PHY KSZ9031 not working with Jetpack 5.0.2

[sevm89]

Dear NVidia Team

In the topic JetPack 5.0 DP - Ethernet Port "RGMII" not working with KSZ9031 we already mentioned that our PHY KSZ9031 does not work anymore with the new JetPack Version. With JetPack 4.6 and older, we did not see any problems.
Now we started testing with the release JetPack 5.0.2 and face again the same problem.
Do we need to change something in our configuration?

Here is our device tree:
out.dts (263.7 KB)

And here the dmesg output:
dmesg.txt (107.3 KB)

We specifically see these errors:

[ 8.251102] nvethernet 2490000.ethernet: Adding to iommu group 31
[ 8.251314] nvethernet 2490000.ethernet: failed to read skip mac reset flag, default 0
[ 8.251319] nvethernet 2490000.ethernet: failed to read MDIO address
[ 8.251325] nvethernet 2490000.ethernet: setting to default DMA bit mask
[ 8.251328] nvethernet 2490000.ethernet: set default TXQ to TC mapping
[ 8.251331] nvethernet 2490000.ethernet: Setting default PTP RX queue
[ 8.263234] nvethernet 2490000.ethernet: failed to get eqos_rx_m clk
[ 8.263240] nvethernet 2490000.ethernet: failed to get eqos_rx_input clk
[ 8.263244] nvethernet 2490000.ethernet: failed to get eqos_tx_divider clk
[ 8.271840] nvethernet 2490000.ethernet: Ethernet MAC address: 48:b0:2d:3c:8c:8e
[ 8.313819] nvethernet 2490000.ethernet: Macsec not supported
[ 8.326269] nvethernet 2490000.ethernet: eth5 (HW ver: 50) created with 1 DMA channels
[ 14.403333] using random self ethernet address
[ 14.411818] using random host ethernet address
[ 15.242098] using random self ethernet address
[ 15.243734] using random host ethernet address
[ 15.629056] nvethernet 2490000.ethernet: failed to connect PHY
[ 15.630774] net eth5: ether_open: Cannot attach to PHY (error: -19)

If you need further information, let us know.
Thank you.

Best regards

[WayneWWW]

Please refer to the bring up guide from Orin for switch and see if you configured everything.

https://docs.nvidia.com/jetson/archives/r35.1/DeveloperGuide/text/HR/JetsonModuleAdaptationAndBringUp/JetsonAgxOrinSeries.html#for-switch

[sevm89]

We added the following lines without success:

/* 1:10G, 0:5G /
nvidia,uphy-gbe-mode = <1>;
/ 0:XFI 10G, 1:XFI 5G, 2:USXGMII 10G, 3:USXGMII 5G */
nvidia,phy-iface-mode = <0>

What are the settings for 1 GBE and RGMII?

Compared to the original device tree of the xavier agx devkit, we just changed the PHY node to the following:

               phy@0 {

                       compatible = "micrel,ksz9031";
                       reg = <0x0>;
                       interrupt-parent = <0xb>;
                       interrupts = <0x34 0x8>;
                       micrel,copper-mode;
                       micrel,reg-init = <0x3 0x12 0x7fff 0x880>;
                       phandle = <0x21>;
              };

and added the fixed-link node:

fixed-link {
speed = <0x3e8>;
full-duplex;
};

Thank you.

[sevm89]

Hi WayneWWW

We found the problem, we had to comment the following line:

// phy-handle = <&phy0>;

After that the Interface works.

Do we have to worry about this messages while booting:

[ 8.999188] arm-smmu 12000000.iommu: Unhandled context fault: fsr=0x80000402, iova=0x859719400, fsynr=0x40000, cbfrsynra=0x81b, cb=0
[ 8.999230] arm-smmu 12000000.iommu: Unhandled context fault: fsr=0x80000402, iova=0x859717e00, fsynr=0x140000, cbfrsynra=0x1b, cb=0
[ 9.028448] mc-err: unknown mcerr fault, int_status=0x00000000, ch_int_status =0x00000000, hubc_int_status=0x00000000 sbs_int_status=0x00000000, hub_int_statu s=0x00000000

Thank you.

[kayccc]

@sevm89

Do we have to worry about this messages while booting:

Any abnormal behavior found on your testing?

[sevm89]

so far not no.