From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from mail.fst.it ([213.255.39.254] helo=fst.it) by pentafluge.infradead.org with smtp (Exim 4.14 #3 (Red Hat Linux)) id 19SbRj-00018s-PK for ; Wed, 18 Jun 2003 12:49:47 +0100 Message-ID: <027f01c33590$8f177db0$a970a8c0@topo> From: "Luca Contini" To: "David Woodhouse" References: <006801c334a8$3aa7c750$a970a8c0@topo> <1055839872.2233.68.camel@passion.cambridge.redhat.com> Date: Wed, 18 Jun 2003 13:55:48 +0200 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit cc: "linux-mtd-request@lists.infradead.org" Subject: Re: Please help, nftl question List-Id: Linux MTD discussion mailing list List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Thank you David. One more question: I want to put a vfat filesystem on top of nftl (actually i'd like to read from a smartmedia formatted with fat12 by a digital camera). Should I create a dev node (e.g. nftla) using major number 93 (the one corresponding to nftl driver) and format this device befor mounting it? thank you very much again Luca ----- Original Message ----- From: "David Woodhouse" To: "Luca Contini" Cc: "linux-mtd-request@lists.infradead.org" Sent: Tuesday, June 17, 2003 10:51 AM Subject: Re: Please help, nftl question > On Tue, 2003-06-17 at 09:12, Luca Contini wrote: > > Hi all, > > does anybody know wich is the role of the nftl in the nand flash driver > > stack? > > I konw there ere for layers: > > 1. File system layer (e.g. JFFS2 or DOS-FAT) > > 2. MTD > > 3. Nand generic driver > > 4. Hardware specific driver > > > > where can the nftl be located in the stack? > > You are confused about the file system layer. > > See http://www.linux-mtd.infradead.org/tech/mtd-upper-layers.fig > (there's a JPEG for the xfig-challenged at > http://www.linux-mtd.infradead.org/~dwmw2/mtd-upper-layers.jpeg ) > > Atop the MTD layer there is _either_ a real file system such as JFFS2 or > YAFFS, which uses MTD devices directly and presents itself as a file > system to the operating system, or there can be a 'translation layer' > which is a kind of pseudo-filesystem which pretends to be a standard > disk drive. The translation layer could be FTL, NFTL, INFTL or the very > simplistic 'mtdblock'. The translation layer registers a block device > with the Linux block layer, and on top of that you can mount 'normal' > file systems such as cramfs, ext3 and FAT (and all the others). > > Also, you can access the 'translated' block devices through their device > nodes (/dev/ftl*, /dev/nftl*, /dev/mtdblock*), and you can access the > MTD devices directly to issue read/write/erase commands via the > /dev/mtdX character device nodes. > > See http://www.linux-mtd.infradead.org/tech/mtd-upper-layers.fig > (there's a JPEG for the xfig-challenged at > http://www.linux-mtd.infradead.org/~dwmw2/mtd-upper-layers.jpeg ) > > I've omitted the lower layers. The whole point of the MTD code is that > you shouldn't need to consider the devices and the 'users' of those > devices in the same breath. If you have a question about what's _below_ > the 'Linux MTD layer' level, ask it separately. It's actually a little > messier and more fluid. > > > is nftl a filesystem or a low level driver? > > No. :) > > -- > dwmw2 > > >