Re: P2020 with BCM53115

[Abraham Yaniv <yaniv_abraham@yahoo.com>]

> Scott Wood <scottwood <at> freescale.com> writes:
> 
> On Mon, 20 Dec 2010 13:08:55 -0500
> "Dry, Craig" <cdry <at> mc.com> wrote:
> 
> > I am trying to bring up a P2020 board which uses the Broadcom BCM53115 
switch in unmanaged mode.  The board is
> patterned after the P2020RDB, except the Vitesse switch has been replaced with 
a BCM53115. There is no
> MDIO connection to the switch, but there is  an SPI connection available if 
needed.
> > 
> > When running Uboot, the BCM53115 works just fine, as I am able to 
download the Linux Kernel and
> P2020RDB.dtb across the network without issue.   But as Linux is booting up, I 
get the message:
> > 
> > mdio_bus mdio <at> ffe24520: error probing PHY at address 0
> > mdio_bus mdio <at> ffe24520: error probing PHY at address 1
> > 
> > I expect these errors occur because the P2020RDB.dts file has these entries.
> > 
> > I'm not sure what to try next to get Linux to use the BCM53115 switch in 
unmanaged mode.
> 
> See the fixed-link property in
> Documentation/powerpc/dts-bindings/fsl/tsec.txt
> 
> -Scott
> 

Hi there Craig!
I have the same setup (Freescale P2020 CPU with Broadcom BCM53115 switch) 
but I'm having a hard time bringing it up.
The P2020's eTSEC2 is connected to port 5 (WAN) of the BCM53115.
I managed to make a link up status on the BCM53115 port 5 but I believe this is a
false link up because no traffic is passing between the P2020 and the BCM53115
and the BCM53115 SerDes/SGMII status register says there is a False Carrier 
situation.
The P2020 eTSEC is configured to SGMII/TBI (ECNTRL[SGMIIM]=1, 
ECNTRL[TBIM]=1, and MACCFG2[I/F]=10).
Is there a chance you can share with me what is the configuration required 
to my BCM53115 to make a true link up?
Thanks a million,
Yaniv