Re: [PATCH resend2] Input: usbtouchscreen - initialize eGalax devices

From: Jiri Kosina <jkosina@suse.cz>
To: Forest Bond <forest@alittletooquiet.net>
Cc: Dmitry Torokhov <dmitry.torokhov@gmail.com>,
	Daniel Ritz <daniel.ritz@gmx.ch>,
	Alan Stern <stern@rowland.harvard.edu>,
	Sergei Shtylyov <sshtylyov@mvista.com>,
	linux-input@vger.kernel.org, linux-usb@vger.kernel.org
Subject: Re: [PATCH resend2] Input: usbtouchscreen - initialize eGalax devices
Date: Mon, 5 Nov 2012 15:19:34 +0100 (CET)	[thread overview]
Message-ID: <alpine.LNX.2.00.1211051516530.24253@pobox.suse.cz> (raw)
In-Reply-To: <20121102195130.GJ12501@alittletooquiet.net>

On Fri, 2 Nov 2012, Forest Bond wrote:

> We have some eGalax devices with class HID and protocol None that both 
> usbhid and usbtouchscreen will bind to, but we only want them bound to 
> usbtouchscreen. Some do in fact work with usbhid, but not all of them 
> do.  OTOH they all work with usbtouchscreen as of commit 
> 037a833ed05a86d01ea27a2c32043b86c549be1b.
> 
> We want to blacklist these devices in usbhid to avoid binding to it if 
> it is loaded first.  But usbhid should continue to handle eGalax devices 
> with class HID and protocol other than None (e.g. Mouse).  They all have 
> the same vendor and product IDs, so we need to be able to blacklist on 
> (VID, PID, protocol) instead of just (VID, PID).

I see, thanks for the explanation.

As this is the first time we are having this problem, I'd just propose to 
put just another switch into hid_ignore() for USB_VENDOR_ID_DWAV, and 
looking at hdev->type there to see whether we should ignore the device or 
not.

If it turns, over time, to be a more general problem for multiple devices, 
we'll just introduce the more general blacklist matching then.

How does that sound?

-- 
Jiri Kosina
SUSE Labs