From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from userp2130.oracle.com ([156.151.31.86]:60086 "EHLO userp2130.oracle.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1726545AbeJPA2o (ORCPT ); Mon, 15 Oct 2018 20:28:44 -0400 Date: Mon, 15 Oct 2018 09:42:20 -0700 From: "Darrick J. Wong" To: Christoph Hellwig Cc: david@fromorbit.com, sandeen@redhat.com, linux-nfs@vger.kernel.org, linux-cifs@vger.kernel.org, Amir Goldstein , linux-unionfs@vger.kernel.org, linux-xfs@vger.kernel.org, linux-mm@kvack.org, linux-btrfs@vger.kernel.org, linux-fsdevel@vger.kernel.org, ocfs2-devel@oss.oracle.com Subject: Re: [PATCH 07/25] vfs: combine the clone and dedupe into a single remap_file_range Message-ID: <20181015164220.GL28243@magnolia> References: <153938912912.8361.13446310416406388958.stgit@magnolia> <153938919123.8361.13059492965161549195.stgit@magnolia> <20181014171927.GD30673@infradead.org> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii In-Reply-To: <20181014171927.GD30673@infradead.org> Sender: linux-nfs-owner@vger.kernel.org List-ID: On Sun, Oct 14, 2018 at 10:19:27AM -0700, Christoph Hellwig wrote: > > unsigned (*mmap_capabilities)(struct file *); > > #endif > > ssize_t (*copy_file_range)(struct file *, loff_t, struct file *, loff_t, size_t, unsigned int); > > - int (*clone_file_range)(struct file *, loff_t, struct file *, loff_t, u64); > > - int (*dedupe_file_range)(struct file *, loff_t, struct file *, loff_t, u64); > > + int (*remap_file_range)(struct file *file_in, loff_t pos_in, > > + struct file *file_out, loff_t pos_out, > > + u64 len, unsigned int remap_flags); > > None of the other methods in this file name their parameters. While > I generally don't like people leaving them out, in the end consistency > is even more important. > > > +int btrfs_remap_file_range(struct file *src_file, loff_t off, > > + struct file *dst_file, loff_t destoff, u64 len, > > + unsigned int remap_flags) > > { > > + if (!remap_check_flags(remap_flags, RFR_SAME_DATA)) > > + return -EINVAL; > > + > > + if (remap_flags & RFR_SAME_DATA) { > > So at least for btrfs there seems to be no shared code at all below > the function calls. This kinda speaks against the argument that > they fundamentally are the same.. They /do/ share/ code -- eventually both btrfs_extent_same and btrfs_clone_files call btrfs_clone. xfs and ocfs2 call the same paths internally too; it's only the vfs helpers that have the extra page cache comparisons if it's a dedup operation. > > +/* > > + * These flags control the behavior of the remap_file_range function pointer. > > + * > > + * RFR_SAME_DATA: only remap if contents identical (i.e. deduplicate) > > + */ > > +#define RFR_SAME_DATA (1 << 0) > > + > > +#define RFR_VALID_FLAGS (RFR_SAME_DATA) > > RFR? Why not REMAP_FILE_* Also why not the well understood > REMAP_FILE_DEDUP instead of the odd SAME_DATA? Sure. I had begin to dislike typing RFR anyway. > > + > > +/* > > + * Filesystem remapping implementations should call this helper on their > > + * remap flags to filter out flags that the implementation doesn't support. > > + * > > + * Returns true if the flags are ok, false otherwise. > > + */ > > +static inline bool remap_check_flags(unsigned int remap_flags, > > + unsigned int supported_flags) > > +{ > > + return (remap_flags & ~(supported_flags & RFR_VALID_FLAGS)) == 0; > > +} > > Any reason to even bother with a helper for this? ->fallocate > seems to be doing fine without the helper, and the resulting code > seems a lot easier to understand to me. (Will respond to these at the current end of the flags thread.) > > @@ -1759,10 +1779,9 @@ struct file_operations { > > #endif > > ssize_t (*copy_file_range)(struct file *, loff_t, struct file *, > > loff_t, size_t, unsigned int); > > - int (*clone_file_range)(struct file *, loff_t, struct file *, loff_t, > > - u64); > > - int (*dedupe_file_range)(struct file *, loff_t, struct file *, loff_t, > > - u64); > > + int (*remap_file_range)(struct file *file_in, loff_t pos_in, > > + struct file *file_out, loff_t pos_out, > > + u64 len, unsigned int remap_flags); > > Same comment here. Didn't we have some nice doc tools to avoid this > duplication? :) We do, but vfs.txt hasn't been ported to any of that. --D