On Thu,  1 Oct 2026 16:54:04 +0530
Hemant Agrawal <[email protected]> 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, non
> fmX-macY shared Ethernet naming, and DMA scatter-gather and
> errata-workaround devargs. Documentation and release notes are updated
> accordingly.
> 
> v19: 
>       addressed AI review comments


No blockers, but there are still some things which are worth addressing.
I demangled what AI was trying to say but still seems confuzuled

Patch 10/26: The new qman_pending_fq_by_cgrid_range() can check a
  whole range of CGRIDs in one scan of the FQID space, but the series
  never uses it that way.

  dpaa_eth_dev_close() loops over every Rx and Tx CGR and calls
  dpaa_cgr_stale_fq_cleanup() for each one.  That calls
  qman_pending_fq_by_cgrid(), which is the range function with a
  range of one CGRID.  So closing a port still scans the full FQID
  space nb_rx_queues + MAX_DPAA_CORES times.

  The Rx CGRIDs are contiguous, and so are the Tx CGRIDs; close
  already releases them with qman_release_cgrid_range().  Make
  dpaa_cgr_stale_fq_cleanup() take a CGRID range, call it once for
  the Rx set and once for the Tx set, then delete the CGRs.  If that
  is not wanted, drop qman_pending_fq_by_cgrid_range() and its
  RTE_EXPORT_INTERNAL_SYMBOL, since nothing outside bus/dpaa uses it.

Patch 24/26 Symbol defined but never used.
 $ git grep rte_dpaa_bus_oldev_enabled
drivers/bus/dpaa/bus_dpaa_driver.h:int rte_dpaa_bus_oldev_enabled(void);
drivers/bus/dpaa/dpaa_bus.c:RTE_EXPORT_INTERNAL_SYMBOL(rte_dpaa_bus_oldev_enabled)
drivers/bus/dpaa/dpaa_bus.c:rte_dpaa_bus_oldev_enabled(void)

* Missing doxygen
  - rte_pmd_dpaa_oldev.h is not in doc/api/doxy-api-index.md.

3. New functions should be in release note.


Pre-existing, not introduced here: in dpaa_dev_init() the "FMC
initializes failed" path does goto free_rx without setting ret, so
probe can succeed with no Rx queues

Reply via email to