[PATCH v9-S2 00/13] net/dpaa2 updates

[Stephen Hemminger]

On Mon, 12 Oct 2026 12:34:31 +0530
Prashant Gupta <prashant.gupta_3 at nxp.com> wrote:

> Changes in v9:
> - Address review comments for meter and policing support.
> - Add meter resource cleanup during device close.
> - Add policing capability checks.
> - Validate burst size ranges before hardware programming.
> - Fix software/hardware state consistency on profile and policy updates.
> - Use standard FDIR reporting for flow-steered packets.
> - Replace global flow miss configuration with a per-device default flow
>   and remove the DPAA2_FLOW_CONTROL_MISS_FLOW environment variable.
> - Document and gate minimum DPNI version requirement for 24-byte flow keys.
> - Rework parser decode infrastructure and parser-result dumping.
> - Parse additional tunnel and fragmented packet types.
> - Fold parser-related fixes into the originating patches.
> - Remove unused VXLAN offset helpers after soft parser removal.
> - Move release note updates into the existing DPAA2 ethernet driver
>   entry under New Features.
> - Fix coding style issues reported during review.
> 
> Series overview:
> - fix an integer overflow in the CCSR region mapping,
> - set Tx confirmation on device init and support a larger burst size,
> - support MPLS and PPPoE flow distribution, meter and policing, and the
>   flow drop action, with a per-device default flow miss action,
> - identify Rx mbuf hash information by FLC and add minimum key size
>   support,
> - restructure parser processing, parse tunnel and fragmented packet
>   types, remove the unused soft parser driver and rename Rx queue flags.
> 
> Gagandeep Singh (7):
>   net/dpaa2: set Tx confirmation on device init
>   net/dpaa2: support larger burst size
>   net/dpaa2: support MPLS and PPPoE flow distribution
>   net/dpaa2: support meter and policing
>   net/dpaa2: support flow drop action
>   net/dpaa2: set default flow miss action per device
>   net/dpaa2: identify Rx mbuf hash information by FLC
> 
> Hemant Agrawal (4):
>   net/dpaa2: fix integer overflow in CCSR region mapping
>   net/dpaa2: add minimum key size support
>   net/dpaa2: restructure dpaa2 parser processing
>   net/dpaa2: parse tunnel and fragmented packet types
> 
> Jun Yang (2):
>   net/dpaa2: remove unused soft parser driver
>   net/dpaa2: rename Rx queue flags
> 
>  doc/guides/nics/dpaa2.rst               |    1 +
>  doc/guides/nics/features/dpaa2.ini      |    4 +-
>  doc/guides/rel_notes/release_26_11.rst  |   15 +-
>  drivers/net/dpaa2/dpaa2_ethdev.c        |   89 +-
>  drivers/net/dpaa2/dpaa2_ethdev.h        |  218 +++-
>  drivers/net/dpaa2/dpaa2_flow.c          |  633 +++------
>  drivers/net/dpaa2/dpaa2_meter.c         |  804 ++++++++++++
>  drivers/net/dpaa2/dpaa2_parse_dump.h    |  374 ------
>  drivers/net/dpaa2/dpaa2_parser_decode.h | 1549 +++++++++++++++++++++++
>  drivers/net/dpaa2/dpaa2_pmd_logs.h      |   67 +
>  drivers/net/dpaa2/dpaa2_recycle.c       |   24 +-
>  drivers/net/dpaa2/dpaa2_rxtx.c          |  351 +++--
>  drivers/net/dpaa2/dpaa2_sparser.c       |  137 --
>  drivers/net/dpaa2/dpaa2_sparser.h       |  202 ---
>  drivers/net/dpaa2/dpaa2_tm.c            |   16 +-
>  drivers/net/dpaa2/mc/dpni.c             |   11 +-
>  drivers/net/dpaa2/mc/fsl_dpni.h         |   25 +-
>  drivers/net/dpaa2/mc/fsl_dpni_cmd.h     |    3 +-
>  drivers/net/dpaa2/meson.build           |    2 +-
>  19 files changed, 3096 insertions(+), 1429 deletions(-)
>  create mode 100644 drivers/net/dpaa2/dpaa2_meter.c
>  delete mode 100644 drivers/net/dpaa2/dpaa2_parse_dump.h
>  create mode 100644 drivers/net/dpaa2/dpaa2_parser_decode.h
>  delete mode 100644 drivers/net/dpaa2/dpaa2_sparser.c
>  delete mode 100644 drivers/net/dpaa2/dpaa2_sparser.h
> 

In general looks good but there still is this open question.

Review: [PATCH v9-S2 00/13] net/dpaa2 updates

Applies on top of v7-S1. v8 comments on release notes, the
duplicated burst check, and 10/13 indent, naming and commit message
are addressed. Two items from v8 remain.


[PATCH 05/13] net/dpaa2: support meter and policing

Warning: no answer yet on meter_lock. meter_create still drops the
lock around dpaa2_mtr_hw_program(). If control operations can be
concurrent, there is a race here; if not allowed, what is the point
of meter_lock? Either hold it across check, program and insert, or
drop the lock.

Info: fsl_dpni.h still has the orphaned comment left behind when
DPNI_POLICER_OPT_DO_NOT_RESET_COUNTERS was removed:

-/**
- * Do NOT reset counters generally for policer update
- */
-