Re: [PATCH v13 00/25] NXP DPAA driver enhancements and fixes

[Stephen Hemminger <stephen@networkplumber.org>]

On Wed, 19 Aug 2026 16:19:39 +0530
Hemant Agrawal <hemant.agrawal@nxp.com> wrote:

> This series collects a set of fixes and enhancements for the NXP DPAA
> bus, mempool, dma, crypto and net drivers targeting 26.11.
> 
> It includes memory-leak and resource-cleanup fixes on the device
> remove/close paths, more robust frame queue and congestion-group
> shutdown, secondary-process safety guards, BPID and cgrid lifecycle
> handling, and several new features: offline (O/H) port device support,
> enhanced virtual storage profile (VSP) port support, fmcless Rx queue
> configuration via devargs, Rx/Tx taildrop threshold devargs, ORP queue
> create and burst enqueue, non fmX-macY shared Ethernet naming, and DMA
> scatter-gather and errata-workaround devargs. Documentation and release
> notes are updated accordingly.
> 
> v13:
> * Addressed mailing-list review feedback on the previous revision.
> * Dropped the Tx port rate limiting patch.
> 
> 
> Gagandeep Singh (3):
>   net/dpaa: fix device remove
>   dma/dpaa: add devargs for SG and errata workaround
>   drivers: add offline (O/H) port device support
> 
> Hemant Agrawal (6):
>   net/dpaa: fix free port resources on close
>   net/dpaa: optimize FM deconfig
>   net/dpaa: support Rx/Tx taildrop threshold devarg
>   net/dpaa: support non fmX-macY type of shared Ethernet name
>   bus/dpaa: improve log macro usages
>   doc: update release notes with NXP DPAA changes
> 
> Jun Yang (13):
>   bus/dpaa: refine fman naming
>   bus/dpaa: scan max BPID from DTS
>   drivers: shutdown DPAA FQ by fq descriptor
>   drivers: add DPAA cgrid cleanup support
>   bus/dpaa: improve FQ shutdown with channel validation
>   drivers: add BMI Tx statistics
>   net/dpaa: optimize FMC MAC type parsing
>   drivers: release DPAA bpid on driver destructor
>   bus/dpaa: orp queue create and burst enqueue
>   net/dpaa: support fmcless rxq number as devargs
>   bus/dpaa: optimize DPAA multi-entry buffer pool operations
>   crypto/dpaa_sec: improve crypto fq resource handling
>   drivers: add dpaax enter destructor to gate EAL memory frees
> 
> Prashant Gupta (2):
>   bus/dpaa: fix FQD dest wq channel decoding
>   drivers: add process-type guards for secondary process
> 
> Vanshika Shukla (1):
>   net/dpaa: enhance VSP port support
> 
>  doc/guides/dmadevs/dpaa.rst               |  30 +-
>  doc/guides/nics/dpaa.rst                  |  39 +-
>  doc/guides/rel_notes/release_26_11.rst    |  34 +
>  drivers/bus/dpaa/base/fman/fman.c         |  42 +-
>  drivers/bus/dpaa/base/fman/fman_hw.c      | 200 +++---
>  drivers/bus/dpaa/base/fman/netcfg_layer.c |   8 +-
>  drivers/bus/dpaa/base/qbman/bman.c        |  59 +-
>  drivers/bus/dpaa/base/qbman/bman_driver.c |  51 +-
>  drivers/bus/dpaa/base/qbman/qman.c        | 264 ++++++--
>  drivers/bus/dpaa/base/qbman/qman.h        |  23 +-
>  drivers/bus/dpaa/base/qbman/qman_driver.c |  29 +-
>  drivers/bus/dpaa/bus_dpaa_driver.h        |  11 +-
>  drivers/bus/dpaa/dpaa_bus.c               | 172 +++--
>  drivers/bus/dpaa/dpaa_bus_base_symbols.c  |   4 +
>  drivers/bus/dpaa/include/fman.h           |  25 +-
>  drivers/bus/dpaa/include/fsl_bman.h       |  49 +-
>  drivers/bus/dpaa/include/fsl_qman.h       |  65 +-
>  drivers/bus/dpaa/rte_dpaa_logs.h          |   2 +
>  drivers/common/dpaax/compat.h             |  22 +-
>  drivers/common/dpaax/dpaax_iova_table.c   |  24 +-
>  drivers/crypto/dpaa_sec/dpaa_sec.c        | 109 ++-
>  drivers/dma/dpaa/dpaa_qdma.c              | 110 ++-
>  drivers/mempool/dpaa/dpaa_mempool.c       |  60 +-
>  drivers/mempool/dpaa/dpaa_mempool.h       |   3 +-
>  drivers/net/dpaa/dpaa_ethdev.c            | 359 ++++++++--
>  drivers/net/dpaa/dpaa_ethdev.h            |  25 +-
>  drivers/net/dpaa/dpaa_flow.c              | 134 ++--
>  drivers/net/dpaa/dpaa_flow.h              |   7 +-
>  drivers/net/dpaa/dpaa_fmc.c               |  84 ++-
>  drivers/net/dpaa/dpaa_oldev.c             | 772 ++++++++++++++++++++++
>  drivers/net/dpaa/meson.build              |  11 +-
>  drivers/net/dpaa/rte_pmd_dpaa_oldev.h     | 149 +++++
>  32 files changed, 2499 insertions(+), 477 deletions(-)
>  create mode 100644 drivers/net/dpaa/dpaa_oldev.c
>  create mode 100644 drivers/net/dpaa/rte_pmd_dpaa_oldev.h
> 

Summary of longish AI review..

Still open

Patch 15 — no response, and unchanged. getenv("DPAA_TX_TAILDROP_THRESHOLD") is still the fallback, and the RST is still a bullet list. This is the one item that got no reply at all, so it may just have been missed rather than rejected.

Patch 19 — marked fixed, but only the commit message was. U48_BY_HI16_LO32() still shifts hi by 32 with no cast, and BIT_SIZE/MAX_U48/HI16_OF_U48/LO32_OF_U48 are still unprefixed in fsl_bman.h. The new message also says the macros were "dropped" when they were moved into the installed header.

Patch 22 — genuinely improved. MAX_NUM_PORTS/MAX_NUM_SUBNETS are prefixed (the collision that actually mattered) and dpaa_ip4_addr_s → dpaa_ip_addr_s. Still unprefixed in the installed header: struct ip_pair_s, struct lgw_subnet_s, and the DPA_ISC_* macros. I'd narrow the remaining ask to just those.

Patch 24 — unchanged in substance, and the contradiction is now sharper than in v12: patch 13 deliberately made dpaa_mpool_finish() the only place rte_dpaa_bpid_info is freed, so that function calls dpaax_enter_destructor() and then rte_free() on EAL memory a few lines later. dpaa_finish() does the same — latch, then dpaa_fm_term() and rte_dpaa_portal_fq_close().

Patch 25 — new inconsistency introduced by the patch 14 fix: the release note still says "Added NXP DPAA DMA scatter-gather support" and "Added scatter-gather (SG) batching... enabled by default", which is exactly the claim patch 14's message was corrected to drop.

Patch 16 (ORP) — on "NXP customer applications use it": qman_enqueue_multi_orp() is tagged __rte_internal, so it isn't reachable from an application at all; internal symbols are restricted to DPDK's own libs and drivers. If NXP customer code is calling it, either it's building in-tree (in which case that user should come with the patch) or the tag is wrong. Worth putting back to them as a question rather than a demand.