From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from mail2.shareable.org ([80.68.89.115]) by bombadil.infradead.org with esmtps (Exim 4.69 #1 (Red Hat Linux)) id 1MlLn0-0002XQ-Ss for linux-mtd@lists.infradead.org; Wed, 09 Sep 2009 11:56:47 +0000 Date: Wed, 9 Sep 2009 12:56:33 +0100 From: Jamie Lokier To: Harald Welte Subject: Re: [RFC] extending nand_ecclayout.eccpos once again Message-ID: <20090909115633.GA1732@shareable.org> References: <20090902093045.GB7377@prithivi.gnumonks.org> <1252390415.5060.38.camel@localhost> <20090908152813.GC5579@shareable.org> <20090909111318.GD4771@prithivi.gnumonks.org> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20090909111318.GD4771@prithivi.gnumonks.org> Cc: Kukjin Kim , Artem Bityutskiy , Jin-Sung Yang , Ilho Lee , Marc Zyngier , linux-mtd@lists.infradead.org, Nicolas Pitre , David Woodhouse List-Id: Linux MTD discussion mailing list List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Harald Welte wrote: > Creating raw images? I'm also not sure how often that is being > done. Why would you precompute an image with ECC and then write > that to NAND rather than having the NAND layer generate the ECC > while writing the image (like writing any other data)? One reason is to be resistant to errors (such as RAM bit-flips) which occur in flight from the time the data was read or generated (perhaps a long time ago or far away), until it is written. Such errors do happen, but it's a bit premature to start coding it in MTD, as it would be far more valuable to have it in ordinary filesystems and copying tools, and that looks quite a long way off :-) -- Jamie