From mboxrd@z Thu Jan 1 00:00:00 1970 From: "H. Peter Anvin" Subject: Re: Proposal: make RAID6 code optional Date: Wed, 22 Apr 2009 11:39:19 -0700 Message-ID: <49EF6457.90505@zytor.com> References: <200904180946.27722.prakash@punnoor.de> <49E98AD2.8060601@msgid.tls.msk.ru> <200904181117.03418.prakash@punnoor.de> <20090418145850.GD28512@mea-ext.zmailer.org> <49EDD11E.2030309@tmr.com> <49EE00F9.6090000@zytor.com> <20090422180051.GD13280@skl-net.de> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <20090422180051.GD13280@skl-net.de> Sender: linux-raid-owner@vger.kernel.org To: Andre Noll Cc: Bill Davidsen , Matti Aarnio , Jesper Juhl , Prakash Punnoor , Michael Tokarev , linux-kernel@vger.kernel.org, linux-raid@vger.kernel.org, neilb@suse.de List-Id: linux-raid.ids Andre Noll wrote: > On 10:23, H. Peter Anvin wrote: > >> We could use vmalloc() and generate the tables at initialization time. >> However, having a separate module which exports the raid6 declaration >> and uses the raid5 module as a subroutine library seems easier. > > Really? Easier than keeping only two 256-byte arrays for exp() and > log() and use these at runtime to populate the (dynamically allocated) > 64K GF multiplication table? That seems to be really simple and would > still shave off 64K of kernel memory for raid5-only users. > Yes, I believe it would be easier than having dynamically allocated arrays. Dynamically generated arrays using static memory allocations (bss) is one thing, but that would only reduce size of the module on disk, which I don't think anyone considers a problem. -hpa