From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: MIME-Version: 1.0 References: <153913023835.32295.13962696655740190941.stgit@magnolia> <153913040858.32295.9474188640729118153.stgit@magnolia> <20181010155055.GS28243@magnolia> In-Reply-To: <20181010155055.GS28243@magnolia> From: Amir Goldstein Date: Wed, 10 Oct 2018 21:28:34 +0300 Message-ID: Subject: Re: [PATCH 14/25] vfs: make remap_file_range functions take and return bytes completed To: "Darrick J. Wong" Cc: Dave Chinner , Eric Sandeen , Linux NFS Mailing List , linux-cifs@vger.kernel.org, overlayfs , linux-xfs , Linux MM , Linux Btrfs , linux-fsdevel , ocfs2-devel@oss.oracle.com Content-Type: text/plain; charset="UTF-8" Sender: owner-linux-mm@kvack.org List-ID: On Wed, Oct 10, 2018 at 6:51 PM Darrick J. Wong wrote: > > On Wed, Oct 10, 2018 at 09:47:00AM +0300, Amir Goldstein wrote: > > On Wed, Oct 10, 2018 at 3:14 AM Darrick J. Wong wrote: > > > > > > From: Darrick J. Wong > > > > > > Change the remap_file_range functions to take a number of bytes to > > > operate upon and return the number of bytes they operated on. This is a > > > requirement for allowing fs implementations to return short clone/dedupe > > > results to the user, which will enable us to obey resource limits in a > > > graceful manner. > > > > > > A subsequent patch will enable copy_file_range to signal to the > > > ->clone_file_range implementation that it can handle a short length, > > > which will be returned in the function's return value. Neither clone > > > ioctl can take advantage of this, alas. > > > > > > Signed-off-by: Darrick J. Wong > > > --- [...] > > Commit message wasn't clear enough on the behavior of copy_file_range() > > before and after the patch IMO. Maybe it would be better to pospone this > > semantic change to the RFR_SHORTEN patch and keep if (cloned == len) > > in this patch? > > There shouldn't be any behavior change here -- all implementations > return a negative error code or the length that was passed in. I'll > clarify that in the commit message. > OK. BTW, you forgot to update Documentation/filesystems/vfs.txt. Also since this series has a potential to break clone/dedup on overlayfs, it would be great if you could run some of the clone/dedupe xfstests with overlay over xfs. For the simple case of running ./check with a local.config file that is not multi section, this just means running ./check -overlay after the first ./check run (-overlay doesn't mkfs the base fs). If this is a problem, let me know once new devel branch is ready and I'll pull it for testing. If I need to pull in extra xfstests, please mention that as well. Thanks, Amir.