topic Re: LS1017A CPU based board cannot find PHY in u-boot in Processor Expert Software

LS1017A CPU based board cannot find PHY in u-boot

xiaosa_en — Wed, 20 Mar 2024 12:39:32 GMT

Dear Experts:

We design a board using LS1017A CPU, connect 2 MARVELL 88e1512 PHY to mac0 and mac1 of the TSN switch.

The serdes config is 99cc. Our SDK used is lsdk21.08.

In u-boot we cannot find PHY, "mii info" command or "mdio list" don't show any devices.

Here's boot log: You can find the error at the last, "felix-switch: Failed to register fixed-link for CPU port"

NOTICE: BL2: Booting BL31
NOTICE: BL31: v2.4(release):LSDK-21.08-1-g5d950d6a2
NOTICE: BL31: Built : 14:15:43, Mar 15 2024
NOTICE: Welcome to ls1028ardb BL31 Phase

U-Boot 2021.04-00001-gec02921ea0-dirty (Mar 19 2024 - 14:16:02 +0800)

SoC: LS1017AE Rev1.0 (0x870b2410)
Clock Configuration:
CPU0(A72):800 MHz
Bus: 300 MHz DDR: 1300 MT/s
Reset Configuration Word (RCW):
00000000: 2000340c 00000030 00000000 00000000
00000010: 00000000 018f0000 0030c000 00000000
00000020: 02003150 00002580 00000000 0020824e
00000030: 00001001 00000008 00000000 00000000
00000040: 00000000 00000000 00000000 00000000
00000050: 00000000 00000000 00000000 00000000
00000060: 00000000 00000000 100e7512 00000000
00000070: cc990000 00000000
Model: NXP Layerscape 1028a RDB Board
Board: LS1017AE Rev1.0-unknown, Version: O, boot from SD
FPGA: v255 (unknown)
SERDES1 Reference : Clock1 = 161.13MHz Clock2 = 161.13MHz
DRAM: 958 MiB
DDR 958 MiB (DDR4, 32-bit, CL=10, ECC off)
GPIO set OK
Using SERDES1 Protocol: 52377 (0xcc99)
PCIe1: pcie@3400000 disabled
PCIe2: pcie@3500000 Root Complex: no link
WDT: Started with servicing (60s timeout)
MMC: FSL_SDHC: 0, FSL_SDHC: 1
Loading Environment from MMC... OK
EEPROM: Read failed.
In: serial
Out: serial
Err: serial
SEC0: RNG instantiated
Net:
Warning: enetc-2 (eth1) using random MAC address - 56:95:41:40:e4:ed
eth1: enetc-2felix-switch felix-switch: Failed to register fixed-link for CPU port
felix-switch felix-switch: Failed to register fixed-link for CPU port

I don't know if our dts files or other config is wrong
Really looking forward to your help! Many thanks!

dts files of U-boot are attached.

BR,

Yuxiao Ji

Re: LS1017A CPU based board cannot find PHY in u-boot

yipingwang — Fri, 22 Mar 2024 08:01:24 GMT

Please define mdio PHY as the following in dts files.

In fsl-ls1028a.dtsi
enetc_mdio_pf3: mdio@0,3 {
compatible = "fsl,enetc-mdio";
reg = <0x000300 0 0 0 0>;
#address-cells = <1>;
#size-cells = <0>;
};

In fsl-ls1028a-rdb.dts

&enetc_mdio_pf3 {

status = "okay";

rdb_phy0: phy@2 {

reg = <2>;

};

sw_phy0: phy@3 {

reg = <0x03>;

};

sw_phy1: phy@4 {

reg = <0x04>;

};

sw_phy2: phy@5 {

reg = <0x05>;

};

sw_phy3: phy@6 {

reg = <0x06>;

};

Re: LS1017A CPU based board cannot find PHY in u-boot

xiaosa_en — Mon, 25 Mar 2024 03:13:21 GMT

Hi yiping:

We have modified the dts files, now it's better, u-boot logs changed to:

Warning: enetc-2 (eth1) using random MAC address - 7a:02:b7:ec:57:6d
eth1: enetc-2Could not get PHY for emdio-3: addr 3

Warning: swp0 (eth2) using random MAC address - da:51:9b:a8:11:cf
, eth2: swp0
Warning: swp1 (eth3) using random MAC address - fe:b4:dd:4f:b8:0f
, eth3: swp1

It seems that we can find PHY at address 4, but PHY @3 is still missing

And the mdio and mii command output is:

=> mdio list
enetc-2:
emdio-3:
4 - Generic PHY <--> swp1
felix-switch:

=> mii info
PHY 0x00: OUI = 0x20F9, Model = 0x00, Rev = 0x00, 10baseT, HDX
PHY 0x01: OUI = 0x20F9, Model = 0x00, Rev = 0x00, 10baseT, HDX

We also did a loop back traffic test, looping back macs at swp0 and swp1, then do ping test, also failed...

Looking forward to your reply, thanks!

BR,

Yuxiao Ji

Re: LS1017A CPU based board cannot find PHY in u-boot

yipingwang — Mon, 25 Mar 2024 06:42:45 GMT

Please add the following in configs/ls1028ardb_tfa_defconfig, then rebuild u-boot.

CONFIG_PHY_MARVELL=y

Please check your hardware design whether swp0 connects to PHY address 0x3.

Re: LS1017A CPU based board cannot find PHY in u-boot

xiaosa_en — Thu, 28 Mar 2024 07:56:27 GMT

hi yiping:

sorry for the delay, but adding the driver didn't solve our problem...

maybe something wrong with our hardware, will check that and if any findings I will contact you later

BR,

Yuxiao Ji