From mboxrd@z Thu Jan 1 00:00:00 1970 From: Pekka J Enberg Subject: Re: Re: [RFC/PATCH] revoke/frevoke system calls V2 Date: Thu, 27 Jul 2006 19:01:55 +0300 (EEST) Message-ID: References: <1154012822.13509.52.camel@localhost.localdomain> <84144f020607270833v4c981d00w8e3e643406aea7a@mail.gmail.com> <1154016589.13509.56.camel@localhost.localdomain> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Cc: linux-kernel@vger.kernel.org, linux-fsdevel@vger.kernel.org, akpm@osdl.org, viro@zeniv.linux.org.uk, tytso@mit.edu, tigran@veritas.com Return-path: Received: from courier.cs.helsinki.fi ([128.214.9.1]:6583 "EHLO mail.cs.helsinki.fi") by vger.kernel.org with ESMTP id S1750818AbWG0QB4 (ORCPT ); Thu, 27 Jul 2006 12:01:56 -0400 To: Alan Cox In-Reply-To: <1154016589.13509.56.camel@localhost.localdomain> Sender: linux-fsdevel-owner@vger.kernel.org List-Id: linux-fsdevel.vger.kernel.org Ar Iau, 2006-07-27 am 18:33 +0300, ysgrifennodd Pekka Enberg: > > Don't device drivers already do that for f_ops->flush (filp_close) and On Thu, 27 Jul 2006, Alan Cox wrote: > ->flush is called when each closing occurs. Yes revoke calls it too, but is that sufficient, or do we need ->revoke? Ar Iau, 2006-07-27 am 18:33 +0300, ysgrifennodd Pekka Enberg: > > vm_ops->close (munmap)? What revoke and frevoke do is basically > > unmap/fsync/close on all the open file descriptors. On Thu, 27 Jul 2006, Alan Cox wrote: > What happens if an app is already blocked on a read when you do a > revoke ? The nasty case answer could be "it completes later on and > returns the users captured password" Ouch. You are right. I need to stick that invalidate_inode_pages2 back in there. The do_fsync call takes care of writes only, obviously. Thanks! Pekka