Nut upsuser - Aug 2014 - USB device descriptor read error with Eaton 5E UPS

Hi,

I recently purchased a new Easton 5E UPS. I plugged the USB cable into
my computer so I could use the NUT UPS tools. However every so often
NUT fails to connect to the device after a boot.

Looking at dmesg I see errors like this:
[    3.277075] usb 1-9: new low-speed USB device number 6 using xhci_hcd
[    3.302091] usb 1-9: device descriptor read/8, error -32
[    3.430241] usb 1-9: device descriptor read/8, error -32
[    3.653526] usb 1-9: new low-speed USB device number 7 using xhci_hcd
[    3.678473] usb 1-9: device descriptor read/8, error -32
[    3.824288] usb 1-9: device descriptor read/8, error -32
[    4.045977] usb 1-9: new low-speed USB device number 8 using xhci_hcd
[    4.071018] usb 1-9: device descriptor read/8, error -32
[    4.199046] usb 1-9: device descriptor read/8, error -32
[    4.422439] usb 1-9: new low-speed USB device number 9 using xhci_hcd
[    4.451464] usb 1-9: device descriptor read/8, error -32
[    4.579618] usb 1-9: device descriptor read/8, error -32

I have tried to plug it into every USB port (including the USB 3.0
ones) and I still get these errors. I discovered that power cycling
the UPS will fix it for a while, and also unplugging and replugging
the USB cable will also fix it sometimes.

I have observed the problem on both an Asus P9D-I motherboard with the
Intel C222 chipset and on the Asus H87M-E motherboard with the Intel
H87 chipset.

I emailed Eaton support but they don't know what the issue is. They
want me to confirm that it is not Linux or PC issue before I return it
for testing under warranty.

They also said that all warranty testing is performed under Windows only :-(.

How can I be sure that this is a problem with the device and not my
computer before I return it?

I am on Ubuntu Server 14.04 x86-64 bit. The kernel version is:
Linux version 3.13.0-32-generic (buildd at kissel) (gcc version 4.8.2
(Ubuntu 4.8.2-19ubuntu1) ) #57-Ubuntu SMP Tue Jul 15 03:51:08 UTC 2014
(Ubuntu 3.13.0-2.57-generic 3.13.11.4)

On Aug 7, 2014, at 7:57 AM, Leith Bade <leith at leithalweapon.geek.nz>
wrote:
> I have tried to plug it into every USB port (including the USB 3.0
> ones) and I still get these errors. I discovered that power cycling
> the UPS will fix it for a while, and also unplugging and replugging
> the USB cable will also fix it sometimes.
Error -32 is EPIPE, which means that the PC didn't see a reply from the
device during the descriptor read. I don't know that we have ever
successfully run one of these issues to ground.

When you plug into the non-3.0 ports, what HCD driver is it using?

I wouldn't recommend using a hub in production, but if you have an older USB
hub lying around, does that help? Shorter cables are also useful, though
low-speed USB 1.1 is pretty resilient to noise.

Also, if you can fix it without rebooting the PC, the issue is probably external
to the PC.
> I have observed the problem on both an Asus P9D-I motherboard with the
> Intel C222 chipset and on the Asus H87M-E motherboard with the Intel
> H87 chipset.
Did you have a chance to check the NUT archives? I vaguely remember some
motherboard-specific threads.

-- 
Charles Lepple
clepple at gmail

On 10 August 2014 09:05, Charles Lepple <clepple at gmail.com>
wrote:
> On Aug 7, 2014, at 7:57 AM, Leith Bade <leith at
leithalweapon.geek.nz> wrote:
>
> Error -32 is EPIPE, which means that the PC didn't see a reply from the
device during the descriptor read. I don't know that we have ever
successfully run one of these issues to ground.
>
Considering Eaton do officially support Linux in their documentation
for this product (as well as make their own Linux UPS client) I will
see if they are able to fix it. Otherwise I will just have to return
it and get another brand.

Also do you know how I can log more detailed debug info from the
Kernel for USB? Perhaps a dump of all USB data to a file somewhere may
make it easier to see exactly what is going on.
> When you plug into the non-3.0 ports, what HCD driver is it using?
>
I generally keep it plugged into the 2.0 ports. Here is the output
from dmesg at boot (when it is working)
[    0.374422] usbcore: registered new interface driver usbfs
[    0.374443] usbcore: registered new interface driver hub
[    0.374503] usbcore: registered new device driver usb:
...
[    1.508833] ehci_hcd: USB 2.0 'Enhanced' Host Controller (EHCI)
Driver
[    1.516042] ehci-pci: EHCI PCI platform driver
[    1.523181] ehci-platform: EHCI generic platform driver
[    1.530293] ohci_hcd: USB 1.1 'Open' Host Controller (OHCI) Driver
[    1.537416] ohci-pci: OHCI PCI platform driver
[    1.544488] ohci-platform: OHCI generic platform driver
[    1.551537] uhci_hcd: USB Universal Host Controller Interface driver
[    1.558906] xhci_hcd 0000:00:14.0: xHCI Host Controller
[    1.565959] xhci_hcd 0000:00:14.0: new USB bus registered, assigned
bus number 1
[    1.573220] xhci_hcd 0000:00:14.0: cache line size of 64 is not supported
[    1.573270] xhci_hcd 0000:00:14.0: irq 44 for MSI/MSI-X
[    1.573418] usb usb1: New USB device found, idVendor=1d6b, idProduct=0002
[    1.580669] usb usb1: New USB device strings: Mfr=3, Product=2,
SerialNumber=1
[    1.587899] usb usb1: Product: xHCI Host Controller
[    1.595059] usb usb1: Manufacturer: Linux 3.13.0-32-generic xhci_hcd
[    1.602239] usb usb1: SerialNumber: 0000:00:14.0
[    1.609600] hub 1-0:1.0: USB hub found
[    1.616605] hub 1-0:1.0: 10 ports detected
[    1.629387] xhci_hcd 0000:00:14.0: xHCI Host Controller
[    1.636300] xhci_hcd 0000:00:14.0: new USB bus registered, assigned
bus number 2
[    1.643425] usb usb2: New USB device found, idVendor=1d6b, idProduct=0003
[    1.650424] usb usb2: New USB device strings: Mfr=3, Product=2,
SerialNumber=1
[    1.657393] usb usb2: Product: xHCI Host Controller
[    1.664350] usb usb2: Manufacturer: Linux 3.13.0-32-generic xhci_hcd
[    1.671278] usb usb2: SerialNumber: 0000:00:14.0
[    1.678293] hub 2-0:1.0: USB hub found
[    1.684992] hub 2-0:1.0: 2 ports detected
...
[    2.797764] usb 1-10: new low-speed USB device number 4 using xhci_hcd
...
[    3.423585] usb 1-10: New USB device found, idVendor=0463, idProduct=ffff
[    3.423599] usb 1-10: New USB device strings: Mfr=1, Product=2,
SerialNumber=0
[    3.423604] usb 1-10: Product: 5E
[    3.423605] usb 1-10: Manufacturer: EATON
[    3.424529] usb 1-10: ep 0x81 - rounding interval to 128
microframes, ep desc says 160 microframes
[    3.472025] hidraw: raw HID events driver (C) Jiri Kosina
...

I guess that means it is using xhci_hcd. All USB ports show up as
usb1-x (on both this motherboard and the other motherboard).

Only other things of interest is that I have only ever had these
plugged into the other USB ports - mouse, keyboard, Sierra 3G modem,
Huawei 3G modem. On the current motherboard various 'virtual' devices
such as mouse, keyboard, floppy, CD drive, and HDD also appear from
the BIOS's remote control features.

However when I first got these problems was with the previous
motherboard and it would occur if nothing else was plugged in too.

> I wouldn't recommend using a hub in production, but if you have an
older USB hub lying around, does that help? Shorter cables are also useful,
though low-speed USB 1.1 is pretty resilient to noise.
>
> Also, if you can fix it without rebooting the PC, the issue is probably
external to the PC.
>
I will try it with a USB hub. Unfortunately I don't have a shorter
cable currently. The one I am using is an aftermarket generic 1m USB
2.0 cable purchased from a local office supply store (Office Works)
since the UPS did not come with a cable :-( This cable appears to be
good quality as it is rather thick.
>
> Did you have a chance to check the NUT archives? I vaguely remember some
motherboard-specific threads.
>
A bit of searching did not turn up much. Only thing I found was a
similar error with a MGE unit a few years ago. Did not say anything
about the motherboard.

Thanks,
Leith Bade
leith at leithalweapon.geek.nz