From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from dell-paw-3.cambridge.redhat.com ([195.224.55.237] helo=passion.cambridge.redhat.com) by pentafluge.infradead.org with esmtp (Exim 3.22 #1 (Red Hat Linux)) id 16Sn5E-0006bt-00 for ; Mon, 21 Jan 2002 22:38:32 +0000 From: David Woodhouse In-Reply-To: <3C487D9A.D85DDC3D@comdev.cc> References: <3C487D9A.D85DDC3D@comdev.cc> <3C48504B.68F8FE7F@comdev.cc> <20020118010528.A22777@swing.comdev.cc> <28158.1011372920@redhat.com> To: Adam Wozniak Cc: linux-mtd@lists.infradead.org Subject: Re: optimum geometry Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Mon, 21 Jan 2002 22:49:33 +0000 Message-ID: <4896.1011653373@redhat.com> Sender: linux-mtd-admin@lists.infradead.org Errors-To: linux-mtd-admin@lists.infradead.org List-Help: List-Post: List-Subscribe: , List-Id: Linux MTD discussion mailing list List-Unsubscribe: , List-Archive: adam.wozniak@comdev.cc said: > I'm beginning to fear that this is far from optimum. What is an > optimum erase block size? As small as possible. The free space required by JFFS2 is measured in blocks. It could probably be reduced to 2 or 3 blocks from the current 5/6 but will always be a straight multiple of blocksize, I think. > I could drop back and rewrite the MTD driver to address each chip > seperately, but I'm concerned that would be both messy and tough on > performance. By treating them as two pairs, you'll halve the write performance and also halve the amount of wasted space. It's a tradeoff which you have to decide on. -- dwmw2