USB not working

[renatovitolo]

Hi shodruk,

please find below the output of the 2 commands. Today I observed a similar initial string of 9-10 reboots with USB not working, followed by 7-8 reboots with a working USB. Again, USB does not work if and only if dmesg displays

xusbps-dr e0002000.usb: Unable to init USB phy, missing?

and lsusb gives error -99, whereas USB does work if and only if dmesg gives

xusbps-ehci xusbps-ehci.0: Xilinx PS USB EHCI Host Controller
xusbps-ehci xusbps-ehci.0: new USB bus registered, assigned bus number 1
...

In the latter case, lsusb gives

Bus 001 Device 002: ID 046d:c52b Logitech, Inc. Unifying Receiver
Bus 001 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub

and both keyboard and mouse of my Logitech mk520 wireless combo work flawlessly with the board. Actually, whenever the USB was working, I was able to connect and use two other (wired, USB) keyboards and two other (wired, USB) mice. For instance, I have just unplugged the combo and plugged an old mouse and lsusb gives

Bus 001 Device 003: ID 0461:4d03 Primax Electronics, Ltd Kensington Mouse-in-a-box

Plugging a USB keyboard gives then

Bus 001 Device 004: ID 045e:00dd Microsoft Corp. Comfort Curve Keyboard 2000 V1.0

So in my case it really looks like a boot-time problem. The only pattern which I have observed is that USB somehow decides to keep functioning for all reboots after a few initial consecutive reboots where it does not, so when the board is "warmed up". Might this be simply a matter of temperature of the board? Having devices plugged or not at boot time does not seem to influence the course of events. I'll keep experimenting....

linaro-nano:~> cat /proc/cmdline
root=/dev/mmcblk0p2 rw rootfstype=ext4 rootwait devtmpfs.mount=1 video=HDMI-A-1:1280x720@60

linaro-nano:~> cat /etc/X11/xorg.conf
Section "Device"
Identifier "Card0"
Driver "modesetting"
Option "ShadowFB" "True"
Option "SWCursor" "True"
Option "HWCursor" "False"
EndSection
Section "Screen"
Identifier "Screen0"
Device "Card0"
SubSection "Display"
#---- Uncomment your preferred mode ----
#Modes "1920x1200"
#Modes "1920x1080"
Modes "1280x720"
#Modes "640x480"
EndSubSection
EndSection

[shodruk]

In my case, I get that error when I connect nothing to USB port.
Strangely, when I connect just a "long USB cable", I never get that error.

without USB cable connected

ehci_hcd: USB 2.0 'Enhanced' Host Controller (EHCI) Driver
ehci-pci: EHCI PCI platform driver
xusbps-dr e0002000.usb: Unable to init USB phy, missing?


with "long USB cable" connected
(just a 1m USB cable + micro USB adapter. it is not connected to any USB device)

ehci_hcd: USB 2.0 'Enhanced' Host Controller (EHCI) Driver
Found SMSC USB3320 ULPI transceiver.
xusbps-ehci xusbps-ehci.0: Xilinx PS USB EHCI Host Controller
xusbps-ehci xusbps-ehci.0: new USB bus registered, assigned bus number 1
xusbps-ehci xusbps-ehci.0: USB 2.0 started, EHCI 1.00
hub 1-0:1.0: USB hub found
USB Video Class driver (1.1.1)
usbhid: USB HID core driver


Adapteva also commented about this issue.
viewtopic.php?f=50&t=1055

What is in just a long USB cable? Capacitance?

Note: I have no issue with USB, with USB hub, keyboard, mouse, etc. conneced.

[renatovitolo]

That's right Shodruk, I get exactly the same behaviour as you!

With the trick of the "long USB cable (nothing attached) before boot", the USB system always comes up properly. Then after boot I unplug the long cable: any USB device (mouse, keyboard, etc) which I plug in afterwards functions properly.

Without long USB cable before boot, the USB system only comes up sporadically and unpredictably, forcing me to reboot several times.

So at the moment this one seems fixed AFAIAC and I can move on.

Thanks again, I owe you a beer!

[cmcconnell]

The long cable solution appears to be working for me too. After enabling tempfs I had no change in the keyboard/mouse behaviour. I have been using the short micro-USB to USB conversion cable supplied with the accessory bundle, to which I had my Logitech radio dongle attached. The success ratio for a working keyboard on bootup was about 1 in 10.

With that cable still plugged in to the parallella, I've now added a long USB extension cable between it and the dongle, and I've just been through a sequence of 7 or 8 startups (sometimes rebooting via the command line, sometimes shutting down and powering off and on) and got a working keyboard each time.

Can't say for certain the problem won't recur, but things are looking promising.

[reid]

Just wanted to share my positive experience with the USB issue, this thread was very helpful to me.

I have an "iogear" 4-port usb/dvi KVM switch, which is powered, and my parallella power-supply is the 3A one that comes with the accessory bundle for a single board. In the as-shipped condition, with the downloaded SD image, USB never worked.

Enabling devtmpfs as described in the speed-up article mostly solved the problem for me.

I only did the devtmpfs steps, i.e. the bit where you edit the devicetree.dt(s|b) file and remove the devtmpfs-disabling boot-time directive. My current situation is that, on cold-boot, initially only the keyboard comes up, but once I log in to the device, I have both keyboard and mouse on the desktop, and can switch away and back on the KVM without losing any functionality.

I have found that if I "warm-start", i.e. issue a console reboot command, then on the second boot, the USB will not come up. I have to unplug the 5v power and plug it back in, and then I get the same keyboard-on-start, mouse-and-keyboard-on-login behavior.

This is a workable solution for me.

[bszente]

The long USB cable solution seems to be working for me as well. Enabling devtmpfs did not change anything so it does not matter. Only the long USB cable solves the problem. I'm wondering if this issue can be fixed by soldering some additional passive electronics (capacitors for example)? Something that behaves much like the long cable

[shodruk]

[rrr]

[bszente]

[rrr]

Thanks for the reply! But yes, I have tried plugging in the long USB cables before connecting the power. I've even tried long USB cables of varying length in case that would make a difference. Then after it start ups, I would switch to a short USB connected to a device like a keyboard and it still would not work.