Current-Users archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: QUick question about USB interfaces

On 10/31/2012 4:38 PM, Paul Goyette wrote:
On Wed, 31 Oct 2012, Dave Burgess wrote:

See the current-users mailing list for my patches to solve this problem.

I can't test them just this second, I'm in the middle of a SUP update....

Yes, I saw that.

I hope it solves your problem.

It's curious, though, how apcupsd requires to connect to a ugen (and needs you to prevent the device from being configured as uhid), while yours (apparently/suspectedly) requires just the opposite behavior!

Well, that was a cluster frack.

I spent most of the weekend fixing the problems I caused trying to install the -current kernel. I'm back on 6.0.0-patch from's daily build sets from the end of October.

Here's what I'm finding:

1) The ugen0 device should iterate through the sub-devices under a generic ugen0 like it does with apcupsctl. The ups-nut package apparently doesn't do that.

2) The uhid0 device is not sufficiently 'keyboardy' enough for the hid device interface to pick up the UPS.

The patches to the usbdevs file that I put up last week should be sufficient for correctly identifying the device, but the interaction with the usb_quirks still seems to demand some attention. The fact that the device connects to FreeBSD as a HID device would imply that we should do the same, but I can't really get that to sync up.

I really would like to get this working with NetBSD, but having it work on CentOS is a work-around I can probably deal with for now. Knowing the UPS operation parameters is the goal. I'll probably get back to it later in the month, but I spent most of the week-end getting back to Wednesday, so I'm a little gun-shy about making any major changes to the OS just this minute :-)

Dave Burgess
Cynjut Consulting Services, LLC

Home | Main Index | Thread Index | Old Index