From owner-freebsd-arch@FreeBSD.ORG Mon May 22 19:17:43 2006 Return-Path: X-Original-To: Arch@freebsd.org Delivered-To: freebsd-arch@FreeBSD.ORG Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id BBDED16A772; Mon, 22 May 2006 19:17:43 +0000 (UTC) (envelope-from mistry.7@osu.edu) Received: from mail.united-ware.com (am-productions.biz [69.61.164.22]) by mx1.FreeBSD.org (Postfix) with ESMTP id B435C43D6D; Mon, 22 May 2006 19:17:40 +0000 (GMT) (envelope-from mistry.7@osu.edu) Received: from [192.168.1.100] (mail.united-ware.com [69.61.164.22] (may be forged)) (authenticated bits=0) by mail.united-ware.com (8.13.4/8.13.6) with ESMTP id k4MJIhgq017021 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO); Mon, 22 May 2006 15:18:49 -0400 (EDT) (envelope-from mistry.7@osu.edu) From: Anish Mistry To: freebsd-usb@freebsd.org, Hans Petter Selasky Date: Mon, 22 May 2006 15:17:29 -0400 User-Agent: KMail/1.9.1 References: <20060522144542.T21787@fledge.watson.org> <86y7wum9ow.fsf@xps.des.no> <200605222037.14030.Danovitsch@vitsch.net> In-Reply-To: <200605222037.14030.Danovitsch@vitsch.net> MIME-Version: 1.0 Content-Type: multipart/signed; boundary="nextPart67549360.mrISMqu43M"; protocol="application/pgp-signature"; micalg=pgp-sha1 Content-Transfer-Encoding: 7bit Message-Id: <200605221518.03687.mistry.7@osu.edu> X-Spam-Status: No, score=-8.5 required=5.0 tests=ALL_TRUSTED,BAYES_50, MYFREEBSD2,MYFREEBSD3 autolearn=failed version=3.1.0 X-Spam-Checker-Version: SpamAssassin 3.1.0 (2005-09-13) on mail.united-ware.com X-Virus-Scanned: ClamAV 0.88.2/1475/Mon May 22 01:09:26 2006 on mail.united-ware.com X-Virus-Status: Clean Cc: "Daan Vreeken \[PA4DAN\]" , Dag-Erling =?iso-8859-1?q?Sm=F8rgrav?= , Robert Watson , Arch@freebsd.org Subject: Re: Resend with subject: No more non-MPSAFE network device drivers, please! X-BeenThere: freebsd-arch@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Discussion related to FreeBSD architecture List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 22 May 2006 19:17:44 -0000 --nextPart67549360.mrISMqu43M Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Content-Disposition: inline On Monday 22 May 2006 14:37, Daan Vreeken [PA4DAN] wrote: > On Monday 22 May 2006 16:10, Dag-Erling Sm=F8rgrav wrote: > > Robert Watson writes: > > > Per discussion at the developer summit, a significant rewrite > > > of the USB code is needed to make it MPSAFE. Volunteers > > > welcome. > > > > HP Selasky claims to have a working Giant-free USB stack: > > (added him to the 'cc' list) > > > http://www.turbocat.net/~hselasky/usb4bsd/index.html > > > > No idea how well it works, though. > > I have been running his USB driver on a production server for > months here without a single problem. I use his ugen implementation > to communicate with a custom made USB device. On FreeBSD's USB > driver a stress test would fail after max 40,000 requests. I've got a patch in a PR that fixes several ugen issues. > With his=20 > USB driver I have successfully made over 30,000,000 requests before > stopping the test manually. > > There is a problem with his code though. (At least, there was back > when I last looked at it, correct me if I'm wrong.) He has > completely rewritten the OHCI, UHCI and EHCI parts, but kept the > original (network, storage, etc) drivers. For the original drivers > to work he has added some wrapper functions. Back when I looked > these wrappers didn't work out of the box for the devices I have > tried. > All "old" drivers need to be converter to "new world order" for > them to work properly. > > That's the only problem I have been having with his driver : You > get a better USB stack, but with less working drivers. > I think it would be great to get his code into -current, but some > developers would have to help out to convert the existing drivers. I'd be willing to help convert drivers if we can get the code into the=20 tree (maybe as usb-ng?). Setting up a bounty for this stuff may be=20 worthwhile since having a solid USB stack would be a boon for=20 everyone. I'm not sure if it's possible with the HPS driver, but it would be=20 nice if every USB device could expose ugen endpoints along with their=20 device specific nodes. =2D-=20 Anish Mistry --nextPart67549360.mrISMqu43M Content-Type: application/pgp-signature -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.3 (FreeBSD) iD8DBQBEcg5rxqA5ziudZT0RAkFEAJ9p0HaaD9uecdAg1HZCRcWP5ntcVACgpyG2 jaEyQZPvohSIxl215FYKYy8= =pFvQ -----END PGP SIGNATURE----- --nextPart67549360.mrISMqu43M--