[EXT] Re: [PATCH v3 3/4] arm64: dts: cn913x: add device trees for topology B boards

Marcin Wojtas mw at semihalf.com
Mon Dec 21 12:46:57 EST 2020


pon., 21 gru 2020 o 18:42 Stefan Chulski <stefanc at marvell.com> napisał(a):
>
> > External Email
> >
> > ----------------------------------------------------------------------
> > > -   cp0_sfp_eth0: sfp-eth at 0 {
> > > -           compatible = "sff,sfp";
> > > -           i2c-bus = <&cp0_sfpp0_i2c>;
> > > -           los-gpio = <&cp0_module_expander1 11
> > GPIO_ACTIVE_HIGH>;
> > > -           mod-def0-gpio = <&cp0_module_expander1 10
> > GPIO_ACTIVE_LOW>;
> > > -           tx-disable-gpio = <&cp0_module_expander1 9
> > GPIO_ACTIVE_HIGH>;
> > > -           tx-fault-gpio = <&cp0_module_expander1 8
> > GPIO_ACTIVE_HIGH>;
> > > -           /*
> > > -            * SFP cages are unconnected on early PCBs because of an
> > the I2C
> > > -            * lanes not being connected. Prevent the port for being
> > > -            * unusable by disabling the SFP node.
> > > -            */
> > > -           status = "disabled";
> >
> > Does the B also have the broken SFP cage?
>
> A and B topologies use same board. So B  also has broken SFP cage.
>
> > Is there a newer revision which fixes the PCB?
>
> We don't have newer revision for DB boards.
>

Question - in such case can we use just 10G interfaces with the
link-IRQ instead (now the ports are disabled)? This is what I did in
edk2 dtb/ACPI and it works just fine for me.

Best regards,
Marcin



More information about the linux-arm-kernel mailing list