh ; Dariusz
> Sosnowski
> Subject: [PATCH v3 0/4] net/mlx5: add support for flow table resizing
>
> Support template
On Wed, Feb 28, 2024 at 4:01 PM Morten Brørup
wrote:
>
> > From: David Marchand [mailto:david.march...@redhat.com]
> > Sent: Wednesday, 28 February 2024 15.19
> >
> > On Tue, Feb 27, 2024 at 6:44 AM Tyler Retzlaff
> > wrote:
> > >
> > > RTE_MARKER typedefs are a GCC extension unsupported by MSVC
NAT64 is treated as a modify header action. The action order and
limitation should be the same as that of modify header in each
domain.
Since the last 2 TAG registers will be used implicitly in the
address backup mode, the values in these registers are no longer
valid after the NAT64 action. The a
The action will handle the IPv4 and IPv6 headers translation. It will
add / remove IPv6 address prefix by default.
To use the user specific address, another rule to modify the
addresses of the IP header is needed.
Signed-off-by: Bing Zhao
---
drivers/net/mlx5/mlx5_flow_hw.c | 20 +++
The NAT64 DR actions can be shared among the tables. All these
actions can be created during configuring the flow queues and saved
for the future usage.
Even the actions can be shared now, inside per each flow rule, the
actual hardware resources are unique.
Signed-off-by: Bing Zhao
---
doc/guid
REG_C_6 is used as the 1st one and since it is reserved internally
by default, there is no impact.
The remaining 2 registers will be fetched from the available TAGs
array from right to left. They will not be masked in the array due
to the fact that not all the rules will use NAT64 action.
Signed-
From: Erez Shitrit
Add support of new action mlx5dr_action_create_nat64.
The new action allows to modify IP packets from version to version, IPV6
to IPV4 and vice versa.
Signed-off-by: Erez Shitrit
Signed-off-by: Bing Zhao
---
drivers/net/mlx5/hws/mlx5dr.h| 29 ++
drivers/net/mlx5/hw
This patch set contains the mlx5 PMD implementation for NAT64.
Series-acked-by: Ori Kam
Update in v4:
1. rebase to solve the conflicts.
2. fix the old NIC startup issue in a separate patch:
https://patches.dpdk.org/project/dpdk/patch/20240227152627.25749-1-bi...@nvidia.com/
Update in
> From: David Marchand [mailto:david.march...@redhat.com]
> Sent: Wednesday, 28 February 2024 15.19
>
> On Tue, Feb 27, 2024 at 6:44 AM Tyler Retzlaff
> wrote:
> >
> > RTE_MARKER typedefs are a GCC extension unsupported by MSVC. Remove
> > RTE_MARKER fields from rte_mbuf struct.
> >
> > Maintain
> On Feb 28, 2024, at 5:44 AM, Abdullah Ömer Yamaç wrote:
>
> While I was implementing the new API, I realized one issue, and it would be
> good to discuss it here. First of all rte_rcu_qsbr_dq_reclaim function checks
> the state of the qsbr values. It means that all threads should report the
Acked-by: Patrick Robb
On 2/28/24 10:56, Konstantin Ananyev wrote:
- Implemented SVE code for comparing signatures in bulk lookup.
- Added Defines in code for SVE code support.
- Optimise NEON code
- New SVE code is ~5% slower than optimized NEON for N2 processor.
Signed-off-by: Yoan Picchi
Signed-off-by: Harjot S
On Wed, Feb 28, 2024 at 3:04 PM Dodji Seketeli wrote:
> > Btw, I see no way to suppress this (except a global [suppress_type]
> > name = rte_mbuf)...
>
> Right.
>
> To avoid having subsequent changes to that type from being "overly"
> suppressed, maybe do something like:
>
> [suppress_type]
>
On Tue, Feb 27, 2024 at 6:44 AM Tyler Retzlaff
wrote:
>
> RTE_MARKER typedefs are a GCC extension unsupported by MSVC. Remove
> RTE_MARKER fields from rte_mbuf struct.
>
> Maintain alignment of fields after removed cacheline1 marker by placing
> C11 alignas(RTE_CACHE_LINE_MIN_SIZE).
>
> Update imp
Hello,
David Marchand writes:
> Hello Dodji,
o/
[...]
> This change is reported as a potential ABI change.
>
> For the context, this patch
> https://patchwork.dpdk.org/project/dpdk/patch/1709012499-12813-21-git-send-email-roret...@linux.microsoft.com/
> removes null-sized markers (those fiel
Adding one new unit test code for validating the features
added as part of GCM with 64 byte AAD.
The new test case adds one new test for GCM algo for both
encrypt and decrypt operations.
Signed-off-by: Nishikant Nayak
Acked-by: Ciara Power
---
v2:
- Removed unused code.
- Added one new u
This patch handles the changes required for updating the common
header fields specific to GEN LCE, Also added/updated of the response
processing APIs based on GEN LCE requirement.
Signed-off-by: Nishikant Nayak
Acked-by: Ciara Power
---
v2:
- Renamed device from GEN 5 to GEN LCE.
- Remov
Adding GEN5 files for handling GEN LCE specific operations.
These files are inherited from the existing files/APIs
which has some changes specific GEN5 requirements
Also updated the mailmap file.
Signed-off-by: Nishikant Nayak
Acked-by: Ciara Power
---
v5:
- Replaced usage of __u8 with uint8
Adding GEN LCE specific macros which is required for updating
the support for GEN LCE features.
Also this patch adds other macros which is being used by GEN LCE
Specific APIs.
Signed-off-by: Nishikant Nayak
Acked-by: Ciara Power
---
v6:
- Removed unused PCI device IDs from the device list.
This patchset adds a new QAT LCE device.
The device currently only supports symmetric crypto,
and only the AES-GCM algorithm.
v6:
- Added documentation and release note changes.
- Removed unused device PCI ID.
v5:
- Fixed compilation issue by replacing __u8 with uint8_t.
v4:
- Fixed cover
Hi,
> -Original Message-
> From: Dariusz Sosnowski
> Sent: Tuesday, February 27, 2024 3:58 PM
> To: Slava Ovsiienko ; Ori Kam ;
> Suanming Mou ; Matan Azrad
>
> Cc: dev@dpdk.org; sta...@dpdk.org
> Subject: [PATCH v2] net/mlx5: fix VLAN handling in meter split
>
> On the attempt to creat
Hi,
> -Original Message-
> From: Michael Baum
> Sent: Wednesday, February 28, 2024 11:51 AM
> To: dev@dpdk.org
> Cc: Matan Azrad ; Dariusz Sosnowski
> ; Raslan Darawsheh ; Slava
> Ovsiienko ; Ori Kam ;
> Suanming Mou ; Alex Vesker ;
> sta...@dpdk.org; Erez Shitrit
> Subject: [PATCH] net/
Hi
> -Original Message-
> From: Bing Zhao
> Sent: Tuesday, February 27, 2024 5:26 PM
> To: dev@dpdk.org; Suanming Mou ; Raslan
> Darawsheh
> Cc: Ori Kam ; Slava Ovsiienko ;
> Matan Azrad ; Dariusz Sosnowski
> ; Gregory Etelson ;
> sta...@dpdk.org
> Subject: [PATCH] net/mlx5: fix the HWS
>
> On Mon, Feb 26, 2024 at 9:20 PM Tyler Retzlaff
> wrote:
> >
> > Use RTE_LOG_LINE_PREFIX instead of RTE_LOG_LINE in macro expansions
> > which allow a prefix and arguments to be inserted into the log line
> > without the need to use the ## args variadic argument pack extension.
> >
> > Signed
Support template table API in PMD.
The patch allows to increase existing table capacity.
Signed-off-by: Gregory Etelson
Acked-by: Dariusz Sosnowski
---
drivers/net/mlx5/mlx5.h | 5 +
drivers/net/mlx5/mlx5_flow.c| 51 +++
drivers/net/mlx5/mlx5_flow.h| 84 +++--
drivers/net/ml
The multi-pattern actions related structures and management code
have been moved to the table level.
That code refactor is required for the upcoming table resize feature.
Signed-off-by: Gregory Etelson
Acked-by: Dariusz Sosnowski
---
drivers/net/mlx5/mlx5_flow.h| 73 ++-
drivers/ne
Modified the conditionals in `flow_hw_table_create()` to use bitwise
AND instead of equality checks when assessing
`table_cfg->attr->specialize` bitmask.
This will allow for greater flexibility as the bitmask may encapsulate
multiple flags.
The patch maintains the previous behavior with single flag
From: Maayan Kashani
Before this patch, ipool size could be fixed by
setting max_idx in mlx5_indexed_pool_config upon
ipool creation. Or it can be auto resized to the
maximum limit by setting max_idx to zero upon
ipool creation and the saved value is the maximum
index possible.
This patch adds ip
Support template table resize API.
Gregory Etelson (3):
net/mlx5: fix parameters verification in HWS table create
net/mlx5: move multi-pattern actions management to table level
net/mlx5: add support for flow table resizing
Maayan Kashani (1):
net/mlx5: add resize function to ipool
drive
On Mon, Feb 26, 2024 at 9:20 PM Tyler Retzlaff
wrote:
>
> Use RTE_LOG_LINE_PREFIX instead of RTE_LOG_LINE in macro expansions
> which allow a prefix and arguments to be inserted into the log line
> without the need to use the ## args variadic argument pack extension.
>
> Signed-off-by: Tyler Retzl
On 2/28/2024 2:27 AM, huangdengdui wrote:
>
>
> On 2024/2/27 0:43, Ferruh Yigit wrote:
>> On 2/26/2024 3:16 AM, Jie Hai wrote:
>>> On 2024/2/23 21:53, Ferruh Yigit wrote:
On 2/20/2024 3:58 AM, Jie Hai wrote:
> Hi, Ferruh,
>
> Thanks for your review.
>
> On 2024/2/7 22:15,
On 2/20/2024 3:33 PM, Sivaprasad Tummala wrote:
> + ops = rte_power_get_ops(env);
> + if (ops->status == 0) {
> + POWER_LOG(ERR, WER,
> + "Power Management Environment(%d) not"
> + " registered\n", env);
'WER' seems typo, causing build er
> -Original Message-
> From: Brian Dooley
> Sent: Wednesday, February 28, 2024 11:33 AM
> To: Ji, Kai ; De Lara Guarch, Pablo
>
> Cc: dev@dpdk.org; gak...@marvell.com; Dooley, Brian
>
> Subject: [PATCH v4] crypto/ipsec_mb: unified IPsec MB interface
>
> Currently IPsec MB provides bo
On Tue, Feb 27, 2024 at 7:15 PM Tyler Retzlaff
wrote:
>
> On Mon, Feb 26, 2024 at 12:54:36PM -0800, Stephen Hemminger wrote:
> > On Mon, 26 Feb 2024 12:19:30 -0800
> > Tyler Retzlaff wrote:
> >
> > > RTE_LOG_LINE cannot be augmented with a prefix format and arguments
> > > without the user of RTE
While I was implementing the new API, I realized one issue, and it would be
good to discuss it here. First of all rte_rcu_qsbr_dq_reclaim function
checks the state of the qsbr values. It means that all threads should
report the quiescent states. It conflicts with my aim.
Let's think about below sc
Currently IPsec MB provides both the JOB API and direct API.
AESNI_MB PMD is using the JOB API codepath while ZUC, KASUMI, SNOW3G
and CHACHA20_POLY1305 are using the direct API.
Instead of using the direct API for these PMDs, they should now make
use of the JOB API codepath. This would remove all u
>
> - Implemented SVE code for comparing signatures in bulk lookup.
> - Added Defines in code for SVE code support.
> - Optimise NEON code
> - New SVE code is ~5% slower than optimized NEON for N2 processor.
>
> Signed-off-by: Yoan Picchi
> Signed-off-by: Harjot Singh
> Reviewed-by: Nathan Br
On Tue, Feb 27, 2024 at 6:23 PM Tyler Retzlaff
wrote:
>
> On Tue, Feb 27, 2024 at 04:18:10PM +0100, David Marchand wrote:
> > Hello Dodji,
> >
> > On Tue, Feb 27, 2024 at 6:44 AM Tyler Retzlaff
> > wrote:
> > >
> > > RTE_MARKER typedefs are a GCC extension unsupported by MSVC. Remove
> > > RTE_MA
When tail pointer of Circ buffer rollsover as the Circ buffer becomes full,
crypto adapter is enqueueing ops beyond the size of the Circ buffer
leading to segfault due to invalid ops access.
Fixed by enqueueing ops from head pointer to (size-head) number of ops
when Circ buffer becomes full and th
Support template table API in PMD.
The patch allows to increase existing table capacity.
Signed-off-by: Gregory Etelson
Acked-by: Dariusz Sosnowski
---
drivers/net/mlx5/mlx5.h | 5 +
drivers/net/mlx5/mlx5_flow.c| 51
drivers/net/mlx5/mlx5_flow.h| 84 --
drivers/net/
The multi-pattern actions related structures and management code
have been moved to the table level.
That code refactor is required for the upcoming table resize feature.
Signed-off-by: Gregory Etelson
Acked-by: Dariusz Sosnowski
---
drivers/net/mlx5/mlx5_flow.h| 73 ++-
drivers/ne
Modified the conditionals in `flow_hw_table_create()` to use bitwise
AND instead of equality checks when assessing
`table_cfg->attr->specialize` bitmask.
This will allow for greater flexibility as the bitmask may encapsulate
multiple flags.
The patch maintains the previous behavior with single flag
From: Maayan Kashani
Before this patch, ipool size could be fixed by
setting max_idx in mlx5_indexed_pool_config upon
ipool creation. Or it can be auto resized to the
maximum limit by setting max_idx to zero upon
ipool creation and the saved value is the maximum
index possible.
This patch adds ip
Support template table resize API.
Gregory Etelson (3):
net/mlx5: fix parameters verification in HWS table create
net/mlx5: move multi-pattern actions management to table level
net/mlx5: add support for flow table resizing
Maayan Kashani (1):
net/mlx5: add resize function to ipool
drive
Replace static array of cache-aligned structs with an lcore variable,
to slightly benefit code simplicity and performance.
RFC v5:
* Fix lcore value pointer bug introduced by RFC v4.
RFC v4:
* Remove strange-looking lcore value lookup potentially containing
invalid lcore id. (Morten Brørup)
Keep per-lcore power intrinsics state in a lcore variable to reduce
cache working set size and avoid any CPU next-line-prefetching causing
false sharing.
Signed-off-by: Mattias Rönnblom
Acked-by: Morten Brørup
---
lib/eal/x86/rte_power_intrinsics.c | 17 +++--
1 file changed, 11 ins
Replace static array of cache-aligned structs with an lcore variable,
to slightly benefit code simplicity and performance.
RFC v3:
* Replace for loop with FOREACH macro.
Signed-off-by: Mattias Rönnblom
Acked-by: Morten Brørup
---
lib/power/rte_power_pmd_mgmt.c | 36 ---
Replace keeping PRNG state in a RTE_MAX_LCORE-sized static array of
cache-aligned and RTE_CACHE_GUARDed struct instances with keeping the
same state in a more cache-friendly lcore variable.
RFC v3:
* Remove cache alignment on unregistered threads' rte_rand_state.
(Morten Brørup)
Signed-off-by
Introduce DPDK per-lcore id variables, or lcore variables for short.
An lcore variable has one value for every current and future lcore
id-equipped thread.
The primary use case is for statically allocating
small chunks of often-used data, which is related logically, but where
there are performan
Add test suite to exercise the API.
RFC v5:
* Adapt tests to reflect the removal of the GET() and SET() macros.
RFC v4:
* Check all lcore id's values for all variables in the many variables
test case.
* Introduce test case for max-sized lcore variables.
RFC v2:
* Improve alignment-relate
This RFC presents a new API for static per-lcore id
data allocation.
Please refer to the API documentation for both a
rationale for this new API, and a comparison to the alternatives
available.
The adoption of this API would affect many different DPDK modules, but
the author updated only a few,
Hi,
> -Original Message-
> From: Dariusz Sosnowski
> Sent: Tuesday, February 27, 2024 3:52 PM
> To: Slava Ovsiienko ; Ori Kam ;
> Suanming Mou ; Matan Azrad
>
> Cc: dev@dpdk.org
> Subject: [PATCH v3 0/4] net/mlx5: connection tracking changes
>
> Patches 1 and 2 contain fixes for existin
Hi,
> -Original Message-
> From: Dariusz Sosnowski
> Sent: Tuesday, February 27, 2024 3:37 PM
> To: Slava Ovsiienko ; Ori Kam ;
> Suanming Mou ; Matan Azrad
>
> Cc: dev@dpdk.org
> Subject: [PATCH v4 1/2] net/mlx5: move meter init functions
>
> Move mlx5_flow_meter_init() and mlx5_flow_m
On Mon, Feb 26, 2024 at 6:30 PM Pavan Nikhilesh Bhagavatula
wrote:
>
> > Uninitialized mbufs are enqueued to eventdev which causes segfault
> > on freeing the mbuf in tx adapter.
> > Fixed by initializing mbufs before enqueuing to eventdev.
> >
> > Fixes: 46cf97e4bbfa ("eventdev: add test for eth
The integrity item uses the DW "oks1" in header layout. It includes the
all supported bits for both inner and outer. When item is integrity
type, the relevant bits are turned on and all DW is submitted.
When user provides more then single integrity item in same pattern, the
last one overrides the
Add support for matching metadata in HWS ingress rules.
It using REG_B matching which is supported for each device supports HWS.
Signed-off-by: Michael Baum
Reviewed-by: Erez Shitrit
---
drivers/net/mlx5/hws/mlx5dr.h | 1 +
drivers/net/mlx5/hws/mlx5dr_definer.c | 32 ++-
Hi Chengwen,
Please see my reply in-line.
Thanks
Amit Shukla
> -Original Message-
> From: fengchengwen
> Sent: Wednesday, February 28, 2024 12:34 PM
> To: Amit Prakash Shukla ; Cheng Jiang
> ; Gowrishankar Muthukrishnan
>
> Cc: dev@dpdk.org; Jerin Jacob ; Anoob Joseph
> ; Kevin Laatz ;
1 - 100 of 110 matches
Mail list logo
