From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1751699AbbCVIPM (ORCPT ); Sun, 22 Mar 2015 04:15:12 -0400 Received: from mail-pd0-f170.google.com ([209.85.192.170]:33848 "EHLO mail-pd0-f170.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751481AbbCVIPG (ORCPT ); Sun, 22 Mar 2015 04:15:06 -0400 From: Ming Lei To: Jens Axboe , linux-kernel@vger.kernel.org Cc: Christoph Hellwig , Al Viro , Maxim Patlasov Subject: [PATCH 0/3] block: loop: switch to VFS ITER_BVEC Date: Sun, 22 Mar 2015 16:14:51 +0800 Message-Id: <1427012095-4095-1-git-send-email-ming.lei@canonical.com> X-Mailer: git-send-email 1.7.9.5 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org This patchset uses VFS ITER_BVEC for reading/writing loop backing file, and basically it is a rewriting of Christoph Hellwig's post in [1]. With this change, loop code becomes much cleaner than before, and it is also a simplication for reading/writing backing file inside kernel. Another benifit is that one extra page copy is avoided for READ in case of none_transfer. This patchset passes xfstest test(ext4, -g auto) over loop block devices. The patch 3/3 depends a bit on Christoph's kiocb split patches which have been in -next for a while, so tree dependency(not build depend, just behaviour depend) should be considered if it is to be merged to 4.0. The next step is to add AIO/O_DIRECT support for reading/writing backing file against this patchset if no one objects, so that double cache can be avoided. [1], loop: convert to vfs_bvec_write http://marc.info/?l=linux-fsdevel&m=142159370107007&w=2 [2], kiocb split patchset http://marc.info/?t=142238145800004&r=1&w=2 Thanks, Ming Lei