From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from pop.scorch.co.nz ([203.167.210.162] helo=firstline.co.nz) by canuck.infradead.org with smtp (Exim 4.63 #1 (Red Hat Linux)) id 1HJPO1-0004ah-Ii for linux-mtd@lists.infradead.org; Tue, 20 Feb 2007 02:26:16 -0500 From: Charles Manning To: linux-mtd@lists.infradead.org Subject: Re: Question about MTD Erase mechanism. Date: Tue, 20 Feb 2007 20:29:33 +1300 References: <9b52d64c0702192241y2c778a54r17cea174eb8781d8@mail.gmail.com> In-Reply-To: <9b52d64c0702192241y2c778a54r17cea174eb8781d8@mail.gmail.com> MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Content-Disposition: inline Message-Id: <200702202029.33870.manningc2@actrix.gen.nz> Cc: Vinit Agnihotri List-Id: Linux MTD discussion mailing list List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , On Tuesday 20 February 2007 19:41, Vinit Agnihotri wrote: > I have 1 question about MTD Erase mechanism. > > Say we have a MTD device with 'n' no. of blocks. Each block is of size > 100Bytes. Once erased we can write 25bytes i.e. min_io_size is > 25bytes. So once a block is erased I can perform 4 writes without > erasing it again only if those 4 writes are not overlapping. But if > any of those write is overlapping i.e. if re-write occurs then entire > block have to be erased. That's a pretty contrived example, and depends on the type of flash etc. > > So my question is who handles these types of conditions. I mean does > hardware handles such conditions or software have to handle such > situations? I mean who handles re-writing & erasures due to that??? Software of course. Hardware is pretty dumb stuff and cannot track and enforce algorithmic rules like that. > > > Thanks & Regards > Vinit. > > ______________________________________________________ > Linux MTD discussion mailing list > http://lists.infradead.org/mailman/listinfo/linux-mtd/