Intermittent (wired) Ethernet issue on Linksys EA8500 with 18.06.1

Installing and Using OpenWrt
1

Hi, I am running 18.06.1 on my Linksys EA8500. Sometimes on reboot I see kernel messages like

[    4.577085] ipq806x-gmac-dwmac 37200000.ethernet eth0: device MAC address 7e:ab:77:6c:f4:4b
[    4.577300] Generic PHY fixed-0:00: attached PHY driver [Generic PHY] (mii_bus:phy_addr=fixed-0:00, irq=POLL)
[    4.694332] ipq806x-gmac-dwmac 37200000.ethernet: Failed to reset the dma
[    4.694392] ipq806x-gmac-dwmac 37200000.ethernet eth0: stmmac_hw_setup: DMA engine initialization failed
[    4.700128] ipq806x-gmac-dwmac 37200000.ethernet eth0: stmmac_open: Hw setup failed

and the ethernet ports don't come up. A power cycle or reboot clears it up. On a good boot I get

[    6.257697] ipq806x-gmac-dwmac 37200000.ethernet eth0: device MAC address 52:7f:37:47:29:3c
[    6.257940] Generic PHY fixed-0:00: attached PHY driver [Generic PHY] (mii_bus:phy_addr=fixed-0:00, irq=POLL)
[    6.266417] dwmac1000: Master AXI performs any burst length
[    6.275077] ipq806x-gmac-dwmac 37200000.ethernet eth0: IEEE 1588-2008 Advanced Timestamp supported
[    6.280505] ipq806x-gmac-dwmac 37200000.ethernet eth0: registered PTP clock
[    6.290366] IPv6: ADDRCONF(NETDEV_UP): eth0: link is not ready
[    7.361821] ipq806x-gmac-dwmac 37200000.ethernet eth0: Link is Up - 1Gbps/Full - flow control off

I updated this system from LEDE 17.01 and I don't recall having any intermittent issues with the wired ethernet driver. Any ideas on troubleshooting this?

Thanks!
Roland

2

Are you still having this problem? Any new information?

I have been working on getting 18.06.1 on my Trendnet tew827dru, which is similar to the ea8500 and have been having problems with it's networking. In my case it looks like my AR8337 switch isn't working properly.

The next time you have a failed boot, try doing a "swconfig list" and "swconfig dev switch0 show" to see if your built-in switch chip is working right.

It could be the driver for the ethernet ports on the ipq8064 SoC, or it might be the switch.

The ea8500 apparently also had some trouble with the built-in switch. See commit c52cd4d8c3ab1fd0db8806a5e37158e7022c1c40 "ipq806x: fix EA8500 switch control". This might be related.

Good luck.

3

This may be related: EA8500 reboot fails, only with power switch on/off successful

4

I haven't rebooted for a while, but no reason to believe it would have changed.

Based on the kernel messages, I'm pretty sure the issue is the SoC Ethernet MAC and not the switch. I think the dwmac reset is too short to cover all cases; for example http://lists.infradead.org/pipermail/barebox/2016-November/028475.html shows a similar problem. When I get a chance to build a new kernel I want to try a similar patch - extend the timeout and print out how long the reset actually takes.

5

Same problem noticed with 18.06.1 on R7800.
An external HP switch (POE+) is connected on lan port.
Just before reboot, messages like below are shown:

kern.err kernel: [ 584.144343] ipq806x-gmac-dwmac 37200000.ethernet eth0: len 1994 larger than size