[PATCH 2/2] arm: mvebu: add support for Methode eDPU

[Pali Rohár]

On Monday 09 May 2022 13:07:00 Robert Marko wrote:
> On Sat, May 7, 2022 at 4:03 PM Pali Rohár <pali at kernel.org> wrote:
> >
> > On Friday 06 May 2022 20:01:40 Robert Marko wrote:
> > > Methode eDPU is an Armada 3720 power board based on the Methode uDPU.
> > >
> > > They feature the same CPU, RAM, and storage as well as the form factor.
> > >
> > > However, eDPU only has one SFP slot plus a copper G.hn port which does not
> > > work under U-boot.
> > >
> > > In order to reduce duplication, split the uDPU DTS into a common one.
> > >
> > > Signed-off-by: Robert Marko <robert.marko at sartura.hr>
> >
> > Send also these changes to Linux kernel as now all 3720 files are same
> > in Linux and U-Boot. We do not want to handle divergence again...
> 
> Yes, I planned to send it the same day but only got around to sending
> it to Linux today.

Perfect!

> > > +&eth1 {
> > > +     fixed-link {
> > > +             speed = <1000>;
> > > +             full-duplex;
> > > +     };
> > > +};
> >
> > Is this U-Boot specific fixed-link property still needed? IIRC Marek's
> > patches for fixing mvneta were merged into master.
> 
> Unfortunately, it's still required as there is no ethernet PHY-s
> present, only an SFP slot.

Ok!

> I am aware of Mareks fixes, he did a nice cleanup of the driver.
> >

> > > +CONFIG_AHCI=y
> >
> > There is SATA/AHCI on this board? I guess that not as it has 2 ethernet
> > serdeses and third seredes is in usb mode. So there is no option for
> > SATA, right?
> 
> That's right, there is no SATA, it's a leftover from the uDPU defconfig.
> >
> > > +CONFIG_CMD_PCI=y
> >
> > And neither PCIe, right?
> 
> Correct, more leftovers from uDPU defconfig.
> >
> > > +CONFIG_E1000=y
> >
> > And neither E1000 PCIe card?
> 
> Correct, more leftovers from uDPU defconfig.
> I have removed all of these in v2.

Ok!

> > Btw, cannot you detect at runtime on which variant of board (uDPU vs
> > eDPU) are you running? If yes, this would allow to use just one u-boot
> > binary and therefore also reduce number of defconfig files. Because
> > u-boot already supports inclusion of more DT files into binary and can
> > do decision at runtime which DT one to use.
> 
> I thought about this, but I don't know how to differentiate them really as there
> is no EEPROM or something like that, or an I2C peripheral difference.

Well, if is is not possible then stick with two configs.

Anyway, Marek already wrote autodetection of more A3720 boards into the
CZ.NIC WTMI firmware. And uDPU is detected based on non-existence of PHY:
https://gitlab.nic.cz/turris/mox-boot-builder/-/blob/master/wtmi/board.c