opi pc - usb disconnects?!

Nix · March 25, 2016

Hi,

Using orange pi pc with armbian 5.05 jessie server image. Everything worked fine for few days, until recently when usb device started disconnecting on its own:

Mar 24 19:58:02 localhost mtp-probe: checking bus 7, device 90: "/sys/devices/platform/sunxi-ohci.3/usb7/7-1"
Mar 24 19:58:02 localhost mtp-probe: bus: 7, device: 90 was not an MTP device
Mar 24 19:58:03 localhost kernel: [  592.325419] usb 7-1: USB disconnect, device number 90
Mar 24 19:58:03 localhost kernel: [  592.332823] ch341-uart ttyUSB0: ch341-uart converter now disconnected from ttyUSB0
Mar 24 19:58:03 localhost kernel: [  592.339829] ch341 7-1:1.0: device disconnected
Mar 24 19:58:03 localhost kernel: [  592.680093] usb 7-1: new full-speed USB device number 91 using sunxi-ohci
Mar 24 19:58:03 localhost kernel: [  592.910999] ch341 7-1:1.0: ch341-uart converter detected
Mar 24 19:58:03 localhost kernel: [  592.949040] usb 7-1: ch341-uart converter now attached to ttyUSB0
Mar 24 19:58:03 localhost mtp-probe: checking bus 7, device 91: "/sys/devices/platform/sunxi-ohci.3/usb7/7-1"
Mar 24 19:58:03 localhost mtp-probe: bus: 7, device: 91 was not an MTP device
Mar 24 19:58:04 localhost kernel: [  594.075309] usb 7-1: USB disconnect, device number 91
Mar 24 19:58:04 localhost kernel: [  594.090754] ch341-uart ttyUSB0: ch341-uart converter now disconnected from ttyUSB0
Mar 24 19:58:04 localhost kernel: [  594.105513] ch341 7-1:1.0: device disconnected
Mar 24 19:58:05 localhost kernel: [  594.460116] usb 7-1: new full-speed USB device number 92 using sunxi-ohci
Mar 24 19:58:05 localhost kernel: [  594.690799] ch341 7-1:1.0: ch341-uart converter detected
Mar 24 19:58:05 localhost kernel: [  594.728941] usb 7-1: ch341-uart converter now attached to ttyUSB0
Mar 24 19:58:05 localhost mtp-probe: checking bus 7, device 92: "/sys/devices/platform/sunxi-ohci.3/usb7/7-1"
Mar 24 19:58:05 localhost mtp-probe: bus: 7, device: 92 was not an MTP device
Mar 24 19:58:06 localhost kernel: [  595.912993] usb 7-1: USB disconnect, device number 92
Mar 24 19:58:06 localhost kernel: [  595.928484] ch341-uart ttyUSB0: ch341-uart converter now disconnected from ttyUSB0
Mar 24 19:58:06 localhost kernel: [  595.943244] ch341 7-1:1.0: device disconnected
Mar 24 19:58:11 localhost kernel: [  600.710093] usb 7-1: new full-speed USB device number 94 using sunxi-ohci
Mar 24 19:58:14 localhost kernel: [  603.360110] usb 7-1: new full-speed USB device number 95 using sunxi-ohci
Mar 24 19:58:16 localhost kernel: [  605.840128] usb 7-1: new full-speed USB device number 96 using sunxi-ohci
Mar 24 19:58:16 localhost kernel: [  605.901337] ch341 7-1:1.0: ch341-uart converter detected
Mar 24 19:58:16 localhost kernel: [  605.938557] usb 7-1: ch341-uart converter now attached to ttyUSB0
Mar 24 19:58:16 localhost mtp-probe: checking bus 7, device 96: "/sys/devices/platform/sunxi-ohci.3/usb7/7-1"
Mar 24 19:58:16 localhost mtp-probe: bus: 7, device: 96 was not an MTP device
Mar 24 19:58:16 localhost kernel: [  606.104080] usb 7-1: USB disconnect, device number 96
Mar 24 19:58:16 localhost kernel: [  606.119031] ch341-uart ttyUSB0: ch341-uart converter now disconnected from ttyUSB0
Mar 24 19:58:16 localhost kernel: [  606.133390] ch341 7-1:1.0: device disconnected
Mar 24 19:58:17 localhost kernel: [  606.480105] usb 7-1: new full-speed USB device number 97 using sunxi-ohci
Mar 24 19:58:24 localhost kernel: [  613.850122] usb 7-1: new full-speed USB device number 98 using sunxi-ohci

Device in question is arduino 2560 mega board. I dont believe it has anything to do with power issues. And arduino board is functional. USB "recovers" only on reboot. Reconnecting device does not help, since after it "disconnects" lsusb is empty.

I am using 2A power supply, and this arduino board is only device connected to it. Am planning to use opi one for it, but for now have extra opi pc at hand

What can be done to fix this issue? Do you need any more information?

tkaiser · March 25, 2016

Try out all different USB ports, try it with another "arduino 2560 mega board", in case these devices don't have fixed cables, try out different cables. How should we fix hardware flaws by writing/reading in a forum?

Nix · March 25, 2016

Try out all different USB ports, try it with another "arduino 2560 mega board", in case these devices don't have fixed cables, try out different cables. How should we fix hardware flaws by writing/reading in a forum?

This is second opi pc I own. First one is running arch linux (some image made by loboris), has two usb devices connected that never caused any problem working 24/7. So hope you understand why my primary suspicion for usb with second opi pc is armbian 5.05. Disconnecting issues started yesterday. It cannot be hardware related since it would happen sooner than one week in.

When it occured yesterday, I tried all three usb ports and lsusb was empty on all three. Even one reboot did not help. Power cycle did help.

Also cables are ok, since I have been using that same cable with laptop connected to arduino and as long as it was connected there were no issues.

tkaiser · March 25, 2016

So when it's Armbian's fault why don't you try to exchange the distro you use? Using your 'line of argument' obviously Armbian is the root cause and has to be replaced (we had USB issues with H3 prior to 5.04 but these are fixed now)

martinayotte · March 25, 2016

I have some USB-RS485 dongle plugged on OPi-PC (with Armbian 4.4.5) and running some code which scan RS485 buses for several days/weeks, and USB never stopped, it still working.

tkaiser · March 25, 2016

I have some USB-RS485 dongle plugged on OPi-PC (with Armbian 4.4.5)

Here as well: one OPi PC serving as NAS with mainline kernel und utilising USB even for networking since months with full USB utilisation all the time. But that's not comparable since he uses apparently kernel 3.4.110. And while it might be possible that there is different USB driver stuff in ours and loboris' kernel I won't look into it based on a single report ignoring all basic rules of fault localisation.

It's never the question whether hardware will die sometimes in the future, it's always just the question when it starts to die. If this isn't taken into account at all... why bother?

Nix · March 25, 2016

Sorry for beathing your air tkaiser. I sincerely apologise for doubting armbian.

tkaiser · March 25, 2016

I sincerely apologise for doubting armbian.

It would be way better for all of us if you would start with proper fault localisation. When the one and only USB device you seem to try out starts to disappear what about connecting OTHER USB devices then? What does lsusb then show?

What about switching OPi PCs (pretty easy, isn't it? Simply exchanging SD cards and peripherals and you're done!) and so one...

You do not really expect that we start wasting our time searching for errors somewhere on the basis of such a 'report' ("there can't be something wrong with the cable now since it worked flawlessly in the past.")

To be clear: It might be possible that there is really a problem in Armbian. If this is the case we want to do everything to fix this as soon as possible (as we did it all the time in the past). But this requires a bit of action from your side to help us understand what's really wrong. Since all I could do now based on your 'report' is to say: Well yes, since you claim your hardware is OK, it must be a software problem, so please return your Armbian copy and you'll get a full refund...