From: Darren Hart Subject: Re: [e2fsprogs] initdir: Writing inode after the initial write? Date: Tue, 04 Dec 2012 09:46:06 -0800 Message-ID: <50BE36DE.6030300@infradead.org> References: <50B967E2.7090703@infradead.org> <92FEB3B3-D4EA-4E84-83F2-F9946D7BCE3B@dilger.ca> <50B990CB.3080607@infradead.org> <85A86E8F-EEB9-495C-AB10-EF3C871EE2B9@dilger.ca> <50BD017F.1070400@infradead.org> <20121204152231.GG29083@thunk.org> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Cc: Andreas Dilger , linux-ext4 To: Theodore Ts'o Return-path: Received: from mga09.intel.com ([134.134.136.24]:26614 "EHLO mga09.intel.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753101Ab2LDRqJ (ORCPT ); Tue, 4 Dec 2012 12:46:09 -0500 In-Reply-To: <20121204152231.GG29083@thunk.org> Sender: linux-ext4-owner@vger.kernel.org List-ID: On 12/04/2012 07:22 AM, Theodore Ts'o wrote: > On Mon, Dec 03, 2012 at 11:46:07AM -0800, Darren Hart wrote: >>> Maybe Ted can confirm whether that is true or not. At least I recall >>> that the block allocator inside libext2fs was horrible, and creating >>> large files was problematic. >> >> Ted, can you confirm? > > The block allocator inside libext2fs is primitive; it will find the > first free block and use it. It should be OK for populating large > flash devices for file system images stored on flash devices (where > seeks don't matter so block group placement isn't a big deal), and > especially for fixed root file system images which are mounted > read-only and which tend to be updated only once in a while (i.e., in > the cases of Android system updates), and so you don't really care > about aligning file writes to eMMC erase blocks. > > It could certainly be made better, and for people who were trying to > use libext2fs with FUSE targetting hard drives, there are ample > opportunities for improvements..... > I think what I'm reading here is that if you care about having a filesystem that makes hardware specific optimizations, you're better off mounting the device and copying the filesystem over. In that case, plan on needing root access. > Creating large files shouldn't be a problem (unless what you mean is > ext4 huge files ala the huge file feature where the number of 512 > blocks exceeds 2**32, in which case you should probably test that case > if you care about it), and it certainly will create extents-based > files. Great, sounds like this approach is still viable. Thanks Ted! -- Darren Hart Intel Open Source Technology Center Yocto Project - Technical Lead - Linux Kernel