[meta-xilinx] ZynqMP - Multiple ETH - Not detected PHY?

[Giordon Stark]

Hi all,

So I definitely got somewhere, but I think there might be issues with
u-boot. A patch like this

+&gem3 {^M
+    local-mac-address = [00 0a 35 00 00 01];
+    phy-handle = <&phy1>;
+    phy1: phy at 7 {
+        reg = <0x7>;
+        ti,rx-internal-delay = <0x8>;
+        ti,tx-internal-delay = <0xa>;
+        ti,fifo-depth = <0x1>;
     };
 };

works perfectly to get GEM3 working. This also works if I do it for gem2
(similar, but with the phy address at 0x4). If I use one or the other,
u-boot doesn't seem to have a problem picking things up and running with
it... but if I *enable both* GEM2 and GEM3 in my project (since we will
require two ETH later), I get errors like

Xilinx Zynq MP First Stage Boot Loader
Release 2017.2   Dec  5 2017  -  15:52:54
NOTICE:  ATF running on XCZU19EG/silicon v3/RTL5.1 at 0xfffea000, with PMU
firmware
NOTICE:  BL31: Secure code at 0x0
NOTICE:  BL31: Non secure code at 0x8000000
NOTICE:  BL31: v1.3(release):7d1a673
NOTICE:  BL31: Built : 14:42:17, Jan  3 2018
PMUFW:  v0.3


U-Boot 2017.01 (Jan 09 2018 - 13:37:43 -0600) gFEX Prototype v3 (ZynqMP SoC)

I2C:   ready
DRAM:  16 GiB
EL Level:       EL2
Chip ID:        xczu19eg
Using default environment

In:    serial at ff000000
Out:   serial at ff000000
Err:   serial at ff000000
Bootmode: QSPI_MODE
Net:   ZYNQ GEM: ff0d0000, phyaddr 4, interface rgmii-id
i2c_mux_set: could not set mux: id: 5 chip: 74 channel: 0
I2C EEPROM MAC address read failed

*Warning: ethernet at ff0d0000 (eth0) using random MAC address -
12:92:73:12:c0:41*
*eth0: ethernet at ff0d0000ZYNQ GEM: ff0e0000, phyaddr 7, interface rgmii-id*
*PHY is not detected*
*GEM PHY init failed*

Hit any key to stop autoboot:  0
Invalid bus 0 (err=-19)
Failed to initialize SPI flash at 0:0 (error -19)
*ZynqMP> mdio list*
*eth0:*
*4 - Marvell 88E1118R <--> ethernet at ff0d0000*
*eth1:*
ZynqMP>

What's strange about this error is that if I enable GEM2 or I enable GEM3
separately in my project, and change my device tree according, bitbake, and
then program the flash -- I'm seeing zero problems with the PHY detection.
It is only when I have *both* PHY enabled that I see a problem. Is this an
issue with u-boot? The `dhcp` command does work fine in the sense that I
have a working PHY to send TFTP communications over -- and I haven't booted
linux yet [and this may not be a problem inside linux, but only inside
u-boot].

Thanks,

Giordon

eth1:
ZynqMP> dhcp
BOOTP broadcast 1
BOOTP broadcast 2
DHCP client bound to address 192.168.1.123 (279 ms)
*** Warning: no boot file name; using 'C0A8017B.img'
Using ethernet at ff0d0000 device
TFTP from server 0.0.0.0; our IP address is 192.168.1.123; sending through
gateway 192.168.1.1


On Wed, Jan 3, 2018 at 9:29 PM Oleg K Dzhimiev <oleg at elphel.com> wrote:

> Hi,
>
> I had this problem with the older Zynq7 series board
>> https://github.com/kratsg/meta-l1calo/commit/940fa27aa6c9a456ccca7ec60351c26287f9f671 and
>> this was the patch to get the PHY working.
>
> If that patch does not work. Try changing:
>
>> compatible = "marvell,88e1116r";
>
> to
>
>> compatible = "ethernet-phy-ieee802.3-c22";
>
>
> Also, see Documentation/devicetree/bindings/net/phy.txt
> <https://github.com/Xilinx/linux-xlnx/blob/master/Documentation/devicetree/bindings/net/phy.txt>
> Example: zynq-parallella.dts
> <https://github.com/Xilinx/linux-xlnx/blob/master/arch/arm/boot/dts/zynq-parallella.dts>
>
> Regards,
> Oleg
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.yoctoproject.org/pipermail/meta-xilinx/attachments/20180110/6633ae4b/attachment.html>