The following reply was made to PR kern/31884; it has been noted by GNATS.

From: Lonnie Mendez <lmendez19@austin.rr.com>
To: Nick Hudson <skrll@netbsd.org>
Cc: Adam Kropelin <akropel1@rochester.rr.com>, gnats-bugs@netbsd.org
Subject: Re: kern/31884: [PATCH] Add quirk for uhid to ignore certain usb devices (blacklist)
Date: Mon, 31 Oct 2005 13:00:38 -0600

 On Monday 31 October 2005 2:13 am, Nick Hudson wrote:
 > On Sunday 30 October 2005 23:25, Adam Kropelin wrote:
 > > On Sat, 29 Oct 2005 08:10:16, Nick Hudson wrote:
 > > > On Friday 21 October 2005 22:33, akropel1@rochester.rr.com wrote:
 > > >> + { USB_VENDOR_DELORME, USB_PRODUCT_DELORME_EARTHMATE,
 > > >>
 > >  >> +	ANY, { UQ_HID_IGNORE }},
 > > >
 > > > Why do you think this should have a quirk?
 > > >
 > > > It is supported by ucycom(4)
 > >
 > > Perhaps the Delorme quirk is not necessary on NetBSD. It was part of
 > > Lonnie's original patch for FreeBSD so I assumed it was applicable to
 > > NetBSD as well.
 > 
 > I think its because FreeBSD doesn't/didn't have uhidev(4) and support for 
 > optional interrupt output pipes.
 
    Hm.  I'll have to try NetBSD with libemul (emul.berlios.de) sometime soon.
 A while back I wasn't able to get it to work with OpenBSD,NetBSD,FreeBSD
 in regards to communication on the interrupt out pipe with libusb; so I settled
 for making the library read-only on the BSDs.
 
    The patch was meant to block uhid from claiming the earthmate interface.
 You can't have a userspace application communicating with the earthmate
 unless uhid releases it's control of the interface, which is what the patch does.
 
 http://mail-index.netbsd.org/tech-kern/2005/05/12/0001.html
 
    Of course, because of this implementation, the patch indeed conflicts with the
 method of going through uhid for communication on NetBSD.