From owner-freebsd-usb@FreeBSD.ORG Wed Jan 9 19:35:39 2008 Return-Path: Delivered-To: usb@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 78C1416A41A for ; Wed, 9 Jan 2008 19:35:39 +0000 (UTC) (envelope-from mi+mill@aldan.algebra.com) Received: from mail5.sea5.speakeasy.net (mail5.sea5.speakeasy.net [69.17.117.7]) by mx1.freebsd.org (Postfix) with ESMTP id 58CF113C458 for ; Wed, 9 Jan 2008 19:35:39 +0000 (UTC) (envelope-from mi+mill@aldan.algebra.com) Received: (qmail 29473 invoked from network); 9 Jan 2008 19:08:59 -0000 Received: from aldan.algebra.com (HELO aldan-mlp) ([216.254.65.224]) (envelope-sender ) by mail5.sea5.speakeasy.net (qmail-ldap-1.03) with SMTP for ; 9 Jan 2008 19:08:59 -0000 From: Mikhail Teterin To: Hans Petter Selasky Date: Wed, 9 Jan 2008 14:08:55 -0500 User-Agent: KMail/1.7.1 References: <200801090114.56195@aldan> <200801091015.17730@aldan> <200801091934.54053.hselasky@c2i.net> In-Reply-To: <200801091934.54053.hselasky@c2i.net> Organization: Virtual Estates, Inc. MIME-Version: 1.0 Content-Type: text/plain; charset="koi8-u" Content-Transfer-Encoding: quoted-printable Content-Disposition: inline Message-Id: <200801091408.57437.mi+mill@aldan.algebra.com> Cc: johnjen@reynoldsnet.org, R.Tobin@ed.ac.uk, usb@freebsd.org, Mikhail Teterin Subject: Re: BlackBerry (Re: using libusb) X-BeenThere: freebsd-usb@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: FreeBSD support for USB List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 09 Jan 2008 19:35:39 -0000 =D3=C5=D2=C5=C4=C1 09 =D3=A6=DE=C5=CE=D8 2008 01:34 =D0=CF, Hans Petter Sel= asky =F7=C9 =CE=C1=D0=C9=D3=C1=CC=C9: > My plan is to extend /dev/usbX so that it becomes a so-called clonable > device. When you open up "/dev/usb0.2.3" for example, you open up the > device having index 2 on USB bus 0 and endpoint 3. This will need some > modifications in libusb. Ugen will still be there, but I plan to move the > functionality over to "usb.c". Accessing endpoints will then work > regardless of what drivers are hooked on. I must admit, that /dev/umass2 looks better than /dev/usb0.2, but I don't=20 really care :-) Other than that you will, probably, have to take over=20 maintainership of the libusb port... Maybe, even to the point of bringing i= ts=20 relevant bits and pieces into the src-tree. There is simply too much stuff being written for Linux that uses the libusb= =20 API to ignore it... It is bad enough, that we can't have video4linux=20 support :( > Some open problems that needs to be resolved: > > Should we allow parallell access to USB interfaces? Not sure, what you mean. Parallel access between multiple interfaces to the= =20 same device (umass and ugen access to the same blackberry), or parallel=20 access between two different user-space processes? > And what about rights management? devfs.conf? -mi