From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S263201AbTFJPas (ORCPT ); Tue, 10 Jun 2003 11:30:48 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S263199AbTFJPaf (ORCPT ); Tue, 10 Jun 2003 11:30:35 -0400 Received: from ns.virtualhost.dk ([195.184.98.160]:21979 "EHLO virtualhost.dk") by vger.kernel.org with ESMTP id S263250AbTFJP2t (ORCPT ); Tue, 10 Jun 2003 11:28:49 -0400 Date: Tue, 10 Jun 2003 17:42:27 +0200 From: Jens Axboe To: Hugh Dickins Cc: Andrew Morton , linux-kernel@vger.kernel.org Subject: Re: [PATCH] loop 6/9 remove LO_FLAGS_BH_REMAP Message-ID: <20030610154227.GE17164@suse.de> References: Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Tue, Jun 10 2003, Hugh Dickins wrote: > Jonah Sherman pointed out back in February how > LO_FLAGS_BH_REMAP is never actually set, since loop_init_xfer only calls > the init for non-0 encryption type. Fix that or scrap it? Let's scrap > it for now, that path (hacking values in bio instead of copying data) > seems never to have been tested, and adds to the number of paths through > loop: leave that optimization to some other occasion. I agree with scrapping it, it's only really interesting for non-transfer block direct remaps which can be done with dm or similar. -- Jens Axboe