[Nut-upsuser] Unable to use nut-2.7.4 with Eaton 5E1500I USB

[Charles Lepple]

On Jun 20, 2017, at 6:00 AM, Manuel Wolfshant wrote:
> 
>    On 06/20/2017 04:08 AM, Charles Lepple wrote:
>> [...]
>> It does not get a length for method 1, so it moves on to method 2, which gets a length, but fails to get the rest of the descriptor. This is what is happening to lsusb, too (tries to fetch 549 bytes, gets only 9).
> and is there anything I can do to fix, short of hoping to receive an updated unbroken firmware for the UPS from a support team which ignores me ?

I can't speak for Eaton, but given what I have seen so far, the symptoms do not necessarily point to the UPS. If I search my copy of the NUT mailing lists for "lsusb" and "incomplete report descriptor", I see two cases of a Liebert GXT-3, one CyberPower PP1100, one older Tripp-Lite unit, and an APC Smart-UPS 1000.

That sample set is not large enough to isolate the problem to a specific kernel, distribution or motherboard. Hopefully Eaton has a larger set of data points.

>>> However I am using the very same package on my personal workstation (which is also running CentOS 6 but has a whole lot of packages installed either from 3rd party repos or built by me  ) since Fri 09 Jan 2015 (I needed it for heimdall ) and I had zero issues so far.
>> Can you be more specific about this? What hardware are you testing libusbx 1.0.19 with?
> Gigabyte F2A88X-D3H with AMD A8-6600K works fine with heimdall which I used to connect to a Samsung S2+ and now to a Samsung A3 (2016) . Some similar board ( I do not recall exactly what model was it, could have been either identical or DH3 which is basically the same motherboard with some minor modifications) was used by my colleague from the IT support dept 2-3 years ago with my phone as well. The packages that I have installed are :

Oh, I thought you meant that libusbx was working fine with another UPS.

I don't know whether the 5E1500I is a 1.5 Mbit/sec or 12 Mbit/sec device, but I suspect that the other USB devices you are testing are faster than that. (Bear in mind that 480 Mbit/sec generally uses a different host controller interface, and 5.0 Gbit/sec uses a different PHY layer as well.)