A new DPDK release candidate is ready for testing:
https://git.dpdk.org/dpdk/tag/?id=v25.03-rc3
There are 71 new patches in this snapshot.
Release notes:
https://doc.dpdk.org/guides/rel_notes/release_25_03.html
Please test and report new issues on https://bugs.dpdk.org
If no maj
18/02/2025 16:44, Kevin Laatz:
> On 31/01/2025 14:49, Bruce Richardson wrote:
> > The requirement to have DPDK in the DSA wq name is given in the idxd
> > chapter section 5.3.3 [1]. However, that information should also be
> > given in section 5.3.1 [2] where we list out the fields to be set and
>
18/02/2025 10:46, Bruce Richardson:
> On Mon, Dec 02, 2024 at 09:35:01AM +, Anatoly Burakov wrote:
> > Support for `--noiommu-mode` flag was added, but documentation for it was
> > not. Add documentation for the flag.
> >
> > Signed-off-by: Anatoly Burakov
>
> Acked-by: Bruce Richardson
Ap
02/12/2024 10:35, Burakov, Anatoly:
> On 12/2/2024 10:31 AM, Anatoly Burakov wrote:
> > Currently, when binding a device to VFIO, the UID/GID for the device will
> > always stay as system default (`root`). Yet, when running DPDK as non-root
> > user, one has to change the UID/GID of the device to m
23/01/2025 11:27, Anoob Joseph:
> > Subject: [PATCH] examples/ipsec-secgw: fix cryptodev and eventdev ID
> >
> > Fixing cryptodev and eventdev ID numbers.
> >
> > Fixes: 0dbe550a4af5 ("examples/ipsec-secgw: initialize event crypto
> > adapter")
> > Cc: sta...@dpdk.org
> >
> > Signed-off-by: Tej
10/03/2025 07:48, Shani Peretz:
> From: Thomas Monjalon
> > 12/01/2025 08:52, Shani Peretz:
> > > This patch corrects the destination IP address mask to restore the
> > > previous implementation's behavior.
> > >
> > > Also it fixes a misuse of rte_flow_item_tcp struct.
> > > Replace it with the a
20/12/2024 00:58, Stephen Hemminger:
> The pdump statistics are used to keep track of the number
> of packets captured, filtered, etc. These need not be cumalative
> and instead should be start counting when capture starts.
>
> This fixes the issue where multiple invocations of dumpcap
> would inc
05/03/2025 21:04, Medvedkin, Vladimir:
> On 05/03/2025 14:45, Andre Muezerie wrote:
> > On Wed, Mar 05, 2025 at 10:08:33AM +, Bruce Richardson wrote:
> >> On Wed, Dec 18, 2024 at 07:21:26AM -0800, Andre Muezerie wrote:
> >>> Test lpm6_perf_autotest was hitting a stack overflow on Windows
> >>>
> > From: Pavan Nikhilesh
> >
> > Fix extra_march_features not being added to mcpu flags when SoC
> > configuration advertises them.
> >
> > Fixes: c02c01dbf907 ("config/arm: prefer strict use of -mcpu if supported")
>
> Acked-by: Jerin Jacob
Applied, thanks.
08/03/2025 10:01, pbhagavat...@marvell.com:
> From: Pavan Nikhilesh
>
> Add crypto as extra march feature to cn9k soc.
>
> Signed-off-by: Pavan Nikhilesh
Applied, thanks.
Remove driver-specific build instructions for the AVX2 and AVX-512 code,
and rely instead on the generic driver build file.
Signed-off-by: Bruce Richardson
---
drivers/net/intel/i40e/meson.build | 26 ++
drivers/net/intel/iavf/meson.build | 25 ++---
d
Remove from remaining net drivers the special-case code to handle AVX2
or AVX512 specific files. These can be built instead using
drivers/meson.build.
Signed-off-by: Bruce Richardson
---
drivers/net/bnxt/meson.build | 10 +-
drivers/net/enic/meson.build | 10 +-
drivers
Add support to the top-level driver build file for AVX2 and AVX512
specific sources. This should simplify driver builds by avoiding the
need to constantly reimplement the same build logic
Signed-off-by: Bruce Richardson
---
drivers/meson.build | 30 ++
1 file changed,
All DPDK-supported versions of clang and gcc have the "-mpclmul" and
"-maes" flags, so we never need to check for those. This allows the SSE
code path to be unconditionally built on x86.
For the AVX512 code path, simplify it by only checking for the
build-time support, and always doing a separate
Since all supported compilers can generate AVX2 code, we will always
enable the build of the AVX2 files on x86. This means that
CC_AVX2_SUPPORT is always true on x86, so it can be removed and a
regular "#ifdef RTE_ARCH_x86" used in its place.
Signed-off-by: Bruce Richardson
---
drivers/net/bnxt/
Remove custom logic for building AVX2 and AVX-512 files. Within the C
code this requires some renaming of build macros to use the standard
defines.
Signed-off-by: Bruce Richardson
---
lib/fib/dir24_8.c | 6 +++---
lib/fib/meson.build | 18 +-
lib/fib/trie.c | 6 +++---
3
On Tue, Mar 18, 2025 at 10:35 AM Bruce Richardson
wrote:
>
> Since all supported compilers can generate AVX2 code, we will always
> enable the build of the AVX2 files on x86. This means that
> CC_AVX2_SUPPORT is always true on x86, so it can be removed and a
> regular "#ifdef RTE_ARCH_x86" used in
On Wed, Mar 05, 2025 at 04:24:58PM +, Bruce Richardson wrote:
> In cases where the number of cores on a given socket is greater than
> RTE_MAX_LCORES, then EAL will be unaware of all the sockets/numa nodes
> on a system. Fix this limitation by having the EAL probe the NUMA node
> for cores it i
Currently packet capturing in Scapy is used by running Scapy's own
class AsyncSniffer in the same shell as packet sending. Because there is
some start up time involved in the AsyncSniffer, doing this every single
time a test case requires it can become slow and unreliable if sniffer
was not ready y
Use the support for building AVX512 code present in lib/meson.build
rather than reimplementing it in the library meson.build file.
Signed-off-by: Bruce Richardson
---
lib/member/meson.build | 46 +++---
1 file changed, 7 insertions(+), 39 deletions(-)
diff --
Use the common support for AVX512 code present in lib/meson.build,
rather than hard-coding it. The only complication is an extra check for
the "-mvpclmulqdq" command-line flag before adding the AVX512 sources.
Signed-off-by: Bruce Richardson
---
lib/net/meson.build | 13 -
lib/net/
remove custom logic for building AVX2 and AVX-512 files.
Signed-off-by: Bruce Richardson
---
lib/acl/meson.build | 54 -
1 file changed, 4 insertions(+), 50 deletions(-)
diff --git a/lib/acl/meson.build b/lib/acl/meson.build
index a80c172812..87e9f25f
Add support to the top-level lib build file for AVX2 and AVX512
specific sources. This should simplify library builds by avoiding the
need to constantly reimplement the same build logic
Signed-off-by: Bruce Richardson
---
lib/meson.build | 34 +-
1 file changed, 3
remove special-case handling for AVX512, and rely on mechanisms in
the drivers meson.build file.
Signed-off-by: Bruce Richardson
---
drivers/event/dlb2/dlb2_sse.c | 4
drivers/event/dlb2/meson.build | 16 ++--
2 files changed, 6 insertions(+), 14 deletions(-)
diff --git a/dri
A number of libs and drivers had special optimized AVX2 and AVX512 code
paths for performance reasons, and these tended to have copy-pasted
logic to build those files. Centralise that logic in the main
drivers/ and lib/ meson.build files to avoid duplication.
v3: add patch for event/dlb2 AVX512 ha
Hi,
From: Gregory Etelson
Sent: Thursday, March 13, 2025 10:33 AM
To: dev@dpdk.org
Cc: Gregory Etelson; Maayan Kashani; Raslan Darawsheh; Michael Baum; Dariusz
Sosnowski; Slava Ovsiienko; Bing Zhao; Ori Kam; Suanming Mou; Matan Azrad
Subject: [PATCH 1/2] common/mlx5: add device duplication funct
Release status meeting minutes 2025-03-18
=
Agenda:
* Release Dates
* Subtrees
* Roadmaps
* LTS
* Defects
* Opens
Participants:
* Broadcom
* ARM
* Intel
* Marvell
* Nvidia
* Red Hat
Release Dates
-
The following are the proposed working dates
On Tue, Mar 18, 2025 at 12:35:01PM +, Medvedkin, Vladimir wrote:
> Acked-by: Vladimir Medvedkin
>
> On 10/03/2025 13:11, Bruce Richardson wrote:
> > When registering interrupts, there was no check if the registration of
> > the interrupt succeeded. Add in such a check, and go to fallback path
> -Original Message-
> From: Pavan Nikhilesh Bhagavatula
> Sent: Thursday, March 13, 2025 10:13 AM
> To: Wathsala Wathawana Vithanage ; Jerin
> Jacob ; Bruce Richardson
> Cc: dev@dpdk.org; nd
> Subject: RE: [PATCH] config/arm: fix march features not being set
>
> > Hi Pavan,
> >
> > I
Acked-by: Vladimir Medvedkin
On 10/03/2025 13:11, Bruce Richardson wrote:
With the fallback interrupt path now enabled on FreeBSD there are
segmentation faults on app exit, due to the alarm interrupt trying to
access invalid pointers. Add checks for null to fix these crashes.
Fixes: cd3b12495
On Mon, Mar 17, 2025 at 10:50:24AM +0100, David Marchand wrote:
> On Fri, Mar 14, 2025 at 6:44 PM Bruce Richardson
> wrote:
> >
> > A number of drivers had special optimized AVX2 and AVX512 code paths
> > for performance reasons, and these tended to have copy-pasted logic
> > to build those files.
> -Original Message-
> From: Thomas Monjalon
> Sent: Friday, March 7, 2025 1:42 AM
> To: annou...@dpdk.org
> Subject: release candidate 25.03-rc2
>
> A new DPDK release candidate is ready for testing:
> https://git.dpdk.org/dpdk/tag/?id=v25.03-rc2
>
> There are 251 new patches in t
On Tue, Mar 18, 2025 at 09:51:05AM +0100, Dariusz Sosnowski wrote:
> > +use std::process::Command;
> > +
> > +pub fn main() {
> > +let mut pkgconfig = Command::new("pkg-config");
> > +
> > +match pkgconfig.args(["--libs", "libdpdk"]).output() {
> > +Ok(output) => {
> > +
On Mon, Mar 17, 2025 at 4:43 PM David Marchand
wrote:
>
> So far, each DPDK library (or driver) exposing symbols in an ABI had to
> maintain a version.map and use some macros for symbol versioning,
> specially crafted with the GNU linker in mind.
>
> This series proposes to rework the whole princi
34 matches
Mail list logo
