Am Wed, Jun 02, 2021 at 10:37:36PM +0200 schrieb Mark Kettenis:
> Linux folks changed the device tree to use 10gbase-r instead of
> 10gbase-kr since "it is more correct". Then the UEFI folks synched
> their device trees to Linux and the 10G ports broke. So accept both
> in the code.
>
> ok?
>
Sigh. sure, ok patrick@.
Though you could probably put it into a single if (!strncmp ||
!strncmp), but I don't mind either way.
>
> Index: dev/fdt/if_mvpp.c
> ===================================================================
> RCS file: /cvs/src/sys/dev/fdt/if_mvpp.c,v
> retrieving revision 1.44
> diff -u -p -r1.44 if_mvpp.c
> --- dev/fdt/if_mvpp.c 12 Dec 2020 11:48:52 -0000 1.44
> +++ dev/fdt/if_mvpp.c 2 Jun 2021 20:34:30 -0000
> @@ -1354,7 +1354,9 @@ mvpp2_port_attach(struct device *parent,
>
> phy_mode = malloc(len, M_TEMP, M_WAITOK);
> OF_getprop(sc->sc_node, "phy-mode", phy_mode, len);
> - if (!strncmp(phy_mode, "10gbase-kr", strlen("10gbase-kr")))
> + if (!strncmp(phy_mode, "10gbase-r", strlen("10gbase-r")))
> + sc->sc_phy_mode = PHY_MODE_10GBASER;
> + else if (!strncmp(phy_mode, "10gbase-kr", strlen("10gbase-kr")))
> sc->sc_phy_mode = PHY_MODE_10GBASER;
> else if (!strncmp(phy_mode, "2500base-x", strlen("2500base-x")))
> sc->sc_phy_mode = PHY_MODE_2500BASEX;
>
