From mboxrd@z Thu Jan 1 00:00:00 1970 From: Christoph Hellwig Subject: Re: [PATCH v8 1/4] fat: add fat_fallocate operation Date: Wed, 10 Dec 2014 02:45:24 -0800 Message-ID: <20141210104524.GA865@infradead.org> References: <001601d01380$2009d000$601d7000$@samsung.com> <003d01d01465$f719f3f0$e54ddbd0$@samsung.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: 'Luk?? Czerner' , 'OGAWA Hirofumi' , 'Andrew Morton' , linux-fsdevel@vger.kernel.org To: Namjae Jeon Return-path: Received: from bombadil.infradead.org ([198.137.202.9]:53800 "EHLO bombadil.infradead.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751607AbaLJKp3 (ORCPT ); Wed, 10 Dec 2014 05:45:29 -0500 Content-Disposition: inline In-Reply-To: <003d01d01465$f719f3f0$e54ddbd0$@samsung.com> Sender: linux-fsdevel-owner@vger.kernel.org List-ID: On Wed, Dec 10, 2014 at 07:42:16PM +0900, Namjae Jeon wrote: > User have to call fallocate again if opening file after close(). > And fallocate() will skip allocation if fallocated blocks is avaiable. > Instead, It gurrantee to be available fallocated blocks between open > and close(). And I updated such limitation of fat fallocate in > Documentation. Documentation doesn't help, this breaks application exposed semantics. If you think there is a use case for non-persistant preallocations please propose a new fallocate flag for it, or given the lack of use case probably just add a fat specific ioctl instead.