From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from relay.sgi.com (relay1.corp.sgi.com [137.38.102.111]) by oss.sgi.com (Postfix) with ESMTP id D10237CA0 for ; Tue, 9 Aug 2016 12:42:01 -0500 (CDT) Received: from cuda.sgi.com (cuda3.sgi.com [192.48.176.15]) by relay1.corp.sgi.com (Postfix) with ESMTP id A19098F8049 for ; Tue, 9 Aug 2016 10:41:58 -0700 (PDT) Received: from mout02.posteo.de (mout02.posteo.de [185.67.36.66]) by cuda.sgi.com with ESMTP id vHuc0doi0fMz8QdR (version=TLSv1.2 cipher=ECDHE-RSA-AES256-GCM-SHA384 bits=256 verify=NO) for ; Tue, 09 Aug 2016 10:41:55 -0700 (PDT) Received: from submission (posteo.de [89.146.220.130]) by mout02.posteo.de (Postfix) with ESMTPS id C922B20B69 for ; Tue, 9 Aug 2016 19:41:53 +0200 (CEST) Date: Tue, 9 Aug 2016 19:41:05 +0200 From: Felix Janda Subject: Re: [PATCH xfsprogs 01/14] configure: use AC_SYS_LARGEFILE Message-ID: <20160809174105.GA5907@nyan> References: <20160809073611.GC25647@infradead.org> MIME-Version: 1.0 Content-Disposition: inline In-Reply-To: <20160809073611.GC25647@infradead.org> List-Id: XFS Filesystem from SGI List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Errors-To: xfs-bounces@oss.sgi.com Sender: xfs-bounces@oss.sgi.com To: Christoph Hellwig Cc: xfs@oss.sgi.com Thanks for reviewing the patch series! > Does this also error out for libraries that don't support large > off_t at all? I think that would be helpful to add if not there yet. I do not quite understand. Do you refer to libraries using libxfs or the external libraries used by xfsprogs? For the latter, none of them exports interfaces using off_t. libblkid has its own blkid_loff_t, which is defined as int64_t. For the former, patch 13 forces any user of libxfs to enable transparent LFS, by for example adding AC_SYS_LARGEFILE. The approach of libblkid is the same as what I was suggesting in a previous patch, but maybe it is good to break applications using libxfs and not transparent LFS. For example this seems to be the case for ceph. It has not enabled transparent LFS but mixes off_t and off64_t. So it is likely that it has some LFS related runtime bugs on 32bit systems. If the xfs header included the off_t size check, building ceph on 32bit systems would lead to a less subtle compile failure. Felix _______________________________________________ xfs mailing list xfs@oss.sgi.com http://oss.sgi.com/mailman/listinfo/xfs