[Nut-upsuser] MGE Nova 1100 using newhidups on netBSD 3.0
NevilTh at nat.com.au
Thu Mar 2 07:44:19 UTC 2006
Well I have built & installed libusb 0.1.11.
No difference in behaviour.
Ran the test suite as suggested and noted still getting the input/output
error reported on many (but not all) of the usb_control_msg calls.
In looking at this ,always appeared to be on calls with large buffers ?
On looking through the doco's found that there was a setting in the request
structure passed to ioctl to allow short xfer's. This was not used so
requests were a larger buffer was provided than the data size returned was
generating the input/output errors (I think).
Anyway changed the code in bsd.c (libusb) function usb_control_msg
>From req.ucr_flags = 0;
To req.ucr_flags = USBD_SHORT_XFER_OK;
And now don't get the input/output errors!.
This hasn't solved the fact that I still get stale data until such time as I
unplug/replug the USB cable but at least after I do that upsc is reporting
UPS related variables which it wasn't before.
bash-3.1# upsc MGENova at localhost
driver.version.data: MGE HID 0.8
ups.mfr: MGE UPS SYSTEMS
ups.model: NOVA 1100 AVR
ups.status: OL CHRG
Still hunting for the unplug/replug problem....
From: Arnaud Quette [mailto:aquette.dev at gmail.com]
Sent: Wednesday, 1 March 2006 3:16 AM
To: Nevil Thatcher
Cc: nut-upsuser at lists.alioth.debian.org
Subject: Re: [Nut-upsuser] MGE Nova 1100 using newhidups on netBSD 3.0
2006/2/28, Nevil Thatcher <NevilTh at nat.com.au>:
> Looking at the libusb project they are recommending to use libusb 0.1.11
> I will download an build that and see if that helps.
a good thing would be to also try the libusb test tools (under the test/
More information about the Nut-upsuser