From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id ; Tue, 2 Apr 2002 22:06:51 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id ; Tue, 2 Apr 2002 22:06:33 -0500 Received: from deimos.hpl.hp.com ([192.6.19.190]:58572 "EHLO deimos.hpl.hp.com") by vger.kernel.org with ESMTP id ; Tue, 2 Apr 2002 22:06:22 -0500 Date: Tue, 2 Apr 2002 19:06:21 -0800 To: Greg KH Cc: irda-users@lists.sourceforge.net, Linux kernel mailing list Subject: Re: [PATCH] : ir257_usb_disconnect_atomic-2.diff Message-ID: <20020402190621.A25089@bougret.hpl.hp.com> Reply-To: jt@hpl.hp.com In-Reply-To: <20020402182413.G24912@bougret.hpl.hp.com> <20020403030038.GA6366@kroah.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i Organisation: HP Labs Palo Alto Address: HP Labs, 1U-17, 1501 Page Mill road, Palo Alto, CA 94304, USA. E-mail: jt@hpl.hp.com From: Jean Tourrilhes Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Tue, Apr 02, 2002 at 07:00:38PM -0800, Greg KH wrote: > On Tue, Apr 02, 2002 at 06:24:13PM -0800, Jean Tourrilhes wrote: > > @@ -1519,33 +1544,47 @@ static void *irda_usb_probe(struct usb_d > > /* > > * The current irda-usb device is removed, the USB layer tell us > > * to shut it down... > > + * One of the constraints is that when we exit this function, > > + * we cannot use the usb_device no more. Gone. Destroyed. kfree(). > > + * Most other subsystem allow you to destroy the instance at a time > > + * when it's convenient to you, to postpone it to a later date, but > > + * not the USB subsystem. > > + * So, we must make bloody sure that everything gets deactivated. > > + * Jean II > > That's one of the next things I'm going to be working on fixing :) By the time you will "fix" that, all the USB driver will be fixed to workaround this issue. Actually, it force to be a bit more careful about the disconnect, and avoid zombies instances all over the place, so is not such a bad thing after all. > The patch looks good. Thanks for setting the proper GFP_* flag. That was Martin... Don't need to thank, because the impact is limited to the IrDA driver... > greg k-h Have fun... Jean