From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jamie Lokier Subject: Re: [PATCH 1/2] fs: add SEEK_HOLE and SEEK_DATA flags Date: Mon, 25 Apr 2011 15:15:40 +0100 Message-ID: <20110425141540.GB20461@shareable.org> References: <1303414954-3315-1-git-send-email-josef@redhat.com> <20110422045054.GB17795@infradead.org> <20110422112852.GB1627@x4.trippels.de> <4DB16B72.1050702@redhat.com> <4DB1AC9D.3010706@oracle.com> <20110424174902.GA20461@shareable.org> <4DB56B11.5090505@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: Sunil Mushran , Markus Trippelsdorf , Christoph Hellwig , Josef Bacik , linux-kernel@vger.kernel.org, linux-btrfs@vger.kernel.org, linux-fsdevel@vger.kernel.org To: Eric Blake Return-path: Content-Disposition: inline In-Reply-To: <4DB56B11.5090505@redhat.com> Sender: linux-kernel-owner@vger.kernel.org List-Id: linux-fsdevel.vger.kernel.org Eric Blake wrote: > On 04/24/2011 11:49 AM, Jamie Lokier wrote: > >> My problem with FIND_* is that we are messing with the well understood > >> semantics of lseek(). > > > > fcntl() looks a better fit for FIND_HOLE/DATA anyway. > > With fcntl(), it would have to be something like: > > off_t offset = start; > if (fcntl (fd, F_FIND_HOLE, &offset) != 0) > ; // find failed > // offset is now set to the next location after start Yes that, or a pointer-to-struct in the style of other fcntl() operations. A struct offers more flexibiliy such as a limit on search distance (may be useful on filesystems where searching reads a lot of metadata and you don't really want to scan all of a large file), and whether to include unwritten prealloc space or written-known-zeros space. I thought of fcntl() because historically it's often how you get quirky information about files and how to read them, on many OSes. -- Jamie