On Thu, 22 Oct 2026 19:44:27 +0200
David Marchand <[email protected]> wrote:
> We have various checks in drivers about Rx and Tx multiqueue
> configuration.
> This series is a followup of the VMDq cleanup, finishing the job with
> DCB and RSS, and then rejecting invalid multiqueue configuration (useful
> from an ABI point of view).
>
> Some drivers misreported their capabilities.
> It is important drivers maintainers review carefully those changes.
>
Review: [PATCH v3 00/13] ethdev multi-queue mode validation
Applied cleanly on main (a635140ba9). Every commit builds with
-Dwerror=true for the touched drivers (except mvneta/mvpp2, no musdk)
and the touched apps and examples. check-doc-vs-code.sh full mode is
clean at every commit.
v2 items addressed: incremental mode now checks the changed drivers
(adding "VMDq = Y" to mlx5.ini is reported with a trusted commit),
intel directory detection (03/13), Cc: stable on 06, 09 and 11, RSS
commit message (12/13). Shared-directory rte_flow checks behave as
described: a doc-only item in cnxk_vf.ini is reported, and an item
missing from every ice doc is reported against ice.
One item left.
Patch 12/13 ethdev: check RSS availability
Warning: the release note still says RSS mode is always rejected. The
code rejects it only with more than one Rx queue and ignores it with
one:
* RSS mode (Rx ``RTE_ETH_MQ_RX_RSS_FLAG``) is rejected with more than
one Rx queue if the driver reports no ``flow_type_rss_offloads``,
and ignored with a single Rx queue.
