somehow ethernet stopped working on any images I’ve tried.
Also, even without SD card and eMMC card on board - if I power rockpi - one led on Ethernet port (yellow) is on.
the same behavior with connected card and booting image - yellow led in on but no ethernet in OS.
Is it normal situation that yellow led is on with powered board?
UPD: yellow led is ON even if lan cable disconnected.
so I think it is some HW issue.
Is there any info how to check circuit and-or check some components on PCB to know what is broken and maybe replace it?
any advices would be helpful.
thanks in advance.
First, make sure your board is 4C. Recently we changed the ethernet phy from 8211E to 8211F because of the chip shortage and EoL. For 8211E, if powering on without software run, it’s definitely hw issue.
is it possible that RTL chip has been broken after running “wrong” image - I’ve compiled couple Armbian Hirsute images for 4C and tested it on my board…
And during first run I noticed that LAN is not working and decide that it maybe because of new image.
And then I realized that Lan now working at all…
thank you!
Im really appreciate your support)
I’m from Russia, so I think it would be faster for me to order chip in Aliexpress. it should take about two weeks for shipping,
I just need confirmation that this chip should be QNF-48 package like on link above.
but if you wiling to send me it - I’m would very grateful for it.
Sorry for bringing up this old thread again, but I just realized that I still have this board.
The RTL chip was replaced at the service center with the RTL8211E chip that you sent me some time ago. However, Armbian still cannot detect the Ethernet interface.
here is dmesg part: root@rockpi-4c:~# dmesg | grep -Ei ‘gmac|stmmac|eth|rtl’ [ 0.000000] psci: probing for conduit method from DT. [ 1.526902] igb: Intel(R) Gigabit Ethernet Network Driver [ 4.229210] rk_gmac-dwmac fe300000.ethernet: IRQ eth_wake_irq not found [ 4.229243] rk_gmac-dwmac fe300000.ethernet: IRQ eth_lpi not found [ 4.229427] rk_gmac-dwmac fe300000.ethernet: PTP uses main clock [ 4.229752] rk_gmac-dwmac fe300000.ethernet: clock input or output? (input). [ 4.229771] rk_gmac-dwmac fe300000.ethernet: TX delay(0x28). [ 4.229787] rk_gmac-dwmac fe300000.ethernet: RX delay(0x11). [ 4.229809] rk_gmac-dwmac fe300000.ethernet: integrated PHY? (no). [ 4.229913] rk_gmac-dwmac fe300000.ethernet: cannot get clock clk_mac_speed [ 4.229927] rk_gmac-dwmac fe300000.ethernet: clock input from PHY [ 4.242851] rk_gmac-dwmac fe300000.ethernet: init for RGMII [ 4.246909] rk_gmac-dwmac fe300000.ethernet: User ID: 0x10, Synopsys ID: 0x35 [ 4.246951] rk_gmac-dwmac fe300000.ethernet: DWMAC1000 [ 4.246966] rk_gmac-dwmac fe300000.ethernet: DMA HW capability register supported [ 4.246978] rk_gmac-dwmac fe300000.ethernet: RX Checksum Offload Engine supported [ 4.246989] rk_gmac-dwmac fe300000.ethernet: COE Type 2 [ 4.247001] rk_gmac-dwmac fe300000.ethernet: TX Checksum insertion supported [ 4.247012] rk_gmac-dwmac fe300000.ethernet: Wake-Up On Lan supported [ 4.247161] rk_gmac-dwmac fe300000.ethernet: Normal descriptors [ 4.247175] rk_gmac-dwmac fe300000.ethernet: Ring mode enabled [ 4.247186] rk_gmac-dwmac fe300000.ethernet: Enable RX Mitigation via HW Watchdog Timer [ 4.339482] rk_gmac-dwmac fe300000.ethernet: No PHY found [ 4.340140] rk_gmac-dwmac fe300000.ethernet: stmmac_dvr_probe: MDIO bus (id: 0) registration failed [ 8.971423] Driver ‘hci_uart_qca’ needs updating - please use bus_type methods [ 18.021762] Bluetooth: BNEP (Ethernet Emulation) ver 1.3
root@rockpi-4c:~# uname -a Linux rockpi-4c 5.15.76-rockchip64 #trunk SMP PREEMPT Tue Nov 1 03:09:53 UTC 2022 aarch64 aarch64 aarch64 GNU/Linux root@rockpi-4c:~# root@rockpi-4c:~# cat /etc/os-release NAME=“Ubuntu” VERSION=“20.04.6 LTS (Focal Fossa)” ID=ubuntu ID_LIKE=debian PRETTY_NAME=“Ubuntu 20.04.6 LTS” VERSION_ID=“20.04” HOME_URL=“``https://www.ubuntu.com/” SUPPORT_URL=“``https://help.ubuntu.com/” BUG_REPORT_URL=“``https://bugs.launchpad.net/ubuntu/” PRIVACY_POLICY_URL=“``https://www.ubuntu.com/legal/terms-and-policies/privacy-policy” VERSION_CODENAME=focal UBUNTU_CODENAME=focal
I reinstalled Armbian from latest image from here
Armbian_22.11.0-trunk_Rockpi-4c_focal_current_5.15.76_xfce_desktop.img.xz
and still only wifi adapter in system..
I cannot really ask for more support since so much time has passed since I created this topic But maybe someone could advise how to bring the Ethernet module back to life?
thank you for your reply RadxaYuntian
sure, Ive already checked these images and for example now it on Armbian_community_26.2.0-trunk.886_Rockpi-4c_resolute_current_6.18.28_xfce_desktop.img.xz
_ _ _ _ _
/\ _ _ _ __ | |__()__ _ _ _ __ ___ _ __ _ __ _ _ _ _ () | _ _
/ _ | '| ’ | ’ \ / ` | ’ \ / / _ \ ’ | ’ \ || | ’ | | | || |
// __| ||||./_,|||| __/|||||_,|||||__|_, |
|__/
v26.2 rolling for Rockpi 4C running Armbian Linux 6.18.28-current-rockchip64
Packages: Ubuntu stable (resolute)
Support: for advanced users (rolling release)
IPv4: (LAN) 192.168.1.202 (WAN) 109.51.117.225
IPv6: (LAN) fdd6:2a64:8ccd:0:122c:6bff:fe4d:ae74
Performance:
Load: 2% Uptime: 19h 48m
Memory usage: 13% of 3.76G
CPU temp: 38°C Usage of /: 26% of 14G
RX today: 2 MiB
Tips:
What’s new in Armbian v26.2? https://tinyurl.com/4a4ktj3u
Commands:
Configuration: armbian-config
Monitoring : htop
Last login: Mon May 11 21:00:09 2026 from 192.168.1.100
root@rockpi-4c:~# uname -a
Linux rockpi-4c 6.18.28-current-rockchip64 #1 SMP PREEMPT Fri May 8 06:40:23 UTC 2026 aarch64 GNU/Linux
root@rockpi-4c:~#
root@rockpi-4c:~# dmesg | grep eth
[ 0.000000] psci: probing for conduit method from DT.
[ 3.472616] rk_gmac-dwmac fe300000.ethernet: IRQ eth_wake_irq not found
[ 3.472629] rk_gmac-dwmac fe300000.ethernet: IRQ sfty not found
[ 3.472673] rk_gmac-dwmac fe300000.ethernet: Deprecated MDIO bus assumption used
[ 3.472712] rk_gmac-dwmac fe300000.ethernet: PTP uses main clock
[ 3.472836] rk_gmac-dwmac fe300000.ethernet: integrated PHY? (no).
[ 3.477876] rk_gmac-dwmac fe300000.ethernet: init for rgmii with delay (tx: 0x28, rx: 0x11)
[ 3.479795] rk_gmac-dwmac fe300000.ethernet: User ID: 0x10, Synopsys ID: 0x35
[ 3.479816] rk_gmac-dwmac fe300000.ethernet: DWMAC1000
[ 3.479822] rk_gmac-dwmac fe300000.ethernet: DMA HW capability register supported
[ 3.479827] rk_gmac-dwmac fe300000.ethernet: RX Checksum Offload Engine supported
[ 3.479832] rk_gmac-dwmac fe300000.ethernet: COE Type 2
[ 3.479836] rk_gmac-dwmac fe300000.ethernet: TX Checksum insertion supported
[ 3.479841] rk_gmac-dwmac fe300000.ethernet: Wake-Up On Lan supported
[ 3.479885] rk_gmac-dwmac fe300000.ethernet: Normal descriptors
[ 3.479890] rk_gmac-dwmac fe300000.ethernet: Ring mode enabled
[ 3.479894] rk_gmac-dwmac fe300000.ethernet: Enable RX Mitigation via HW Watchdog Timer
[ 3.558565] rk_gmac-dwmac fe300000.ethernet: No PHY found
[ 3.558948] rk_gmac-dwmac fe300000.ethernet: error -ENODEV: MDIO bus (id: 0) registration failed
[ 24.798954] rockchip-pm-domain ff310000.power-management:power-controller: sync_state() pending due to fe300000.ethernet
root@rockpi-4c:~#
and RTL chip was very hot BEFORE replacing.. now it is ok.
Also I checked this issue with ChatGPR and Cursor - there is short report from Cursor for example:
Short Investigation Report
What was done
Verified network interfaces: only lo and wlan0, no eth0.
Collected kernel logs: rk_gmac-dwmac initializes, then fails with No PHY found and MDIO bus ... registration failed.
Checked live DT node /proc/device-tree/ethernet@fe300000: Ethernet enabled (status=okay, phy-mode=rgmii), but no explicit PHY/MDIO child nodes.
Decompiled active DTB and inspected Ethernet section: GMAC config present, but still no explicit PHY node (ethernet-phy@...) / phy-handle.
Checked runtime MDIO bus: /sys/bus/mdio_bus/devices is empty (no PHY detected at any address).
Checked GPIO debug: reset-related GPIO exists, but no evidence of successful PHY bring-up from software logs.
Checked pinctrl debug: GMAC groups exist in definitions, but no active GMAC/MDIO pin usage shown at runtime.
Checked clocks: GMAC-related clocks listed, so MAC block is present from kernel perspective.
Current conclusion
Main failure point is PHY not responding on MDIO (hardware-level symptom).
Linux MAC driver loads, but cannot discover RTL8211E, so eth0 is never created.
What still needs to be done
Hardware checks on replaced RTL8211E:
PHY rails (3.3V / core rail)
RESET_N behavior (must go high after reset pulse)
25 MHz reference clock presence
MDC toggling + MDIO bidirectional traffic during boot
strap resistors / PHY address straps and interface straps
continuity + shorts on MDIO/MDC/RESET traces and nearby passives
Possible software-side issues (less likely, but possible)
Wrong/incomplete DTB for this board revision (rk3399-rock-pi-4c.dtb missing proper PHY description).
Pinctrl state not actually applied to GMAC pins at runtime.
PHY reset GPIO polarity/timing mismatch (snps,reset-* params wrong for this hardware state).
PHY address mismatch if auto-scan behavior is limited by current driver/DT setup.
Kernel/DT regression in this Armbian community build (can test by booting older known-good image/kernel).
Maybe someone from Radxa team can look at this and advice me how to proceed next?
so far Im going to order new RTL8211E chip from aliexpress and replace it again in the other Service Center.
This looks like your hardware has other issues, that prevents RTL8211E from being detected. At this point we cannot tell what caused it, and ordering new chips may not help.
hmm..
ok.. but it is strange..
both LEDs on ETH port works. also there is link to LAN switch when I connect LAN cable to board - and LEDs shows some activity..
so it look like that HW should work.. but..
I reinstalled RTL chip - soldered new one - and I receive the same dmesg messages.
But I noticed one thing.
is it possible to know the nominal of resistor near RTL chip which I pointed on the picture?
this is very small black “brick“ from the right side of the block with number “19“
this resistor is absent on my board - so I can assume this is can be the issue..
it can be 0 Ohm if it is a jumper or probably 10k if it is some kind of pull-up..
You can check the SMD file for part number, and then check schematic for reference value. However, what actually used on the product is defined in the BOM, which is not open.