[PATCH phy-next 0/5] Lynx 28G SerDes: 25GbE support

Josua Mayer josua at solid-run.com
Wed May 13 07:51:02 PDT 2026


Am 11.05.26 um 17:00 schrieb Vladimir Oltean:
> This is the remainder of "Lynx 28G improvements part 2", previously
> submitted here:
> https://lore.kernel.org/linux-phy/176850672122.1082429.444623229961712368.robh@kernel.org/
>
> but split up into smaller portions (merged separately):
> - https://lore.kernel.org/linux-phy/20260226182853.1103616-1-vladimir.oltean@nxp.com/
> - https://lore.kernel.org/linux-phy/20260321011451.1557091-1-vladimir.oltean@nxp.com/
>
> What remains is the highlight feature (patch 5/5): support for dynamic
> protocol changes to/from 25GBase-R, required by SFP28 modules. These are
> used with the NXP LX2160A and the (Ethernet) dpaa2-mac consumer.
>
> Patches 1-4 handle the following situation: with current device trees,
> the driver will think 25GBase-R will work on a lane, but it may work or
> may not. This is because not all lanes support this protocol. So we
> modify the SerDes compatible strings to identify them, and we use a
> driver-internal database to figure out on which lanes does each SerDes
> instance support this protocol.
>
> On current device trees, 25GbE is not supported.
>
> Detailed change log in patches. Summary:
> - reworded commit messages
> - change match condition from dt-bindings change 2/5
> - patch 3/5 is new (reject probing on devices with no OF node)
>
> Ioana Ciornei (1):
>   phy: lynx-28g: add support for 25GBASER
>
> Vladimir Oltean (4):
>   dt-bindings: phy: lynx-28g: add compatible strings per SerDes and
>     instantiation
>   dt-bindings: phy: lynx-28g: add constraint on LX2162A lane indices
>   phy: lynx-28g: require an OF node to probe
>   phy: lynx-28g: probe on per-SoC and per-instance compatible strings
>
>  .../devicetree/bindings/phy/fsl,lynx-28g.yaml |  48 +++-
>  drivers/phy/freescale/phy-fsl-lynx-28g.c      | 221 +++++++++++++++++-
>  2 files changed, 254 insertions(+), 15 deletions(-)
>
Tested-by: Josua Mayer <josua at solid-run.com>


More information about the linux-phy mailing list