From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1754858AbaHLTOv (ORCPT ); Tue, 12 Aug 2014 15:14:51 -0400 Received: from zeniv.linux.org.uk ([195.92.253.2]:41470 "EHLO ZenIV.linux.org.uk" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753970AbaHLTOt (ORCPT ); Tue, 12 Aug 2014 15:14:49 -0400 Date: Tue, 12 Aug 2014 20:14:45 +0100 From: Al Viro To: David Herrmann Cc: linux-kernel@vger.kernel.org, Greg Kroah-Hartman , Linus Torvalds , Andrew Morton , Tejun Heo , linux-fsdevel@vger.kernel.org, Theodore Tso , Christoph Hellwig Subject: Re: [PATCH RFC 0/4] VFS revoke() Message-ID: <20140812191444.GW18016@ZenIV.linux.org.uk> References: <1407869648-1449-1-git-send-email-dh.herrmann@gmail.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <1407869648-1449-1-git-send-email-dh.herrmann@gmail.com> User-Agent: Mutt/1.5.21 (2010-09-15) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Tue, Aug 12, 2014 at 08:54:04PM +0200, David Herrmann wrote: > Open questions: > * Obviously, we need to protect all calls into file->f_op->xyz(). This series > provides enter_file() and leave_file() that can be used like: > if (enter_file(file)) > retval = file->f_op->xyz(file, ...); > else > retval = -ENODEV; > Question is, should we do this at the time we actually invoke those > callbacks or should we do it at the syscall-entry time? The former, obviously. The real bitch is around mmap - you don't need that for evdev, but generic solution would better deal with that kind of crap. I'll post a review once I'm done with Eric's patchset; hopefully tonight or tomorrow morning.