Pi Zero ethernet

[PlanB]

A Pi zero arrived here the other day so I plugged in an sd card from a Pi3 (originally a PiB+ sd with with updates till it worked in the pi3) and a WiFi dongle and SSH from my PC worked straight up.

Emboldened by this early success I thought I'd try ethernet via http://www.ebay.com.au/itm/381475334140 ... EBIDX%3AIT The only thing that seems to work on this is the keyboard. It can't find the router on the ethernet port & if I plug the WiFi dongle into one of the ubs ports the Pi zero confuses it's fixed ethernet ip (10.0.0.198) with it's fixed WIFI ip (10.0.0.99) & tries to find the router on WiFi via 10.0.0.198

Do I need a driver or something for this hybrid usb/ethernet hub? I think this is the first time in my life where a wired connection has been harder to achieve than a wireless one, strange days indeed.

[DougieLawson]

You need to send that pile of junk back to where it came from and buy a decent ethernet adapter that isn't based on the exceedingly crappy DM9601 chipset.

[PlanB]

Bugger. I guess that's why http://www.ebay.com.au/itm/191019692396 ... EBIDX%3AIT don't work neither? Any recommendations?

[rpdom]

Can you show the output of the "lsusb" (LiSt USB) command when the unit is plugged in?
Then we can check that it is being detected properly and which chipsets it uses.

Even if it does use the slightly-dodgy DM9601 ethernet chip, it should still work. It won't provide a fast connection, but it should provide one. The problems with the driver for that crashing the system have been fixed.

[PlanB]

It says.....
0fe6:9700 Kontron (industrial Computer Source / ICS Advent) DM9601 Fast Ethernet Adapter

[aBUGSworstnightmare]

I've already used a Digitus DN-3023 Gigabit Ethernet Adapter (RJ-45 to USB 3.0) with my Zero without problems. This dongle seems to use ASIX AX88179 chipset.

[fruitoftheloom]

I've already used a Digitus DN-3023 Gigabit Ethernet Adapter (RJ-45 to USB 3.0) with my Zero without problems. This dongle seems to use ASIX AX88179 chipset.

The Adapters with Realtek Chipset are more expensive:

https://www.amazon.co.uk/Smays-Micro-US ... B009XU6N54

https://www.amazon.co.uk/TeckNet-Gigabi ... B00E44F3BO


.

[karrika]

Can you show the output of the "lsusb" (LiSt USB) command when the unit is plugged in?
Then we can check that it is being detected properly and which chipsets it uses.

Even if it does use the slightly-dodgy DM9601 ethernet chip, it should still work. It won't provide a fast connection, but it should provide one. The problems with the driver for that crashing the system have been fixed.

When was it fixed? I accidentally plugged in my to a PiZero with an updated Jessie image for testing the audio patches a week ago and it did not work. The ASIX based Ethernet is working now.

[DougieLawson]

Don't know if this is still available viewtopic.php?f=63&t=131379 but I'd go with the magic Google Chromecast power supply and ethernet if I wanted to run a Zero with a wired connection.

[FTrevorGowen]

...
Even if it does use the slightly-dodgy DM9601 ethernet chip, it should still work. It won't provide a fast connection, but it should provide one. The problems with the driver for that crashing the system have been fixed.

Interesting ... Do you have any idea when? I last checked my sample(s) some months ago with an older Pi Zero (and, probably, older kernel etc.) Another thing for my "to do list" - I'm currently adding to/tidying up my, relocated web pages (locally) and hope to update the main site soon (including my USB-to-LAN adapters page)
Trev.

[gregeric]

It wasn't the driver that was fixed, rather the Pi's dwc_otg usb driver was made robust against the quirks of the dm9601:
https://github.com/raspberrypi/linux/is ... -226159623
https://github.com/raspberrypi/linux/co ... b41eaf7ec2
https://github.com/raspberrypi/linux/co ... 50e7fd151d

Just checked - still working for me (you see a message in the kernel log from dwc_otg "trimming transfer length", where previously there would have been a panic).

[FTrevorGowen]

It wasn't the driver that was fixed, rather the Pi's dwc_otg usb driver was made robust against the quirks of the dm9601:
https://github.com/raspberrypi/linux/is ... -226159623
https://github.com/raspberrypi/linux/co ... b41eaf7ec2
https://github.com/raspberrypi/linux/co ... 50e7fd151d
Just checked - still working for me (you see a message in the kernel log from dwc_otg "trimming transfer length", where previously there would have been a panic).

Thanks for that info. - looks like it happened "just after" I re-tested (with an "earlier" version of Jessie etc.) . I'll re-test, especially the variant that's part of/combined with a USB hub (3 ports available)** sometime this week (hopefully) and then I think it's time to update the (external) webpages even though some other work will still be "in progress".
Trev.
** Shown, with the other devices I have here:
https://www.cpmspectrepi.uk/raspberry_p ... pters.html
(apart from a Realtek chipset example, recently acquired and tried)

[PlanB]

So do I have any chance of getting the 9601 chipset working without upgrading to Jessie? Jessie frightens the life out of me because of stuff like viewtopic.php?f=63&t=110606 and this thing is looking like it should work, no kernel crashes, route -n, ifconfig etc all look normal but it can't find the router

[DougieLawson]

Get a fresh new unused SDCard and try it.

[gregeric]

So do I have any chance of getting the 9601 chipset working without upgrading to Jessie?

Bite the bullet and upgrade.

If you absolutely must stick with what you have, you could re-compile your current kernel version after patching in the dwc_otg fix. Or do an rpi-update to get a new kernel.

[PlanB]

I'm getting no indication that there is a driver problem here, no kernel panics, the iptable looks OK, /var/log/messages are sweet. I've got this nagging suspicion something else is amiss.

[PlanB]

For what it's worth the updated wheezy sd from my Pi3 works OK with the 9601 chipset, you don't have to go to jessie. The particular combined usbhub/ethernet unit I had was a dud but a 9601 based ethernet only adapter was fine.

[FTrevorGowen]

...
Thanks for that info. - looks like it happened "just after" I re-tested (with an "earlier" version of Jessie etc.) . I'll re-test, especially the variant that's part of/combined with a USB hub (3 ports available)** sometime this week ...
...

O.K., FWIW I've now managed to re-test two devices with a Pi Zero v1.3:
1) The 3-port hub/LAN combination - PowerSync Tech, ID 0fe6:9700 Kontron (LAN) & ID 1a40:0101 Terminus Tech. Ltd. (Hub). I was able to successfully update/upgrade Raspbian 4.4.11+ #888 23/5/2016 to 4.4.13+ #894 13/6/2016 and, subsequently ...
2) performed a quick headless test (ping etc.) using a niceEshop VK-QF9700 LAN only device (same ID as the LAN part of the first device).
Trev.
Edit(05/08/2016): The external webpages have been updated and some screenshots/notes w.r.t. the above can be found here:
https://www.cpmspectrepi.uk/raspberry_p ... ver_issues

[mc1970]

I also have the same "ID 0fe6:9700 Kontron (Industrial Computer Source / ICS Advent) DM9601 Fast Ethernet Adapter" combination and have upgraded to the latest Jessie. The network hangs still occur when you try some higher network throughput. Just having it plugged in with the occasional light load seems ok.

Kernel messages show :-
Nov 12 17:55:22 raspberrypi kernel: [ 155.977629] dwc_otg: DEVICE:002 : update_urb_state_xfer_comp:742:trimming xfer length
Nov 12 17:55:24 raspberrypi kernel: [ 158.035333] ERROR::handle_hc_chhltd_intr_dma:2214: handle_hc_chhltd_intr_dma: Channel 1, DMA Mo
de -- ChHltd set, but reason for halting is unknown, hcint 0x00000002, intsts 0x06000021
Nov 12 17:55:24 raspberrypi kernel: [ 158.035333]
Nov 12 17:55:24 raspberrypi kernel: [ 158.035410] ERROR::handle_hc_chhltd_intr_dma:2214: handle_hc_chhltd_intr_dma: Channel 7, DMA Mo
de -- ChHltd set, but reason for halting is unknown, hcint 0x00000002, intsts 0x06000001
Nov 12 17:55:24 raspberrypi kernel: [ 158.035410]
Nov 12 17:55:24 raspberrypi kernel: [ 158.035459] ERROR::handle_hc_chhltd_intr_dma:2214: handle_hc_chhltd_intr_dma: Channel 0, DMA Mo
de -- ChHltd set, but reason for halting is unknown, hcint 0x00000002, intsts 0x06000001
Nov 12 17:55:24 raspberrypi kernel: [ 158.035459]
Nov 12 17:55:24 raspberrypi kernel: [ 158.121047] usb usb1-port1: disabled by hub (EMI?), re-enabling...
Nov 12 17:55:24 raspberrypi kernel: [ 158.121088] usb 1-1: USB disconnect, device number 2
Nov 12 17:55:24 raspberrypi kernel: [ 158.121472] dm9601 1-1:1.0 eth0: unregister 'dm9601' usb-20980000.usb-1, Davicom DM96xx USB 10/
100 Ethernet
Nov 12 17:55:24 raspberrypi kernel: [ 158.291052] Indeed it is in host mode hprt0 = 00021d01
Nov 12 17:55:24 raspberrypi kernel: [ 158.531064] usb 1-1: new full-speed USB device number 3 using dwc_otg
Nov 12 17:55:24 raspberrypi kernel: [ 158.534215] Indeed it is in host mode hprt0 = 00021d01
Nov 12 17:55:24 raspberrypi kernel: [ 158.825434] usb 1-1: New USB device found, idVendor=0fe6, idProduct=9700
Nov 12 17:55:24 raspberrypi kernel: [ 158.825465] usb 1-1: New USB device strings: Mfr=0, Product=2, SerialNumber=0
Nov 12 17:55:24 raspberrypi kernel: [ 158.825479] usb 1-1: Product: USB 2.0 10/100M Ethernet Adaptor
Nov 12 17:55:24 raspberrypi kernel: [ 158.916166] dm9601 1-1:1.0 eth0: register 'dm9601' at usb-20980000.usb-1, Davicom DM96xx USB 10
/100 Ethernet, 00:e0:4c:53:44:58
Nov 12 17:55:25 raspberrypi kernel: [ 159.811396] dm9601 1-1:1.0 eth0: link up, 100Mbps, full-duplex, lpa 0xFFFF
Nov 12 17:55:26 raspberrypi kernel: [ 160.371385] dm9601 1-1:1.0 eth0: kevent 4 may have been dropped
Nov 12 17:55:26 raspberrypi kernel: [ 160.379399] dm9601 1-1:1.0 eth0: kevent 4 may have been dropped
Nov 12 17:55:26 raspberrypi kernel: [ 160.387392] dm9601 1-1:1.0 eth0: kevent 4 may have been dropped
Nov 12 17:55:26 raspberrypi kernel: [ 160.397416] dm9601 1-1:1.0 eth0: kevent 4 may have been dropped
Nov 12 17:55:26 raspberrypi kernel: [ 160.439408] dm9601 1-1:1.0 eth0: link up, 100Mbps, full-duplex, lpa 0xFFFF
Nov 12 17:55:26 raspberrypi kernel: [ 160.571686] dm9601 1-1:1.0 eth0: link up, 100Mbps, full-duplex, lpa 0xFFFF

[gregeric]

The first line of that log is the patch to the Pi's dwc_otg (USB) driver for the DM9601 doing its job, whereas before you would get kernel panics.

The rest of the log, I dunno.

Do you see the same in other Pi models with the DM9601?

[rpdom]

Possibly power issues causing the device to reset?

[DougieLawson]

viewtopic.php?p=1028165#p1028165

[gregeric]

Possibly power issues causing the device to reset?

That did go through my mind, though it baffles me why the Pi & connected USB peripherals seem so susceptible, so was loathe to suggest it.

I've been using the DM9601 with Pi2 (just for testing) on and off for months since the fix - no issues here, or seen reported elsewhere.

@Dougie - the device works OK with its intrinsic limitations (100Mbps bridged over USB 1 full speed) elsewhere on OpenWRT, x86 Debian etc, even Windows. And since the fix to dwc_otg, on the Pi likewise.

[karrika]

I have exactly the same problem with my DM9601 PiZero combo. A simple speedtest.py disconnects it. With the ASIX AX88772 and PiZero I have no issues. The network is up for days without problems.