[Nut-upsuser] Junda-tech

[Charles Lepple]

On Oct 17, 2017, at 8:48 AM, LLSJ Krüger <llsjk at lakruger.za.org> wrote:
> Results of 'lsusb -vvv -d 3344:'
[...]
>           wDescriptorLength     136
>          Report Descriptors:
>            ** UNAVAILABLE **
> 

Can you please re-run lsusb, possibly as root, to grab the contents of the "Report Descriptors" section? 

(Running "usbhid-ups" should have detached the kernel driver already, which should allow lsusb to retrieve that descriptor -- but if you have disconnected+reconnected the USB cable since testing, you will need to re-run "usbhid-ups" before "lsusb".)

> ran usbhid-ups -DDDD -u root -x explore -x vendorid=3344 -x port=auto -a
> myups >& /tmp/junda-tech.txt
> 
> and then
> 
> gen-usbhid-subdriver.sh < /tmp/junda-tech.txt' :
> 
[...]
> /* --------------------------------------------------------------- */
> /* HID2NUT lookup table                                            */
> /* --------------------------------------------------------------- */
> 
> static hid_info_t jundatech_hid2nut[] = {
> 
>   { "unmapped.ups.powersummary.capacitymode", 0, 0,
> "UPS.PowerSummary.CapacityMode", NULL, "%.0f", 0, NULL },
>   { "unmapped.ups.powersummary.designcapacity", 0, 0,
> "UPS.PowerSummary.DesignCapacity", NULL, "%.0f", 0, NULL },
>   { "unmapped.ups.powersummary.fullchargecapacity", 0, 0,
> "UPS.PowerSummary.FullChargeCapacity", NULL, "%.0f", 0, NULL },
>   { "unmapped.ups.powersummary.presentstatus.acpresent", 0, 0,
> "UPS.PowerSummary.PresentStatus.ACPresent", NULL, "%.0f", 0, NULL },
>   { "unmapped.ups.powersummary.presentstatus.charging", 0, 0,
> "UPS.PowerSummary.PresentStatus.Charging", NULL, "%.0f", 0, NULL },
>   { "unmapped.ups.powersummary.presentstatus.discharging", 0, 0,
> "UPS.PowerSummary.PresentStatus.Discharging", NULL, "%.0f", 0, NULL },
>   { "unmapped.ups.powersummary.rechargeable", 0, 0,
> "UPS.PowerSummary.Rechargeable", NULL, "%.0f", 0, NULL },
>   { "unmapped.ups.powersummary.remainingcapacity", 0, 0,
> "UPS.PowerSummary.RemainingCapacity", NULL, "%.0f", 0, NULL },
>   { "unmapped.ups.powersummary.imanufacturer", 0, 0,
> "UPS.PowerSummary.iManufacturer", NULL, "%.0f", 0, NULL },
>   { "unmapped.ups.powersummary.iproduct", 0, 0,
> "UPS.PowerSummary.iProduct", NULL, "%.0f", 0, NULL },
>   { "unmapped.ups.powersummary.iserialnumber", 0, 0,
> "UPS.PowerSummary.iSerialNumber", NULL, "%.0f", 0, NULL },

This part looks promising - the UPS appears to implement a portion of the HID Power Device Class specification, rather than just being a HID-enabled USB-to-serial adapter.

In the generated code, the capitalized names are HID PDC names, which you can search for in the nut/drivers/ directory to see how other vendors use them. The script just maps them to generic numbers, but the values are most useful if mapped to proper types. For instance:

$ git grep UPS.PowerSummary.PresentStatus.ACPresent

drivers/apc-hid.c:  { "BOOL", 0, 0, "UPS.PowerSummary.PresentStatus.ACPresent", NULL, NULL, HU_FLAG_QUICK_POLL, online_info },

drivers/belkin-hid.c:  { "BOOL", 0, 0, "UPS.PowerSummary.PresentStatus.ACPresent", NULL, NULL, HU_FLAG_QUICK_POLL, liebert_online_info },

...

This indicates that the APC models tend to follow the spec, but that we had to add a custom mapping function for Belkin/Liebert since their boolean True value is often 1e-7 instead of 1.

I would recommend checking the values to make sure that they match the labels, but static attributes such as CapacityMode, DesignCapacity, FullChargeCapacity and Rechargeable can get a HU_FLAG_STATIC or HU_FLAG_SEMI_STATIC flag (see drivers/usbhid-ups.h) so that the driver does not waste time polling them. Some values like iManufacturer, iProduct and iSerialNumber are likely duplicates of the values in the USB device descriptor, and do not need to be included.

>>> 
>> Which part of the website?
> As below except with some minor changes, like -a instead of -s

Ah, right - I will add a note to the text that `-s` might not be available yet.