Re: [PATCH] net/mlx5: fix flex tunnel validation

Hi, On 30/04/2025 2:03 PM, Dariusz Sosnowski wrote: Pattern template validation did not take into account the configured tunnel mode of the flex item. This caused validation failures for templates with flex item registered as tunnel headers. Specifically, any item after such flex item was still

Re: [PATCH] net/mlx5: fix support for meter flow action in HWS

Hi, On 04/05/2025 8:23 AM, Gregory Etelson wrote: METER flow action is not supported in MLX5 HWS mode. Application must use METER_MARK flow action. The patch removes METER action from HWS code. Fixes: 48fbb0e93d06 ("net/mlx5: support flow meter mark indirect action with HWS") Signed-off-by:

Re: [PATCH] net/mlx5: add match on IPv6 fragment extension header

Hi, On 27/04/2025 2:25 PM, Maayan Kashani wrote: Add HWS support for IPv6 fragment extension header matching for HWS. Supported field is next_header. Signed-off-by: Maayan Kashani Acked-by: Dariusz Sosnowski - Patch applied to next-net-mlx, Kindest regards Raslan Darawsheh

Re: [PATCH 1/2] net/mlx5: non template - fix validation for GENEVE options

Hi, On 27/04/2025 2:22 PM, Maayan Kashani wrote: For non-template API on top of HWS, geneve options parser is created by the pmd and not by the user, therefore during validation , the parser is not yet created. The fix is to ignore the validation of geneve options in case the rule is a non-temp

Re: [PATCH 1/2] common/mlx5: read SWS capability bits

Hi, On 27/04/2025 2:19 PM, Maayan Kashani wrote: SWS will be disabled in future HW generation. Checking SWS capability bits and returning relevant error. if user configuration is not supported, will be added in the follow up commit. Signed-off-by: Maayan Kashani Acked-by: Dariusz Sosnowski

[PATCH] net/mana: check for vendor_part_id when probing RDMA device

From: Long Li The RDMA kernel driver may expose two MANA RDMA devices, for RC and RAW QP types. The purpose is to support fast service mode at SOC. Depending on kernel version, the probe may fail if DPDK picks up the wrong device. Add check for vendor_part_id and RAW QP capability when probing t

Re: [PATCH] net/ice: update log message

On Fri, May 09, 2025 at 04:25:38PM +, Pillai, Dhanya R wrote: > Clarified log message for DDP package load failure > > Signed-off-by: Pillai, Dhanya R > >From v1: Acked-by: Anatoly Burakov Applied to dpdk-next-net-intel. Thanks, /Bruce

Re: [PATCH] net/ice: enlarge limits to 64 bits (ice_stats_get)

On Mon, May 12, 2025 at 05:04:20PM +0100, Bruce Richardson wrote: > On Thu, Apr 03, 2025 at 09:50:41AM -0400, LukeSeewald wrote: > > There are various counters/stats coming from ice_stats_get that appear > > to only use 32 or 40 bits and once it hits that limit it "overflows" > > and resets. This w

RE: [PATCH v2 2/2] ethdev: remove callback checks from fast path

> From: Sunil Kumar Kori > Sent: Monday, 12 May 2025 17.07 > > rte_eth_fp_ops contains ops for fast path APIs. Each API > validates availability of callback and then invoke it. > These checks impact data path performace. > > Hence removing these NULL checks instead using dummy > callbacks. > >

[PATCH] net/tap: fix add qdisc failed when create tap

From: luyan The errno variable was assigned a positive value in the previous handling, but here it is compared against -EEXIST, When the tap already exists, this would falsely report an error. Signed-off-by: luyan --- drivers/net/tap/tap_tcmsgs.c | 4 ++-- 1 file changed, 2 insertions(+), 2 de

Re: [PATCH] net/ice: enlarge limits to 64 bits (ice_stats_get)

On Thu, Apr 03, 2025 at 09:50:41AM -0400, LukeSeewald wrote: > There are various counters/stats coming from ice_stats_get that appear > to only use 32 or 40 bits and once it hits that limit it "overflows" > and resets. This was causing large spikes (sudden 2^42 increase from > one second to the nex

RE: [PATCH v2 1/2] ethdev: remove unnecessary type conversion

> From: Sunil Kumar Kori > Sent: Monday, 12 May 2025 17.07 > > As rte_eth_rx_queue_count() returns signed value to represent > the error cases but internally invoked callback is returning > unsigned value. Hence unnecessary type conversion is done. > > To avoid this typecasting from signed to un

Re: [PATCH] build: fix disabling developer mode

On Mon, May 12, 2025 at 5:28 PM Bruce Richardson wrote: > On Thu, May 08, 2025 at 06:42:29AM +0200, David Marchand wrote: > > A recent change broke disabling developer mode: > > $ meson setup plop -Ddeveloper_mode=disabled > > ... > > drivers/meson.build:150:14: ERROR: Unknown variable "fs". > > >

Re: [PATCH] build: fix disabling developer mode

On Thu, May 08, 2025 at 06:42:29AM +0200, David Marchand wrote: > A recent change broke disabling developer mode: > $ meson setup plop -Ddeveloper_mode=disabled > ... > drivers/meson.build:150:14: ERROR: Unknown variable "fs". > > Fixes: 8fa3af6c05f8 ("build: add generic support for base code in d

[PATCH v2 1/2] ethdev: remove unnecessary type conversion

From: Sunil Kumar Kori As rte_eth_rx_queue_count() returns signed value to represent the error cases but internally invoked callback is returning unsigned value. Hence unnecessary type conversion is done. To avoid this typecasting from signed to unsigned, fixed return type of callback functions.

[PATCH v2 2/2] ethdev: remove callback checks from fast path

From: Sunil Kumar Kori rte_eth_fp_ops contains ops for fast path APIs. Each API validates availability of callback and then invoke it. These checks impact data path performace. Hence removing these NULL checks instead using dummy callbacks. Signed-off-by: Sunil Kumar Kori --- lib/ethdev/ethde

[PATCH] doc: reword contributor's guidelines

I have revised sections 9-12 for grammar and clarity. Signed-off-by: Nandini Persad --- doc/guides/contributing/linux_uapi.rst| 32 ++-- doc/guides/contributing/patches.rst | 177 +++--- doc/guides/contributing/stable.rst| 163 doc/guides/c

Re: [PATCH 1/4] bus/pci: Use force-noreplace flag when mapping PCI resources

On Thu, May 08, 2025 at 01:32:50PM +0200, Burakov, Anatoly wrote: > On 5/6/2025 7:40 PM, Jake Freeland wrote: > > When mapping PCI resources in secondary processes, use the > > RTE_MAP_FORCE_ADDRESS_NOREPLACE flag to indicate that the > > mapping must be made at the provided address. > > > > Witho

Re: [PATCH 1/4] bus/pci: Use force-noreplace flag when mapping PCI resources

On Thu May 8, 2025 at 6:32 AM CDT, Anatoly Burakov wrote: > On 5/6/2025 7:40 PM, Jake Freeland wrote: > > When mapping PCI resources in secondary processes, use the > > RTE_MAP_FORCE_ADDRESS_NOREPLACE flag to indicate that the > > mapping must be made at the provided address. > > > > Without this

[DPDK/eventdev Bug 1705] DPDK fails to compile on Ubuntu 24.04 with BullseyeCoverage

https://bugs.dpdk.org/show_bug.cgi?id=1705 Bug ID: 1705 Summary: DPDK fails to compile on Ubuntu 24.04 with BullseyeCoverage Product: DPDK Version: 25.03 Hardware: x86 OS: Linux Status: UNCONFIRM

[PATCH v2] cnxk: use stdatomic API

From: Pavan Nikhilesh Replace GCC inbuilt atomic functions with rte_atomic_xxx API. Signed-off-by: Pavan Nikhilesh --- drivers/crypto/cnxk/cn10k_cryptodev_ops.c | 68 +-- drivers/crypto/cnxk/cn9k_cryptodev_ops.c | 25 + drivers/net/cnxk/cn10k_ethdev.c |

[PATCH v3 10/13] net/ixgbe: use common Rx rearm code

The ixgbe driver has implementations of vectorized mbuf rearm code that is identical to the ones in the common code, so just use those. Signed-off-by: Anatoly Burakov --- drivers/net/intel/ixgbe/ixgbe_rxtx.h | 2 +- drivers/net/intel/ixgbe/ixgbe_rxtx_vec_neon.c | 66 +--- d

[PATCH v3 12/13] net/intel: add common Rx mbuf recycle

Currently, there are duplicate implementations of Rx mbuf recycle in some drivers, specifically ixgbe and i40e. Move them into a common header. While we're at it, also support no-IOVA-in-mbuf case. Signed-off-by: Anatoly Burakov --- drivers/net/intel/common/recycle_mbufs.h | 67 +++

Re: [PATCH v3 01/13] net/ixgbe: remove unused field in Rx queue struct

On Mon, May 12, 2025 at 01:54:27PM +0100, Anatoly Burakov wrote: > The `rdh` (read head) field in the `ixgbe_rx_queue` struct is not used > anywhere in the codebase, and can be removed. > > Signed-off-by: Anatoly Burakov > --- > drivers/net/intel/ixgbe/ixgbe_rxtx.c | 9 ++--- > drivers/net/i

[PATCH v3 13/13] net/intel: add common Tx mbuf recycle

Currently, there are duplicate implementations of Tx mbuf recycle in some drivers, specifically ixgbe and i40e. Move them into a common header. Signed-off-by: Anatoly Burakov --- drivers/net/intel/common/recycle_mbufs.h | 98 +++ drivers/net/intel/common/tx.h

[PATCH v3 11/13] net/intel: support wider x86 vectors for Rx rearm

Currently, for 32-byte descriptor format, only SSE instruction set is supported. Add implementation for AVX2 and AVX512 instruction sets. This implementation similarly constant-propagates everything at compile time and thus should not affect performance of existing code paths. To improve code reada

[PATCH v3 09/13] net/iavf: use common Rx rearm code

The iavf driver has implementations of vectorized mbuf rearm code that is identical to the ones in the common code, so just use those. Signed-off-by: Anatoly Burakov --- drivers/net/intel/iavf/iavf_rxtx.h| 4 +- drivers/net/intel/iavf/iavf_rxtx_vec_avx2.c | 3 +- drivers/net/in

[PATCH v3 07/13] net/intel: generalize vectorized Rx rearm

There is certain amount of duplication between various drivers when it comes to Rx ring rearm. This patch takes implementation from ice driver as a base because it has support for no IOVA in mbuf as well as all vector implementations, and moves them to a common file. The driver Rx rearm code used

[PATCH v3 08/13] net/i40e: use common Rx rearm code

The i40e driver has an implementation of vectorized mbuf rearm code that is identical to the one in the common code, so just use that. In addition, the i40e has an implementation of Rx queue rearm for Neon instruction set, so create a common header for Neon implementations too, and use that in i40

[PATCH v3 06/13] net/iavf: use the common Rx queue structure

Make the iavf driver use the new common Rx queue structure. Because the iavf driver supports both 16-byte and 32-byte descriptor formats (controlled by RTE_LIBRTE_IAVF_16BYTE_RX_DESC define), the common queue structure has to take that into account, so the ring queue structure will have both, whil

[PATCH v3 05/13] net/ice: use the common Rx queue structure

Make the ice driver use the new common Rx queue structure. Because the ice driver supports both 16-byte and 32-byte descriptor formats (controlled by RTE_LIBRTE_ICE_16BYTE_RX_DESC define), the common queue structure has to take that into account, so the ring queue structure will have both, while t

[PATCH v3 02/13] net/iavf: make IPsec stats dynamically allocated

Currently, the stats structure is directly embedded in the queue structure. We're about to move iavf driver to a common Rx queue structure, so we can't have driver-specific structures that aren't pointers, inside the common queue structure. To prepare, we replace direct embedding into the queue str

[PATCH v3 04/13] net/i40e: use the common Rx queue structure

Make the i40e driver use the new common Rx queue structure. Because the i40e driver supports both 16-byte and 32-byte descriptor formats (controlled by RTE_LIBRTE_I40E_16BYTE_RX_DESC define), the common queue structure has to take that into account, so the ring queue structure will have both, whil

[PATCH v3 03/13] net/ixgbe: create common Rx queue structure

In preparation for deduplication effort, generalize the Rx queue structure. Most of the fields are simply moved to common/rx.h, clarifying the comments where necessary. There are some instances where the field is renamed when moving, to make it more consistent with the rest of the codebase. Speci

[PATCH v3 01/13] net/ixgbe: remove unused field in Rx queue struct

The `rdh` (read head) field in the `ixgbe_rx_queue` struct is not used anywhere in the codebase, and can be removed. Signed-off-by: Anatoly Burakov --- drivers/net/intel/ixgbe/ixgbe_rxtx.c | 9 ++--- drivers/net/intel/ixgbe/ixgbe_rxtx.h | 1 - 2 files changed, 2 insertions(+), 8 deletions(-)

RE: [PATCH v1 1/1] bbdev: fix possible null pointer dereference

> -Original Message- > From: Maxime Coquelin > Sent: 12 May 2025 17:23 > To: Nicolas Chautru ; dev@dpdk.org > Cc: Hemant Agrawal ; hernan.var...@intel.com > Subject: Re: [PATCH v1 1/1] bbdev: fix possible null pointer dereference > Importance: High > > > > On 5/2/25 12:10 AM, Nicolas

Re: [PATCH] net/virtio: support Rx/Tx burst mode info

Hi Roger, On 4/16/25 7:35 PM, Roger Melton wrote: Return burst mode according to the selected Rx/Tx burst function name. Update 25.07 release notes with this information. Signed-off-by: Roger Melton --- doc/guides/rel_notes/release_25_07.rst | 3 + drivers/net/virtio/virtio_ethdev.c |

Re: [PATCH v1 1/1] bbdev: fix possible null pointer dereference

On 5/2/25 12:10 AM, Nicolas Chautru wrote: This change addresses the CID 457446: Null pointer dereferences (REVERSE_INULL). Not expected to happen in practice, but the config may be potentially undefined through API. Coverity issue: 457446 Fixes: 61aa25b96e14 ("bbdev: add trace points") Sig

[PATCH v2 13/13] net/intel: add common Tx mbuf recycle

Currently, there are duplicate implementations of Tx mbuf recycle in some drivers, specifically ixgbe and i40e. Move them into a common header. Signed-off-by: Anatoly Burakov --- drivers/net/intel/common/recycle_mbufs.h | 98 +++ drivers/net/intel/common/tx.h

[PATCH v2 12/13] net/intel: add common Rx mbuf recycle

Currently, there are duplicate implementations of Rx mbuf recycle in some drivers, specifically ixgbe and i40e. Move them into a common header. While we're at it, also support no-IOVA-in-mbuf case. Signed-off-by: Anatoly Burakov --- drivers/net/intel/common/recycle_mbufs.h | 67 +++

[PATCH v2 11/13] net/intel: support wider x86 vectors for Rx rearm

Currently, for 32-byte descriptor format, only SSE instruction set is supported. Add implementation for AVX2 and AVX512 instruction sets. This implementation similarly constant-propagates everything at compile time and thus should not affect performance of existing code paths. To improve code reada

[PATCH v2 10/13] net/ixgbe: use common Rx rearm code

The ixgbe driver has implementations of vectorized mbuf rearm code that is identical to the ones in the common code, so just use those. Signed-off-by: Anatoly Burakov --- drivers/net/intel/ixgbe/ixgbe_rxtx.h | 2 +- drivers/net/intel/ixgbe/ixgbe_rxtx_vec_neon.c | 66 +--- d

[PATCH v2 09/13] net/iavf: use common Rx rearm code

The iavf driver has implementations of vectorized mbuf rearm code that is identical to the ones in the common code, so just use those. Signed-off-by: Anatoly Burakov --- drivers/net/intel/iavf/iavf_rxtx.h| 4 +- drivers/net/intel/iavf/iavf_rxtx_vec_avx2.c | 3 +- drivers/net/in

[PATCH v2 08/13] net/i40e: use common Rx rearm code

The i40e driver has an implementation of vectorized mbuf rearm code that is identical to the one in the common code, so just use that. In addition, the i40e has an implementation of Rx queue rearm for Neon instruction set, so create a common header for Neon implementations too, and use that in i40

[PATCH v2 07/13] net/intel: generalize vectorized Rx rearm

There is certain amount of duplication between various drivers when it comes to Rx ring rearm. This patch takes implementation from ice driver as a base because it has support for no IOVA in mbuf as well as all vector implementations, and moves them to a common file. The driver Rx rearm code used

[PATCH v2 06/13] net/iavf: use the common Rx queue structure

Make the iavf driver use the new common Rx queue structure. Because the iavf driver supports both 16-byte and 32-byte descriptor formats (controlled by RTE_LIBRTE_IAVF_16BYTE_RX_DESC define), the common queue structure has to take that into account, so the ring queue structure will have both, whil

[PATCH v2 05/13] net/ice: use the common Rx queue structure

Make the ice driver use the new common Rx queue structure. Because the ice driver supports both 16-byte and 32-byte descriptor formats (controlled by RTE_LIBRTE_ICE_16BYTE_RX_DESC define), the common queue structure has to take that into account, so the ring queue structure will have both, while t

[PATCH v2 04/13] net/i40e: use the common Rx queue structure

Make the i40e driver use the new common Rx queue structure. Because the i40e driver supports both 16-byte and 32-byte descriptor formats (controlled by RTE_LIBRTE_I40E_16BYTE_RX_DESC define), the common queue structure has to take that into account, so the ring queue structure will have both, whil

[PATCH v2 03/13] net/ixgbe: create common Rx queue structure

In preparation for deduplication effort, generalize the Rx queue structure. Most of the fields are simply moved to common/rx.h, clarifying the comments where necessary. There are some instances where the field is renamed when moving, to make it more consistent with the rest of the codebase. Speci

[PATCH v2 02/13] net/iavf: make IPsec stats dynamically allocated

Currently, the stats structure is directly embedded in the queue structure. We're about to move iavf driver to a common Rx queue structure, so we can't have driver-specific structures that aren't pointers, inside the common queue structure. To prepare, we replace direct embedding into the queue str

[PATCH v2 01/13] net/ixgbe: remove unused field in Rx queue struct

The `rdh` (read head) field in the `ixgbe_rx_queue` struct is not used anywhere in the codebase, and can be removed. Signed-off-by: Anatoly Burakov --- drivers/net/intel/ixgbe/ixgbe_rxtx.c | 9 ++--- drivers/net/intel/ixgbe/ixgbe_rxtx.h | 1 - 2 files changed, 2 insertions(+), 8 deletions(-)

Re: [PATCH v2 2/2] net/mlx5: enhance log file path

Hi Experts, is there any chance to review and accept this patch? Brs, Yang Ming

Re: [PATCH v2 2/2] crypto/ipsec_mb: fix QP release in secondary

On 5/7/2025 11:25 PM, Ji, Kai wrote: Hi Yangming, PID check is implemented here: https://github.com/DPDK/dpdk/blob/75f179ebe347b6098cf3af26d3d3b7168fe3fe24/drivers/crypto/ipsec_mb/ipsec_mb_ops.c#L376

RE: [PATCH v1 1/1] bbdev: fix possible null pointer dereference

> -Original Message- > From: Nicolas Chautru > Sent: Thursday 1 May 2025 23:11 > To: dev@dpdk.org; maxime.coque...@redhat.com > Cc: hemant.agra...@nxp.com; Vargas, Hernan ; > Chautru, Nicolas > Subject: [PATCH v1 1/1] bbdev: fix possible null pointer dereference > > This change addresses