> From: Stephen Hemminger [mailto:step...@networkplumber.org]
> Sent: Friday, 22 August 2025 20.20
>
> Now that static_assert is available, better than pre-processor
> use of #if condition.
>
> Signed-off-by: Stephen Hemminger
> ---
Acked-by: Morten Brørup
> From: Stephen Hemminger [mailto:step...@networkplumber.org]
> Sent: Friday, 22 August 2025 20.20
>
> Add new compare functions for common small key sizes.
>
> Bugzilla ID: 1775
> Suggested-by: Morten Brørup
> Reported-by: Mattias Rönnblom
>
> Signed-off-by: Stephen Hemminger
> ---
Acked-by
> From: Stephen Hemminger [mailto:step...@networkplumber.org]
> Sent: Friday, 22 August 2025 20.20
>
> Make comparison of sizes compatible across platforms.
> Keep the special case code for 16 bytes for x86 and arm64 but
> also add simple xor for others.
>
> Need to keep rte_hash_k32_cmp_eq() exp
I forgot to add my tag to the previous email.
Reviewed-by: Patrick Robb
I think this patch looks good. Thanks Thomas and thanks Ivan for your
insights.
Reviewed-by: Patrick Robb
On Wed, Jul 30, 2025 at 8:59 AM Thomas Wilks wrote:
> From: Alex Chapman
>
> This patch adds the required functionality for the RSS key_update,
> RETA, and hash test suites. This includes
On Wed, Jul 30, 2025 at 8:59 AM Thomas Wilks wrote:
>
> +def configure_random_reta(self, testpmd: TestPmdShell, queue_number:
> int) -> list[int]:
> +"""Configure RETA to have random order of queues.
> +
> +Args:
> +testpmd: The testpmd instance that will be used t
On Mon, Aug 25, 2025 at 8:58 PM Ajit Khaparde
wrote:
>
> On Mon, Aug 25, 2025 at 2:28 AM David Marchand
> wrote:
> >
> > There is no good reason to keep code for older versions of DPDK in the
> > main branch.
> ACK.
> >
> > Signed-off-by: David Marchand
> Reviewed-by: Ajit Khaparde
>
> > ---
>
From: Long Li
If a device is hot removed and hot plugged, it needs the same driver
parameters that are passed to EAL. However, during device removal, all
EAL driver parameters are freed as part of the cleanup.
Cache those driver parameters for future hot plug events. Because we don't
know which
The DPDK DQ driver made an implicit assumption that buffers will be
returned from the NIC in the order which they are posted. While this is
generally the case, there are certain edge cases, such as HW GRO, where
buffers may be completed out of order. In such cases, the driver would
have returned th
Prior to this patch, the DQ format made use of two separate arrays which
dictate the TX submission queue, named tx_ring and sw_ring. The tx_ring
represented the actual hardware descriptor ring, while the sw_ring
represented a ring of DMA'd buffers to be used in transmission.
Given that the DQ form
This would allow for separate control of datapath-related logs vs
non-datapath logs.
Signed-off-by: Joshua Washington
Reviewed-by: Ankit Garg
---
drivers/net/gve/gve_logs.h | 3 +++
drivers/net/gve/gve_rx_dqo.c | 5 +++--
drivers/net/gve/gve_tx_dqo.c | 11 ++-
3 files changed, 12 in
When creating new RX rings, one less than the number of buffers in the
ring need to be allocated. It is possible that only a part of the
allocation is successful, resulting in a failure to create the rings.
In this case, the driver should free the buffers which were successfully
allocated to avoid
Both RX and TX processing on DQ were originally implemented with the
asusmption that descriptor completions will be written by the hardware
in the order that they are posted. In certain cases, such as RSC on RX
and double completions on TX, this will not necessarily be the case.
Depends-on: series
From: Long Li
In most cases, netvsc is used with a VF device. The application generally
runs with better performance when all the device memory is allocated on
VF's NUMA node, as the VF device carries most of the data packets.
But sometimes netvsc may run on a different NUMA node than that of th
#
August 21, 2025
Attendees
1. Patrick Robb
2. Luca Vizzarro
3. Maxime Coquelin
4. Dean Marx
#
Minutes
==
On Mon, Aug 25, 2025 at 2:28 AM David Marchand
wrote:
>
> There is no good reason to keep code for older versions of DPDK in the
> main branch.
ACK.
>
> Signed-off-by: David Marchand
Reviewed-by: Ajit Khaparde
> ---
> drivers/net/bnxt/bnxt_cpr.h | 2 --
> drivers/net/bnxt/bnxt_et
On Sun, Aug 24, 2025 at 08:38:56PM -0700, Stephen Hemminger wrote:
> Superseded by common rte_bsd_queue.h
>
> Signed-off-by: Stephen Hemminger
> ---
> lib/eal/windows/include/sys/queue.h | 840
> 1 file changed, 840 deletions(-)
> delete mode 100644 lib/eal/windows/
Removed use of non-standard array range initialization, which is
non-standard and is not supported by MSVC.
Fixed macro mlx5_malloc_numa_tolerant: added missing backslash.
Fixed macro pool_malloc: added missing backslash and added prefix
MLX5_ to "NUMA_TOLERANT" parameter.
Added temporary workar
Removed array range initializations as they are non-standard and not
implemented by MSVC.
Signed-off-by: Andre Muezerie
---
app/test/test_cryptodev_zuc_test_vectors.h | 156 -
1 file changed, 153 insertions(+), 3 deletions(-)
diff --git a/app/test/test_cryptodev_zuc_test_vec
Now that all issues related to drivers are addressed they can be
enabled to be built with MSVC.
Signed-off-by: Andre Muezerie
---
drivers/meson.build | 4
1 file changed, 4 deletions(-)
diff --git a/drivers/meson.build b/drivers/meson.build
index b62880db02..d98ef27f66 100644
--- a/drivers
This patchset solves the last issues preventing drivers from being
compiled with MSVC and enables drivers to be built with it.
v5:
- Remove unnecessary parentheses from macros
v4:
- Add explicit casts to avoid warnings in mlx5 driver
(https://bugs.dpdk.org/show_bug.cgi?id=1758).
- Add cast to dri
25/08/2025 16:13, Gavin Li:
> Previously, mac_own was a bit-field indicating the MAC addresses owned by
> the PMD, specifically marking those added to VFs by the PMD and allowing
> them to be flushed during exit.
>
> With this commit, the bit-field now applies to all MAC addresses added by
> the P
25/08/2025 16:13, Gavin Li:
> Rules for multicast MAC addresses are intended to filter multicast traffic
> and are managed through multicast MAC add/remove APIs.
You should name the functions to be explicit.
Also it is not only managed via DPDK API,
as you say below we also retrieve kernel address
Rules for multicast MAC addresses are intended to filter multicast traffic
and are managed through multicast MAC add/remove APIs. In mlx5_dev_spawn
function, devices (PF, VFs, and SFs) retrieve the netdev-configured MAC
addresses via netlink and store them in the PMD device data, which
includes mul
Previously, mac_own was a bit-field indicating the MAC addresses owned by
the PMD, specifically marking those added to VFs by the PMD and allowing
them to be flushed during exit.
With this commit, the bit-field now applies to all MAC addresses added by
the PMD, including those for PF, VF, and SFs.
A bug was introduced by commit 2d0665a7f771 ("net/mlx5: align PF and
VF/SF MAC address handling"). Flows for multicast MAC addresses would
not be created anymore.
This patch series are to fix the issue.
Minggang Li (Gavin) (2):
net/mlx5: update how MAC address bit-fields are used
net/mlx5: ad
https://bugs.dpdk.org/show_bug.cgi?id=1703
Patrick Robb (pr...@iol.unh.edu) changed:
What|Removed |Added
Status|UNCONFIRMED |RESOLVED
Resolution|--
https://bugs.dpdk.org/show_bug.cgi?id=1702
Patrick Robb (pr...@iol.unh.edu) changed:
What|Removed |Added
Status|UNCONFIRMED |RESOLVED
Resolution|--
On Mon, Aug 25, 2025 at 12:35:48PM +, Ciara Loftus wrote:
> The SSE paths have been removed from the i40e, iavf and ice Intel
> drivers. Add a release note documenting their removal.
>
> Signed-off-by: Ciara Loftus
> ---
> doc/guides/rel_notes/release_25_11.rst | 5 +
> 1 file changed, 5
Reviewed-by: Wathsala Vithanage
> -Original Message-
> From: Shani Peretz
> Sent: Tuesday, August 5, 2025 1:23 AM
> To: dev@dpdk.org
> Cc: getel...@nvidia.com; bi...@nvidia.com; Shani Peretz
> ; Wathsala Vithanage ;
> sta...@dpdk.org; Bruce Richardson ; Dhruv
> Tripathi
> Subject: [PATC
The SSE paths have been removed from the i40e, iavf and ice Intel
drivers. Add a release note documenting their removal.
Signed-off-by: Ciara Loftus
---
doc/guides/rel_notes/release_25_11.rst | 5 +
1 file changed, 5 insertions(+)
diff --git a/doc/guides/rel_notes/release_25_11.rst
b/doc/g
Modern x86 systems all support AVX2, if not AVX-512, so the SSE path
is no longer widely used. This change will not result in any feature
loss, as in the rare cases where the SSE path would have been used in
the absence of AVX2 or AVX-512, the fallback scalar path will be used
which has feature par
Modern x86 systems all support AVX2, if not AVX-512, so the SSE path
is no longer widely used. This change will not result in any feature
loss, as in the rare cases where the SSE path would have been used in
the absence of AVX2 or AVX-512, the fallback scalar path will be used
which has feature par
Modern x86 systems all support AVX2, if not AVX-512, so the SSE path
is no longer widely used. This change will not result in any feature
loss, as in the rare cases where the SSE path would have been used in
the absence of AVX2 or AVX-512, the fallback scalar path will be used
which has feature par
This patch removes the SSE Rx and Tx paths from the i40e, iavf and ice
driverss. Each of these drivers have faster vector paths (AVX2 and AVX-512)
which have feature parity with the SSE paths. In environments where AVX2 or
AVX-512 are not supported, the scalar path will still be used, which also
ha
Fixes a typo in the kernel header comment for struct vduse_iova_info
where "IOVA region" was incorrectly repeated.
Signed-off-by: Ashwini Sahu
---
kernel/linux/uapi/linux/vduse.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/kernel/linux/uapi/linux/vduse.h b/kernel/linux/ua
Fixes a typo in the kernel header comment for struct vduse_iova_info
where "IOVA region" was incorrectly repeated.
Signed-off-by: Ashwini Sahu
---
kernel/linux/uapi/linux/vduse.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/kernel/linux/uapi/linux/vduse.h b/kernel/linux/ua
Hi, Morten
> -Original Message-
> From: Morten Brørup
> Sent: Saturday, July 19, 2025 5:39 PM
> To: Shani Peretz
> Cc: dev@dpdk.org; Stephen Hemminger
> Subject: RE: [RFC PATCH 0/5] Introduce mempool object new debug
> capabilities
>
> > From: Morten Brørup [mailto:m...@smartsharesyste
Hi all,
Here is a new stable release:
https://fast.dpdk.org/rel/dpdk-23.11.5.tar.xz
The git tree is at:
https://dpdk.org/browse/dpdk-stable/?h=23.11
Xueming Li
---
.mailmap | 12 +
VERSION|
Acked-by: Chengwen Feng
On 8/25/2025 4:32 PM, Ashwini Sahu wrote:
> Fixes a typo in the kernel header comment for struct vduse_iova_info
> where "IOVA region" was incorrectly repeated.
>
> Signed-off-by: Ashwini Sahu
On 8/25/2025 11:38 AM, Stephen Hemminger wrote:
> The version of sys/queue.h on Linux from glibc was derived
> from an older version of BSD and is missing several key macros
> such as LIST_FOREACH_SAFE. This leads to drivers re-implementing
> the macros (sometimes badly) and other bugs.
> Introduce
There is no good reason to keep code for older versions of DPDK in the
main branch.
Signed-off-by: David Marchand
---
drivers/net/bnxt/bnxt_cpr.h | 2 --
drivers/net/bnxt/bnxt_ethdev.c | 5
drivers/net/bnxt/bnxt_reps.c | 7 +-
drivers/net/bnxt/tf_ul
Hi,
Please try to fix your email client so that we can follow the replies
inlined.
On 8/22/25 9:33 AM, Bathija, Pravin wrote:
Dear Maxime, Bruce, David, Chenbo,
I would like to follow up on the status of the patches and when we should
expect these to be merged into the mainline.
Please let m
From: Luca Vizzarro
Paul is giving up his role of co-maintainer of DTS,
electing Luca Vizzarro in his place.
Signed-off-by: Luca Vizzarro
Acked-by: Paul Szczepanek
Acked-by: Patrick Robb
---
MAINTAINERS | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/MAINTAINERS b/MAINTAI
Fixes a typo in the kernel header comment for struct vduse_iova_info
where "IOVA region" was incorrectly repeated.
Signed-off-by: Ashwini Sahu
---
kernel/linux/uapi/linux/vduse.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/kernel/linux/uapi/linux/vduse.h b/kernel/linux/ua
Added the option to jump to a specific group.
snippets that are not supported on the root table can now call
create_jump_flow to jump to other group.
Added three new snippets that uses the jump flow:
- match Network Service Header (NSH) snippet
- match RoCE IB BTH opcode/dest_qp snippet
- Switch G
This snippet demonstrates port affinity matching in bonded ports,
ensuring packets are sent back through their ingress port.
The NR_QUEUES constant was moved to common.h to support mapping
four TX queues (0-3) to two port affinities (0,1→1 and 2,3→2).
Signed-off-by: Shani Peretz
---
examples/fl
v3:
add comma in the meson.build file to fix build error
v2:
sent mistakely as RFC - remove RFC tag, no functional changes
v1:
This patch series adds port affinity matching for bonded ports and
refactors queue management to support the new port affinity snippets.
This refactoring also enables oth
When test vector plaintext exceeds buffer size, only the first
max_buffer_size bytes are processed, causing incorrect digest
verification (computed vs expected mismatch).
This patch fixes this issue by checking that the plaintext size is
larger than the buffer size and returns an error with a log.
https://bugs.dpdk.org/show_bug.cgi?id=1772
xuemi...@nvidia.com (xuemi...@nvidia.com) changed:
What|Removed |Added
Resolution|--- |DUPLICATE
On 8/25/25 06:38, Stephen Hemminger wrote:
Use the new rte_bsd_queue.h where needed but most files already
get the needed macros from rte_tailq.h
Signed-off-by: Stephen Hemminger
...
diff --git a/lib/mbuf/rte_mbuf_dyn.c b/lib/mbuf/rte_mbuf_dyn.c
index 5987c9dee8..9a20033021 100644
--- a/lib
51 matches
Mail list logo