NXP QorIQ Layerscape DPAA2 DPIO service
Support code for the Data Path I/O portal hardware inside NXP's QorIQ Layerscape ARM SoCs (such as the LS1088A and LS2088A), which is part of the second-generation Data Path Acceleration Architecture used to move packets between CPUs, the on-chip Ethernet engine, and other accelerators in networking and edge appliances.
recommendation
It should stay because the Layerscape chips that depend on it (LS1088A, LS2088A and relatives) are still listed as Active products by NXP in 2025, and the code itself saw upstream maintenance commits as recently as 2023 and 2024. No removal discussion turned up, though the analysis notes a couple of search tools timed out, so that conclusion is slightly hedged.
repository signals
sources
- lore.kernel.org
drivers/soc/fsl/dpio/dpio-service.c was still being touched by upstream code changes on 2024-10-31, indicating the code is not abandoned.
- lore.kernel.org
drivers/soc/fsl/dpio/dpio-driver.c saw upstream cleanup in 2023, showing maintenance activity rather than removal.
- cateee.net
CONFIG_FSL_MC_DPIO is the 'QorIQ DPAA2 DPIO driver' and remains present through current kernel series, confirming the directory is an in-tree hardware driver.
- nxp.com
NXP lists LS2088A as 'Active' and documents a DPAA2 reference manual for it, showing relevant DPAA2 hardware remained commercially active into the 2025 timeframe.
- nxp.com
NXP lists LS1088A as 'Active' and explicitly says the processor family uses DPAA2, supporting ongoing real-world deployment of DPAA2-based systems.
codex reasoning notes (technical)
Local code inspection via shell (`rg`) shows a real fsl-mc driver entry point and identifies this as DPAA2 DPIO service code. Lore evidence came from `lore_file_timeline` on dpio-driver.c and dpio-service.c; the cited lore URLs show maintenance in 2023 and 2024. Driver identity/support window came from web search to LKDDb. Deployment evidence came from web search to official NXP product pages showing LS2088A and LS1088A still marked Active and tied to DPAA2. I did not find removal evidence in the sampled lore timeline; a dedicated `lore_regex` removal-subject scan timed out and `lei` was sandbox-blocked, so 'no removal discussion' is an inference with reduced confidence.