From: Jan Kara Subject: Re: [PATCH 6/8] nowait aio: ext4 Date: Mon, 10 Apr 2017 17:13:59 +0200 Message-ID: <20170410151359.GH3224@quack2.suse.cz> References: <20170403185307.6243-1-rgoldwyn@suse.de> <20170403185307.6243-7-rgoldwyn@suse.de> <20170404075853.GB28522@quack2.suse.cz> <20170404084122.GA10252@infradead.org> <20170410074539.GA18250@infradead.org> <20170410123750.GE3224@quack2.suse.cz> <20170410143943.GA2930@infradead.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: Jan Kara , Goldwyn Rodrigues , linux-fsdevel@vger.kernel.org, jack@suse.com, linux-block@vger.kernel.org, linux-btrfs@vger.kernel.org, linux-ext4@vger.kernel.org, linux-xfs@vger.kernel.org, sagi@grimberg.me, avi@scylladb.com, axboe@kernel.dk, linux-api@vger.kernel.org, willy@infradead.org, tom.leiming@gmail.com, Goldwyn Rodrigues To: Christoph Hellwig Return-path: Content-Disposition: inline In-Reply-To: <20170410143943.GA2930@infradead.org> Sender: linux-btrfs-owner@vger.kernel.org List-Id: linux-ext4.vger.kernel.org On Mon 10-04-17 07:39:43, Christoph Hellwig wrote: > On Mon, Apr 10, 2017 at 02:37:50PM +0200, Jan Kara wrote: > > I don't understand here. Do you want that all filesystems support NOWAIT > > direct IO? > > No. Per-file_system_type is way to coarse grained. All feature flag > needs to be per-file_operation at least for cases like ext4 with our > without extents (or journal) XFS v4 vs v5, different NFS versions, etc. Ah, I see your point now. Thanks for patience. I think we could make this work by making generic_file_write/read_iter() refuse NOWAIT IO with EOPNOTSUPP and then only modify those few filesystems that implement their own iter helpers and will not initially support NOWAIT IO. Sounds easy enough. Honza -- Jan Kara SUSE Labs, CR