Re: [PATCH v2 0/6] net/nfb: ethernet enhancements

[Stephen Hemminger <stephen@networkplumber.org>]

On Tue, 17 Feb 2026 12:03:18 +0100
spinler@cesnet.cz wrote:

> From: Martin Spinler <spinler@cesnet.cz>
> 
> This set includes minor enhancements to the Ethernet
> configuration and overall usage.
> 
> ---
> Depends-on: series-37385 ("net/nfb: rework to real multiport")
> 
> v2:
> * P2/7: close MDIO handles in error path of nfb_nc_eth_init
> * P2/7: add bounds check for eth_node population
> * P2/7: add check for nfb_mdio_if_read_pma() return value
>         in nfb_mdio_cl45_pma_get_speed_capa()
> * P2/7: add include guard to nfb_mdio.h
> * P3/7: move nfb_eth_dev_set_link() (PMA configuration)
>         before MAC enable/disable call to avoid inconsistency on failure
> * P4/7: add check for nfb_mdio_if_write_pma() return value
> * P5/7: set dev_info->max_mtu
> * P5/7: add frame_len_max_cap to struct pmd_priv; used in MTU configuration
> * P5/7: extract nfb_mac_read_frame_len_max_cap() from nfb_eth_mtu_set()
> * P5/7: check for MTU overflow
> * P6/7: use direct value for stats->ierrors from rx_cntrs
> * P7/7: remove this patch completely and distribute release notes update
>         to matching patches
> 
> Martin Spinler (6):
>   net/nfb: use MAC address assigned to card
>   net/nfb: get correct link speed
>   net/nfb: support real link-up/down config
>   net/nfb: support setting RS-FEC mode
>   net/nfb: support setting Rx MTU
>   net/nfb: read total stats from macs
> 
>  doc/guides/rel_notes/release_26_03.rst |   3 +-
>  drivers/net/nfb/meson.build            |   1 +
>  drivers/net/nfb/nfb.h                  |   8 +
>  drivers/net/nfb/nfb_ethdev.c           | 247 +++++++++++++++++++++----
>  drivers/net/nfb/nfb_mdio.c             |  44 +++++
>  drivers/net/nfb/nfb_mdio.h             |  39 ++++
>  drivers/net/nfb/nfb_stats.c            |  40 ++--
>  7 files changed, 335 insertions(+), 47 deletions(-)
>  create mode 100644 drivers/net/nfb/nfb_mdio.c
>  create mode 100644 drivers/net/nfb/nfb_mdio.h
> 

Queued all these to next-net with small change to avoid overflow when
computing mtu.