Return-Path: linux-nfs-owner@vger.kernel.org Received: from zeniv.linux.org.uk ([195.92.253.2]:40896 "EHLO ZenIV.linux.org.uk" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750781AbaLOGQI (ORCPT ); Mon, 15 Dec 2014 01:16:08 -0500 Date: Mon, 15 Dec 2014 06:16:02 +0000 From: Al Viro To: Omar Sandoval Cc: Andrew Morton , Trond Myklebust , Christoph Hellwig , David Sterba , linux-fsdevel@vger.kernel.org, linux-mm@kvack.org, linux-nfs@vger.kernel.org, linux-kernel@vger.kernel.org Subject: Re: [PATCH 3/8] swap: don't add ITER_BVEC flag to direct_IO rw Message-ID: <20141215061601.GT22149@ZenIV.linux.org.uk> References: <5f9e8a7dcdf08bd2dd433f1a42690ab8e67e7915.1418618044.git.osandov@osandov.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii In-Reply-To: <5f9e8a7dcdf08bd2dd433f1a42690ab8e67e7915.1418618044.git.osandov@osandov.com> Sender: linux-nfs-owner@vger.kernel.org List-ID: On Sun, Dec 14, 2014 at 09:26:57PM -0800, Omar Sandoval wrote: > The rw argument to direct_IO has some ill-defined semantics. Some > filesystems (e.g., ext4, FAT) decide whether they're doing a write with > rw == WRITE, but others (e.g., XFS) check rw & WRITE. Let's set a good > example in the swap file code and say ITER_BVEC belongs in > iov_iter->flags but not in rw. This caters to the least common > denominator and avoids a sweeping change of every direct_IO > implementation for now. Frankly, this is bogus. If anything, let's just kill the first argument completely - ->direct_IO() can always pick it from iter->type. As for catering to the least common denominator... To hell with the lowest common denominator. How many instances of ->direct_IO() do we have, anyway? 24 in the mainline (and we don't give a flying fuck for out-of-tree code, as a matter of policy). Moreover, several are of "do nothing" variety. FWIW, 'rw' is a mess. We used to have this: READ: O_DIRECT read WRITE: O_DIRECT write KERNEL_WRITE: swapout These days KERNEL_WRITE got replaced with ITER_BVEC | WRITE. The thing is, we have a bunch of places where we explicitly checked for being _equal_ to WRITE. I.e. the checks that gave a negative on swapouts. I suspect that most of them are wrong and should trigger on all writes, including swapouts, but I really didn't want to dig into that pile of fun back then. That's the main reason why 'rw' argument has survived at all...